article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the goal of high - energy neutrino astronomy is to provide a new view of the universe by detecting the messengers emitted from its most violent regions .
the emission of high - energy neutrinos necessarily implies the presence of highly relativistic baryons at the acceleration sites and consequently provides incontrovertible evidence for the acceleration of charged cosmic rays .
the observable neutrino flux is expected to be generated mainly through charged pion production in collisions of high - energy protons from the cosmic ray accelerators with the ambient gas or with radiation fields @xcite .
these neutrinos will point back to even very distant sources , as they are neither absorbed nor deflected , a property that makes these particles unique astronomical messengers @xcite .
high - energy neutrinos can be of galactic and extragalactic origin .
supernova remnants and micro - quasars are examples for candidate sources in our galaxy , while gamma - ray bursts and active galactic nuclei represent promising potential extragalactic sources @xcite . due to the extremely low cross section of neutrino interactions ,
neutrino detectors need to instrument very large volumes and should be built in a low background environment .
the current neutrino telescopes exploit the idea , proposed by markov @xcite , of instrumenting a large volume of water or ice , in order to detect the charged leptons ( in particular muons ) emerging from charged - current neutrino - nucleon interactions . for a list of recent reviews
see @xcite .
the antares detector @xcite ( see fig . [ fig : schematic ] for a schematic view ) is located at a depth of @xmath0 in the mediterranean sea ( @xmath1 , @xmath2 ) , @xmath3 from the french city of toulon .
it is equipped with 885 optical sensors arranged on 12 flexible lines .
each line comprises up to 25 detection storeys , each equipped with three downward - looking 10-inch photomultipliers ( pmts ) , oriented at 45@xmath4 relative to the vertical .
each pmt is installed in an optical module ( om ) that consists of a 17-inch glass sphere in which the optical connection between the pmt and the glass is assured by an optical gel .
each line is roughly @xmath5 long and is held tight by a buoy at its top .
the spacing between storeys is @xmath6 .
the distance between adjacent lines is of the order of @xmath7 .
antares contains in addition a line ( il07 ) with oceanographic sensors dedicated to the measurement of environmental parameters .
antares therefore represents an important multidisciplinary deep - sea research infrastructure delivering unique data to marine biologists and oceanographers .
the construction of the antares detector took place in several sea campaigns starting in the year 2006 and has been completed in may 2008 with the deployment and connection of the last 2 lines . due to this long construction phase and the modularity of the detector ,
the commissioning phase comprised several detector configurations with different numbers of active lines .
the neutrino detection relies on the emission of cherenkov photons by high - energy muons originating from charged - current neutrino - nucleon interactions in or around the instrumented volume . from the pmt positions and the relative arrival times of the cherenkov photons at the pmts , and making use of the characteristic emission angle of cherenkov radiation , the trajectory of the muon can be reconstructed .
the direction of incident neutrinos can be inferred with an energy dependent precision that is expected to be better than 0.3 degrees for @xmath8 .
as the detector lines move with the deep - sea current , the position of the oms need to be monitored to ensure this excellent angular resolution .
the position of the oms is determined every 2 minutes by means of an acoustic triangulation system , while the orientation of each storey is measured with a compass and a tiltmeter .
the timing calibration @xcite , which is also crucial for the angular resolution and very stable in time , is monitored regularly in - situ with dedicated pulsed light sources distributed along the lines .
the main goal of a neutrino telescope is to detect high - energy neutrinos from extraterrestrial sources .
however , the signal observed by antares is dominated by muons that are generated in cosmic ray interactions in the atmosphere above the detector and which have sufficient energy to reach the detector at its average installation depth of @xmath9 below sea surface . the muon flux measured at
the antares site is an important test beam to study detector systematics and to validate the reconstruction algorithms employed .
two different studies of the depth - intensity relation for muons have been carried out . in the first ,
the attenuation of the muon flux as a function of depth is observed as a reduction in the rate of photon coincidences between adjacent storeys along the detection lines @xcite .
this method has the advantage that it does not rely on any track reconstruction method and therefore allows for testing directly the response of the detector .
the second method is based on a standard tracking algorithm that allows for reconstructing the ( average ) zenith angle of the incident muon ( bundle ) which is then used to compute the track length from the sea surface to the detector .
this track length is usually called `` equivalent slant depth '' . taking into account the known angular distribution of the incident muons ,
a depth - intensity relation can be extracted @xcite .
the results are in good agreement with previous measurements as can be seen from fig .
[ fig : muonflux ] .
the rather large error band is mainly due to the systematic uncertainty on the determination of the absorption length of light in water and of the angular acceptance of the oms at large angles , which becomes important for muons traversing the detector from above due to the downward - pointing setup of the oms .
the prediction of a neutrino flux from extraterrestrial sources is a direct consequence of the observation of high - energy particle and gamma radiation impinging on the earth s atmosphere @xcite .
while both electrons and charged hadrons can be present at cosmic acceleration sites , only in the case of hadron acceleration will the energy escaping from the source be distributed between the cosmic ray component , gamma rays and neutrinos .
a spatially unresolved , hence diffuse , flux of such high - energy neutrinos resulting from different cosmic sources has been predicted by various authors .
there are two relevant upper bound estimates : waxman and bahcall ( w&b ) @xcite use as a constraint the cosmic ray flux measurements at energies @xmath10 ; mannheim , protheroe and rachen ( mpr ) @xcite consider the diffuse @xmath11-ray flux in addition . for sources that are assumed to be transparent to neutrons ,
the resulting upper limits are shown in fig .
[ fig : diffuse ] .
the search method for the diffuse neutrino flux exploits the fact that the atmospheric neutrino flux , which constitutes the main background in this search , has been measured to exhibit a @xmath12 dependence at high energies . the predicted diffuse flux of cosmic neutrinos ,
however , is expected to follow the much harder energy spectrum of its parent hadron distribution , i.e. a spectrum with a spectral index close to -2 .
this prediction results from the fact that the only known mechanism that can accelerate cosmic ray particles up to the highest observed energies is the so called fermi acceleration expected to occur in hydrodynamical shock fronts . to separate atmospheric and diffuse cosmic neutrino fluxes , a robust energy estimator for high - energy muon
neutrino events has been developed for antares .
the algorithm is based on the average number of hit repetitions ( @xmath13 ) in the oms due to the different arrival times of so called direct and delayed photons .
the number of hit repetitions for a specific om in a single event is defined as the number of hits measured in a time interval up to @xmath14 after and including the first hit that is used for the muon track reconstruction .
the estimator @xmath13 is calculated as the average number of repetitions , dividing the sum of the number of repetitions in the individual oms by the number of all oms that contribute at least one hit selected by the track reconstruction algorithm .
direct photons reach an om without being scattered on their way from their cherenkov vertex along the muon track , whereas scattered cherenkov photons or photons induced by electromagnetic showers along the muon track are referred to as delayed with arrival time differences up to hundreds of nanoseconds with respect to direct photons . for high muon energies ( @xmath15 )
energy loss contributions due to radiative processes start to dominate and increase linearly with the muon energy , thus leading to additional delayed light in the detector due to electromagnetic showers .
this is exploited to select neutrino events and to finally discriminate between atmospheric background and cosmic neutrinos . using a large set of atmospheric muon and neutrino monte carlo events , for which the detector response was fully simulated
, the event selection has been optimised before the signal region was uncovered for the data .
the atmospheric neutrino background was modelled following the bartol flux parametrisation @xcite with an additional high - energy component induced by the decay of charmed mesons ( prompt component)@xcite .
the signal neutrino flux @xmath16 was modelled with a @xmath17 spectral shape .
the discrimination of signal and background neutrinos is achieved by a single cut on the hit repetition value @xmath13 of the above defined energy estimator .
this cut has been optimised to maximise the detector sensitivity using monte - carlo events only .
the expected number of remaining atmospheric neutrino events is @xmath18 .
the systematic uncertainties are dominated by the uncertainty on the atmospheric neutrino flux and the detector acceptance including its dependence on environmental parameters .
data taken during the years 2007 to 2009 with an equivalent live time of 334 days and with several different detector configurations are used for the analysis . after applying the same cut on the data , 9 neutrino candidate events remain , in full agreement with the background assumption .
this result translates into an upper limit ( 90% c.l . ) for the diffuse cosmic neutrino flux of @xmath19 .
this limit @xcite holds for an assumed @xmath17 signal spectrum and in a neutrino energy range @xmath20 .
models yielding spectral shapes different from the generic @xmath17 have been tested and some of them could be excluded ( cf .
@xcite and ref .
therein ) at the 90% confidence level .
a search for cosmic sources of muon neutrinos has been carried out using data collected in the years between early 2007 until the end of 2010 , corresponding to an integrated live time of 813 days
. an earlier version of this analysis using data collected in 295 days can be found in @xcite .
the muon reconstruction algorithm employed is based on a maximum likelihood fit and yields the track direction , a reconstruction quality parameter based on the reduced log - likelihood of the track fit and an estimate of the angular error . using a full detector simulation ,
an average angular resolution of @xmath21 , defined as the median angle between the neutrino and the reconstructed muon direction , has been determined for a @xmath17 neutrino spectrum .
optimising the limit setting and discovery potential at the same time , only upward - going events with a good reconstruction quality and an angular error estimate better than @xmath22 are selected .
the resulting event sample consists of 3058 up - going event candidates .
mc simulations indicate that 84% of these can be expected to be neutrinos , while 16% are expected to be misreconstructed down - going atmospheric muons .
as the signature of a point source is a cluster of events at a given celestial position , two different analyses , both using an unbinned maximum likelihood method employing the angular resolution and the declination - dependent rate of background events , are applied to the up - going event sample , in order to search for such clusters . in each case , the likelihoods for the assumption that the cluster under inspection is induced by pure background or by an additional source contribution of unknown intensity are computed and their ratio is used to distinguish between signal - like clusters and clusters induced by background fluctuations . in the first analysis the full sky is searched for point sources , while in the second analysis neutrinos are searched for only in the direction of 50 predefined celestial positions , corresponding to well known galactic and extragalactic astrophysical objects that could be powerful cosmic accelerators and potential neutrino sources @xcite .
these candidate sources were selected according to their observed gev to tev gamma - ray fluxes , whereby no temporal coincidence of the gamma - ray emission with the analyzed data was required .
neither search yielded a significant excess of events over the expectation from pure atmospheric neutrino background .
the post - trial probability of the most signal - like cluster of events at @xmath23 for the full sky search to be compatible with a fluctuation of the atmospheric neutrino background is 2.5% ( p - value ) and consequently the hypothesis of a neutrino point source at these coordinates must be rejected as insignificant .
the candidate search yields the object hess j1023 - 575 as the source associated with the most signal - like cluster .
the probability , however , for a background fluctuation resulting in a cluster at least as significant as the observed one , is 41% .
consequently , as no excess neutrino signal has been observed with sufficient significance , upper limits have been set on the flux for each individual source candidate assuming an @xmath17 spectrum .
the individual limits and the antares sensitivity as a function of declination and computed as the median of the limits are reported in fig .
[ fig : limits ] .
the limit computation is based on a large number of simulated experiments in which systematic uncertainties on the angular resolution and the acceptance are taken into account .
the obtained upper limits are more stringent than those from previous experiments in the northern hemisphere ( observing the southern sky ) and competitive with those set by the icecube observatory @xcite for declinations @xmath24 .
the various experiments are sensitive in different energy ranges , even though they all set limits on @xmath17 spectra .
for such a spectrum , antares detects most events at energies in a broad range around 10 tev , which is a relevant energy range for several galactic source candidates . by design
, neutrino telescopes constantly monitor at least one complete hemisphere of the sky and are thus well set to detect neutrinos produced in transient astrophysical sources .
the importance of the irreducible neutrino background originating from earth s atmosphere can be drastically reduced by selecting a narrow time window around the assumed neutrino production period .
blazars are particularly interesting as potential neutrino point sources as their enormous energy output in the form of electromagnetic radiation and their relativistic outflow of collimated streams of matter make them good candidate sources of ultra high - energy cosmic rays . as a consequence , neutrinos and gamma - rays
may be produced in interactions of accelerated hadrons with intense ambient photon fields or matter .
the gamma - ray light curves of blazars measured by the lat instrument on - board the fermi satellite reveal important time variability on a timescale of hours up to several weeks , with high - state intensities mostly several times larger than the typical flux of the source in its quiescent state @xcite . assuming a hadronic origin of the observed gamma - rays , it is assumed in the following that the observed time - variable gamma - ray fluxes and the expected associated neutrino fluxes are proportional .
therefore , high states of gamma - ray activity in a source are used to define time windows for the neutrino search from this source .
the antares data used in this first coincidence search corresponds to the period from september 6th to december 31st , 2008 and were taken with the complete detector in its 12-line setup .
periods with high bioluminescence - induced optical background have been excluded , which results in an effective live time of 60.8 days .
the algorithm for a point source search with time dependence factorizes the probability of a given event to be signal or background into a directional and a time component .
the probability density function describing the background contribution is derived from data using the observed declination distribution of the selected neutrino candidate events and the time distribution of all reconstructed muons . the probability density for the signal
is described by the telescope s point spread function and by the gamma - ray light curve of the source under inspection . using a method that maximizes the likelihood ratio of signal and background , the average number of signal events required to achieve a @xmath25 discovery with 50% probability can be computed as a function of the flare duration . for a test flare with a constant flux light curve from a source at a declination of @xmath26 , it has been shown ( cf .
@xcite for details ) that the discovery potential of the telescope improves at least by a factor of 2 for flare durations shorter than a day compared to the standard time integrated point source analysis .
this time - dependent analysis has been applied to bright and variable blazars reported in the first year fermi lat catalogue @xcite and in the lat bright agn sample @xcite . only sources in the field of view of antares were selected whose average high - state photon flux for gamma energies above @xmath27 in bins of one day duration was greater than @xmath28 and that showed significant time variability in the studied time period .
the resulting list includes six flat spectrum radio quasars and four bl lacs .
the most significant source is the blazar 3c279 , which has a pre - trial p - value of 1% .
the unbinned method described above finds one high - energy neutrino event located at @xmath29 from the source location during a large flare in november 2008 . fig .
[ fig : fermi ] shows the time distribution of the fermi gamma - ray light curve of 3c279 and the time of the coincident neutrino event @xcite .
this event has been reconstructed with 89 hit optical modules distributed over 10 out of 12 detection lines .
the error estimate on the reconstructed direction , derived from the maximum - likelihood track fit , is @xmath30 .
the post - trial probability is computed taking into account the search for ten different objects . the final probability to find a signal at least as significant as the one observed amounts to 10% .
hence , the observed neutrino can be attributed to the atmospheric background .
the energy information of the neutrinos has not been used in this analysis up to now .
the antares neutrino telescope started routine data taking in a configuration with 5 installed lines in 2007 .
since may 2008 this first - generation telescope is complete and has in the meanwhile recorded a large neutrino sample of high quality .
the feasibility of installation and operation of a particle physics detector in the hostile environment of the deep sea has been demonstrated and first results have been obtained .
the search for a cosmic diffuse high - energy neutrino flux and a search for steady point sources both resulted in stringent and competitive upper limits for the flux of cosmic neutrinos . in order to increase the discovery potential ,
a first step towards multi - messenger searches has been made by using fermi - lat gamma - ray flux data to narrow the search time windows for neutrino production .
the main assumption made is that the gamma - ray and neutrino fluxes are proportional .
one neutrino has been found in coincidence with a gamma - ray flare of a flat spectrum radio quasar during the last 4 months of the year 2008 , but can be explained by the atmospheric background .
the successful operation of the antares neutrino telescope is an important step towards km3net @xcite , a future multi-@xmath31-scale high - energy neutrino observatory and marine science infrastructure proposed for construction in the mediterranean sea .
the authors acknowledge the financial support of the funding agencies : centre national de la recherche scientifique ( cnrs ) , commissariat lnergie atomique et aux energies alternatives ( cea ) , agence national de la recherche ( anr ) , commission europenne ( feder fund and marie curie program ) , rgion alsace ( contrat cper ) , rgion provence - alpes - cte dazur , dpartement du var and ville de la seyne - sur - mer , france ; bundesministerium fr bildung und forschung ( bmbf ) , germany ; istituto nazionale di fisica nucleare ( infn ) , italy ; stichting voor fundamenteel onderzoek der materie ( fom ) , nederlandse organisatie voor wetenschappelijk onderzoek ( nwo ) , the netherlands ; council of the president of the russian federation for young scientists and leading scientific schools supporting grants , russia ; national authority for scientific research ( ancs ) , romania ; ministerio de ciencia e innovacin ( micinn ) , prometeo of generalitat valenciana ( gva ) and multidark , spain .
we also acknowledge the technical support of ifremer , aim and foselev marine for the sea operation and the cc - in2p3 for the computing facilities .
9 m. sikora et al . , _ astrophys .
j. lett . _
* 320 * ( 1987 ) l81 ; f.w .
stecker et al . , _ phys .
* 66 * ( 1991 ) 2697 ; * 69 * ( 1992 ) 2738(e ) ; k. mannheim , p. biermann , _ astron .
_ * 253 * ( 1992 ) l21 .
k. greisen , _ ann . rev .
sci . _ * 10 * ( 1960 ) 63 .
gaisser et al . , _ phys .
* 258 * ( 1995 ) 173 - 236 .
markov , i.m .
zheleznykh , _ nucl .
_ * 27 * ( 1961 ) 385 .
becker , _ phys .
* 458 * ( 2008 ) 173 - 246 ; l.a .
anchordoqui , t. montaruli , _ ann .
nucl . part .
* 60 * ( 2010 ) 129 - 162 ; u.f .
katz , ch .
spiering , arxiv:1111.0507v1 [ astro-ph.he ] . m. ageron et al .
instrum . meth .
a _ * 656 * ( 2011 ) 11 - 38 .
aguilar et al . , _ astropart .
* 34 * ( 2011 ) 539 - 549 .
aguilar et al . , _ astropart .
phys . _ * 33 * ( 2010 ) 86 - 90 .
aguilar et al . , _ astropart .
_ * 34 * ( 2010 ) 179 - 184 .
stecker , _ astrophys .
j. _ * 228 * ( 1979 ) 919 .
e. waxman , j. bahcall , _ phys .
* 59 * ( 1998 ) 023002 .
j. bahcall , e. waxman , _ phys .
* 64 * ( 2001 ) 023002 .
k. mannheim , r.j .
protheroe , j.p .
rachen , _ phys .
* 63 * ( 2000 ) 023003 .
barr et al . , _ phys .
* 70 * ( 2004 ) 023006 .
bugaev et al . , _ nuovo cimento c _ * 12 * ( 1998 ) 41 .
v. agrawal et al . , _ phys .
* 53 * ( 1996 ) 1314 .
aguilar et al . , _ phys .
b _ * 696 * ( 2011 ) 1622 .
s. adrian - martinez et al . , _ astrophys .
j. lett . _
* 743 * ( 2011 ) l14-l19 .
s. adrian - martinez et al . , arxiv:1112.0478v1 [ astro-ph.he ] , 12 .
r. abbasi et al . , _ astrophys .
j. _ * 732 * ( 2011 ) 18 . a.a .
abdo et al . , _ astrophys .
j. _ * 722 * ( 2010 ) 520 .
s. adrian - martinez et al . , arxiv:1111.3473v1 [ astro-ph.he ] .
abdo et al . , _ astrophys .
j. supp . _
* 188 * ( 2010 ) 405 .
abdo et al . ,
_ astrophys .
j. _ * 715 * ( 2010 ) 429 .
the km3net consortium , conceptual and technical design report , + isbn 978 - 90 - 6488 - 031 - 5 and isbn 978 - 90 - 6488 - 033 - 9 , www.km3net.org . | the antares detector is the most sensitive neutrino telescope observing the southern sky and the world s first particle detector operating in the deep sea .
it is installed in the mediterranean sea at a depth of @xmath0 . as an example of early results ,
the determination of the atmospheric muon flux is discussed and a good agreement with previous measurements is found . furthermore , the results of a search for high - energy events in excess of the atmospheric neutrino flux are reported and significant limits are set on the diffuse cosmic neutrino flux in the multi - tev to pev energy range . using data from more than 800 days of effective data taking , partly during the construction phase ,
a first analysis searching for point - like excesses in the neutrino sky distribution has been performed .
the resulting sensitivity of antares is reported and compared to measurements of other detectors . a method employed for a first search for neutrinos from fermi - detected gamma - ray flaring blazars in the last 4 months of 2008
is described and the results are reported .
no significant neutrino signal in excess of that expected from atmospheric background has been found . |
the past decade has seen the discovery of a large number of massive young star clusters throughout the local universe .
these systems are large and young enough that they contain statistically significant numbers of massive stars , affording us key insights into their initial mass function and structural properties .
of greatest interest to dynamicists are those systems in which stellar dynamical time scales are short enough that the cluster can undergo significant structural change during the lifetimes of the most massive stars . in such clusters ,
dynamical evolution opens up entirely new avenues for stellar and binary evolution , allowing the formation of stellar species completely inaccessible by standard stellar and binary evolutionary pathways . from this perspective
, the clusters listed by portegies zwart and mcmillan ( these proceedings ) represent an ideal combination of properties , having ages of less than a few million years and relaxation times of less than a few tens of millions of years . in these clusters , dynamical evolution ,
traditionally regarded as a `` slow '' process , actually occurs much more rapidly than the stellar evolution of even the most massive stars .
thus the dynamics controls the early phases of these stars lives .
we focus here on perhaps the most dramatic environmental modification of standard stellar evolution repeated physical collisions between stars .
the scenario described below is fast becoming the `` standard '' paradigm by which runaway stellar mergers might occur , perhaps forming intermediate - mass black holes ( imbhs ) in sufficiently dense systems .
the possibility of multiple mergers of massive stars in dense stellar systems was first demonstrated in n - body simulations of r136 by portegies zwart et al .
subsequently ebisuzaki et al .
( 2001 ) suggested that the ultraluminous x - ray source m82-x1 might be the result of such a process ; this possibility was explored in detail by portegies zwart et al .
portegies zwart & mcmillan ( 2002 ) and mcmillan & portegies zwart ( 2003 ) have explored the possibility of imbh formation in young clusters in the galactic center .
recently , grkan et al . (
2004 ) have reaffirmed the basic process using monte - carlo simulations , and have carried out systematic studies of runaway mergers in galactic nuclei . in this paper
we consider how a young star cluster might come to be in such a high - density state , and look critically at the key physical processes needed for a runaway merger to occur .
we then turn briefly to the possibility of an imbh in the center of the milky way galaxy ( maillard et al .
2004 ) , showing how theory and recent observations may provide a consistent picture of the galactic center .
consider a massive object of mass @xmath0 and radius @xmath1 moving through a field of background stars of total mass density @xmath2 and velocity dispersion @xmath3 .
if @xmath4 and @xmath5 are large compared to the masses and radii of other stars , and all velocities are small enough that gravitational focusing dominates the total cross section , the object s collision cross section is @xmath6 , nearly independent of the properties of the other stars .
the rate of increase of the object s mass due to collisions then is @xmath7 if the object initially has @xmath8 and we adopt a simple mass radius relation @xmath9 , then for the object to reach @xmath10 in 3 myr ( to form an imbh within the lifetime of a massive star ) , the local density must satisfy @xmath11 such a density is much higher than the mean density of any known star cluster , young or old . for comparison ,
the average density of the arches cluster is @xmath12 , that of a fairly compact globular cluster is @xmath13 , while even the most concentrated globular cluster cores have @xmath14 .
mergers would be enhanced if the cluster were born very centrally concentrated , as suggested by portegies zwart et al .
( 2004 ) and merritt et al .
( 2004 ) . as a simple limiting model of such a cluster ,
consider the nearly isothermal system of total mass @xmath15 and half - mass radius @xmath16 , described by the density profile @xmath17 for @xmath18 .
densities exceeding @xmath19 are found for @xmath20 , where @xmath21 where the cluster velocity dispersion is @xmath22 .
however , the total mass contained within this radius is just @xmath23 we conclude that , for reasonable cluster parameters , there is too little initial mass in the high - density region to accomplish the task of forming a @xmath24 object in the time available .
thus collisions in a static cluster core can not lead to the formation of an ultramassive object . however , cluster dynamical evolution can result in conditions much more favorable for a runaway merger to occur .
the evolution of a cluster is governed by its half - mass relaxation time , the time scale on which two - body encounters transport energy around the system : @xmath25 ( heggie & hut 2003 ) . here , @xmath26 is the number of stars in the system , @xmath27 is the mean stellar mass , taken here to be @xmath28 , @xmath29 is the mean cluster density , and @xmath30 . for an equal - mass system , the time scale for dynamical evolution the core collapse time is about @xmath31 , too long to cause significant structural change within a few million years .
however , the presence of even a modest range in masses greatly accelerates the process of core collapse ( spitzer 1987 ) .
the time scale for a star of mass @xmath32 to sink to the cluster center as equipartition reduces its velocity is @xmath33 where @xmath34 is the local relaxation time .
portegies zwart & mcmillan ( 2002 ) find that the most massive ( @xmath35 ) stars segregate rapidly to the cluster center , forming a dense stellar subcore on a time scale @xmath36 .
a central density increase of 23 orders of magnitude is typical , boosting even a relatively low - density core into the range where collisions become common , and greatly increasing the supply of raw material to form a collision runaway . in systems with @xmath37 myr ( @xmath38 myr ) , essentially all the massive stars in the cluster reach the center before exploding as supernovae , and hence can participate in the runaway process .
the arches and westerlund i fall into this category ; the quintuplet , ngc 3603 , and r136 all come close . in these cases ,
the maximum mass of the runaway is limited primarily by the total number of massive stars in the system anywhere from a few percent to several tens of percent of the total , depending on the cluster mass function . in less dense or more massive clusters
, the longer half - mass relaxation time means that only a fraction of the massive stars initially present in the system can reach the center in the time available , but the total supply of mass may still ensure that a runaway can occur .
we can estimate the amount of mass made available by mass segregation as follows . again adopting an isothermal model to simplify the calculation , and taking the segregation time scale from eq .
[ tseg ] , it is easily shown that a star of mass @xmath32 can sink to the center within time @xmath39 if its initial distance @xmath40 from the cluster center satisfies @xmath41 the fraction of stars of mass @xmath32
satisfying this relation is @xmath42 choosing ( again for simplicity ) a mass function @xmath43 for @xmath44 , we determine the total stellar mass potentially available for mergers as @xmath45 where the lower mass limit @xmath46 is somewhat greater than the mean stellar mass ( @xmath47 here ) , but is otherwise unimportant so long as it is much less than the upper limit of @xmath48 .
the result for the chosen mass function is @xmath49 comfortably above the value needed for imbh formation .
we have assumed here that the cluster relaxation time is long enough that @xmath50 for all @xmath32 of interest .
this is the case for @xmath51 myr .
an appropriately modified version of this analysis for @xmath52 myr shows the available mass leveling off at a fixed fraction of the cluster mass as @xmath53 , as expected .
given that dynamical evolution can concentrate enough mass in a cluster core for collisions to occur at a significant rate , we can then ask ( i ) if the collisions actually lead to mergers , and ( ii ) under what circumstances a runaway merger can occur . the answer to the first question
is provided by the sph simulations of freitag & benz ( 2001 ; see also lombardi et al .
2003 ) , who find that the low relative velocities typical of these systems ensure that the colliding stars usually merge with minimal mass loss . in small systems ( containing less than a few tens of thousands of solar masses ) ,
collision rates are significantly enhanced by the fact that the massive object tends to form binaries , which are then perturbed into eccentric orbits by encounters with other stars ( portegies zwart & mcmillan 2002 ) .
binary - induced mergers increase the collision cross section , but they still require high central densities before the ( three - body ) binary formation rates become significant . in larger systems ,
unbound collisions appear to be the norm .
thus collisions naturally involve the most massive stars in the cluster , and lead to the production of even more massive objects .
the merger products are generally out of thermal equilibrium and often rapidly rotating , with the result that their subsequent stellar evolution is currently poorly understood . in our simulations
we generally assume that the merged object evolves along a suitably rejuvenated ( non - rotating ) track appropriate to its mass ( portegies zwart et al .
this is at best a crude approximation , but we note that it probably underestimates the radius of the merger product during the out - of - equilibrium phase and hence the likelihood of a runaway .
acting in the opposite sense is the fact that , while stellar mass loss rates are very uncertain , ultramassive stars probably have very strong winds .
van beveren ( these proceedings ) points out that if the wind mass loss rate exceeds the accretion rate due to mergers , then the entire runaway process may fail .
our simulations generally yield net merger accretion rates of @xmath54 , suggesting that a high ( but perhaps not impossibly so ) mass - loss rate is needed to shut the process down .
of course , it must be conceded that next to nothing is known about the detailed evolution and ultimate fate of stars hundreds or thousands of times more massive than the sun . nevertheless
, the estimates presented here make it clear that dynamical evolution in dense stellar systems can easily produce conditions suitable for repeated stellar collisions .
the collision runaway at the center of such a system should be extremely luminous and eminently observable during its short lifetime .
observations of the cores of dense young star clusters in our galaxy and beyond may thus shed light on the structure and lifetimes of such ultramassive stellar objects .
if our evolving cluster happens to reside close to the galactic center , then dynamical friction will tend to drive it inward , raising the possibility that mass segregation can lead to a collision runaway en route , and that the resulting imbh can be transported rapidly toward the center by the much more massive cluster ( mcmillan & portegies zwart 2003 ) .
figure [ fig1 ] illustrates this possibility .
it shows the mass @xmath4 and galactocentric distance @xmath5 of the runaway formed in a cluster of initial mass @xmath55 , placed in a circular orbit of radius 10 pc , as it spirals inward and is disrupted by the galactic field .
we clearly see the inward transport of the growing merger product , terminating with the dissolution of the cluster at @xmath56 myr .
subsequently , the black hole sinks more slowly to the center , eventually reaching @xmath57 at @xmath58 myr .
the early ( @xmath59 myr ) portion of the figure is from an n - body simulation ; the remainder is based on a simple analytical model of the galactic potential ( sanders & lowinger 1972 ) .
the recent report by maillard et al .
( 2004 ) of a possible imbh near the galactic center , at the heart of the swarm of stars known as irs13 , provides the exciting prospect of confronting our models directly with reality .
portegies zwart et al .
( 2005 , in preparation ) find that the observed properties of irs13 are completely consistent with theoretical expectations for a dense cluster remnant .
they further estimate that the total imbh mass resulting from runaway collisions in the cluster population which produced the inner bulge is consistent with the @xmath60 supermassive black hole at the galactic center .
this work was supported by nasa atp grant nag5 - 10775 , the royal netherlands academy of sciences ( knaw ) , the dutch organization of science ( nwo ) , and by the netherlands research school for astronomy ( nova ) . | close encounters and physical collisions between stars in young dense clusters may lead to the formation of very massive stars and black holes via runaway merging .
we examine critically some details of this process , using n - body simulations and simple analytical estimates to place limits on the cluster parameters for which it expected to occur . for small clusters ,
the mass of the runaway is effectively limited by the total number of high - mass stars in the system . for sufficiently dense larger clusters ,
the runaway mass is determined by the fraction of stars that can mass segregate to the cluster core while still on the main sequence .
the result is in the range commonly cited for intermediate - mass black holes , such as that recently reported in the galactic center . |
ordinary thermodynamics describing the macroscopic phenomenology of homogeneous equilibrium systems leads to the following _ stability conditions _ : @xmath0 these conditions ensure that the pressure @xmath1 of a system decreases as its volume @xmath2 increases and that the system warms up when absorbing energy at fixed volume ( positive specific heat at constant volume @xmath3 ) .
one of the major achievements of mathematical physics was to show that ordinary matter , composed of a large number of electrons and protons , behave according to eqs .
( [ eq : dp / dv<0],[eq : cv>0 ] ) @xcite .
however , it has recently turned out that systems not satisfying the condition ( [ eq : cv>0 ] ) , actually exist in nature , and that a corresponding thermodynamic description should yet be possible in certain cases @xcite .
the prototypical examples of such unstable systems are stars which are known to expand and cool down as their energy increases @xcite . within the canonical ensemble
the specific heat is related to the fluctuations of energy via the relation @xmath4 . from this expression
it is evident that specific heat is necessarily positive in the canonical ensemble .
negative specific heat , however may appear within the microcanonical ensemble .
this can happen due to different mechanisms : * the ergodic properties of the system may depend on energy .
hence , at different energies , different parts of phase space are accessible .
this may lead to negative specific heat and other thermodynamical anomalies .
an example was given by hertel and thirring @xcite .
* the system might be far from the thermodynamic limit .
systems that are stable in the thermodynamic limit , such as lennard - jones gases , may display negative specific if only their size is small enough @xcite . *
long ranged forces might prevent the thermodynamic limit to exist at all .
but even when an equilibrium state exists such systems remain _ nonextensive _ and may show negative specific heat @xcite . in this work we will present yet another mechanism that leads to negative specific heat and other thermodynamic anomalies for systems in _ contact _ with a _ heat bath_. apart from extensivity and short ranged interactions , another assumption is customarily made in the statistical mechanics of canonical systems , namely _ weak coupling _ between system and its environment . when the coupling to the environment is not negligible , violations of condition ( [ eq : cv>0 ] ) may appear even if system and environment , as a whole , are in a canonical state @xcite .
hence , yet another item can be added to our list of exceptions : * systems that strongly interact with their environment may display thermodynamic anomalies for example , a single free particle , which would display a positive specific heat when weakly coupled to a bath , may display negative specific heat when in strong interaction with the bath @xcite .
similar effects were observed for a two level system coupled to a harmonic oscillator @xcite .
we assume that the total system @xmath5 , is in _ weak _ contact with a _ super - bath _ that provides the temperature concept ( @xmath6 , @xmath7 is boltzmann constant . )
this total system is therefore described by the _ canonical _
statistics @xmath8 , where @xmath9 is the hamiltonian of the total system @xmath5 and @xmath10 . accordingly
, its specific heat can not be negative .
however , if the coupling between @xmath11 and @xmath12 is non - negligible , the overall canonical state does not factorize into the product of two canonical states for @xmath11 and @xmath12 respectively : @xmath13 .
although the system is in perfect thermal equilibrium with its environment , it is not in a canonical state and therefore negative entropies and violations of the inequalities ( [ eq : dp / dv<0],[eq : cv>0 ] ) may occur .
these violations though are not a sign of any instability because as a part of a stable total system , the system of interest itself is also stable .
there are few available exact solutions of open systems displaying this kind of anomalies .
one known example is the damped free particle @xcite and a two - level fluctuator in contact with a single oscillator @xcite .
numerical investigations of thermodynamic anomalies in the context of the casimir effect , the multichannel kondo effect and of mesoscopic superconductors containing magnetic impurities have been reported recently in the literature @xcite . in this work
we study the thermodynamics of a subchain of a longer chain of spins interacting according to the isotropic xy model with free ends @xcite .
the spin chain is composed of two parts .
one subchain is defining the system @xmath14 , while the rest of the chain comprises the bath , @xmath15 ( see fig .
[ fig : lattice ] ) .
being this an exactly solvable model , we are able to analytically find the relevant thermodynamical functions for the subchain . far from being a purely academic problem
, the study of the equilibrium properties of spin-@xmath16 chains has been recently attracting a great deal of attention .
spin systems not only are the basis of the physics of magnetic materials @xcite but they might have an enormous impact with regard to the development of quantum technologies @xcite .
we believe that the study of the thermodynamics of small quantum systems will help to understand problems of quantum information and vice versa @xcite . .
the chain is composed of two parts : the system of interest @xmath11 , made of the first left - most @xmath17 spins ( green squares ) , and the `` bath '' @xmath12 , made of the remaining @xmath18 spins ( yellow circles ) .
when the interaction between the two subchains is non negligible , anomalies may occur in the thermodynamics of the system of interest @xmath11.,width=283 ] in sec .
[ sec : tdopen ] we briefly review the generalities of the thermodynamics of open systems . the specific model studied in this paper
is described in sec .
[ sec : model ] , while its thermodynamics is illustrated in sec .
[ sec : thermodynamics ] .
various anomalies ranging from negative entropy and negative specific heat , to negative susceptibility are observed for system size , @xmath17 , of the order of unity . in sec .
[ sec : rho ] we study the reduced density matrix of the open system @xmath14 and show that it departs from the canonical form in the strong coupling regime . the spectrum of the reduced density matrix is analyzed and it is observed that , regardless of the strong coupling , a quantum phase transition occurs at zero temperature .
conclusions will be drawn in sec .
[ sec : conclusions ] .
consider the following hamiltonian : @xmath19 describing a systems of interest @xmath14 interacting with a second system , the bath @xmath20 , via the interaction energy term @xmath21 .
the compound system @xmath5 is in weak contact with a large super - bath at temperature @xmath22 .
hence it is described by the canonical density matrix : @xmath23 where @xmath24 is the total system partition function , with @xmath25 denoting the trace over the total system hilbert space .
the partition function of the open quantum system , @xmath11 , is given by the ratio of the total system partition function @xmath26 and the bare bath partition function @xmath27 , i.e. @xcite : @xmath28 where @xmath29 with @xmath30 denoting the trace over the bath hilbert space .
according to the rules of statistical mechanics , the helmholtz free energy of the open quantum system is then : @xmath31 from this free energy , all the relevant thermodynamic functions of the open quantum system can be derived . in particular ,
the entropy and the specific heat are given by @xmath32 , and @xmath33 , respectively .
thanks to eq .
( [ eq : z = ztot / zb ] ) the system s free energies is the difference of the free energy of the total system and the bare bath , i.e. : @xmath34 due to the linearity of the derivative , this implies according relations for the entropy and the specific heat of the system , which take the form : @xmath35 while @xmath36 and @xmath37 are positive numbers , there is no reason why their difference should be positive as well .
hence , coupling the bath @xmath20 strongly to the system @xmath14 may result in an overall decrease of specific heat .
the same holds for the entropy and other thermodynamic quantities that linearly depend on the free energy .
these relations will be exemplified with the model described below .
we consider the isotropic @xmath38 model of @xmath39 interacting spins on a one dimensional lattice of equally spaced sites with _ free ends _ : @xmath40 where @xmath41 and @xmath42 are the pauli matrices of the @xmath43 spin . dropping the label @xmath44
these are : @xmath45 the first term in the hamiltonian ( [ xymodel ] ) accounts for the _ zeeman _ interaction between each spin and an applied magnetic field @xmath46 pointing in the @xmath47-direction .
the second term describes the nearest neighbor interaction with strength @xmath48 .
the isotropic @xmath38 model ( [ xymodel ] ) is an exactly solvable model . by means of the jordan - wigner transformation @xcite ,
the hamiltonian ( [ xymodel ] ) is mapped onto a system of @xmath39 free fermionic eigenmodes , of energies @xmath49 see [ app : jw ] .
thereby , the spectrum of the hamiltonian ( [ xymodel ] ) is expressed in terms of the occupation numbers of each fermionic mode @xmath50 , as @xmath51 where each @xmath50 can only be @xmath52 or @xmath53 .
given the spectrum , the partition function , @xmath54 , is calculated as : @xmath55 according to the formula @xmath56 , the free energy of an isotropic @xmath38 chain of @xmath39 spins then becomes : @xmath57 where @xmath58 denotes the contribution from the @xmath59 fermionic mode .
we now consider the part of the chain consisting of the first @xmath17 spins , counted , say , from the left end of the chain , as the system of interest , @xmath11 , and the rest of @xmath60 spins as the bath , @xmath20 , see fig .
[ fig : lattice ] .
the total hamiltonian @xmath61 , can be recast in the usual system - bath form @xmath62 of eq .
( [ htotal ] ) , with @xmath63 @xmath64 @xmath65 the regime of strong coupling holds when the system energy is comparable to the interaction energy .
weak coupling is achieved in this problem when at least one of the following holds @xmath66 , @xmath67 , @xmath68 . from the eq .
( [ eq : f = ftot - fb ] ) , the thermodynamic free energy of the open system @xmath11 is the difference between the free energy of the total chain of length @xmath39 and the free energy of the bare bath @xmath12 , which is itself a @xmath38 chain of a certain length @xmath18 .
its free energy is then readily obtained by replacing @xmath39 with @xmath18 in eq .
( [ eq : fn ] ) .
we thus obtain for the system of interest @xmath11 : @xmath69 using eqs . ( [ eq : fn],[eq : fk ] ) : @xmath70
we discuss now the explicit results for the relevant thermodynamical functions , such as entropy , specific heat , magnetization and susceptibility of the open system defined in eqs .
( [ htotal ] ) and ( [ hs ] ) .
see also fig .
[ fig : lattice ] .
we begin our discussion with studying the entropy of @xmath11 : @xmath71 from eq .
( [ eq : fn ] ) , follows @xmath72 where @xmath73 and @xmath74 denote the total system and the bare bath entropies , respectively , and @xmath75 is the single mode entropy of the total chain of length @xmath39 .
likewise @xmath76 is the single mode entropy of the bath @xmath12 .
( black solid line ) , @xmath77 ( red solid line ) , @xmath78 ( black dashed line ) , @xmath79 ( red dashed line ) .
top panel : @xmath80 , bottom panel @xmath81 .
entropy becomes negative at low temperature for the lower magnetic field @xmath81 , in the strongly coupled case , i.e. @xmath82,width=302 ] in fig .
[ fig : s ] the temperature dependence of the entropy per spin @xmath83 of the system @xmath11 , is shown for different system sizes @xmath17 of @xmath11 and fixed total size @xmath39 .
the magnetic field is @xmath80 and @xmath81 in the top and bottom panel of fig .
[ fig : s ] , respectively . for comparison ,
the graphs showing the temperature dependence of entropy for the same system sizes in absence of the bath , are shown as well .
we shall use the label @xmath84 for the bare system .
notably the presence of the bath does not affect much the entropy of the system at high temperatures , whereas its effect is most prominent at low temperatures .
this high temperature behavior can be understood by looking at eq .
( [ eq : s_2 ] ) .
the entropy @xmath85 of each mode tends to @xmath86 for large temperatures , thus the entropy of the total system tends to @xmath87 .
likewise , the entropy of the bare bath tends to @xmath88 and that of the system @xmath11 tends to @xmath89 , which is the same as the entropy of the bare system @xmath84 at high temperature , i.e. , @xmath90 .
thus , for increasing temperature the effect of the bath on the entropy of the system @xmath11 becomes less relevant .
with reference to both panels of fig .
[ fig : s ] , we see that at low temperature , the effect of the bath becomes increasingly relevant as the system size decreases .
as the system shrinks , the relative effect of the interaction with the bath , which is a _ surface _ effect , becomes more important ( strong coupling ) .
in particular , we notice the pronounced peak in the entropy for the smallest value of @xmath17 ( @xmath91 ) in the case @xmath80 , top panel of fig . [
fig : s ] .
a different situation occurs in the bottom panel of fig .
[ fig : s ] where the magnetic field is chosen to be @xmath81 and all other parameters are kept unchanged as compared to fig .
[ fig : s ] .
the major difference is the presence of a region of _ negative entropy_. this is a feature that is often found in open systems , in the regime of strong coupling @xcite for chain of @xmath92 spins coupled to a bath of @xmath93 spins ( black solid line ) , as compared to an isolated chain with @xmath92 ( red dashed line ) at fixed @xmath94 .
the two entropies coincide only for small enough values of @xmath95 .
while the entropy of the uncoupled chain is always positive , the entropy of the coupled chain may become negative for large enough values of @xmath95.,width=302 ] fig .
[ fig : sj ] shows the entropy of a larger chain ( @xmath92 ) coupled to an equally large bath ( @xmath93 ) as a function of coupling strength @xmath95 , as compared to the entropy of the same chain with no bath . for small enough @xmath95
the two entropies are approximately equal , while thir difference is apparent for large enough coupling .
the entropy of the uncoupled chain is always positive , whereas the entropy of the coupled chain may become negative for large enough coupling @xmath95 .
the specific heat at constant magnetic field @xmath46 is obtained from the entropy via the standard formula : @xmath96 evidently @xmath97 is the difference between the total system specific heat , @xmath98 , and the bare bath specific heat , @xmath99 @xmath100 where the single mode specific heat is : @xmath101 the specific heat of each single mode vanishes at high temperature and so do the total chain specific heat , the specific heat of the bath , the specific heat of the system @xmath11 , and the specific heat of the bare system @xmath84 .
( black solid line ) , @xmath77 ( red solid line ) , @xmath78 ( black dashed line ) , @xmath79 ( red dashed line ) .
top panel : @xmath80 , bottom panel @xmath81 .
negative specific heat regions appear at low temperature in the case @xmath81 ( bottom ) and at intermediate temperatures in the case @xmath80.,width=302 ] fig .
[ fig : c ] shows the specific heat for the same parameters as in fig .
[ fig : s ] , that is @xmath102 , @xmath103 , with @xmath80 ( top panel ) and @xmath104 ( bottom panel ) . quite interestingly , for the smallest system ( @xmath91 ) which is more affected by the bath , the specific heat displays a very pronounced positive peak at very low temperature followed by a _
negative _ peak at intermediate temperatures , for the case @xmath80 .
this appearance of a negative specific heat at intermediate temperature was observed also for a free particle coupled to a minimal bath composed of a single oscillator @xcite .
the behavior changes if the magnetic field is decreased to a value smaller than @xmath48 , e.g. @xmath81 , see bottom panel in fig .
[ fig : c ] . in case of
very small @xmath17 , i.e. @xmath91 , there is a negative peak at very low temperature and a positive one at intermediate temperature .
a similar situation was also observed for the case of a two - level system coupled to a minimal bath composed of a single oscillator @xcite .
the sign of the specific heat is given by the sign of the derivative of @xmath105 with respect to temperature @xmath106 .
thus , whenever the entropy displays a region where it decreases for increasing temperature , correspondingly the specific heat displays a region of negative values .
therefore , the presence of a positive peak in the entropy at low temperature leads to the appearance of negative and positive peaks in the specific heat at low and intermediate @xmath106 s , respectively ( compare the top panels in figs . [
fig : s ] , [ fig : c ] ) . likewise
the presence of a negative peak in the entropy at low @xmath106 leads to a negative specific heat region at low temperature whose extension is smaller than that of negative entropy ( see also @xcite ) . of further interest
is the magnetization of the system @xmath11 as a function of temperature @xmath106 and magnetic field
@xmath46 : @xmath107 it is given by the difference of the magnetizations of the total chain and the part of the chain representing the bath , reading : @xmath108 where the single mode magnetization , @xmath109 is : @xmath110 ( black solid line ) , @xmath77 ( red solid line ) , @xmath78 ( black dashed line ) , @xmath79 ( red dashed line ) .
top panel : @xmath80 , bottom panel @xmath81.,width=302 ] at high temperature the single mode magnetization tends to zero as the thermal agitation wins over magnetic ordering .
[ fig : m ] shows the magnetization per spin for the same parameters reported in figs .
[ fig : s ] , [ fig : c ] , i.e. , @xmath102 , @xmath103 and @xmath80 ( top panel ) , @xmath81 ( bottom panel ) . with reference to the top panel
, we see that in absence of a bath the magnetization per spin of the shortest chain , @xmath91 , is much larger than the magnetization of the longer chain @xmath111 , for thermal energies ( @xmath112 ) up to the order of @xmath48 .
when the system is put in contact with the bath , no relevant change in the magnetization is observed for the longer chain @xmath111 . for the smallest chain however
, a quick drop of the magnetization occurs with increasing temperature , and values close to those pertaining to larger chains are reached already at temperatures of the order @xmath113 . the bottom panel of fig .
[ fig : m ] shows the magnetization per spin for the same parameters as of the top panel but for @xmath81 .
in contrast to the case @xmath80 , and with reference to the smallest chain , @xmath91 , we observe a dramatic enhancement of several orders of magnitude of magnetization at values of @xmath114 , due to the coupling to the bath . by taking the partial derivative of the magnetization @xmath115 with respect to magnetic field ,
one obtains the magnetic susceptibility : @xmath116 one finds : @xmath117 where the single mode susceptibility reads : @xmath118 fig .
[ fig : x ] displays the susceptibility for the same set of parameter values used in figs .
[ fig : s ] , [ fig : c ] , [ fig : m ] , i.e. , @xmath102 , @xmath103 and @xmath119 .
note the enhancement of susceptibility at low temperature in both cases due to the presence of the bath . as expected , this is more pronounced for the shortest chain . of particular interest
is the behavior of the susceptibility for values of @xmath120 and the smallest chain ( bottom panel of fig .
[ fig : x ] ) .
note the large _ negative _ enhancement at low temperature for @xmath91 .
not only the entropy and the specific heat may display anomalous behavior in open systems and take on negative values , but also the susceptibility may do so .
( black solid line ) , @xmath77 ( red solid line ) , @xmath78 ( black dashed line ) , @xmath79 ( red dashed line ) .
top panel : @xmath80 , bottom panel @xmath81.,width=302 ] the system internal energy is obtained from the thermodynamic relation : @xmath121 evidently , also the internal energy of the system is given by the difference of the respective quantities for the total system and the bare bath , resulting to : @xmath122 where the single mode energy is given by : @xmath123 using eqs .
( [ eq : fk],[eq : sk ] ) for @xmath124 and @xmath125 we obtain @xmath126 where @xmath127 is the fermi distribution giving the average occupation number of the @xmath128th fermionic eigenmode . in fig .
[ fig : u ] we show the curve @xmath129 , parameterized by the temperature @xmath106 , for the fixed parameters @xmath102 , @xmath91 and @xmath80 ( top ) , @xmath81 ( bottom ) .
note that the relation of entropy and internal energy is not one - to one everywhere , but there is a region of internal energies belonging to different entropies and vice - versa .
this is a specific feature of strongly coupled open systems , which may not appear in the weak coupling regime .
the graph @xmath129 contains many informations .
each point on the plotted curves corresponds to a given temperature @xmath106 , which is identical to the slope of the graph at the very same point .
@xmath130 cusps appear in the curve @xmath129 for those values of @xmath106 for which the entropy @xmath131 has a local extremum . by comparison with fig .
[ fig : s ] , we see that for @xmath91 , @xmath80 , first the entropy goes through a maximum and subsequently a minimum , as the temperature is increased .
these two extrema correspond to the two cusps one encounters in the graph @xmath129 as one follows the curve in the @xmath132-@xmath105 plane starting at @xmath133 .
note that the curve continuously changes its slope when passing through the cusps .
( bottom ) , and @xmath80 ( top ) . for both values of @xmath81 ( bottom ) and
@xmath80 ( top ) cusps appear where the entropy has a maximum as a function of temperature.,width=302 ] the bottom panel of fig .
[ fig : u ] shows the curve @xmath129 @xmath102 , @xmath91 and @xmath81 .
only one cusp appears here in correspondence to the single extremum ( a minimum ) of the graph @xmath131 , compare to fig .
[ fig : s ] . in presence of vanishingly weak coupling ,
the specific heat is positive , meaning that the entropy is a strictly increasing function of @xmath106 .
this precludes the possibility of having local extrema in the entropy , which in turn excludes the appearance of cusps in the @xmath134 plot .
in this section we study the density matrix of the system of interest @xmath11 .
we recall that a super - bath thermalizes the total system @xmath5 .
this means that the total system is in a thermal gibbs state : @xmath135 the density matrix @xmath136 of @xmath11 is obtained by tracing out the bath degrees of freedom from @xmath137 : @xmath138 for the sake of simplicity , we limit our discussion to the case of a small system with @xmath91 , where the effects of the coupling to the bath @xmath12 are maximal . in this case of a system composed of two spins the reduced density matrix can be calculated by means of two - point correlators , according to the general formula @xmath139 where @xmath140 , denotes the identity operator of the hilbert space of the @xmath141 spin ( @xmath142 ) , and @xmath143 denotes quantum expectation values with respect to @xmath137 given in eq .
( [ eq : rhon ] ) . in the present case it is
@xmath144 with @xmath25 being the trace over the total system s hilbert space .
( [ eq : bloch4 ] ) , we find the reduced density matrix ( in the basis @xmath145 of the common eigenvectors of @xmath146 and @xmath147 ) , as : @xmath148 where @xmath149 using eq .
( [ eq : bloch4 ] ) one obtains , after standard but tedious algebra : @xmath150 diagonalization of @xmath136 is straightforward .
we obtain the following eigenvectors @xmath151 become @xmath152 the corresponding eigenvalues are : @xmath153 where we have introduced the abbreviation : @xmath154 it is interesting to study the spectrum of the reduced density matrix at zero temperature , i.e. , for @xmath155 . .
parameters are : @xmath156 .
beyond @xmath157 only a single nonvanishing eigenvalue exists . at higher temperatures
the steps are washed out.,width=302 ] fig .
[ fig : spectrum ] shows the spectrum as a function of the applied magnetic field for @xmath158 , @xmath159 , and @xmath102 .
for @xmath160 the only nonzero eigenvalue is @xmath161 , meaning that the open system @xmath14 is in the pure state @xmath162 .
the same holds for @xmath163 too , in that case the ground state is @xmath164 .
this is an interesting results : even though the system is ( strongly ) coupled to the bath @xmath12 , its ground state is the same pure state @xmath165 ( or @xmath166 ) that the system would have in absence of the bath , if the magnetic field is strong enough .
the fact that the bath does not alter the ground state for @xmath167 is another interesting aspect of a system strongly coupled to its environment .
this happens because for @xmath167 the total system ground state is a pure factorized state given by the product of single spin states all pointing in the same direction ( parallel to @xmath46 for @xmath163 , antiparallel to @xmath46 for @xmath160 ) . by tracing out the bath spins ,
a pure factorized state given by aligned spins remains for the subsystem @xmath11 .
on the other hand , when @xmath168 , the spins in the ground state of the total system are entangled .
thus tracing out the bath degrees of freedom leads now to a mixed state for the subsystem @xmath11 .
note that the transition of the system density matrix takes place at the same parameter value @xmath169 at which the infinite chain ( @xmath170 ) undergoes a _
quantum phase transition _ @xcite . the abrupt steps appearing in the spectrum , as displayed in fig . [ fig : spectrum ] , correspond to the values of @xmath46 for which the energy of a fermionic eigenmode vanishes , i.e , @xmath171 , see eq .
( [ eq : lambda ] ) .
these steps stem from the terms @xmath172 in eqs .
( [ eq : xx],[eq : z ] ) , which at zero temperature tend to unit steps : @xmath173 where @xmath174 is the heaviside function .
the summation over all these steps , then originates the staircase structure of the spectrum . in the thermodynamic limit @xmath175 ,
both width and height of the steps shrink and a continuous curve results .
an analytical expression for the @xmath176 s can then easily be found by replacing sums with integrals ( not reported ) .
it is apparent that , in general , the reduced density matrix is not of the form @xmath177 , corresponding to the uncoupled case . in the same basis @xmath145 of eq .
( [ eq : rho ] ) , this canonical @xmath178 reads : @xmath179 with @xmath180 $ ] .
the analysis carried out in the previous section shows that , at high temperature , the thermodynamic behavior of the system is not affected by the coupling to the bath .
this suggests that at high temperature the reduced density matrix @xmath136 should tend to the uncoupled system density matrix @xmath178 .
this is indeed the case .
taylor expansion around @xmath181 reveals that @xmath136 and @xmath178 coincide up to second order in @xmath182 .
the ( possible ) appearance of thermodynamic anomalies due to a violation of the the usual _ weak coupling _
assumption have been highlighted .
negative specific heats and entropies were reported already in the literature @xcite , here , for the first time , we described anomalies of the susceptibility .
apart from the fact that the susceptibility may become negative , its value can also be enhanced by several orders of magnitude , in the strong coupling regime .
the coupling strength of the xy model is determined by the exchange energy @xmath48 .
this energy has to be compared with thermal energy and the bulk energy of the system , which grows with the system size @xmath17 .
therefore , the anomalies may only emerge for relatively small systems at sufficiently low tempratures . for large systems and at high temperatures , the canonical density matrix @xmath183/z_s$ ]
describes the equilibrium properties of the system excluding the presence of any anomaly .
this is a well known generic feature of any spatially extended system interacting with its environment by short range interactions @xcite .
the analysis of the reduced density matrix shows that the thermal equilibrium properties of open quantum systems may grossly differ from those resulting from a canonical gibbs state .
this is true even in the thermodynamic limit of the bath @xmath184 , as long as the coupling remains strong , ( @xmath185 is sufficiently large and @xmath17 sufficiently small ) . in regard to the case @xmath91
, we found that the ground state changes from a pure to a mixed state , a situation akin to a _
quantum phase transition _ @xcite .
this transition takes place at the same value @xmath186 of the magnetic field at which the usual phase transition in the closed isotropic xy model , is observed .
finally we note that for the even smaller system composed of a single spin the above described anomalies continue to exist .
we have restricted ourselves to the properties of spin chains and therefore disregarded the case of an open single spin system .
the authors thank peter hnggi for his constant support , for many original ideas that he has generously shared with us and for the lively scientific atmosphere at his `` lehrstuhl '' .
we wish him many active years to come .
pt would like to take this opportunity to once again express his thanks for ref .
@xcite which he was given as a birthday present and which triggered his interest and the interest of many others in the thermodynamics of small quantum systems .
p.t gratefully acknowledge financial support by the dfg via the collaborative research center sfb-486 , project a-10 , via the project no .
1517/262 , by the german excellence initiative via the _ nanosystems initiative munich _ ( nim ) and by the volkswagen foundation ( project i/80424 ) .
is indebted to jos luis garca - palacios for useful discussions and acknowledges financial support from fis2008 - 01240 and fis2009 - 13364-c02 - 01 ( micinn ) .
in this appendix we briefly review the solution of the isotropic xy model , hamiltonian ( [ xymodel ] ) @xcite .
new operators are defined by means of the jordan - wigner transformation @xcite : @xmath187 where @xmath188 is the @xmath43 spin lowering operator @xmath189 the operators @xmath190 are fermionic operators satisfying the canonical anti - commutation rules : @xmath191 in terms of these operators the hamiltonian ( [ xymodel ] ) is expressed as : @xmath192 next , following known procedures , yet new operators are defined , via the discrete sine fourier transform : @xmath193 the new operators also obey canonical fermionic anti - commutation rules : @xmath194 expressing the hamiltonian in terms of the @xmath195 s , one gets : @xmath196 _ i.e. _ a free fermionic hamiltonian , with single mode energies @xmath197 , eq .
( [ eq : lambda ] ) . denoting the fock state associated to the operators
@xmath198 as @xmath199 , that is : @xmath200 with @xmath201 , the eigenvalues of the hamiltonian read : @xmath202 | the exactly solvable model of a one dimensional isotropic xy spin chain is employed to study the thermodynamics of open systems . for this purpose the chain is subdivided into two parts , one part is considered as _ the system _ while the rest as _ the environment _ or _ bath_. the equilibrium properties of the system display several anomalous aspects such as negative entropies , negative specific heat , negative susceptibilities in dependence of temperature and coupling strength between system and bath .
the statistical mechanics of this system is studied in terms of a reduced density matrix . at zero temperature and for a certain parameter values
we observe a change of the ground state , a situation akin to a quantum phase transition . |
thermally assisted magnetization reversal has been the subject of intensive theoretical and experimental study for many decades . aside from the relevance of the subject to emerging magnetic technology such as heat - assisted magnetic recording @xcite and thermal stability of magnetic random access memory @xcite , fundamental physics of magnetization reversal process driven by white - noise
is very rich .
classical transition - rate theory of kramer @xcite has supplied a framework in understanding thermal activation of a single domain magnetic element @xcite .
namely , the thermal switching probability @xmath0 can be described by the nel - brown ( nb ) relaxation - time formula , @xmath1 , where the relaxation time is @xmath2 , @xmath3 is an attempt frequency , @xmath4 is the energy barrier , and @xmath5 is the temperature . for a multi - domain structure , the energy surface becomes extremely complicated and identifying energy barriers are numerically non - trivial . nevertheless , with recent development of micromagnetic modeling
, one can understand thermal reversal reasonably well for a not - too - complicated structure @xcite .
an implicit and yet critical assumption in the nb theory is that magnetization dynamics is governed by a torque from an effective magnetic field @xmath6 , where @xmath7 is the total magnetic energy , i.e. , the effective field is derivable from the derivative of an energy function with respect to the magnetization vector .
therefore , an energy barrier is well - defined in the nb relaxation formula .
if the torque is not derivable from an energy function , one would expect breakdown of the nb formula .
recently , a new class of torques , called spin transfer torque ( stt ) , has been proposed @xcite and verified experimentally @xcite .
stt is derived from a spin polarized current in magnetic multilayers . for a spin valve structure ,
stt is written as @xcite @xmath8 where @xmath9 represents the strength of stt ; it is proportional to the current density .
@xmath10 is gyromagnetic ratio , @xmath11 is a _ unit _
vector representing the direction of the magnetization of the pinned layer , @xmath12 is the magnetization vector of the free layer and @xmath13 is the saturation magnetization .
if we define an effective field @xmath14 from stt , it is evidently that @xmath15 can _ not _ be written as a total derivative of a function with respect to the magnetization vector , i.e. , there is no well - defined energy associated with the field @xmath16 .
recent experiments @xcite on the thermal effect of the spin torque had also indicated that the thermally assisted magnetization reversal can not be simply described by the nb formula .
et al _ @xcite found that the activation energy strongly depends on the magnitude as well as the direction of the current . to capture the gloss features of the observed experiments
, they have to introduce an effective temperature unrelated to the true temperature in the nb formula .
their proposed effective temperature was then interpreted via a possible magnetic heating and magnetic excitations from the spin transfer torque .
the current directional dependence of the effective temperature indicated that the heating is not of the ordinary current - induced joule heating .
however , no attempt has been made to mathematically link the effective temperature with the spin transfer torque of eq .
the problem of thermally assisted escape process driven by a non - gradient driven force , not derivable from a potential , is an unresolved outstanding problem in statistical physics .
while there are already some efforts to formulate the escape time in this case , the general conclusion is that the law of escape time lacks universality and a variety of scaling relations exist @xcite .
the standard treatment of the thermal escape problem in the presence of a non - gradient field would start from the fokker - planck ( fp ) equation and one numerically solves for the probability distribution @xcite .
this procedure involves proper averaging over the possible escape trajectories .
let us consider the total work done by the conservative torque and the non - conservative stt along an arbitrary trajectory , @xmath17 where we have assumed that the magnetization vector starts at an initial equilibrium point @xmath18 and reaches to an energy saddle point @xmath19 , and we have defined the energy barrier from the conservative torque @xmath20 .
one immediately realizes that @xmath21 defined above depends on the escape trajectory . in the absence of stt ,
one relies on the assumption that the most probable path of the escape is through a minimum energy barrier , i.e. , @xmath22 would be an energy saddle point . in the presence of stt , such assumption breaks down in general . in the present paper
we do not intend to address the general problem of the thermal escape in an open system , rather we want to answer a very focused question : to what extent one can formulate the thermal agitation in terms of a simplified nel - brown activation process and to what accuracy one can analyze the relevant experimental data through a simple effective formula we will develop in the later sections .
the paper is organized as follows . in sec .
ii , we propose the stochastic landau - lifshitz equation and its corresponding fokker - planck equation in the presence of the current . in sec .
iii , we introduce a stationary solution for the probability density of magnetization by identifying an effective barrier or an effective temperature associated with spin torques . in sec .
iv , we present a numerical calculation to demonstrate the validity of our theory in several realistic cases .
finally , we compare our theory with existing experiments and summarize our theory .
let us first explicitly propose the following generalized stochastic landau - lifshitz ( ll ) equation that describes dynamics of the magnetization vector subject to a stt at finite temperatures @xmath23 + \mbox{\boldmath $ \gamma $ } _ s,\ ] ] where @xmath24 is the damping constant , @xmath25 is the effective magnetic field including the external field , the anisotropy field , the exchange field , and the demagnetization field , and @xmath26 is a fluctuating field with a gaussian stochastic process whose statistical properties are defined as @xmath27 where @xmath28 and @xmath29 are cartesian indices , @xmath30 represents the strength of the thermal fluctuations whose value will be determined later .
@xmath31 denotes an average taken over all realization of the fluctuating field . in the absence of the spin torque , the above equation is the standard stochastic ll equation .
note that we have conveniently dropped the customary renormalized gyromagnetic ratio @xmath32 when compared with the standard landau - lifshitz - gilbert equation .
the critical assumption of our proposed stochastic ll equation , eq .
( 3 ) , is that _ the spin torque does not contain a fluctuating field @xmath26_.
the justification for this choice is that the spin torque comes from the conduction electrons whose transport properties are less affected by thermal fluctuations since the fermi level is much higher than the thermal energy .
therefore , the thermal fluctuation would not appear to affect @xmath9 which represents the strength of the spin torque .
we believe that our proposed stochastic ll equation captures the main random processes induced by thermal fluctuation .
nevertheless , one could , in principle , have introduced a second random field or torques to take into account the fluctuation of the spin torque . in our proposed ll equation , the thermal effect on the spin torque
is only encoded in the dependence of the magnetization vector . to establish the thermal properties from the above stochastic equation
, one must take a proper thermal average .
fortunately , much of theoretical work on the stochastic ll in the absence of the spin torque had been carried out @xcite . here
we will follow and generalize the procedure pioneered by brown @xcite .
we define @xmath33 as a non - equilibrium probability density for magnetic orientation vectors associated with the stochastic eq .
the rate equation for @xmath33 is @xmath34 where the probability current density @xmath35 , @xmath36 is a short notation for the gradient operator on the magnetization vector @xmath37 , and @xmath38 is the diffusion constant whose value is determined by fluctuation - dissipation theorem . in the present case , @xmath38 is related to @xmath30 by @xmath39 .
the above rate equation , eq . ( 5 ) , is a simple statement for the angular momentum conservation : the change of probability density in an enclosed small volume ( first term ) has to be balanced by the net probability in - flowing flux ( second term ) plus the probability density loss via spin diffusion ( third term ) . by inserting eq .
( 3 ) into eq . ( 5 ) , after a straightforward but rather tedious algebra manipulation @xcite , the resulting equation is the fokker - planck equation @xmath40 p \right\ } \end{array}\ ] ] in the absence of the spin torque , i.e. , @xmath41 , the thermal equilibrium distribution density @xmath42 demands to take the form of the boltzmann distribution function , i.e. , @xmath43 where @xmath5 is the temperature and @xmath44 is the energy defined by @xmath45 . inserting this equilibrium @xmath46 into eq . ( 6 ) , one finds that @xmath47 this is the well - known dissipation - fluctuation relation .
we now postulate that _ the fluctuating field is independent of the spin torque_. this hypothesis is consistent with our notion that the spin torque is a deterministic action so that the spin torque does not alter the randomness induced by the thermal fluctuation . with this identification , the stochastic ll equation , eq .
( 3 ) , and its corresponding fokker - planck equation , eq .
( 6 ) , completely determine the dynamics of the magnetization vector at finite temperature @xmath5 .
before we numerically solve the above stochastic ll equation for a number of interesting cases , we should first look for a stationary solution in eq .
( 6 ) , i.e. , @xmath48 is independent of time . without the spin torque ,
this solution is known as the equilibrium boltzmann distribution function mentioned earlier . with the spin torque ,
the system is no more in an equilibrium state because the system is subject to the spin torque and thus it is not a closed system . for an open system
, the law of thermal dynamics does not require the minimum free energy and the concept of thermal equilibrium breaks down . nevertheless , it is still meaningful to obtain a stationary solution @xmath49 where fokker - planck probability density is time - independent @xmath50 .
thus , @xmath51 p_s \right\ } = 0 \end{array}\ ] ] unfortunately , the above eigenstate problem for an arbitrary field @xmath25 is generally difficult to solve .
one can immediately verify that the boltzmann probability density @xmath52 is no more a solution of the above equation . to make progress , we need to consider a special case below .
first , we recall that the nel - brown formula for the thermal agitation is in fact most useful where the energy barrier constructed by @xmath25 is much higher than the thermal energy @xmath53 . in this limit
, the probability density will be very small if the direction of the magnetization vector is away from the energy minimum .
here we should also consider this case .
we now tentatively seek a solution of @xmath49 in the form of @xmath54 where we have introduced an effective temperature @xmath55 that will be determined next . by placing this @xmath49 into eq .
( 8) and by utilizing the identity @xmath56 , we find @xmath57 p_s \right\ } = 0 \end{array}\ ] ] clearly , the above equation does not necessarily have a solution for an arbitrary effective field .
however , as we point out earlier , we have limited ourselves to a high barrier case such that the magnetization vector at the stationary condition is nearly at the direction of @xmath58 . for concreteness ,
let us consider a most experimentally relevant geometry where @xmath59 and @xmath60 where @xmath61 is the external field which is applied at x - direction , @xmath62 is the anisotropy field , and @xmath63 is the demagnetization field perpendicular to the plane of the film . in this case , the energy minimum are at @xmath64 , @xmath65 .
we simply keep the first order in @xmath66 and @xmath67 , and set @xmath68 up to the first order , we find @xmath69 and @xmath70 . by placing these expressions into eq .
( 9 ) and carrying out the divergence @xmath36 for @xmath66 and @xmath67 components we find ( one noticed that the divergence operator on @xmath49 produces higher orders since eq . (
11 ) is already the first order in @xmath66 and @xmath67 , thus it is consistent with our approximation by neglecting the terms of divergence on @xmath49 ) , @xmath71 or @xmath72 where we defined a critical spin torque @xmath73 . coincidentally , this critical spin torque is precisely the minimum spin torque required to switch the magnetization at zero temperature @xcite . we should point out that the concept of the effective temperature introduced here should be understood in terms of the stationary solution of the probability density .
the thermally averaged dynamical variable , for example , the magnetization vector @xmath74 would behave as if the temperature of the system is @xmath55 .
however , the magnetic temperature which is defined through the thermal fluctuation remains to be @xmath5 .
an alternative understanding of this effective temperature is to rewrite the stationary solution by @xmath75 where @xmath76 is an effective activation energy @xmath77 in other word , we can state that the spin torque alters the magnetic energy and thus there will be an effective energy barrier associated with the spin current .
therefore , it is equivalent to think of the effect of the spin torque via the modification of the temperature or of the energy barrier . to conclude this section ,
we have found a stationary solution @xmath49 of the stochastic ll equation . since the life time or the relaxation time @xmath78 is inversely proportional to the probability density @xmath49
, we can write the generalized brown - nel formula below @xmath79 where @xmath3 is an attempt frequency and @xmath4 is the energy barrier .
the stationary solution , @xmath80 $ ] , is based on the assumption of high energy barrier assumption . in general cases , one should start the calculation of magnetization dynamics from our generalized ll equation , eq . ( 3 ) . since the stationary solution is simple and easy to use in analyzing experimental data , it would be necessary to establish the range of validity of eq .
( 15 ) for various interesting experimental situations .
once its validity is established , we expect that our stationary solution would be serving as a starting point to understand various thermal agitation phenomena in the present of the current . in this section
, we numerically integrate eq .
( 3 ) and compare the result with @xmath49 . as a function of the inverse of the temperature for several values of the spin - transfer torques .
the external field of @xmath81 ( oe ) is applied along @xmath82 direction.,width=264 ] the standard white - noise spectrum , eq .
( 4 ) where @xmath30 is given by eq ( 7 ) , is used for the modeling of the temperature dependence of random fields .
the calculation procedure is same as that for the standard llg equation , except a spin torque is added to the equation of the motion . a magnetic layer ,
whose lateral size is 64nm@xmath8364 nm and whose thickness is 2.5 nm , is treated as a single macrospin .
the in - plane uniaxial anisotropy field @xmath62 is @xmath84(oe ) and the saturation magnetization is @xmath85 ( oe ) .
these parameters are reasonably consistent with the experiments performed by the cornell group @xcite .
the gilbert damping constant was taken as @xmath86 throughout the modeling .
the magnetization of the free layer is initially saturated at @xmath87 direction . at @xmath88
, we apply a magnetic field at @xmath82 direction whose magnitude is close to but less than the anisotropy field @xmath89 . at the same time , @xmath90 is also applied to the system .
equation ( 3 ) was numerically integrated in time using the stochastic heun method with a 0.3ps time step . a smaller time step has been tested and it yields nearly identical results in all the cases presented in the paper . with above specified parameters and procedure ,
we first determine the probability @xmath0 that the magnetic layer has been reversed within the waiting time @xmath91 . by performing up to @xmath92 independent runs for each set of parameters ( each run starts at @xmath88 and ends at the time that the magnetization has either been just switched or ends at the time up to @xmath93 , whichever is smaller ) and then by recording the number of them that the magnetization is switched at time interval ( @xmath91 , @xmath94 ) , we obtained a simulated @xmath0 that is fitted by a simple exponential function , i.e. , @xmath95 where @xmath78 is the fitting parameter for the relaxation time .
we have found that the fit works remarkably well for any values of @xmath9 we have considered . in fig . 1 , we show the fitted relaxation time @xmath78 as a function of the temperature for a fixed applied field @xmath96 .
two features are immediately seen : first the data fall on a straight line for any fixed @xmath9 ; this indicates the thermally assisted reversal can be described by an activation process , i.e. , @xmath97 .
therefore , it is meaningful to introduce an effective activation energy , see eq .
( 14 ) , in accordance with the nel - brown law of thermal agitation .
the second conclusion is that the slope , or the activation energy depends on stt : the positive @xmath9 favors a lower energy barrier .
all these features are well described by eqs .
( 13)-(15 ) . as a function of @xmath90 at three different external fields .
inset : @xmath98 vs @xmath99 for several different values of stt.,width=264 ] in fig . 2 ,
the effective activation energies are shown to be linearly dependent on the current and they vanish at almost the same point @xmath100 for different external fields [ note that @xmath100 is weakly dependent on the external field , see the definition of @xmath100 after eq .
( 13 ) ] . in the insert of fig . 2 , we have shown the activation energy as a function of the magnetic field for several different stt . the activation energy can be fitted by @xmath101 where @xmath102 is the switching field at zero temperature , @xmath103 is the energy barrier at zero magnetic field field , and @xmath104 is a constant , which has been argued to be 1.5 or 2 .
the exponent @xmath105 for the external field applied parallel to the easy axis .
these simulated results confirm our analytical result , eq .
a number of experiments on spin torque induced thermal agitation had been carried out .
it would be interesting to see whether our prediction , eq .
( 15 ) , agrees with these existing data .
the phenomenon that we want to compare first is so - called `` telegraph noises '' or dwell times .
experimentally , one simultaneously applies an external magnetic field and a spin current to a spin valve structure so that the magnetization direction of the free layer is fluctuating from one direction to another due to thermal agitation @xcite . the dwell time @xmath106 @xmath107 is defined as an average time the magnetization of the free layer is parallel ( antiparallel ) to that of the fixed layer . in general , @xmath108
however , by adjusting the magnetic field or the spin current , one is able to obtain an equal dwell time for parallel and antiparallel states , @xmath109 . from eq .
( 15 ) for the parallel and the antiparallel states , the condition of the equal dwell time is @xmath110 where @xmath111 and @xmath112 are the critical currents for the magnetization switching from antiparallel to parallel alignments and vice versa ; their magnitudes are not necessary the same , i.e. , @xmath113 in a typical experimental geometry @xcite . to compare our prediction , eq .
( 17 ) , with experiments , we plot the h - i phase diagram of equal dwell time @xmath109 in fig . 3 . .
the coupling field is @xmath114 , the critical currents are @xmath115 for the transition from ap to p alignments and @xmath116 from p to ap alignments .
line : eq .
( 17 ) except @xmath61 being replaced by @xmath117 .
square : experimental data @xcite.,width=264 ] it is noted that we have shifted the external field by @xmath118 to take into account of the magnetic coupling between the free and fixed layers .
the coupling may come from either the exchange or dipolar couplings . since we assume that the magnetization of the fixed layer is held at the direction of @xmath119 , the free layer receives an effective coupling field that will be added to @xmath61 in eq .
evidently , the agreement between our results and the experimental data is excellent @xcite .
next we compare the ratio of the dwell times of antiparallel and parallel states for a fixed magnetic field .
again , from eq .
( 15 ) , we have @xmath120 .
the parameters are same as those in fig.4.,width=264 ] in fig . 4
, we show the ratio of the dwell times for a fixed magnetic field as a function of the spin torque by using the same set of experimental parameters as in fig .
3 . the results are consistent with experimental data ( however , the data points in ref.@xcite are rather scattered so that i do not include those data in the figure ) . up till now
, we have studied the thermal activation by _ abruptly _ introducing an external field and a stt at t=0 . in experiments ,
there are ramping times , e.g. , stt is gradually increasing at rate of , say @xmath121 @xcite and care must be taken when one compares our theory , eq .
( 15 ) , with experiments . in the current ramping period ,
@xmath9 is not a constant and thus the activation energy , eq .
( 16 ) , varies with time . in this case
, one should utilize the differential form of the switching probability instead of @xmath122 , @xmath123 by assuming a linear ramping of stt , i.e. , @xmath124 is a constant and by placing eqs . ( 15 ) and ( 16 ) into eq .
( 19 ) , we integrate eq .
( 19 ) from @xmath88 to @xmath125 and find the average switching stt @xmath126 @xmath127\ ] ] where @xmath128 $ ] .
the variance of the switching stt is found as @xmath129 these relations , eqs .
( 20 ) and ( 21 ) , are consistent with earlier studies on the similar energy barrier formalism for a very different physical system @xcite . on the sweeping rate @xmath130 at finite temperature .
inset : @xmath131 and @xmath132 as a function of temperature for the sweeping rate of stt at @xmath133.,width=264 ] numerically , the finite ramping rate can be rather easily handled . in determining @xmath134 at finite temperature ,
we ramp @xmath9 with a fixed rate . at a certain time , the magnetization vector switches and we record the value of @xmath9 . by repeating the above procedure 800 times , we are able to establish the switching @xmath9 histograms from which the mean switching @xmath131 and its standard deviation @xmath132 are obtained . at temperature between @xmath135 and @xmath136
, we calculated the mean and standard deviation of the distributions at a sweeping rate between @xmath137 and @xmath138 . as expected for a thermally activated process
, @xmath131 increases with decreasing temperature and with increasing sweeping rate .
figure 5 shows temperature and sweeping rate dependence of @xmath131 and @xmath132 .
a logarithmic dependence of @xmath131 on the sweeping rate has been found , which is in a good agreement to eq.(20 ) .
moreover , the inset of fig .
5 describes the temperature dependence of @xmath139 and @xmath132 .
we have verified that the temperature dependence of @xmath132 is a linear relationship and @xmath131 monotonically decreases with increasing @xmath5 .
myers _ et al _
@xcite have discovered that the thermal activation driven by spin torque is qualitatively different from that driven by the magnetic field .
they have suggested an activation energy whose form is similar to ours , except that they have postulated an arbitrary exponent , i.e. , @xmath140 .
although they have stated that @xmath141 might be 1.5 , most of experimental data shown in the paper can be used to determine the value of @xmath142 .
one set of data , fig .
1(d ) of ref .
[ 11 ] , shows that @xmath143 linearly increases with @xmath144 as predicted by our eq .
if one uses different scaling relation , e.g. , @xmath145 , one would obtain @xmath146 that would disagree with the experimental data .
therefore , the existing data support the linear scaling between the activation energy and the spin torque .
in summary , we have extended the law of thermal agitation to include the spin transfer torque driven by the spin polarized current in magnetic multilayers .
although the concept of the energy barrier or the temperature in the nel - brown formula breaks down in the presence of the spin transfer torque , we are able to re - establish the nel - brown formula by properly introducing an activation energy or an effective temperature to replace the true energy barrier or true lattice temperature .
our formalism is further supported by numerical solutions and is in agreement with experimental results . | we propose a generalized stochastic landau - lifshitz equation and its corresponding fokker - planck equation for the magnetization dynamics in the presence of spin transfer torques .
since the spin transfer torque can pump a magnetic energy into the magnetic system , the equilibrium temperature of the magnetic system is ill - defined .
we introduce an effective temperature based on a stationary solution of the fokker - planck equation . in the limit of high energy barriers ,
the law of thermal agitation is derived .
we find that the nel - brown relaxation formula remains valid as long as we replace the temperature by an effective one that is linearly dependent of the spin torque .
we carry out the numerical integration of the stochastic landau - lifshitz equation to support our theory .
our results agree with existing experimental data . |
an appealing physical idea of the weak measurement tomography johansen , lundeen1,lundeen2,elefante , salvail offers a possibility of reconstruction of the wave function in a single experimental setup that involves a specific system - pointer coupling ; the so - called direct state tomography ( dst ) . basically , the scheme consists in successive measurements of two complementary observables of the system , where only the first one is weakly coupled to the measurement apparatus lundeen1,lundeen2,elefante , salvail .
recently , this procedure was generalized to arbitrary coupling strengths @xcite , and it was argued that the strong measurements expectably outperform the weak ones both in precision and accuracy . since in the framework of weak measurements
the efficiency is traded for accuracy , the error estimation analysis becomes vital .
typically , the experimental performance of dst in the case of weak measurements , is compared either with results of strong ( projective ) tomography @xcite .
an alternative method of estimation of the fidelity of a reconstructed state was introduced in @xcite . however , non of the above - mentioned approaches analyzed the global intrinsic error estimation helstrom , englert , checos . in this letter
we show that conveniently reformulating the approach vallone as a mutually unbiased bases ( mub)-like reconstruction scheme in non - orthogonal bases @xcite one can carry out the mean square error ( mse ) analysis of dst , including the weak measurement limit , in the framework of measurement statistics @xcite .
in particular , we exemplify on the single qubit case that non - orthogonal bases appear as effective projective states , in such a way that a weak coupling corresponds to projection into near - parallel bases .
this allows us to reformulate the accuracy analysis in terms of measured probabilities . and
thus , estimate the intrinsic statistical errors finding the minimum mse using the crmer - rao lower bound .
following the general idea of dst we consider an unknown state @xmath0 of the system ( one qubit ) interacting with a pointer ( another qubit ) initially prepared in the eigenstate state @xmath1 of the pauli operator @xmath2 , according to @xmath3where @xmath4 $ ] is the measurement strength .
after the interaction the system is postselected in the state @xmath5 , and the pointer is measured in the bases @xmath6 : @xmath7and the following probabilities are retrieved @xmath8where @xmath9 and @xmath10 are the normalization constants . in the framework of dst @xcite ,
@xcite the wave function is reconstructed in the basis @xmath11 of eigenstates of @xmath12 as a linear combination of the probabilities @xmath13 . on the other hand , the probabilities ( [ pik ] ) can be considered as projections of the initial state @xmath14 into the set @xmath15 so that @xmath16it is worth noticing that similar effective projection states naturally appear in experiments @xcite . explicitly , the effective projection states ( [ psiik ] ) have the form @xmath17where @xmath18 , and satisfy the condition @xmath19defining the so - called equidistant bases @xcite .
introducing @xmath20 we rewrite elements @xmath21 as follows @xmath22for @xmath23 so that @xmath24 and the limit @xmath25 of almost parallel states ( close to @xmath26 ) corresponds to the weak measurement case , @xmath27
. an important feature of the probabilities @xmath13 , @xmath23 is the relation @xcite @xmath28which reflects a statistical dependence on the measurements in the non - orthogonal and computational bases .
the bases ( [ bases ] ) form an informationally complete set for @xmath29 ) according to @xcite @xmath30where @xmath31 , is the corresponding @xmath32-th biorthogonal basis , @xmath33 . in the limit @xmath34 ( @xmath35 )
the expression above is converted to the standard ( orthogonal ) mub tomographic expression @xcite allowing the maximum information gain @xcite .
each projector in the set @xmath36 can be considered as a single output channel of an effective measuring apparatus .
an estimation procedure consists in a repetitive measurement on each of @xmath37 identical copies of the system , i.e. the pointer is postselected in every basis the same number of times , obtaining frequencies @xmath38 , where @xmath39 is the number of projections into @xmath40 .
the corresponding statistics of outcomes is binomial - like @xcite @xmath41where @xmath42 , @xmath43 , @xmath44 and the condition ( [ s ] ) is satisfied . for the computational basis , where @xmath45 , the statistics is obviously binomial .
the expectation values corresponding to the probability distribution ( binomial ) are of the form @xmath46 .
\label{expectation}\ ] ] following general ideas @xcite , we compute the estimation error as the average squared of the hilbert - schmidt distance between the true @xmath0 and estimated @xmath47 system states , @xmath48\rangle . \label{e}\ ] ] it depends on @xmath49 and the inner product between all the projectors appearing in ( [ reconstruccion ] ) .
taking into account ( [ s ] ) we obtain for the difference between true ( @xmath13 ) and estimated ( @xmath50 ) probabilities , @xmath51 , the following relations @xmath52substituting the above relations into ( [ e ] ) and averaging for many repetitions , we obtain for the average quadratic error @xmath53where the explicit form of the coefficients @xmath54 ( in matrix form ) is given in appendix . employing the cramr - rao lower bound we minimize the possible mean square error ( mse ) per trail @xcite @xmath55where @xmath56 $ ] and @xmath57 is the fisher matrix per trail , @xmath58being @xmath59 the likelihood . after straightforward but lengthy calculations ( see appendix ) we find that the lower bound for the estimation error per trial in terms of measured probabilities is given by @xmath60 .
\label{error2}\end{aligned}\]]it is easy to see that at @xmath34 ( corresponding to @xmath35 ) the mean hilbert - schmidt distance for mubs is recovered @xcite .
the lower bound ( [ error2 ] ) can still be averaged over the space of quantum states .
we will consider pure and mixed states separately .
let us first consider an arbitrary pure state @xmath61 with projections @xmath62 and @xmath63 on the basis @xmath64 , that can be taken as @xmath65 due to invariance of the averaging procedure under unitary transformations .
it is straightforward to check that @xmath66thus , the averaged , over the space of pure states , mse takes the form @xmath67where the double brackets mean averaging both over the sample and over the space of states . for @xmath34 , corresponding to the standard mub tomography , @xmath68 checos , while in the limit @xmath69 the lower bound of the mse diverges as @xmath70 , which qualitatively coincides with results of @xcite . in order to average over mixed states we use the eigenvalue distribution based on the bures metric @xcite , @xmath71making use of the spectral decomposition @xmath72 , where the eigenstates can be parametrized as @xmath73with @xmath74 $ ] , and @xmath75 $ ] we perform integration of ( [ error2 ] ) with the measure @xmath76 .
the result of such integration can be found analytically in terms of special functions and studied in the limit cases . due to its cumbersome form
we do not present the explicit expression , but instead plot it in fig .
[ error1qb ] . over a sample of @xmath77 random statesas a fucntion of @xmath78 : ( red ) dashed line for pure states , ( blue ) continuous line for mixed states .
average error for sic - povm does not depend on @xmath79 and is represented as a constant ( red ) dashed line @xmath80 for pure states and ( blue ) continuous line @xmath81 for random mixed states . the vertical line at @xmath82 shows the bound where the @xmath83.,width=302 ] in fig .
[ error1qb ] we plot @xmath84 , where the average is taken for a sample of @xmath77 pure and mixed random states following the routine introduced in @xcite . for pure states
, the plot of the square root of equation ( [ erroran ] ) perfectly coincides with the numerical results .
the mixed states are produced according to the bures metric .
as it is expected , the best estimation is obtained for mub tomography , with @xmath85 for pure states , and @xmath86 for mixed states .
one can also clearly see that the stronger the measurements are , the smaller the estimation errors are .
the performance of dst can be also compared with a tomographic scheme based on symmetric informationally complete positive operator valued measure ( sic - pomv ) measurements @xcite . for a single qubit
a set of projectors @xmath87 such that @xmath88 and @xmath89 span the density matrix@xmath90where the probabilities @xmath91 are the outcomes associated with measurement of the operator @xmath92 , @xmath93 .
the corresponding mse lower bound has the form ( [ mse ] ) , where the components of the matrix @xmath94 are @xmath95 , @xmath96 , and the fisher matrix elements per trail are @xmath97 which leads to @xmath98 for pure states @xcite . in fig.[error1qb ] we plot @xmath99 for sic - povm tomography as ( red ) dashed constant line for pure states and as a ( blue ) continuous constant line for mixed states , produced according to the bures metric , obtaining in this case @xmath100 by averaging over @xmath77 randomly generated states .
one can observe that dst outperforms sic - povm qubit tomography for @xmath101 , which is indicated in fig .
[ error1qb ] as a vertical ( magenta ) dotted - dashed line .
we have shown that the performance of the dst protocol can be analyzed in a similar way as in the standard projection - based reconstruction schemes . in the framework of our approach
we have been able to determine the estimation error for any measurement strength , including the weak measurement case . in addition , an explicit analytic form for the minimum square error have been found in the pure and mixed states .
the proposed scheme can be extended to higher dimensions and composite many - particle systems .
in this appendix we briefly deduce eq .
( [ error2 ] ) . taking into account the overlaps @xcite @xmath102 , @xmath103 into ( [ e ] ) and ( reconstruccion ) one arrives to @xmath104 the fisher matrix ( per trial ) ( [ fisher ] )
is obtained directly form the likelihood @xmath105 in particular , one has @xmath106where the two first terms correspond to the mub tomography @xcite , the third term appears due to dependence of the sum of probabilities in the non - orthogonal bases on @xmath107 , and the last term comes form the normalization factor @xmath108 .
the main difference with the mub case consists in appearing elements in @xmath57 outside of the main diagonal , which is a consequence of the relation ( [ s ] ) : @xmath109for the non - orthogonal bases @xmath110 , the elements are similar to the mub case , normalized by the factor @xmath111 : @xmath112substituting the explicit forms of @xmath94 and @xmath57 into ( [ mse ] ) one obtains ( [ error2 ] ) .
j. erhart , _ et al _ , nat
* 8 * , 185 ( 2012 ) ; l.a .
rozema , _ et al _ , phys .
. lett . * 109 * , 100404 ( 2012 ) ; s .- y .
baek , _ et al _ , sci . rep . * 3 * , 2221 ( 2013 ) ; m. ringbauer , _ et al _ , phys .
lett . * 112 * , 020401 ( 2014 ) .
miszczak , int . j. mod
c * 22 * , 897 - 918 ( 2011 ) ; j.a .
miszczak , z. puchala , and p. gawron , _ qi : quantum information package for mathematica _ , http://zksi.iitis.pl/wiki/projects:mathematica-qi ( 2010 ) . | we show that reformulating the direct state tomography ( dst ) protocol in terms of projections into a set of non - orthogonal bases one can perform an accuracy analysis of dst in a similar way as in the standard projection - based reconstruction schemes .
i.e. in terms of the hilbert - schmidt distance between estimated and true states .
this allows us to determine the estimation error for any measurement strength , including the weak measurement case , and to obtain an explicit analytic form for the average minimum square errors . |
a knot @xmath0 in @xmath1 is a _ ribbon knot _ if it bounds an immersed disk in @xmath1 with only ribbon singularities .
recall that @xmath0 is a _ slice knot _ if it bounds a smoothly embedded 2-disk @xmath2 in @xmath3 .
every ribbon knot is slice .
a notorious question in knot theory is whether the converse is true .
we know all ribbon knots with crossing number @xmath4 10 : there are 21 of them .
symmetric unions yield another way of showing that knots are ribbon .
all 21 ribbon knots with crossing number @xmath4 10 are symmetric unions.@xmath5 as an example we show the symmetric union for @xmath6 which was missing from my list in 1998 .
let @xmath7 be an unoriented knot diagram and @xmath8 the diagram @xmath7 reflected at an axis in the plane .
if , as in fig . [ sudef ] , we insert the tangle @xmath9 and twist tangles @xmath10 we call the result a _ symmetric union _ of @xmath7 and @xmath8 with twist parameters @xmath10 , @xmath11 .
the _ partial knot _
@xmath12 of the symmetric union is the knot given by the diagram @xmath7 . 1 . ) as an example we consider again the diagram in fig .
[ 10_87 ] . here
the partial knot is the knot @xmath13 : 2 . )
if all @xmath14 then we get the well - known symmetric ribbon for @xmath15 .
the properties in theorem [ matrix ] were already known to kinoshita and terasaka in the case that @xmath16 ( only one twist tangle inserted ) .
the theorem is proved for instance by using the matrix definition of the alexander polynomial and the goeritz matrix .
[ matrix ] the alexander polynomial of a symmetric union depends only on the parity of the twist numbers @xmath10 .
the determinant of a symmetric union @xmath0 is independent of the twist numbers and equals the square of the determinant of the partial knot : @xmath17 . @xmath5 analyzing the knot groups ( modulo squares of meridians ) of @xmath0 and @xmath12 we can prove : if the partial knot of a symmetric union @xmath0 is non - trivial , then @xmath0 is non - trivial.@xmath5 partial knots are not unique : different partial knots can result in the same symmetric union .
in this talk we are interested mainly in the ribbon property of symmetric unions : symmetric unions are ribbon knots.@xmath5 the proof uses the same construction as for @xmath18 with additional twists in the ribbon , see fig .
[ bandbeisp ] .
* * knots with @xmath19 or @xmath20 : * the knot @xmath21 and the kinoshita - terasaka knot are symmetric unions of the trivial knot , see fig .
hence they have determinant 1 . for the kinoshita - terasaka knot
the twist parameter is an even number , hence its alexander polynomial is equal to 1 . +
* * families of knots with the same polynomial invariants : * t.kanenobu used in 1986 symmetric unions to construct knots with the same jones or homfly polynomial . in 2004 he used the following chiral knot in order to present a knot whose chirality is not detected by the links / gould invariant . + * * tanaka s example : * toshifumi tanaka used in 2004 a symmetric union to construct a counterexample for a conjecture of t.fiedler on the jones polynomial of ribbon knots , see fig .
recall the definition of the bridge number of a knot : if @xmath22 is a unit vector in @xmath23 and @xmath0 is a knot , then let @xmath24 be the number of maxima of the orthogonal projection of @xmath0 on the line spanned by @xmath25 .
then the _ bridge number _ of @xmath0 is : @xmath26 because knots with bridge number 1 are trivial , the simplest cases occur for bridge number 2 .
they were studied by * bankwitz / schumann ( viergeflechte , 1934 ) , * schubert ( knoten mit 2 brcken , 1956 ) and * conway ( rational knots , 1970 ) .
their double branched coverings are lens spaces @xmath27 with @xmath28 equal to the determinant of the knot . in the plait normal form @xmath29
, also called _ conway notation _ , the numbers of half - twists @xmath30 are related to the parameters @xmath28 and @xmath31 by the continued fraction @xmath32 = \frac{p}{q}$ ] .
we sometimes call @xmath33 the `` fraction '' of @xmath0 .
the following theorem gives a necessary condition for 2-bridge ribbon knots .
let @xmath0 be a ribbon knot with @xmath34 and double branched covering @xmath35 .
then : @xmath36.@xmath5 here @xmath37 denotes the triangle @xmath38 .
the value of @xmath39 is computed by counting lattice points similar to pick s formula : interior points count as 1 , boundary points as 1/2 and vertices different from ( 0,0 ) as 1/4 . in the example of figure [ gordon ] - for @xmath40
the area of the triangle is 23 and the value of @xmath39 is @xmath41 - for @xmath42 these values are 92 and @xmath43 .
all knots found to satisfy the necessary condition are the following 2-bridge knots ( and their mirror images ) : * family 0 : @xmath44 , with parameters @xmath45 , * family 1 : @xmath46 , with @xmath47 , * family 2 : @xmath48 , with @xmath47 , and each knot in this list is a ribbon knot .
hence this is a complete list of known 2-bridge ribbon knots .
it is a conjecture ( since 1974 ) that there are no other 2-bridge ribbon knots . 1 . )
amphicheiral 2-bridge knots have ( an even length ) palindromic conway notation : for instance @xmath49
. hence for each @xmath50 there are exactly as many knots of family 0 with minimal crossing number @xmath51 as there are amphicheiral 2-bridge knots with minimal crossing number c. 2 . )
families 1 and 2 are not disjoint : knots of family 1 with @xmath52 equal those of family 2 with @xmath52 or @xmath53 .
our main theorem is : [ maintheorem ] every knot contained in one of the three families is a symmetric union .
proof : the proof is given by the following deformations of knot diagrams .
for family 0 we have @xmath54 which is a symmetric union ( this family was already considered by kanenobu in 1986 ) . for families 1 and 2
we have the following diagrams : @xmath5 the surprise is that it is necessary to insert extra 2a/-2a twists to obtain the symmetric diagram for family 1 .
there is another characterisation of the three families ( this is exercise 18 e ) in the unpublished notes of l.siebenmann : _ exercices sur les noeuds rationnels _ ) : [ fractions ] the following conditions for a 2-bridge knot @xmath0 with double branched cover @xmath35 are equivalent : * @xmath0 is contained in one of the three families of 2-bridge ribbon knots of casson and gordon . * the parameter
@xmath31 satisfies * * @xmath55 , with @xmath56 , or * * @xmath57 , with @xmath58 , or * * @xmath57 , with @xmath59 , @xmath60 odd , or * * @xmath61 , with @xmath62 , @xmath63 odd .
@xmath5 * example : * several of the conditions in b ) can apply . for the knot
@xmath64 we have simultaneously @xmath65 as in ii ) and @xmath66 as in iv ) .
this description allows an elegant computation of the partial knot s fraction in the symmetric union representation in theorem [ maintheorem ] .
[ partialfraction ] let @xmath0 be a knot contained in one of the three families and @xmath12 be its partial knot as constructed in theorem [ maintheorem ] . if @xmath67 then @xmath68 where @xmath60 is as defined in lemma [ fractions ] . different possibilities for @xmath60 result in the same partial knot.@xmath5 * example : * by ii ) for @xmath64
we get the partial knot @xmath69 . by iv )
we get the partial knot @xmath70 .
these fractions denote the same knot @xmath71 .
* questions : * is there a geometric explanation for the result in theorem [ partialfraction ] ? are partial knots unique for 2-bridge ribbon knots ?
[ conj1 ] every 2-bridge ribbon knot is contained in one of the three families of casson and gordon . for the number of 2-bridge ribbon knots for each crossing number @xmath74 we found ( knots which belong to both family 1 and family 2 are shown together with family 1 ; the list can be extended to @xmath75 ) : 1 .
use eisenstein method mentioned in the article of siebenmann to improve computation .
2 . all ribbon knots and symmetric unions with minimal crossing number @xmath76 are known .
extend this list to knots with minimal crossing number 11 ( is conway knot ribbon ? ) .
3 . prove conjecture [ conj1 ] .
4 . prove conjecture [ conj2 ] .
casson , a./gordon , c.mca .
: _ cobordism of classical knots _ , a la recherche de la topologie perdue , ed . by guillou and marin ,
progress in mathematics , volume 62 , 1986 .
kawauchi , a . : a survey of knot theory , birkhuser , 1996 .
lamm , c . : _
symmetric unions and ribbon knots _ , osaka j. math . *
37 * , 537 - 550 ( 2000 ) .
siebenmann , l . : _
exercices sur les noeuds rationnels _ , unpublished notes ( 1975 ) . | symmetric unions have been defined as generalizations of kinoshita - terasaka s construction in 1957 .
they are given by diagrams which look like the connected sum of a knot and its mirror image with additional twist tangles inserted near the symmetry axis . because all symmetric unions are ribbon knots , we can ask how big a subfamily of ribbon knots they form .
it is known that all 21 ribbon knots with crossing number less or equal 10 are symmetric unions . in this
talk we extend our knowledge about symmetric unions : we prove that the family of symmetric unions contains all known 2-bridge ribbon knots .
the question , however , whether the three families of 2-bridge ribbon knots , found by casson and gordon in 1974 , are a complete list of all 2-bridge ribbon knots , is still open . 0.5 cm |
the formation of primordial black holes ( pbhs ) has been postulated in many theories of the early universe ( for a recent review see ref . ) .
black holes of mass @xmath1 continually emit hawking radiation @xcite with a temperature of @xmath2 in the form of all available fundamental particle species .
the emitted particles decay quickly on astrophysical timescales into @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 , @xmath8 and @xmath9 .
pbhs with an initial mass@xcite of @xmath10 g should be expiring today with a burst of high energy particles including gamma - rays .
the current upper limit on the number expiring today per volume per unit time is@xcite @xmath11 where @xmath12 is the density enhancement of pbhs in the local region .
typically @xmath12 is @xmath13 ( for clustering in the galactic halo ) or larger .
such pbh bursts may be detectable by the fermi gamma - ray space telescope observatory s large area telescope ( lat ) .
conversely , non - detection by the lat may lead to tighter bounds on the pbh distribution .
in the standard model@xcite , the total number of photons emitted per second by a @xmath14 gev black hole scales as@xcite @xmath15 the number of photons per second per unit area reaching the earth from a pbh bursting at a distance @xmath16 from the earth is @xmath17 let us assume an ideal detector of effective area @xmath18 which can detect every photon that falls on it .
( if the detector is non - ideal then the efficiency can be incorprated into the value of @xmath18 . ) if the detector requires @xmath19 photons over time @xmath20 to distinguish an incoming event as a burst , then to detect a burst we require @xmath21 .
that is the pbh must be closer than @xmath22 what @xmath23 maximizes the chance of detection ?
the remaining lifetime@xcite of a pbh of temperature @xmath23 is @xmath24 s where @xmath25 weights the number of emitted species .
( the remaining lifetime of a 300 gev , 1 tev , or 5 tev black hole is 1 hour , 1 minute , or 1 second , respectively . )
taking @xmath26 , a pbh will be detected by the ideal detector if it is closer than @xmath27 thus the detectability is maximized for the lowest @xmath23 black hole visible above the background and/or by using the longest detector exposure time . for a detector of angular resolution @xmath28 to resolve the pbh above the gamma - ray background @xmath29
, we also require that @xmath30 .
the pbh will be resolved above the observed ( egret ) extragalactic background@xcite @xmath31 at energy @xmath32 by the ideal detector if the pbh is closer than @xmath33 and @xmath32 is less than the average energy@xcite of the pbh photons @xmath34 gev . the isotropic diffuse gamma - ray background , which is an upper limit on the extragalactic background , recently measured@xcite by the lat at mid - galactic latitudes is consistent with the earlier egret measurements eq . ( [ aba : eq6 ] ) ,
although the extragalactic component may@xcite be weaker by up to a factor of 2 . for a given detector , the scanned volume of space is then @xmath35 where @xmath36 is the detector acceptance angle ( field of view ) and @xmath37 .
extensive air shower arrays characteristically have @xmath38 , large @xmath36 and small @xmath28 but very high threshold energy ( typically @xmath39 tev ) and hence are background - limited , while atmospheric cerenkov detectors@xcite characteristically have @xmath40 and small @xmath28 but high threshold energy ( typically @xmath41 gev although the whipple sgarface system@xcite has a threshold of 100 mev ) and very small @xmath36 ( @xmath42 sr ) .
in contrast , the fermi lat has@xcite a smaller @xmath43 but large @xmath44 sr , finer source position angular resolution ( @xmath45 ) , low energy thresholds ( down to 20 mev ) , good time resolution and is essentially background - free with respect to burst sensitivity .
additionally , most of the photons emitted by expiring @xmath46 tev pbhs are in the lat energy range ( 20 mev - 300 gev ) .
the fermi lat offers greater sensitivity to local pbh bursts than ground - based detectors .
we have proposed@xcite spectral lag measurements ( the temporal delay between high and low energy pulses ) of the incoming light curve in two different energy bands as a method to identify pbh bursts .
a pbh burst arriving at the detector will exhibit positive to negative evolution with increasing energy because the black hole temperature and @xmath47 increase over time as the black hole loses mass .
because spectral lag measurements require counts in only two energy bands , and not the full spectrum , spectral lag can be measured even for weak events that last for very short time scales .
work is in progress to calculate quantitative values for the pbh spectral lags for the characteristics of the fermi lat .
9 b. j. carr , in _ inflating horizons of particle astrophysics and cosmology _ , ( universal academy press inc , tokyo , 2006 ) . s. w. hawking , _ nature _ * 248 * , 30 ( 1974 ) ; s. w. hawking , _ commun .
* 43 * , 199 ( 1975 ) ; s. w. hawking , _ phys .
d _ * 13 * , 191 ( 1976 ) . j. h. macgibbon , b. j. carr , and d. n. page , _ phys . rev
. d _ * 78 * , 064043 ( 2008 ) .
j. h. macgibbon and b. j. carr , _ astrophys .
j. _ * 371 * , 447 ( 1991 ) .
j. h. macgibbon and b. r. webber , _ phys .
d _ * 41 * , 3052 ( 1990 ) .
j. h. macgibbon , _ phys .
d _ * 44 * , 376 ( 1991 ) .
p. shreekumar _
_ , _ astrophys
. j. _ * 494 * , 523 ( 1998 ) .
a. a. abdo _
_ , _ astrophys .
j. _ * 703 * , 1249 ( 2009 ) .
a. w. strong , i. v. moskalenko , and o. reimer , _ astrophys .
j. _ * 613 * , 956 ( 2004 ) .
l. a. antonelli _
_ , _ the next generation of cherenkov telescopes . a white paper for the italian national institute for astrophysics ( inaf )
_ arxiv:0906.4114 .
s. lebohec , f. krennrich , and g. sleege _ astropart .
phys . _ * 23 * , 235 ( 2005 ) .
w. b. atwood _
_ , _ the large area telescope on the fermi gamma - ray space telescope mission _ arxiv:0902.1089
. t. n. ukwatta _
_ , _ aip conf
* 1133 * , 440 ( 2009 ) . | primordial black holes ( pbhs ) , which may have been created in the early universe , are predicted to be detectable by their hawking radiation .
the fermi gamma - ray space telescope observatory offers increased sensitivity to the gamma - ray bursts produced by pbhs with an initial mass of @xmath0 g expiring today .
pbhs are candidate progenitors of unidentified gamma - ray bursts ( grbs ) that lack x - ray afterglow .
we propose spectral lag , which is the temporal delay between the high and low energy pulses , as an efficient method to identify pbh evaporation events with the fermi large area telescope ( lat ) .
unf - astro-3 - 1 - 10c |
abelian dominance and monopole dominance in ma gauge suggest the existence of an effective monopole action also in @xmath0 qcd .
there are two independent ( three with one constraint @xmath1 ) currents in the case of @xmath0 .
applying the same inverse monte - carlo method developed in the case of @xmath2 qcd@xcite , we have derived the infrared effective monopole action starting from the vacuum ensemble \{@xmath3 } defined from the thermalized abelian link fields after the ma abelian projection@xcite .
effective monopole actions can be derived similarly for the blocked monopole currents @xmath4@xcite .
this corresponds to the block - spin transformation of the monopole currents on the dual lattice .
the form of the action is restricted to 27 two - point interactions up to @xmath5 distance between monopole currents and the most leading 4 and 6 point interactions .
the lattice size is @xmath6 for @xmath7 .
the results are as follow : 1 .
the monopole actions are fixed clearly when we pay attention to two independent monopole currents : @xmath8 as well as one kind of monopole currents : @xmath9 .
2 . the scaling @xmath10 for fixed @xmath11 looks good both for one - current and two - current cases . ( fig.[g1 ] ) 3 .
monopole condensation occurs in the large @xmath12 region .
entropy dominates over the energy : ln7@xmath13 in the one - current case .
( fig.[g1 ] ) 4 .
the quadratic monopole actions are dominant for large @xmath12 in both cases .
= 0.4 in the infrared regions , quadratic monopole interactions are dominant .
when we restrict ourselves to quadratic monopole interactions alone , the block - spin transformation can be done analytically as done in @xmath2 qcd@xcite .
we can obtain also the perfect operator to evaluate the static quark - antiquark potential @xmath14 on the coarse @xmath12 lattice . in the framework of the string model ,
the quantum effects are seen to be small in the infrared region and the static potential is evaluated by the classical contributions alone : @xmath15 where @xmath16 and @xmath17 is the source term corresponding to the wilson loop .
we get the rotational invariance @xmath18 and @xmath19 , where @xmath20 and @xmath21 are dimensional constants parametrizing the monopole action on the fine @xmath22 lattice .
@xmath23 is equivalent to the result from dgl theory@xcite .
abelian and monopole dominances are shown numerically in ma gauge .
however , monopole condensation after abelian projection in some gauges _ can not explain color confinement_. abelian neutral but color non - singlet states are not confined .
* gauge independence of monopole condensation is essential for color confinement . * here we prove the gauge independence of the monopole dominance if abelian dominance is gauge independent .
first we note that abelian wilson loops without any gauge fixing are known to give the same string tension @xcite as that from non - abelian wilson loops in the strong coupling expansion .
recently , ogilivie@xcite proved the same fact for any @xmath24 using the character expansion@xcite .
abelian dominance comes from the elitzur s theorem@xcite and then it must be gauge independent .
abelian dominance leads us to abelian effective action .
we express the abelian action in terms of the villain form @xmath25}.$ ] the general villain action can be expressed as follows : @xmath26 } \\ & & \hspace*{-.5cm}=e^{-f^{'}[-i\delta/\delta b ] } { \rm e}^{-(d\theta + 2\pi n , d(d\theta + 2\pi n ) ) + i(b , d\theta + 2\pi n)}\bigg|_{b=0},\end{aligned}\ ] ] where @xmath27=[d,\delta]=0 $ ] are satisfied in the large @xmath24 scaling region . the abelian wilson loop @xmath28 is estimated with this action , where @xmath29 is the electric current .
when we use the bkt transformation@xcite , we get an action in terms of monopole currents : @xmath30}\sum_{k\in { \bf z } , dk=0 } { \rm e}^{-\frac{1}{4 } ( \delta b , ( \delta d ) ^{-1 } \delta b ) } \\ & & { \rm e}^{\frac{1}{2 } ( i b , 4\pi \delta \delta^{-1}k + i ( \delta d ) ^{-1 } dj ) } { \rm e}^{-4\pi^{2 } ( k , \delta^{-1}d k ) } \\ & & { \rm e}^{2\pi i ( \delta \delta^{-1}k , s ) } { \rm e}^{- \frac{1}{4 } ( j , ( \delta d ) ^{-1 } j ) } \bigg|_{b=0 } \end{aligned}\ ] ] 1)electric - electric current @xmath31 interactions ( with no monopole @xmath32 ) come from the exchange of regular photons and have no line singularity leading to a linear potential .
hence the linear potential of abelian wilson loops is due to the monopole contribution alone .
* monopole dominance is proved from abelian dominance . *
2)the linear potential comes only from @xmath33 .
the surface independence of the static potential is assured due to the 4-d linking number .
to confirm the above results numerically , we analyze gauge dependence of abelian projection by langevin equation with stochastic gauge fixing term@xcite , @xmath34 where @xmath35 is fictious time , @xmath36 is gaussian white noise and @xmath37 is gauge fixing term .
@xmath38 ( @xmath39 ) corresponds to the ma gauge fixing ( no gauge fixing ) . as a first test
, we perform numerical simulations on a @xmath45 lattice with @xmath46 , @xmath47 , and @xmath48 . we show the abelian heavy quark potentials for different @xmath49 in fig.[pot - abel ] together with that of non - abelian potential , which are essentially same from @xmath38 to 2.5 , and show the confinement linear potential behavior .
note that the string tension obtained by the standard gauge fixing procedure is larger than that of @xmath38 .
this is probably due to the gribov copies@xcite .
bali et al .
@xcite have observed that when the number of copies decreases the string tension becomes small .
this is very consistent with our data .
99 h.shiba and t.suzuki , phys .
* b351*(1995 ) 519 .
s.kato @xmath50 @xmath51 , nucl . phys . * b520*(1998 ) 323 .
degrand and d. toussaint , phys . rev .
* d22 * ( 1980 ) 2478 .
ivanenko @xmath50 @xmath51 , phys . lett .
* 252b * ( 1990 ) 631 .
s.fujimoto @xmath50 @xmath51 , in this proceedings .
t. suzuki , prog.theor.phys .
* 80 * ( 1988 ) 929 ; s. maedan and t. suzuki , prog.theor.phys .
* 81 * ( 1988 ) 229 .
t. suzuki and i. yotsuyanagi , phys .
rev . * d42 * ( 1990 ) 4257 m.faber @xmath50 @xmath51 , hep - lat/9810008 .
ogilvie , phys.rev .
d59 ( 1999 ) 074505 .
s.elitzur , phys . rev .
* d12 * ( 1975 ) 3978 .
beresinskii , _ sov .
jetp _ * 32 * ( 1970 ) 493 ; j.m
. kosterlitz and d.j .
thouless , _
j. phys._,*c6 * ( 1973 ) 1181 .
d.zwanziger , nucl .
b192 * ( 1981 ) 259 .
m.mizutani and a.nakamura , vistas in astronomy , * 37 * , 305 , 1993 , pergamon press ; nucl.phys .
b(proc.suppl.)34 , 253 , 1994 .
s.hioki et al . , phys.lett .
* b272 * ( 1991 ) 326 .
g.s.bali et al.,phys.rev .
d54 ( 1996 ) 2863 . | three topics concerning infrared effective lattice qcd are discussed .
( 1)perfect lattice action of infrared @xmath0 qcd and perfect operators for the static potential are analytically given when we assume two - point monopole interactions alone .
the assumption seems to be justified from numerical analyses of pure @xmath0 qcd in maximally abelian gauge .
( 2)gauge invariance of monopole dominance can be proved theoretically if the gauge invariance of abelian dominance is proved .
the gauge invariance of monopole condensation leads us to confinement of abelian neutral but color octet states after abelian projection .
( 3)a stochastic gauge fixing method is developed to study the gauge dependence of the abelian projection , which interpolates between the maximally abelian ( ma ) gauge and no gauge fixing .
abelian dominance for the heavy quark potential holds even in the gauge which is far from maximally abelian one . |
it is found in many cases that hamiltonian chaotic dynamics possesses in many cases a kinetics that does nt obey the gaussian law process and that fluctuations of the observables can be persistent , i.e. there is no any characteristing time of the fluctuations decay .
this type of dynamics can be characterized by the so called polynomial complexity rather than an exponential one .
more accurately , one can introduce some complexity function and entropy function based on the dynamical process of separation of trajectories in phase space by a finite distance during a finite time .
the new approach to the problem of complexity and entropy covers different limit cases , exponential and polynomial , depending on the local instability of trajectories and the way of the trajectories dispersion .
the complexity of dynamical systems begins from the description of chaotic trajectories in phase space ( @xcite ) .
the notion of complexity has a rigorous meaning and it presents a quantity that characterizes systems and can be measured .
in an oversimplified way one can say that the less predictable is a system , the larger complexity should be assigned to the system .
the original version of the dynamics complexity was closely linked to the system s instability and entropy @xcite .
a typical situation of the chaotic dynamics could be associated with a positive kolmogorov - sinai entropy , or be similar to the anosov - type systems .
this type of randomness and complexity of the systems can be characterized by the exponential divergence of trajectories in phase space . as long as investigation of chaotic dynamics reveals new and more detailed pictures of chaos , the simplified old version of the complexity appears to be constrained to be applied to typical systems .
let us mention that the typical hamiltonians do not possess ergodicity , the boundary of islands in phase space make the dynamics singular in their vicinity , and even zero measure phase space domains in the sinai billiard are responsible for the anomalous kinetics @xcite .
attempts to find adequate complexity definitions for realistic chaotic dynamics were subject to many publications and reviews @xcite .
the basic idea of new developments for chaotic systems is to involve a finite time of the systems unstable evolution into a definition of the complexity or entropy . in some sense
, our paper is a continuation of these attempts .
there are numerous observations that the hamiltonian systems referred as the chaotic ones , _ do not have _ exponential dispersion of trajectories for arbitrary long time intervals
. these pieces of trajectories , called flights , appear with a probability that is not exponentially small ( see for example @xcite ) .
a similar type of random dynamics with zero lyapunov exponent appears in billiards and maps with discontinuities ( @xcite ) .
the behavior of systems with zero lyapunov exponents definitely have some level of complexity and some value of entropy in a physical sense , but the same systems with zero lyapunov exponent can not be applied to by the regular notion of the kolmogorov - sinai entropy of the standard definitions of complexity .
the most appropriate thing to say about such systems is that the proliferation of an indefiniteness has an algebraic dependence on time rather than the exponential one .
moreover , some systems behave in a mixed way : partly with an exponential growth of their enveloping ( coarse - grained ) phases volume and partly with its algebraic growth with time .
all the above mentioned facts lead us to a necessity to introduce a new notion of complexity and entropy and this is the goal of this paper .
our scheme is based on a similar one to the bowen idea of the @xmath2-separation of trajectories , which explicitly imposes the instability features of the dynamics .
however , instead of the complexity @xmath11 we introduce a complexity function @xmath0 which is similar to a propagator @xmath3 with a replacement @xmath12 where @xmath5 is the natural length of trajectories between @xmath13 .
the corresponding entropy @xmath1 can be defined as @xmath14 . in this form
the complexity and entropy describe the system evolution in space - time without assumption of the space - time separation .
another important change is related to the choice of _ basic variables _ which can be @xmath15 , @xmath16 instead of @xmath17 .
similar logarithmic variable @xmath18 instead of @xmath4 appeared in @xcite for a definition of entropy when lvy distributions is considered .
we show that for some typical cases @xmath19 possesses invariants @xmath9 similar to the riemann invariants for the simple wave propagations .
the invariants @xmath9 do not depend on @xmath2 and they represent space - time fractal dimensions of the dynamical system . in the proposed way of description of random dynamics , the notions of _ complexity function _ and _ entropy function _ appear to be constructive tools of the description of dynamics with zero or nonzero lyapunov exponents , with mixing or weak mixing properties , and with normal or anomalous transport .
a similar approach can be developed for a system with mixed features when the basic variables are @xmath20 or @xmath21 . in all considered cases
the _ entropy is an additive function _ of the corresponding basic variables . as an important consequence of the new notions , we consider directional complexity and directional entropy generalizing the notion introduction in @xcite and @xcite .
consider a dynamical system on the 2@xmath22-dimensional torus @xmath23 and let the dynamics of a particle be defined by an evolution operator @xmath24 , \label{eq : eq2.1}\ ] ] which preserves the measure @xmath25 = const ( phase volume ) , and @xmath26 are generalized momentum and coordinate .
what kind of dynamics should be considered as a simple one and what as a complex one ?
we do not assume that there is the only definition of complexity which particularly depends on how a notion of it will be applied to the dynamics .
a more or less typical definition of the complexity depends on how trajectories are mixed in phase space due to the dynamics ( [ eq : eq2.1 ] ) .
the stronger is mixing , the more complex is the dynamics .
one can immediately comment on some weak features of this type of approach which deals with global phase space and global mixing .
the process of mixing can be non - uniform in time and it can have local space - different rates .
we call these features of mixing as space - time non - uniformities and , speaking about the space , we have in mind the full phase space or its part where the dynamics is ergodic .
the former comment leads us to a possibility of such definition of the complexity which could embrace non - uniformity of space - time dynamic processes represented by trajectories .
space - time non - uniformity suggests that vicinities of any trajectory considered at different points @xmath27 with @xmath28 , @xmath29 may have very different dynamics of trajectories ( see fig .
[ fig1 ] ) .
a strong `` inconvenience '' of this conclusion becomes clear if we assume the `` vicinity '' as an infinitesimal ball of the radius @xmath30 around a point @xmath31 : it is difficult , if not impossible , to describe a trajectory finite - term behavior on the basis of the information about the trajectory from an infinitesimal domain of phase space .
the necessary data should arrive from finite pieces of trajectories which make a possible definition of complexity to be space - time non - local
. we should be ready to have a situation with an exponential divergence of trajectories at some small parts of the phase space , and to have a sub - exponential divergence at other parts .
the importance of observing a trajectory with some precision @xmath2 during a finite time @xmath32 was discussed in detail by grassberger and procaccia in [ gp ] and [ p ] .
they also consider fluctuations of the lyapunov exponents obtained from a finite time observation .
these works also presented a space - time partitioning as a way to obtain correlation properties of trajectories .
our analysis here will be extended , comparing to [ gp ] and [ p ] , in two directions : it will be applied to dynamical systems that may have a sub - exponential divergence of trajectories at some phase space domains , and it will deal with area - preserving hamiltonian dynamics , which permits the use of some important results on the poincar recurrences .
here we present a brief review of a few important definitions of complexity of dynamics that involve a bunch of orbits and their comparable behavior in phase space .
historically , the first notion of complexity was introduced by hedlund and morse [ he ] for symbolic systems .
let @xmath33 be an admissible sequence and @xmath34 be the set of all admissible sequences .
one can say that @xmath35 are coordinates , @xmath36 is a piece of an admissible trajectory of length @xmath22 , and @xmath37 is the phase space .
dynamics is defined by a shift operator @xmath38 the complexity @xmath39 of an individual orbit going through a point @xmath40 is the number of different words @xmath41 of the length @xmath22 @xmath42 in the sequence @xmath43 .
if we take into account that each word @xmath44 indicates a piece of the symbolic phase space , i.e. an element of the phase space partition , then @xmath39 is the number of different cells in the phase space available by the orbit going through the initial point @xmath43 .
it was shown in [ he ] that if @xmath45 then @xmath43 is eventually periodic , i.e. @xmath46 const and , for example , there is no such @xmath43 that @xmath47 .
the complexity @xmath39 may grow exponentially with @xmath22 @xmath48 and in this case @xmath49 is the topological entropy .
there are examples of @xmath50 i.e. of the sub - exponential complexity ( see more in [ ti ] , [ fer ] , and references therein ) . a transition from the symbolic complexity to the complexity of dynamical systems with arbitrary topological phase space was suggested in [ bla ] .
let a topological space @xmath51 and a continuous map @xmath52 @xmath53 generate a dynamical system @xmath54 , and let @xmath55 is a finite cover of the phase space by , say open , subsets .
for any initial point @xmath56 one can consider different itineraries @xmath57 such that @xmath58 .
then the topological complexity @xmath59 is the number of different possible itineraries of the temporal length @xmath22 for points @xmath60 in @xmath51 .
evidently , this complexity depends on the covering system @xmath61 . if the system is chaotic ,
then @xmath62 where @xmath49 is the topological entropy provided that @xmath61 is chosen in a right way .
let us emphasize that the topological complexity deals with all orbits of a dynamical system .
there are other possibilities to introduce a complexity such as kolmogorov complexity ( see for example [ br ] ) , a measure theoretical complexity [ fer1 ] , etc . , which will not be discussed here
. the definition of @xmath63-complexity will be the most important for our following generalizations , and we discuss it in more detail
. in @xcite the authors introduced notions of @xmath2-capacity and @xmath2-entropy in space of curves which can or can not be solutions of a differential equation .
let @xmath64 , @xmath65 be a space of continuous curves endowed with the following ( chebyshev ) metric @xmath66 where @xmath67 is a metric in the space @xmath68 .
let @xmath69 be the maximal number of curves which are @xmath2-pairwise @xmath70-disjoint .
then @xmath71 is called the @xmath2-capacity .
let @xmath72 be the minimal number of sets of @xmath70-diameter @xmath73 , needed to cover the set @xmath74 .
then @xmath75 is called the @xmath2-entropy .
if one assumes that @xmath76 is a piece of an orbit of a dynamical system and replaces the continuous time by discrete one , then one comes ( as it was done by bowen [ b ] ) to the following definitions .
let us consider first a dynamical system generated by an evolution operator @xmath77 on the phase space @xmath68 .
let @xmath78 be a subset of initial points ( it could be invariant or not ) and @xmath79 an orbit segment of temporal length @xmath22 going through an initial point @xmath80 .
two segments @xmath81 and @xmath82 , @xmath83 , are said to be @xmath84-separated if there exists @xmath85 , @xmath86 , such that dist@xmath87 , where dist means distance in the phase space @xmath68 .
the maximal number of distinct segments of orbits with accuracy @xmath2 is defined by @xmath88 and is said to be @xmath84-complexity of the set @xmath89 .
evidently @xmath90 is the @xmath2-capacity .
it was shown in [ b ] ( see also @xcite for compact invariant subset of initial points ) that @xmath91 is the topological entropy of the dynamical system @xmath92 on the set @xmath89 , and in [ t ] that @xmath93 is the upper box dimension of the set @xmath89 .
thus , we may assume that if @xmath94 , @xmath95 then @xmath96 where @xmath97 is a subexponential function of @xmath98 and @xmath22 .
we will call the number @xmath99 the bowen or @xmath63-complexity of the set @xmath89 .
the definition ( [ eq : eq3.6 ] ) is fairly general and can be applied to non - hamiltonian and non - compact dynamics .
the definition considers a set of orbits without details of their separation process with a time - interval @xmath22 that typically is large . to illustrate the property ( [ eq : eq3.7 ] ) , consider one - dimensional mixing dynamics on the interval @xmath100 $ ] , @xmath101 $ ] with an exponential divergence of trajectories .
let @xmath102 is the initial distance at @xmath103 between two trajectories and @xmath104 is the distance at time @xmath32
. then @xmath105 i.e. for @xmath106 two trajectories are separated .
the number of such trajectories is @xmath107 where @xmath108 is the length of a small initial interval @xmath89 .
the properties ( [ eq : eq3.6 ] ) and ( [ eq : eq3.7 ] ) follow directly from ( [ eq : eq3.11 ] ) for @xmath109 . if the set @xmath89 has the box - dimension @xmath110 , then ( [ eq : eq3.10 ] ) should be replaced by @xmath111 and correspondingly , instead of ( [ eq : eq3.11 ] ) @xmath112 for more general situations we may assume @xmath113 where @xmath114 is a slow varying function of @xmath115 and @xmath32 compared to the main multipliers .
the expression ( [ eq : eq3.14 ] ) shows in an explicit way how the bowen complexity depends on the time interval @xmath32 , accuracy @xmath2 to determine the location of trajectories , and the domain @xmath89 of a set of initial conditions .
the dependence on @xmath89 can be eliminated from ( [ eq : eq3.11 ] ) or ( [ eq : eq3.13 ] ) by choosing a normalized complexity @xmath116 per unit volume : @xmath117 which is possible due to the uniformity of mixing in the considered dynamical system
. * remark . *
complexity of an orbit .
if one chooses an orbit @xmath118 in the capacity of the set @xmath89 of initial points , one will arrive to a definition of the @xmath2-complexity of the orbit @xmath119 .
it is simple to see that this definition is an analog of the symbolic complexity of a symbolic system described above .
it is not difficult to show that for any small @xmath120 , @xmath121 i.e. , the complexity of the closure of an orbit asymptotically behaves in the same way as the complexity of the orbit . furthermore making use of the definition of complexity for an arbitrary set @xmath89
, one may introduce the complexity of a measure .
* definition 1 . * given an invariant measure @xmath10 , the quantity @xmath122 ia called the complexity of the measure @xmath10 .
we shall use this definition bellow .
expression ( [ eq : eq3.6 ] ) can be interpreted in a way that may be generalized to much more complicated situations .
let again @xmath68 be the phase space , @xmath123 its phase volume , and @xmath124 be the phase volume of a set of initial conditions @xmath125 at time @xmath126 and consider their evolution @xmath127 up to time @xmath32 .
let @xmath128 be a minimal enveloping @xmath127 convex phase volume .
then for systems with exponential divergence of trajectories @xmath129 to find how many different states can occupy the volume @xmath130 , one should define an `` elementary '' minimal volume of one state , i.e. @xmath131 . then @xmath132 where maximum is considered with respect to different sets @xmath89 in @xmath124 .
expressions ( [ eq : eq3.11]),([eq : eq3.12 ] ) permit an important physical interpretation .
hamiltonian chaotic dynamics preserves the phase volume , i.e. @xmath133 . the enveloped or coarse - grained phase volume @xmath128 grows approximately as ( [ eq : eq3.16 ] ) .
the number of states in @xmath128 depends on the definition of a state in the enveloped phase volume .
let one state occupies an elementary volume @xmath134 .
then the number of states that occupy the volume @xmath128 is simply @xmath135 let us emphasize that this interpretation stops working when @xmath136 ( in fact , when @xmath137 ) . now instead of the @xmath2-separated trajectories we can introduce @xmath134-separated ones , associated to a partitioning of @xmath138 into elementary cells .
two trajectories will be @xmath134-separated over the time interval @xmath32 if they do not stay at the same cell @xmath139 during @xmath32 .
if the volume of an elementary cell is @xmath140 , then @xmath141 and we arrive at the connection @xmath142 we also can introduce an entropy @xmath143 for the @xmath134-separated states , i.e. @xmath144 with a condition for the states - separation @xmath145 it is essential that the entropy @xmath143 is defined relatively to a definition of the elementary phase volume @xmath134 which depends on the system and the type of the evolution process .
non universality of the entropy @xmath143 will be more evident in the next sections where the algebraic complexity will be considered .
here we only would like to mention that partitioning of phase space depends on the level of information we want to ignore in the description of dynamics and on the level of information we would like to keep about system trajectories .
a few other comments also are useful .
@xmath134-partitioning resembles a procedure of coarse - graining , which is typical in statistical mechanics .
@xmath134-partitioning is not the same as @xmath2-partitioning since trajectories from neighboring cells can be arbitrarily close to each other during an arbitrarily large @xmath32 , although we will never know it . at the same time
there are common features between @xmath134- and @xmath2-partitioning .
the formal expressions ( [ eq : eq3.13 ] ) for @xmath146 and ( [ eq : eq3.19 ] ) for @xmath147 are the same up to a const ( compare to ( [ eq : eq3.20]),([eq : eq3.22 ] ) ) .
both definitions have the same limitation for the bounded hamiltonian dynamics : @xmath148 @xmath149 the constraints ( [ eq : eq3.23 ] ) do not depend on time .
thus , it follows from ( [ eq : eq3.23 ] ) the existence of @xmath150 such that trajectories or their segments not separated during the @xmath151 will be non - distinguishable .
this property of the definitions ( [ eq : eq3.13]),([eq : eq3.19 ] ) eliminates a significant part of the dynamics with non - uniform mixing in phase space .
the goal of this chapter is to introduce a new definition of a complexity function ( cf ) rather than just complexity , in order to be able to characterize a system with at least two different time scales of a `` complex '' dynamics . as before
, we deal with area - preserving dynamics of systems in a metric phase space @xmath68 endowed with a distance @xmath152 and discrete or continuous time @xmath32 .
the dynamical system @xmath153 defines a distance @xmath154 between two trajectories at time @xmath32 that were initially at points @xmath155 .
one can also introduce a natural length @xmath156 along the trajectory initially at @xmath4 .
we shall need the following definition : two trajectories with initial points @xmath83 will be @xmath157-indistinguishable if @xmath158 we now introduce a notion of complexity that is based on the verification of divergence of trajectories from fixed several ones .
we start with a definition of local complexity .
consider a small domain @xmath78 with diameters @xmath159 , @xmath160 where @xmath161 , and fix some number @xmath162 .
let us pick a point @xmath163 and call the corresponding trajectory the basic one .
a set @xmath164 is said to be locally ( @xmath2 , t)-separated if * for every @xmath60 there is @xmath165 such that @xmath166 and @xmath167 * for every pair ( @xmath85 , @xmath168 ) , @xmath169 , @xmath170 , one has @xmath171 if ( [ distance ] ) is not valid then a pair of trajectories corresponding to the pair @xmath172 is @xmath2-indistinguishable , and it should be treated as one trajectory during the time @xmath32 . * definition 2 . *
the number @xmath173 is called the local complexity . as a function of @xmath32
it is said to be the local complexity function .
a set @xmath174 is called ( @xmath2 , t)-optimal if it is locally ( @xmath2 , t)-separated and @xmath175 .
it is simple to see that @xmath176 and @xmath177 so , if we are interested in the separation of a bunch @xmath174 of @xmath178 trajectories with initial points @xmath179 , @xmath180 and their evolution @xmath181 , then after time @xmath32 we find out that there are @xmath182 @xmath2-separated ( from the basic one ) trajectories- denote this set by @xmath183- and @xmath184 indistinguishable ( from the basic one ) trajectories ( see fig . [ new1 ] ) . of the size
< < @xmath2 during time @xmath32 .
the optimal ( maximal ) number of @xmath185separated trajectories @xmath186 .
diameter of all 9 domains of @xmath187 is @xmath2 . te initial bunch
@xmath174 has 10 trajectories with @xmath188 of the @xmath185separated ones .
two top examples show the process of evolution that makes a pair of trajectories ( in)distinguishable.,scaledwidth=50.0% ] if @xmath183 is locally separated then we obtain an estimate @xmath189 if , in addition , @xmath183 is ( @xmath2 , t)-optimal then @xmath190 from physical point of view it is natural to put the following restriction @xmath191 see subsection d of section 3 .
now consider a partitioning of the full phase volume @xmath68 by a set of domains @xmath192 , and in each @xmath193 select points @xmath194 that make a set of basic trajectories @xmath195 .
let us assume for the sake of definiteness that @xmath196 because of ( [ kk ] ) , every two points belonging to @xmath2-neighborhoods of different basic points are ( @xmath2 , t)-separated for any @xmath197 .
we consider the finite set @xmath198 , @xmath199 .
every point in @xmath174 belongs to the one of sets @xmath193 , so we have a partition @xmath200 , @xmath201 .
we call the set @xmath174 semi - locally ( @xmath2 , t)-separated , or simply ( @xmath2 , t)-separated , if every @xmath202 is locally ( @xmath2 , t)-separated ( with respect to the basic point @xmath194 ) . * definition 3 . * the number @xmath203 is called the semi - local complexity , or simply complexity . as a function of @xmath32
it is said to be the complexity function .
if we choose @xmath204 initial points at each set @xmath193 and consider @xmath205 of them corresponding to trajectories ( @xmath2 , t)-separated from the basic one , then we may form the sum @xmath206 if these @xmath207 points form an optimal set for every @xmath85 then this inequality becomes the equality : @xmath208 the complexity functions @xmath209 and @xmath210 show a level of time - proliferation of @xmath2-separated trajectories from the initial set of a large number of indistinguishable points .
if we are interested in the only typical ( for some measure ) orbits we may adjust definitions above to this situation by assuming that @xmath174 always consists of typical points for this measure and the points @xmath194 are also typical ( see below ) .
one of the main points we are interested in is behavior of complexity functions in a neighborhood of the sticky set .
it is well - known that a `` standard '' trajectory in chaotic sea behaves in the intermittent way : after relatively short chaotic burst it is attracted to the sticky set for a long time , then comes back to mixing part of the chaotic sea , etc .
if our consideration is restricted to a neighborhood of one ( or several ) basic orbit , then fast separated pieces of orbits correspond to a mixing type of behavior and their initial points are situated much `` far '' from the sticky set than initial points of slow - separated pieces of orbits . by using this observation , we may eliminate fast - separated points ( see the next section ) i.e. , in fact we may choose such initial points in @xmath174 which practically belong to the sticky set .
in other words , we may in principle calculate local and semi - local complexity of a measure concentrated on the sticky set .
more rigorously , assume that an invariant measure @xmath10 is given .
then in definition 2 we consider only sets @xmath211 such that * @xmath174 is locally ( @xmath2,t)-separated ; * @xmath212 , @xmath213 ( where @xmath214 is the smallest closed set of full measure i.e. , the set on which the measure @xmath10 is concentrated )
. the maximal number of elements in such sets @xmath174 will be called the local ( @xmath2,t)-complexity of measure @xmath10 .
we denote it by @xmath215 .
similarly if in definition 3 we consider only sets @xmath174 containing points from @xmath214 , then we obtain the semi - local ( @xmath2,t)-complexity of measure @xmath10 , @xmath216 it is useful to introduce the following quantity @xmath217 \approx p_\mu ( \epsilon , t , x_0 ) \delta t,\ ] ] where @xmath218 .
this quantity gives a probability to diverge by distance @xmath2 from the basic orbits during the time interval @xmath219 $ ] , and @xmath220 is the corresponding probability density function . similarly , @xmath221 \approx p_\mu ( \epsilon , t , b ) \delta t\ ] ] gives the probability to diverge from basic orbits going through @xmath222 during the time interval @xmath219 $ ] , and @xmath223 is the corresponding probability density function . from now , we will choose initial points @xmath224 in @xmath174 in such a way that the distances between @xmath224 and the basic point @xmath225 are the same for all @xmath226 , and we denote it by @xmath227 moreover , for the sake of simplicity we omit the argument @xmath89 in @xmath209 and @xmath228 in @xmath210 .
so @xmath229 and @xmath230 .
as usual in numerical simulations we will assume that randomly chosen points @xmath225 , @xmath194 and @xmath174 are typical with respect to some measure @xmath10 we are interested in .
the values of @xmath231 in ( [ n0x0 ] ) , ( [ n0x01 ] ) and in ( [ n0b ] ) , ( [ n0b1 ] ) depend on the choice of @xmath227,@xmath2 , @xmath225 ( or @xmath222 ) and @xmath174 .
the smaller is @xmath227 , the longer @xmath32 should be considered until the maximal values of @xmath232 or @xmath233 will be achieved .
this makes the limit @xmath234 fairly simple .
understanding a way to work with the parameter @xmath2 is more complicated .
consider one trajectory @xmath235 that starts at @xmath4 and has a natural length @xmath236 .
let @xmath32 be fairly big and select a set of points @xmath60 along a trajectory and , approximately , almost uniformly distributed . we can operate with points @xmath60 in the same way as with the basic points @xmath194 in ( [ n0b1 ] ) . as a result
, we obtain the quantity @xmath237 where points @xmath60 belong to the same trajectory .
that means that while @xmath238 characterizes the @xmath2-divergence from @xmath4 during @xmath32 , @xmath239 characterizes the @xmath2-divergence from @xmath60 : points @xmath4 are taken in different places of the phase space and points @xmath60 are taken along the only trajectory .
it is natural to believe that for an appropriately typical set of @xmath240 and of @xmath241 and fairly large @xmath32 , the equality @xmath242 holds where subscripts @xmath85 are omitted .
this equality can be treated as an analog of the ergodic theorem .
a corresponding simulation for @xmath243 was performed in [ lz ] for a system of tracer dynamics in the field of point vortices .
the basic trajectory was created by a tracer , and a few `` host tracers '' were considered within a small distance @xmath244 from the basic trajectory .
each time , when a host tracer moves at a distance @xmath2 from the basic one , it was removed and replaced by a new host tracer at a distance @xmath227 from the basic one ( fig .
[ new2 ] ) .
-separation from the basic trajectory.,scaledwidth=50.0% ] the scheme of calculation of @xmath243 is similar to one used for calculation of the lyapunov exponents , except for some details : 1 .
the value of @xmath2 was much less than the distance typically used to evaluate the lyapunov exponents ; 2 .
trajectories of some tracers were @xmath2-separated after a very long time .
just these trajectories correspond to events of our main interest and their statistics was collected .
the scheme of obtaining of @xmath243 provides simultaneously two different distributions : `` probability '' @xmath243 to have @xmath2-separation at time @xmath32 , and @xmath245 as a `` probability '' to have @xmath2-separation after the travel over natural length @xmath246 along the trajectory .
both of these probabilities can not be obtained by a simple transformation of variables @xmath247 since the trajectories can be of a fractal type and the variables may not be transformable .
these necessitates to use more general distribution than @xmath245 or @xmath248 . this will be fulfilled in the following section .
the heuristic consideration of the previous section can be formalized in a more accurate way and linked to the distribution function of exit time .
let us return back to the set @xmath89 with a basic trajectory that starts at @xmath80 , and let @xmath249 be the set of @xmath2-separated during the time interval @xmath250 trajectories .
assume that the system @xmath251 is transitive and the semi - orbit @xmath252 is dense in @xmath253 .
it follows from that : * proposition 1 . * for any @xmath254 there is @xmath255 such that for any @xmath256 there exists @xmath257 such that if dist@xmath258 , then dist@xmath259 , @xmath260 .
proposition 1 states the existence of the indistinguishable trajectories , and the @xmath2-separation of orbits can be interpreted as an exit time event of @xmath261 from the @xmath227-vicinity of @xmath262 , @xmath263 .
let @xmath264 be a piece of the basic trajectory @xmath265 with the length @xmath266 .
given this basic orbit @xmath264 , one may study a distribution of pairs with respect not only to exit times but also to positions in space measured by a coordinate @xmath5 along the orbit @xmath267 . indeed , let us measure the position of the point @xmath268 on @xmath267 by the coordinate @xmath5 equal to the length @xmath269 of the piece of orbit @xmath270 with the initial point @xmath4 and the final point @xmath271 . because of the proposition 1 , for an arbitrary point @xmath272 the ball @xmath273 of radius @xmath227 centered at @xmath274 contains at least one point @xmath275 .
one may characterize the trajectory @xmath276 by the same parameter @xmath5 as @xmath277 .
let @xmath249 be a @xmath185-separated set and @xmath278 .
fix a coordinate @xmath279 of @xmath224 by choosing the initial point at @xmath224 to start counting @xmath279 .
now , introduce into consideration the pair @xmath280 where @xmath281 is the first time for which @xmath282 , i.e. the exit time from the @xmath2-neighborhood of the basic trajectory @xmath264 , and @xmath283 , the length of the piece of the orbit between the initial and final points .
for two different @xmath224 and @xmath60 that are close to the basic @xmath4 and escape from the @xmath2-vicinity of @xmath235 at close times @xmath284 and @xmath285 one can expect that the orbit segment s length @xmath286 will also be close .
but if we take another @xmath287 as a basic one , other two initial points @xmath288 in the @xmath227-vicinity of @xmath287 such that the @xmath2-escape time for them will be correspondingly @xmath284 , @xmath285 , i.e. the same as for the initial basic point @xmath4 , then it could not be that @xmath289 are the same .
this property is a result of the non - uniformity of behavior of orbits in the phase space and , speaking further , it is a result of the fractal space - time properties of trajectories .
consider a small domain @xmath290 , a typical point @xmath291 and a typical close point @xmath292 at @xmath227-distance from @xmath4 .
the @xmath2-separation of the pair @xmath293 , @xmath294 occurs at some time @xmath32 and distance @xmath295 .
here we omitted @xmath2 as an argument of @xmath296 .
furthermore , let us remark that the distance @xmath5 depends on @xmath274 .
but we neglect this dependence assuming implicitely that for a typical @xmath274 at the @xmath227distance from @xmath4 the value of @xmath5 is asymptotically the same .
the distance @xmath5 of the @xmath157-separation is said to be the length of a flight , i.e. a length of the path that two nearby trajectories are flying together .
let us treat @xmath4 as the basic point and consider the set @xmath297 of points @xmath227-close to @xmath4 .
this set is called locally ( @xmath2 , t , s)-separated if it is locally ( @xmath2 , t)-separated ( see subsection a ) and moreover @xmath298 where @xmath299 . * definition 4 . * the number @xmath300 is called the local ( @xmath2 , t , s)-complexity . as a function of @xmath32 , @xmath5 , it is called the local flight complexity function ( fcf ) . similarly to the definition of @xmath301 in ( [ n0b ] ) , ( [ n0b1 ] ) we can consider a collection of flights and their lengths and time intervals from different domains @xmath302 . as a result
we have the semi - local fcf @xmath303 where @xmath304 are basic points . as in subsection
a we may choose @xmath204 initial points in @xmath193 for any @xmath85 , select the maximal ( @xmath2 , t , s)-separated subset consisting of @xmath305 points and form the sum @xmath306 the sum is equal to the semi - local fcf if initial points form an optimal set . as in subsection c ,
let us assume that there exists a trajectory @xmath262 dense in our phase space @xmath68 .
then , as it was said there , every piece of a trajectory @xmath307 lies in an @xmath2-vicinity of a piece @xmath277 , @xmath257 , and , thus has the coordinate @xmath308 .
it allows us to extend the notion of ( @xmath2 , t , s)-complexity as follows .
we say that a set @xmath297 is ( @xmath2 , t , s)-separated if it is ( @xmath2 , t)-separated and @xmath309 where @xmath310 is the @xmath5-coordinate of the piece @xmath311 .
* definition 5 . * the number @xmath312 is called the flight complexity . as a function of @xmath32 , @xmath5 , it is called the flight complexity function .
this quantity has a simple meaning : let us take a large enough number @xmath313 of @xmath227-close pairs within @xmath68 with a fairly typical distribution of the initial conditions @xmath314 , @xmath315 .
@xmath0 is the number of trajectories that are mutually @xmath2-separated @xmath316 during time @xmath32 at distance @xmath5 therefore @xmath317 is the number of trajectories that are @xmath2-separated during time @xmath318 at distance @xmath319 .
one can introduce the function @xmath320\ ] ] where @xmath321 , is called the density of the fcf . the main goal of this paper is to connect the notion of complexity with the space - time local instability properties of systems . a necessity of such a notion appears due to the specific features of hamiltonian systems to have chaotic dynamics strongly non - uniform in phase space and strongly intermittent in time , and particularly due to the existence of dynamical traps [ za2 ] .
let us consider again a small domain @xmath89 and pick some @xmath178 trajectories in @xmath89 with a characteristic distance @xmath227 between them .
due to the instability these @xmath178 trajectories will fill at time @xmath32 an enveloped phase volume @xmath322 because of the larger distance between trajectories .
after coarse - graining of @xmath323 we arrive at @xmath324 but without empty space ( `` bubbles '' ) presented in @xmath325 .
this is a typical physical situation of growth of the coarse - grained phase volume and the problem is reduced to the estimate of the growth .
the presence of dynamical traps makes the growth of @xmath326 sensitively dependent on all its parameters @xmath32 , @xmath89 and @xmath227 or , in other words , expanding of @xmath130 depends on the initial coordinate and observational time .
for example , if @xmath89 is taken in the stochastic sea , @xmath327 but if @xmath89 is taken inside a dynamical trap , @xmath328 . we need to make the only step to arrive to our definition of complexity .
indeed , one can consider the phase volume partitioning by elementary cells @xmath134 and introduce the complexity @xmath329 where we use @xmath4 as a coordinate of @xmath89 and @xmath2 defines a diameter of the coarse - graining and does not enter into any physical results .
the density of the fcf ( 49 ) can now be considered as a distribution function of having displacement @xmath5 at a time instant @xmath32 and the corresponding entropy can be easily introduced ( see the next section ) .
one more useful connection appears from the expression @xmath330 that is a probability to be trapped in a tube of the diameter @xmath2 supported by the domain @xmath89 during time @xmath32 .
respectively @xmath331 is a probability to exit from the tube during some time @xmath332 , and for @xmath333 and @xmath334 we arrive at @xmath335 of the probability density to return to any small domain @xmath336 with @xmath337 and no dependence on @xmath336 . due to the kac lemma @xmath338 for the bounded hamiltonian dynamics . in this chapter
we considered the case of continuous time .
but all ideas , definitions and results can be generalized to the case of discrete time .
one can choose different analogues of the length of the piece of an orbit- for example in an euclidean space one may consider the length of union of segments joining consecutive points of the orbit .
the main ideas are independent of the choice of the definition of the lengths .
the function @xmath0 characterizes local @xmath2-divergence of trajectories and it can be considered as a new characteristic of the dynamics . its main role is to describe evolution of a typical pair of orbits taken apart of a small distance @xmath227 . let us show how the @xmath339-complexity works .
following a general physical approach let us define @xmath340 as @xmath341-entropy of the dynamics since @xmath0 is a number of @xmath342-separated states . physically speaking , due to ( [ eq : eq4.11 ] ) the number of separated states can not be more than @xmath343 ( compare to ( [ volume ] ) , i.e. for the ergodic dynamics @xmath344 due to a mixing property in phase spaces or , more specifically , due to an instability and separation of any initially close orbits
, one can expect the growth of @xmath1 with time until the max @xmath345 is reached .
the definition ( [ eq : eq5.1 ] ) permits to estimate from @xmath346 some fine properties of the evolution of @xmath347 as a function of time @xmath32 and the length of separation @xmath348 . for the case of the anosov - type systems
@xmath349 i.e. @xmath5 has a `` good '' dependence on time and it may be expunged from @xmath346
. then @xmath350 where @xmath351 and ( [ eq : eq5.4 ] ) coincides with ( [ eq : eq3.13 ] ) ( and the arrow means `` is completely defined by '' ) . similarly @xmath352 and we arrive to the standard expressions @xmath353 @xmath354 which deliberately are written in the form of partial derivatives .
consider fully opposite case of an integrable system in the annulus @xmath355 , @xmath356 , @xmath357 , defined by the equation @xmath358 i.e. the map @xmath359 is @xmath360 .
let us calculate @xmath146 where @xmath89 is the interval @xmath361 , @xmath362 .
since @xmath363 , the segments @xmath364 , @xmath365 are @xmath157-separated iff @xmath366 , i.e. @xmath367 . the maximal number of such points on the interval [ 1,2 ] is @xmath368 , and hence @xmath369 since the number of @xmath2-different sections @xmath361 on the interval @xmath370 $ ] is @xmath371 , then one has @xmath372 thus , integrable nonlinear system on the annulus is a system with linear growth of complexity . by the way , @xmath373 is the dimension of the annulus .
the corresponding entropy is @xmath374 and @xmath375 this shows that in a polynomial case it is possible to consider the expression @xmath376 ( equals to 1 for ( [ eq : eq5.10 ] ) ) which provides the information on the frequency ( equals 1 ) of the rotation in the annulus .
there are many other less trivial systems for which growth of complexity is subexponential ( billiards in polygons , interval exchange transformations , etc . ) .
complexity defined in ( [ eq : eq5.4 ] ) grows exponentially with time reflecting an existence of a positive lyapunov exponent and exponential divergence of trajectories in phase space .
this is not the case for zero lyapunov exponent systems and for diffusional type processes . to consider large scale processes such as the diffusion for systems of that kind ,
let us use the partition function @xmath377 @xmath378 with a normalization constant that can be chosen in such a way that @xmath379 in numerous cases the process of mixing in phase space and the corresponding separation of trajectories has algebraic dependence on time . particularly , it is related to the dynamics with the fractal time [ @xmath380 and the diffusional type of processes . for this case ,
the complexity is polynomial in time and @xmath381 with some exponent @xmath382 and characteristic time scale @xmath383 . as in ( [ eq : eq5.12 ] ) , the entropy growth rate can be defined in the @xmath384 scale @xmath385 and the exponent @xmath382 characterizes the corresponding rate of the entropy growth .
as it appeared in [ @xmath380 , chaotic dynamics of real hamiltonian systems does not display , in general , exponential growth of the complexity , and the number @xmath386 of @xmath2-separated orbits grows asymptotically as polynomial in time and the length of separation @xmath387 . in this case
the complexity can be written as @xmath388 where @xmath389 are characteristic scales of length and time , @xmath390 , and @xmath391 is a slow varying function
. we will be interested in the limit @xmath392 the limit ( [ eq : eq5.18 ] ) corresponds to the increasing of the number of @xmath2-separated states with the growth of time and decreasing of this number with the growth of the separation distance .
for example @xmath393 and @xmath394 correspond to the normal one - dimensional diffusion , while other powers appear for the anomalous diffusion .
following the definition ( [ eq : eq5.1 ] ) let us introduce the @xmath342-entropy : @xmath395 this expression will be analyzed in the next section .
here we would like to mention that the complexity in the form ( [ eq : eq5.17 ] ) suggest that the corresponding partition function @xmath396 has a self - similar dependence on @xmath397 in the limit ( [ eq : eq5.14 ] ) , and consequently for its moment we have @xmath398 with @xmath61 that may depend slowly on @xmath32 .
the equation ( [ eq : eq5.21 ] ) describes the anomalous diffusion with the transport exponent @xmath399 that can be obtained from @xmath400 ( see more in @xcite ) .
the new parameters @xmath401 , and @xmath10 can be considered as new dynamical invariants intimately linked to the @xmath342-complexity and the corresponding entropy .
their interpretation is given in the next section .
dependence of the @xmath342-complexity on two variables @xmath5 and @xmath32 makes it possible to have `` mixed '' complexities of the following types : @xmath402 @xmath403 i.e. algebraic in one variable and exponential in the second one .
the case @xmath404 shows exponential divergence in time and a fractal length structure with a box dimension @xmath405 .
the case @xmath406 does not have yet analogy or physical interpretation .
similar definitions can be done for the entropy : @xmath407 @xmath408 on the basis of all definitions of entropy we can make an important remark about its extensivity .
typically , the extensivity is considered with respect to the coordinate variable in statistical physics .
dynamical systems analysis provides an extensivity with respect to the number of degrees of freedom , or with respect to time .
here we would like to mention that instead of @xmath409 can be @xmath384 or @xmath410 , or different combinations .
for example , in the case of @xmath411 the entropy @xmath412 is additive in the space of variables @xmath413 while in the case ( [ eq : eq5.17 ] ) the entropy is additive in space @xmath414 .
our final remark for this chapter is that @xmath415-complexity is considered for a system with evolution and it may happen very typically , that there are two time intervals : at the beginning the evolution of the complexity is exponential and , after a fairly long time it becomes algebraic [ enz ] .
similar pattern occurs for the sinai billiard where a distribution of the poincar recurrences follows exponential law for the beginning and then decays as @xmath416 .
consider new variables @xmath417 and rewrite ( [ eq : eq5.17 ] ) as @xmath418 where @xmath419 is the slow function of @xmath420 . neglecting the derivatives of @xmath419 , we can conclude from ( [ eq : eq6.2 ] ) that the entropy is constant along the traveling wave - front @xmath421 where @xmath422 is the wave speed .
due to ( [ eq : eq5.22 ] ) @xmath423 and the transport exponent can be interpreted as the double speed of the traveling wave of the entropy or complexity .
this property appears because of the nontrivial coupling between time and phase space of dynamical systems .
indeed , any kind of randomness , which is assumed for kinetics in stochastic field , differs from the randomness of chaotic dynamics since the randomization of trajectories occurs as a result of nonlinearity and dynamical instability .
the structure of the entropy ( [ eq : eq6.2 ] ) or complexity function ( [ eq : eq5.17 ] ) is invariant , with respect to the renormalization of @xmath424 and @xmath425 , namely @xmath426 if we neglect variations of @xmath427 , and if @xmath428 satisfy the equation @xmath429 equation ( [ eq : eq6.6 ] ) shows a connection between parameters @xmath430 of the dynamical origin , transport exponent @xmath10 , and velocity @xmath422 of the traveling wave of the complexity .
the parameter @xmath422 also defines a direction in @xmath397 coordinates of the traveling wave propagation or , in other words , directional complexity and entropy . for different dynamical models scaling parameters @xmath428
were defined in [ zen , kz , bkwz ] .
in fact , @xmath9 or @xmath431 are not constants , and wave propagation can be in different directions .
this situation is an analog to multi - fractal space - time structure . for any direction with `` velocity '' @xmath422 in @xmath432 space
, there is a corresponding curve in @xmath17 space and an isoline of the section of the surface @xmath433 by a plane @xmath433 = const .
the fractional exponents @xmath9 of the complexity space - time dependence are receiving a new meaning as entropy / complexity characteristics @xmath434 along which the traveling waves have a constant velocity . in this way we have a generalization of the usual notion of entropy for a finite system .
this analogy can be advanced further by comparing waves of complexity / entropy to simple waves in fluid dynamics .
then the velocity @xmath422 is the riemann invariant and so is @xmath10 .
we arrive to an interesting conclusion : the transport exponent @xmath10 can be interpreted as a riemann invariant of the complexity ( entropy ) simple wave in the @xmath397-space . in the following chapter
it will be discussed a possibility of anisotropic phase space with different values of @xmath10 in different directions .
correspondingly , we will have different riemann invariants .
the case ( [ eq : eq5.17]),([eq : eq5.19 ] ) of the self - similar behavior of the complexity function shows explicitly that the entropy can be written in the form @xmath435 up to a slow varying function @xmath419 . the expression ( [ eq : eq7.1 ] )
permits to consider joint space - time variables as a 2-dimensional space @xmath432 where entropy propagates along a line in the direction with tangent @xmath422 . as it was mentioned in the previous section
, it can be different riemann invariants and so is for @xmath436 where @xmath437 is a unit vector along some direction . to formalize the notion of the directional complexity and entropy ,
let us use an idea of milnor [ m1,m2 ] for construction of directional windows in space - time ( see also [ acmf ] ) .
consider dynamical systems generated by a map @xmath438 , the @xmath70-dimensional euclidean space . denote by @xmath439 a compact subset of @xmath440 and by @xmath437 a unit vector in @xmath441 such that the projection of @xmath437 onto the last coordinate axis is non - negative .
given @xmath442 let @xmath443 denotes the following window in @xmath444 : @xmath445 for example if @xmath446 , @xmath447 , @xmath448 $ ] , then @xmath449 if @xmath450 , @xmath451 , @xmath452 , then @xmath453 let us generalize the definition 2 in section ( [ sec : sec4.1 ] ) of the @xmath2-separability of orbits as their distinguishability through the window @xmath443 . * definition 6
* two orbits @xmath454 and @xmath455 , @xmath456 , are @xmath457-separated at time @xmath458 if @xmath459 and dist@xmath460 , where dist is the distance in @xmath440 .
now we may define the maximal number of @xmath457-separated orbits : @xmath461 and call it * @xmath462-complexity in the direction @xmath437*. of course , the complexity depends on the set @xmath439 of initial points .
but the rate of its growth with @xmath458 may be independent of @xmath439 .
if this growth is exponential we may define the directional topological entropy .
the topological entropy in the direction @xmath437 is defined by @xmath463 where supremum is taken over all compact subsets @xmath439 of @xmath440 .
note that the limit in @xmath2 exists since @xmath464 is a nondecreasing function of @xmath439 and @xmath2 .
moreover @xmath465 $ ] . for some systems it may happen that @xmath466 . in this case we may compute the following quantity .
the density of topological entropy in the direction @xmath437 is defined by @xmath467 let us remark that definitions of directional topological entropy and its density in [ acfm ] is similar to the definitions above . for the web map @xmath468
shows the complexity traveling waves.,scaledwidth=50.0% ] here we show two numerical examples that demonstrate the existence of the directional complexity and complexity / entropy traveling waves .
both examples are for the web map of 4th - fold symmetry [ z ] .
the map acts in @xmath469-space : @xmath470 with a parameter @xmath471 and @xmath472 .
it was shown in [ wz ] that transport exponent @xmath10 can be different in different directions of @xmath469-space ( see fig .
[ fig3 ] ) .
dependence of transport exponents on the the direction also appears to be recognizable for tracers in a 3-vertex flow @xcite .
after we have made a link between the @xmath342-complexity and transport , one may expect that the presence of the directional anisotropy indicates the anisotropy of the complexity and entropy as well .
more accurately we should say that for an infinite dense set of values of @xmath471 the trajectories are sticky to some islands and , due to that , the distribution function has power tails [ zen ] .
we assume that the complexity function should be of the polynomial type for these cases .
different transport exponents for different directions in such a case were demonstrated in [ wz ] by simulations ( see fig .
[ fig5 ] ) . const . for the same distribution as in fig .
4 shows the lines of the wave propagation.,scaledwidth=50.0% ] figure [ fig4 ] provides more information since it shows in 3-dimensional plot a dependence of the normalized number of trajectories @xmath473 as a function of their position @xmath474 at time instant @xmath32 in the log - log - log scales .
the plot clearly indicates directional waves as ripples of @xmath475 .
figure [ fig5 ] shows different slopes of the isolines @xmath475 = const that correspond to different velocities of the wave propagation . along the @xmath476-direction and @xmath10
= 1.50 along the diagonal direction [ wz].,scaledwidth=50.0% ] for dynamical systems in the circle rotation numbers reflect some properties of behavior of orbits in space - time .
it is interesting to study their relation to directional complexity . for
that we consider the map @xmath477 of the form @xmath478 where @xmath479 is a 1-periodic smooth function such that @xmath480 and @xmath481 is a parameter .
if @xmath482 then @xmath483 is one - to - one and the map @xmath484 mod 1 can be treated as either a continuous map of the circle or a piecewise continuous map of the interval @xmath485 $ ] .
we choose the last interpretation .
for @xmath486 we just have an interval exchange transformation of the interval .
for @xmath487 this transformation becomes nonlinear but topologically very similar to that for @xmath486 .
consider the case @xmath486 more carefully .
the orbit @xmath488 for @xmath486 satisfies the equation @xmath489 therefore the rotation number of the map @xmath483 equals @xmath490 . in any window @xmath491 , @xmath492 , @xmath493 , there are points @xmath494 for any @xmath495 , but if @xmath496 , @xmath497 , then for any fixed @xmath498 there is @xmath499 such that there are no points @xmath494 , @xmath500 , in this window .
more precise , if @xmath501 , @xmath502 $ ] , then @xmath503 for all @xmath504 , and the set @xmath505 does not contain points @xmath494 , @xmath506 , @xmath507 .
if @xmath508 then the set @xmath505 contains points @xmath494 for any @xmath509 but the complexity @xmath510 is still bounded ; in fact , @xmath511 .
the case @xmath512 is similar to the just described one and we omit its consideration .
now let @xmath513 .
we consider the rotation interval of the map @xmath483 , i.e. the set of accumulation points of sequences @xmath514 for all @xmath515 , ( see for instance @xcite , @xcite , @xcite ) .
this interval can be nontrivial , say @xmath516 $ ] , @xmath517 , and for every @xmath518 $ ] there is an infinite @xmath483-invariant set @xmath519 such that @xmath520 implies @xmath521 .
moreover @xmath519 may contain uncountably many points ( @xcite ) .
trajectories in @xmath519 may behave chaotically and the complexity @xmath522 , @xmath496 , @xmath497 , may grow exponentially with @xmath458 , i.e. @xmath523 .
but if @xmath524 $ ] then @xmath522 will be bounded ( as above ) and @xmath525 will be 0 .
thus the behavior of directional complexity has nontrivial dependence on the direction in space - time .
the authors of @xcite studied observable values of rotation numbers .
they found numerically that even if @xmath526 one is able to observe the only one ( or a small number ) of rotation numbers corresponding to preferable directions in space - time .
nonergodicity of phase space of hamiltonian dynamics is due to the presence of islands .
the boundaries of the islands are sticky and this imposes strong nonuniformity of the distribution function in space and in time .
such properties of the dynamics necessitates a new approach to the problem of complexity , entropy and related notions of the unstable dynamics . in this paper
the notion of complexity function is introduced as a characteristic number of mutually @xmath2-separated trajectories during time @xmath32 .
this notion resembles the similar features as a propagator function , i.e. it depends on the initial and final space - time .
in fact , the complexity function shows how fast trajectories disperse from each other due to the local instability properties of a system .
the entropy can be introduced in a natural way as a logarithm of the complexity function but the choice of variables can be different depending on the type of local instability . in this way we were able to include into the same scheme systems with strong intermittency and systems with zero lyapunov exponent . in this article we introduced local and semi - local and flight complexity functions that can be calculated for specific systems .
the question arises : why do not try to calculate just @xmath2-complexity function and work directly with it ? we think that this is impossible .
the main reason becomes clear from the following consideration .
fix @xmath527 and assume that we found an optimal set @xmath297 , @xmath528 , of initial @xmath63-separated points .
if we are interested in the complexity function @xmath529 as a function of @xmath22 , then , say , for @xmath530 we either need to find a new optimal set or to use points in @xmath174 .
of course the second option is preferable .
so , the problem appears : knowing that @xmath174 is ( @xmath2 , n)-separated , select the maximal subset in @xmath174 that is ( @xmath2 , @xmath531)-separated .
let us reformulates this problem as follows .
define the graph @xmath532 with vertices @xmath533 such that the edge @xmath534 exists if and only if the points @xmath535 and @xmath224 are ( @xmath2 , @xmath531)-separated .
the problem is to find a subgraph of @xmath532 with maximal number of vertices such that every two vertices of this subgraph are connected by an edge .
this is the well - known clique problem which is happened to be so called np - complete ( see for instance @xcite ) .
specialists believe that np - complete problems have no good algorithms , i. e. one can not solve such a problem by using best computers even for not so large values of @xmath531 .
we would like to thank m. edelman for the help in preparing the figures and m. courbage for interesting discussions .
gz was supported by the u.s .
navy through grant nos .
n00014 - 96 - 1 - 0055 and n00014 - 97 - 1 - 0426 , and the u.s .
department of energy through grant no .
de - fg02 - 92er54184 .
v. a. was partially supported by conacyt grant 485100 - 5 - 36445-e .
saum2001 m. abel , l. biferale , m. cencini , m. falcione , d. vergni and a. vulpiani , exit - time and @xmath2-entropy for dynamical systems , stochastic processes , and turbulence , physica d * 147 * , 12 - 35 ( 2000 ) .
v.s . afraimovich and g.m .
zaslavsky , sticky orbits of chaotic hamiltonian dynamics , lecture notes in physics * 511 * , 59 - 82 ( 1998 ) .
v. afraimovich , m. courbage , b. fernandez and a. morante , directional entropy in lattice dynamical systems , in : * mathematical problems of nonlinear dynamics * lerman .
shilnikov , eds , nizhny novgorod university press ( 2002 ) .
l. ascola , j. llibre , m. misiurewicz , combinatorical dynamics and entropy in dimension one , in * advanced series in nonlinear dynamics 5 * , world scientific , singapore , 1993 .
p. ashwin , x .- c .
fu , and j.r .
terry , riddling and invariance for discontinuous maps preserving lebesque measure , nonlinearity * 15 * , 633 - 645 ( 2002 ) .
r. badii and a. politi , * complexity * , cambridge univ .
press , cambridge , 1999 .
s. benkadda , s. kassibrakis , r. white , and g.m .
zaslavsky , self - similarity and transport in the standard map , pre * 55 * 4909 - 4917 ( 1997 ) .
f. blanchard , b. host , and a. maass , topological complexity , erg .
* 20 * , 641 - 662 ( 2000 ) .
f. blanchard , e. glasner , s. kolyada , and a. maass , on li - yorke pairs , to appear in j. fr reine und ungewandte matematic .
g. boffetta , m. cencini , m. falcioni , and a. vulpiani , predictability : a way to characterize complexity , preprint , 2001 .
g. boffetta , m. cencini , m. falcioni , and a. vulpiani , predictability : a way to characterize complexity , physics reports * 356 * , 367 - 474 ( 2002 ) . r. bowen , topological entropy for noncompact sets , trans .
ams * 84 * , 125 - 136 ( 1973 ) .
brudno , entropy and complexity of the trajectories of a dynamical system , trans .
moscow math .
soc . * 44 * , 127 - 148 ( 1983 ) .
p. collet and j .- p .
eckmann , the definition and measurement of the topological entropy per unit volume in parabolic pdes , nonlinearity * 12 * , 451 - 473 ( 1999 ) .
m. courbage and d. hamden , decay of correlations and mixing properties in a dynamical system with zero k - s entropy , ergod .
th . & dynam .
* 17 * , 87 - 103 ( 1997 ) .
e. i. dinaburg , relation between diferent entropy characteristics of dynamical systems , mat .
izvestia ansssr , * 35 * , 324 - 366 ( 1971 ) .
s. ferenczi , measure - theoretic complexity of ergodic systems , israel j. math .
* 100 * , 189 - 207 ( 1997 ) .
s. ferenczi , complexity of sequences and dynamical systems , discrete mathematics * 206 * , 145 - 154 ( 1999 ) . m. r. garey , d. s. johnson , computers and intractability .
a guide to the theory of np - completeness , w. h. freeman and company , ny , twenty - first printing 1999 .
p. grassberger and i. procaccia , measuring the strangeness of strange attractors , physica d * 9 * , 189 - 208 ( 1983 ) .
p. grassberger and i. procaccia , dimensions and entropis of strange attractors from a fluctuating dynamics approach , physica d * 13 * , 34 - 54 ( 1984 ) .
hedlund and m. morse , symbolic dynamics , am . j. math . *
60 * , 815 - 866 ( 1938 ) .
keener , chaotic behavior in piece - wise continuous difference equations , trans .
ams * 261 * , 589 - 604 ( 1980 ) .
n. kopell and deruelle , bounds on complexity in reaction - diffusion systems , siam j. appl . math . * 46 * , 68 - 80 ( 1986 ) .
l. kuznetsov and g.m .
zaslavsky , passive particle transport in three - vortex flow , phys .
e * 61 * , 377 - 379 ( 2001 ) .
a. n. kolmogorov , v. m. tikhomirov , @xmath2-entropy and @xmath2-capacity of sets in functional spaces , uspekhi mat .
nauk , * 14 * , 3 - 86 ( 1959 ) .
x. leoncini and g.m .
zaslavsky , _ jets , stickiness , and anomalous transport _ , phys .
e * 65 * , 046216 ( 2002 ) .
d. lind and b. markus , * an introduction to symbolic dynamics and coding * , cambridge university press ( 1995 ) .
j.h . lowenstein and f.
vivaldi , anomalous transport in a model of hamiltonian round - off , nonlinearity * 11 * , 1321 - 1350 ( 1998 ) .
malkin , entropy and rotation sets of continuous circle maps , ukrainian math . j. * 35 * , 280 - 285 ( 1983 ) .
malkin , rotation intervals and the dynamics of lorenz - type mappings , selecta mathematica sovietica * 10 * , 267 - 275 ( 1991 ) .
j. milnor , directional entropies of cellular automaton - maps , in * disordered systems and biological organization * , springer , berlin ( 1986 ) , 113 - 115 .
j. milnor , on the entropy geometry of cellular automata , complex systems * 2 * , 357 - 385 ( 1988 ) .
montroll and m.f .
shlesinger , maximum entropy formalism , fractals , scaling phenomena , and @xmath536 noise : a tale of tails , j. stat .
* 32 * , 209 - 230 ( 1983 ) .
i. procaccia , the static and dynamic invariants that characterize chaos and the relations between them in theory and experiments , physica scripta * t9 * , 40 - 46 ( 1985 ) .
saum and t.r .
young , observed rotation numbers in families of circle maps , int . j. bifurcation and chaos * 11 * , 73 - 89 ( 2001 ) .
f. takens , distinguishing determinstic and random systems , in * nonlinear dynamics and turbulence * ( eds .
g.i barenblatt , g. iooss , d.d .
joseph ) , pitman ( 1983 ) , 314 - 333 .
tikhomirov , uspehi matem .
nauk , * 18 * , 55 ( 1963 ) .
h. weitzner and g.m .
zaslavsky , directional fractional kinetics , chaos * 11 * , 384 - 396 ( 2001 ) .
zaslavsky , chaotic dynamics and the origin of statistical laws , physics today * 8 * , 39 - 45 ( 1999 ) .
zaslavsky , * physics of chaos in hamiltonian systems * , imperial college press ( london , 1999 ) .
zaslavsky , dynamical traps , physica d , to appear in 2002 .
zaslavsky and m. edelman , weak mixing and anomalous kinetics along filamented surfaces , chaos * 11 * , 295 - 305 ( 2001 ) .
zaslavsky and m. edelman , pseudochaos ( not published ) .
zaslavsky , m. edelman , and b. nyazov , self - similarity , renormalization , and phase space nonuniformity of hamiltonian chaotic dynamics , chaos * 7 * , 159 - 181 ( 1997 ) .
zaslavsky , chaos , fractional kinetics and anomalous transport , phys .
reports , * 371 * , 461 - 580 ( 2002 ) | new notions of the complexity function @xmath0 and entropy function @xmath1 are introduced to describe systems with nonzero or zero lyapunov exponents or systems that exhibit strong intermittent behavior with `` flights '' , trappings , weak mixing , etc .
the important part of the new notions is the first appearance of @xmath2-separation of initially close trajectories .
the complexity function is similar to the propagator @xmath3 with a replacement of @xmath4 by the natural lengths @xmath5 of trajectories , and its introduction does not assume of the space - time independence in the process of evolution of the system .
a special stress is done on the choice of variables and the replacement @xmath6 , @xmath7 makes it possible to consider time - algebraic and space - algebraic complexity and some mixed cases .
it is shown that for typical cases the entropy function @xmath8 possesses invariants @xmath9 that describe the fractal dimensions of the space - time structures of trajectories .
the invariants @xmath9 can be linked to the transport properties of the system , from one side , and to the riemann invariants for simple waves , from the other side .
this analog provides a new meaning for the transport exponent @xmath10 that can be considered as the speed of a riemann wave in the log - phase space of the log - space - time variables .
some other applications of new notions are considered and numerical examples are presented . |
the non - hermitian evolution of the different systems considered in this work has been calculated with a master equation approach , in which the different dissipative channels are described by proper lindblad terms . for a given hamiltonian @xmath23
, we calculate the temporal evolution of the density matrix @xmath100 through the equation @xmath101 + \sum_{i } \frac{\kappa_i}{2 } \left(2\hat{a}_i \hat{\rho } \hat{a}_i^\dag - \left\{\hat{a}_i^\dag \hat{a}_i , \hat{\rho}\right\ } \right ) + \frac{\gamma_\mathrm{m}(n_\mathrm{th}+1)}{2 } \left(2\hat{b } \hat{\rho } \hat{b}^\dag - \left\{\hat{b}^\dag \hat{b } , \hat{\rho}\right\ } \right ) + \\&+\frac{\gamma_\mathrm{m } n_\mathrm{th}}{2 } \left(2\hat{b}^\dag \hat{\rho } \hat{b } - \left\{\hat{b } \hat{b}^\dag , \hat{\rho}\right\ } \right ) + \frac{\gamma}{2 } \left(2\hat{\sigma}_- \hat{\rho } \hat{\sigma}_+ - \left\{\hat{\sigma}_+ \sigma_- , \hat{\rho}\right\ } \right ) + \frac{\gamma^*}{2 } \left(\hat{\sigma}_z \hat{\rho } \hat{\sigma}_z -\hat{\rho } \right ) \end{split}\end{aligned}\ ] ] where the sum refers to all the optical cavities considered , @xmath102 is decay rate of the i - th cavity , @xmath103 is the decay rate of the mechanical mode , @xmath104 is the average phonon number of the external bath , and @xmath84 and @xmath105 are the decay rate and pure dephasing rate of the emitter , respectively .
the brackets [ , ] and \ { , } indicate commutation and anti - commutation of the operators , respectively . the master equation has been solved numerically with the opensource python framework qutip @xcite . for the numerical calculations , the dimensions of the fock spaces of the optical cavities and mechanical resonator need to be truncated .
for the calculations of the emitter - phonon swapping , the fock spaces of the mechanical oscillator and optical cavities all have dimensions of 2 .
we verified that no appreciable numerical differences arise for larger fock spaces dimensions . for the cooling calculations ,
the fock space of the mechanical resonator has a dimension of 15 , to ensure that a thermal state with @xmath92 can be properly described .
the general model describing mode field coupling in a two - cavity system is schematically depicted in fig .
[ figsuppinfo_model_2_cavities.png ] .
we consider two identical optomechanical systems , denoted left ( l ) and right ( r ) , each composed of an optical cavity with frequency @xmath7 and a mechanical resonator with frequency @xmath14 . in each system , the cavity and the resonator are dispersively coupled at a rate @xmath52 .
we describe the fields in the two optical cavities with the annihilation operators @xmath106 and @xmath107 , and the two mechanical resonators with annihilation operators @xmath108 and @xmath109 .
the two optical cavities are coupled with a rate @xmath40 , while the two mechanical resonators are coupled with a rate @xmath110 .
finally , a two - level emitter is placed in the left cavity , and interacts with the field @xmath106 with a coupling rate @xmath19 . .
each cavity is dispersively coupled to a separate mechanical resonator with an optomechanical coupling rate @xmath52 .
the resonators are identical and have frequency @xmath14 .
the two mechanical resonators are additionally coupled to each other at a rate @xmath110 .
an emitter is placed in the left optical cavity and interacts with one of its optical modes with a coupling rate @xmath19 . ]
the full hamiltonian reads @xmath111 \hat{a}_\mathrm{l}^\dag \hat{a}_\mathrm{l } + \left[\omega_\mathrm{c } - g_0 \left(\hat{b}_\mathrm{r}^\dag + \hat{b}_\mathrm{r } \right ) \right ] \hat{a}_\mathrm{r}^\dag \hat{a}_\mathrm{r } + \omega_\mathrm{m } \left(\hat{b}_\mathrm{l}^\dag \hat{b}_\mathrm{l } + \hat{b}_\mathrm{r}^\dag \hat{b}_\mathrm{r}\right ) + \frac{\omega_\mathrm{a}}{2}\hat{\sigma}_z + \\ & + j \left(\hat{a}^\dag_r \hat{a}_\mathrm{l } + h.c.\right ) + j_\mathrm{m } \left(\hat{b}^\dag_r \hat{b}_\mathrm{l } + h.c.\right)+ g\left(\hat{\sigma}_+ \hat{a}_\mathrm{l } + h.c .
\right ) . \end{split}\ ] ] we now introduce mechanical supermodes @xmath112 . substituting these in the hamiltonian
, we obtain @xmath113 \hat{a}_\mathrm{l}^\dag \hat{a}_\mathrm{l } + \left[\omega_\mathrm{c } - \frac{g_0}{\sqrt{2 } } \left(\hat{x}_+ - \hat{x}_- \right ) \right ] \hat{a}_\mathrm{r}^\dag \hat{a}_\mathrm{r } + \\ & + \left(\omega_\mathrm{m}+j_\mathrm{m } \right ) \hat{b}_+^\dag \hat{b}_+ + \left(\omega_\mathrm{m}-j_\mathrm{m}\right ) \hat{b}_-^\dag
\hat{b}_- + \frac{\omega_\mathrm{a}}{2}\hat{\sigma}_z + \\ & + j \left(\hat{a}^\dag_r \hat{a}_\mathrm{l } + h.c.\right ) + g\left(\hat{\sigma}_+ \hat{a}_\mathrm{l } + h.c .
\right ) , \end{split}\ ] ] where we have defined the dimensionless position operators of the mechanical supermodes @xmath114 .
the mechanical mode @xmath115 interacts dispersively with the modes @xmath107 and @xmath106 with equal and opposite coupling rate @xmath53 . in the following
, we will neglect the presence of the other mechanical supermode , @xmath116 . for simplicity ,
we replace @xmath117 and @xmath118 .
we moreover define the operator @xmath119 .
the new hamiltonian reads @xmath120 where we also performed a unitary transformation @xmath121 , with @xmath122 $ ] . assuming a quasi - static approximation for @xmath59 , valid in the limit @xmath123 , we can diagonalize the optical part of the hamiltonian @xcite .
the optical supermodes are defined by @xmath124 where @xmath125 and @xmath126 are operators defined by @xmath127 in the supermode basis , the hamiltonian is @xmath128 up to the first order in @xmath61 , @xmath125 and @xmath126 read @xmath129 by inserting these expansions in the hamiltonian in eq .
[ eq_hamiltonian2cavites ] , and expressing again @xmath59 as a function of @xmath15 and @xmath130 , we get @xmath131 the second row describes the rabi interaction of the emitter with the two optical supermodes , with a coupling rate @xmath132 . as mentioned in the main text , in the two - cavity system this interaction is unavoidable and is due to the fact that both optical supermodes have nonzero field in both cavities for all finite detunings .
the third row describes the tripartite interaction between the emitter , the phonon and the optical supermodes .
the model for the three - cavity case is schematically depicted in fig . [ figsuppinfo_model_3_cavities ] .
the three identical optical cavities , denoted left ( l ) , target ( t ) and right ( r ) are dispersively coupled at a rate @xmath52 with a separate mechanical resonator , in similar fasion as in the previous section .
the three mechanical resonators are assumed identical and with frequency @xmath14 .
we describe the three optical cavities with the annihilation operators @xmath106 , @xmath133 and @xmath107 , and the three resonators with annihilation operators @xmath108 , @xmath134 and @xmath109 . the optical cavity t is coupled to the cavities l and r with a rate @xmath40 , while the mechanical resonator @xmath134 is coupled with a rate @xmath110 to the resonators @xmath109 and @xmath108 .
finally , a two - level emitter is placed in the target cavity , and interacts with the field @xmath133 with a coupling rate @xmath19 . .
each cavity is dispersively coupled to a separate resonator with an optomechanical coupling rate @xmath52 .
the resonators are identical and have frequency @xmath14 .
the central mechanical resonator is additionally coupled to the other two mechanical resonators with a rate @xmath110 .
an emitter is placed in the central optical cavity and interacts with its optical mode with a coupling rate @xmath19 . ]
the full hamiltonian reads @xmath135
\hat{a}_\mathrm{l}^\dag \hat{a}_\mathrm{l}+ \left[\omega_\mathrm{c } - g_0 \left(\hat{b}_\mathrm{t}^\dag + \hat{b}_\mathrm{t } \right ) \right ] \hat{a}_\mathrm{t}^\dag \hat{a}_\mathrm{t } + \left[\omega_\mathrm{c } - g_0 \left(\hat{b}_\mathrm{r}^\dag + \hat{b}_\mathrm{r } \right ) \right ] \hat{a}_\mathrm{r}^\dag \hat{a}_\mathrm{r } + \\ & + \omega_\mathrm{m } \left(\hat{b}_\mathrm{l}^\dag \hat{b}_\mathrm{l } + \hat{b}_\mathrm{t}^\dag \hat{b}_\mathrm{t } + \hat{b}_\mathrm{r}^\dag
\hat{b}_\mathrm{r}\right ) + \frac{\omega_\mathrm{a}}{2}\hat{\sigma}_z + \\ & + j \left[\hat{a}_\mathrm{t}^\dag\left(\hat{a}_\mathrm{r}+ \hat{a}_\mathrm{l}\right ) + h.c.\right ] + j_\mathrm{m } \left[\hat{b}_\mathrm{t}^\dag\left(\hat{b}_\mathrm{r } + \hat{b}_\mathrm{l } \right ) + h.c.\right]+ g\left(\hat{\sigma}_+ \hat{a}_\mathrm{t } + h.c .
\right ) . \end{split}\ ] ] the mechanical part of this hamiltonian can be diagonalized by introducing the modes @xmath136 the transformed hamiltonian reads @xmath137 + g\left(\hat{\sigma}_+ \hat{a}_\mathrm{t } + h.c .
\right ) , \end{split}\end{aligned}\ ] ] where we have defined the dimensionless position operators of the mechanical supermodes @xmath114 and @xmath138 .
we now focus only on the mechanical mode @xmath139 with the assumption that , under resonant conditions , the terms involving the other mechanical modes are negligible ( this assumption is numerically verified in section [ sec_othermechanialmodes ] ) . for simplicity , we redefine @xmath140 . moreover , we perform a unitary transformation @xmath121 , with @xmath141 $ ] .
the new hamiltonian reads @xmath142 + g\left(\hat{\sigma}_+ \hat{a}_\mathrm{t } + h.c .
\right ) , \end{split}\end{aligned}\ ] ] where we defined @xmath143 .
we note that the selected mechanical mode has equal and opposite dispersive coupling with the optical cavities l and r , at a rate @xmath53 , while it does not affect the optical cavity t. we again assume a quasi - static approximation for @xmath59 , valid in the limit @xmath123 , to diagonalize the optical part of the hamiltonian @xcite .
the three optical supermodes are defined in this case by @xmath144 where @xmath145 , @xmath126 , @xmath72 and @xmath146 are operators . in particular , @xmath147 the functions @xmath146 and @xmath72 satisfy @xmath148 .
their expressions are more complicated and not reported here , as they are not needed in the following .
the hamiltonian in the supermode basis reads @xmath149 .
\end{split}\end{aligned}\ ] ] up to the first order in @xmath61 , @xmath126 and @xmath145 read @xmath150 by inserting these expansions in the hamiltonian in eq .
[ eq_hamiltonian3cavities_4 ] , and expressing again @xmath59 as a function of @xmath15 and @xmath130 , we get @xmath151 . \end{split}\end{aligned}\ ] ] the first term of the second row describes the mode field coupling between the emitter , the phonon and the mode @xmath152 , with a coupling rate @xmath153 .
the second term of the second row describes a rabi interaction between the emitter and the _ other _ two optical supermodes . as confirmed by the numerical simulations based on the full hamiltonian in eq .
[ eq_hamiltonian3cavities_3 ] ( see main text ) , the effect of these terms is negligible when the spectral separation between the supermodes is much larger than the mechanical frequency ( @xmath154 ) .
however , for finite optical linewidths ( @xmath81 ) , the presence of the modes @xmath60 will introduce additional decay channels for the qe , as explained below .
as mentioned in the main text , for the calculation in fig . 3b we assume a conservative value of @xmath74 = 0.05 ghz for the rate of the emitter decay into the leaky modes ( _ i.e. _ , all the optical modes different from the three supermodes of the system ) .
this value corresponds to the one measured for an nv center placed in a single defect cavity ( similar to the one used here to realize a three - cavity system ) and out of resonance with the optical mode @xcite .
however , significantly smaller radiative decay rates could in principle occur in the proposed structure , as we explain in the following . to evaluate the expected decay rate , we perform finite element method calculations ( comsol ) of the power emitted by an electric dipole ( fig . [ figsuppinfo - decayratein3cav ] ) placed in either the target ( blue line ) or the left cavity ( green line ) of the structure considered in the main text .
we then normalize the calculated powers by the power emitted by the same dipole in bulk diamond . in this way we obtain a prediction for the radiative decay rate of a qe ( with the same position and polarization of the dipole ) normalized to the radiative decay rate of the same qe in bulk . in agreement with the spatial patterns of the optical supermodes
( see also sec .
[ sec_design ] of this s.i . )
, a dipole placed in the lateral cavity can feed all the three optical supermodes , while one that is located in the target cavity does not emit into the mode @xmath152 . in particular
, the calculation shows that the radiative decay rate of a qe placed in the cavity t is reduced by almost a factor @xmath155 with respect to the decay rate in bulk .
as the measured lifetime of an nv center in bulk is about 12 ns @xcite , a radiative decay rate of the order of @xmath74 = 0.1 mhz is in principle expected for the same emitter placed in the target cavity .
we also note that the emission of a dipole in the target cavity is dominated by the coupling with the nonresonant coupled modes for the chosen set of parameters ( see sec . [ sec_additionaldecays ] ) , while the emission into leaky modes is expected to be even smaller .
we therefore conclude that the assumed value of 50 mhz is indeed a conservative , upper estimate of the qe decay rate . of this si .
the powers ( @xmath156 ) are normalized by the power emitted in bulk diamond ( @xmath157 ) . ]
as shown in the main text , for a proper parameter choice the hamiltonian of the three - cavity system provides a tripartite interaction between the supermode @xmath152 , the qe and the mechanical mode .
the interaction between the qe and the modes @xmath60 is suppressed for large supermode separation ( @xmath154 ) .
however , these optical modes are still coupled with the emitter ( since their fields at the emitter position is not zero ) , as shown in fig .
[ figsuppinfo - decayratein3cav ] . for large optical losses @xmath160 , the supermodes @xmath60
are weakly coupled to the qe and therefore they introduce additional decay channels ( denoted @xmath161 ) .
these additional decay channels become relevant when the other radiative decay channel ( @xmath84 ) is small and the optical linewidth is significant .
this effect is important in understanding some of the features shown in the main text about the swapping fidelity and the ground - state cooling ( see fig .
3(c - d ) and accompanying discussion in the main text ) .
@xmath161 can be calculated as follows .
if an optical mode with losses @xmath81 is resonant with a qe and coupled with it at a rate @xmath162 , the qe decays into the optical mode at a rate @xmath163 . in the presence of a large spectral detuning @xmath164 , the decay rate is modified into @xmath165 @xcite . in our case ,
@xmath166 and we replace @xmath167 , consistently with the notation in the main text where @xmath19 is the coupling between the qe and the uncoupled cavity . therefore , @xmath168 with respect to the qe - phonon swapping this additional decay channel is negligible when @xmath169 , which corresponds to the vertical dashed - dotted line in fig .
3c of the main text .
in the three - cavity system the mode field coupling is due to a particular optical supermode ( @xmath152 ) , which , when the system is not perturbed by the mechanical displacement , features zero electric field in the central cavity and equal and opposite field amplitudes in the lateral cavities ( see the first of eqs .
[ eq_hamiltonian3cavities_defcoupledmodes ] for @xmath170 ) . in order to enhance the mode field coupling it is therefore necessary to selectively pump this mode , without feeding the other two optical supermodes .
this is possible by pumping the two lateral cavities at a frequency @xmath7 with equal field amplitude @xmath171 and opposite phases : excitation of modes @xmath60 is then symmetry - forbidden .
focusing only on the optical part of the hamiltonian , we have @xmath172 + \mathcal{e}\left(\hat{a}^\dag_r - \hat{a}^\dag_l + h.c.\right ) , \end{split}\end{aligned}\ ] ] where we switched to a frame rotating at frequency @xmath7 . by writing the equations of motion for the three cavity field amplitudes , and assuming equal cavity losses @xmath81
, we can calculate the steady state population under this pumping scheme , namely @xmath173 and @xmath174 . by comparing these results with eqs .
[ eq_hamiltonian3cavities_defcoupledmodes ] for @xmath170 , we see that only the mode @xmath152 is fed , while the population of the other two modes remains strictly zero .
interestingly , in the limit of validity of the coupled mode theory , this result ( _ i.e. _ the zero population of the supermodes @xmath158 and @xmath159 ) is independent of @xmath81 .
in the previous section we explained how the mode @xmath152 can be selectively pumped .
however , solving a full quantum model ( eq .
[ eq_hamiltonian3cavities_3 ] ) with a large pumping of the optical cavities is computationally extremely challenging , because of the large fock space dimensions required .
we therefore solve a transformed version of the hamiltonian in eq .
[ eq_hamiltonian3cavities_3 ] , obtained through a displacement of the cavity operators .
that is , we replace the operators by the sum of a steady - state amplitude and a fluctuating operator , _ i.e. _ @xmath175 , where @xmath176 l , r , t .
the transformed hamiltonian reads @xmath177 \end{split}\end{aligned}\ ] ] where we defined @xmath178 and we have used the fact that , with the pumping scheme described in the previous section , @xmath179 and @xmath180 = 0 .
we note that this hamiltonian is still equivalent to the original one , since no terms have been neglected .
the hamiltonian has been solved numerically ( with the method discussed in sec .
[ sec_mastereq ] ) by either setting @xmath34 = 0 ( for the case in which the mode @xmath152 is not externally pumped ) or setting a fixed value of @xmath34 ( continuous optical pumping of the system ) . to address the case of a square - pulse excitation , we first calculate the time evolution of the ( classical ) cavity fields amplitudes upon the external pumping described above , neglecting the presence of the qe and the mechanical mode . from the amplitudes @xmath181 and @xmath182 ,
we calculate the population of the mode @xmath152 ( @xmath183 ) , which is used as a time - dependent parameter in solving the hamiltonian in eq .
[ eq_hamiltonian3cavities_linearized ] .
and the population of the mode @xmath152 .
the dashed black line indicates the condition @xmath184 .
( b ) fidelity of the emitter - phonon swapping versus the emitter decay rate @xmath84 and the emitter pure dephasing rate @xmath105 .
the horizontal and vertical dashed black line indicate the conditions @xmath185 and @xmath186 , respectively . ]
as mentioned in the main text , the mode field coupling rate can be enhanced by increasing the population of the mode @xmath152 .
this is beneficial for , @xmath187 , the emitter - phonon swapping fidelity .
the colormap in fig .
[ figsuppinfo - fidelityvsncav_dephasing]a shows the swapping fidelity ( calculated with the linearized hamiltonian of the three - cavity system , eq .
[ eq_hamiltonian3cavities_linearized ] ) versus the emitter decay rate @xmath84 and the population of the mode @xmath152 .
all the other parameters are the same as in the main text .
as expected , near - unity fidelity is obtained when @xmath188 . due to the coherent nature of the mfc , pure dephasing of the qe
is expected to further decrease the swapping fidelity . as shown in fig .
[ figsuppinfo - fidelityvsncav_dephasing]b , the qe decay @xmath84 and pure dephasing rates @xmath105 play a similar role in determining the swapping fidelity , and @xmath189 is required as well to achieve large fidelity .
we note that for the system considered for the calculations shown in this work ( nv centers in diamond ) , the pure dephasing rate is typically of the same order of magnitude of the decay rate in bulk @xcite .
, _ i.e. _ considering all mechanical modes .
the parameters of the system are the same as in fig .
3a of the main text .
( a - b ) vacuum oscillations of the system ( _ i.e. _ the optical cavities are not pumped ) . ( a ) the mechanical modes @xmath108 and @xmath109 are equally populated ( blue solid and blue dotted lines ) and reach a maximum occupation probability of 0.5 , while the mechanical mode @xmath134 is not populated ( cyan solid line ) .
( b ) when switching to the supermode basis ( through eqs .
[ eq_hamiltonian3cavities_coupledmechmodes ] ) , the mode @xmath139 behaves exactly as shown in previous calculations in which the other supermodes were neglected .
( c - d ) same as in panels ( a - b ) but with an external pumping of the optical cavities , such that the optical supermode @xmath152 contains @xmath190 photons . ] in the main text , and in the previous sections of this supplemental information , we assumed that , once the mechanical modes are hybridized , only the mechanical supermode of interest is relevant for the system dynamics . to verify this assumption ,
we perform additional numerical calculations based on the hamiltonian in eq .
[ eq_hamiltonian3cavities_1 ] , _ i.e. _ by considering explicitly the three mechanical uncoupled modes .
we use the same parameters as in the main text , and @xmath191 mhz .
initially , we assume no mechanical losses ( @xmath192 ) . after calculating the time evolution of the modes @xmath108 , @xmath109 and @xmath134
, we derive the time evolution of the three mechanical supermodes based on eqs .
[ eq_hamiltonian3cavities_coupledmechmodes ] . as shown in fig .
[ figsuppinfo_other2mechanicalmodes ] , both in the unpumped and the pumped case the dynamics of the mechanical mode @xmath139 matches that found previously , where the other mechanical modes were neglected .
the population of the other two mechanical supermodes ( @xmath116 and @xmath115 ) is always zero ( not shown in these plots ) .
this is due to the fact that the other two mechanical supermodes interact with the optical cavities in a way which does not lead to any tripartite interaction ( see eq .
[ eq_hamiltonian3cavities_2 ] ) , and their presence is therefore negligible for proper choice of frequencies .
indeed , we also verified ( fig .
[ figsuppinfo - fidelityvs_jm_gammam]a ) that the swapping fidelity does not depend on the mechanical interaction @xmath110 even for large mechanical losses @xmath193 .
decrease of the fidelity is only observed if the mechanical losses become larger than @xmath95 .
the fact that the the presence of the other two mechanical supermodes can be completely neglected is also confirmed by fig .
[ figsuppinfo - fidelityvs_jm_gammam]b : here , we compare the swapping fidelity versus @xmath103 for the case of a three - cavity system ( _ i.e. _ a horizontal cut of panel a ) with the same graph calculated for a system composed by one optical cavity and one mechanical resonator , where the mfc is introduced _ ad - hoc _ ( eq . 2 of the main text ) .
the identical behaviours indicate that a large @xmath103 introduces only additional losses but does not make the system interact with the other mechanical supermodes .
the small deviation in the fidelity for @xmath194 is due to the additional losses present in the three - cavity system due to the optical modes @xmath60 .
and the mechanical losses @xmath103 .
all the other parameters are the same as in the main text and the qe losses have been neglected . the system is pumped with @xmath195 .
the dashed - dotted vertical line indicates the condition @xmath196 .
( b ) horizontal cut of panel ( a ) for the lowest value of @xmath110 ( solid blue line ) compared with the fidelity calculated for the same mechanical losses but in the mfc model with one cavity and one resonator ( green dashed line ) , see eq . 2 of main text .
-component of the electric field is shown .
the mode frequencies and the optical losses are indicated in the figure .
( b - c ) three identical cavities on a nanobeam , separated by 2 periods .
the position of the three cavities is marked by the vertical grey dashed lines .
( b ) electric field pattern of the three optical supermodes when the mechanical mode is at rest .
( c ) displacement pattern of the three mechanical supermodes .
( d ) frequencies of the three optical supermodes as a function of the cavities separation ( _ i.e. _ number of periods ) .
( e ) values of the optical interaction @xmath40 ( extrapolated from the plot in panel ( d ) , see text ) versus the cavity separation . ]
the structure proposed in the main text is composed of three in - line defect cavities in a photonic crystal diamond nanobeam .
simulations have been performed with a finite element method commercial software ( comsol ) .
we first designed the single cavity structure ( fig .
[ figsuppinfo_designcavities_coupledmodes]a ) , which is realized by quadratically tapering the lattice constant across 9 lattice periods , from the unperturbed value of @xmath197 nm to the central value of @xmath198 nm .
the design is inspired by the one proposed recently by lee _
_ et al.__@xcite .
the holes have elliptical shape , with semiaxes equal to @xmath199 and @xmath200 .
the width of the nanobeam ( along the @xmath201-direction ) is @xmath202 nm , and the thickness is 200 nm .
this structure supports an optical mode with frequency @xmath203 thz and @xmath204 ( fig .
[ figsuppinfo_designcavities_coupledmodes]a , top panel ) , and a co - localized mechanical mode with frequency @xmath205 ghz ( fig .
[ figsuppinfo_designcavities_coupledmodes]a , bottom panel ) .
the dispersive optomechanical coupling between the optical and mechanical modes has been calculated by evaluating the moving - boundary and photoelastic contribution separately ( as discussed in ref . ) and amounts to @xmath206 mhz .
we notice that we did not perform any systematic optimization on our design , and therefore the optical q factor and optomechanical coupling rate could be further increased by carefully optimizing the design parameters ( _ e.g. _ , by tapering also the holes semiaxes ) as discussed by other authors @xcite . after having chosen a design for the single cavity , we considered three identical defect cavities on the same nanobeam , separated by an equal distance @xmath207 , defined as the number of unperturbed periods between two adjacent cavities . for each distance @xmath207 , we calculated the optical and mechanical supermodes . figure [ figsuppinfo_designcavities_coupledmodes]b shows the @xmath208 field component for the three optical supermodes .
note that the optical mode of interest , @xmath152 , has no field in the central cavity , differently from the other two modes .
the modes @xmath158 and @xmath159 differ in the relative sign between the field in the lateral cavities and the field in the central cavity , as expected from the results of the coupled - mode theory ( eq . [ eq_hamiltonian3cavities_defcoupledmodes ] ) .
figure [ figsuppinfo_designcavities_coupledmodes]c shows the displacement pattern of three mechanical supermodes .
similarly to the optical case , the mechanical supermodes @xmath116 and @xmath115 have opposing oscillation phase in the central and the lateral cavities .
figure [ figsuppinfo_designcavities_coupledmodes]d shows the frequencies of the three optical supermodes versus the cavity separation @xmath207 .
the frequency of the mode @xmath152 is expected to be independent of the cavity interaction @xmath40 ( and therefore the cavity distances ) . the small frequency deviations observed can be due to either the finite mesh size ( which makes the cavities slightly different from each other ) or to the breakdown of the coupled - mode theory for very short cavity distances .
we notice that this does not affect our theoretical model , since the mechanical movement does not change the rate @xmath40 . in fig .
[ figsuppinfo_designcavities_coupledmodes]e we show the estimated optical interaction rate @xmath40 as a function of the inter - cavity distance . for each distance @xmath207 ,
@xmath40 has been estimated by the formula @xmath209 . in fig .
2i of the main text we show the expected pattern of the mode @xmath152 for the case in which the system is mechanically perturbed , such that the lateral cavities are detuned by @xmath210 .
the shown electric field pattern has been calculated analytically from the formula @xmath211 where @xmath212 , with @xmath213 , are the electric field patterns of the supermodes when the mechanical mode is at rest ( fig . [ figsuppinfo_designcavities_coupledmodes]b ) , and
the coefficients @xmath214 are calculated analytical from the coupled mode theory .
the interaction described in this work relies on a variation of the electric field at the emitter position upon a mechanical displacement .
a small field variation occurs in principle also in a simple fabry - perot ( fp ) cavity when one of the mirrors is displaced by a mechanical resonator @xcite .
however , the interaction strength in this case is much smaller than the one obtained in the two and three cavity systems . in particular , for an emitter placed in a node of the @xmath215-th fp mode , the tripartite coupling rate reads @xmath216 , where @xmath19 is the maximum rabi coupling between the qe and the @xmath215-th mode , @xmath52 is the dispersive coupling induced by the mirror movement and @xmath217 is the fundamental cavity frequency .
the coupling rate bears similarity with the one derived for the two- and three - cavity systems , but with the important difference that the role of the intercavity interaction rate @xmath40 is taken by @xmath217 , which , in the fp cavity , corresponds also to the frequency spacing between the unperturbed modes .
this , in the visible and near - ir regime , severely limits the achievable values of @xmath218 , especially since any effort to reduce @xmath217 reduces both @xmath19 and @xmath52 .
10ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12
& 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | we propose a novel type of optomechanical coupling which enables a tripartite interaction between a quantum emitter , an optical mode and a macroscopic mechanical oscillator .
the interaction uses a mechanism we term _ mode field coupling _ : mechanical displacement modifies the spatial distribution of the optical mode field , which in turn modulates the atom - photon coupling rate . in properly designed multimode optomechanical systems , we can achieve situations in which mode field coupling is the only possible interaction pathway for the system .
this enables , for example , swapping of a single excitation between emitter and phonon , creation of nonclassical states of motion and mechanical ground - state cooling in the bad - cavity regime .
importantly , the emitter - phonon coupling rate can be enhanced through an optical drive field , allowing active control of strong atom - phonon coupling for realistic experimental parameters .
interfacing different quantum systems , such as atoms , photons , and phonons , is a key requirement for quantum information processing .
the well - established framework of cavity quantum electrodynamics ( cqed ) interfaces photons ideal for communication to natural or artifical atoms ( quantum emitters , qes ) , whose strong nonlinearities enable quantum processing .
mechanical resonators have recently come to the forefront due to their large coherence times and their interaction with photons in cavity optomechanical systems @xcite .
moreover , the creation of nonclassical states in macroscopic mechanical systems is appealing for fundamental studies of quantum physics @xcite . in these contexts , establishing an efficient and controllable interaction between phonons and qes would be highly beneficial , as it would enable using the qe nonlinearity for the creation and manipulation of phononic quantum states @xcite .
different approaches have been proposed to realize such an interaction .
first , a phonon can directly couple to a solid - state qe through mechanical strain @xcite . despite the large coupling rates obtainable in specific systems , this mechanism is difficult to engineer and to dynamically control . a second approach couples mechanical modes dispersively to an optical cavity , which in turn interacts with a qe @xcite .
tripartite entanglement and atom - assisted optomechanical cooling are predicted in so - far elusive regimes when the optomechanical interaction is nonlinear at the quantum level @xcite or when the emitter - field coupling rate approaches the emitter frequency @xcite .
additionally , qe - phonon interaction occurs in molecules and solids when the electronic and vibrational degrees of freedom are coupled , leading to inelastic scattering processes .
natural raman transitions have been used to transfer a photon s quantum state to an optical phonon in diamond @xcite , but the extremely high frequency and large dissipation limit general application for quantum processing . with respect to the cavity
stimulates transitions from @xmath0 to @xmath1 ( @xmath2 ) .
( c ) time evolution dictated by the hamiltonian in eq .
[ eq_hamitiltonian_vfc ] , with the qe excited at @xmath3 and @xmath4 . ] here , we propose a novel optomechanical effect that provides an explicit , engineerable , and optically controllable interaction between a qe and a macroscopic mechanical oscillator .
the interaction arises from a mechanically - induced modification of the spatial distribution of the optical field ( fig .
[ fig1_intro]a ) , which in turn modulates the qe - photon coupling rate .
we term this interaction _ mode field coupling _ ( mfc ) .
we show that in simple multicavity optomechanical systems mfc is the only possible interaction for the system , enabling , _
e.g. _ , qe - phonon excitation swapping and mechanical ground - state cooling in the bad - cavity regime .
importantly , the interaction strength can be controlled and enhanced by the optical field intensity , resulting in optically - controlled emitter - phonon coherent manipulation .
this coupling , and the resulting hamiltonian , share important traits with raman - like processes in trapped ions @xcite , which has proven powerful in controlling the motional state of single ions .
mfc has however two distinct features : it involves large - mass macroscopic resonators , and its rate is nonetheless large enough to overcome the large decoherence typical of solid - state qes .
_ model .
_ we consider a standard cqed setup , in which a two - level qe couples to an optical cavity mode through the hamiltonian @xmath5 , where @xmath6 ( @xmath7 ) denotes the qe ( optical mode ) frequency , @xmath8 is the photon annihilation operator , @xmath9 are pauli operators describing the qe and @xmath10 .
we initially neglect any loss , focusing on conservative interactions . the qe - field coupling rate @xmath11 is determined by the emitter s transition dipole moment @xmath12 and the electric field per photon @xmath13 of the optical mode at the emitter position .
next , we consider a mechanical oscillator with frequency @xmath14 and phonon annihilation operator @xmath15 . in a standard
dispersively coupled optomechanical system , the resonator s displacement @xmath16 affects the optical cavity frequency .
this interaction is quantified by the coupling rate @xmath17 , where @xmath18 is the zero - point motion amplitude . here
we consider a fundamentally different situation , in which the mechanical displacement induces a variation of the spatial distribution of the cavity field ( fig .
[ fig1_intro]a ) , while @xmath7 is negligibly affected .
as a direct consequence the emitter - cavity coupling rate @xmath19 becomes dependent on mechanical position . up to first order in @xmath20 , @xmath21 , where we defined the mfc coupling rate @xmath22 . inserting this expression in the hamiltonian @xmath23 leads to the appearance of a tripartite interaction between the qe , the optical field , and the mechanical resonator . in the specific case that at mechanical equilibrium the field at the emitter s position vanishes ( fig .
[ fig1_intro]a ) , @xmath24 and the only possible interaction channel for the system is the tripartite one .
the interaction hamiltonian reads @xmath25 this hamiltonian allows swapping the excitation between the three quantum systems under particular resonant conditions . for @xmath26 ( @xmath27 ) ,
the dominant term is @xmath28 ( @xmath29 ) , describing phonon and qe excitation upon annihilation of a photon ( excitation of the qe due to photon and phonon annihilation ) and the reverse process .
depending on the photon energy , therefore , the transitions @xmath30 are realized ( fig .
[ fig1_intro]b ) , where @xmath31 ( @xmath19 ) denotes the qe excited ( ground ) state , and @xmath32 the phonon number .
figure [ fig1_intro]c shows the lossless time evolution described by eq .
[ eq_hamitiltonian_vfc ] for @xmath4 , with only the qe excited at @xmath3 .
the excitation oscillates , at a frequency @xmath33 , between the qe and the state formed by one photon and one phonon .
next , we consider pumping the cavity with a large coherent field to an average photon number @xmath34 , writing the cavity field as @xmath35
. neglecting for now the fluctuations @xmath36 ( valid for @xmath37 , the hamiltonian reads @xmath38 which describes a coherent qe - phonon interaction , with a coupling rate controlled by @xmath39 .
thus , the optical intensity can enhance the qe - phonon coupling and , in particular , overcome system losses .
_ creating large field variation .
_ as the mode field is the solution of an eigenvalue problem @xcite , we look for a mechanical perturbation that induces strong changes of the eigenvector without affecting the eigenvalue .
such an effect is maximized for quasi - degenerate unperturbed eigenvalues , which can be obtained by coupling two or more cavities such that hybridized modes ( ` supermodes ' ) with well - defined symmetry are formed . near a symmetry point , _ i.e. _ , an anticrossing , an odd perturbation breaks symmetry .
this localizes the supermodes in one of the cavities , resulting in a large variation of the local mode field .
indeed , we find an example of mfc ( fig .
[ fig2_2and3cavities]a ) in two identical optical cavities coupled with rate @xmath40 ( a _ membrane - in - the - middle _ setup @xcite ) .
a mechanical displacement that induces opposite detuning @xmath41 to each cavity affects the spatial distribution of the supermode amplitudes , and thereby their coupling rate ( @xmath42 ) with an emitter placed in one cavity ( fig . [
fig2_2and3cavities]c ) .
for @xmath43 the supermode frequencies are constant ( fig .
[ fig2_2and3cavities]b ) , _ i.e. _ , dispersive coupling is absent .
the mfc coupling rate @xmath44 scales as @xmath45 ( see supplemental information @xcite ) : for weakly interacting cavities ( @xmath46 ) , small deviations from the condition @xmath47 quickly lead to localization of the supermodes into the individual cavities . in the two - cavity system , however , the tripartite mfc interaction competes with the rabi emitter - photon interaction as @xmath48 ( fig . [ fig2_2and3cavities]c ) .
this direct qe - photon interaction can be suppressed by introducing an additional optical cavity with identical frequency ( fig . [ fig2_2and3cavities]d ) .
the middle cavity ( t ) , which contains the qe , interacts with both lateral cavities with rate @xmath40 , leading to the formation of three supermodes @xmath49 , @xmath50 and @xmath51 @xcite .
the mode @xmath51 has opposite fields in the lateral cavities and zero field in t ( blue line in fig .
[ fig2_2and3cavities]d ) , and therefore does not interact with the qe . we now consider a mechanical mode that detunes only the frequencies of the lateral cavities by @xmath41 , while leaving t unperturbed .
this could be realized for example by rigidly connecting the two membranes .
more generally , it can be obtained by dispersively coupling each optical cavity , at a rate @xmath52 , to a mechanical oscillator @xcite .
if these three oscillators are coupled mechanically , one resulting mechanical supermode has equal and opposite dispersive interaction with the lateral cavities with a rate @xmath53 , and zero interaction with t. the frequency of @xmath51 is unaltered by such detuning ( fig .
[ fig2_2and3cavities]e ) , while its field in t assumes a finite value ( fig .
[ fig2_2and3cavities]d , dashed line ) , which translates in a large modulation of the coupling rate @xmath54 between @xmath51 and the qe around the value @xmath55 ( fig .
[ fig2_2and3cavities]f ) .
therefore , the interaction between the emitter , the mode @xmath51 and the selected mechanical mode will be described by the hamiltonian in eq .
[ eq_hamitiltonian_vfc ] .
figure [ fig2_2and3cavities](g - h ) shows an implementation of this model in a photonic crystal nanobeam .
cavities are defined by local variations of the periodicity , which results in co - localized and dispersively coupled optical and mechanical resonances @xcite .
three defect cavities are placed on the same nanobeam , leading to both optical and mechanical hybridization .
inter - cavity separation controls the optical interaction rate @xmath40 .
the electric field of @xmath51 ( fig .
[ fig2_2and3cavities]g ) is zero in the central cavity when the mechanical mode is at rest .
figure [ fig2_2and3cavities]h shows the mechanical mode that provides the required detuning on the lateral cavities . upon mechanically - induced detuning of the lateral cavities , the mode @xmath51 acquires a finite electric field in the central cavity ( fig .
[ fig2_2and3cavities]i ) . _ full model and numerical calculations .
_ we now analyse the three - cavity system in detail , and show that it behaves as predicted by the hamiltonian in eq .
[ eq_hamitiltonian_vfc ] . for simplicity
, we consider only one of the hybridized mechanical supermodes , described by the operator @xmath15 , frequency @xmath14 and dispersively coupled to the lateral cavities at a rate @xmath53 .
the validity of this approach is justified in @xcite . in a frame rotating at @xmath7 , the hamiltonian is @xmath56 + g\left ( \hat{a}_\mathrm{t } \hat{\sigma}_+ + h.c .
\right ) , \end{split}\end{aligned}\ ] ] where we defined @xmath57 .
the first two terms describe the mechanically - induced detuning on the lateral cavities .
the second row describes optical mode coupling and the rabi interaction between emitter and cavity t. assuming a quasi - static approximation @xcite for the mechanical motion ( valid for @xmath58 ) , we can treat @xmath59 quasi - statically and diagonalize the optical part of the hamiltonian by introducing three optical supermodes , @xmath60 and @xmath51 .
up to the first order in @xmath61 , we obtain an interaction hamiltonian @xmath62 . \end{split}\end{aligned}\ ] ] the first term of eq .
[ eq_full3cavities_hamiltonian_coupledmodes_firstorder ] shows the tripartite interaction explicitly , with @xmath63 .
the last term describes a rabi interaction between the emitter and the supermodes @xmath60 with coupling rate @xmath64 .
a pure tripartite interaction can therefore be obtained for large supermode separation ( @xmath58 ) and tripartite resonance ( @xmath65 ) . additionally , to let the emitter interact with the supermodes ( and not the uncoupled modes ) , we require @xmath66 . to verify that the predicted coherent emitter - phonon interaction occurs in a realistic scenario
, we numerically solve @xcite the master equation derived from the full hamiltonian in eq .
[ eq_full3cavities_hamiltonian ] . out of the many possible systems
, we consider the structure of fig .
[ fig2_2and3cavities](g - h ) made in diamond with a nitrogen vacancy ( nv ) center as emitter .
the simulated parameters for this system are \{@xmath7 , @xmath14 , @xmath19 , @xmath52 } = @xmath67\{4.7@xmath68 , 14 , 20 , 0.004 } ghz .
we consider the case @xmath69 and @xmath70 , corresponding to one period cavity separation @xcite .
the unitary evolution of the system starting with the emitter excited ( fig .
[ fig3_numericalcalculations1]a ) agrees perfectly with that of the mfc hamiltonian ( fig .
[ fig1_intro]c ) and verifies the predicted qe - phonon oscillation period @xmath71 = 4.5 @xmath72s .
this confirms that for realistic choices of parameters purely tripartite interaction is obtained in the three - cavity system . in order to overcome losses , unavoidable in an experimental setting
, the coupling rate can be enhanced by selectively pumping the supermode @xmath51 @xcite .
figure [ fig3_numericalcalculations1]b shows the evolution of the same system as in fig . [ fig3_numericalcalculations1]a , now with dissipations introduced through lindblad operators @xcite , with a cavity decay rate @xmath73 = 1 ghz and a conservative emitter decay rate @xmath74 = 0.05 ghz ( measured for nv centers in photonic crystal structures @xcite ) . in fact , much smaller radiative decay rates are in principle expected here @xcite .
the mode @xmath51 is continuously pumped to a steady - state population of @xmath39 = @xmath75 , shown to be experimentally feasible in diamond @xcite .
we note that the emitter is not directly affected by the large optical intensity as the field is zero at the emitter position . at @xmath76
the qe is excited and interacts with the mechanical mode with a coupling rate @xmath77 . at @xmath78 the pump
is switched off , suppressing the interaction and leaving the system in a long - lived nonclassical state with phonon population @xmath79 0.8 .
the swapping fidelity can be made arbitrarily close to one ( fig .
[ fig3_numericalcalculations1]c ) by reducing the qe decay rate ( so that @xmath80 ) and the optical losses @xmath81 of the supermodes @xmath60 , which introduce additional decay for the qe due to the finite optical linewidth .
this decay is negligible when @xmath82 ( vertical dashed - dotted line in fig .
[ fig3_numericalcalculations1]c ) @xcite .
the influence of @xmath39 and qe dephasing on the swapping fidelity is discussed in the supplemental information @xcite .
, parameters in text ) .
( a ) lossless evolution starting from an excited qe without optical pumping .
( b ) as in ( a ) but assuming losses for the qe and optical cavity .
the mode @xmath51 is continuosly pumped until t @xmath83 8 ns , swapping the excitation from emitter to resonator and creating a nonclassical motional state .
( c ) fidelity of the qe - phonon swapping versus @xmath84 and @xmath81 for @xmath85 .
the horizontal dashed line indicates the condition @xmath86 , while the vertical dashed - dotted line indicates @xmath87 .
( d ) cooling through mfc .
steady - state mean phonon number versus @xmath84 and @xmath39 for continuous optical pumping and @xmath88 . ]
the proposed qe - phonon interaction can also be used to cool the mechanical resonator to its ground state .
the cooling cycle is triggered by a red - detuned cavity photon which excites the qe upon annihilation of a phonon .
the excitation is subsequently dissipated through the qe decay .
differently from standard optomechanical cooling @xcite , this mechanism can achieve ground state cooling in the bad cavity regime ( @xmath89 ) , while the sideband resolved regime is required only for the qe ( @xmath90 ) .
figure [ fig3_numericalcalculations1]d shows the steady - state phonon population in the three - cavity system as a function of @xmath84 and @xmath34 , for @xmath88 and for finite mechanical losses ( @xmath91 khz ) and thermal phonon occupation ( @xmath92 ) . as expected , ground - state cooling is possible for @xmath93 .
phonon population lower than 0.1 can be achieved with @xmath94 and realistic qe decay rates .
the phonon population increase for large @xmath34 is attributed to the onset of ultra - strong coupling , as @xmath95 approaches @xmath14 . for small @xmath84 ,
the qe total decay rate is dominated by the additional emission into the supermodes @xmath60 ( which read @xmath96 ghz @xcite ) , which explains the saturation of the phonon population for @xmath97 . in conclusion
, we have introduced a new kind of emitter - photon - phonon interaction in hybrid - optomechanical systems , based on mechanically - induced variation of the electric field spatial pattern .
the coupling rate can be particularly strong in multicavity systems with small coupling rate , as it scales inversely with the coupling rate @xmath40 . for large optical drives
, this mechanism leads to an emitter - phonon coherent interaction whose coupling strength is controlled by the optical intensity .
emitter - phonon excitation swapping and mechanical ground - state cooling are possible with feasible experimental parameters .
the proposed interaction strength is much larger than effects obtainable in single - mode systems , which require the ultra - strong coupling regime ( @xmath98 ) to have comparable rates @xcite .
differently from strain - based methods @xcite , the proposed coupling mechanism is not limited to a specific choice of the solid - state emitter and material system , and it could even be applied to atoms trapped near a mechanical resonator @xcite .
moreover , it provides strong quantum nonlinearity without requiring the single - photon strong optomechanical coupling regime ( @xmath99 ) . in perspective
, the optically controlled coherent emitter - interaction introduced here paves the way for , _
e.g. _ , control of spontaneous phonon emission , creation of nonclassical states of motion and phonon lasing .
the authors acknowledge l. midolo for first pointing out the possibility of displacement - induced field variations in optomechanical cavities .
this work is part of the research programme of the foundation for fundamental research on matter ( fom ) , which is financially supported by the netherlands organisation for scientific research ( nwo ) .
e.v . gratefully acknowledges an nwo - vidi grant for financial support .
27ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) |
even though the first transiting extrasolar planet ( tep ) was discovered only as recently as 1999 @xcite , the number of recognised teps and the size of the literature on them has grown very quickly .
the known planetary systems now display a bewildering diversity of qualities : from super - earths with extremely short periods ( corot-7b ; * ? ? ?
* ) to massive planets on wide and highly eccentric orbits ( hd80606b ; * ? ? ?
* ) ; a system of six planets showing strong mutual gravitational perturbations ( kepler-11 ; * ? ? ?
* ) ; circumbinary planets @xcite ; and ones with a size inexplicably approaching twice that of jupiter ( wasp-17b ; * ? ? ?
* ; * ? ? ?
part of the reason for the known tep population being such a menagerie is the variety of methods deployed in catching them
. many of the teps orbiting brighter stars ( including the first , hd209458b ) were found by radial velocity surveys and later shown to transit .
others ( such as the second known tep ogle - tr-56b ; @xcite ) were unearthed via deep photometric searches and subsequently confirmed spectroscopically .
the dominant population of teps currently stems from large - scale photometric surveys of stars of intermediate brightness ( @xmath7@xmath8 ) coupled with spectroscopic follow - up programs , such as those of hat @xcite and wasp @xcite .
this population is strongly biased towards larger planets with orbital periods @xmath9d , as these _ hot jupiters _ show the greatest photometric variability .
more recently , surveys such as corot and _ kepler _ have found success by obtaining uninterrupted photometry of stars using space satellites .
the objects found in this way like the deep ground - based surveys
mostly orbit fainter stars , which hinders attempts to obtain follow - up observations to refine their physical properties or even spectroscopically confirm the planetary nature of the transiting body .
an important problem with the study of teps is that their physical properties can not be obtained simply by measuring some observable quantities and putting them through standard formulae : the number of quantities measurable directly from observations is one too few .
an _ additional constraint _ is therefore needed , which is usually obtained by requiring the physical properties of the host star to match the predictions of stellar theory . in practise
, there are several ways of implementing this constraint , and multiple sets of theoretical predictions from which to choose .
as different researchers select different approaches , their results are inhomogeneous and therefore non - trivial to compare .
the diversity of the known teps , coupled with the varied methods used in their analysis , makes a consistent picture of their properties more difficult to attain .
the current series of papers is an attempt to perform a uniform analysis for all suitable tep systems , in order to produce homogeneous measurements of their physical properties .
such results are valuable in statistical investigations of the characteristics of this population ( such as work similar to that by @xcite ) , as well as providing an independent confirmation ( or otherwise ) of published numbers . in some cases it is possible to achieve improved results simply by considering all available data rather than concentrating on only one set of observations .
in paperi @xcite i presented the methods used for analysing the transit photometry of tep systems , and applied them to the fourteen object which had good light curves at that point .
paperii @xcite discussed the more thorny issue of how to turn the measured photometric and spectroscopic parameters into physical properties ; the primary problem here is the need to incorporate the predictions of theoretical stellar evolutionary models ( or some other additional constraint ) into the process .
this was achieved for the fourteen teps by using three sets of stellar models , plus an empirical stellar mass radius relation as an alternative constraint . in paperiii
@xcite the net was widened to include the full analysis of 30 teps with five sets of stellar models , plus a description of the methods used to incorporate contaminating light and orbital eccentricity into the analysis .
paperiv @xcite further extended the project to include 58 teps , concentrating on those which had space - based light curves from corot , _ kepler _ , or the nasa epoxi satellites .
an improved empirical constraint was also used ( based on stellar density , temperature and metal abundance ) , calibrated on the measured properties of 180 stars in 90 detached eclipsing binaries . in this work
i present new photometric analyses of 30 more teps , and measure the physical properties of these plus a further eight objects previously considered in the current project .
this brings the total number of teps within the _ homogeneous studies _ project to 82 . whilst this is by far the largest source of homogeneous physical properties
, it still represents less than half of the known tep systems .
the physical properties , follow - up status , sky positions and discovery rate of the known teps are finally discussed in sections [ sec : properties ] to [ sec : discpos ] . as with previous papers in this
series , extensive tables of intermediate results are available for inspection in the online - only supplementary information .
i have studied the light curves of each tep system using the methods discussed in detail in paperi .
the jktebop code was used to find the best - fitting model for each light curve , from which the photometric parameters were extracted .
jktebop represents the star and planet as biaxial spheroids with shapes governed by the mass ratio .
the results in this work are all very insensitive to the adopted mass ratio , in part because the projected surfaces of the stars suffer the least distortions from sphericity close to the times of transit and occultation .
the correct terminology for transits ( primary eclipses ) and occultations ( secondary eclipses ) was discussed in sect.4 of paperiv .
the main parameters for jktebop are the orbital inclination ( @xmath10 ) , and fractional radii of the star ( @xmath11 ) and planet ( @xmath12 ) , defined as : @xmath13 where @xmath14 is the orbital semimajor axis , and @xmath15 and @xmath16 are the volume - equivalent stellar and planetary radii . in this work
i fitted for the sum and ratio of the fractional radii : @xmath17 as these are less strongly correlated . in all cases
the time of transit midpoint ( @xmath18 ) was included as a fitted parameter . when modelling one transit , the orbital period ( @xmath19 ) was fixed to a value taken from the literature . when modelling multiple transits at the same time ,
@xmath20 was included as a fitted parameter .
limb darkening ( ld ) was considered in detail .
every light curve was solved with each of five different ld laws ( see paperi ) and with three different approaches to the ld coefficients : ( 1 ) both fixed ( ` ld - fixed ' ) ; ( 2 ) the linear coefficients ( @xmath21 ) fitted and the nonlinear coefficients ( @xmath22 ) fixed but perturbed0.05 in paperiv , rather than the true value of @xmath230.10 .
i thank the referee of a previous paper @xcite for bringing this to my attention .
] by @xmath230.10 in the error analysis simulations ( ` ld - fit / fix ' ) ; ( 3 ) both coefficients fitted ( ` ld - fitted ' ) . in a few cases where the photometric data are of limited quality
the ld - fitted alternatives were not calculated .
initial or fixed values for the ld coefficients were obtained by bilinear interpolation to the correct @xmath24 and @xmath25 within tabulated theoretical predictions .
once the best of the three options ( ld - fixed , ld - fit / fix , ld - fitted ) was determined , the final light curve parameters were taken to be the weighted mean of the relevant results for the four nonlinear laws .
errorbars for the parameters were obtained from 1000 monte carlo ( mc ) simulations @xcite , for each of the solutions with the adopted approach to ld .
the largest of these was retained , and an additional contribution added to reflect any variation in the parameter values from the four nonlinear ld laws .
finally , alternative errorbars were calculated using a residual permutation ( rp ) algorithm ( as implemented in paperi ) , which accounts for correlated noise .
the rp errorbars were adopted if they were greater than the mc equivalents .
any contaminating ` third ' light ( e.g. * ? ? ?
* ) , denoted by @xmath26 , was included as a constraint following the procedure of @xcite .
a detailed investigation of the issue of third light was given in paperiii .
orbital eccentricity was treated similarly @xcite , with constraints on the combinations @xmath27 and @xmath28 preferred over those for eccentricity ( @xmath29 ) and periastron longitude ( @xmath30 ) directly ( see paperiii ) .
numerical integration over long exposure times is important for several teps , most notably those observed in long cadence by the _ kepler _
satellite , and was dealt with in jktebop following the scheme given in paperiv .
the analysis of transit light curves gives the orbital ephemeris ( @xmath20 , @xmath18 ) and the photometric parameters ( @xmath11 , @xmath12 and @xmath10 ) . for each system
we also have the orbital velocity amplitude of the host star ( @xmath31 ) .
these measured parameters alone do not lead to a unique solution for the physical properties of the system , so an additional constraint has to be sought from elsewhere .
the possible constraints are outlined below in each case , the physical properties calculated were the mass , radius , surface gravity and density of the star ( @xmath32 , @xmath15 , @xmath33 , @xmath34 ) and of the planet ( @xmath35 , @xmath16 , @xmath36 , @xmath37 ) , a surrogate for the planetary equilibrium temperature : @xmath38 and the @xcite number : @xmath39 the set of physical constants used in the process above were tabulated in sect.3.3 of paperiv@xmath5 and @xmath40@xmath2 , respectively .
this does not affect any of the calculations in paperiv or other analyses by the current author .
] , and are very close to the values proposed to the astronomical community by @xcite . the standard approach in obtaining the additional constraint is to use the predictions of theoretical stellar evolutionary models .
these provide the mass and radius of the host star , guided by determinations of its effective temperature ( @xmath24 ) , metal abundance ( @xmath41 $ ] ) and light - curve - derived density @xcite .
the full physical properties of the system are then straightforward to calculate .
i adopted the procedure introduced in paperii , which begins with estimating the velocity amplitude of the _ planet _
( @xmath42 ) .
the other measured quantities ( @xmath20 , @xmath31 , @xmath11 , @xmath12 , @xmath10 , @xmath29 ) were used to determine the ensuing physical properties using standard formulae ( e.g. * ? ? ? * ) .
the value of @xmath42 could then be iteratively refined to minimise the figure of merit @xmath43 ^ 2 + \left[\frac{{\ensuremath{t_{\rm eff}}}^{\rm ( obs)}-{\ensuremath{t_{\rm eff}}}^{\rm ( calc)}}{\sigma({\ensuremath{t_{\rm eff}}}^{\rm ( obs)})}\right]^2\ ] ] where @xmath44 and @xmath45 are quantities determined from observations . @xmath46 and
@xmath47 were obtained for a given @xmath42 value by interpolating in the tabulations of theoretical models for the calculated mass and observed @xmath41 $ ] of the star .
the outcome of this procedure is a set of physical properties corresponding to the best agreement between observed and calculated quantities .
the solution control parameter , @xmath42 , represents the entirety of the input from theoretical stellar models .
because the evolution of a star is highly non - linear with time , the above procedure was carried out for a range of ages starting from 0.01gyr and incrementing in 0.01gyr chunks until the star was significantly evolved ( surface gravity @xmath48 ) .
this led to a set of overall best physical properties plus a model - dependent estimate of the age of the system .
the statistical uncertainties on the input parameters were propagated using a perturbation analysis @xcite implemented in the jktabsdim code .
this had the additional benefit of yielding a complete error budget for every output parameter , allowing identification of how best to improve our understanding of each tep .
the use of theoretical stellar models also incurs a systematic error due to our incomplete understanding of stellar physics .
i therefore ran jktabsdim solutions for each of five sets of theoretical models ( see paperiii ) : _ claret _
@xcite , _ teramo _
@xcite , _ vrss _ @xcite and _ dsep _ @xcite .
this yielded five sets of values and statistical errorbars for each output parameter . for the final result for each parameter
i have taken the unweighted mean of the five values .
the largest of the individual statistical errors was adopted as the final statistical error , and the standard deviation of the five values was taken to represent the systematic error .
three of the derived physical properties have no dependence on stellar theory so are quoted without systematic errorbars : @xmath36 @xcite ; @xmath34 to a good approximation @xcite ; and @xmath50 ( paperiii ) .
[ cols="<,<,^,^ , > , < , > , < , > , < , < " , ] most of the teps in the current work would benefit from further observations of some sort , and this is summarised in table[tab : absdim : obs ] . in many cases
the dominant uncertainty stems from the quality of the light curve .
this remains true for many of the corot systems , even though they have space - based data . in each case
this is due to their relative faintness ( remember that the aperture of the corot telescope is only 27 cm ) and/or that few transits were seen due to either a short observing sequence or a long orbital period .
additional rv measurements are useful too . in many circumstances , particularly for the fainter objects , the rvs are good enough to unambiguously confirm the planetary nature of a system but are the dominant source of uncertainty in the planetary masses . now over 200 teps are known it seems appropriate to concentrate follow - up resources on measuring the physical properties of a golden subset of these to high precision .
an additional requirement of rvs is definition of the orbital shape ( @xmath29 and @xmath30 ) , and imprecise measurements of these quantities compromise measurements of the photometric parameters ( in particular @xmath11 )
. the physical properties of quite a few of the teps are also limited by the precision of the @xmath24 and @xmath41 $ ] measurements available . in many cases this can be improved , but in some cases this is not an option because the errorbars are already close to the limit set by our understanding of low - mass stars ( taken to be 50k in @xmath24 and 0.05dex in @xmath41 $ ] ) .
there is no immediate prospect of lowering these thresholds ; in fact there is evidence that they are already slightly optimistic @xcite .
quite a few of the teps have ephemerides which will become uncertain over the timescale of a few years .
to investigate this i compiled a catalogue of ephemerides of all known teps and identified the first predicted times of transit which were uncertain by one hour , and by half of one transit duration .
a list of objects for which one of these dates is earlier than the year 2022 is given in table[tab : absdim : eph ] .
the _ kepler _ planets are not included in this analysis because they continue to be observed and will benefit from significantly improved ephemerides once these data become available .
the list of the other planets is dominated by corot objects , and it is notable that the ephemerides for corot-4 , corot-14 and corot-17 are already uncertain by more than one hour .
hat - p-31 is present in the list due to the fact that there is no published follow - up photometry of this object at all , only the discovery data from the hat survey cameras .
further photometric observations of these objects , such as those obtained by the terms project @xcite , are needed before the predictive power of their ephemerides deteriorates much further .
whilst scientifically less relevant than the analyses described above , the discovery rate of the transiting planets is of sufficient interest to be worthy of discussion .
fig.[fig : discrate ] shows the discovery rate of the known teps . for discovery time
i have taken the date at which important details of that system became available in a refereed journal article .
if the article was lodged on the arxiv preprint server prior to this , i have instead taken the date at which this preprint became public .
the population of known teps continues to increase at an exponential rate ; a plot of fig.[fig : discrate ] but with a logarithmic ordinate axis ( not shown ) exhibits an approximately straight line .
the early front - runners were the ogle survey ( including the follow - up observations which were performed by a range of associated and independent consortia ) , but the superwasp consortium @xcite held the lead during the years 2009 - 11 .
the greatest number of known teps is now attributable to data obtained by the _
kepler _ satellite , and
once again this number includes both associated and independent groups of researchers .
_ kepler _ has now been used to find over 2000 candidate teps @xcite and possible as many as 5000 @xcite ; the rate of false positives is probably small but remains under discussion @xcite .
the candidate list provides the ammunition to continue the exponential rise in the number of known teps .
fig.[fig : skypos ] shows the sky positions of the known teps , overlaid with the celestial equator and galactic plane , and is a good tool to illustrate the strength and diversity of the selection effects which afflict this population .
the most obvious feature in fig.[fig : skypos ] is a strong clustering around 19hr right ascension and @xmath51 declination , which is the _ kepler _ field .
the _ kepler _ satellite @xcite is currently the most successful tep discovery machine , and has a modest field of view ( 105deg@xmath49 ) .
this concentration of points is the result .
additional clusters of teps occur around the intersections of the galactic plane with the celestial equator , the two areas targeted by the corot satellite @xcite .
corot aims for these positions in order to obtain a high surface density of stars ( to give plenty of targets in its small 8.2deg@xmath49 field of view ) which are easily followed up using ground - based telescopes in both hemispheres .
two other small groups of teps come via the ogle survey @xcite observations of the galactic centre ( 18hr r.a . and @xmath52 dec ) and carina regions ( 11hr r.a . and @xmath53 dec ) . the wider spread of objects in fig.[fig : skypos ] is dominated by hat @xcite in the north and superwasp @xcite in the south .
these , plus other similar projects , operate multiple small wide - field telescopes with very coarse pixel scales .
they are thus able to cover large amounts of sky but must avoid the crowded fields towards the galactic plane , due to problems with blending .
blending makes it more difficult to measure good photometry and also elevates the rate of false alarms in a search for teps .
the surveys operating small ground - based telescopes tend to point away from the galactic plane , which introduces a bias towards intrinsically faint stars .
this is advantageous because it reduces the fraction of giants in the studied sample ; these surveys can not find planets around giant stars but are able to detect larger transiting object which are of lesser scientific interest .
conversely , the _ kepler _ satellite studies stars with fainter apparent magnitudes which are located close to the galactic plane .
these stars have slightly different population characteristics to those studied by other surveys with different designs ; @xcite show that the host stars of the _ kepler _ planet candidates are preferentially metal - rich .
such biases must be accounted for when studying the population characteristics of the known teps and their host stars .
the _ homogeneous studies _ project aims to provide measurements of the physical properties of a large sample of transiting extrasolar planetary systems , using consistent methods and with careful attention paid to the estimation of robust statistical and systematic errors .
the transit light curves are modelled using the jktebop code , and uncertainties are gauged with monte carlo and residual - permutation algorithms .
attention is paid to the treatment of limb darkening , orbital eccentricity , and contaminating ` third ' light . to the resulting photometric parameters
are added measured spectroscopic quantities : the @xmath24 , @xmath41 $ ] and velocity amplitude of the host star .
one additional constraint is needed , and is supplied either by one of five sets of theoretical stellar evolutionary models or by a semi - empirical calibration of low - mass star properties based on detached eclipsing binaries .
the statistical errors are propagated using a perturbation algorithm , leading to complete error budgets for each output quantity .
the inclusion of multiple stellar models allows systematic errors to be deduced too .
the current paper has presented complete analyses of thirty tep systems , and updated the physical properties of eight more systems based on newly published spectroscopic results .
combined with previous work , this yields a total of 82 transiting planets and host stars with homogeneously - measured physical properties . in many cases these results
are based on more numerous datasets than previous measurements , so should be preferred over literature values even if homogeneity is not of specific importance to the matter in hand .
headline results from the current paper are summarised below .
analyses in the current work were the first to consider all available good light curves for hat - p-3 ( five datasets ) , hat - p-6 ( three ) , hat - p-9 ( four ) , hat - p-14 ( four ) , wasp-12 ( four ) and wasp-14 ( four ) . for wasp-14
this has resulted in a change in planetary radius ( @xmath16 ) from @xmath54@xmath2 to @xmath55@xmath2 , moving this object away from the parameter space occupied by theoretical predictions .
the current study is also the first to present results based on short - cadence data for kepler-14 , kepler-15 and koi-135 .
the measured mass of kepler-14a has decreased by @xmath56 , and the measured radii of both components in the kepler-15 system are both significantly larger .
the density deduced for kepler-15b is lower by a factor of two compared to previous studies .
additional data was modelled for koi-196 , koi-204 , koi-254 , koi-423 and koi-428 . in two cases ( koi-423 and koi-428 ) the available data is more than a factor three more extensive than used in published studies . for koi-428 the radii of the components are larger than previously found ;
the star is now comfortably the largest known known to host a tep at @xmath57@xmath5 .
the physical properties of ogle - tr-56 have been heavily revised based on the first complete analysis of newly - published high - quality photometry @xcite .
the radius of the planet is @xmath58@xmath2 , which is much larger than all previous determinations ( 1.231.38@xmath2 ) .
this radius makes ogle - tr-56b one of the largest known planets , as expected for its high equilibrium temperature and hot and massive host star .
its measured density is smaller by a factor of three .
previous analyses were based either on survey - quality photometry , a single transit light curve which is now known to be strongly affected by correlated noise , or on unnecessary assumptions . two morals can be deduced from this .
firstly , one should always use a newly - derived stellar density in calculating revised physical properties of a tep system . secondly
, results based on only one transit light curve can not be definitive as systematic errors may lurk undetected in the data .
new results have also been calculated for ogle - tr-111 , ogle - tr-113 , ogle - tr-132 and ogle - tr - l9 .
these comprise the first complete analyses of recently published high - quality photometry for each system .
for wasp-13 and wasp-21 my analysis has resulted in the upward revision of the planetary radii , making them two of the least dense planets known . for wasp-13 @xmath16 changes from @xmath59@xmath2 to @xmath60@xmath2 , and for wasp-21
the change is from @xmath61@xmath2 to @xmath62@xmath2 .
finally , the inclusion of an improved @xmath24 measurement makes wasp-2 one of the best - understood tep systems .
previously found correlations between @xmath20 and planet gravity , and @xmath20 and planet mass , have been revisited and found to be of improved statistical significance ( @xmath63 and @xmath64 respectively ) .
the division of teps into two classes based on their safronov number is shown to be spurious .
i recommend further observations of specific types for the majority of the objects studied here .
these are summarised in sect.[sec : followup ] .
in particular , some orbital ephemerides are of low precision and will soon be of limited use .
the affected systems would benefit from new transit photometry to solve this problem . finally , the discovery rate and distribution of sky positions of the known teps are plotted and discussed .
the main results from this work will be made available in a convenient format in the _ transiting extrasolar planets catalogue _
( tepcat ) .
i acknowledge financial support from stfc in the form of an advanced fellowship .
i am grateful to barry smalley for many conversations and to conrad vilela for discussions about the _ kepler _ archive .
i am grateful for data provided by joel hartmann ( hat - p-6 and hat - p-9 ) , neale gibson ( hat - p-3 ) , guillaume hbrard ( corot-18 ) , jason dittmann ( hat - p-9 ) , pedro sada ( hat - p-9 ) , kamen todorov ( hat - p-9 ) and monika lendl ( ogle - tr - l9 ) .
i thank the cds , mast , ias and nsted websites for archiving the many datasets now available for transiting planets .
the following internet - based resources were used in research for this paper : the eso digitized sky survey ; the nasa astrophysics data system ; the simbad database operated at cds , strasbourg , france ; and the ar@xmath65iv scientific paper preprint service operated by cornell university . | i measure the physical properties of 38 transiting extrasolar planetary systems , bringing the total number studied within the _ homogeneous studies _ project to 82 .
transit light curves are modelled using the jktebop code , with careful attention paid to limb darkening , orbital eccentricity and contaminating light .
the physical properties of each system are obtained from the photometric parameters , published spectroscopic measurements and five sets of theoretical stellar model predictions .
statistical errors are assessed using monte carlo and residual - permutation algorithms and propagated via a perturbation algorithm .
systematic errors are estimated from the interagreement between results calculated using five theoretical stellar models .
the headline result is a major upward revision of the radius of the planet in the ogle - tr-56 system , from @xmath0@xmath1@xmath2 to @xmath3@xmath2 ( statistical and systematic errors , respectively ) .
its density is three times lower than previously thought .
this change comes from the first complete analysis of published high - quality photometry .
significantly larger planetary radii are also found for kepler-15 , koi-428 , wasp-13 , wasp-14 and wasp-21 compared to previous work .
i present the first results based on _ kepler _ short - cadence data for kepler-14 , kepler-15 and koi-135 .
more extensive long - cadence data from the _ kepler _ satellite is used to improve the measured properties of koi-196 , koi-204 , koi-254 , koi-423 and koi-428 .
the stellar component in the koi-428 system is the largest known to host a transiting planet , at @xmath4@xmath5 .
detailed analyses are given for hat - p-3 , hat - p-6 , hat - p-9 , hat - p-14 and wasp-12 , based on more extensive datasets than considered in previous studies .
detailed analyses are also presented for the corot systems 17 , 18 , 19 , 20 and 23 ; kepler- 7 , 12 and 17 ; koi-254 ; ogle - tr- 111 , 113 , 132 and l9 ; and tres-4 .
i revisit the correlations between orbital period and surface gravity , and orbital period and mass of the transiting planets , finding both to be significant at the @xmath6 level .
i conclude by discussing the opportunities for follow - up observations , the sky positions and the discovery rate of the known transiting planets .
stars : planetary systems stars : fundamental parameters |
metabolomics is the comprehensive study of all small organic compounds in a biological specimen @xcite .
metabolite concentrations in body fluids such as urine , serum , and plasma have proven valuable in predicting disease onset and progression @xcite .
metabolomic data can be generated by a variety of methods of which mass spectrometry and @xmath0h nmr spectroscopy are the most common . while its high sensitivity makes mass spectrometry the preferred method for discovery projects , the high reproducibility of nmr data @xcite is ideal for applications in precision medicine .
@xmath0h nmr allows for the simultaneous detection of all proton - containing metabolites present at sufficient concentrations in biological specimens .
furthermore , nmr signal volume scales linearly with concentration .
the complete set of nmr signals acquired from a given specimen is called its metabolite `` fingerprint '' . due to differences in ph ,
salt concentration , and/or temperature , relative displacement of signal positions occurs across spectra .
binning schemes can compensate for this effect by splitting spectra into segments called bins and summing up signal volumes contained therein .
equal - sized bins are commonly used , albeit other schemes such as adaptive binning have been suggested @xcite . binned fingerprint data is the typical starting point for subsequent multivariate data analysis including biomarker discovery and specimen classification using methods such as support vector machines @xcite , ridge @xcite , and lasso regression @xcite .
metabolite fingerprints need to be scaled to a common unit .
typical examples include mmol metabolite per ml plasma , mmol metabolite per mmol creatinine in urine , or the relative contribution of a bucket to the total spectral intensity of an nmr spectrum . in practice
this is achieved by dividing each spectrum by the unit defining quantity : the intensity of an nmr reference such as tsp , the intensity of creatinine , or the total spectral intensity .
this scaling of the raw data defines the measurement , but it also serves a second purpose : scaling corrects for unwanted experimental and physiological variability in the raw spectra . correcting unwanted variability is usually called normalization .
instrument performance can vary and so can patient - specific parameters like the fluid balance , which is affected by drinking , respiration , defecation , perspiration , or medication , thus altering urine metabolite concentrations without reflecting disease state @xcite .
all scales have their pros and cons : the nmr reference can normalize for spectrometer performance @xcite but not for unwanted variability in urine density .
choosing creatinine as a standard for urine assumes the absence of inter - individual differences in the production and renal excretion of creatinine @xcite .
in fact , creatinine production and excretion is affected by sex , age , muscle mass , diet , pregnancy , and , most importantly , renal pathology @xcite .
normalization to a constant total spectral intensity or spectral mapping to a reference spectrum assume that the total amount of metabolites is constant over time and across patients and that spectra are not contaminated by signals that do not represent metabolites .
however , this is not always the case @xcite .
in fact , for urinary specimens of patients suffering from proteinuria the additional protein signals greatly increase total spectral area , and scaling to a constant total intensity systematically underestimates metabolite abundances
. similarly , excessive glucose uptake , for example by a glucose infusion , leads to high total spectral intensities that are dominated by glucose and its metabolites .
while the high values for these metabolites correctly reflect the metabolic state for these patients , their influence on the total spectral intensity leads to systematic underestimation of metabolites not related to glucose metabolism . in general , the preferred scaling protocol depends on the specific data set to be investigated .
however , for some data sets it is not possible to use the same scale for all patients in the cohort .
we will describe such data sets below . in this case
different protocols need to be used for different patients , introducing new challenges to data analysis . in this contribution
, we first studied how the choice of scale affects statistical analysis , the selection of biomarkers , and patients diagnosis by these biomarkers .
we report on two supervised metabolomics data analysis scenarios , namely urine and plasma biomarker discovery in 1d @xmath0h nmr metabolite fingerprints for the early detection of acute kidney injury onset after cardiac surgery @xcite . in both applications we tested metabolites for differential abundance using alternative scaling protocols .
we observed pronounced disagreements between the lists of significantly differential metabolites depending on how the same data was scaled .
more importantly , the different scalings led to inconsistencies in the classification of individual patients . in view of these observations ,
reproducibility of metabolic studies is only possible if the exact same scaling protocols are used . to overcome this problem
, we extended zero - sum regression @xcite , which has recently been demonstrated to be invariant under any rescaling of data @xcite , to logistic zero - sum regression and compared it to two standard methods for constructing multivariate signatures : lasso logistic regression and support vector machines in combination with _
t_-score based feature filtering .
unlike the latter methods , logistic zero - sum regression always identifies the same biomarkers regardless of the scaling method .
consequently , prior data normalization may be omitted completely .
we make logistic zero - sum regression available as an _ r _ package and as a high - performance computing software that can be downloaded at https://github.com/rehbergt/zerosum .
the first data set comprised 1d @xmath0h nmr fingerprints of @xmath1 = 106 urine specimens , which had been collected from patients 24 h after cardiac surgery with cardiopulmonary bypass ( cpb ) use at the university clinic of erlangen @xcite . of these 106 ,
34 were diagnosed with acute kidney injury ( aki ) 48 h after surgery .
the challenge in this data is to define a urinary biomarker signature that allows for the early detection of aki onset ( acute kidney injury network ( akin ) stages 1 to 3 ) . the second data set consisted of 85 edta - plasma specimens , which had been collected 24 h post - op from a subcohort of the original cohort of 106 patients undergoing cardiac surgery with cpb use and which had been subjected to 10 kda cutoff filtration @xcite . in total , 33 patients out of these 85 patients were diagnosed with postoperative aki @xcite .
again our goal is to detect biomarkers for an earlier detection of aki .
a total of 400 @xmath2l of urine or edta - plasma ultrafiltrate was mixed with 200 @xmath2l of phosphate buffer , ph 7.4 , and 50 @xmath2l of 0.75% ( w ) 3-trimethylsilyl-2,2,3,3-tetradeuteropropionate ( tsp ) dissolved in deuterium oxide as the internal standard ( sigma - aldrich , taufkirchen , germany ) .
nmr experiments were carried out on a 600 mhz bruker avance iii ( bruker biospin gmbh , rheinstetten , germany ) employing a triple resonance ( @xmath0h , @xmath3c , @xmath4p , @xmath5h lock ) cryogenic probe equipped with @xmath6-gradients and an automatic cooled sample changer .
for each sample , a 1d @xmath0h nmr spectrum was acquired employing a 1d nuclear overhauser enhancement spectroscopy ( noesy ) pulse sequence with solvent signal suppression by presaturation during relaxation and mixing time following established protocols @xcite .
nmr signals were identified by comparison with reference spectra of pure compounds acquired under equal experimental conditions @xcite .
the spectral region from 9.5 to @xmath70.5 ppm of the 1d spectra was exported as even bins of 0.001 ppm width employing amix 3.9.13 ( bruker biospin ) .
the data matrix was imported into the statistical analysis software _
r _ version 3.3.2 . for the urinary spectra , the region 6.5 4.5 ppm , which contains the broad urea and water signals ,
was excluded prior to further analysis . for plasma spectra , the region 6.2 4.6 ppm , containing the urea and remaining water signals ,
was removed prior to analysis .
in addition , the regions 3.82 3.76 ppm , 3.68 3.52 ppm , 3.23 3.2 ppm , and 0.75 0.72 ppm , corresponding to filter residues and free edta , were excluded prior to classification for the plasma specimens .
we compare our scaling- and normalization - invariant approach to standard analysis strategies that were applied to data preprocessed by four state - of - the - art normalization protocols .
for creatinine normalization of urinary data we divided each bucket intensity by the summed intensities of the creatinine reference region ranging from 3.055 to 3.013 ppm of the corresponding nmr spectrum . when scaling to the total spectral intensity we summed over all signal intensities from 9.5 to 0.5 ppm after exclusion of the water and urea signals and divided each bucket intensity by this sum .
when scaling to the signal intensity of the nmr reference compound , in our case tsp , we summed up the intensities of the tsp buckets ranging from @xmath70.025 to 0.025 ppm and divided each bucket intensity by this sum .
this normalization method corrects for differences in spectrometer performance @xcite , but not for differences in fluid intake .
finally , we also evaluated probabilistic quotient normalization ( pqn ) @xcite .
pqn follows the rationale that changes in concentration of one or a few metabolites affect only small segments of the spectra , whereas specimen dilution , for example due to differences in fluid intake in case of urinary specimens , influences all spectral signals simultaneously @xcite .
following @xcite we first normalized each spectrum to its total spectral intensity .
then , we took the median across all these spectra as the reference spectrum and calculated the ratio of all bucket intensities between raw and reference spectra .
the median of these ratios in a sample was our final division factor . for subsequent analysis , only the spectral region between 9.5 and 0.5
ppm was taken into account . to compensate for slight shifts in signal positions across spectra due to small variations in sample ph , salt concentration , and/or temperature , bucket intensities across ten buckets
were fused together in one bucket of 0.01 ppm width by summing up the individual bucket intensities .
the resulting intensities were @xmath8 transformed to correct for heteroscedasticity , and all subsequent data analysis was performed with these values .
we compared logistic zero - sum regression to two classification algorithms for normalized data , namely a support vector machine @xcite employing a linear kernel function combined with univariate feature filtering ( f - svm ) and the least absolute shrinkage and selection operator ( lasso ) logistic regression @xcite . in the lasso logistic regression and the zero - sum models , the penalizing parameter @xmath9 that balances
the bias variance trade - off was optimized in an internal cross - validation .
lasso models were trained utilizing the _ r _ package _ glmnet _ @xcite .
the zero - sum models were trained by employing our _ zerosum _ software , which is implemented in two versions : first , a version optimized for high - performance computing ( hpc ) , and second , an _ r _
package for ordinary desktop computers .
we utilized the _ zerosum _ hpc implementation throughout the article .
the parameter @xmath9 induces sparseness of the models and circumvents over - fitting for lasso and zero - sum models . similarly , we optimized the parameters of f - svm .
this was done in a two - fold nested cross - validation , i.e. , in each cross - validation step we screened for the best feature threshold , allowing for @xmath10 included features , and the matching best cost parameter , which was screened in @xmath11 . for training the svm we utilized the _
r _ package _
metabolic biomarkers can be metabolites or just spectral features .
moreover , they can be identified as stand - alone predictors or as part of a multivariate biomarker signature . before we focus on signatures we study the effect of scales on stand - alone biomarkers .
more precisely , we focus on spectral buckets with differential intensities between two classes of samples , i.e. , from patients who developed acute kidney injury after cardiac surgery versus from those that did not .
bin intensities were normalized to ( a ) a constant total spectral intensity , ( b ) a constant creatinine intensity , ( c ) a median reference spectrum employing the most probable quotient and ( d ) a constant intensity of the nmr reference .
for scoring normalized bin intensities as potential biomarkers we used benjamini - hochberg ( b / h ) corrected @xmath12-values from the moderated _ t_-statistics implemented in the _ r _ package limma @xcite and @xmath8 fold changes between the classes .
figure [ t - test ] shows plots of spectral bin positions against @xmath13(@xmath12-values ) ( a - d ) and @xmath8 fold changes ( e - h ) between aki and non - aki urine specimens . from left to right the plots correspond to ( a , e ) normalization by total spectral intensity , ( b , f ) normalization to creatinine , ( c , g ) pqn , and ( d , h ) normalization to tsp .
the red line in plots ( a - d ) marks a significance level of 0.01 , corresponding to a false discovery rate ( fdr ) below 1@xmath14 .
significant buckets are plotted in red in all eight plots . *
normalization to total spectral intensity : * on this scale we observed almost exclusively negative fold changes indicating lower bucket intensities for aki in comparison to non - aki .
242 features were significant with features corresponding to carnitine , 2-oxoglutaric acid , and glutamine ranking highest .
supplementary table s1 ( a ) lists raw and b / h - adjusted @xmath12-values as well as metabolite assignments of the top ten significant buckets .
technical artifacts of the scale become apparent too : the spectral region from 4 to 3.5 ppm , highlighted as a blue band , hardly comprised any significant features .
this region is dominated by signals from sugars such as d - mannitol , which had been used as a pre - filling material for the tubes of the cpb machine @xcite . as d - mannitol exhibits a rather large number of nmr signals ,
as highlighted in an exemplary urine spectrum shown in figure [ aki_urin ] , it comprised between @xmath15 and @xmath16 of the total spectral areas . on a scale that makes the total spectral area the same across all samples
these signals can not serve as biomarkers . however , all patients take up d - mannitol during surgery , and the amount of d - mannitol still found in urine 24 h past surgery is modulated by actual kidney function .
d - mannitol entangles kidney function with the total spectral area .
therefore , the spectral area can not be recommended for scaling biomarkers , at least not in this context .
* normalization to creatinine : * for this scale we observed different analysis results .
@xmath8 fold changes were predominantly positive indicating higher metabolite levels in aki patients .
204 buckets were significant , and the top ten are listed in supplementary table s1 ( b ) .
the biomarker ranked highest was tranexamic acid , which is given in cardiac surgery to prevent excessive blood loss .
strikingly , the region between 4 and 3.5 ppm ( blue band ) containing the d - mannitol signals was highly significant on this scale
. however , this scale confounds with aspects of kidney function . in response to kidney disease
urinary creatinine excretion rates can be affected @xcite .
thus , normalization to creatinine can also obscure other metabolite biomarkers if their excretion correlates with that of creatinine .
* probabilistic quotient normalization : * on this scale we observed both positive and negative fold changes in almost equal numbers .
only 89 features were now significant , and the top ten are listed in table s1 ( c ) , with carnitine and glutamine as the leading biomarkers .
again , the blue shaded region now covers significant features , although the number is much lower than after creatinine scaling . due to the strong conceptual similarity between pqn and normalization to total spectral intensity
, the d - mannitol artifact also compromises the use of pqn .
* tsp normalization : * on this scale we did not detect any significant biomarkers .
the scaling method corrects only for differences in spectrometer performance , not for changes in global metabolite concentration . due to large variability in urine density throughout the cohort
, it can not be used in this context .
we observed only one bucket at 3.715 ppm , identified as an overlap of propofol - glucuronide , broad protein signals , and tentatively d - glucuronic acid , which obtained a significant @xmath12-value on three scales ( figure [ venn_aki_classif]a ) .
the plasma data biomarker discovery was not consistent across scales either ( supplementary table s2 , figure s1 , and figure [ venn_aki_classif]b ) . scaling to total spectral area and scaling to tsp predominantly identified metabolites that accumulate in the blood of patients developing aki .
one might explain this observation by a reduced glomerular filtration in these patients .
however , the pqn data immediately challenged this interpretation , as it identified a large set of down - regulated metabolites in aki .
creatinine normalization is not common for plasma metabolomics and was not investigated here .
exemplary 1d @xmath0h nmr spectrum of a urine specimen collected 24 h after surgery from an aki patient .
the lower histogram illustrates the common binning procedure.,scaledwidth=50.0% ] number of significant buckets and their overlap between the different scaling methods for ( a ) the aki urine and ( b ) the aki plasma data set.,title="fig:",scaledwidth=24.0% ] number of significant buckets and their overlap between the different scaling methods for ( a ) the aki urine and ( b ) the aki plasma data set.,title="fig:",scaledwidth=24.0% ] biomarkers can be combined to biomarker signatures in multivariate analysis .
machine - learning algorithms are used to learn these signatures from training data .
we tested two of these algorithms , namely a linear svm with @xmath17-score based feature filtering ( f - svm ) and standard logistic lasso regression .
both methods implement linear signatures of the form @xmath18 where @xmath19 is the raw signal from bucket @xmath20 in patient @xmath21 , @xmath22 the weight of this feature , and @xmath23 the scaling factor used to normalize sample @xmath21 .
both methods were used to learn signatures from data normalized in four different ways , namely scaling by total spectral area , scaling to creatinine , pqn , and scaling to tsp .
the performance of the algorithms was tested in cross - validation .
figure [ roc_aki]a shows roc curves of signatures that aim to predict aki from urinary fingerprints using f - svm signatures .
the areas under the roc curves ( auc - roc ) are summarized in table [ roc_table ] .
the higher this area the better the prediction of patient outcome .
performances ranged from an auc of 0.72 for tsp - scaled data to an auc of 0.83 for pqn data . in line with the results for stand - alone biomarkers , the f - svm algorithm picked different features across scaling methods .
not a single bucket was chosen simultaneously for all four scales ( figure [ roc_aki]a bottom row ) .
figure [ roc_aki]b and table [ roc_table ] show the corresponding results for standard lasso logistic regression .
the lasso signatures performed slightly better and were more consistent .
nevertheless , there is still a dependence of the chosen biomarkers on the scale .
similar results were observed for the plasma data set ( supplementary figure s2 ) . the urinary aki data set : receiver operating characteristic ( roc ) curves for two classification approaches , ( a ) svm in combination with @xmath17-test based feature filtering and ( b ) lasso , after application of four different normalization strategies : scaling to total spectral area ( red solid line ) , scaling to creatinine ( blue dashed line ) , probabilistic quotient normalization ( pqn ) ( green dotted line ) , and scaling to tsp ( yellow dashed - dotted line ) .
the bottom row shows the number of features included in the respective classification models in venn diagrams .
the corresponding models were built by averaging over all models of the outer cv loop . ] from a clinical perspective , varying biomarkers constitute only a minor problem , as long as they agree in their predictions of outcome
. however , they usually do not agree .
for the urinary data set , lasso signatures yielded conflicting predictions in 16% of patients . for f - svm signatures ,
the percentage increased to 30% .
figure [ classification_summary_aki ] summarizes the predictions of aki onset after cardiac surgery patient by patient for the urinary data set .
the row `` patient outcome '' shows patients in blue who did not develop aki ( akin stage 0 ) and in red those that developed aki .
furthermore , the yellow dashed lines highlight patients with akin stage 2 and 3 , a more severe manifestation of kidney injury .
the latter patients constitute the high - risk group where early detection of aki onset can save lives @xcite .
the true outcomes 48 h after surgery are contrasted to the predictions 24 h earlier . shown
are predictions of f - svm and lasso signatures for the four scaling methods .
we observed that predictions frequently changed with scale and that signatures learned on data that was scaled to , e.g. , creatinine did not properly identify the highest risk group . for f -
svm only the signature on pqn data identified all high - risk patients correctly .
the lasso identified high - risk patients for total spectral area and pqn data correctly .
we discuss some misclassifications in more detail .
patients aki-35 and aki-106 developed severe aki after 48 h but were not predicted to do so by two of the f - svm signatures , namely those for creatinine and total spectral area scaled data .
a key observation of this paper is that these misclassifications could be traced back to data normalization . in theory , the predictions could have been saved by readjusting the scaling of these fingerprints ( however , the necessary scaling factor is not evident _ a priori _ ) .
figure [ prob_vs_gamma ] shows the prediction probabilities of these signatures on the @xmath24-axis .
if this probability was above 0.5 we predicted an onset of aki , otherwise we did not .
the @xmath25-axis shows possible multiplicative scale adjustments @xmath26 .
a value of @xmath27 indicates the actual scale used for prediction .
values of @xmath28 correspond to a down - scaling of all buckets by the factor @xmath26 , while values of @xmath29 correspond to an up - scaling by @xmath26 .
the plot thus shows the probability of developing aki as a function of the scale - readjustment factor @xmath26 .
the dashed lines correspond to f - svm and the solid lines to lasso signatures .
colors indicate the underlying scaling methods of the signatures : total spectral area ( red ) , creatinine ( blue ) , pqn ( green ) , and tsp ( yellow ) . on the original scale ( @xmath27 )
patient aki-35 shows prediction probabilities below 0.5 for f - svm - creatinine ( blue dashed line ) and f - svm - total - area ( red dashed line ) .
let @xmath30 be the ratio of the creatinine signal ( from 3.055 to 3.013 ppm ) and the remaining spectral area after exclusion of the d - mannitol region ( 4.0 to 3.5 ppm ) . for aki-35 we observed @xmath31 , while the median over all aki predicted samples was @xmath32 . in line with this
, up - scaling the fingerprint rescues the prediction for this patient ( blue dashed line in figure [ prob_vs_gamma]a ) .
similarly , @xmath33 , the ratio between d - mannitol and the remaining spectral area , was @xmath34 , while the median over all aki predicted samples was 2.42 , which explains why the f - svm - total - area prediction could be rescued by a down - scaling of the fingerprint ( red dashed line in figure [ prob_vs_gamma]a ) .
similar problems can be observed for patient aki-106 , and for the lasso - creatinine signature in patients aki-70 and aki-4 . for the plasma data , we observed inconsistent predictions for 19% ( 6% ) of the patients for f - svm ( lasso ) ( supplementary figure s3 ) . here , no method identified the high - risk group completely , with best identification shown for lasso on pqn data . in summary , we observed inconsistencies in the selection of biomarkers and more importantly in the prediction of patient outcome across the four scaling methods .
these inconsistencies were most pronounced for the filter - based feature selection in f - svm and affected urinary fingerprints more than plasma fingerprints . still they could be observed for all analysis scenarios . for several patients
a simple scale adjustment would have saved the prediction .
the next section will show that zero - sum logistic regression resolves these inconsistencies fully and leads to more accurate predictions . the probability of aki prediction as a function of the scale readjustment factor @xmath26 ( see text ) .
the dashed lines correspond to f - svm signatures and the solid lines to lasso signatures .
colors indicate the underlying scaling methods of the signatures : total spectral area ( red ) , creatinine ( blue ) , pqn ( green ) , and tsp ( yellow ) .
the black dotted line gives the zero - sum predictions , which are independent of @xmath26 .
, title="fig : " ] the probability of aki prediction as a function of the scale readjustment factor @xmath26 ( see text ) .
the dashed lines correspond to f - svm signatures and the solid lines to lasso signatures .
colors indicate the underlying scaling methods of the signatures : total spectral area ( red ) , creatinine ( blue ) , pqn ( green ) , and tsp ( yellow ) .
the black dotted line gives the zero - sum predictions , which are independent of @xmath26 .
, title="fig : " ] the probability of aki prediction as a function of the scale readjustment factor @xmath26 ( see text ) .
the dashed lines correspond to f - svm signatures and the solid lines to lasso signatures .
colors indicate the underlying scaling methods of the signatures : total spectral area ( red ) , creatinine ( blue ) , pqn ( green ) , and tsp ( yellow ) .
the black dotted line gives the zero - sum predictions , which are independent of @xmath26 .
, title="fig : " ] the probability of aki prediction as a function of the scale readjustment factor @xmath26 ( see text ) .
the dashed lines correspond to f - svm signatures and the solid lines to lasso signatures .
colors indicate the underlying scaling methods of the signatures : total spectral area ( red ) , creatinine ( blue ) , pqn ( green ) , and tsp ( yellow ) .
the black dotted line gives the zero - sum predictions , which are independent of @xmath26 .
, title="fig : " ] zero - sum regression @xcite is a novel machine - learning algorithm that is insensitive to rescaling the data @xcite .
it allows for a selection of biomarkers that does not depend on the units chosen .
the classifications of patients that result from these signatures do not depend on any scaling of the data either .
in fact , patients can be classified with spectral data that were not normalized at all . for the reader s convenience we review the concept here in a nutshell : let @xmath35 be metabolomics data , where @xmath36 is the logarithm of the intensity for bucket @xmath37 in sample @xmath38 and @xmath39 the corresponding clinical response of patient @xmath21 .
in regression analysis the data sets need to be normalized to a common unit .
note that the data are on a logarithmic scale , therefore the scaling to a common unit becomes a shifting of the spectrum @xmath40 by some sample specific value @xmath41 .
thus for normalized data the regression equation reads @xmath42 now note that the equation becomes independent of the normalization @xmath41 if and only if the regression coefficients @xmath22 sum up to zero , i.e. , @xmath43 this is the idea of zero - sum regression . in the machine - learning context of high - content data ,
zero - sum regression can be combined with lasso or elastic - net regularization and shows predictive performances that were not compromised by the zero - sum constraint @xcite . in
many biomarker discovery challenges the response @xmath24 is not continuous but binary .
we thus need to extend the concept of zero - sum regression from linear regression to classification .
we do this by introducing logistic zero - sum regression . in standard logistic regression
the log - likelihood of normalized data reads @xmath44 where we abbreviated @xmath45 . in this generalized linear model
the log - likelihood again becomes independent of the normalization @xmath41 if the regression coefficients @xmath22 add up to zero .
this statement also holds if we add the penalizing term @xmath46 , which corresponds to the elastic - net regularization penalty .
the parameter @xmath9 is calibrated in cross - validation @xcite , while @xmath47 is usually fixed to a specific value in @xmath48 $ ] .
here , @xmath49 corresponds to ridge regression @xcite and @xmath50 to the lasso @xcite .
thus , logistic zero - sum regression amounts to finding coefficients @xmath51 that minimize @xmath52 for all applications throughout the article , we have chosen @xmath50 .
as proposed by @xcite we use the quadratic approximation of eq .
( [ loglilogist ] ) locally at the current parameters @xmath53 of the optimization algorithm : @xmath54 with @xmath55 @xmath56 is independent of @xmath57 and can thus be neglected .
an efficient coordinate - descent algorithm for zero - sum regression can be constructed by incorporating the zero - sum constraint into the log - likelihood by substituting @xmath58 with @xmath59 @xcite . calculating the partial derivative of the resulting log - likelihood with respect to @xmath60 ( @xmath61 ) and solving for @xmath60 yields an update scheme for @xmath60 and @xmath58 : @xmath62 the update value for @xmath58
is given by @xmath59 .
after each update the approximation has to be renewed .
since we update @xmath60 and @xmath58 simultaneously the number of updates scales quadratically with the number of features .
therefore , the minimization problem ( [ loglizsum ] ) will be computationally demanding if @xmath25 becomes high - dimensional .
we performed a nested cross - validation ( cv ) , where the inner cv uses 10 folds to determine a suitable regularization strength @xmath9 and the outer cv is a leave - one - out cv to evaluate the prediction accuracy of the resulting models .
we are evaluating two different data sets , each within a leave - one - out cv , one data set with 4 and the other with 3 different normalizations , and therefore we have to do 107@xmath634 + 86@xmath633=686 independent inner cvs .
each of these cvs is performed on an approximated regularization path of length 500 , however , the algorithm stops when overfitting occurs . in total
we performed @xmath64 model fits .
this number presents an obvious computational challenge . to avoid numerical uncertainties in the feature selection we used the very precise convergence criterion of @xmath65 , which additionally increases the computational effort . on a standard server
( 2 intel xeon x5650 processors with 6 cores each ) , the complete calculation took about two days .
we were able to bring the compute time down to 12 minutes by developing an hpc implementation of _ zerosum _ and executing it on 174 nodes of the supercomputer qpace 3 operated by the computational particle physics group .
qpace 3 currently comprises 352 intel xeon phi 7210 ( `` knights landing '' ) processors , connected by an omnipath network .
each processor contains 64 compute cores , and each of these compute cores contains two 512-bit wide vector units . to run efficiently on this machine ,
the _ zerosum _ c code was extended to use avx512 vector intrinsics for the calculation of the coordinate - descent updates .
openmp is used to parallelize the inner cv so that the data has to be stored in memory only once and can be accessed from all folds . additionally , we provide a new _ zerosum _ _ r _ package , which is a wrapper around the hpc version and can easily be used within _
r _ package also includes functions for exporting and importing all necessary files for / from the hpc implementation . for users without access to hpc facilities
, our package can be run on a regular workstation with reduced convergence precision and a shorter regularization path in less than one hour .
receiver operating characteristic ( roc ) curves for logistic zero - sum regression , after application of four different normalization strategies , scaling to total spectral area ( red solid line ) , scaling to creatinine ( blue dashed line ) , probabilistic quotient normalization ( pqn ) ( green dotted line ) , and scaling to tsp ( yellow dashed - dotted line ) for ( a ) the urinary aki and ( b ) the plasma aki data set .
the right column shows the number of features included in the respective classification models in venn diagrams .
the corresponding models were built by averaging over all models of the outer cv loop .
, title="fig : " ] receiver operating characteristic ( roc ) curves for logistic zero - sum regression , after application of four different normalization strategies , scaling to total spectral area ( red solid line ) , scaling to creatinine ( blue dashed line ) , probabilistic quotient normalization ( pqn ) ( green dotted line ) , and scaling to tsp ( yellow dashed - dotted line ) for ( a ) the urinary aki and ( b ) the plasma aki data set . the right column shows the number of features included in the respective classification models in venn diagrams .
the corresponding models were built by averaging over all models of the outer cv loop .
, title="fig : " ] logistic zero - sum regression is not affected by changes in scale . actually , it will always yield the same result regardless of the scaling method used .
figure [ roc_zerosum ] gives the corresponding roc curves and venn diagrams for the four normalization strategies investigated .
since zero - sum models are scale independent , this classification approach always identified the same biomarkers and assigned the same weights to them independently of the normalization method chosen .
supplementary file 1 ( supplementary file 2 ) lists all nmr buckets and corresponding metabolites for the urinary ( plasma ) data set with regression weights not equal to zero in at least one classification - normalization approach . again
consider the urinary aki data set .
metabolites with large absolute regression weights ( @xmath66 ) in the zero - sum models included both endo- as well as exogenous compounds .
the largest positive regression weight was assigned to a bucket at 3.285 ppm , which was identified as a superposition of myo - inositol , taurine , 4-hydroxy - propofol-4-oh - d - glucuronide , and tentatively to d - glucuronic acid as well as an unknown metabolite .
other important buckets with large positive regression weights comprised propofol - glucuronide , 4-hydroxy - propofol-1-oh - d - glucuronide , isobutyrylglycine , broad protein signals , and probably 2-oxoisovaleric acid , indoxyl sulfate , as well as unidentified metabolites . buckets with large negative regression weights correspond to a superposition of carnitine and 4-hydroxy - propofol-4-oh - d - glucuronide , n - acetyl - l - glutamine , and probably 4-hydroxyphenylacetic acid , as well as unidentified metabolites .
positive / negative regression coefficients indicate an up-/down - regulation of the corresponding metabolite in the aki in comparison to the non - aki group .
consequently , we can report an up - regulation of exogenous compounds such as glucuronide conjugates of propofol , an anesthetic agent which had been administered during the cardiac surgery and is mainly metabolized by glucuronidation @xcite .
elevated urinary levels of propofol conjugates in the aki group point to a delayed excretion of exogenous compounds caused by reduced glomerular filtration or prolonged administration @xcite .
higher absolute concentrations of propofol - glucuronide ( @xmath12-value @xmath67 ) have already been reported for the 24 h urine nmr fingerprints in @xcite .
an up - regulation of carnitine in the non - aki group is indicated by its large negative regression weight .
this observation has already been reported in @xcite , where an up - regulation of carnitine , whose main function is the transport of long - chain fatty acids into the mitochondria for subsequent beta - oxidation @xcite , in the non - aki group has been discussed as a successful protective response against ischemic injury . also , logistic zero - sum regression returned the same predictions for all patients regardless of the applied normalization strategy , summarized in figure [ classification_summary_aki ] , as well as in supplementary figure s3 . furthermore , the zero - sum constraint did not compromise predictive performance for the urinary and plasma aki data set , respectively ( table [ roc_table ] , figures [ roc_zerosum]a and [ roc_zerosum]b ) .
auc values ranged among the highest for the urinary aki data set , and yielded the largest value among all three classification methods in combination with three different normalization strategies for the plasma aki data set .
most importantly , zero - sum regression reliably identified patients that developed severe aki ( stage 2 and 3 , highlighted by yellow dashed area )
. only lasso on pqn data gave competitive results .
with respect to auc , zero - sum was still superior .
moreover , a sample - specific rescaling @xmath26 has no effect on the corresponding zero - sum prediction probabilities , as illustrated in figure [ prob_vs_gamma ] by the black dotted lines .
we first pointed out some intrinsic problems in metabolomics biomarker detection .
the identification of biomarkers strongly depends on the method selected for scaling the reference profiles . for stand - alone biomarkers , accordance across scales can be as low as 0% .
the same is true for multivariate biomarker signatures when filtering approaches like filtering by the _
t_-statistics are used , while wrapping approaches like the lasso yield somewhat more consistent yet still not scale - independent signatures .
more importantly , the prediction of a patient s outcome can change depending on the scaling method employed .
in fact , for some cases we could attribute failing prediction to inappropriate scaling of the raw data . to overcome this problem
we suggest logistic zero - sum regression , which provides a completely scale - independent analysis .
it always selects the same biomarkers , assigns the same weights to them , and predicts the same outcome no matter whether the raw data was scaled relative to total spectral area , creatinine , tsp , any other reference point , by pqn , or not at all .
furthermore , zero - sum regression was among the best aki predictors both for urinary and plasma fingerprints . also , the number of chosen features did not change significantly compared to other signatures , allowing for transfer of zero - sum signatures into clinical practice .
the interplay of statistical analysis and normalization protocols has been described by several authors @xcite . proposed
strategies to overcome these issues included the selection of the most `` robust '' normalization @xcite or the parallel application of several normalization strategies and a subsequent meta analysis of the individual results @xcite .
clearly , these strategies deal with the problem , but unlike zero - sum regression do not solve it .
the zero - sum signatures did not identify completely new biomarkers .
in fact , most of the underlying metabolites of the signatures were described before .
however , we believe that zero - sum signatures make better use of these biomarkers by combining and weighing them differently .
predictions are no longer confined to a specific normalization protocol but hold in general .
biological interpretation becomes easier as there are fewer user - dependent choices to make .
it is a set of metabolites that hold predictive information and not their abundance relative to an arbitrarily chosen reference point .
moreover , signatures can be validated across studies and tested in different labs even if scaling protocols do not match or if no scaling is applied at all .
this fact has the potential to greatly enhance the reproducibility of clinical trials .
the urinary aki data set is a good example for data where a proper normalization can not replace independence of normalization .
we believe that this data set just can not be normalized properly : varying urine density precludes tsp normalization , the strong confounding of spectral areas by d - mannitol signals and by proteinuria eliminates both normalization to the total spectral area and pqn , while creatinine excretion is entangled with kidney function and varies strongly across patients , disqualifying it from use as a common reference .
moreover , correcting for artifacts like d - mannitol requires that the existence of an artifact is known for a specific patient .
interestingly , the lasso - tsp signature worked surprisingly well in spite of strong variability in urinary density across patients .
coincidently , however , this signature had weights adding up to almost ( yet not exactly ) zero .
lasso is an algorithm used in artificial intelligence .
here it detected the intrinsic normalization problems with the tsp - normalization and `` intelligently '' decided for a scale - independent zero - sum type signature @xcite . our _ zerosum _ package can be used on a desktop computer for small data sets and reduced precision .
large clinical studies with several thousand patients require the use of an hpc infrastructure .
we here offer code that can be used for such studies . in conclusion , we provide a high - performance classification framework independent of prior data normalization , which reduces the number of user - adjustable parameters and should be ideally suited for the transfer of metabolic signatures across labs .
furthermore , we expect that the results presented here also hold for metabolomics data generated by other methods such as mass spectrometry .
the authors thank drs . gunnar schley , carsten willam , and kai - uwe eckardt for providing the urine and plasma specimens of the aki data sets .
financial support : e : med initiative of the german ministry for education and research ( grant 031a428a ) and the german research foundation ( sfb / trr-55 ) .
none .
anderson , p. e. , mahle , d. a. , doom , t. e. , reo , n. v. , delraso , n. j. , raymer , m. l. , _ dynamic adaptive binning : an improved quantification technique for nmr spectroscopic data _ , metabolomics , 2011 , 7(2):179 - 190 .
altenbuchinger , m. , rehberg , t. , zacharias , h. u. , oefner , p. j. , dettmer - wilde , k. , holler , e. , weber , d. , gessner , a. , hiergeist , a. , spang , r. , _ reference point insensitive molecular data analysis _ , bioinformatics , 2016 , doi:10.1093/bioinformatics / btw598 .
dawiskiba , t. , deja , s. , mulak , a. , zbek , a. , jawie , e. , paweka , d. , banasik , m. , mastalerz - migas , a. , balcerzak , w. , kaliszewski , k. , skra , j. , bar , p. , korta , k. , pormaczuk , k. , szyber , p. , litarski , a. , mynarz , p. , _ serum and urine metabolomic fingerprinting in diagnostics of inflammatory bowel diseases _ , world journal of gastroenterology , 2014 , 20(1):163 - 174 .
de meyer , t. , sinnaeve , d. , van gasse , b. , tsiporkova , e. , rietzschel , e. r. , de buyzere , m. l. , gillebert , t. c. , bekaert , s. , martins , j. c. , van criekinge , w. , _ nmr - based characterization of metabolic alterations in hypertension using an adaptive , intelligent binning algorithm _ , 2008 , analytical chemistry , 80(10):3783 - 3790 .
dieterle , f. , ross , a. , schlotterbeck , g. , senn , h. , _ probabilistic quotient normalization as robust method to account for dilution of complex biological mixtures .
application in 1h nmr metabonomics _
, analytical chemistry , 2006 , 78(13):4281 - 90 .
elliott , p. , posma , j. m. , chan , q. , garcia - perez , i. , wijeyesekera , a. , bictash , m. , ebbels , t. m. d. , ueshima , h. , zhao , l. , van horn , l. , daviglus , m. , stamler , j. , holmes , e. , nicholson , j. k. , _ urinary metabolic signatures of human adiposity _ , science translational medicine , 2015 , 7(285):285ra62 .
gronwald , w. , klein , m. s. , kaspar , h. , fagerer , s. r. , nrnberger , n. , dettmer , k. , bertsch , t. , oefner , p. j. , _ urinary metabolite quantification employing 2d nmr spectroscopy _ , analytical chemistry , 2008 , 80(23):9288 - 9297 .
gronwald , w. , klein , m. s. , zeltner , r. , schulze , b .-
reinhold , s. w. , deutschmann , m. , immervoll , a .-
k . , bger , c. a. , banas , b. , eckardt , k .- u .
, oefner , p. j. , _ detection of autosomal dominant polycystic kidney disease by nmr spectroscopic fingerprinting of urine _ , kidney international , 2011 , 79:1244 - 1253 .
hochrein , j. , klein , m. s. , zacharias , h. u. , li , j. , wijffels , g. , schirra , h. j. , spang , r. , oefner , p. j. , gronwald , w. , _ performance evaluation of algorithms for the classification of metabolic ^1^h nmr fingerprints _ , journal of proteome research , 2012 , 11:6242 - 6251 .
hochrein , j. , zacharias , h. u. , taruttis , f. , samol , c. , engelmann , j. , spang , r. , oefner , p. j. , gronwald , w. , _ data normalization of @xmath68h - nmr metabolite fingerprinting datasets in the presence of unbalanced metabolite regulation _ , journal of proteome research , 2015 , 14(8):3217 - 3228 .
keun , h. c. , ebbels , t. m. , antti , h. , bollard , m. e. , beckonert , o. , schlotterbeck , g. , senn , h. , niederhauser , u. , holmes , e. , lindon , j. c. , nicholson , j. k. , _ analytical reproducibility in 1h nmr - based metabonomic urinalysis _
, chemical research in toxicology , 2002 , 15(11 ) , 1380 - 1386 .
ross , a. , schlotterbeck , g. , dieterle , f. , senn , h. , _ nmr spectroscopy techniques for application to metabonomics _
, in lindon , j. c. , nicholson , j. k. , holmes , e. , ( eds . ) , _ nmr spectroscopy techniques for application to metabonomics _ ,
elsevier bv : amsterdam , the netherlands , 2007 , pp .
55 - 112 .
ritchie , m.e . ,
phipson , b. , wu , d. , hu , y. , law , c.w . , shi , w. , smyth , g.k . ,
_ limma powers differential expression analyses for rna - sequencing and microarray studies _ , 2015 , nucleic acids research 43(7 ) , e47 .
saccenti , e. , _ correlation patterns in experimental data are affected by normalization procedures : consequences for data analysis and network inference _ , journal of proteome research , 2016 , doi 10.1021/acs.jproteome.6b00704 .
viant , m. r. , lyeth , b. g. , miller , m. g. , berman , r. f. , _ an nmr metabolomic investigation of early metabolic disturbances following traumatic brain injury in a mammalian model _ ,
nmr in biomedicine , 2005 , 18(8 ) , 507 - 516 .
zacharias , h. u. , schley , g. , hochrein , j. , klein , m. s. , kberle , c. , eckardt , k .- u .
, willam , c. , oefner , p. j. , gronwald , w. , _ analysis of human urine reveals metabolic changes related to the development of acute kidney injury following cardiac surgery _ , metabolomics , 2013 , 9(3):697 - 707 .
zacharias , h. u. , hochrein , j. , klein , m. s. , samol , c. , oefner , p. j. , gronwald , w. , _ current experimental , bioinformatic and statistical methods used in nmr based metabolomics _
, current metabolomics , 2013 , 1(3):253 - 268(16 ) .
zacharias , h. u. , hochrein , j. , vogl , f. c. , schley , g. , mayer , f. , jeleazcov , c. , eckardt , k .- u .
, willam , c. , oefner , p. j. , gronwald , w. , _ identification of plasma metabolites prognostic of acute kidney injury after cardiac surgery with cardiopulmonary bypass _ , journal of proteome research , 2015 , 14(7):2897 - 2905 . | * motivation : * metabolomics data is typically scaled to a common reference like a constant volume of body fluid , a constant creatinine level , or a constant area under the spectrum .
such normalization of the data , however , may affect the selection of biomarkers and the biological interpretation of results in unforeseen ways .
* results : * first , we study how the outcome of hypothesis tests for differential metabolite concentration is affected by the choice of scale .
furthermore , we observe this interdependence also for different classification approaches .
second , to overcome this problem and establish a scale - invariant biomarker discovery algorithm , we extend linear zero - sum regression to the logistic regression framework and show in two applications to @xmath0h nmr - based metabolomics data how this approach overcomes the scaling problem .
* availability : * logistic zero - sum regression is available as an _ r _ package as well as a high - performance computing implementation that can be downloaded at https://github.com/rehbergt/zerosum . |
a gauged linear sigma model ( glsm ) in two dimensions @xcite is capable of describing a curved geometry ( typically a sympletic or khler quotient space ) through gauge theory language .
more precisely , the isometry , the isometry @xmath5 is required to be sympletic such that vector fields @xmath6 s generating it give @xmath7 where @xmath8 is a sympletic form on the sympletic manifold @xmath9 . ] of the transverse internal space gets partially gauged and chiral multiplets on the worldsheet are coupled to corresponding gauge fields .
then , the curved geometry @xmath10 arises as the supersymmetric vacuum ( moduli space ) of the 2d gauge theory .
note that @xmath11 stands for @xmath12 complexified @xmath13 s with moment maps @xmath14 @xmath15 parameterizing @xmath16 represents the lowest component of an @xmath17=(2,2 ) chiral superfield @xmath18 . also , @xmath19 s are called fayet - iliopoulos ( fi ) parameters which bring in @xmath12 khler classes .
modulo the following gauge symmetry @xmath20 is exactly the vacuum manifold denoted as @xmath10 or @xmath21 .
similarly , a hyperkhler quotient space is defined by @xmath22 ( @xmath23 ) where @xmath24 is generated by vector fields @xmath6 s with @xmath25 ( @xmath26 : metric ) and @xmath27 .
three complex structures @xmath28 @xmath29 are @xmath30 for each @xmath31 described by @xmath32 @xmath33 and transform as a triplet under @xmath34 .
our convention is as follows .
a quaternion @xmath35 with a pure imaginary part @xmath36 @xmath37 consists of a pair of complex numbers @xmath38 : @xmath39 for @xmath40 quaternions , @xmath12 moment maps associated with @xmath41 under a charge matrix @xmath42 will read @xmath43 or @xmath44 here , triplets @xmath45 s are given level sets . by definition @xmath22 has real @xmath46 dimensions . in this note , we propose two kinds of glsms which have , respectively , 8d hyperkhler manifolds and calabi - yau ( cy ) 4-folds as their moduli spaces . in the former case ,
we have followed pioneering works @xcite .
we extend their @xmath17=(4,4 ) models to include generic 8d hyperkhler geometries dual to type iib @xmath15-brane configurations @xcite .
moreover , taking an infra - red ( ir ) limit leads to frozen kinetic terms of vector - multiplets . by integrating them out , a nonlinear sigma model ( nl@xmath47 m ) can be realized in the higgs branch ; that is , the quotient space metric is now pulled back onto worldsheet s kinetic term .
the latter case is an @xmath17=(2,2 ) model which provides instead a cy 4-fold at ir .
it is not the traditional one @xcite which executes a sympletic quotient because it possesses all key features of 3d @xmath17=2 chern - simons - matter theories on a stack of m2-branes probing toric cy 4-folds .
namely , usually f - term conditions ( defining a master space ) and d - term ones as a whole give a cy 3-fold as the story happens in 4d @xmath17=1 superconformal field theories ( scft ) .
however , because of the appearance of one @xmath3 @xmath4 @xmath48 here , the d - term constraint associated with @xmath48 becomes redundant due to @xmath2 fayet - iliopoulos ( fi ) parameters and a 4-fold emerges thereof .
remarkably , we find these properties definitely show up in our model .
we demonstrate this mechanism via one explicit example a 4-fold arising from @xmath49 . in section 2 , a taub - nut ( alf ) space constructed by means of hyperkhler quotient and its corresponding glsm
are reviewed .
then , in section 3 we write down our model mentioned above whose higgs branch probes generic 8d hyperkhler manifolds .
section 4 is devoted to the model probing cy 4-folds .
finally , our results are summarized in section 5 .
to begin with , let us see how a multi - centered taub - nut ( or alf ) space can be constructed by means of hyperkhler quotient @xcite .
after then , we go to acquaint ourselves with the glsm description of it according to @xcite .
the very quotient procedure is exactly carried out in the higgs branch of the proposed glsm at ir .
non - zero hyper - multiplets take their values in a hyperkhler manifold thereof . as said in section 1 , from @xmath50
one can construct a hyperkhler manifold @xmath22 with @xmath41 being triholomorphic , i.e. it does not alter the hyperkhler structure .
we will mainly adopt the following triholomorphic right multiplication @xmath51 \nonumber\end{aligned}\ ] ] whose moment map reads @xmath52 in addition , a right multiplication @xmath53 by a unit quaternion @xmath54 ( @xmath55 ) @xmath56 does not vary the hyperkhler structure @xmath28 of @xmath57 , while a left multiplication @xmath58 which keeps the flat metric @xmath59 invariant but rotates @xmath28 plays the role of @xmath60 ( @xmath61-symmetry ) under which @xmath62 transforms as a doublet .
two operations commute . from now on ,
we usually do not distinguish between @xmath63 being pure imaginary and @xmath64 . picking up @xmath65 quaternions @xmath66 @xmath67
parameterizing @xmath68 , we want to yield a multi - centered taub - nut through dividing @xmath69 by the following triholomorphic isometry ( @xmath70 ) : @xmath71 there are @xmath72 corresponding triplet moment maps : @xmath73 put into a taub - nut form the metric of @xmath74 ( @xmath75 ) : @xmath76.\end{aligned}\ ] ] by fixing the level set at @xmath77 such that @xmath78 , it is straightforward to show that becomes @xmath79 as a final step , we need to express in terms of @xmath80 and a @xmath81-invariant @xmath82 under moment maps in , i.e. @xmath83 henceforth , instead of @xmath84 s we use variables like @xmath85 the square @xmath86 is completed and then dropped due to its @xmath81 variance .
finally , we are left with @xmath82 and @xmath87 .
further taking care of the completion of @xmath88 , one arrives at @xmath89 where @xmath90 is evaluated at @xmath91 .
take as the prototype the simplest one - nut @xmath92 .
let us review its glsm following harvey and jensen @xcite .
introduce @xmath17=(4,4 ) superfield conventions necessary for later convenience : + + @xmath17=(4,4 ) vector - multiplet @xmath93 + ( @xmath94 is an @xmath17=(2,2 ) twisted chiral superfield ) + @xmath17=(4,4 ) hyper - multiplet @xmath95 + @xmath17=(4,4 ) linear hyper - multiplet @xmath96 + the glsm lagrangian @xmath97 for an one - nut @xmath92 consists of @xmath98 fi terms are included in @xmath99 with @xmath100 where superscripts are cartesian labels . to find out the vacuum , one has to expand into ( bosonic ) component fields and @xmath101 are as follows : @xmath102 ] .
terms from @xmath103 in order are listed below ( wess - zumino gauge of @xmath104 ) @xmath105 immediately , a d - term potential ( by integrating out @xmath106 in @xmath104 ) @xmath107 and f - term potentials ( by integrating out @xmath108 , @xmath109 , @xmath110 and @xmath111 ) @xmath112 follow .
further by taking an infra - red limit ( @xmath113 ) which decouples the vector - multiplet kinetic term , we find from to that the @xmath114 branch is essentially excluded because of @xmath115 though irrelevant here . ] .
we finally obtain the supersymmetric vacuum satisfying @xmath116 which is just the standard moment map considered in .
we observe that the massive gauge field @xmath117 ( mass square @xmath118 ) manages to @xmath119 ( gauge away ) one unwanted angular variable variant under .
given that the metric @xmath120 of a quaternion is flat , we express it in a taub - nut form with @xmath121 denoting the triholomorphic @xmath13 such that @xmath122 defining @xmath123 one obtains @xmath124 further by completing @xmath125 , @xmath126 note that a remnant @xmath127 is gauged away by @xmath117 . to conclude , we have found that at ir ( @xmath113 ) the higgs branch vacuum manifold manifests itself as a hyperkhler quotient space .
integrating out @xmath117 then results in a nl@xmath47 m with an explicit @xmath128 target metric .
we are mainly interested in 8d toric hyperkhler manifolds @xmath129 which can be obtained by hyperkhler quotient of quaternions .
while @xmath130=cone(@xmath131 ) is expressed as a cone , the base seven - manifold @xmath131 is tri - sasakian and einsteinian with constant sectional curvature @xmath132 .
recently , this kind of geometry has been studied quite intensively in the context of ads / cft because 11d supergravity solutions of the type @xmath133 are dual to various 3d @xmath17=3 scfts newly found in @xcite called eschenburg space @xmath134 ( up to an orbifold identification ) is shown to be dual to new @xmath17=3 scfts in @xcite . ] . as mentioned before
, @xmath41 is part of the triholomorphic isometry of @xmath135 and preserves its hyperkhler structure .
with respect to the remaining triholomorphic @xmath136 , the metric of @xmath130 can be put into the following form ( @xmath137 $ ] ) : @xmath138 where @xmath139 and @xmath140 ( cartesian label ) is the matrix inverse of @xmath141 . ] .
conventionally , @xmath130 is embedded in m - theory and occupies ( 345678910 ) , i.e. @xmath142 and @xmath143 while circles @xmath144 stands for @xmath145 .
in fact , this metric is so rigid in the sense that it is fully determined once a proper 2 by 2 symmetric matrix @xmath146 gets specified @xcite . certainly , adding a constant part @xmath147 to @xmath146 is still a solution .
issues about @xmath147 will be addressed below . in generic non - degenerate cases where the holonomy of @xmath130 is exactly @xmath0 ( instead of @xmath148 ) , a fraction @xmath149 out of full 32 susy remains and there exist three real covariantly constant spinors rotated by @xmath150 @xmath61-symmetry of @xmath130 . since covariantly constant spinors and the triholomorphic isometry commute , the same amount of susy survive the duality map which utilizes the above @xmath136 , say , @xmath151m - theory / two - torus@xmath152 @xmath153 type iib string theory " .
consequently , via the duality chain @xmath130 gets dual to properly - oriented iib @xmath15-branes attached on a circle @xmath154 ( t - dualized @xmath155 ) . how a reduced amount of holonomy from @xmath0 to @xmath156 which results in totally @xmath157 susy occurs ?
this becomes possible when @xmath130 just reduces to two orthogonal taub - nut spaces occupying @xmath158 and @xmath159 , respectively ( up to some orbifold identification ) .
there may be two kind of situations .
one is when @xmath160 happens to factorize into a diagonal form .
the other lies in zooming in on the @xmath161-@xmath162 region ( @xmath163 ) such that @xmath147 is gotten rid of and @xmath164 may get properly diagonalized .
as advertised , some comments follow from the relation @xmath165 where we have dropped the subscript @xmath166 for asymptotic @xmath167 and @xmath146 at infinity .
though simply results from translating the asymptotic metric of @xmath136 in into its complex moduli @xmath167 , this correlation does imply an @xmath168 covariance of both 5-brane charge vectors and the iib axio - dilaton @xmath169 gets identified with the m - theory torus s complex moduli @xmath170 via the aforementioned duality chain . ] . as a matter of fact
, @xmath147 plays a role of fixing the orientation of 5-branes and subsequently determines the form of @xmath164 .
therefore as far as preserving six supercharges is concerned , @xmath147 and @xmath164 are not independent . more precisely , the normalization is that when @xmath171 ( @xmath172 ) a @xmath1735-brane . ]
is placed relative to a ( 1,0)5 ( ns5 ) brane occupying ( 012345 ) by an angle @xmath174 uniformly on @xmath175- , @xmath176- and @xmath177-plane . in other words , we have measured @xmath178 between two kinds of 5-branes according to @xcite @xmath179 combining and , we are able to tell why it is @xmath168 rather than @xmath180 that remains as a symmetry of type iib string theory by supersymmetry arguments . because a correct amount of susy should be respected in 11d @xmath181 due to its hyperkhler nature , dual iib @xmath15-branes are asked to have specific orientations .
that is @xmath168-invariant leads to @xmath168-covariant @xmath182 and @xmath183 in order to maintain definite 3/16 susy necessarily .
equipped with the above warm - up , we are ready to write down a glsm which provides triple moment maps from its d- and f - term conditions .
it realizes a 8d hyperkhler metric in its higgs branch thereof at ir upon integrating out auxiliary fields . to begin with , @xmath184 where @xmath185 this is an extension of the original work of okuyama @xcite .
our convention is as follows : @xmath186 bold symbols ( e.g. @xmath187 , @xmath188 , @xmath189 , @xmath190 , @xmath191 , etc . )
denote @xmath9-vectors , while @xmath192 represents the usual vector inner product .
also , @xmath193 will be identified with the @xmath147 part in .
we have turned off fi parameters which spoil the cone structure of @xmath130 when one zooms in on the vicinity around the origin .
in addition to couplings constants , @xmath194 are the only parameters in .
it becomes transparent later that @xmath194 is responsible for @xmath9 distinct charge vectors of iib @xmath1955-branes .
we should expand @xmath196 into components fields , integrate out auxiliary @xmath106 and @xmath197 fields . further imposing the infra - red limit @xmath198 freezes kinetic terms of vector - multiplets .
henceforth , we are left with ( bosonic part only ) @xmath199 and potentials @xmath200 a susy vacuum accompanied by @xmath201 implies @xmath9 sets of triple moment maps : @xmath202 from @xmath203 .
it is now clear that the model itself performs a hyperkhler quotient on @xmath204 quaternions by imposing which kill @xmath9 pairs of @xmath205 . as before ,
let us rewrite the flat metric @xmath206 of a quaternion into a taub - nut form with @xmath84 denoting the triholomorphic @xmath13 , i.e. @xmath207 where @xmath208 according to @xcite , @xmath164 corresponds to @xmath9 dual iib @xmath2095-branes localized at @xmath210 w.r.t .
@xmath211 which can be adjusted to @xmath212 .
since only @xmath213 and @xmath214 are invariant under the action of moment maps in , our strategy is to let @xmath9 massive gauge fields gauge away remaining @xmath81-variant angular variables . to evaluate the term @xmath215
we adopt variables @xmath216 such that every @xmath217 splits into a sum @xmath218 in dealing with , we complete in order @xmath219 @xmath220 from @xmath221 and gauge them away . finally by substituting the outcome into the desired metric form
as can be reached , i.e. @xmath222 this is exactly a nl@xmath47 m which pulls back the metric of @xmath130 .
let us propose an @xmath17=(2,2 ) glsm whose ir moduli space becomes a cy 4-fold .
our superfield conventions are inherited from section 3 .
the lagrangian under consideration is @xmath223 where @xmath224 note that @xmath225 denotes the superpotential and the gauge invariance in the first term of @xmath103 is maintained by @xmath226 bi - fundamental chiral fields @xmath227 charged under @xmath12 @xmath13 s according to a given quiver diagram with @xmath12 nodes .
the subscript @xmath228 runs over arrows representing bi - fundamental fields in the quiver diagram , while @xmath229 or @xmath230 specifies a node on which the head or tail of @xmath228 ends .
in addition , @xmath225 subject to holomorphy is again related to the given quiver @xmath231 technology one can read off @xmath225 from a given quiver .
see @xcite for an excellent review . ] .
note that only @xmath101 of the previous @xmath17=(4,4 ) multiplet @xmath232 enters @xmath196 with the lowest component @xmath233 ( @xmath234 ) .
it is apparent that both @xmath196 and the resultant ir moduli space are characterized only by @xmath235 and the responsible quiver diagram .
first of all , let us expand superfields in @xmath196 into components and integrate out auxiliary @xmath106 and @xmath197 fields . taking @xmath236 freezes kinetic terms of fields @xmath237 .
what we are left with are ( bosonic part only ) + 1 .
kinetic terms of @xmath238 ( bosonic component of @xmath239 + 2 . a remnant : @xmath240 3 .
a potential : @xmath241 where @xmath242 the vacuum manifold @xmath243 is determined by @xmath244 several comments follow .
a distinguishing feature departing from those dealt with in previous sections is that by means of @xmath245 s e.o.m .
one gauge field @xmath246 in can be dualized into a scalar @xmath247 . in the context of 3d yang - mills - chern - simons theory
, @xmath248 is referred to as a @xmath249 @xmath4 .
together with @xmath250 we can understand that the model itself executes rather a @xmath251 quotient over complex variables @xmath252 than a hyperkhler one .
let us analyze the moduli space in detail from both field - theoretical and geometrical grounds .
the constraint @xmath253 in characterizes the decoupled diagonal @xmath13 and implies as well that @xmath254 or the vector @xmath235 is orthogonal to @xmath255 . to this end , one has only @xmath256 linearly independent d - term conditions @xcite : @xmath257 this fact is consistent with that we have a dualized @xmath258 and leads naturally to @xmath259 which is a cy 4-fold by definition emerging from a 3-fold .
namely , @xmath260 is a cy 3-fold because its derivation is just the same with that of 4d @xmath17=1 scfts on a bunch of d3-branes probing cy@xmath261 cones .
+ + 2 . geometrically
, an ungauged @xmath13 @xmath48 suggests the existence of an @xmath262 in the vacuum moduli space correspondingly . according to @xmath250
, we can thus describe @xmath243 as a 3-fold fibered on a real line @xmath263 with a circle bundle fibered over them with non - trivial first chern class @xcite .
we find that the field @xmath263 in ( or @xmath250 ) serves as @xmath2 fi parameters and plays the very role of giving varying khler classes for the fibered 3-fold .
+ + 3 . let us elaborate on arguments about the @xmath264 @xmath262 bundle associated with @xmath48 .
for conceptual convenience , we can first think of it as @xmath245 charged logarithmically under @xmath12 @xmath13 s : @xmath265 consequently , defines a one - parameter gauge transformation @xmath266 .
we find that @xmath245 is especially helpful when one wants to determine the correct periodicity of @xmath262 .
in fact , there may remain some discrete amount of gauge symmetry @xmath267 of @xmath258 such that @xmath267 shrinks the circumference of @xmath262 to @xmath268 .
this is apparent from the expression of @xmath269 .
let us put things in an inverse way . from @xmath258
is spontaneously broken once @xmath245 acquires a vev @xmath270 .
but if there exists @xmath271 with @xmath272 @xmath273 , the gauge fixing becomes incomplete due to a factor @xmath274 .
this soon means that the gauge symmetry of @xmath258 is only spontaneously broken down to a discrete extent . to conclude , we should at the end divide @xmath243 further by @xmath267 .
all these are directly reminiscent of the novel mechanism in newly - developed 3d @xmath17=2 chern - simons - matter theories @xcite .
therefore , it turns out that , up to a discrete quotient @xmath267 , @xmath259 where additional real two dimensions result effectively from degrees of freedom of @xmath275 and @xmath245 are called fayet - iliopoulos and stuckelburg fields , respectively . ] .
let us review quickly stuffs about toric varieties which are highly helpful when one lifts a toric cy@xmath261 to a toric cy@xmath276 .
in general , a toric variety @xmath104 is expressed as @xmath277 and can be summarized pictorially by a toric diagram is determined by partial resolutions . ] consisting of @xmath40 @xmath278-vectors @xmath279 @xmath280 subject to @xmath281 by imposing the calabi - yau condition : @xmath282 @xmath283 , these @xmath284 s can be written as @xmath285 and are called toric data of a toric cy @xmath278-fold . for @xmath278=3 , toric data are plotted on a plane , while @xmath278=4 toric data are encoded in 3d lattice points which define a convex polytope @xmath286 . this is why we adopt the name @xmath286 in our title page because cy 4-folds are under consideration .
let us talk about the geometric meaning of toric data .
every @xmath287 assigns a shrinking 1-cycle out of @xmath288 at the @xmath228-th facet of the boundary @xmath289 where @xmath290 denotes a cone in @xmath291 over which @xmath288 is fibered over . in @xmath278=3 cases , lines where two facets meet will altogether constitute a @xmath292 diagram which in turn represents a tree formed by multiple semi - infinite ns5-branes .
although we will not go further to details , interested readers are encouraged to consult @xcite .
though lifting 3-folds to 4-folds has been studied quite heavily to date , let us just pick up one canonical example @xmath49 discussed also in @xcite . here
, @xmath293 is embedded in the hopf @xmath13 bundle of @xmath294 whose action is @xmath295 a stack of d3-branes probing @xmath49 @xcite has nine bi - fundamental chiral superfields on its worldvolume gauge theory with a superpotential ( though we will focus on its abelian version ) @xmath296 .
three nodes are labeled by integers @xmath297 .
three sets of chiral superfields are denoted as @xmath298 @xmath299 .
, height=302 ] in figure [ fff1 ] , three nodes denote three gauge groups and hence @xmath300 d - term conditions .
in addition , @xmath301 renders four more constraints @xmath302 instead , we can @xmath303 via @xmath304 , @xmath305 and @xmath306 @xmath299 accompanied by a d - term like condition : @xmath307 @xmath308 can then be thought of as @xmath309 .
alternatively , the following exact sequence @xmath310 reveals the same thing .
it merely says mathematically @xmath311 can be regarded as @xmath312 where the action of @xmath313 is spelt out by @xmath314 .
more precisely , one can express @xmath315 as a cone @xmath316 in @xmath317 .
its dual cone ( toric data ) @xmath318 in @xmath319 is generated by six lattice points , i.e. @xmath320 : @xmath321 . by definition ker(@xmath320 ) is the image of @xmath322 , so the cokernel of @xmath322 is isomorphic to @xmath320 ( @xmath314 : 1 by 6 matrix ) . to get a 3-fold
, we need to incorporate d - term constraints as well . in summary
, we can obey three steps below @xcite : + 1 .
introduce @xmath146 ( 5 by 6 matrix ) with @xmath323 and @xmath104 ( 2 by 5 matrix ) .
@xmath104 encodes two d - term constraints on five independent variables as indicated in + 2 .
concatenate @xmath324 ( 2 by 6 matrix ) and @xmath314 to yield @xmath325 ( 3 by 6 matrix ) + 3 .
computing the cokernel of @xmath326 gives us desired toric data in terms of a map @xmath327 : @xmath328 + to have a 4-fold we need to use a new @xmath104 involving only one linearly - combined d - term constraint ( irrelevant to @xmath48 ) out of three nodes and repeat the above procedure . instead of this approach , a more convenient one figured out in @xcite is to directly perform one more quotient on previous @xmath329 by another @xmath330 : @xmath331 .
translating this into the action on @xmath332 , one easily realizes that it is the combination @xmath333 orthogonal to @xmath334 with @xmath335 .
see table 1 for informations about @xmath336 responsible for the @xmath228-th node .
.nine bi - fundamental chiral matters charged under three gauge groups in the case of @xmath49 [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] arbitrary @xmath235 encountered in meet the requirement @xmath254 . also , we may demand @xmath337 technically in order to avoid a further discrete quotient @xmath267 explained around .
all in all , via @xmath338 we end up with an eight cone over a seven sasaki - einstein base denoted as @xmath339 .
its toric data are @xcite : @xmath340 these six vectors do form a 3d crystal as was shown in @xcite with an additional vector @xmath341 standing for a resolution , provided @xmath342 .
see @xcite for more details .
in this short note we have studied two kinds of glsms whose vacuum moduli spaces are real eight - dimensional hyperkhler and calabi - yau manifolds , respectively .
the higgs branch of the former @xmath17=(4,4 ) glsm manifests itself as a generic 8d hyperkhler quotient space at the infra - red regime . on the other hand ,
the latter @xmath17=(2,2 ) glsm has a more flexible superpotential @xmath225 as well as chiral matter contents spelt out by a quiver diagram .
usually , d- and f - term constraints give rise to a toric cy 3-fold just as in familiar 4d @xmath17=1 scft setups . nevertheless ,
owing to a novel mechanism introduced in section 4 , 3-folds are promoted to 4-folds exactly the same way as what happens in recently - developed 3d @xmath17=2 chern - simons - matter theories which have toric cy 4-folds as their moduli spaces . in words
, in addition to the diagonal gauge field @xmath343 , an ungauged dual photon @xmath344 as well does not enter the quotient process due to @xmath2 fayet - iliopoulos parameters .
therefore , the moduli space dimension is enhanced from six to eight .
the extra two dimensions effectively result from degrees of freedom of a chiral superfield @xmath101 .
parameters @xmath345 in our 2d lagrangian satisfying @xmath254 are arbitrary integers .
their role is analogous to 3d quiver chern - simons levels .
we discussed the whole story via a very well - developed 3-fold @xmath49 .
we would like to thank amihay hanany , kazuo hosomichi , hiroaki kanno , muneto nitta , hiroshi ohki , yutaka sakamura and especially kazumi okuyama for enlightening discussions at different intermediate stages .
tst is grateful to organizers of the wonderful workshop on branes , strings and black holes " at yukawa institute of theoretical physics .
tst is supported in part by the postdoctoral program at riken .
999 = -2pt e. witten , nucl .
b * 403 * , 159 ( 1993 ) [ arxiv : hep - th/9301042 ] .
j. a. harvey and s. jensen , jhep * 0510 * ( 2005 ) 028 [ arxiv : hep - th/0507204 ] .
k. okuyama , jhep * 0508 * ( 2005 ) 089 [ arxiv : hep - th/0508097 ] .
k. okuyama , phys .
b * 668 * ( 2008 ) 153 [ arxiv:0807.0047 [ hep - th ] ] .
d. forcella , a. hanany , y. h. he and a. zaffaroni , jhep * 0808 * ( 2008 ) 012 [ arxiv:0801.1585 [ hep - th ] ] .
j. p. gauntlett , g. w. gibbons , g. papadopoulos and p. k. townsend , nucl .
b * 500 * ( 1997 ) 133 [ arxiv : hep - th/9702202 ] .
g. w. gibbons and p. rychenkova , commun .
* 186 * ( 1997 ) 585 [ arxiv : hep - th/9608085 ] .
e. witten , jhep * 0906 * ( 2009 ) 067 [ arxiv:0902.0948 [ hep - th ] ] .
y. imamura and k. kimura , prog .
* 120 * ( 2008 ) 509 [ arxiv:0806.3727 [ hep - th ] ] .
d. l. jafferis and a. tomasiello , jhep * 0810 * ( 2008 ) 101 [ arxiv:0808.0864 [ hep - th ] ] .
s. hohenegger and i. kirsch , jhep * 0904 * ( 2009 ) 129 [ arxiv:0903.1730 [ hep - th ] ] .
d. gaiotto and d. l. jafferis , `` notes on adding d6 branes wrapping @xmath346 in @xmath347 , '' arxiv:0903.2175 [ hep - th ] . y. hikida , w. li and t. takayanagi , jhep * 0907 * ( 2009 ) 065 [ arxiv:0903.2194 [ hep - th ] ] .
d. martelli and j. sparks , phys .
d * 78 * ( 2008 ) 126005 [ arxiv:0808.0912 [ hep - th ] ] .
m. aganagic and c. vafa , arxiv : hep - th/0110171 .
m. aganagic , arxiv:0905.3415 [ hep - th ] .
o. aharony , o. bergman , d. l. jafferis and j. maldacena , jhep * 0810 * ( 2008 ) 091 [ arxiv:0806.1218 [ hep - th ] ] . | we propose two kinds of gauged linear sigma models whose moduli spaces are real eight - dimensional hyperkhler and calabi - yau manifolds , respectively . here , hyperkhler manifolds have @xmath0 holonomy in general and are dual to type iib @xmath15-brane configurations
. on the other hand , calabi - yau fourfolds are toric varieties expressed as quotient spaces .
our model involving fourfolds is different from the usual one which is directly related to a symplectic quotient procedure .
remarkable features in newly - found three - dimensional chern - simons - matter theories appear here as well , such as @xmath2 fayet - iliopoulos parameters , one @xmath3 @xmath4 and its residual discrete gauge symmetry .
riken - th-173 new gauged linear sigma models for + 8d hyperkhler manifolds + and + calabi - yau crystals yutaka baba and ta - sheng tai _ theoretical physics laboratory , riken , wako , saitama 351 - 0198 , japan _ |
the most luminous active galactic nuclei ( agn ) are known as quasars .
broad absorption line ( bal ) quasars are those quasars which show ultraviolet absorption troughs thousands of kms@xmath1 wide ( e.g. , lynds 1967 ; allen et al .
2011 ) , widths for which the accretion process in quasars is thought to be ultimately responsible .
the traditional minimum velocity width for a bal trough is 2000 km s@xmath1 at 10% depth below the continuum ( weymann et al .
objects with narrower intrinsic absorption troughs , down to 500 km s@xmath1 wide , are often referred to as mini - bal quasars ( e.g. , hamann & sabra 2004 ) .
objects with intrinsic troughs @xmath2500 km s@xmath1 wide are referred to as narrow absorption line ( nal ) quasars ( see , e.g. , ganguly & brotherton 2008 ) .
intrinsic troughs are those which arise from gas connected with the accretion process onto the quasar , as opposed to arising elsewhere in the host galaxy , but it is not always possible to determine the origin of a given absorption system . if studying intrinsic quasar absorption regardless of velocity width is the goal , less restrictive minimum width definitions for bal troughs can be established ( hall et al .
2002 ; trump et al .
however , such definitions also run the risk of increased contamination from intervening absorption systems ( knigge et al .
intervening absorption systems unrelated to the quasar will generally have widths @xmath2500 km s@xmath1 , but blending and clustering can produce apparently broader absorption profiles , especially in spectra where the resolution or signal - to - noise ratio ( or both ) is relatively low .
bal quasars are often subdivided into three subtypes depending on the ionization stages seen in absorption ( e.g. , hall et al . 2002 ) .
high - ionization bal quasars ( hibals ) have absorption from civ@xmath31548,1550 ( all wavelengths in ) , nv@xmath31238,1242 , and ovi@xmath31031,1037 , and are the most common subtype .
low - ionization bal quasars ( lobals ) have high - ionization absorption plus absorption from cii@xmath41334 , aliii@xmath31854,1862 , and/or mgii@xmath32796,2803 .
iron low - ionization bal quasars ( felobals ) have high- and low - ionization absorption plus absorption from excited states of feii and/or feiii , and are the least common subtype .
bal troughs are found over a wide range of velocities relative to the quasar , both collectively and in individual objects .
the traditional velocity range over which the strength of bal troughs detected in civ@xmath31548,1550 is evaluated is @xmath525000 km s@xmath1 to @xmath53000 km s@xmath1 ( weymann et al . 1991 ) , where we adopt the convention that negative velocities denote absorption blueshifted from the quasar redshift .
the high - velocity cutoff was established to avoid confusion with siiv / oiv]@xmath41400 emission and siiv@xmath31393,1402 bal troughs ; the low - velocity to avoid strong associated narrow - line civ complexes . in this paper
we adopt the convention that quasars with troughs @xmath62000 km s@xmath1 wide , regardless of the trough velocity offset , are bal quasars . in that case ,
the highest velocity known for a bal outflow is either @xmath556000 km s@xmath1 in pg 2302 + 029 ( jannuzi et al . 1996 ) or @xmath566000 km s@xmath1 in h 1414 + 089 ( foltz et al . 1983 ) .
extending the velocity limits within which bal troughs are measured to the widest possible wavelength range accessible in a given quasar sample is necessary if the goal is to study all broad intrinsic absorption in quasars .
one drawback of that approach is contamination of low - velocity troughs with complexes of associated narrow - line absorption at the systemic redshift ; studies at spectral resolution sufficient to resolve such complexes can help account for this effect .
another drawback is contamination of high - velocity troughs in one transition with low - velocity troughs of a shorter - wavelength transition ( e.g. , high - velocity civ@xmath31548,1550 with low - velocity siiv@xmath31393,1402 ) .
physical considerations based on elemental abundances and ionization fractions can be used to correcly identify troughs in such cases ; for example , civ absorption can be seen without accompanying siiv , but the reverse is not true ( see , e.g. , gibson et al . 2010 ) .
as long as care is taken to account for the above effects , the ideal starting velocity for measuring the strength of bal troughs is easy to define : the starting velocity for a given trough should be chosen to encompass the entire trough in that transition . in principle , studying bal troughs may require starting velocities which are redshifted relative to the quasar rest frame .
hall et al .
( 2002 ) presented two quasars with mgii troughs showing absorption both blueshifted and redshifted relative to the quasar rest frame : sdss [email protected] ( j1125 ) and sdss [email protected] ( j1128 ) .
we suggested that these are systems where the quasar s extended ultraviolet continuum source is seen through an outflow which is dominated along our line of sight by its rotational velocity .
this possibility had earlier been raised in the context of nal outflows by ganguly et al .
( 2001 , their 5.3 ) . here
we use data from the sloan digital sky survey ( sdss ; york et al .
2000 ) and the sdss - iii ( eisenstein et al . 2011 ) baryon oscillation spectroscopic survey ( boss ; dawson et al .
2013 ) , discussed in [ data ] , to present seventeen quasars with broad absorption troughs along our line of sight which are redshifted relative to the quasar s systemic redshift and are seen in multiple transitions including civ , siiv and nv ( [ cands ] ) .
we consider various properties of this population of objects in [ props ] , discuss several possible explanations for these objects in [ poss ] , and discuss some implications and tests of those explanations in [ end ] .
notes on confirmed , candidate and rejected objects are presented in the appendices .
the sdss - i and sdss - ii surveys used two fiber - fed , double spectrographs to obtain resolution @xmath82100 spectra over 38009200 for @xmath910@xmath10 galaxies and @xmath910@xmath11 quasar candidates ( stoughton et al .
2002 ; abazajian et al .
the sdss - iii initiative consists of four surveys including the boss .
the boss will obtain spectra of @xmath91.5@xmath1210@xmath10 luminous red galaxies and @xmath91.5@xmath1210@xmath11 quasars at @xmath13 ( dawson et al . 2013 ) . all known @xmath13 quasars in the boss footprint which are point sources in sdss imaging ( fukugita et al .
1996 ; gunn et al . 1998 ; aihara et al .
2011a , b ) , including bal quasars , are being targeted ( ross et al . 2012 ) .
in addition , @xmath92000 known bal quasars with sdss spectra are being targeted for reobservation via a boss ancillary project ( dawson et al .
together , the above observations will enable unprecedented studies of bal quasar variability on multi - year timescales ( e.g. , filiz ak et al .
2012 ) .
the boss uses the same 2.5 m sloan foundation telescope ( gunn et al .
2006 ) as sdss - i / ii did , but the fiber - fed spectrographs have been upgraded with 1000 2@xmath14 optical diameter fibers instead of 640 3@xmath14 ones , improved optics , higher throughput gratings over a wider spectral range of 360010400 at a resolution @xmath15 , and new ccds with improved blue and red response ( smee et al .
the first boss spectra have been publicly released as part of the sdss data release nine ( dr9 ; ahn et al .
2012 ) . the boss spectra shown herein are from the boss reduction pipeline version v5_4_45 ( bolton et al .
2012 ) for all but one object . that version of the pipeline has a systematic flux excess shortward of 4100 at a level of @xmath162.5% of the flux at 5600 , increasing to @xmath1612.5% at 3600aa ( 2.4.1 of pris et al .
2012 ) ; this fact rarely affects the conclusions we draw from our spectra , but should be kept in mind .
all sdss spectra shown herein are the improved sky - subtraction versions of wild & hewett ( 2010 ) .
the boss and sdss spectra shown herein have not been corrected for galactic extinction . in all spectra , we interpolated over narrow regions near strong night sky lines if large flux residuals from the sky subtraction were present .
@lrrccclccc@ name ( sdss j ) & ra & dec & @xmath17 & redshift & redshift & bal & trough & first & @xmath18 + & ( j2000 ) & ( j2000 ) & mag . &
@xmath19 & source & type & shape & mjy beam@xmath1 & + + [email protected] & 7.104254 & 1.101175 & 20.509 & [email protected] & hw10 & lo & @xmath21 & @xmath20.73 & @xmath527.00 + [email protected] & 27.124227 & 1.504184 & 20.415 & [email protected] & ciii ] pca & lo & _ v _ & @xmath21.06 & @xmath526.40 + [email protected] & 121.437461 & 26.684147 & 21.745 & [email protected] & ciii ] pca & lo & _ v _ & @xmath20.96 & @xmath524.81 + [email protected] & 127.078399 & 36.466304 & 19.767 & [email protected] & full pca & lo & @xmath22 & @xmath20.93 & @xmath526.52 + [email protected] & 127.626083 & 16.912417 & 19.160 & [email protected] & inspection & lo & @xmath21 & @xmath20.99 & @xmath527.17 + [email protected] & 145.287167 & @xmath52.495776 & 20.197 & [email protected] & ciii ] pca & hi & @xmath22 & @xmath21.01 & @xmath526.90 + [email protected] & 154.942000 & 5.256583 & 20.978 & [email protected] & ciii ]
pca & felo & @xmath21 & @xmath21.00 & @xmath525.34 + [email protected] & 158.551375 & 7.334333 & 18.161 & [email protected] & hw10 & lo@xmath23 & @xmath22 & @xmath21.01 & @xmath526.85 + [email protected] & 176.729393 & 33.130585 & 19.294 & [email protected] & ciii ] pca & lo ?
& _ v _ & @xmath20.95 & @xmath527.26 + [email protected] & 176.983335 & @xmath52.839839 & 19.278 & [email protected] & hw10 & lo & _ v_/@xmath24 & @xmath21.03 & @xmath527.17 + [email protected] & 200.887578 & 0.776082 & 20.286 & [email protected] & inspection & lo ? & @xmath21 & @xmath20.98 & @xmath526.02 + [email protected] & 219.938667 & 4.735889 & 20.850 & [email protected] & ciii ] pca & lo ? & _ v _ & @xmath21.00 & @xmath525.56 + [email protected] & 220.231641 & 31.847709 & 20.295 & [email protected] & ciii ] pca & lo & @xmath22 & @xmath20.96 & @xmath526.42 + [email protected] & 247.024167 & 47.737667 & 18.481 & [email protected] & hw10 & hi & @xmath22 & @xmath20.93 & @xmath526.38 + [email protected] & 257.472000 & 27.087944 & 20.560 & [email protected] & ciii ] pca & lo & @xmath21 & @xmath20.98 & @xmath526.36 + [email protected] & 261.018500 & 31.594333 & 19.600 & [email protected] & ciii ] pca & lo & @xmath21 & @xmath20.95 & @xmath526.86 + [email protected] & 329.267750 & @xmath50.371583 & 20.071 & [email protected] & ciii ] pca & lo & _ v _ & @xmath20.73 & @xmath526.18 + + [email protected] & 171.358833 & 0.483694 & 17.896 & [email protected] & hw10 & felo & _ v _ & @xmath20.99 & @xmath525.54 + [email protected] & 172.117958 & 1.227194 & 18.366 & [email protected] & hw10 & felo & _ v _ & @xmath20.98 & @xmath525.16 + + [email protected] & 12.625542 & 2.654167 & 19.987 & [email protected] & full pca & lo & @xmath21 & @xmath20.93 & @xmath525.94 + [email protected] & 189.754167 & 1.803722 & 19.973 & [email protected] & inspection & hi & _ v _ & @xmath20.76 & @xmath525.85 + [email protected] & 199.155275 & @xmath50.610007 & 18.049 & [email protected] & hw10 & fe?lo & @xmath22 & [email protected] & @xmath525.55 + [email protected] & 205.682805 & 36.383885 & 21.340 & [email protected] & inspection & hi & @xmath22 & @xmath20.95 & @xmath525.14 + [email protected] & 248.332333 & 19.149083 & 17.412 & [email protected] & mgii pca & lo ? & _ v _ & [email protected] & @xmath528.31 + [email protected] & 256.235083 & 23.473806 & 18.918 & [email protected] & inspection & lo & @xmath21 & @xmath20.94 & @xmath528.22 + [email protected] & 323.425278 & 7.235772 & 21.416 & [email protected] & inspection & hi & _ v _ & @xmath20.83 & @xmath524.55 + + the @xmath17 column gives the psf magnitude in the @xmath25 band .
redshifts ( see [ zsys ] ) are given with uncertainties derived from the emission lines used to measure the redshifts ; _ not _ included are the systematic uncertainties related to the shifts of those lines from the systemic redshift of each quasar .
redshift sources are inspection ( for our own inspection redshifts ) , full , ciii ] or mgii pca ( principal component analysis ; pris et al .
2012 ) , and hw10 ( hewett & wild 2010 ) .
the bal type is hi for high - ionization , lo for low - ionization , and felo for iron low - ionization ( [ intro ] ) .
trough shapes are discussed in [ uvw ] .
the first column gives the peak flux or limit in the first catalog ( becker , white & helfand 1995 ; hodge et al .
the @xmath18 column gives the absolute @xmath25-band magnitude from pris et al .
( 2012 ) ( schneider et al .
( 2010 ) for j1628 ) .
+ @xmath23 j1034 has blueshifted low - ionization absorption but no clear redshifted low - ionization absorption .
[ tab1 ] over the course of the sdss and now the boss , we have accumulated a sample of bal quasars whose rest - frame spectra exhibit troughs with absorption at wavelengths longer than the rest frame wavelength of one or more transitions .
these objects were found by visual inspection of over 100,000 quasar spectra by the authors , including over 87,000 quasars in the data release nine quasar catalog ( pris et al .
2012 , hereafter dr9q ) . as with any qualitative selection criteria
, while we can not guarantee that our inspections are complete , our numbers serve as lower limits to the frequency of bal quasars with redshifted troughs in bal quasar samples .
( we can state that at an intermediate stage in the preparation of this paper , nine out of the ten redshifted bal quasars known at the time were recovered by visual inspection of all visually identified bal quasars in dr9q . ) in table 1 we give basic parameters for seventeen confirmed cases of redshifted absorption in civ , two previously known objects with redshifted absorption in mgii , and seven candidate cases of redshifted absorption .
detailed notes on each confirmed case are given in appendix [ notes ] .
we discuss our reservations about the candidate objects in appendix [ maybe ] and present a number of rejected candidates in appendix [ noway ] . the spectrum of the most striking object is shown in fig .
[ f_zoom ] and discussed below .
spectra of all objects with redshifted high - ionization absorption are shown in fig .
[ f_cands ] ( boss discoveries , with close - ups of the same spectra in fig .
[ f_closeups ] ) and fig .
[ f_cands2 ] ( sdss discovery spectra in green , with second - epoch spectroscopy from the boss in black ) .
sdss [email protected] ( j2157 ; fig .
[ f_zoom ] ) is the most dramatic example we have found of redshifted absorption .
the quasar has a redshift @xmath26 , measured from its broad siiii]@xmath41892+ciii]@xmath41908 emission as detailed in [ zsys ] . because j2157 has a prominent blue wing of broad civ emission , for comparison to it in fig .
[ f_zoom ] we plot the composite spectrum of the quartile of quasars with the largest civ blueshifts in the study of richards et al .
( 2002 ) . of their four composites ,
the one plotted best matches the emission lines of j2157 in regions not affected by absorption .
the absorption troughs in j2157 in nv@xmath31238,1242 , siiv@xmath31393,1402 and civ@xmath31548,1550 begin at a small blueshift ( @xmath27=@xmath28 km s@xmath1 in civ ) and extend smoothly to a large redshift ( @xmath27=9050 km s@xmath1 in civ ) at more or less constantly decreasing depth
. there is also redshifted absorption in pv@xmath31118,1128 ( shaded in fig .
[ f_zoom ] ) and in aliii@xmath31854,1862 .
j2157 exhibits strong , narrow ly@xmath29 emission accompanied by narrow oi@xmath41302+siii@xmath41304 , heii@xmath41640 , mgii@xmath32796,2803 , and possibly civ emission .
the presence of redshifted absorption in multiple transitions makes the identification of the troughs in j2157 and other such objects unambiguous .
the variety of the redshifted absorption troughs profiles , widths , and velocity ranges rule out an explanation of the troughs as gaps between emission features ( but see appendix [ maybe ] for additional candidates where that explanation can not be excluded ) .
the identification of these troughs as redshifted depends on the systemic redshift of these quasars being known to an uncertainty less than the trough widths .
the redshifted absorption extends to @xmath27@xmath63000 km s@xmath1 in all confirmed cases and up to @xmath27@xmath610,000 km s@xmath1 in a few cases ( see [ ai ] ) .
such values are generally larger than the uncertainties in redshift measurements made using low - ionization broad emission lines . to estimate those uncertainties we use the results that , on average , narrow [ oiii ] in quasars is blueshifted by an average of @xmath545@xmath205 km s@xmath1 from the host galaxy redshift defined by caii k absorption ( hewett & wild 2010 ) , mgii is redshifted relative to [ oiii ] by 97@xmath20269 km s@xmath1 ( richards et al .
2002 ) , and the ciii ] complex complex can introduce offsets of order a few 100 km s@xmath1 into the determination of the systemic redshift from it .
this complex is a blend of ciii]@xmath41908 , siiii]@xmath41892 , and feiiiuv34 @xmath4@xmath4@xmath4 1895,1914,1926 , plus aliii@xmath31854,1862 if the line profiles are sufficiently broad . at densities
@xmath3010@xmath11@xmath31 there will also be significant contributions from [ ciii]@xmath41907 and [ siiii]@xmath41883 ( osterbrock & ferland 2006 ) . ]
is blueshifted relative to mgii by @xmath5827@xmath20604 km s@xmath1 ( shen et al . 2007 ) .
therefore , relative to the host galaxy redshift , mgii redshifts should on average be redshifted by 52@xmath20269 km s@xmath1 , and ciii ] redshifts should on average be blueshifted by @xmath5775@xmath20661 km s@xmath1 .
we do not correct redshifts we obtain from those transitions for the above offsets , but those offsets and scatter should be kept in mind as limits on the current accuracy of our redshifts .
we examined the spectra of our objects to determine ` inspection redshifts ' which were then compared to the various redshifts tabulated in dr9q , and , for quasars in the sdss data release seven ( dr7 ; abazajian et al .
2009 ) quasar catalog ( schneider et al . 2010 ) , to the redshifts computed by hewett & wild ( 2010 ) .
these objects have unusual spectra which affect the accuracy of redshifts measured by fitting the overall spectrum with a principal component analysis ( pca ) reconstruction with the redshift as a free parameter ( the dr9q full pca redshifts ) , but which rarely affect fits to the ciii ] emission region .
therefore , the default redshifts adopted for our objects are the redshifts measured from the peak of a pca reconstruction of the ciii ] emission line ( see dr9q ) .
the available ciii ] pca redshifts agreed very well with our inspection redshifts except in two cases , with an average @xmath32 } - v_{insp } = ( 60 \pm 270)$ ] km s@xmath1 excluding the two special cases .
those cases are the candidates j0050 ( @xmath33 km s@xmath1 ) and j1704 ( @xmath34 km s@xmath1 ) . in the former case
we believe the ciii ] pca redshift to be incorrect and in the latter we believe the true redshift is bracketed by our inspection redshift and the ciii ] pca redshift ( see appendix [ maybe ] ) . for objects without ciii ] pca redshifts available , we use redshifts from hw10 whenever those redshifts were plausible .
the remaining cases are discussed individually in appendix [ notes ] .
all of our seventeen confirmed objects have maximum redshifted trough velocities of 3170 km s@xmath1 or greater , equivalent to requiring deviations from the mean of the ciii ] redshift distribution of 3.6@xmath35 or greater ( we do not quote probabilities because we do not know how gaussian the distribution is , especially in the tails ) . overall , our objects do not have maximum redshifted trough velocities small enough that the redshifted absorption could be spurious , arising from an outlying redshift measurement from the ciii ] emission line .
finally , a k - s test reveals no statistically significant difference between the redshift distribution of our high - ionization objects and that of the dr9q bal sample .
we discuss here several objects of particular interest .
notes on all objects with redshifted high - ionization absorption are given in appendix [ notes ] .
sdss [email protected] ( j0941 ; fig .
[ f_cands]e ) , at @xmath19=3.446 from its dr9q ciii ] pca redshift , is morphologically classified as a galaxy in the sdss .
its image appears extended in the nw - se direction .
these properties make j0941 a prime candidate for being a binary quasar system where the bal outflow of one quasar is silhouetted in front of a second quasar ; see [ binary ] .
note that j0941 was not targeted as a boss quasar candidate because all such candidates are required to be unresolved ( ross et al .
it was targeted as part of the cmass ( `` constant mass '' ) galaxy sample designed to select galaxies at @xmath36 ( dawson et al .
it is unlikely to be a lensed quasar superimposed on a @xmath37 galaxy : the rest - frame equivalent width of broad ciii ] in the observed spectrum is typical , so the observed flux must be dominated by the quasar and not a galaxy at @xmath37 .
if j0941 is a lensed quasar , the lensing galaxy must be of sufficiently high redshift or low optical luminosity ( or both ) to leave no signature in the boss spectrum .
sdss [email protected] ( j1019 ; fig .
[ f_cands]f ) has ly@xmath29 which peaks at @xmath19=2.4685 , but we adopt the dr9q ciii ] pca redshift of @xmath19=2.452 as systemic .
j1019 has mostly redshifted absorption in a wide range of transitions : feiii multiplets uv48 , uv50 and uv34 , aliii , alii@xmath41670 , civ , siiv , cii@xmath41334 , siii@xmath41304 , siii@xmath41263 , nv , and apparently ciii*@xmath41175 , pv@xmath31118,1128+feiii@xmath41122 , siv@xmath41062+siv*@xmath41072 and probably ovi@xmath31031,1037 .
the last five transitions are in the ly@xmath29 forest and are thus less certain identifications .
if confirmed , pv and siv would indicate a high column density absorber ( borguet et al .
absorption from feiiiuv34 ( ep 3.73ev ) and feiiiuv48 ( ep 5.08ev ) has been seen before in bal quasars ( hall & hutsemkers 2003 ) , but this is the first reported case of a trough from feiiiuv50 ( ep 7.86ev ) .
absorption from feii is not clearly detected despite the detection of siii and the similarity of the two elements abundances and ionization potentials for ionization stages i iii .
we discuss this quasar further in [ dist ] .
sdss [email protected] ( j1628 fig .
[ f_cands2]d ) at @xmath19=1.5949 has broad civ absorption blueshifted by up to @xmath59810 km s@xmath1 and a redshifted civ trough extending to 12400 km s@xmath1 .
we have previously outlined two alternative explanations for the latter trough ( hall et al .
it could be a high velocity ( @xmath546000 km s@xmath1 ) aliii trough without accompanying mgii .
the required velocity is less than the maximum velocity known for civ , but considerably larger than the maximum velocity of 25000 km s@xmath1 previously identified for aliii ( in sdss j154303.24 + 264052.3 ; allen et al . 2011 ) .
furthermore , the lack of bal quasars with aliii seen between 25000 km s@xmath1 and 45000 km s@xmath1 makes this explanation unlikely
. the trough could also be heii@xmath41640 absorption associated with the blueshifted civ outflow .
no unambiguous detection of heii absorption in a bal quasar has been reported , but it has been seen in disc - wind outflows in galactic objects ( hartley et al .
such absorption would indicate gas of high density and with a moderately high ionization parameter ( wampler , chugai & petitjean 1995 ) .
both of the above hypotheses remain technically viable for this object ; however , the confirmed existence of redshifted civ absorption in other quasars makes these alternate explanations less likely.463718.4 ( j1048 ) at @xmath38 has blueshifted siiv and civ absorption plus absorption which could be civ redshifted by @xmath39 km s@xmath1 or heii@xmath41640 absorption in the blueshifted outflow .
the putative redshifted civ absorption in j1048 appears less prominent in the spectrum presented by ( 2010 ) , however .
the trough may have varied , but the significance of its detection is difficult to evaluate since no error array accompanies either published spectrum . ]
in addition to the new discoveries above , j1125 and j1128 from hall et al .
( 2002 ) remain cases of redshifted mgii absorption,0058 and 3c 288.1 were candidates for redshifted absorption .
sdss spectra obtained of those sources reveal that the absorption is not redshifted in either case .
it is at the systemic redshift as given by mgii for un j1053@xmath50058 and by numerous narrow lines for 3c 288.1 . ] as discussed in
[ intro ] . in [ low ]
we discuss new boss spectra obtained for both objects .
our sample includes four quasars with sdss spectra of redshifted civ absorption and two with sdss spectra of redshifted mgii absorption , all of which now also have boss spectra .
figure [ f_cands2 ] shows the unscaled sdss and boss spectra for the redshifted civ absorption objects . in the case of j1147 ,
the sdss spectrophotometry disagrees with both the sdss photometry and the boss spectrophotometry , and so we plot only the boss spectrum .
( there are also two quasars with candidate redshifted civ absorption which have sdss spectra ; see figure [ f_cands3 ] . ) between spectroscopic epochs separated by years in the rest frame , quasars can vary in flux level and continuum slope as well as in absorption properties ( e.g. , foltz et al .
1987 ; gibson et al . 2010 ; rodrguez hidalgo , hamann & hall 2011 ) . to compare absorption troughs in these quasars sdss and boss spectra
we scale and tilt the boss spectra to match the sdss spectra .
we scale by a constant factor to account for flux variability and tilt through multiplication by a power - law to account for continuum slope differences ( apparent or real ) . for display purposes ,
we normalize the spectra separately for each figure rather than using a single global fit .
the results of this approach are shown for j1034 in figure [ fj1034 ] and for j1628 in figure [ fj1628 ] . in each figure , the left- and right - hand panels show the unscaled and rescaled boss spectra ( black ) , respectively , along with the unscaled sdss spectrum ( green ) . in the right - hand panels
, around each boss spectrum we show in grey the @xmath40 statistical uncertainty range , where @xmath35 accounts for both the sdss and scaled boss uncertainties .
if the normalized absorption troughs in the two spectra were identical , the grey region would include 95% of the points in the green spectrum . for j1034 , in the 1135 rest - frame days between the sdss and boss spectra the civ absorption at the systemic redshift and at redshifted velocities weakened in relative strength , resulting in a stronger narrow emission peak for civ .
the normalized blueshifted absorption did not change significantly .
the changes could be due to a strengthening of the underlying civ emission line plus a velocity - dependent broad - line region covering factor . in any case
, j1034 shows that redshifted and systemic absorption can vary in normalized spectra independently of the blueshifted absorption . for j1628 , in the 1559 rest - frame days between the sdss and boss
spectra the civ emission line appears to have strengthened somewhat ( between 1520 and 1560 rest - frame ) .
the absorption at blueshifted velocities @xmath41
km s@xmath1 did not change significantly .
the redshifted absorption trough has shifted to smaller redshifted velocities , by about 2000 km s@xmath1 on average . whether this is a case of deceleration or just covering factor variations as a function of velocity
can not be determined without further observations .
if this is a case of gas undergoing decelerating infall due to outward radiative acceleration , the acceleration amounts to @xmath42 cm s@xmath43 . for comparison
, @xmath44 is predicted in the main acceleration region of a disc wind in the model of murray et al .
if the observed absorption arises in decelerating infalling gas , and assuming constant ionizing luminosity , the deceleration should be seen to increase with time as the gas moves to smaller radii .
an additional result of this approach is shown in fig .
[ fjothers ] .
no significant variations are seen in the absorption in j1147 over 886 rest - frame days .
finally , in figure [ fj0028three ] we show one sdss and two boss epochs of spectroscopy for j0028 , with no scaling . in the 571 rest - frame days between the sdss observation and the first boss observation of j0028 ( green to black spectra in fig .
[ fj0028three ] ) , the flux level of the unabsorbed continuum brightened by @xmath950% , the flux levels within the two siii troughs and the cii trough increased by nearly 50% , but the flux levels within the siiv , civ and nv troughs , both redshifted and blueshifted , did not change significantly except in the most highly blueshifted parts of the siiv and civ troughs . in the 52 rest - frame days between the two boss observations of j0028 ( black to magenta spectra in fig .
[ fj0028three ] ) , the flux level across most of the spectrum decreased by @xmath910% . between both pairs of spectra
, the trough variability showed no evidence for any change with velocity across the systemic redshift dividing redshifted and blueshifted absorption .
both quasars with previously known redshifted mgii absorption , j1125 and j1128 , have also been observed by the boss ( figure [ fj11 ] ) .
we have adopted the redshifts from hewett & wild ( 2010 ) for both these objects .
the redshift for j1128 is almost identical to that used in hall et al .
the redshift for j1125 ( @xmath45 , primarily from [ oii ] ) is 338 km s@xmath1 smaller than the @xmath46 used in hall et al .
( 2002 ) . that latter redshift corresponds to narrow balmer - line absorption which was recognized in hall et al .
( 2002 ) and assumed to be host - galaxy absorption .
however , a small but growing number of quasars are now known to exhibit balmer - line absorption in their outflows ( hall 2007 ; aoki 2010 ; and references therein ) .
furthermore , [ oii ] is offset by only @xmath521@xmath205 km s@xmath1 on average from host - galaxy absorption in sdss quasars ( hewett & wild 2010 ) .
therefore the balmer - line absorption is most likely part of the redshifted absorption in this object , and we no longer adopt the redshift of that absorption as systemic . both quasars show broad balmer - line emission , narrow [ oii ] and [ oiii ] emission , and relatively narrow ( few 1000 km s@xmath1 wide ) absorption from many ground - state ( feii ) and excited - state ( feii * ) transitions of singly ionized iron ( moore 1950 ) , as well as absorption in mgii .
superpositions of these narrow lines form the apparent broad troughs in the spectra .
there is also narrow feii emission in the spectra ; for example , the narrow emission lines flanking h@xmath47 in j1128 are slightly redshifted feii ( vron - cetty , joly & vron 2004 ) . as in the previous section , to compare absorption properties at both epochs we used a constant times a power law to rescale the boss spectra . for these objects we scaled to the sdss spectra in continuum regions near 2100 and 2910 rest frame .
the resulting spectra are shown in fig .
[ fj11s ] ; again , the green spectrum precedes the black spectrum . around each boss spectrum ( black ) , we show in grey the @xmath40 statistical uncertainty range . in both objects ,
the two normalized spectra show small but significant differences near mgii which can not be matched by smooth continuum variations . the mgii absorption , emission , or both must have varied . in j1125 ,
the blueshifted mgii absorption has weakened more than the redshifted absorption has .
the underlying mgii emission may have strengthened relative to the continuum ( figure [ fj11 ] is consistent with constant emission - line flux and weakening continuum flux ) , but different behaviour in blueshifted and redshifted absorption would still be required if that were the case .
stronger emission and apparently weakened blueshifted absorption are also seen in h@xmath47 and h@xmath48 ( not shown ) .
also , the feii * absorption between 23002630 has apparently weakened , with most of the scaled boss spectrum @xmath49 above the sdss spectrum . in j1128 ,
there have been @xmath50 decreases in the strengths of the redshifted absorption troughs in mgii and feii * multiplets uv1 , uv62 and uv63 , visible just to the right of the vertical lines marking those transitions in the bottom panel of figure [ fj11s ] .
there has been no significant weakening in the blueshifted absorption in those transitions . as with civ absorption in the previous section
, it appears that redshifted mgii absorption can vary independently of blueshifted absorption .
we found no x - ray observations of useful depth of our objects in the _ chandra _ , _ xmm - newton _ , _ suzaku _ , _ bepposax _ , _ swift xrt _ , or pointed _ rosat _ archives , which we searched using heasarc browse .
we searched for counterparts of our objects within 2 arcseconds in the wide - field infrared survey explorer ( wise ; wright et al .
2010 ) all - sky source catalog .
fourteen objects with confirmed civ absorption were detected ( all except j0805 , j0828 and j1709),8 to the west which has a photometric redshift of @xmath51 ( oyaizu et al .
along with four of seven candidates ( all except j1342 and j2133 ) and all objects with confirmed or candidate redshifted mgii absorption .
three of the five non - detections are the three faintest objects in our sample , all with @xmath52 .
none of our objects exhibit obviously unusual @xmath53 , @xmath54 or @xmath55 colors as compared to other quasars in dr9q . a full multiwavelength study of these objects to determine how well their spectral energy distributions match those of standard bal and non
- bal quasars will be very useful , but is beyond the scope of this work .
we note in passing that none of our confirmed objects and only two of our candidates have radio detections in first ( table 1 ) , one with candidate redshifted civ absorption and one with candidate redshifted mgii absorption .
the detection rate of dr9q quasars at @xmath56 ( where boss spectra cover siiv to civ ) is 3.26% , so between zero and two radio - detected objects in our redshifted civ sample is statistically consistent with expectations .
the absorption profiles of the redshifted troughs ( or the redshifted parts of the troughs ) in these objects resemble those of blueshifted troughs in many ways .
when present , troughs from low - ionization species ( ionization stages i to iii ) are generally narrower and weaker than those from high - ionization species ( stages iv and higher ) ; e.g. , aliii in j2157 ( fig .
[ f_zoom ] ) , or siii@xmath41304 and cii@xmath41334 in j1019 ( fig .
[ f_cands]f and [ f_closeups]f ) .
the lack of objects with civ absorption much stronger than absorption in other transitions is unlikely to be due purely to the selection effect of typically requiring absorption in multiple troughs to identify a quasar as having redshifted absorption .
we have found no cases where strong , redshifted civ is present but redshifted siiv is not ; only j1324 ( appendix [ maybe ] ) comes close .
the residual intensities are similar in transitions with very different transition probabilities from species of very different abundances , indicative of partial covering of the emission regions at optical depth @xmath57 ( e.g. , arav et al .
1999 ; crenshaw , kraemer & george 2003 ) .
because the continuum emission region and the broad emission line region ( belr ) have very different sizes , the bal gas may have different covering factors of each region .
examination of these objects spectra near civ shows that both the continuum source and the belr can be fully covered by gas at systemic and blueshifted velocities .
however , in no case is the continuum source fully covered by redshifted absorbing gas .
moreover , the redshifted gas need not cover the civ belr at all in any of these objects ( though if it does not , the continuum source would be close to fully covered in j0830 , j1019 and j1709 ) .
that possibility means that the redshifted absorbing gas could be located interior to the belr ( e.g. arav et al . 1999 ) , though it does not need to be ( we can not rule out equal covering of the continuum source and the belr ) . the only exception to the above is the candidate j0050 ( appendix [ maybe ] ) , in which the absorption has nearly full coverage of both the continuum and the civ belr . broadly speaking , we can classify our quasars with high - ionization troughs into three categories based on the shapes of their absorption troughs : _
v_-shaped , @xmath22-shaped , or @xmath21-shaped . see [ ai ] for a discussion of trough shapes in terms of quantitative absorption strengths .
six of our quasars , including j2157 , have _ v_-shaped absorption troughs which are strongest at their short - wavelength ends and which extend smoothly to longer wavelengths at decreasing depths .
( in some cases the civ trough is interrupted by narrow emission , but the siiv trough shows the underlying _ v_-shaped trough . ) these quasars are j0148 ( fig .
[ f_cands]a and [ f_closeups]a ) , j0805 ( fig . [ f_cands]b and [ f_closeups]b ) , j1146 ( fig .
[ f_cands]g and [ f_closeups]g ) , j1147 ( a borderline case ; see fig .
[ f_cands2]c and fig . [ fjothers ] ) , j1439 ( fig . [ f_cands]i and [ f_closeups]i ) and j2157 ( fig .
[ f_zoom ] ) plus , if confirmed , the candidates j1239 , j1633 , and j2133 ( appendix [ maybe ] ) .
our two low - redshift quasars with redshifted mgii absorption ( j1125 and j1128 ; fig . [ fj11s ] ) also have _ v_-shaped troughs , albeit quite narrow ones .
five of our quasars have what appear to be separate blueshifted and redshifted absorption troughs , with the same structure in civ and siiv ; we refer to such troughs as @xmath22-shaped .
these quasars are j0828 ( fig .
[ f_cands]c and [ f_closeups]c ) , j0941 ( fig .
[ f_cands]e and [ f_closeups]e ) , j1034 ( fig .
[ f_cands2]b ) , j1440 ( fig .
[ f_cands]j and [ f_closeups]j ) and j1628 ( fig . [ f_cands2]d ) plus , if confirmed , the candidates j1316 and j1342 ( appendix [ maybe ] ) . the blueshifted absorption is stronger than the redshifted absorption in all cases , although the relative strengths are nearly equal in j0941 and j1324 . in the case of j1034 ,
the weakness of the civ emission line between the two troughs means the absorption is likely to be continuous over a span of @xmath611000 km s@xmath1 .
j1034 may have less absorption at the systemic redshift or it may have a higher covering factor of the continuum than of the emission - line region .
six of our quasars have troughs which have relatively sharp edges at both their short- and long - wavelength ends and which reach maximum depth near the trough center ( apart from occasional weak , narrow emission in civ ) .
we refer to such troughs as @xmath21-shaped .
these quasars are j0028 ( fig .
[ f_cands2]a ) , j0830 ( fig .
[ f_cands]d and [ f_closeups]d ) , j1019 ( fig .
[ f_cands]f and [ f_closeups]f ) , j1323 ( fig .
[ f_cands]h and [ f_closeups]h ) , j1709 ( fig .
[ f_cands]k and [ f_closeups]k ) and j1724 ( fig .
[ f_cands]l and [ f_closeups]l ) plus , if confirmed , the candidates j0050 and j1704 ( appendix [ maybe ] ) . note that four of our quasars have absorption only at redshifted velocities : j1019 , j1146 , j1709 and j1724 plus , if confirmed , the candidate j0050 ( appendix [ maybe ] ) . among bal quasars with redshifted civ absorption ,
lobal quasars are greatly overrepresented .
of our 17 confirmed cases of redshifted civ absorption , 12 to 14 also have redshifted low - ionization absorption , and one more ( j1034 ) has blueshifted aliii absorption but no clear redshifted aliii absorption .
the resulting lobal fraction is 56%92% ( 90% confidence range ) .
lobal quasars make up only @xmath95% of the population of all sdss bal quasars ( and thus only 1%2% of the entire sdss quasar population ; trump et al .
2006 ; allen et al .
however , lobal quasars constitute a larger fraction of the bal quasar population in near - infrared - selected quasar samples : a factor of two larger comparing the studies of dai , shankar & sivakoff ( 2008 ) and dai , shankar & sivakoff ( 2012 ) , and a remarkable 20 lobal quasars out of 21 bal quasars in glikman et al .
our lobal classifications are based on the detection of either cii@xmath41334 or aliii@xmath31854,1862 troughs , and there are no plausible alternate identifications for those troughs . with only one felobal quasar in our sample , we can only say that they do not appear to be overrepresented among bal quasars with redshifted civ absorption to the same extent that lobal quasars are .
siiv absorption is also more common among bal quasars with redshifted civ absorption than among all bal quasars .
( this is unlikely to be a selection effect ; see [ uvw ] . )
allen et al .
( 2011 ) find siiv absorption in only 42.5%@xmath201.4% of bal quasars with civ absorption .
redshifted siiv absorption is seen in all but one of our bal quasars with redshifted civ absorption and spectral coverage of siiv .
the exception is the candidate object j1342 ( appendix [ maybe ] ) , and even it has blueshifted siiv absorption .
@lllrrllrrr@ quasar & plate - mjd - fiber & shape & & & civ & civ & civ + & & & & & ai & ai@xmath58 & ai@xmath59 + + j0028 & 0689 - 52262 - 363 & @xmath21 & @xmath60@xmath20780 & @xmath61 & 0 & 3310@xmath20830 & 1382@xmath20658 & 843@xmath20270 & 2225@xmath20940 + ... & 3586 - 55181 - 0548 & @xmath21 & @xmath62@xmath20880 & @xmath61 & 0 & 3510@xmath20120 & 2206@xmath20384 & 1381@xmath2092 & 3586@xmath20477 + ... & 4220 - 55447 - 0790 & @xmath21 & @xmath63@xmath20440 & @xmath61 & 0 & 3510@xmath20250 & 2203@xmath20776 & 1483@xmath20255 & 3686@xmath201029 + j0148 & 4273 - 55506 - 0334 & _ v _ & @xmath64@xmath20160 & @xmath65@xmath20120 & 2020@xmath20210 & 6090@xmath201080 & 538@xmath20159 & 664@xmath20383 & 1203@xmath20543 + j0805 & 4458 - 55536 - 0158 & _ v _ & @xmath66@xmath2070 & @xmath61@xmath2070 & 2140@xmath20120 & 3930@xmath20160 & 1040@xmath20145 & 327@xmath20145 & 1367@xmath20290 + j0828 & 3762 - 55507 - 0690 & @xmath22 & @xmath67@xmath2085 & @xmath68@xmath20210 & 1270@xmath20400 & 6130@xmath201500 & 869@xmath20200 & 296@xmath20229 & 1166@xmath20513 + j0830 & 4489 - 55545 - 0364 & @xmath21 & @xmath69@xmath2070 & @xmath61 & 0 & 3170@xmath2090 & 743@xmath2072 & 1326@xmath20102 & 2069@xmath20175 + j0941 & 3782 - 55244 - 0417 & @xmath22 & @xmath70@xmath20880 & @xmath71@xmath201720 & 6370@xmath202200 & 14780@xmath20340 & 3301@xmath20712 & 1718@xmath20793 & 5019@xmath201505 + j1019 & 4802 - 55652 - 0629 & @xmath21 & ... & ... & 800@xmath2070 & 8300@xmath20160 & 0 & 2672@xmath20100 & 2672@xmath20100 + j1034 & 0999 - 52636 - 503 & @xmath22 & @xmath72@xmath2080 & @xmath73@xmath2070 & 1730@xmath20250 & 7600@xmath2070 & 2110@xmath2066 & 758@xmath20100 & 2868@xmath20165 + ... & 4852 - 55689 - 0592 & @xmath22 & @xmath74@xmath2070 & @xmath75@xmath2070 & 2360@xmath2070 & 6840@xmath2070 & 2035@xmath207 & 81@xmath2028 & 2116@xmath2022 + j1146 & 4614 - 55604 - 0726 & _ v _ & ... & ... & 1420@xmath20880 & 7980@xmath20540 & 0 & 176@xmath20159 & 176@xmath20159 + j1147 & 0329 - 52056 - 176 & _
v_/@xmath24 & @xmath76@xmath2080 & @xmath61 & 0 & 8900@xmath20120 & 1706@xmath20352 & 1366@xmath2010 & 3071@xmath20356 + ... & 3790 - 55208 - 0290 & _ v_/@xmath24 & @xmath77@xmath2070 & @xmath61 & 0 & 10560@xmath2070 & 1694@xmath2070 & 1780@xmath20335 & 3474@xmath20406 + j1323 & 4006 - 55328 - 0170 & @xmath21 & @xmath78@xmath2070 & @xmath61 & 0 & 5600@xmath2070 & 849@xmath2026 & 1391@xmath2010 & 2240@xmath2026 + ... & 4050 - 55599 - 0624 & @xmath21 & @xmath79@xmath20160 & @xmath61 & 0 & 5520@xmath201080 & 1195@xmath20218 & 1338@xmath20605 & 2533@xmath20823 + j1439 & 4780 - 55682 - 0992 & _
v _ & @xmath80@xmath2070 & @xmath61 & 0 & 9400@xmath201800 & 306@xmath2033 & 1602@xmath20356 & 1908@xmath20389 + j1440 & 3868 - 55360 - 0886 & @xmath22 & @xmath81@xmath201070 & @xmath82@xmath2085 & 1800@xmath201270 & 12600@xmath20400 & 3062@xmath20167 & 1307@xmath20509 & 4369@xmath20676 + j1628 & 0625 - 52145 - 121 & @xmath22 & @xmath83@xmath20240 & @xmath84@xmath20100 & 5610@xmath20590 & 12400@xmath201000 & 2034@xmath20292 & 934@xmath20215 & 2967@xmath20506 + ... & 6322 - 56190 - 0599 & @xmath22 & @xmath85@xmath2080 & @xmath86@xmath202000 & 2710@xmath20120 & 11100@xmath201200 & 2058@xmath20141 & 1167@xmath20282 & 3225@xmath20422 + j1709 & 5014 - 55717 - 0763 & @xmath21 & ... & ... & 700@xmath2070 & 4430@xmath2070 & 0 & 1804@xmath2048 & 1804@xmath2048 + j1724 & 5002 - 55710 - 0600 & @xmath21 & ... & ... & 0@xmath20640 & 9350@xmath20540 & 0 & 3812@xmath20618 & 3812@xmath20618 + j2157 & 4197 - 55479 - 0030 & _ v _ & @xmath28@xmath2085 & @xmath61 & 0 & 9050@xmath2070 & 1329@xmath2034 & 2627@xmath20187 & 3956@xmath20221 + + j0050 & 4306 - 55584 - 0846 & @xmath21 & ... & ... & 0@xmath20200 & 1780@xmath2070 &
0 & 1397@xmath2010 & 1397@xmath2010 + j1239 & 0521 - 52326 - 221 & _ v _ & @xmath87@xmath20160 & 0@xmath20200 & 2360@xmath20250 & 11670@xmath20590 & 1226@xmath20431 & 1152@xmath20155 & 2378@xmath20276 + ... & 4754 - 55649 - 0294 & _ v _ & @xmath88@xmath2070 & 0@xmath20200 & 1770@xmath20160 & 11640@xmath20160 & 1139@xmath2055 & 1766@xmath2035 & 2905@xmath2021 + j1342 & 3987 - 55590 - 0124 & @xmath22 & @xmath89@xmath201170 & @xmath90@xmath20120 & 920@xmath2070 & 6990@xmath2080 & 1385@xmath20129 & 793@xmath2014 & 2178@xmath20117 + j1633 & 4061 - 55362 - 0256 & _ v _ & @xmath68@xmath2070 & @xmath61 & 0 & 7300@xmath202800 & 930@xmath2014 & 601@xmath20121 & 1531@xmath20135 + j1704 & 1687 - 53260 - 150 & @xmath21 & @xmath91@xmath2070 & @xmath61 & 0 & 2410@xmath20240 & 4128@xmath20142 & 560@xmath2066 & 4688@xmath20209 + ... & 4177 - 55688 - 0220 & @xmath21 & @xmath92@xmath20350 & @xmath61 & 0 & 2550@xmath20120 & 4428@xmath20288 & 447@xmath2092 & 4875@xmath20380 + j2133 & 4088 - 55451 - 0426 & _ v _ & @xmath93@xmath20100 & @xmath61 & 0 & 5200@xmath2080 & 891@xmath20260 & 854@xmath20190 & 1809@xmath20495 + + the plate - mjd - fiber column gives the sdss / boss spectroscopic plate number , mjd of observation , and fiber number .
sdss and boss observations can be distinguished because sdss uses a three - digit fiber number ( 001 to 600 ) while boss uses a four - digit fiber number ( 0001 to 1000 ) .
trough shapes are discussed in [ uvw ] .
the velocity ranges of absorption in civ are measured relative to the redshift given in table 1 , but do not include the systematic uncertainties on those systemic redshifts .
entries with a 0 in both the blueshifted and redshifted columns represent absorption troughs which are continuous across the systemic redshift .
absorption indices are discussed in [ ai ] ; the values here were calculated using @xmath94 km s@xmath1 .
[ tab2 ] the balnicity index ( bi ) , which is essentially a modified equivalent width , was introduced by weymann et al .
( 1991 ) as a measure of the strength of a blueshifted absorption trough in a quasar spectrum . the absorption index ( ai ; hall et al . 2002 ; trump et al .
2006 ) is a modification of the bi which extends to narrower troughs and to blueshifted velocities closer to systemic .
we use the ai to measure the strength of blueshifted absorption , and we define ai@xmath58 as the corresponding measurement for redshifted absorption and ai@xmath59 as the corresponding measurement for absorption at all velocities . with @xmath95 being the normalized rest - frame . ]
spectrum as a function of velocity from the systemic redshift , and with negative velocities representing blueshifted absorption , we have : @xmath96 ~c ' ~dv \\ { \rm ai}^+(v_{min } ) = \int_{0}^{c } [ 1-f(v)/f_{int } ] ~c ' ~dv \\ { \rm ai}^{\rm tot}(v_{min } ) = \int_{-c}^{c } [ 1-f(v)/f_{int } ] ~c ' ~dv \ ] ] where @xmath97 in contiguous intervals of width @xmath98 or greater within the integration limits wherein the quantity in brackets is everywhere positive ; otherwise , @xmath99 .
although we will write ai instead of ai@xmath100 in general , in the definition we use ai@xmath100 to emphasize that ai values depend on the @xmath98 used . in hall et al .
( 2002 ) and dr9q , @xmath94 km s@xmath1 and @xmath101 have been used .
for ease of comparison with dr9q , we adopt those values also .
km s@xmath1 and used @xmath101 to calculate @xmath102 but @xmath103 to calculate ai values , so that the ai would be a true equivalent width measured in km s@xmath1 .
] we have given the limits to the integrals as the speed of light to emphasize that the ideal ai would measure the absorption in a given transition at all outflow velocities , even though in practice limits on the integrals will be set by confusion with absorption from other transitions .
we present civ absorption indices for our quasars in table 2 .
ai values were calculated automatically in the dr9q only for five confirmed and three candidate objects , due to low signal - to - noise ratios or poor continuum fits in other objects .
we therefore calculated our own ai , ai@xmath58 and ai@xmath59 values for all our quasars to ensure internally consistent measurements .
we first performed weighted - average smoothing on each spectrum , using a smoothing box size of the minimum odd number of pixels required to yield an average signal - to - noise ratio @xmath10410 in the 13251625 wavelength range ; typically , this was 7 pixels ( each boss pixel spans 69 km s@xmath1 in the observed frame ) .
we then normalized each smoothed spectrum by a fifth - order polynomial continuum fit to wavelengths usually free from strong emission or absorption ( see appendix a of hall et al . 2002 ) , adjusted as necessary for each object individually .
the ai values calculated from these normalized continua are expected to be underestimated relative to ai values from dr9q because dr9q accounts for emission lines in the normalizing continuum and we do not .
we find this to be the case for seven out of eight objects where a comparison is possible .
it is of course preferable to account for emission lines , but to do so accurately for these unusual objects will require study of their continuum and emission - line properties which is beyond the scope of this work .
we calculated ai@xmath59 between @xmath515000 km s@xmath1 and 15000 km s@xmath1 in general , but adjusted the limits for individual quasars as needed to avoid obvious intervening narrow absorption and to include all intrinsic absorption from the transition under consideration .
we do not use the traditional blueshifted limit of @xmath525000 km s@xmath1 to avoid counting redshifted siiv as blueshifted civ .
we estimated the uncertainties for the velocity limits and ai , ai@xmath58 and ai@xmath59 values by repeating our measurements on spectra normalized using a different method that of filiz ak et al .
( 2012 ) and then calculating the dispersion between those two measurements of each quantity .
the average uncertainty on the ai values is @xmath920% . for the velocity uncertainties , we added in quadrature a base uncertainty of 69 km s@xmath1 , equivalent to one spectral pixel .
the average velocity uncertainty is @xmath9450 km s@xmath1 ; note that this is in addition to the systematic uncertainty on the quasar redshift ( the velocity zeropoint ) . in the left panel of figure [ f_rsai ]
we compare the cumulative distribution functions of non - zero civ ai values for visually identified bal quasars from dr9q ( solid line ) with those of redshifted - trough bal quasars ai values ( short - dashed line ) , ai@xmath58 values ( dotted line ) , and ai@xmath59 values ( long - dashed line ) .
although the distribution of ai@xmath59 in our small sample of bal quasars with redshifted troughs appears to lack the tail to high values seen in the ai distribution of all dr9q visual bal quasars , the difference is not statistically significant : there is a 52% chance of identical intrinsic distributions .
( we calculate all probabilities here using a two - sample kuiper s test ( press et al .
2007 ) , a variant of the k - s test which is equally sensitive to deviations at all values of the distribution . )
the differences between the ai distribution in all visual bal quasars and the ai or ai@xmath58 distributions of bal quasars with redshifted troughs are only marginally statistically significant , due to our small sample size ( 3.1% or 0.12% chances of identical intrinsic distributions , respectively , equivalent to @xmath105 or @xmath106 ) .
the ai and ai@xmath58 distributions of bal quasars with redshifted troughs are themselves statistically indistinguishable ( 67% chance of identical intrinsic distributions ) .
note , however , that despite commonly having low - ionization absorption ( [ lobals ] ) , redshifted - trough bal quasars are completely inconsistent with being a random sample of lobal quasars , as traced by dr9q quasars with aliii absorption ( dot - dashed line ) .
lobal troughs are preferentially seen in standard bal quasars when the civ absorption is strong ( see , e.g. , figures 5 and 6 of allen et al .
2011 ) , but the civ ai distribution of redshifted - trough bal quasars lacks the tail to large ai values seen in standard bal quasars , let alone in lobal quasars . in the right panel of figure [ f_rsai ]
we plot the redshifted absorption index ai@xmath58 versus the ( blueshifted ) absorption index ai for confirmed and candidate quasars with redshifted absorption .
there is no overall correlation between ai@xmath58 and ai .
there is a tendency for differently shaped troughs to inhabit different regions of the diagram . in @xmath22-shaped troughs ,
the redshifted absorption is always weaker than the blueshifted absorption .
the opposite is true in most _ v_-shaped and especially in most @xmath21-shaped troughs .
selection effects may help explain the latter : @xmath21-shaped and _
v_-shaped troughs with only a small redshifted component might have been missed in the assembly of our sample , given redshift uncertainties and the difficulty of detecting weak absorption atop broad emission lines .
however , selection effects can not explain the lack of @xmath22-shaped troughs with stronger redshifted than blueshifted absorption .
more objects and better systemic redshift determinations will be useful in interpreting this diagram . if the separation of different trough shapes in this diagram persists with more objects added and after more study of selection effects , that may suggest that different mechanisms are responsible for different trough shapes . to summarize , in our objects the ai and ai@xmath58 values are uncorrelated overall , but the ai@xmath59 values are a statistical match to the ai distribution for standard bal quasars .
this result is somewhat surprising , given that our objects are predominantly lobals ( see [ lobals ] ) and that lobals have stronger average civ absorption than hibals .
this contradiction indicates that there is likely a different origin for at least the low - ionization absorption in our objects as compared to standard bal quasars . for an approximate estimate of how common bal quasars with redshifted troughs are , we begin with the sdss dr5 bal catalog of gibson et al .
( 2009 ) and the sdss dr6 bal catalog of allen et al .
the strength of a bal trough can be measured by its balnicity index ( [ ai ] ) .
gibson et al .
( 2009 ) quote values for both the traditional balnicity index ( bi ) , measured between blueshifts of @xmath53000 km s@xmath1 and @xmath525000 km s@xmath1 , and a modified balnicity index bi@xmath107 measured between blueshifts of 0 km s@xmath1 and @xmath525000 km s@xmath1 .
allen et al .
( 2011 ) only quote bi values . we have found four sdss quasars with redshifted civ absorption ( figure [ f_cands2 ] ) and two sdss quasars with redshifted mgii absorption ( figure [ fj11 ] ) which we can compare to sdss bal quasar samples . of 4664 quasars with civ bi@xmath107@xmath60 in gibson et al .
( 2009 ) , 3 have redshifted absorption ( j0028 does not make that cut ) . of 3880 quasars with civ bi@xmath60 in gibson
( 2009 ) , only j1628 has redshifted absorption .
of 3317 quasars with civ bi@xmath60 in allen et al .
( 2011 ) , only j1034 has redshifted absorption .
thus , in the sdss at most about 1 in 1600 bal quasars with blueshifted civ absorption also has redshifted civ absorption ( 0.064% , with a 90% confidence range of 0.028%@xmath50.124% , based on the bi@xmath107 statistics and determined following gehrels 1986 ) .
the quasars we have discovered with redshifted civ absorption in the boss dr9q are only marginallly consistent with that result . of our seventeen quasars with redshifted absorption , two
were only included in the boss as ancillary targets because of that absorption ( j1034 and j1628 ) , one was targeted as a galaxy ( j0941 ) , and three others have only redshifted absorption .
the remaining eleven have a parent sample of 7228 visually identified bal quasars at @xmath56 ( where boss spectra cover siiv to civ ) in the dr9q ( pris et al .
thus , approximately 1 in 660 dr9q quasars with blueshifted absorption also has redshifted absorption ( 0.15% , with 90% confidence range 0.11%@xmath50.21% ) .
note that the overall incidence of quasars with redshifted bal troughs among boss dr9q quasar targets at @xmath56 is approximately 1 in 5000 ( 14/69674 ) . in summary ,
the frequency of bal quasars with both blueshifted and redshifted civ absorption differs by a factor of two between the sdss and the boss surveys , although the 90% confidence ranges overlap due to small - number statistics .
if the difference in frequency is confirmed , it might mean that the incidence of redshifted absorption troughs decreases with increasing quasar uv luminosity , as the mean luminosity of boss quasars is lower than that of sdss quasars by about a factor of four .
however , a k - s test between the @xmath18 distributions of our sample and of other boss bal quasars reveals no statistically significant discrepancy ( 22.5% chance of identical intrinsic distributions ) .
the statistics are somewhat different for redshifted mgii absorption .
of 332 quasars with mgii bi@xmath107@xmath60 in gibson et al .
( 2009 ) , two have redshifted absorption ( or three , if the candidate j1316 is confirmed ; see appendix [ maybe ] ) . of 244 quasars with mgii bi@xmath60 in gibson
( 2009 ) , two have redshifted absorption . of 222 quasars with mgii bi@xmath60 in allen et al .
( 2011 ) , two have redshifted absorption . thus , approximately 1 in 130 bal quasars with blueshifted mgii absorption also has redshifted mgii absorption ( 0.60% , with a 90% confidence range of 0.19% to 1.35% , based on the bi@xmath107 statistics of the two confirmed objects ) .
we are unwilling to place much emphasis on this higher abundance of redshifted mgii troughs based on two objects , especially since the 90% confidence ranges overlap , but it is clearly worthy of further investigation . in fig .
[ f_j1019zoom ] we show a detailed view of the spectrum of j1019 ( [ cases ] ) .
j1019 may have absorption in ciii * , a set of transitions arising from levels with an excitation potential ( ep ) 6.5 ev above the ground state .
ciii * can be a useful density diagnostic due to the range of critical densities of its different levels ( section 4 of gabel et al .
j1019 also has absorption in feiiiuv34 @xmath4@xmath4@xmath41895,1914,1926 ( ep 3.73ev , @xmath108@xmath31 ) , feiiiuv48 @xmath4@xmath4@xmath42062,2068,2079 ( ep 5.08ev , @xmath109@xmath31 ) and feiiiuv50 ( ep 7.86ev , @xmath110@xmath31 ) , a multiplet spanning 19871996 . the feiii troughs appear to extend to redshifted velocities as large as seen in civ and larger than seen in aliii .
the short - wavelength end of the feiii uv34 trough appears to be filled in by emission from a combination of ciii ] , siiii ] , and feiii uv34 itself .
strong absorption from feiii without strong accompanying feii absorption is rare in part because it occurs only for absorbers with column density too low to form a hydrogen ionization front ( outside of which feii will form ) and with electron densities sufficiently high ( @xmath111@xmath31 ) to increase the recombination of feiv to feiii ( rogerson et al . 2011 , section 4 ) .
such densities have also been seen in outflowing gas in ngc 4051 ( king , miller & raymond 2012 ) .
j1019 and possibly j1316 ( appendix [ maybe ] ) represent the first reported detections of feiii uv50 absorption in quasars .
its rarity is explained by the relative populations of the ground states of the uv34 , uv48 and uv50 multiplets at @xmath112@xmath31 and @xmath113 k ; namely , 1:0.16:0.016 ( bautista & pradhan 1998 ) .
thus , the density in the medium causing the absorption at redshifted velocities of thousands of km s@xmath1 in j1019 is higher than the gas - phase densities found everywhere else in a galaxy , except in accretion discs ( e.g. , fig . 1 of hopkins & quataert 2010 ) .
even protostellar cores in molecular clouds only reach gas - phase densities of @xmath114@xmath31 ( di francesco et al .
if this dense gas is in fact accreting onto the quasar , it must have been compressed at some stage in the accretion process .
this lower limit on the density @xmath115 enables us to set an upper limit on the distance @xmath116 of the absorbing gas from the ionizing continuum source by rearranging the definition of the ionization parameter @xmath117 : @xmath118 where @xmath119 is the number of hydrogen - ionizing photons emitted per second .
j1019 is almost exactly an order of magnitude less luminous than the quasar studied by moe et al .
( 2009 ) , so its value of @xmath119 should also be an order of magnitude less : @xmath120 . with feiii most abundant at @xmath121 ( de kool et al .
2002 ) , , but the relative strengths of absorption from feiii and from feii in the object can not be reproduced at such parameters . to see feiii uv50 at @xmath122 over the width of the trough in j1019 requires @xmath123@xmath31 , which is sufficient to generate a hydrogen ionization front and accompanying feii absorption unless @xmath124 ( figure 6 of de kool et al .
we adopt a slightly lower value of @xmath125 to place a conservative upper limit on the distance of the absorbing gas . ]
we conservatively estimate that the absorbing gas in this object is located at @xmath126 pc , with an uncertainty in this upper limit of about a factor of two .
high resolution spectroscopy of j1019 ( and possibly j1316 ) will be extremely useful for measuring or setting limits on the column densities in the feiii lines to obtain more accurate value of @xmath115 and @xmath117 and thus of @xmath116 .
having established the reality and incidence of redshifted absorption in bal quasars , we now consider possible explanations for redshifted absorption , more than one of which may be at work in our sample .
we consider gravitational redshifts , infall , rotating outflows , binary quasars , and the relativistic doppler shift .
where relevant , as a fiducial example we consider a 10@xmath127 @xmath128 black hole surrounded by a standard shakura & sunyaev ( 1973 , hereafter ss73 ) disc with @xmath129 accreting at a fraction @xmath130 of the eddington rate .
the disc s surface temperature profile is approximately @xmath131 ( ss73 eq .
3.5 , but see blackburne et al .
2011 ) , so the physical diameter of the region of an ss73 accretion disc which emits significant continuum at wavelength @xmath4 can be written as @xmath132 for a black hole of mass @xmath133 .
the crossing time of the 1550 continuum region of this fiducial bh and accretion disc for gas moving at @xmath134 percent of the speed of light is 210/@xmath134 days .
for example , gas moving at 300 km s@xmath1 across the line of sight , a plausible speed in the core of a massive galaxy , would have a crossing time of 5.75 years .
dorodnitsyn ( 2009 , 2010 ) have shown that various wind velocity laws plus gravitational redshifting can lead to a variety of redshifted and blueshifted troughs being produced by spherically symmetric winds , provided they are launched within @xmath135 of the black hole . however , we can rule out gravitational redshifts as the sole explanation for redshifted absorption in our objects .
the gravitational redshift of a transition seen in absorption or emission at radius @xmath136 around a schwarzschild black hole is given by @xmath137 a purely gravitational redshift explanation for our objects would require absorption from gas at @xmath138 to explain redshifted velocities between 3000 and 14,700 km s@xmath1 .
such radii are comparable to the ultraviolet continuum emission region sizes of quasars .
unshielded gas at such small radii would be too highly ionized by that continuum ( or too dense , in the accretion disc itself ) for ions such as civ to exist ( e.g. , murray & chiang 1998 ) .
however , gravitational redshifting may contribute to the observed redshifted velocities of absorption from a rotating , shielded wind launched from within or just outside the continuum emission region ; see
[ rotvel ] .
simple infall of gas on large scales is a reasonable explanation for absorption at redshifted velocities characteristic of dark matter halos ; i.e. up to a few 100 km s@xmath1 in galaxies - shaped troughs ) .
some snrs are likely to be created at relatively small distances from quasars as a result of star formation in the outer regions of quasar accretion discs ( collin & zahn 1999 ) , which would increase the covering factor of snrs around quasars .
however , the density in such discs is likely to be high enough to greatly suppress the expansion of such snrs ( 2.1 of thompson , quataert & murray 2005 ) .
snrs can have expansion velocities of 10,000 km s@xmath1 or more , but the gas moving at such velocities is typically too highly ionized to absorb at uv wavelengths ( hayato et al .
the swept - up gas which can show absorption in ions such as ovi typically does so over a velocity range of only hundreds of km s@xmath1 ( jenkins , wallerstein & silk 1976 ) .
furthermore , absorption from cool supernova ejecta free - expanding at thousands of km s@xmath1 in the interior of a snr is seen only in elements such as ca , si and fe in ionization stages i to iv ( winkler et al .
2005 ) , which can produce spectra very different from those of our objects or of standard bal quasars ( wu et al .
1997 ; fesen et al .
thus , a silhouetted snr explanation seems unlikely for these objects , although it should be kept in mind . ] and up to a few 1000 km s@xmath1 for galaxies in galaxy clusters .
however , the _ widths _ of the absorption troughs in our objects rule out direct infall explanations except in the narrowest cases ( @xmath22000 km s@xmath1 or so ) .
absorption from an infalling galaxy ( or the quasar host galaxy ) could conceivably produce such linewidths if its gas is affected by a merger or a starburst - driven wind .
iono et al .
( 2007 ) find gas motion spanning a range of 1200 km s@xmath1 in the central kpc of the merging binary agn ngc 6240 .
steidel et al .
( 2010 ) find that oi@xmath41302 absorption in lyman - break galaxies can span a velocity range of at least 700 km s@xmath1 infall to @xmath51000 km s@xmath1 outflow ( their figures 6 , 9 and 10 ) . in our sample , only j0050 ( see appendix [ maybe ] ) , j1125 , and perhaps j1128 have such narrow troughs .
absorption from a galaxy is a possible explanation for j0050 .
however , the feii absorption blanketing the spectrum of j1125 and j1128 is much stronger than is ever seen in star - forming galaxies ( farrah et al .
2005 ) . also
, neither j1125 nor j1128 host luminous starbursts detectable in the far - ir ( farrah et al . 2012 ) . to generate absorption at infall velocities @xmath63000
km s@xmath1 toward the uv continuum sources in these quasars requires gas infall along our line of sight ( and others , if the infall is not purely radial ) down to radii small enough to generate such velocities . to a good approximation ,
the maximum velocity gain achievable through infall to radius @xmath136 is the escape velocity from that radius ( @xmath139 ) , so if redshifted absorption at velocity @xmath140 is caused by gas radially infalling towards the quasar , then @xmath141 .
the maximum observed @xmath142 ( @xmath143 km s@xmath1 ) in our sample requires infall to @xmath144 .
this calculation agrees within a factor of two with the 3-d simulations of kurosawa & proga ( 2009 ) , in which infall velocities of 4500 km s@xmath1 to 7000 km s@xmath1 can be reached at @xmath145 ( their figures 7 and 8) , and of barai , proga & nagamine ( 2012 ) , in which infall velocities of 1500 km s@xmath1 to 2000 km s@xmath1 can be reached at @xmath146 ( their figure 3 ) .
such velocities are reached in their models if the infall is weakly rotating and is viewed close to perpendicular to its rotation vector .
the corresponding 2-d simulations of proga , ostriker & kurosawa ( 2008 ) yield very similar maximum infall velocities ( proga , personal communication ) .
similarly , the 2-d simulations of li , ostriker & sunyaev ( 2012 ) show that free - falling inflows can reach @xmath147 .
however , it remains to be seen whether or not these simulations can reproduce the large optical depths observed at these redshifted velocities in our objects .
simulations of gas inflow on larger scales in galaxy nuclei have been performed by , e.g. , levine et al .
( 2008 ) , hopkins & quataert ( 2010 ) , hobbs et al .
( 2011 ) and gaspari , ruszkowski & oh ( 2013 ) .
levine et al .
( 2008 ) find that a flattened but thick circumnuclear disc ( @xmath148 ) is formed , containing large - scale structure such as spiral waves and bars .
however , the average inflow velocity is only 10% of the rotational velocity at each radius in their simulation domain ( @xmath149pc ) .
hopkins & quataert ( 2010 ) find inflow velocities of 0.010.3 times the rotational velocity in their simulations ( their 4 ) . at scales of @xmath9110pc
, they find that eccentric discs or single - armed spirals can form and drive gas inwards .
however , even in the extreme case of plunging hyperbolic orbits for such gas , the maximum inflow velocity achieved will be only @xmath150 times the keplerian velocity at the innermost radius of the flow plus the turbulent velocity with which the gas began its infall ( @xmath21000 km s@xmath1 ) .
hobbs et al .
( 2011 ) argue that supersonic turbulence can lead to the formation of dense filaments which fall ballistically towards the black hole .
their model may be a promising way to explain our objects , but their study is confined to scales @xmath61pc , making direct comparison difficult .
gaspari et al .
( 2013 ) make a similar argument using simulations that extend down to sub - parsec scales , but do not report gas infall velocities .
the absorbing gas in our objects has avoided overionization and is experiencing a net outward force from scattered uv radiation , so why is it infalling ? one possible answer
is that the absorbing gas is unusually dense , which might also help explain the high fraction of lobals in our sample ( [ lobals ] ) .
denser gas will have a lower ionization parameter , all else being equal , and is plausibly more likely to form optically thick structures .
such structures are less likely to be driven into outflow by a quasar s radiation pressure ( in the absence of velocity gradients , only the inner face of an opticallly thick gas clump will scatter uv lines to provide an outward force , but the entire clump mass will feel the inward force of gravity ) .
thus , dense gas clumps formed by swept - up lower - density gas may end up in infall ( nayakshin & zubovas 2012 ) . in this scenario ,
more variability should occur at larger redshifted velocities than at smaller ones , because gas with larger infall velocities is likely to have larger transverse velocities as well .
( note that standard bal quasars exhibit greater trough variability at higher trough outflow velocities , which is likely due to higher transverse velocities in such troughs ; see capellupo et al .
2011 and filiz ak et al . 2012 . ) for troughs broader than can reasonably be produced outside the nucleus of a galaxy ( @xmath61000 km s@xmath1 or so ) , the longer the troughs are seen _ not _ to vary , the more ordered the infall has to be , and the less likely infall becomes as a hypothesis .
a key test of the infall hypothesis will be to compute absorption profiles for the simulations to determine if they can match the apparent optical depths of the ionization stages seen in these objects ( e.g. , sim et al .
2012 ) . in observed cases where the absorption profiles are smooth over a large velocity range , the infall motions must be ordered so that our line of sight intercepts most of the infall trajectory . in that case
the gas must have lost considerable angular momentum and have avoided shocking to unobservably high temperatures and ionization stages during the accretion process .
gas with such properties will be rare , but redshifted
bal troughs are also rare .
therefore , an explanation postulating infall velocities of @xmath63000 km s@xmath1 in some of these objects remains potentially viable .
however , in this scenario , infalling redshifted troughs and outflowing blueshifted troughs should have nothing to do with each other , so there is no reason to expect their absorption index distributions to be statistically indistinguishable , as observed ( [ ai ] ) , nor to expect the distribution of their summed absorption indices ( ai@xmath59 ) to be a statistical match to the ai distribution for standard bal quasars .
furthermore , we can argue against infalling gas in some objects following 4.2 of martin & bouch ( 2009 ) , who point out that the covering factor of optically thick structures in a wind changes in a predictable way with radius . in bal quasars where civ and siiv are of comparable depth , the absorbing gas is almost certainly optically thick , with the depth of the troughs determined by the covering factor of the absorbing structures ( crenshaw et al .
2003 ; arav et al .
2008 ) . regardless of whether these structures are outflowing or infalling ,
if they are located in a gaseous halo with pressure profile @xmath151 then their volume will adjust at different radii so as to maintain pressure equilibrium ( @xmath152 ) .
we can thus write @xmath153 where we assume @xmath154 , appropriate for a monatomic , ideal gas undergoing adiabatic compression . for clouds that keep the same shape and orientation as their volume changes , the covering factor @xmath155
is given by the area of the clouds ( @xmath156 ) relative to the area of the sphere at that radius : @xmath157 .
for an isothermal sphere @xmath158 , yielding @xmath159 .
for the covering factor to decrease with decreasing radius would require an implausibly steep pressure profile with @xmath160 . for radially infalling gas , we have @xmath161 , which yields @xmath162 for the isothermal case and @xmath163 in general . _ thus , unless @xmath160 , troughs from infalling gas undergoing adiabatic compression should have covering factors which increase with redshifted velocity _ until the gas leaves our line of sight or the ion in question vanishes due to increasing ionization ( from shocks , photoionization of lower - density fragments created from instabilities in the main clump , or both ) .
conceptually , gas clumps free - falling toward a quasar do not shrink as fast as the surface area around the quasar shrinks ( unless the pressure increases implausibly fast with decreasing radius ) , leading to an increasing covering factor with increasing redshifted velocity .
this effect can be seen in the simulated clumpy accretion flows of barai et al .
( 2012 , their figure 2 or figure 7 ) .
note that the ionization parameter @xmath117 depends on the ionizing luminosity @xmath164 , the distance @xmath136 , and the gas density @xmath165 : @xmath166 .
thus , under our assumptions the ionization parameter will increase with decreasing radius for @xmath167 , following @xmath168 in the isothermal case . with @xmath161
, we have @xmath169 in the isothermal case and @xmath170 in general .
there is no clear sign in our objects for an overall trend of increasing ionization ( as traced by the aliii / civ ratio ) with increasing redshifted velocity .
it can be is difficult to determine the extent of redshifted troughs of aliii due to confusion with the broad emission complex at 1900 . nonetheless , in some objects ( e.g. , j0830 and probably j1709 ) the troughs of aliii , nv and civ all appear to have the same redshifted extent ,
while in others ( e.g. , j1019 and j1724 ) , aliii is limited to less redshifted velocities . in summary :
for the case of adiabatic compression , infalling gas is ruled out as an explanation for redshifted troughs with _
v_-shaped profiles , but infall remains a possibility for those @xmath21 and @xmath22-shaped profiles wherein the covering factor initially increases with increasing redshifted velocity .
such an increase is found in seven out of the eleven objects with @xmath21 or @xmath22-shaped profiles in our sample ( j0828 , j0941 , j1019 , j1146 , j1440 , j1709 and j1724 ) and in the candidates j0050 and j1342 , and may be found in other objects ( e.g. , j0830 ) depending on the exact values of their systemic redshifts .
the above conclusion does depend on our assumption of constant cloud shape .
abandoning that assumption , reconciling the case of adiabatic compression with our observations would require @xmath171 between 2 and @xmath172 ( to match the behaviour of @xmath117 with velocity ) and clouds which radially elongate in those respective cases by a factor of @xmath610@xmath173=16 to @xmath610@xmath174=5 per factor of ten decrease in radius ( to match the behaviour of @xmath175 with velocity ) .
detailed simulations involving heating and cooling may also show infall to still be a possible explanation for our objects . as mentioned in
[ intro ] , ganguly et al .
( 2001 ) and hall et al .
( 2002 ) have noted that redshifted absorption can occur when an extended emission source is seen through a rotating outflow .
specifically , the outflow must have a rotational velocity which dominates the component along our line of sight of its poloidal ( nonazimuthal ) velocity . to yield redshifted velocities of @xmath1041000 km s@xmath1
, the inner radius of the outflow should also be no more than @xmath910 times larger than the radial extent of the continuum source @xmath176 ( hall et al .
2002 , 6.5.2 and thereafter ) . in a rotating disc - wind model , discs viewed at high inclination ( close to edge - on )
will always be more likely to produce redshifted absorption .
unobscured high inclination sightlines will be rare in models where the obscuring torus ( or equivalent ) is coplanar with the disc , but are possible with tilted but not twisted discs as obscurers ( e.g. , lawrence & elvis 2010 ) .
the rotating wind scenario is consistent with the result that the ai@xmath59 values for these objects have the same statistical distribution as in standard bal quasars .
conceptually , one can imagine gas which is distributed over a range of blueshifted velocities in standard bal quasars being seen near its launch radius , where it is rotating over a range of blue- and red - shifted velocities .
the high fraction of lobals in our sample ( [ lobals ] ) might be understandable in this scenario if the absorbing gas is of relatively high density , resulting in outflows with lower terminal velocities .
that would enable redshifted velocities due to rotation to be dominant over a larger range of outflow distances in these objects than in standard bal quasars . to estimate the relevant properties of a rotating disc - wind
, we consider our fiducial 10@xmath127 @xmath128 black hole and ss73 disc with @xmath177 cm = 92@xmath178 .
we adopt an inner wind launching radius of @xmath179 cm = 1255@xmath178 ( elvis 2000 ; murray & chiang 1998 ) , at which @xmath180 km s@xmath1 . for a wind terminal velocity
greater than 8500 km s@xmath1 ( the escape velocity from @xmath181 ) , we find that redshifted velocities of up to 600 km s@xmath1 in civ can be reached in either the elvis or murray et al .
model ( assumed launching angles of 27@xmath182 and 6@xmath182 , respectively ) .
larger redshifted velocities can be produced if the inner wind launching radius is closer to the continuum emission region . that could arise
if the continuum region is larger than predicted by ss73 , which observationally does seem to be the case : quasar half - light radii at rest - frame 1736 are larger than predicted by a factor of @xmath95 ( jimnez - vicente et al .
2012 ) to @xmath915 ( blackburne et al .
2011 ) , on average . increasing the continuum source @xmath176 by a factor of @xmath95
makes redshifted velocities of up to 2150 km s@xmath1 possible .
a factor of @xmath915 increase , so that @xmath183 , makes a redshifted velocity of up to @xmath184 km s@xmath1 possible .
such a velocity is within a factor of two or three of the largest redshifted velocities seen in our sample . equating the highest redshifted velocities in our sample with @xmath185 implies a wind launch radius as small as @xmath186 in this model .
such an @xmath187 would be within the observationally determined uv continuum emission region .
the simulations of proga & kallman ( 2004 ) indicate that winds can be launched from radii as small as @xmath188 , but how much civ absorption is likely to arise at those radii is an open question .
the redshifted velocities from a rotating disc wind might also increase with @xmath189 .
for an ss73 disc at fixed eddington ratio , @xmath190 ( equation [ e_rcon ] ) ; if @xmath181 scales the same way then @xmath191 , increasing the maximum redshifted velocities by a factor of 2.15 for the most massive black holes with @xmath192 @xmath128 .
however , if @xmath181 scales linearly with @xmath189 then the maximum redshifted velocity is independent of @xmath189 .
two properties of our objects absorption profiles are relevant to the rotating wind model .
all our objects except j0028 ( and the candidates j1342 and j1704 ) have civ absorption which extends farther in the redshifted velocity direction than in the blueshifted velocity direction .
azimuthal asymmetries in rotating winds can in principle explain this result ; such asymmetries will eventually disappear due to rotation , and objects with winds hosting such asymmetries will still over the long term have larger blueshifted than redshifted velocities . also , rotating winds launched at large radii can not have redshifted trough depths greater than 50% , as seen in j0830 , j1019 , j1709 , j1724 and probably j2157 .
because at most half of the outflow is rotating away from the absorber , only half of the continuum source will be seen through redshifted absorption .
both of these properties may be explainable with gravitational redshifting .
if the absorption arises at small enough radii , gravitational redshifting will shift the absorption profile to more redshifted velocities , making it possible for more than half of the continuum source to be seen through redshifted absorption .
the 3800 km s@xmath1 average difference between maximum absolute blueshifted and redshifted velocities would require some absorption to arise at an average @xmath193 .
however , the lack of cases of absorption extending farther in the blueshifted direction than in the redshifted direction , and the lack of any blueshifted absorption in some cases , might require a failed wind in these objects . in other words , in this model the redshifted absorption might have to arise at such small radii that the observed velocities are due solely to rotation and gravitational redshifting , with no actual outflowing velocity component to extend the profiles farther in the blueshifted direction than in the redshifted direction .
some numerical disc - wind outflow models have predicted absorption troughs spanning both blueshifted and redshifted velocities , at least for some parameter choices . for examples ,
see figures 7 and 9 of knigge , woods & drew ( 1995 ) , figure 1 of proga et al .
( 2002 ) , figure 1 of proga ( 2003 ) , and figure 3 of proga & kallman ( 2004).2000 km s@xmath1 around a systemic redshift have been seen in nova - like variables in our galaxy ( e.g. , figures 1c and 2c of hartley et al . 2002 ) .
those authors attributed the absorption to rotationally broadened photospheric features from an accretion disc , which is not a viable explanation for quasars because the broadening will be @xmath1940.1@xmath195 in the ultraviolet continuum emitting region of quasars ( few tens of @xmath178 ) . ] furthermore , some profiles in proga et al .
( 2002 ) show distinct blueshifted and redshifted absorption troughs ( their figure 1 , panels c , d , h and i ) , similar to those seen in j1034 and j1628 .
these troughs arise at relatively large inclination angles ( @xmath196 ) when considering only absorption from wind radii @xmath210 times the continuum source size . when the same models are extended to follow the wind to larger radii ( proga 2003 ) , additional absorption at small line - of - sight velocities is seen between the blueshifted and redshifted absorption , resulting in an absorption profile spanning both blueshifted and redshifted velocities .
therefore , one possibility for producing distinct blueshifted and redshifted absorption troughs is a rotating wind more ionized at large radii than in the models of proga ( 2003 ) .
such a situation could occur if the density in the wind were to drop off more rapidly with distance from the black hole than assumed in those models ( due to more efficient acceleration , for example ) , or perhaps if the overall density in the wind were lower .
such density profiles being uncommon could help explain the rarity of bal quasars with redshifted absorption
. further work on simulated line profiles including gravitational redshifting and resolved continuum emission regions ( e.g. , proga , rodriguez - hidalgo & hamann 2012 ) will be useful in untangling how redshifted absorption troughs arise .
absorption at only systemic and redshifted velocities can be produced by an azimuthally asymmetric outflow with absorbing gas present in the receding part of the silhouetted rotating outflow but not the approaching part .
that scenario makes a specific prediction that the parcels of redshifted absorbing gas will each move to larger redshifted velocities as the flow rotates , so that the low - velocity end of the trough moves to higher velocities until the trough disappears .
any new absorption components rotating into view should appear as blueshifted absorption .
note that the component of the poloidal velocity perpendicular to our line of sight can complicate the above picture by limiting the range of line - of - sight velocities through which a gas element flows as it crosses the continuum source .
azimuthal asymmetries can also produce variability in objects in which rotating disc winds produce both blueshifted and redshifted troughs .
the variability of the absorption in j1125 ( fig .
[ fj11s ] ) is potentially consistent with the model : the blueshifted absorption weakened more than the redshifted , consistent with rotating absorbing gas moving from blueshifted to redshifted velocities .
however , the variability of the blueshifted and redshifted absorption in j1034 ( fig . [ fj1034 ] ) , j1128 ( fig . [ fj11s ] ) , and j1628 ( fig .
[ fj1628 ] ) does not follow the above pattern , arguing against the rotating disc - wind model .
in the first two cases the redshifted absorption weakened while the blueshifted did not ( although that behaviour might be explained by poloidal velocity , as mentioned above ) . in the case of j1628 ,
the redshifted absorption trough moved toward zero velocity , consistent with a decelerating inflow but not a rotating disc .
we can write the time required for gas with an azimuthal velocity equal to the keplerian velocity at its radius of @xmath197 to pass in front of the continuum - emitting region of the accretion disc ( @xmath198 ) in units of the orbital timescale at @xmath176 as @xmath199/\pi}{({r_{\rm con}/r})^{3/2}}.\ ] ] as @xmath136 increases , this ratio declines from 0.5 at @xmath200 to a minimum value of 0.412 at 1.2@xmath200 before increasing to @xmath90.5 at 2.3@xmath200 and @xmath91 at 10@xmath200 .
( note that the above is a lower limit to the true crossing time , as angular momentum conservation lowers the rotational velocity of outflowing gas far from its launch radius below the keplerian velocity at that radius .
however , such gas is unlikely to produce redshifted absorption . ) for the 1550 continuum - emitting region around our fiducial @xmath201@xmath128 black hole , @xmath202 years and @xmath203 years , where @xmath204 is the inner edge of our fiducial wind . for @xmath200 five times larger , @xmath205 years and @xmath206 years . because @xmath190 ( equation [ e_rcon ] ) and @xmath207 , @xmath208 .
assuming @xmath209 , both @xmath210 and @xmath211 are a factor of ten longer for the most massive black holes ( @xmath212@xmath128 ) .
thus , variability on a timescale dependent on the continuum source size and black hole mass is expected in this model , with the most common variability pattern being absorption appearing at blueshifted velocities and moving to redshifted velocities .
note that variability can be avoided if the winds have sufficiently small azimuthal variations in their covering factors and/or transmitted flux over more than one orbital time .
( and large azimuthal variations in @xmath213 can occur without causing variability as long as @xmath57 is always satisfied . ) such high levels of symmetry would only be possible for @xmath21-shaped and @xmath22-shaped troughs , as azimuthal asymmetry is required to explain _
v_-shaped troughs in this model .
whether such high levels of symmetry are possible is an open question , but some bal outflows do appear to have substantial azimuthal symmetry ( e.g. , gibson et al . 2010 ) . a rotating disc wind model can in principle produce the redshifted _ velocities _ observed in these objects , but such a model has difficulty explaining the observed variability in several of our objects .
gravitational redshifting may help explain the prevalence of absorption that extends farther in the redshifted velocity direction than the blueshifted , as well as the four or five cases of redshifted troughs with depths greater than 50% of the continuum , but overall , a simple rotating disc wind model does not cleanly explain all objects in our sample . determining how often , if ever , rotating outflows can exhibit the parameters required to match these redshifted troughs will be a useful goal for analytic and numerical simulations .
quasars inhabit massive dark matter halos , which are highly clustered ( e.g. , tegmark & peebles 1998 ; shen et al .
2007 ) . therefore ,
close binary quasar systems should exist , and in fact numerous examples are known ( e.g. , hennawi et al .
a binary quasar system can produce a quasar with a redshifted bal trough if the following conditions are met : + 1 ) the quasar physically farthest from us is at a redshift given by the emission lines seen in the sdss spectrum .
+ 2 ) the quasar physically closest to us : + @xmath214 emits many fewer photons along our sight - line than does the background quasar ( so that its broad emission lines are not obvious in the combined spectrum ) , + @xmath214 produces a bal outflow , + @xmath214 and is oriented such that the background quasar backlights at least part of the outflow with a relative radial velocity vector that produces redshifted absorption . ) . ]
this scenario is viable in cases where the quasars are far enough apart ( @xmath215 pc or so ) that they do not dramatically disrupt each other s accretion discs , fueling , or bal outflows .
a large projected area of the flow on the sky increases the chance of backlighting , and some bal outflows are known to extend to kpc scales ( e.g. , edmonds et al .
2011 , and references therein ) .
different parts of the outflow along the same sightline could have different redshifts or blueshifts .
the binary scenario may be favored for _
v_-shaped troughs , which have the appearance a standard bal outflow might have if it were pointed away from the observer and then backlit .
we inspected available images of our objects for evidence supporting a binary quasar scenario for their spectra .
the sdss imaging pipeline classifies all objects as unresolved except for j0941 , which is morphologically extended ( [ cases]).62 away which was not selected as a quasar candidate in the sdss , and therefore is likely to be a late - type star . ]
near - ir imaging from ukidss ( warren et al . 2007 ) confirms that j0941 is not a pure point source , but yields no evidence for binarity in j0805 , j1019 , j1439 or j2157 .
also , deep coadds of sdss imaging data in stripe 82 ( annis et al . 2011 ) show no evidence for binarity in j0028 or j2157
. however , the angular separation of any putative binary quasar could be much smaller than the value needed to make sdss or ukidss images elongated , and the foreground quasar could also be obscured . to confirm or reject the binary scenario for
individual objects would require high - resolution imaging , multiwavelength photometry , or near - ir spectroscopy to search for narrow line emission at two redshifts .
note that there is no evidence for two narrow - line redshifts in our two low - redshift quasars with redshifted mgii absorption ( figure [ fj11 ] ) .
the conditions outlined in this section are very similar to those proposed by civano et al .
( 2010 ) as one scenario to explain the @xmath216 agn cxoc j100043.1@xmath7020637 ( hereafter cxoc j1000 ) .
that object hosts two compact optical sources in the same galaxy and has a broad - line redshift which is higher by @xmath91200 km s@xmath1 than the narrow - line redshift .
it also exhibits a strong x - ray iron emission line accompanied by redshifted iron absorption .
in contrast to our objects , the absorption in cxoc j1000 is deepest at the longest wavelengths ( greatest redshifted velocities ) .
civano et al .
( 2010 , 2012 ) suggest that the foreground object in that system could be a heavily obscured type 2 ( narrow - line ) agn with an outflowing , high - velocity , highly ionized wind backlit by the x - ray continuum of a background type 1 ( broad - line ) agn .
the velocity vector of the backlit portion of the wind must be oriented so as to produce a redshift .
the large velocity difference between the redshfts of the two agn could arise from a gravitational interaction .
another effect may be at work in this model , for both cxoc j1000 and any binary quasar in our sample .
if the wind from the foreground object has a sufficiently large velocity , the relativistic doppler shift can produce significantly redshifted absorption even when the backlit wind is directed near the plane of the sky .
we discuss the relativistic doppler shift in [ doppler ] .
we can estimate whether or not the abundance of bal quasars with redshifted troughs is consistent with a scenario explaining them as silhouetted outflows in binary quasar systems .
we start with 69674 @xmath56 boss quasars examined for redshifted troughs . at @xmath56 ,
5 kpc is a separation small enough for the system to appear unresolved in sdss imaging .
approximately [email protected]% of luminous sdss quasars have binary companions at projected radii between @xmath95 and @xmath9100 kpc ( myers et al .
2007 , 2008 ; hennawi et al . 2010 ) . the binary fraction at projected
radii @xmath305 kpc is likely to be a factor of @xmath910 lower , based on how long merging or interacting galaxies spend at a given projected separation ( liu et al .
thus , we expect only @xmath93 binary quasars at @xmath305 kpc separation in the 69674 boss dr9q quasars . even if we make the extreme assumption that all quasars in such close binaries host bal outflows with covering factor @xmath175=100% out to a radius of 5 kpc , only half of such binaries will have an optically faint or obscured quasar as the foreground object and an optically brighter quasar as the background object , as required in our scenario .
thus , while a small number of silhouetted outflows may be present in our sample , to explain our entire sample with them would require binary quasars at projected radii @xmath305 kpc to be a factor of @xmath910 more abundant than current studies indicate .
furthermore , their outflows would have to have a geometry and distribution of covering factors able to explain why both blueshifted and redshifted absorption are seen in 10 out of 13 objects targeted by the boss as @xmath217 quasar candidates .
that fraction is much larger than expected if blueshifted and redshifted absorption are uncorrelated .
a final possibility for producing redshifted absorption is the relativistic doppler shift ( einstein 1905 ) . in its own rest frame
, an ion absorbs at the rest wavelength of a given transition . in a frame
in which the ion is moving ( in any direction ) then in addition to the normal doppler shift time dilation means that the ion will absorb photons of a lower frequency ( longer wavelength ) .
normally the time dilation factor is negligible , but it has been observed in the galactic microquasar ss 433 in emission lines which arise from two oppositely directed jets moving at 0.26@xmath195 near to the plane of the sky ( abell & margon 1979 ) . if there is absorbing gas in some or all of these bal quasars moving at a sufficiently high speed over a range of angles close to the plane of the sky , the resulting absorption will be seen at a range of velocities which can span both blueshifts and redshifts relative to the quasar rest frame . in appendix [ rds ]
we discuss some underappreciated aspects of the relativistic doppler shift relevant to line _ absorption _
( for example , absorption at a given blueshifted wavelength does not uniquely determine the velocity of the outflow even if its angle to the line of sight is known ) .
few papers on the effects of the relativistic doppler shift on absorption have appeared in the astronomical literature ( sher 1968 ; hovsepyan 1998 ; wang 2011 ) , in contrast to the numerous papers applying it to line _ emission _ in quasars ( beginning with mathews 1982 ) , in this section we consider only the potential application of the relativistic doppler shift to the absorption in our objects .
we use @xmath218 to denote the angle relative to our line of sight , with @xmath219 pointing away from us and @xmath220 denoting the plane of the sky . as an example , for a flow reaching the highest speeds seen in quasar outflows ( @xmath221 ) , a range of angles centered slightly toward the line of sight ( @xmath222 ) would be sufficient to explain the trough in j2157 ( approximately @xmath52000 km s@xmath1 to 10,000 km s@xmath1 ) . for
a flow reaching a top speed of only @xmath223 , reproducing the same trough would require twice as wide a range of angles , centered slightly away from the line of sight ( @xmath224 ) .
similarly , gas at a range of speeds from @xmath225 at @xmath220 would produce redshifted troughs spanning 0 to + 6000 km s@xmath1 , but an identical trough could be produced by gas at lower velocities directed slightly away from the observer ( @xmath226 ) .
producing redshifted troughs via the relativistic doppler effect does not require large gas inflow velocities , but does require even larger velocities nearly transverse to our line of sight to a quasar s continuum source .
one scenario which could generate large transverse velocities is a nearly vertical outflow launched from small radii , where the initial rotational velocity and radiative acceleration are both large , and viewed at a large inclination angle .
the relativistic doppler effect also could be a factor if any object is involved in a binary quasar scenario ( [ binary ] ) . in that scenario , the outflow velocity as well as the rotational velocity can contribute to the transverse velocity . in either case , large transverse velocities should result in variability in the redshifted troughs on the crossing timescale of the continuum source by the gas .
gas moving at @xmath227 crosses the 1550 continuum region of our fiducial @xmath201 @xmath128 bh + ss73 disc in 21 days
. the crossing time would be at most 2.2 times longer at 2800 than at 1550 for an ss73 disc , but could be 5 to 15 times longer at both wavelengths ( 105 to 315 days ) given the observationally inferred sizes of quasar continuum source regions ( blackburne et al .
2011 and [ rotvel ] ) .
for the most massive black holes ( 10@xmath228 @xmath128 ) , assuming a continuum source 5 to 15 times larger than a ss73 disc that nonetheless scales as @xmath190 ( equation [ e_rcon ] ) , the continuum source crossing time at @xmath227 would be 6 to 18 years .
we have presented observations of seventeen quasars with redshifted broad absorption in civ and siiv ( and often in numerous other transitions ) and new observations of two previously known quasars with redshifted broad absorption in mgii .
roughly 1 in 1000 bal quasars has redshifted civ absorption ( [ frac ] ) , although such absorption appears to be more common in the sdss - iii / boss dr9 quasar catalog ( 1 in 660 ) than in the sdss - i / ii dr7 quasar catalog ( 1 in 1600 ) . why these objects are so rare , given that infalling matter is ultimately the power source for all quasars , is not clear
. the absorption in these objects can extend to velocities of @xmath610,000 km s@xmath1 and appears just as saturated as absorption in standard , blueshifted bal troughs does .
the trough shapes can be classified as either @xmath21-shaped , _ v_-shaped or @xmath22-shaped ( [ uvw ] ) .
although absorption at the systemic redshift in these objects sometimes reaches zero flux within the noise , none of their redshifted absorption troughs do so , meaning that the redshifted absorption does not fully cover the continuum source . the redshifted absorption does not need to cover the broad emission line region to explain the observations ( with one exception ) , but we can not rule out equal covering of both the continuum and emission ine regions .
we have used the absorption index ( [ ai ] ) to measure the strengths of the blueshifted absorption ( ai ) , the redshifted absorption ( ai@xmath58 ) , and the total absorption ( ai@xmath59 ) in these objects .
the ai and ai@xmath58 values are uncorrelated overall , but their distributions are statistically indistinguishable .
the ai@xmath59 distribution is statistically consistent with the ai distribution of standard bal quasars , although the ai@xmath59 distribution lacks a tail to very strong troughs . in other words ,
the redshifted velocities of these troughs are unusual , but their overall strengths are not . based on our sample , at least 50% of quasars with redshifted civ absorption are lobal quasars , as compared to 5% of standard quasars being lobal quasars ( [ lobals ] )
. however , the ai@xmath59 distribution of these objects is not at all consistent with the ai distribution of standard lobal quasars ( [ ai ] ) .
the low - ionization gas in these objects appears to have a different origin than in standard lobal quasars . in one or two cases where high - excitation feiii absorption is seen ( [ dist ] )
, we infer that the absorber has a very high density ( @xmath112@xmath31 ) and a small distance from the ionizing continuum source ( @xmath126 pc ) .
we have discussed possible explanations for redshifted troughs including gravitational redshifts ( [ grav ] ) , infall ( [ infall ] ) , rotating disc - wind outflows ( [ rotation ] ) , outflows from one member of a binary quasar pair seen in absorption against the other ( [ binary ] ) , and the relativistic doppler effect in outflows moving close to transverse to our line of sight ( [ doppler ] ) .
none of these explanations by themselves seems likely to explain all our objects .
the binary quasar scenario can explain the observed range of trough shapes , velocities and absorption indices .
however , a factor of @xmath910 excess in the numbers of close binary quasars over the predictions of current studies would be needed to explain our sample entirely as binary quasars ; this scenario can not be the dominant explanation .
as alluded to in the title of our paper , we are left with two main scenarios to explain these objects .
it is possible that some of our objects will be explained by one scenario and some by the other .
infalling gas undergoing adiabatic compression is expected to have an increasing covering factor with infall velocity , making it a plausible explanation only for some @xmath21-shaped and @xmath22-shaped troughs .
the infall velocities observed in our objects have been seen in simulations of gas infall in quasars , but whether the gas at those velocities can match the large optical depths observed in civ and other species is not clear . in this scenario
, the high fraction of low - ionization redshifted troughs would be due to preferential infall of high - density gas .
however , the match between the distributions of ai@xmath59 in bal quasars with redshifted absorption and ai in standard bal quasars is not explained by this model .
rotation - dominated disc winds can explain the distributions of absorption indices in these objects , at least for @xmath21-shaped and @xmath22-shaped troughs .
rotation plus gravitational redshifting can in principle explain the range of trough shapes and velocity extents in this scenario .
low - ionization absorption might be preferentially seen in this model due to winds with relatively high - density gas having relatively low terminal velocities ( perhaps not even exceeding the escape velocity ) , enabling redshifted velocities due to rotation to dominate the absorption profile .
however , the variability properties available for a subset of our objects ( see below ) do not match the predictions of this scenario .
the observed variability of bal troughs is believed to be due primarily to tranverse motions of absorbing gas across the continuum source ( e.g. , gibson et al . 2010 ; hall et al . 2011 ) , although cases are known where the variability in narrower quasar outflows are more consistent with changes in the ionization of the absorbing gas ( misawa et al .
2007 ; hamann et al . 2011 ) . in this section
we discuss what each scenario considered herein for the origin of redshifted bal troughs might predict for the variability of those troughs .
future spectroscopic observations of these objects will be required to test these predictions . in all cases ,
what matters for variability is the transverse motion of the absorber and its size relative to the continuum source , not the distance of the absorber from the continuum source ; as both are at the same cosmological distance from us , only their relative sizes matter . in each of our suggested models for redshifted troughs , strong variability on timescales of less than a year
is possible because large transverse motions are possible .
not observing rapid variability in any of these objects over time would constrain the rotating wind model to ever greater azimuthal symmetry , the infall scenario to more ordered infall , the binary quasar scenario to smaller transverse velocities , and would rule out a substantial contribution from a relativistic doppler shift . in the infall scenario ,
more variability should occur at larger redshifted velocities , because gas with larger infall velocities is likely to have larger transverse velocities as well . in the quasar j1628 ,
a shift of the absorption to smaller average redshifted velocities has been seen .
if the absorption is due to gas undergoing decelerating infall due to outward radiative acceleration , future observations should show a continued shift to smaller average redshifted velocities and increasing deceleration .
variability in the rotation - dominated wind scenario has been discussed in [ rotabs ] . the pattern expected for the simplest variability scenario
( changes in absorption strength propagating from blueshifted to redshifted velocities ) has been seen in only one of the four cases where two epochs of high s / n spectroscopy are available .
a lack of variability in _ v_-shaped outflows on rest - frame timescales of two years or more would rule out this scenario for such objects . in the binary quasar scenario , variability at blueshifted and redshifted velocities
will be uncorrelated .
the level of variability in this scenario depends on the transverse velocity of the silhouetted outflow relative to the background quasar .
the variability could be comparable to that in standard bal quasars , but might be lower if bal outflows decrease substantially in velocity at large outflow distances due to the sweeping up of ambient gas ( faucher - giguere , quataert & murray 2011 ) . in the case of absorption redshifts arising from a relativistic doppler shift ,
the variability of such absorption is likely to be more substantial than the variability of a standard bal trough , as the large transverse velocity of the gas causes it to traverse the continuum source quickly .
a final unusual aspect of these quasars with redshifted broad absorption is that most of them have strong and relatively narrow ly@xmath29 emission ( nine of fifteen objects with ly@xmath29 coverage have such emission ) .
although it may be possible to produce apparent narrow ly@xmath29 emission via ly@xmath29 and nv absorption of underlying broad emission ( e.g. , hall et al .
2004c ) , absorption in those lines is not always strong in these objects .
a full comparison of the ly@xmath29 properties of these objects and of dr9q quasars is beyond the scope of this paper , but the prevalence of strong , narrow ly@xmath29 emission argues against a rotating accretion - disc scenario for redshifted civ absorption , as there is no obvious explanation for such ly@xmath29 emission in such scenarios .
our objects ly@xmath29 fwhms are near the upper end of the 5002000 km s@xmath1 range of the fwhms of ly@xmath29 emission from the regions of neutral gas near high - redshift galaxies known as ly@xmath29 blobs ( 2006 ; smith & jarvis 2007 , and references therein ) .
the fluxes of our objects narrow ly@xmath29 lines are about an order of magnitude larger than in the most luminous ly@xmath29 blobs known ( steidel et al . 2000 ) , so the lines must arise predominantly from illumination by the quasar .
the strong , relatively narrow ly@xmath29 emission in our quasars might arise due to larger than usual amounts of neutral gas in the quasar host galaxy and surrounding environs .
an excess of such gas would be consistent with the infall explanation for these objects .
narrow - band imaging or longslit spectroscopy of these objects to ascertain the spatial extent of their ly@xmath29 emission might constrain the origin of the gas responsible .
a variety of further observations of these objects would be useful to determine their nature .
near - infrared spectroscopy could provide accurate redshifts and possible evidence for binarity via narrow emission lines in the rest - frame optical .
continued monitoring of these objects spectra at observed optical wavelengths will establish the variability properties of the full sample .
high - resolution uv / optical spectroscopy of some objects may better constrain the column densities of the absorbing gas and thereby its physical properties and location as well .
x - ray observations could help discriminate between the different models we have put forward .
quasars observed to have the same strong x - ray absorption seen in standard bal quasars with the same trough strength and ionization range distributions as our sample are candidates for having rotating disc winds . in the infall scenario , however , the redshifted gas observed in front of these quasars uv emitting regions may not cover the smaller x - ray emitting region , or may have the same covering factor for both regions ( if the absorbing gas consists of many small clouds , for example ) .
thus , in this scenario quasars with only redshifted absorption may have x - ray properties similar to those of normal , non - bal quasars ( i.e. , no significant x - ray absorption , normal x - ray spectral shape , and nominal optical to x - ray flux ratio ) . larger samples of quasars with redshifted troughs would better establish how common they are and how their trough properties compare to those of standard , blueshifted - trough bal quasars .
the goal of the boss is to obtain at least 150,000 spectra of @xmath217 quasars , and the dr9q catalog contains only @xmath962,000 such objects ; therefore , an increase in the sample size a factor @xmath92.5 can be expected in the boss alone .
follow - up spectroscopy of candidate cases and a dedicated search for redshifted mgii absorption extending to lower redshifts ( especially where [ oii ] emission at the systemic redshift can be seen ) could increase this number still further .
furthermore , the existence of redshifted bal troughs raises the possibility that the processes producing them could also create narrower redshifted troughs .
that possibility may complicate the interpretation of the velocity distribution of narrow absorbers around quasar redshifts solely in terms of infall and clustering ( e.g. , wild et al . 2008 ) .
regardless of the origin of redshifted bal troughs , they provide a rare but useful probe of quasars and their environs .
any which are due to infall of gas offer a test of models of gas dynamics and agn fueling and feedback in galaxies during epochs of agn activity .
any which are due to rotating accretion disc winds offer a test of models of such winds .
any which are due to binary quasars offer a new view of bal outflows ; namely , along lines of sight which do not intersect the quasar which launched them .
such cases offer probes of bal outflow properties at known projected distances from their launch site , enabling new tests of models of bal dynamics .
we thank d. proga , m. bautista , d. edmonds and n. murray for discussions , and the referee for a careful review .
pbh thanks nserc for its research support , the institute of astronomy at the university of cambridge for hosting his sabbatical , and the aspen center for physics ( nsf grant # 1066293 ) for its hospitality .
wnb and nfa are supported by nsf grant ast-1108604 .
this research has made extensive use of nasa s astrophysics data system bibliographic services .
and of the atomic line list at http://www.pa.uky.edu/@xmath9peter / atomic/. funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s .
department of energy office of science .
the sdss - iii web site is http://www.sdss3.org/. sdss - iii is managed by the astrophysical research consortium for the participating institutions of the sdss - iii collaboration including the university of arizona , the brazilian participation group , brookhaven national laboratory , university of cambridge , carnegie mellon university , university of florida , the french participation group , the german participation group , harvard university , the instituto de astrofisica de canarias , the michigan state / notre dame / jina participation group , johns hopkins university , lawrence berkeley national laboratory , max planck institute for astrophysics , max planck institute for extraterrestrial physics , new mexico state university , new york university , ohio state university , pennsylvania state university , university of portsmouth , princeton university , the spanish participation group , university of tokyo , university of utah , vanderbilt university , university of virginia , university of washington , and yale university .
k. n. , adelman - mccarthy j. k. , ageros m. a. , allam s. s. , allende prieto c. , an d. , anderson k. s. j. , anderson s. f. , et al .
, 2009 , apjs , 182 , 543 g. o. , margon b. , 1979 , nature , 279 , 701 c. p. , alexandroff r. , allende prieto c. , anderson s. f. , anderton t. , andrews b. h. , aubourg . , bailey s. , et al . , 2012 , astrophys . j. suppl .
ser . , 203 , 21 h. , allende prieto c. , an d. , anderson s. f. , aubourg .
, balbinot e. , beers t. c. , berlind a. a. , et al . , 2011a , apjs , 193 , 29 , 2011b , apjs , 195 , 26 j. t. , hewett p. c. , maddox n. , richards g. t. , belokurov v. , 2011 , mnras , 410 , 860 j. , soares - santos m. , strauss m. a. , becker a. c. , dodelson s. , fan x. , gunn j. e. , hao j. , et al .
, 2011 , arxiv e - prints k. , 2010 , publ .
jpn . , 62 , 1333 n. , becker r. h. , laurent - muehleisen s. a. , gregg m. d. , white r. l. , brotherton m. s. , de kool m. , 1999 , apj , 524 , 566 n. , moe m. , costantini e. , korista k. t. , benn c. , ellison s. , 2008 , astrophys .
j. , 681 , 954 p. , proga d. , nagamine k. , 2012 , mon . not .
r. astron .
soc . , 424 , 728 m. a. , pradhan a. k. , 1998 , apj , 492 , 650 r. h. , white r. l. , helfand d. j. , 1995 , apj , 450 , 559 j. a. , pooley d. , rappaport s. , schechter p. l. , 2011 , astrophys .
j. , 729 , 34 a. s. , schlegel d. j. , aubourg e. , bailey s. , bhardwaj v. , brownstein j. r. , burles s. , chen y .- m .
, et al . , 2012 ,
aj , 144 b. , edmonds d. , arav n. , benn c. , chamberlain c. , 2012 , arxiv e - prints d. m. , hamann f. , shields j. c. , rodrguez hidalgo p. , barlow t. a. , 2011 , mnras , 413 , 908 g. , saez c. , brandt w. n. , giustini m. , garmire g. p. , 2009 ,
apj , 706 , 644 f. , elvis m. , lanzuisi g. , aldcroft t. , trichas m. , bongiorno a. , brusa m. , blecha l. , et al . , 2012 ,
apj , 752 , 49 f. , elvis m. , lanzuisi g. , jahnke k. , zamorani g. , blecha l. , bongiorno a. , brusa m. , et al . , 2010 , apj , 717 , 209 s. , zahn j .- p . , 1999 , astron .
, 344 , 433 d. m. , kraemer s. b. , george i. m. , 2003 , annu .
astrophys . ,
41 , 117 x. , shankar f. , sivakoff g. r. , 2008 , apj , 672 , 108 , 2012 , astrophys .
j. , 757 , 180 k. s. , schlegel d. j. , ahn c. p. , anderson s. f. , aubourg . , bailey s. , barkhouser r. h. , bautista j. e. , et al .
, 2013 , astron .
j. , 145 , 10 m. , becker r. h. , arav n. , gregg m. d. , white r. l. , 2002 , apj , 570 , 514 j. , evans ii n. j. , caselli p. , myers p. c. , shirley y. , aikawa y. , tafalla m. , 2007 , protostars and planets v , 17 a. v. , 2009 , mnras , 393 , 1433 , 2010 ,
mnras , 406 , 1060 d. , borguet b. , arav n. , dunn j. p. , penton s. , kriss g. a. , korista k. , costantini e. , et al . , 2011 , arxiv:1106.2084 a. , 1905 , annalen der physik , 322 , 891 d. j. , weinberg d. h. , agol e. , aihara h. , allende prieto c. , anderson s. f. , arns j. a. , aubourg e. , et al . , 2011 ,
aj , 142 , 72 m. , 2000 , apj , 545 , 63 d. , surace j. a. , veilleux s. , sanders d. b. , vacca w. d. , 2005 , apj , 626 , 70 d. , urrutia t. , lacy m. , efstathiou a. , afonso j. , coppin k. , hall p. b. , lonsdale c. , et al . , 2012 ,
apj , 745 , 178 c .- a . , quataert e. , murray n. , 2011 , arxiv:1108.0413 r. a. , gerardy c. l. , mclin k. m. , hamilton a. j. s. , 1999 , astrophys .
j. , 514 , 195 n. , brandt w. n. , hall p. b. , schneider d. p. , anderson s. f. , gibson r. r. , lundgren b. f. , myers a. d. , et al . , 2012 , astrophys .
j. , 757 , 114 c. b. , weymann r. j. , morris s. l. , turnshek d. a. , 1987 , apj , 317 , 450 c. b. , weymann r. j. , peterson b. m. , sun l. , malkan m. a. , chaffee frederic h. j. , 1986 , apj , 307 , 504 c. b. , wilkes b. , weymann r. , turnshek d. , 1983 , pasp , 95 , 341 m. , ichikawa t. , gunn j. e. , doi m. , shimasaku k. , schneider d. p. , 1996
, aj , 111 , 1748 j. r. , kraemer s. b. , crenshaw d. m. , george i. m. , brandt w. n. , hamann f. w. , kaiser m. e. , kaspi s. , et al .
, 2005 , astrophys .
j. , 631 , 741 s. , maiolino r. , juarez y. , nagao t. , marconi a. , bianchi s. , schneider r. , mannucci f. , et al . , 2010 , a&a , 523 , a85 r. , bond n. a. , charlton j. c. , eracleous m. , brandt w. n. , churchill c. w. , 2001 , apj , 549 , 133 r. , brotherton m. s. , 2008 , apj , 672 , 102 m. , ruszkowski m. , oh s. p. , 2013
, arxiv e - prints n. , 1986 , apj , 303 , 336 r. r. , brandt w. n. , gallagher s. c. , hewett p. c. , schneider d. p. , 2010 ,
apj , 713 , 220 r. r. , jiang l. , brandt w. n. , hall p. b. , shen y. , wu j. , anderson s. f. , schneider d. p. , et al . , 2009 ,
apj , 692 , 758 e. , urrutia t. , lacy m. , djorgovski s. g. , mahabal a. , myers a. d. , ross n. p. , petitjean p. , et al . , 2012 , astrophys .
j. , 757 , 51 j. e. , carr m. , rockosi c. , sekiguchi m. , berry k. , elms b. , de haas e. , ivezi . , et al . , 1998 , aj , 116 , 3040 j. e. , siegmund w. a. , mannery e. j. , owen r. e. , hull c. l. , leger r. f. , carey l. n. , knapp g. r. , et al .
, 2006 , aj , 131 , 2332 p. b. , 2007 , aj , 133 , 1271 p. b. , anderson s. f. , strauss m. a. , york d. g. , richards g. t. , fan x. , knapp g. r. , schneider d. p. , et al . , 2002 , apjs , 141 , 267 p. b. , anosov k. , white r. l. , brandt w. n. , gregg m. d. , gibson r. r. , becker r. h. , schneider d. p. , 2011 ,
mnras , 411 , 2653 p. b. , hoversten e. a. , tremonti c. a. , vanden berk d. e. , schneider d. p. , strauss m. a. , knapp g. r. , york d. g. , et al . , 2004a ,
aj , 127 , 3146 p. b. , hutsemkers d. , 2003 , in active galactic nuclei : from central engine to host galaxy , collin s. , combes f. , shlosman i. , eds . , p. 209
p. b. , knapp g. r. , richards g. t. , strauss m. a. , anderson s. f. , schneider d. p. , vanden berk d. a. , york d. g. , et al .
, 2004b , in multiwavelength agn surveys , r. mjica & r. maiolino , ed . , pp . 247252 p. b. , kowalski p. m. , harris h. c. , awal a. , leggett s. k. , kilic m. , anderson s. f. , gates e. , 2008 , aj , 136 , 76 p. b. , snedden s. a. , niederste - ostholt m. , eisenstein d. j. , strauss m. a. , york d. g. , schneider d. p. , 2004c
, aj , 128 , 534 f. , kanekar n. , prochaska j. x. , murphy m. t. , ellison s. , malec a. l. , milutinovic n. , ubachs w. , 2011 , mnras , 410 , 1957 f. , sabra b. , 2004 , in astronomical society of the pacific conference series , vol .
311 , agn physics with the sloan digital sky survey , g. t. richards & p. b. hall , ed . , pp .
203212 l. e. , drew j. e. , long k. s. , knigge c. , proga d. , 2002 , mnras , 332 , 127 a. , yamaguchi h. , tamagawa t. , katsuda s. , hwang u. , hughes j. p. , ozawa m. , bamba a. , et al . , 2010 , astrophys .
j. , 725 , 894 j. f. , myers a. d. , shen y. , strauss m. a. , djorgovski s. g. , fan x. , glikman e. , mahabal a. , et al . , 2010 , astrophys .
j. , 719 , 1672 j. f. , strauss m. a. , oguri m. , inada n. , richards g. t. , pindor b. , schneider d. p. , becker r. h. , et al . , 2006 , aj , 131 , 1 p. c. , foltz c. b. , chaffee f. h. , 2001 , aj , 122 , 518 p. c. , wild v. , 2010 , mnras , 405 , 2302 a. , nayakshin s. , power c. , king a. , 2011 , mon . not
r. astron .
soc . , 413 , 2633 j. a. , becker r. h. , white r. l. , richards g. t. , zeimann g. r. , 2011 , arxiv:1103.5749 p. f. , quataert e. , 2010 , mnras , 407 , 1529 y. i. , 1998 , physics uspekhi , 41 , 941 d. , wilson c. d. , takakuwa s. , yun m. s. , petitpas g. r. , peck a. b. , ho p. t. p. , matsushita s. , et al .
, 2007 , apj , 659 , 283 b. t. , hartig g. f. , kirhakos s. , sargent w. l. w. , turnshek d. a. , weymann r. j. , bahcall j. n. , bergeron j. , et al .
, 1996 , apjl , 470 , l11 e. b. , wallerstein g. , silk j. , 1976 , astrophys . j. suppl .
ser . , 32 , 681 j. , mediavilla e. , muoz j. a. , kochanek c. s. , 2012 , astrophys .
j. , 751 , 106 s. , behar e. , 2010 , space sci .
, 157 , 249 a. l. , miller j. m. , raymond j. , 2012 , apj , 746 , 2 c. , scaringi s. , goad m. r. , cottis c. e. , 2008 , mnras , 386 , 1426 c. , woods j. a. , drew j. e. , 1995 , mnras , 273 , 225 r. , proga d. , 2009 , apj , 693 , 1929 a. , elvis m. , 2010 , apj , 714 , 561 k. m. , moore j. r. , 2004 , apj , 611 , 107 r. , gnedin n. y. , hamilton a. j. s. , kravtsov a. v. , 2008 , apj , 678 , 154 j. , ostriker j. , sunyaev r. , 2012 , apj , submitted x. , shen y. , strauss m. a. , hao l. , 2011 , apj , 737 , 101 c. r. , 1967 , apj , 147 , 396 r. , oliva e. , ghinassi f. , pedani m. , mannucci f. , mujica r. , juarez y. , 2004 , a&a , 420 , 889 c. l. , bouch n. , 2009 , astrophys . j. , 703 , 1394 w. g. , 1982 , astrophys .
j. , 258 , 425 t. , eracleous m. , charlton j. c. , kashikawa n. , 2007 , apj , 660 , 152 m. , arav n. , bautista m. a. , korista k. t. , 2009 , apj , 706 , 525 c. e. , 1950 , an ultraviolet multiplet table , sections 1 - 2 .
nbs circular 488 , washington : us government printing office n. , chiang j. , 1998 , apj , 494 , 125 n. , chiang j. , grossman s. a. , voit g. m. , 1995 , apj , 451 , 498 a. d. , brunner r. j. , richards g. t. , nichol r. c. , schneider d. p. , bahcall n. a. , 2007 , astrophys .
j. , 658 , 99 a. d. , richards g. t. , brunner r. j. , schneider d. p. , strand n. e. , hall p. b. , blomquist j. a. , york d. g. , 2008 , apj , 678 , 635 s. , zubovas k. , 2012 , arxiv e - prints k. k. , fynbo j. p. u. , mller p. , sommer - larsen j. , ledoux c. , 2006 , astron .
astrophys . , 452 , l23 d. e. , ferland g. j. , 2006 , astrophysics of gaseous nebulae and active galactic nuclei , second edition .
sausalito : university science books , pp .
125126 h. , lima m. , cunha c. e. , lin h. , frieman j. , sheldon e. s. , 2008 , apj , 674 , 768 i. , petitjean p. , aubourg . , bailey s. , ross n. p. , myers a. d. , strauss m. a. , anderson s. f. , et al . , 2012 , astron .
astrophys . , 548 , a66 w. h. , teukolsky s. a. , vetterling w. t. , flannery b. p. , 2007
, numerical recipes , 3rd edition .
cambridge : cambridge university press , p. 731
d. , 2003 , apjl , 592 , l9
d. , kallman t. r. , 2004 , apj , 616 , 688 d. , kallman t. r. , drew j. e. , hartley l. e. , 2002 , apj , 572 , 382 d. , ostriker j. p. , kurosawa r. , 2008 , apj , 676 , 101 d. , rodriguez - hidalgo p. , hamann f. , 2012 , in astronomical society of the pacific conference series , vol .
460 , agn winds in charleston , chartas g. , hamann f. , leighly k. m. , eds .
, p. 171
g. t. , vanden berk d. e. , reichard t. a. , hall p. b. , schneider d. p. , subbarao m. , thakar a. r. , york d. g. , 2002 , aj , 124 , 1 g. , nardini e. , salvati m. , elvis m. , fabbiano g. , maiolino r. , pietrini p. , torricelli - ciamponi g. , 2011 , mnras , 410 , 1027 p. , hamann f. , hall p. , 2011 ,
mnras , 411 , 247 j. a. , hall p. b. , snedden s. a. , brotherton m. s. , anderson s. f. , 2011 , new astronomy , 16 , 128 n. p. , myers a. d. , sheldon e. s. , yche c. , strauss m. a. , bovy j. , kirkpatrick j. a. , richards g. t. , et al .
, 2012 , apjs , 199 , 3 d. p. , richards g. t. , hall p. b. , strauss m. a. , anderson s. f. , boroson t. a. , ross n. p. , shen y. , et al . , 2010 , aj , 139 , 2360 n. i. , sunyaev r. a. , 1973 , a&a , 24 , 337 z. , huiyuan w. , hongyan z. , tinggui w. , peng j. , 2011 , arxiv:1106.1587 y. , strauss m. a. , oguri m. , hennawi j. f. , fan x. , richards g. t. , hall p. b. , gunn j. e. , et al .
, 2007 , aj , 133 , 2222 d. , 1968 , jrasc , 62 , 105 s. a. , proga d. , kurosawa r. , long k. s. , miller l. , turner t. j. , 2012 , arxiv e - prints s. , gunn j. e. , uomoto a. , roe n. , schlegel d. , rockosi c. m. , carr m. a. , leger f. , et al . , 2012 , arxiv e - prints d. j. b. , jarvis m. j. , 2007 , mnras , 378 , l49
c. c. , adelberger k. l. , shapley a. e. , pettini m. , dickinson m. , giavalisco m. , 2000 , astrophys .
j. , 532 , 170 c. c. , erb d. k. , shapley a. e. , pettini m. , reddy n. , bogosavljevi m. , rudie g. c. , rakic o. , 2010 , apj , 717 , 289 c. , lupton r. h. , bernardi m. b. , blanton m. r. , burles s. , castander f. j. , connolly a. j. , eisenstein d. j. , et al .
, 2002 , aj , 123 , 485 m. , peebles p. j. e. , 1998 , apjl , 500 , l79 t. a. , quataert e. , murray n. , 2005 , astrophys .
j. , 630 , 167 f. , cappi m. , reeves j. n. , palumbo g. g. c. , braito v. , dadina m. , 2011 , apj , 742 , 44 j. r. , hall p. b. , reichard t. a. , richards g. t. , schneider d. p. , vanden berk d. e. , knapp g. r. , anderson s. f. , et al . , 2006 , apjs , 165 , 1 s. , uttley p. , 2008
, mnras , 390 , 421 m .- p . , joly m. , vron p. , 2004
, a&a , 417 , 515 e. j. , chugai n. n. , petitjean p. , 1995
, apj , 443 , 586 c. , 2011 , annalen der physik , 523 , 239 s. j. , hambly n. c. , dye s. , almaini o. , cross n. j. g. , edge a. c. , foucaud s. , hewett p. c. , et al .
, 2007 , mnras , 375 , 213 r. j. , morris s. l. , foltz c. b. , hewett p. c. , 1991 , apj , 373 , 23 v. , hewett p. c. , 2010 , arxiv:1010.2500 v. , kauffmann g. , white s. , york d. , lehnert m. , heckman t. , hall p. b. , khare p. , et al . , 2008 , mnras , 388 , 227 p. f. , long k. s. , hamilton a. j. s. , fesen r. a. , 2005 , astrophys .
j. , 624 , 189 e. l. , eisenhardt p. r. m. , mainzer a. k. , ressler m. e. , cutri r. m. , jarrett t. , kirkpatrick j. d. , padgett d. , et al . , 2010 , aj , 140 , 1868 c .- c . , crenshaw d. m. , hamilton a. j. s. , fesen r. a. , leventhal m. , sarazin c. l. , 1997 , astrophys
. j. lett . , 477 , l53 j. , brandt w. n. , hall p. b. , gibson r. r. , richards g. t. , schneider d. p. , shemmer o. , just d. w. , et al . , 2011 ,
apj , 736 , 28 d. g. , adelman j. , anderson j. e. , anderson s. f. , annis j. , bahcall n. a. , bakken j. a. , barkhouser r. , et al . , 2000 ,
aj , 120 , 1579
* sdss [email protected] * ( j0028 ; fig .
[ f_cands2 ] ) has a redshift of @[email protected] ( hewett & wild 2010 ) , from emission features of ly@xmath29 , siiii ] , and ciii ] .
( only a visual redshift is available in dr9q . )
j0028 has absorption troughs spanning the systemic redshift in civ , siiv , nv , and pv .
the siiv trough is continuous , but the civ trough is interrupted by weak , narrow civ emission at the systemic redshift .
there is also emission and absorption in siii , oi , and cii .
the three epochs of spectroscopy on this quasar are shown in figure [ fj0028three ] and are discussed in [ rpt ] . * sdss [email protected] * ( j0148 ; fig .
[ f_cands ] ) has a dr9q ciii ] pca redshift of @[email protected] .
it has strong , narrow ly@xmath29 emission and accompanying narrow civ emission which appears to interrupt a bal trough spanning both blueshifted and redshifted velocities , as seen in siiv .
similar troughs interrupted by narrow emission are seen in ovi and nv .
there may be some feii * absorption longward of civ , but feii * can not explain the redshifted absorption seen in ovi , nv and siiv . at the blue end of the siiv trough where the absorption is strongest ,
absorption is also seen in pv , siv , possibly ciii ] * , cii , and aliii .
there also appears to be strong , blueshifted emission of civ and siiv . * sdss [email protected] * ( j0805 ; fig .
[ f_cands ] ) has a dr9q ciii ] pca redshift of @[email protected] .
its spectrum is remarkably similar to that of j0148 , except that the cii absorption is weaker and only weak blueshifted emission is discernible in civ and none at all in siiv . * sdss [email protected] * ( j0828 ; fig .
[ f_cands ] ) has a dr9q full pca redshift of @[email protected] , which is a better match to the spectrum than the dr9q ciii ] pca @xmath19=2.370 .
the true systemic redshift may be even smaller , as j0828 has strong ly@xmath29 emission with an apparent peak at @xmath19=2.361 , but we adopt a systemic @xmath19=2.366 to be conservative .
it has strong , blueshifted absorption in nv , cii , siiv , civ , aliii , and mgii .
it has redshifted absorption in siiv and civ , and probably in nv , cii , and aliii . * sdss [email protected] * ( j0830 ; fig .
[ f_cands ] ) has narrow ly@xmath29 emission at @[email protected] ; there is also narrow emission at the same redshift in nv , siiv , and civ .
the dr9q full pca redshift is @[email protected] ( no ciii ] pca redshift is available ) .
we adopt the narrow emission redshift as systemic to be conservative about the extent of the redshifted absorption .
there is redshifted absorption in ciii * , nv , cii , siiv , civ , aliii , and possibly ly@xmath29 , extending to at least @xmath19=2.465 in all troughs from at least @xmath19=2.417 in ciii * , nv , siiv and civ ( @xmath51530 to + 2650 km s@xmath1 , @xmath229 km s@xmath1 ) and from @xmath19=2.43 in ly@xmath29 , cii and aliii ( @xmath5390 to + 2650 km s@xmath1 , @xmath230 km s@xmath1 ) .
ly@xmath29 absorption is confused with broad and narrow ly@xmath29 emission , making the absorption identification uncertain ; however , the redshifted velocity of the end of the trough for other ions matches that of the putative ly@xmath29 trough . * sdss [email protected] * ( j0941 ; fig .
[ f_cands ] ) has @[email protected] from its dr9q ciii ] pca redshift .
it has a red continuum and striking , narrow emission lines of ovi , ly@xmath29 , nv , siiv ( very weak ) , and civ .
j0941 has blueshifted absorption in ovi , pv , nv , siiv , and civ .
it has redshifted civ absorption which is roughly symmetric around the emission redshift with the blueshifted civ absorption .
there appears to be accompanying redshifted siiv absorption , and possibly redshifted pv and nv absorption , but there is no sign of redshifted ovi absorption .
sdss j0941 is unique in our sample in that it is morphologically classified as a galaxy ; see [ cases ] . * sdss [email protected] * ( j1019 ; fig .
[ f_cands ] ) has a dr9q ciii ] pca redshift of @xmath19=2.452 .
it has broad emission in nv , aliii , ciii ] , feiii uv34 , uv50 and uv48 , and possibly mgii .
blue wings of siiv and civ emission also appear to be present .
there is redshifted absorption ( from @xmath19=2.46 to @xmath19=2.54 , or @xmath27=+800 km s@xmath1 to @xmath27=+7500 km s@xmath1 ) in feiii , aliii , alii , civ , siiv , cii , siii and nv . shortward of ly@xmath29 , there is absorption consistent with redshifted ciii * , pv+feiii , siv and probably ovi , but given the resolution of the boss spectrum , ly@xmath29 forest contamination makes those identifications less certain .
the civ and feiii absorption extend to @xmath19=2.55 ( @xmath27=+8300 km s@xmath1 ) , and may extend to @xmath19=2.57 ( @xmath27=+10100 km s@xmath1 ) .
the feiii absorption is seen in multiplets uv34 , uv48 , and uv50 . as discussed in [ cands ] , bal troughs in uv34 and uv48
have been seen before , but this is the first reported case of uv50 absorption .
an intervening absorber at @xmath19=1.6135 is responsible for narrow absorption at 4050 , 4200 and 4365 .
* sdss [email protected] * ( j1034 ; fig .
[ f_cands2 ] ) at @xmath231 ( hewett & wild 2010 ) has civ absorption both blueshifted ( @xmath93100 km s@xmath1 wide up to @xmath232 km s@xmath1 ) and redshifted ( @xmath95200 km s@xmath1 wide , up to @xmath233 km
s@xmath1 ) in its sdss and boss spectra , plus blueshifted aliii and mgii .
in addition , its boss spectrum shows both blueshifted and redshifted siiv . in both civ and siiv ,
the blueshifted absorption is much stronger than the redshifted absorption .
there is no sign of [ oii]@xmath33727,3729 emission in its boss spectrum . * sdss [email protected] * ( j1146 ; fig .
[ f_cands ] ) has a dr9q ciii ] pca redshift of @xmath19=2.780 .
it has weak , redshifted absorption in nv , siiv , civ , and probably aliii and ovi .
j1146 has the shallowest absorption and the smallest ai@xmath59 value in our sample .
* sdss [email protected] * ( j1147 ; fig .
[ f_cands2 ] ) has @[email protected] ( hewett & wild 2010 ) from siiii]+ciii ] emission .
( the dr9q ciii ] pca redshift is @xmath19=2.5600 , but despite the small difference our inspection shows that the hw10 redshift provides the better fit . )
however , the continuum slope of the sdss spectrum of this object is in error .
the @xmath234 and @xmath235 colors synthesized from the spectrum are much redder than the corresponding colors from the sdss imaging and the boss spectrum .
therefore , we only plot the boss spectrum in figure [ f_cands2 ] .
j1147 shows absorption in siiv and civ which is strongest at @xmath19=2.540 ( @xmath51340 km s@xmath1 ) and extends to redshifted velocities , generally weakening with increasing redshifted velocity .
there is no clear redshifted nv trough , but it may be concealed by narrow nv emission ; such emission is seen in civ .
narrow ly@xmath29 , nv , siii , and cii absorption are seen at @xmath19=2.540 and at @xmath19=2.610 ( + 4530 km s@xmath1 ) . * sdss [email protected] * ( j1323 ; fig .
[ f_cands ] ) is at @[email protected] from our inspection of siiii]+ciii ] ; no dr9q ciii ] pca redshift is available , and the dr9q full pca redshift of @xmath19=2.376 is spuriously low .
it has troughs in pv , nv , siiv , civ and possibly cii that span both blueshifted and redshifted velocities .
j1323 has blueshifted civ emission but no sign of blueshifted siiv emission .
j1323 has two epochs of boss spectroscopy , but the second - epoch spectrum has a considerably lower signal - to - noise ratio than the first , and so no notable trough variability is distinguishable between the two epochs . *
sdss [email protected] * ( j1439 ) has a dr9q ciii ] pca redshift of @[email protected] , which matches the redshift of the peak of its narrow ly@xmath29 emission .
any mgii emission present is swamped by noise around 9800 ( observed ) ; there is also poor night sky line subtraction near 5892 ( observed ) .
the spectrum of j1439 resembles that of j2157 .
j1439 appears to have redshifted absorption in ciii * , nv , siiv , and civ beginning at its systemic redshift and extending to @xmath236 ( @xmath237 km s@xmath1 ) in all three lines , and possibly to @xmath238 ( + 11200 km s@xmath1 ) in civ .
there are also relatively narrow absorption systems at @xmath19=2.465 in ly@xmath29 , civ and possibly siiv , and at @xmath19=2.4873 in ly@xmath29 , civ , siii , cii , and possibly nv ; the latter redshift may mark the short - wavelength end of the broad absorption troughs . * sdss [email protected] * ( j1440 ; fig .
[ f_cands ] ) , at a dr9q ciii ] pca redshift of @[email protected] , has blueshifted and redshifted absorption in ovi , nv , siii , cii , siiv , and civ .
the troughs are split by weak emission just longward of the systemic redshift in each transition . * sdss [email protected] * ( j1628 ; fig .
[ f_cands2 ] ) at @xmath239 ( hewett & wild 2010 ) has broad civ absorption blueshifted by up to 9500 km s@xmath1 and a redshifted civ trough extending to 12400 km s@xmath1 in its sdss spectrum and to 11100 km s@xmath1 in its boss spectrum .
we previously suggested the latter trough could instead be either a very high velocity ( @xmath240 km s@xmath1 ) aliii trough without accompanying mgii , or heii@xmath41640 absorption associated with the @xmath241 km s@xmath1 civ outflow ( hall et al .
these hypotheses are ruled out by the detection of redshifted siiv absorption , at velocities matching those of the redshifted civ absorption , in the boss spectrum of this object .
j1628 also exhibits narrow absorption at @xmath19=1.4967 ( @xmath511600 km s@xmath1 ) in civ , alii , feii , and mgii and narrow feii+mgii absorption at @xmath242 .
no [ oii ] emission is detected in the boss spectrum of this object . * sdss [email protected] * ( j1709 ; fig .
[ f_cands ] ) has strong , narrow ly@xmath29 at @xmath243 , consistent with the dr9q ciii ] pca redshift of @xmath244 , which we adopt as systemic .
a redshifted trough is seen in ovi , ciii * , nv , siiv and civ , extending to + 4430 km s@xmath1 ( slightly narrower in siiv ) .
weak , redshifted absorption is seen in aliii@xmath31854,1862 . * sdss [email protected] * ( j1724 ; fig .
[ f_cands ] ) has narrow ly@xmath29 emission at @[email protected] , but we adopt the dr9q ciii ] pca redshift of @[email protected] as systemic .
there is absorption which is mostly redshifted in nv , cii , siiv , civ , aliii , and possibly mgii .
the siiv and civ absorption extends from @xmath19=2.51 to @xmath19=2.62 ( @xmath245=9350 km s@xmath1 ) , and the nv absorption is at least that wide .
the cii and aliii troughs are narrower ( 4200 km s@xmath1 ) .
the mgii region is at the far red end of the spectrum , but there is a suggestion of a trough in the same velocity range as cii and aliii and a depth of 50% of the continuum .
below ly@xmath29 , there is strong , broad ciii*@xmath41175 and pv / feiii @xmath41116 emission in this object , along with redshifted absorption from ovi , siv@xmath41062+siv*@xmath41072 , pv@xmath31118,1128 and ciii*. there is a narrow ly@xmath29 absorber at @xmath19=2.53 .
the apparent narrow ly@xmath29 absorber longward of that is cii from a narrow intervening system with civ , siiv and possibly mgii at @xmath19=2.224 .
there is also an intervening system with siiv , civ , aliii , and mgii at 1.9066 .
* sdss [email protected] * ( j2157 ; fig .
[ f_cands ] ) has a dr9q ciii ] pca redshift of @xmath26 , which we adopt as systemic .
( we do not adopt the dr9q mgii pca redshift of @xmath246 because mgii is weak and affected by night sky emission line residuals . )
it has absorption troughs in nv , siiv and civ composed of absorption beginning at a small blueshift ( @xmath247 km s@xmath1 in civ ) and extending smoothly to a large redshift ( @xmath27=9050 km s@xmath1 in civ ) at generally decreasing depth .
the region of the blueshifted trough at @xmath248 either has a different range of ionization stages present or constitutes a separate , narrow absorption system ; absorption is seen there in ly@xmath29 , siii and cii as well as in nv , siiv , and civ .
there also appears to be redshifted pv and aliii absorption .
the absorption in aliii is narrower than in the other ions , which is typical for standard bal quasars which exhibit aliii troughs .
the spectrum exhibits strong , narrow ly@xmath29 emission accompanied by narrow oi@xmath41302+siii@xmath41304 , heii@xmath41640 , mgii and possibly civ emission .
in this section we discuss a number of quasars which have candidate redshifted high - ionization absorption features in their spectra . their spectra are shown in figure [ f_1633 ] and figure [ f_cands3 ] .
we discuss the candidate shown in figure [ f_1633 ] at some length before discussing the others . * sdss [email protected] * ( j1633 ; fig .
[ f_1633 ] ) has a dr9q mgii pca redshift of @xmath249 , which we adopt as systemic ( no ciii ] pca redshift is available ) .
it has a complex of three narrow absorption systems at blueshifts @xmath22000 km s@xmath1 in civ , nv , cii@xmath41334 and siiv .
it also has what appear to be broad troughs extending longward of the rest wavelengths of siiv ( halfway to the position of civ ) and civ ( to the position of heii ) .
these troughs extend to at least @xmath27=7300 km s@xmath1 and possibly to @xmath27=14100 km s@xmath1 in civ .
there may be a redshifted component of cii@xmath41334 absorption , but there is no clear sign of redshifted absorption in ly@xmath29 or nv .
j1633 also exhibits an intervening mgii+feii system at @xmath250 and an intervening civ system at @xmath251 .
j1633 was detected in the radio by the first survey ( becker et al .
1995 ) and has a deconvolved size of [email protected] with major axis oriented 86.1@xmath182 e of n. the reality of the putative redshifted siiv and civ absorption in j1633 is unclear , as follows .
the ultraviolet spectrum of a quasar can be thought of as a roughly power - law continuum plus strong , broad emission lines and a pseudocontinuum of weaker and/or blended broad emission lines , balmer continuum emission , hi two - photon emission ( hall et al .
2004a ) , etc . the combination of unusual emission - line profiles and strong emission in the pseudocontinuum can cause gaps between emission complexes to mimic absorption features .
j1633 is a potential example of this phenomenon .
its spectrum appears explainable by strong siiv and civ emission which is almost entirely blueshifted from the systemic wavelength , plus a complex of narrow nv , siiv and civ absorption systems near the systemic redshift , plus an unusual spectral shape between civ and ciii]@xmath41908 .
cases of entirely blueshifted civ emission are rare but not unknown ( e.g. , fig . 3 of leighly & moore 2004 ) . in particular
, the spectrum of j1633 strongly resembles the composite spectrum of ` phl 1811 analog ' quasars presented in fig .
8 of wu et al .
the uv spectra of phl 1811 analogs exhibit weak high - ionization emission lines but strong feiii 2080 and uv feii emission like that seen in j1633 . as an example , the candidate phl 1811 analog sdss [email protected] shows a strong emission complex between civ and ciii ] .
we show in fig .
[ f_1633 ] the spectrum of j1633 as a thick black line , with a scaled and heavily smoothed spectrum of the unusual quasar sdss [email protected] overplotted as a thin red line .
we used a scaling @xmath252 to roughly account for the effects of different power - law continuum slopes and reddening in the two objects .
the scaling demonstrates that a quasar without any obvious absorption can match the continuum shape of j1633 at wavelengths between civ and mgii .
( this figure also demonstrates that if j1633 lacked narrow absorption troughs , it would likely qualify as a phl 1811 analog according to the criteria of wu et al .
( 2011 ) . )
a slightly shallower scaling would better match the continuum near siiv , indicating that at wavelengths just longward of civ j1633 has weaker emission than sdss [email protected] does .
possible contributors for the strong emission between civ and ciii ] include feii , heii , alii , niv ] and aliii .
however , the exact origin is immaterial for our purposes : we have demonstrated that j1633 can plausibly be explained without invoking redshifted absorption troughs , and so we exclude it from further discussion here . nonetheless , it would be worth obtaining further spectroscopy of this object , to shorter rest - frame wavelengths if possible , to characterize its spectral shape more accurately and to search for absorption variability .
turning to figure [ f_cands3 ] , we have the candidate * sdss [email protected] * ( j0050 ; fig .
[ f_cands3 ] ) .
it has strong , narrow ly@xmath29 emission at @[email protected] , consistent with the dr9q mgii pca redshift of @xmath19=2.112 and the dr9q full pca redshift of @xmath19=2.118 ; we adopt the latter as systemic .
( the dr9q ciii ] pca redshift of @xmath19=2.129 is spuriously large . ) a redshift consistent with the redshift of the peaks of civ and mgii .
it has a redshifted trough seen in ly@xmath29 , nv , cii , siiv and civ from @xmath27=0 to + 1780 km s@xmath1 ) .
the relatively narrow velocity range means that it is either a mini - bal quasar or has a complex of unresolved narrow absorption systems along our line of sight which appear redshifted due to their peculiar velocities ( wild et al .
in fact , within the broad redshifted trough , there is narrow aliii , feiii , and mgii absorption at @xmath19=2.129 ( + 1060 km s@xmath1 ) .
the broad absorption does not reach zero depth in ly@xmath29 , but nearly does in nv .
the absorber might be too highly ionized to show strong ly@xmath29 , or might not fully cover the ly@xmath29 emission region at the velocities in question .
the line widths and strengths of the absorption features in j0050 are probably too strong for the absorber to be an intervening system with a redshifted peculiar velocity of 1060 km s@xmath1 relative to the quasar .
however , the redshifted velocity is one of the smallest in our sample , and the ly@xmath29 , ciii ] and mgii lines are all affected by absorption , increasing the uncertainty of this object s systemic redshift so that a redshift high enough to eliminate the redshifted absorption can not be excluded . until a more robust systemic redshift is determined ( e.g. , through near - infrared spectroscopy of rest - frame optical narrow emission lines ) , we consider this quasar merely a candidate for having redshifted absorption . * sdss [email protected] * ( j1239 ; fig .
[ f_cands3 ] ) may be a case of a quasar with an unusual continuum shape and relatively narrow absorption at the systemic redshift .
it has an sdss spectrum which differs from the boss spectrum in having a redder continuum shape and a weaker ly@xmath29+nv emission line .
j1239 has strong narrow high- and low - ionization absorption from a potential damped ly@xmath29 absorber at @[email protected] .
it has a strong 1900 emission complex at a redshift consistent with the absorption redshift .
ly@xmath29 , nv , civ , heii and probably mgii emission are blueshifted by 3000 km s@xmath1 to @[email protected] .
the spectrum also exhibits a strong uv feii pseudocontinuum .
there appears to be broad redshifted absorption in civ , siiv and possibly nv extending from at least @xmath253 km s@xmath1 to @xmath254 km s@xmath1 ( @xmath255 km s@xmath1 ) , which would still extend to redshifted velocities even if we adopted the hewett & wild ( 2010 ) redshift of [email protected] for this object .
our lack of certainty in the reality and extent of these putative redshifted troughs may be due to the low signal - to - noise ratio of the spectra .
nonetheless , as with j1633 , a combination of blueshifted siiv and civ emission and strong feii pseudocontinuum could explain the spectrum of this object without redshifted absorption troughs , so for now we consider it merely a candidate .
* sdss [email protected] * ( j1316 ; fig .
[ f_cands3 ] ) at @xmath256 from [ oii ] emission ( hewett & wild 2010 ) is our only candidate with mgii absorption .
this object is not a boss ancillary target , but has three sdss spectra ( one of which is very noisy ) .
there is some flux variability between the two best sdss spectra , but no statistically significant trough variability between the normalized spectra .
the spectrum of this object presented in figure 3 of hewett , foltz & chaffee ( 2001 ) was taken in between the last two sdss spectral epochs and confirms the absorption features bracketing the systemic redshift of mgii , extending from @xmath257 km s@xmath1 to @xmath258 km s@xmath1 in a @xmath22 shape .
the sdss spectra show that these are not nal systems ; both `` troughs '' are smooth over their @xmath92000 km s@xmath1 widths .
the blueshifted trough has been included in the bal quasar samples of trump et al .
( 2006 ) and gibson et al .
we classify this object as having candidate redshifted absorption because we do not have confirmation of the redshifted trough in other transitions .
in fact , this object was mentioned in hall et al .
( 2002 ) , where we suggested that both apparent mgii troughs could be gaps between feii multiplets seen in emission , as could the apparent troughs near 4000 observed .
however , those latter features could be redshifted feiii absorption , and there may be redshifted balmer - line absorption as well .
the high densities required for such absorption ( hall 2007 ) could yield a very useful distance constraint for the absorber in this object ( see [ dist ] ) , if the absorption is confirmed . * sdss [email protected] * ( j1342 ; fig .
[ f_cands3 ] ) has blueshifted troughs in civ , siiv , ly@xmath29+nv , and ovi ( the latter being particularly strong relative to civ ) .
our inspection redshift of @xmath259 agrees with the dr9 pipeline redshift and is equal within the errors to the pca redshift of @xmath260 .
it also appears to have a weak , redshifted civ trough .
however , the significance of that putative trough is low due to the noise in the spectrum , and there is no redshifted trough apparent in any other ion .
therefore we consider j1342 as only a candidate for redshifted absorption .
* sdss [email protected] * ( j1704 ; fig .
[ f_cands3 ] ) has ly@xmath29 emission which peaks at @[email protected] .
hewett & wild ( 2010 ) give @[email protected] , which is clearly too low a redshift .
there is narrow ly@xmath29 absorption at @[email protected] , consistent with the redshift of the broad ciii ] line , and so we adopt that as the systemic redshift . however , the dr9q ciii ] redshift is @xmath19=3.491 , in which case its absorption troughs would be purely blueshifted .
examination of the ciii ] emission region suggests that the true systemic redshift lies somewhere between these two values .
therefore , we consider this object only a candidate for redshifted absorption until a more robust systemic redshift is available .
both sdss and boss spectra are available for j1704 , which is slightly fainter in the boss epoch .
there is no significant trough variability between the two epochs .
j1704 has troughs extending across our adopted systemic redshift in siiv , civ and aliii from at least @xmath261 km s@xmath1 to @xmath262 km s@xmath1 , and blueshifted troughs in aliii , alii , cii , oi / siii , siii , nv , ciii * and ovi , with spans of up to @xmath263 km s@xmath1 . because the redshifted portions of the civ , siiv and aliii troughs in this object extend only to @xmath9 + 2500 km s@xmath1 , it is possible that they are blends of narrow associated absorbers which appear redshifted due to their peculiar velocities ( e.g. , wild et al .
that possibility can be tested with high resolution spectroscopy .
it is also possible that there is some feii * absorption in this object between civ and ciii ] .
however , the wavelengths and strengths of those feii * transitions are not a good match to the apparent redshifted absorption in j1704 , which does match siiv , civ and aliii in wavelength . finally , for * sdss [email protected] * ( j2133 ; fig .
[ f_cands3 ] ) we conservatively adopt a systemic redshift of @xmath19=2.165 from our inspection of the spectrum ( the dr9q pca and mgii pca redshifts are @[email protected] ) .
j2133 has troughs at the adopted systemic redshift , plus candidate redshifted troughs , in civ , siiv , and nv . it is listed as a candidate primarily because the low signal - to - noise ratio of its spectrum makes both its systemic redshift and the reality of the putative redshifted troughs uncertain .
for example , while no dr9q ciii ] pca redshift is available for this object , its dr9q full pca redshift is @xmath19=2.096 , which if correct would make its troughs even more redshifted .
in this section we discuss a number of quasars which have apparent redshifted absorption features in their spectra which are not real . in figure [ f_candfe2 ]
we show spectra of quasars wherein absorption from excited - state feii ( feii * ) resembles redshifted high - ionization absorption . the quasar sdss j074850.40 + 442439.0 ( hereafter j0748 )
is shown in the top panel .
it is an felobal with apparent redshifted absorption in civ but not in siiv .
however , closer inspection of its spectrum from 1350 to 1750 ( middle panel ) reveals that the apparent redshifted civ absorption is part of a complex of feii * absorption with many strong transitions between civ and ciii ] but only one or two strong transitions between siiv and civ ( see 5.1.1 of hall et al . 2002 ) .
such feii * absorption can blend together to masquerade as a redshifted civ trough in other quasars such as sdss [email protected] and sdss 144424.54 + 013457.0 ( not shown ) and sdss j105442.44 + 060610.8 ( bottom panel ) . in that panel
the dotted lines show feii * absorption at the redshift of strongest absorption in other transitions ( dashed lines ) .
some of the feii * absorbs the civ emission line , weakening it relative to the siiv emission line , and some absorbs part of the continuum between rest - frame 1600 and 1750 . in this object
there is little evidence for additional feii * absorption between 2330 and 2630 which should be present , as it arises from the same lower levels as the observed absorption .
most likely , the absorption is present but is blended with feii emission so as to mimic a relatively smooth continuum .
feii and feiii emission may be at work in sdss [email protected] ( j2131 ; figure [ fj2131rej ] ) , which we considered , but ultimately rejected , as a redshifted - absorption candidate .
this object has a secure redshift from aliii+siiii]+ciii ] emission .
an extrapolation of the continuum longward of ciii ] to shorter wavelengths generates the appearance of strong , redshifted troughs of civ , siiv , and nv . however , the detailed shape and position of the putative troughs do not agree with that identification .
( e.g. , what looks like a moderately broad trough at 4700 in the smoothed spectrum is actually the blended siiv doublet of an intervening metal - line absorber at @xmath264 . ) therefore , the unusual feature in this spectrum is not redshifted absorption but the huge apparent emission feature from 17002300 .
this feature is somewhat reminiscent of that in sdss j033810.85 + 005617.6 ( 5.3.5 of hall et al .
2002 ) , but is likely due to some combination of blended feii and feiii emission and an extinction curve steeper than that of the smc ( 5.3 of hall et al .
2002 ) .
another source of apparent redshifted absorption features is unrecognized flaws in the data .
such flaws are a possibility in any pipeline , and spurious broad features have previously been seen in some sdss spectra ( see , e.g. , hall et al .
2008 ) .
experience has shown that spectra from certain sdss fibers , including # 374 , are more prone to spurious features .
for example , sdss [email protected] ( sdss plate - mjd - fiber 2022 - 53827 - 374 and boss 3973 - 55323 - 0536 ) had an apparent redshifted civ trough in sdss which was not confirmed by boss spectroscopy .
inspection of the individual spectroscopic exposures in the sdss revealed that the trough was indeed spurious .
the apparent trough is located in the wavelength range where the coverage of the red and blue arms of the sdss spectrographs overlap .
a decrease in flux at the trough wavelengths was present in all of the spectra from the blue arm but none of the spectra from the red arm .
the pixels from the blue arm were not rejected , resulting in a spurious trough .
other examples of spurious troughs involving sdss fiber # 374 include sdss [email protected] and sdss j171216.27 + 660212.0 . as another example of spurious features tied to certain fibers , the sdss spectrum of the bl lac object om 280 ( sdss j115019.21 + 241753.8 , plate - mjd - fiber 2510 - 53877 - 086 ) appears to show a trough at @xmath94150 .
this trough is spurious , likely arising from a region of problematic flatfielding ( e.g. , due to a localized variable feature in the flatfield ) in the ccd in the blue arm of sdss spectrograph # 1 .
this region is visible as a circular depression on the reduced 2-d spectrograms of that ccd in that wavelength range , centered near fiber 81 .
the affected wavelength region is interpolated over for fibers 7884 , but its effects can sometimes be seen in other nearby fibers such as # 86 ( used for the spectrum of om 280 ) , # 76 ( producing a similar spurious trough in sdss j115013.88 + 234602.1 ) , and possibly # 73 , in which case it could explain the apparent trough in the sdss spectrum of the bl lac object pks0138@xmath5097 ( shaohua et al .
2011 ) .
lastly , inspection of both spectroscopy and imaging can also identify cases of apparent redshifted absorption caused by superpositions .
for example , the star - quasar blend sdss j120749.19 + 082406.7 has an apparent redshifted aliii trough from the stellar mg b feature .
sher ( 1968 ) discuss the implications of the relativistic doppler effect in terms of the radial and total velocities of the source .
we present a similar discussion in terms of the total velocity of the gas and the angle of its motion relative to our line of sight .
for simplicity we consider an observer at the systemic redshift @xmath265 of the quasar ( but physically distant from it ) with the same line of sight to the quasar as us .
the only difference between such an observer and us is that the wavelengths seen on earth are all a factor of @xmath266 longer .
consider gas moving with speed @xmath140 at an angle @xmath267 to the observer s line of sight in the observer s frame , where @xmath268 and @xmath269 points directly toward the observer ( in other words , the observer looks in the direction @xmath270 to see the gas ) . in its own frame
, such gas will absorb at the rest wavelength of a given transition , @xmath271 . in the observer s frame , those absorbed photons have wavelength @xmath272 following foltz et al .
( 1986 ) , we rewrite this by defining @xmath273 where @xmath274 . because @xmath275 refers to motion toward the observer ,
a negative line - of - sight velocity @xmath276 indicates motion toward the observer .
redshifted absorption ( @xmath277 ) can be produced by gas moving at relatively low velocity more or less directly away from the observer ( @xmath278 ) .
it can also be produced by gas moving purely in the plane of the sky ( @xmath279 ) with @xmath280 .
that is the relativistic doppler effect : due to time dilation , gas moving transversely in the observer s frame absorbs at a lower frequency ( longer wavelength ) than stationary gas ( figure [ figrel ] ) .
as there is nothing singular about @xmath279 , redshifted absorption can also be produced by gas moving in directions @xmath275 ( toward the observer ) as long as @xmath47 is large enough that in the observer s frame the blueshifted velocity of the absorbing gas is counteracted by its increased time dilation .
absorption at the rest wavelength of a transition can be produced by gas at rest .
it can not be produced by gas moving away from the observer .
however , absorption at the rest wavelength can be produced by gas moving toward the observer at any @xmath48 , provided that the gas moves at an angle @xmath281 blueshifted absorption can only be produced by gas moving toward the observer ( @xmath282 ) .
for any given @xmath283 , there is a minimum @xmath267 at which gas can be moving and still produce absorption at @xmath284 : @xmath285 however , for any @xmath286 , there are _ two _ values of @xmath48 which yield absorption at the same @xmath287 . consider gas moving at a given @xmath286 and photons with @xmath288 .
for @xmath289 just below unity , relatively low - velocity gas ( for which time dilation is negligible ) can absorb those photons when its motion redshifts them to the rest wavelength of the transition in the absorbing gas frame .
extremely high - velocity gas can also absorb those photons because the highly blueshifted velocity of the absorbing gas in the observer s frame is counteracted by its large time dilation . for smaller @xmath289 ,
there continue to be two velocities of @xmath48 at which absorption will occur , but time dilation is not negligible for either of them . only for @xmath269
does this effect vanish . as a final aside
, we note that detections of relativistic quasar outflows via highly ionized x - ray absorption troughs , both broad and narrow , have been reported over the past decade ( e.g. , chartas et al .
2009 ; kaspi & behar 2010 ; tombesi et al .
while the significance of narrow x - ray absorption line detections has been questioned , in part due to their variability between different observations ( vaughan & uttley 2008 ) , strong variability would be natural for such high - velocity outflows and has been observed in mrk 766 ( risaliti et al .
furthermore , an additional potential source of variability is that if the outflow varies in @xmath218 even at fixed @xmath48 , @xmath287 will vary .
lastly , a given ratio of observed energy to rest energy ( @xmath290 ) for such an outflow yields a maximum angle it can have to the line of sight , which may be a useful constraint for models of such outflows . | we report the discovery in the sloan digital sky survey and the sdss - iii baryon oscillation spectroscopic survey of seventeen broad absorption line ( bal ) quasars with high - ionization troughs that include absorption redshifted relative to the quasar rest frame .
the redshifted troughs extend to velocities up to @xmath0 km s@xmath1 and the trough widths exceed 3000 km s@xmath1 in all but one case .
approximately 1 in 1000 bal quasars with blueshifted civ absorption also has redshifted civ absorption ; objects with civ absorption present only at redshifted velocities are roughly four times rarer . in more than half of our objects ,
redshifted absorption is seen in cii or aliii as well as civ , making low - ionization absorption at least ten times more common among bal quasars with redshifted troughs than among standard bal quasars . however , the civ absorption equivalent widths in our objects are on average smaller than those of standard bal quasars with low - ionization absorption .
we consider several possible ways of generating redshifted absorption .
the two most likely possibilities may be at work simultaneously , in the same objects or in different ones .
rotationally dominated outflows seen against a quasar s extended continuum source can produce redshifted and blueshifted absorption , but variability consistent with this scenario is seen in only one of the four objects with multiple spectra .
the infall of relatively dense and low - ionization gas to radii as small as 400 schwarzschild radii can in principle explain the observed range of trough profiles , but current models do not easily explain the origin and survival of such gas .
whatever the origin(s ) of the absorbing gas in these objects , it must be located at small radii to explain its large redshifted velocities , and thus offers a novel probe of the inner regions of quasars .
[ firstpage ] galaxies : nuclei - quasars : general - quasars : absorption lines |
a quantum system inevitably couples to the environmental degree of freedom , which forms an open quantum system @xcite .
it is of great significance to understand the influence of environment on the quantum system . as a paradigmatic model to study the open quantum systems ,
the spin - boson model has drawn persistent attentions @xcite . in the spin - boson model ,
the quantum system is simplified as a single spin ( qubit ) , while the environment is abstracted into a bosonic bath with an infinite number of modes .
the coupling between the qubit and the environment is characterized by a spectral function @xmath3 which is proportional to @xmath4 .
the spectral exponent @xmath5 varies the spin - boson model into three different types : sub - ohmic ( @xmath6 ) , ohmic ( @xmath7 ) , and super - ohmic ( @xmath8 ) . despite its simple form ,
there still exist great challenges to analyze the spin - boson model , due to the continuous bosonic bath which leads to an infinite number of degree of freedoms .
the difficulties are aggravated by the infrared divergence of the sub - ohmic and ohmic spin - boson model which has close relations with the quantum phase transition ( qpt ) @xcite .
many advanced numerical approaches have been applied to this model , such as the numerical renormalization group ( nrg ) @xcite , quantum monte carlo simulation ( qmc ) @xcite , sparse polynomial space approach @xcite , exact diagonalization in terms of shift boson @xcite , and variational matrix product state method @xcite .
it is generally accepted that there exists a second - order qpt for sub - ohmic baths and the kosterlitz - thouless qpt for the ohmic bath .
it has been argued for a long time that the qpt of the present quantum model is in the same universality class as the thermodynamic phase transition of the one - dimensional ising model with long - range interactions @xcite .
this quantum - to - classical correspondence was supported by most numerical approaches , but also questioned by a berry phase term emerged in the action @xcite .
most recently , we have developed a displaced fock state ( dfs ) method @xcite to analytically study the sub - ohmic spin - boson model and present evidence of the gaussian criticality persistent in the whole sub - ohmic bath regime .
the variational study based on the polaronic unitary transformation by silbey and harris @xcite has inspired a lot of studies by means of coherent states and various extensions @xcite .
zheng _ et al .
_ reproduced the results of silbey - harris ansatz by unitary transformation without variational procedures . the zeroth - order approximation in dfs method @xcite
can also recover the famous silbey - harris results .
a generalized silbey - harris ansatz was proposed by chin _ et al . _ which correctly describes a continuous transition with mean - field exponents for @xmath9 @xcite .
however , it failed to give reliable critical points for @xmath10 @xcite .
recently , the single coherent states ansatz @xcite was improved by simply adding other coherent states on the equal footing @xcite and by superpositions of two degenerate single coherent states @xcite , which are generally termed as multi - coherent - states ( mcs ) ansatz in this paper .
actually , the mcs in the single - mode version has been proposed ten years before by ren _
et al . _
@xcite independently .
et al . _
@xcite have studied the novel environmental entanglement and spin coherence in the ohmic spin - boson model by increasing the number of the coherent states without much more difficulties . very interestingly , the mcs ansatz was shown to have fast convergence and can give results with very high accuracy .
the variational study using mcs with unconstrained parameters have not been studied in the spin - boson model , which may hopefully shed light on the quantum criticality of the spin - boson model . among all single coherent ansatz , any correlations among bosons
are not included , so in principle the non - mean field exponent can not be given . while the correlations among phonons should be certainly embodied in the mcs ansatz .
recently , diagrammatic multiscale methods anchored around local approximations , where the particle correlations are self - consistently taken into account , indeed capture the well known non - mean - field nature of a lattice model @xcite .
so it is expected that mcs would give the precise description for the quantum criticality also , which motivate the present study for the qpt in both sub - ohmic and ohmic spin - boson model within the mcs ansatz without imposing any limits on the variational parameters .
the rest of paper is organized as follows . in sec .
ii , the spin - boson hamiltonian is introduced briefly .
the asymmetrical mcs ansatz is proposed in sec .
iii , and the self - consistent equations for the variational parameters are derived .
the numerical results for the order parameter and the entanglement entropy are presented and discussed in sec .
iv , and conclusions are given in the last section .
the spin - boson hamiltonian can be written as @xmath11 where @xmath12 ( @xmath13 ) are the pauli matrices , @xmath14 is the tunneling amplitude between the spin - up state @xmath15 and the spin - down state @xmath16 , @xmath17 ( @xmath18 ) is the bosonic annihilation ( creation ) operator which can ( create ) a boson with frequency @xmath19 , @xmath20 is the corresponding coupling strength between the qubit and the bosonic bath , which is determined by the spectral density @xmath3 , @xmath21 where @xmath22 is a dimensionless coupling constant , @xmath23 is the cutoff frequency which is set to be @xmath24 throughout this paper , @xmath25 is a step function .
the ground - state in the generalized silbey - harris ansatz @xcite can be written in the bases of spin - up state @xmath15 and spin - down state @xmath16 as @xmath26 b\exp \left [ \sum_{k=1}^{n_b}g_k\left ( b_k^{\dag } -b_k\right ) \right ] \end{array } \right ) , \ ] ] where @xmath27 ( @xmath28 ) is related to the occupation probabilities of spin - up ( spin - down ) state , while @xmath29 ( @xmath30 ) are the corresponding bosonic displacements of the @xmath31th mode .
it can be reduced to the original silbey - harris ansatz if set @xmath32 , and @xmath33 .
note that the nonlocal correlations among phonons are not included in this ansatz , so non - mean - field nature can not be described in this ansatz . in the recent analytic dfs method
@xcite , correlations for more than one phonon can be fully considered step by step .
this is a very clean and rigourous approach where the correlations among phonons are explicitly shown .
however , even the nearly converged results for the magnetic order parameter up to the third - order dfs still can not give the non - mean - field nature for @xmath34 . note that in the dfs , the number of the parameters for the self - consistent solutions , which are required in the discretization in the continuous integral , increase exponentially with the approximation order , so it becomes extremely difficult to explore the further corrections .
it is very interesting to note that the correlations among more phonons can be also included in the mcs ansatz @xcite .
more importantly , the number of variational parameters only increases with the number of the coherent states linearly , so the high order of approximations can be practically performed .
it is expected that the mcs can provide insights upon the nontrivial critical behavior of the spin - boson model .
a general form of the mcs in the bases of spin - up state @xmath15 and spin - down state @xmath16 can be written as @xmath35 |0\rangle \\ \sum_{n=1}^nb_n\exp \left [ \sum_{k=1}^{n_b}g_{n , k}\left ( b_k^{\dag } -b_k\right ) \right ] |0\rangle \end{array } \right ) , \label{gs_wave}\ ] ] where @xmath36 ( @xmath37 ) are related to the occupation probabilities of spin - up ( spin - down ) state , while @xmath38 ( @xmath39 ) are the corresponding bosonic displacements of the @xmath31th mode , and @xmath40 is the vacuum state of the bosonic bath .
@xmath41 is the number of the coherent states , and @xmath42 is the number of the discrete bosonic modes .
a generalized silbey - harris ansatz @xcite can thus be recovered if set @xmath43 .
the symmetric mcs ansatz ( @xmath44 and @xmath45 ) can only be applied to the delocalized phase . in the spin - boson model ,
since the qpt may occur with a symmetry breaking , the asymmetric wavefunction [ gs_wave ] ) should be generally employed .
the energy expectation value can be expressed as @xmath46 where @xmath47 with @xmath48 , \\
g_{m , n } & = & \exp \left [ -\frac 12\sum_k\left ( g_{m , k}-g_{n , k}\right ) ^2\right ] , \\
\gamma _ { m , n } & = & \exp \left [ -\frac 12\sum_k\left ( f_{m , k}-g_{n , k}\right ) ^2\right ] ,
\\ \alpha _ { m , n } & = & \sum_k\left [ \omega _ kf_{m , k}f_{n , k}+\frac{\lambda _
k}% 2\left ( f_{m , k}+f_{n , k}\right ) \right ] , \\
\beta _ { m , n } & = & \sum_k\left [ \omega _ kg_{m , k}g_{n , k}-\frac{\lambda _
k}% 2\left ( g_{m , k}+g_{n , k}\right ) \right ] .\end{aligned}\ ] ] the parameters @xmath49 , @xmath50 , @xmath51 , and @xmath52 are determined by minimizing the energy expectation value @xmath53 with respect to the variational parameters .
the total number of variational parameters is @xmath54 . for @xmath36 and @xmath37
, we have @xmath55 and for @xmath38 and @xmath39 , we obtain @xmath56 \}=0 , \\ & & \sum_n\{-\delta \gamma _ { n , i}a_nf_{n , k } \nonumber \\ & & + b_ng_{i , n}\left [ 2\left ( \beta _ { i , n}+\omega _ k - e\right ) g_{n , k}-\lambda _ k\right ] \}=0 . \label{gg}\end{aligned}\ ] ] in practise , these parameters can be obtained by solving the coupled equations self - consistently , which in turn give the gs energy and wavefunction .
generally , the largest number of the coherent states in the practical calculations can be reached up to @xmath57 in this paper . in the dfs study
@xcite , we have rigourously proved that all summation over @xmath31 is related to @xmath58 and can be transformed into the continuous integral @xmath59to be free from any approximation except for the finite number of the mcs , the exact summation over @xmath31 should be performed .
we here use the gaussian - logarithmical discretization developed in ref .
@xcite to calculate the continuous spectral numerically exactly by checking the convergence .
the widely used logarithmical discretization usually overestimate the critical points @xcite . as a function of @xmath60 for various numbers of coherent states .
the inset is a enlarged view for @xmath61 and @xmath62 curves . @xmath63 and @xmath64 .
, width=302 ] in order to check the validity of the mcs ansatz with unconstrained the parameters , we compare the ground state energy calculated by eq .
( [ ehd ] ) for the different number of coherent states ( @xmath41 ) .
we present the relative difference @xmath65 as a function of the coupling strength @xmath22 in fig .
[ energy ] for @xmath66 and @xmath67 .
it is shown that @xmath68 becomes smaller rapidly with increasing @xmath41 .
@xmath69 is always less than @xmath70 in the whole coupling regime , demonstrating that an excellent convergence is achieved for @xmath57 . in other words
, @xmath57 is sufficient large to describe the ground - state in this model . in the next section
, we will study the criticality of the spin - boson model with the mcs by using the numerically exact gaussian - logarithmical integrations .
it is well - known that the magnetization @xmath71 acts as an order parameter in the sub - ohmic spin - boson model , which can be written as @xmath72 figure .
[ pz ] shows the value of @xmath71 as a function of @xmath22 for different numbers of coherent states .
it is obvious that there exists a critical coupling strength @xmath73 which separates the delocalized phase with zero magnetization from the localized one with nonzero magnetization . in the weak coupling regime ,
the magnetization @xmath71 is always zero , indicating that the spin stay in the spin - up and the spin - down states with the same probability . with increasing coupling strength
, @xmath71 will tend to @xmath74 ( due to the degeneracy , another branch will tend to @xmath75 symmetrically ) .
the critical coupling strength @xmath73 increases with the value of @xmath5 , as shown in fig .
[ pz ] . for same spectral exponent @xmath5 , the obtained @xmath76 increases with the number of coherent state .
the ohmic case was studied by the symmetric mcs ansatz for @xmath77 in ref .
it should be noted that the symmetric mcs is invalid in the localized phase . as a function of @xmath60 for @xmath78 , @xmath79 , @xmath80 , and @xmath24 with @xmath81 , and @xmath82 .
@xmath67.,width=302 ] we then pay particular attention to @xmath66 , which is typical value greater than @xmath79 . as shown in fig .
[ s075 ] , a fast convergence for the magnetization can be achieved with increasing number of the coherent states . in principle , the true critical coupling strength @xmath73 should only be obtained by the infinite number of the coherent states in the mcs , which definitely can not be realized in the practical calculations .
it also becomes extremely difficult to solve the self - consistent equations [ [ aa]]-[[gg ] ] for an unbounded large number of coherent states , especially near critical points .
fortunately , a perfect linear scaling @xmath73 as a function of the inverse coherent state number @xmath0 is shown in the inset of fig .
[ s075 ] . in this way
, @xmath73 is extrapolated to @xmath83 when @xmath84 , i.e. @xmath85 . very interestingly , this value for @xmath73 is consistent excellently with @xmath86 obtained by qmc @xcite , the only numerical method where the discretization of the bath is not needed in literature , to the best of our knowledge . as a function of @xmath22 with seven numbers of coherent states for @xmath66 and @xmath87 .
inset shows the critical coupling strength @xmath73 as a function of the inverse coherent state number @xmath0 .
a very nice linear fitting yields the extrapolated limiting value @xmath88 as @xmath85 . ,
width=302 ] the magnetization shows a power law behavior near the critical point , namely @xmath89 .
the classical counterpart of the sub - ohmic spin - boson model is the one - dimensional ising model with long - range interaction according to the quantum - to - classical mapping @xcite .
it is predicted that a continuous phase transition with mean - field behavior undergoes for @xmath9 and non - mean - field behavior for @xmath90 .
most previous numerical approaches demonstrated the validity of the quantum - to - classical mapping .
surprisingly , the dfs method @xcite gives the robust mean - field exponent @xmath91 in the whole sub - ohmic regime @xmath92 , which is in sharp contrast with the previous conclusions .
therefore , the critical behavior of the sub - ohmic spin - boson model needs further extensive direct studies where the berry phase and topological effects are not missed @xcite .
the studies based on the feynman path - integral representation of the partition function may not satisfy this requirement . as a function of @xmath93 for @xmath66 and @xmath67 with the number of coherent state @xmath43 ( square ) , @xmath94 ( cycle ) and @xmath95 ( triangle ) .
the power law curves with @xmath96 is denoted by the dashed line .
, width=302 ] ( a ) and @xmath7 ( b).,width=566 ] figure . [ criticale ] displays the log - log plot of magnetization as a function of @xmath97 for @xmath66 for @xmath98 , and @xmath95 .
it is surprising to observe that all curves show a very nice power - law behavior over more than @xmath99 decades with an exponent @xmath100 .
it is strongly suggested that even @xmath101 coherent states can hardly modify the exponent @xmath102 . in other words , the number of coherent state @xmath41 has negligible effect on the critical exponent @xmath103 , unlike the critical coupling strength @xmath73 which is quite sensitive on @xmath41 .
so here we provide another piece of evidence for the breakdown of the quantum - to - classical mapping for @xmath2 , besides the dfs study @xcite .
now we move to the ohmic bath case ( @xmath7 ) . to demonstrate the difference between the @xmath7 and @xmath6 baths , we study the entanglement entropy between the qubit and the bath .
in the spin - boson model , entanglement entropy can be obtained as @xcite @xmath104 where @xmath105 in the exact nrg study @xcite , @xmath106 for @xmath107 .
so in the delocalized phase , the entailment entropy should converge to @xmath108 at the strong coupling , which is @xmath109 for @xmath67 .
the entanglement entropy as a function of the coupling strength is given in fig .
[ entanglement ] for @xmath66 and @xmath110at @xmath67 .
we observe that the entanglement entropy exhibits a cusp for @xmath6 , and shows a very steep drop for @xmath7 at the critical points .
with increase of the number of the coherent states , the entanglement entropy jump more steeply at the critical point .
interestingly , the entanglement entropy curves for @xmath7 resemble very much curves for unbiased case in fig . 4 of ref .
@xcite by the bethe ansatz study at extremely low temperature .
it is expected that the entanglement entropy will vertically jump to zero in the limit of @xmath111 .
this may be a signature of the kosterlitz - thouless type qpt from a second - order qpt .
the kosterlitz - thouless phase transition , of infinite order , for the ohmic bath ( @xmath7 ) @xcite , can be only reached in the infinite number of coherent states .
such a sudden jump is not , but perhaps related to the sudden jump of the superfluid density in thin films of @xmath112 described by the two - dimensional xy model @xcite , which is worthy of a further study .
in this work , by means of the asymmetric mcs ansatz , we extensively analyze the ground state of the spin - boson model , especially the quantum criticality .
this direct study to the quantum model does not miss the berry phase or topological defects . without any artificial approximations except for the finite number of the coherent states , we find that the magnetic exponent @xmath113 at the critical points is robust in the whole sub - ohmic bath regime , consistent with the recent dfs study and in contrast with most numerical studies in literature .
it is strongly suggested that the well - known quantum - to - classical mapping is broken down for @xmath2 . the asymptotical behavior in the infinite number of the coherent states are also analyzed .
the converged critical strengths for @xmath114 agrees well with the qmc formulated on feynman path - integral representation of the partition function . for @xmath7 , more steep jump of the entanglement entropy at the critical point for the larger number @xmath41 of the coherent states
is observed .
it is expected that the vertical jump of the entanglement entropy would occur at the kosterlitz - thouless phase transition points in the large @xmath115 limit .
this work is supported by national natural science foundation of china under grant no .
11474256 , and national basic research program of china under grant no .
@xmath116 corresponding author .
z. g. lu and h. zheng , phys .
b. * 75 * , 054302(2007 ) ; q. wang , a. y. hu , and h. zheng , phys .
rev . b. * 80 * , 214301(2009 ) ; q .-
tong , j .- h . an , h .-
luo , and c. h. oh , phys . rev .
b * 84 * , 174301 ( 2011 ) . | a variational approach based on the multi - coherent - state ansatz with asymmetric parameters is employed to study the ground state of the spin - boson model . without any artificial approximations except for the finite number of the coherent states , we find the robust gaussian critical behavior in the whole sub - ohmic bath regime . the converged critical coupling strength can be estimated with the @xmath0 scaling , where @xmath1 is the number of the coherent states .
it is strongly demonstrated the breakdown of the well - known quantum - to - classical mapping for @xmath2 .
in addition , the entanglement entropy displays more steep jump around the critical points for the ohmic bath than the sub - ohmic bath . |
historic light curves of some bright bl lac objects have shown that fast luminosity fluctuations ( typical of this class of agns ) are frequently superinposed on long - term trends of relatively large amplitude .
the origin of these long - term changes is not fully understood .
an interesting hypothesis is that they can be related to changes of the jet orientation due , for instance , to the presence of a massive black hole binary system in the nucleus of the object ( abraham and carrara 1998 ; lehto and valtonen 1996 ) .
we expect , therefore , that radio jets associated with these sources should exibit large structural changes , detectable in vlbi images . to this aim we started to observe some bl lac objects , which are intensively monitored in the optical , with the european vlbi network ( evn ) .
one of them is on 231 ( w com , 1219 + 285 ) , discovered by wolf ( 1916 ) as a variable star and identified as the counterpart of a radio source by biraud ( 1971 ) and browne ( 1971 ) . the historic light curve ( tosti et al .
1999a ) , although with gaps before 1970 , shows that the mean source luminosity had a minimum in the early seventies and since that epoch has increased to reach its highest level in the spring of 1998 , when the source also had a very bright outburst ( massaro et al .
1999 ) .
the vlbi structure of on 231 has been known for almost twenty years . the first 5 ghz image ( observation epoch 1982.25 )
was presented by weistrop et al .
( 1985 ) and showed a core , emitting about one third of the total flux , and a multicomponent jet elongated in position angle ( pa ) @xmath0110@xmath1 .
the same structure was confirmed by subsequent observations ( gabuzda et al .
1992 , gabuzda et al .
1994 , gabuzda & cawthorne 1996 ) . a recent image at 15ghz can be found in kellermann et al .
( 1998 ) in which the jet can be tracked up to about 10 mas from the nucleus . the innermost structure at 22 ghz , with a resolution of about 0.35 mas , can be seen in the three vlba maps covering the period from 1995.31 to 1997.58 presented by jorstad et al .
radio images of on 231 on the arcsecond scale ( de bruyn & schilizzi 1984 , kollgaard et al .
1992 , perlman & stocke 1994 ) show emission from an extended faint region directed southwest ( centred at pa @xmath0310@xmath1 ) with respect to the core .
we used the evn@xmath2merlin facility to observe on 231 at 1.6 ghz and at 5 ghz . in this paper
we present images made at the two frequencies which show relevant new features in the source structure compared to the previous published images .
we propose some possible scenarios to relate these changes with the long - term luminosity evolution .
on 231 is one of the few bl lac objects whose variations have been known for about one century , altough before the 1970s it was observed only occasionally . for the past six years
it has been intensively monitored , in particular by the groups of perugia and rome .
our photometric data up to 1998 have already been published together with those of other collaborating groups in several papers ( tosti et al .
1998 , tosti et al .
1999b , massaro et al .
additional recent data of the 1999 observational season have been obtained with the automatic telescope of the perugia observatory and the reflector of the vallinfreda astronomical station , near rome .
observational equipment and data reduction procedures are described in detail in the above references .
the up - to - date historic light curve is reported in the upper panel of fig.1 .
the sparse observations of on 231 before 1970 show magnitude variations generally in the interval @xmath3 .
it had a minimum around 1975 and after a few years a brightening phase began .
this brightening trend is clearer if the data are averaged over time intervals longer than the typical fluctuation duration . in fig.1 ( lower panel )
we plotted the mean magnitude computed over three year intervals : the brightening trend after 1975 is evident despite the gap from 1990 to 1995 .
the luminosity evolution in the period 19941999 is shown in detail by the light curve in the r band of fig .
apart from the time intervals in which on 231 is not visible because of its proximity to the sun , this light curve shows that the source behaviour after 1995 was characterized by a series of major bursts , at least one per year with the present time sampling , on which rapid variations are superinposed .
these events had a typical time - scale of about five - six months and flux changes of a factor of about 3 . in spring 1998 , on 231 had an exceptional outburst and reached the highest recorded brightness level .
in this episode the energy spectral index in the optical band hardened from 1.4 to 0.5 at the maximum and the linear polarisation increased up to about 10 % with a pa in the range of 130@xmath1160@xmath1 ( massaro et al .
1999 ) , as expected in the case of synchrotron radiation from freshly accelerated electrons . after the burst ,
the mean source luminosity decreased and was comparable to that observed in the previous years , but with variations on shorter time scales .
radio light curves at various frequencies in the range from 4.8ghz to 22ghz ( aller et al . 1999 ,
tersranta et al . 1998 ) show a behaviour not correlated with that in the optical .
fig.3 gives an up - to - date version of the multiband light curve to the end of 1999 taken from the university of michigan radio astronomy observatory database .
on 231 showed a general decreasing trend of the radio luminosity starting from about 1980 .
after a four year long flare with a peak in 1992 , the source flux started to increase again in 1995 .
the comparison with the optical light curve , despite the different behaviour in the first portion , indicates that in 1995 on 231 likely entered a new phase in which both the optical and radio flux levels started to increase .
recall that our present knowledge about possible correlations between the long - term behaviour of bl lac objects in the radio and optical bands is rather poor because of the small number of studied sources and of the irregular time coverage .
it is therefore hard to reach a firm conclusion regarding whether the increase of the radio and optical fluxes of on 231 after 1995 is just a coincidence or it has a real origin .
the latter possibility can be adopted as a working hypothesis , and can be used to formulate some scenarios for the evolution of the jet geometry , as we will discuss in the last section .
on231 was observed with the evn plus merlin on 05 june 1998 at 1.6ghz and on 17 february 1999 at 5ghz .
the data were recorded with the markiiia terminals at the stations in mode b ( 28mhz bandwidth ) . during each experiment ,
the source was tracked for about 11 hours together with the calibration sources oq208 and 3c286 , observed for 13minutes scans , regularly spaced over the experiment .
the vlbi data were correlated at the markiiia correlator of the max - planck - institut fr radioastronomie ( bonn , germany ) .
the raw data , output from the correlator , were integrated for 4 seconds . of the 11 stations of the evn , namely jodrell bank and cambridge ( uk ) ,
westerbork ( nl ) , effelsberg ( d ) , medicina and noto ( i ) , onsala ( s ) , sheshan ( shanghai , china ) , nanshan ( urumqi , china ) , torun ( pl ) , simeiz ( ukraine ) , nine gave fringes at 1.6ghz ( all but sheshan and torun ) . at 5ghz effelsberg
could not observe because of snow and simeiz could not take part in the observations . at the jodrell bank observatory ,
the lovell telescope was used at 1.6ghz and the mark2 telescope at 5ghz .
the correlator output was calibrated in amplitude and phase using @xmath4 is the nrao s _ astronomical image processing system _ ] and imaged using difmap ( shepherd et al .
total power measurements taken at the same times as the vlbi observations and the gain curve of the telescopes were applied in the amplitude calibration process .
the self - calibration procedure , which uses closure amplitudes to determine telescope amplitude corrections , gave calibration factors that were within 10% of unity for all the telescopes but cambridge , simeiz and torun , which were in the range 15 - 20% .
the source was also tracked by merlin .
initial values for the telescope and correlator gains were determined from a short observation of a bright , unresolved calibration source .
the primary calibrator 3c 286 was scheduled during the observations in order to fix the absolute flux density scale .
images from the merlin lhc polarization data were produced using @xmath4 .
an @xmath4 analysis path for processing joint evn@xmath2merlin data was then followed and the relative images were made using difmap .
the complex structure of on 231 is evident from the two full - resolution vlbi images of figures 4 and 5 at 1.6ghz and 5ghz respectively .
the source shows a straight collimated east - west jet at pa @xmath0 110@xmath1 which bends twice by @xmath5 , in a spiral - like structure .
this extension south of the jet was not present in previously published images , but it was partially detected by observations with the vlba at 2.3ghz in january 1997,285 are available in the radio reference frame image database ( rrfid ) of the united states naval observatory ( usno ) .
http://maia.usno.navy.mil/rorf/rrfid.html ] included in the radio reference frame image database ( rrfid ) .
charlot , sol and vicente ( 1997 ) obtained a 5ghz image with evn observations in february 1994 .
the achieved mean beam size of about 20 mas was not adequate to resolve the jet but good enough to show the presence of a component south - east of the core , which is likely the jet extention seen in our images .
the jet shows a multicomponent structure and the most relevant difference with respect to literature images ( gabuzda et al .
1992 , 1994 , gabuzda & cawthorne 1996 ) is that the brightest component is no longer at the western extreme of the source , but rather in a central location . in fact
, we clearly detect an extension on the western side of the brightest component at both frequencies .
such an extension is also visible on a larger scale .
the phase referenced merlin image at 5 ghz shows that the source is resolved by the interferometer at a 50 mas resolution scale . in fig .
6 we can see an extension of the brightness distribution on both sides of the central bright component , along the source major axis at pa@xmath0110@xmath1 , plus an extension on the south - eastern part of the source .
we performed several tests to ensure that the western extension detected on the mas scale ( component c0 in the 5 ghz evn image ) is real . in the process of producing the hybrid radio images at both frequencies ,
the self - calibration was done , windowing out the clean components from the western emission .
the images we obtained were worse than those presented in the paper , showing a poor fit between model and visibilities on the longest baselines .
in particular , the amplitude gain corrections required for the two chinese stations ( which allowed high resolution in the east - west direction ) were of the order of @xmath025% at 5 ghz . in the normal process , i.e. cleaning on the emission from the western area persisting in the residual map , the corrections were @xmath6 , comparable to those applied to the other stations and in agreement with the corrections required when self - calibrating the data set of the calibrator source oq 208 .
we note that the images were made usinge both the analysis packages @xmath4 and difmap , obtaining an almost identical structure for on 231 .
these results raise the question of a proper identification of the source core .
two hypotheses can be considered : @xmath7 the core is the brightest component and we observed the development of a new component on the opposite side of the jet , @xmath8 the core is located at the western extreme of the jet and we observed an exceptionally bright intermediate component . to tackle the problem , we produced two further images by combining evn and merlin data at both frequencies .
the images were made with the same resolution and were used to map the distribution of the spectral index ( @xmath9 ) , reported in grey scale in fig .
7 , superposed on the contour map at 5ghz .
the uncertainties in the phase centre locations in the two images ( lost in processing the data ) , are at the pixel level and can slightly affect the spectral index estimate .
of course , we assumed that the mean flux of the various components did not change dramatically in the eight months between the observations , as expected from the behaviour at these intermediate radio frequencies , shown in fig.3 .
the measured overall flux densities at both frequencies in the merlin images and evn@xmath2merlin images agree to better than 1% .
we measured total fluxes of 903.3 mjy and 755.5 mjy at 1.6ghz and at 5ghz respectively , corresponding to an overall spectral index of 0.16 .
the only component showing a flat spectrum ( @xmath10 ) is the brightest one , labelled @xmath11 in fig .
we point out that the value for the east - southern rim could be affected by the relatively low signal to noise ratio .
the spectral index values of the other , more prominent components , labelled @xmath12 , @xmath13 and @xmath14 , are 0.60 , 0.57 and 0.89 , respectively .
this result suggests that the core is likely the brightest component in @xmath11 , i.e. component @xmath15 in fig . 5 , rather than @xmath16 at the western edge of @xmath11 , although we can not rule out the latter hypothesis .
we performed a test to confirm the identification of @xmath15 with the core . making use of the @xmath4 task jmfit
, we evaluated the angular distances between the main jet components in the low resolution evn+merlin images .
one would expect that if the core is actually located at the western edge of @xmath11 , the angular distance from the bright , isolated , component @xmath13 in fig.7 ( corresponding to @xmath17 in fig5 ) and the core component @xmath11 , which is a blend of @xmath18 and @xmath19 , should be greater in the image at 5ghz than in the image at 1.6ghz .
in fact , in the former case , @xmath16 , supposed to be the core , should contribute much more than @xmath19 ( almost equal in flux density ) to the total flux density of @xmath11 due to its flat or inverted spectral index between 1.6ghz and 5ghz , producing a shift of the centre of the gaussian component(s ) fitting @xmath11 . to check this
we have tried to fit the brightness distribution of the main part of the jet region with a three - gaussian component model , leaving their central positions as free parameters .
the position of the most eastern component @xmath13 was then taken as reference point .
in the 1.6ghz case we found that the centres of the two gaussians other than @xmath13 , were located in pa@xmath20 at distances of 10.4 and 12.6 mas respectively . the solution for the brightness distribution at 5ghz failed to converge to the three gaussian components of the input model .
the output was a two component model separated by 10.4 mas each other , equal to the separation found for a pair of components in the previous solution .
the above solutions were found to be a stable convergence of the model fit process achieved by jmfit in both cases , and suggest @xmath15 as the core of on231 .
a further point in favour of this suggestion comes from the analysis of the behaviour of the spectral indices of the individual components @xmath18 and @xmath19 , which give a total flux density of 454mjy at 5ghz for @xmath11 ( from the model in table 1 ) and conspire to give a flat spectrum to it .
let us take @xmath16 as the core component in on231 with an inverted spectral index of + 2.5 , as expected for a synchrotron self - absorbed source , and a spectral index @xmath21 for @xmath19 .
@xmath15 will then contribute for about 260mjy to the total flux density of @xmath11 at 1.6ghz , implying it has an almost flat spectral index , rather unusual for a jet component .
a more normal inverted spectral index for @xmath16 , say @xmath22 , makes the spectral index of the two other components even flatter .
all these results indicate that the core is likely the brightest component @xmath15 in @xmath11 rather than @xmath12 at the western edge , even though we can not rule out the alternative hypothesis .
we will further discuss this point in section 4.1 .
the total jet extent along the axis in our images is @xmath23 15 mas , while it is shorter in the previous images , ranging from about 9 to 11 mas .
a weak component at 14 mas is visible in the 8.4ghz image by gabuzda & cawthorne ( 1996 ) and a much farther one , at about 25 mas along the jet axis at pa @xmath0 110@xmath1 , was reported by gabuzda et al .
we found no evidence for emission at such long distances from the source centre .
we stress that according to the core identification given above , the extension of the eastern jet would reduce to about 13mas , closer to the previous estimates .
the occurrence of various emitting components in the jet of on 231 has been recognized from early vlbi images .
weistrop et al . (
1985 ) reported the detection of at least three components .
gabuzda et al .
( 1992 ) detected four components , later increased to eight / nine by gabuzda et al .
( 1994 ) and gabuzda & cawthorne ( 1996 ) . from these observations , which span a time interval of about three years
, the authors estimated component velocities @xmath24 in the range 0.62 2.42 , increasing with the distance from the core .
jorstad et al .
( 2001 ) at 22ghz found a component moving at 1.5 2.3 in the inner part of the jet .
we present in fig.5 the most resolved source structure at 5ghz that can be modelled with five gaussian components , with a good formal agreement between data and model , with @xmath25 quite close to unity . in table1 , together with the flux of each component , we report the angular distances from the centre of the brightest component @xmath15 , the corresponding pas , the major axis , the axial ratios and the orientation of the best fitted gaussians . the component @xmath15 is the one giving the major contribution to the flux of the flat spectrum component @xmath11 in the map of fig.7 . taking into account that both components are in the same position despite the different resolution , we can identify @xmath15 as the core of on 231 .
the model parameters of table 1 fit well the visibilities of the longest baselines at 1.6ghz , while it does not account for the total flux density on the short baselines .
* table1 - modelfitting 5ghz data set * [ cols="^,>,^,>,^,^ , > " , ] @xmath26 : components are ordered from west to east in the following , we derived some estimates for the source orientation with respect to the line of sight and for the intrinsic plasma speed based on a comparison of our images and previously published data used to derive proper motion estimates .
we also investigate the possibility that the steep - spectrum emission to the west of @xmath15 is indicative of the counter - jet .
for our purposes we used the full resolution 5ghz image reported in fig.5 and the components from the model fitting given in table1 .
it is not easy to support a proper association between any of the components in our model and those from observations made roughly 10 years ago by gabuzda et al .
the brightness profile along the jet in our images is quite different from those resulting from previous observations , as shown in fig.8 .
note that the peak flux of the component @xmath15 , which we identify as the core of the radio emission , decreased from 250 mjy in the early observations by gabuzda et al .
( 1992 ) to 170 mjy in our data with a ratio in agreement with the total flux decrease in this time interval ( fig.3 ) .
a plausible association could be made between our component @xmath17 and the @xmath27 of gabuzda et al .
( 1994 ) . according to their study ,
this component moved eastward at @xmath0 0.37 mas / year in 1.88 years . assuming that its velocity remained constant in the following ten years , it should now be at about 11.3 mas from the core , compatible with the distance of @xmath17 from the component @xmath15 .
a proper motion @xmath28 0.37 mas / year corresponds to @xmath29 1.66 for h@xmath30 = 100 km s@xmath31 mpc@xmath31 and q@xmath30=0.5 ( same cosmology as in gabuzda et al .
1994 ) . using the standard formulae given in pearson & zensus ( 1987 )
, the derived apparent superluminal speed implies a minimum value for the intrinsic plasma speed @xmath32 0.84 , which can be obtained if the viewing angle @xmath33 .
the maximum allowed angle is @xmath0 62@xmath1 , which is a large value for bl lac , even considering that it is an upper limit .
let us now test whether it is plausible that the steep - spectrum emission to the west of @xmath15 is a counter - jet ; i.e. the doppler de - boosted jet that is receding from the observer . if we assume that the intrinsic source structure is symmetrical , an estimate of @xmath34 and @xmath35 can be derived .
if @xmath15 is the core , in order to estimate the jet to counter - jet brightness ratio @xmath36 , we should compare the flux of the component @xmath16 to the sum of the fluxes from @xmath19 , @xmath17 and @xmath37 .
this leads to a ratio @xmath38 , corresponding to @xmath39 cos@xmath35 @xmath0 0.11 . as indicated above ,
based on the observed apparent superluminal motion , @xmath40 0.84 ; this implies a @xmath41 value of @xmath42 0.13 , or a minimum viewing angle of @xmath43 82@xmath1 .
this value is far too large to be plausible ; in other words , the emission to the west of @xmath15 is far too bright to be a counter - jet . we discuss alternative interpretations of this emission below
the intensive optical monitoring of some bright bl lac objects shows the existence of long - term trends , which can be related to structure modifications of the source .
our recent vlbi images of on 231 were obtained during the very bright and active phase in the optical beginning in 1995 , following the twenty year increase in the mean luminosity , and show significant changes with respect to several previous vlbi images taken since 1987 . as already discussed in sect .
4 we identified the source core with the brightest component on the basis of its flat spectrum .
the most important consequence of this result is that we detected a new component , not visible in older images , at pa @xmath070@xmath1 , opposite to the main jet .
the absence of this component in the subsequent 22 ghz map ( epoch 1997.58 ) by jorstad et al .
( 2001 ) could be due to its rather steep spectrum .
we stress , however , that those images indicate a brightening of a component at about 2 mas east of the core whose flux density remained the strongest .
parsec - scale jets of bl lac objects are generally one - sided , and this is explained by the strong doppler boosting of the jet pointing to the observer line of sight and by the corresponding deboosting of the jet at the opposite direction . in a few cases , however , indication for double jet and complex structures have been reported in the literature , like those found in 1308 + 326 by pyatumina , aller & aller ( 1999 ) , who analysed a series of 8 ghz images from the geodetic vlbi data base . in particular , they noticed the appearance of components in different directions .
furthermore , jet structures in bl lac sources are frequently not very well aligned and show wiggles , suggesting either a strong interaction with the surrounding medium or some kind of motion , regular ( e.g. precession ) or not , of the jet itself .
we can consider a few possible scenarios for interpreting both the time evolution of the optical luminosity and the changes of the jet structure of on 231 on the parsec scale .
a first possibility is that the jet , pointing very close to the observing direction , undergoes strong instabilities and oscillations , and therefore the new western component is the result of one episodic large amplitude displacement .
again , emitting blobs from the nuclear source can be distributed around a mean direction of the motion and one of them could be displaced toward west at an angle greater than that of other blobs .
future vlbi images will then show the evolution of this unique component , while the others will continue to move along the former direction . it will be interesting to verify if the occurrence of new very bright optical outbursts are related to the presence of components on the west side ( or other directions ) of the core .
another possible scenario is that of a slowly precessing jet approaching the observer s line of sight over the past few decades .
the progressive increase of the beaming factor would then be responsible for the mean brightening optical trend shown by this source , and the minimum angular distance was likely reached in 1997 - 98 when the source was observed at the highest level . after this phase , the apparent jet direction changed to the opposite side as in our images .
a more detailed model , which takes into account the evolution of the emitting electron population and the radiation transfer , is necessary to explain the absence of a positive correlation between the radio and optical light curves , but it is beyond the aim of the present paper . in this framework , we can expect that in the next five or ten years a further development of the western jet and a fading of the eastern jet will be oserved . the absence of radio emission on the north side of the jet axis , even on the arsecond scale ( kollgaard et al .
1992 , perlman & stocke 1994 ) , indicates that the precession cone axis lies south - southwest respect to the line of sight . if the above core identification is not correct and the core corresponds to the weak component at the western extreme , it is necessary to understand why a jet component achieved a spectrum flatter than the core itself .
a possible explanation is that the jet suffered a strong interaction with the surrounding matter and changed its direction , as in the recent model proposed by pohl & schlickaiser ( 2000 ) .
a relativistic hadronic blast wave is emitted by the central engine and a radiation burst is produced when it interacts with near dense clouds .
synchrotron emitting electrons can be originated from the decay of charged pions produced in the nuclear collisions between the blast wave particles and those of the cloud .
we expect , therefore , that gamma rays having energies greater than about 50 mev are copiously produced in such an event .
on 231 is known to have been detected by egret ( mukherjee et al .
1997 ) , but unfortunately no data after 1995.5 are available .
the extension of the jet after the bend is clearly detected in our images for the first time .
since this is not a new emitting region , it was not detected in previous images due to sensitivity limits . in the above scenario of a precessing jet , it could be the trace of the jet itself , rotating clockwise from south to north .
a second possibility is that the bending of the jet is the consequence of a strong interaction with the near environment .
finally , we would like to stress the relevance of a coordinated study of multifrequency variability with vlbi mapping , particularly for bl lac objects like on 231 which have the synchrotron peak in the near ir - optical band .
our data suggest that the occurrence of a long - term trend in the optical luminosity and of periods of enhanced activity could be related to changes in the innermost radio structure
. a better understanding of these phenomenon requires both an intense optical monitoring and vlbi mapping , spaced not longer than one year apart .
such a program can be realised only for a small sample of sources and on 231 should be one of the primary targets .
the roma and perugia groups acknowledge the financial support by the italian ministry for university and research ( murst ) under the grant cofin 98 - 02 - 32 .
we acknowledge support from european commission s tmr - lsf programme , contract no .
this research has made use of the data from the university of michigan radio astronomy observatory which is supported by funds from the university of michigan .
this research has made use of the united states naval observatory ( usno ) radio reference frame image database ( rrfid ) . | recent radio images of the bl lac object on 231 ( w com , 1219 + 285 ) show remarkable new features in the source structure compared to those previously published .
the images were obtained from observations made with the european vlbi network plus merlin at 1.6 ghz and 5 ghz after the exceptional optical outburst occurred in spring 1998 .
the up - to - date b band historic light curve of on 231 is also presented together with the r band luminosity evolution in the period 19941999 .
we identify the source core in the radio images with the brightest component having the flattest spectrum .
a consequence of this assumption is the existence of a two sided emission in on 231 not detected in previous vlbi images .
a further new feature is a large bend in the jet at about 10 mas from the core .
the emission extends for about 20 mas after the bend , which might be due to strong interaction with the environment surrounding the nucleus .
we suggest some possible interpretations to relate the changes in the source structure with the optical and radio flux density variation in the frame of the unification model . |
what features of an edge - weighted tree identify it uniquely up to isomorphism , perhaps within some class of such trees ? here an _ edge - weighted tree _ is a connected , acyclic finite graph @xmath0 with vertex set @xmath8 and edge set @xmath9 which is equipped with a function @xmath10 .
the value of @xmath11 for an edge @xmath12 is called the _ weight _ or the _ length _ of @xmath13 .
two such trees @xmath14 and @xmath15 are isomorphic if there is a bijection @xmath16 such that : * @xmath17 if and only if @xmath18 , * @xmath19 for all @xmath17 .
the question above is , more formally , one of asking for a given class of edge - weighted trees @xmath20 about the possible sets @xmath21 and functions @xmath22 such that for all @xmath23 we have @xmath24 if and only if @xmath14 and @xmath15 are isomorphic .
if the class @xmath20 consists of edge - weighted trees for which all edges have length @xmath25 ( we will call such objects _ combinatorial trees _ for the sake of emphasis ) , then determining whether two trees in @xmath20 are isomorphic is just a particular case of the standard graph isomorphism problem .
the general graph isomorphism problem has been the subject of a large amount of work in combinatorics and computer science
@xcite already speaks of the `` graph isomorphism disease '' and , in particular , there are many results on reconstructing the isomorphism type of a graph from the isomorphism types of subgraphs of various sorts ( see , for example , the review @xcite ) .
there is also a substantial volume of somewhat parallel research on graph isomorphism in computational chemistry ( see , for example , @xcite for a review ) .
there seems to be considerably less work on determining isomorphism ( in the obvious sense ) of edge - weighted graphs ; of course , in order for two edge - weighted graphs to be isomorphic the underlying combinatorial graphs must be isomorphic , but this does not imply that the best way for checking that two edge - weighted graphs are isomorphic proceeds by first determining whether the underlying combinatorial graphs are isomorphic and then somehow testing whether some isomorphism of the combinatorial graphs is still an isomorphism when the edge - weights are considered .
we begin with a discussion of previous results that address various aspects of the problem of determining when two edge - weighted or combinatorial trees are isomorphic .
a result in @xcite gives the following criterion for a bijection @xmath16 , where @xmath14 and @xmath15 are combinatorial trees , to be an isomorphism : + if @xmath26 is any sequence from @xmath27 such that @xmath28 and @xmath29 then @xmath30 .
the above result is elegant , but , of course , one does not need to apply it to all possible bijections to determine whether two combinatorial trees are isomorphic : there is a much more explicit and efficient procedure , which we now describe for the sake of completeness . first of all , suppose that @xmath14 and @xmath15 have distinguished vertices @xmath31 and @xmath32 and , in addition to the requirements in the above definition of an isomorphism @xmath33 , we require that @xmath33 maps @xmath31 to @xmath32 ; that is , we have rooted trees and we require that an isomorphism maps the root of one tree to the root of the other .
the presence of a root allows us to think of a combinatorial tree as a directed graph , where the head of an edge is the vertex that is closer to the root and the tail is the vertex farther from the root .
the children of a vertex are the adjacent vertices that are farther from the root and , more generally , the descendants of a vertex @xmath34 are those vertices @xmath35 such that the path from the root to @xmath35 passes through @xmath34 .
the subtree spanned by a vertex @xmath34 and its descendants contains no other vertices and can be thought of as a combinatorial tree rooted at @xmath34 , and we call this subtree the subtree below @xmath34 .
then , two rooted , combinatorial trees @xmath14 and @xmath15 are isomorphic if the two roots have the same number of children , say @xmath36 , and there is an ordering of these children for each tree such that the subtree below the @xmath37 child of the root of @xmath14 is isomorphic ( as a rooted , combinatorial tree ) to the subtree below the @xmath37 child of the root of @xmath15 .
this observation can be turned into an efficient algorithm ( see , for example , ( * ? ? ?
* example 3.2 ) ) .
now , two combinatorial trees are isomorphic if there is some choice of roots such the resulting rooted , combinatorial trees are isomorphic .
a _ center _ of a combinatorial tree is a vertex @xmath38 such that @xmath39 where @xmath40 is the number of edges in he unique path between @xmath34 and @xmath35 for @xmath41 , and a combinatorial tree has either a unique center or two centers that are adjacent .
it is therefore possible to determine if two combinatorial trees are isomorphic by rooting each of them at their various centers and checking if any two such rooted , combinatorial trees are isomorphic . we , however , are interested in whether there are `` statistics '' of a more numerical character that can be used to decide tree isomorphism . for combinatorial trees ,
one somewhat obvious possibility is the multiset of eigenvalues of some matrix associated with the tree such as the adjacency matrix or the distance matrix .
unfortunately , the results of @xcite show that not only is the isomorphism type of a tree not uniquely determined by the spectrum of its adjacency matrix but for various ensembles of combinatorial trees if one picks a tree uniformly at random from those in the ensemble with @xmath1 vertices , then the probability there is another tree in the ensemble with an adjacency matrix that has the same spectrum converges to one as @xmath42 .
the results of @xcite can be used to show that an analogous phenomenon is present when one considers the spectrum of the matrix of leaf - to - leaf distances .
two trees have adjacency matrices with the same spectrum if and only if the characteristic polynomials of the adjacency matrices are equal . given some irreducible representation of the symmetric group on the number of letters equal to the dimension of a square matrix , the immanantal polynomial of the matrix is constructed in the same manner as the characteristic polynomial except that the determinant is replaced by a similarly defined object for which the sign character is replaced by the character of the representation .
one might hope that the immanantal polynomials are more successful at deciding isomorphism of combinatorial trees , but a result of @xcite shows that if the adjacency matrices of two combinatorial trees have the same characteristic polynomials , then they have the same immanantal polynomials for every irreducible representation .
we note that @xcite already contains an example of two combinatorial trees with adjacency matrices that are explicitly shown to have the same immanantal polynomial .
the greedoid tutte polynomial of a combinatorial tree @xmath0 encodes for each @xmath43 and @xmath44 the number of subtrees of @xmath0 that have @xmath43 internal vertices and @xmath44 leaves .
it was conjectured in @xcite that this information identifies the isomorphism type of a combinatorial tree . however , it was shown in @xcite that there are infinitely many pairs of nonisomorphic caterpillars that share the same greedoid tutte polynomial : a caterpillar is a combinatorial tree that consists of some number of internal vertices along a single path and leaves that are each adjacent to one of the internal vertices .
this contrasts with the situation for rooted , combinatorial trees ; it is shown in @xcite that there is a two - variable polynomial defined for all rooted , directed graphs ( and hence , in particular , for rooted , combinatorial trees ) such that two rooted , combinatorial trees have the same polynomial if and only if they are isomorphic .
the polynomial in @xcite is defined recursively , but it is not hard to see that it encodes in a compact manner the total number of vertices in the tree , the number of children of the root , the number of vertices in each of the subtrees below the children of the root , and so on .
the chromatic symmetric function of a graph was introduced in @xcite .
a proper coloring of a finite graph is a function @xmath45 from the vertices of the graph to @xmath46 such that adjacent vertices are assigned different values .
we can introduce an equivalence relation on the proper colorings by declaring that two colorings @xmath47 and @xmath48 are equivalent if there is a bijection @xmath49 such that @xmath50 . for a graph with @xmath36 vertices , each equivalence class gives rise to a partition @xmath51 of @xmath36 by taking , for any @xmath45 in the equivalence class , @xmath52 to be the @xmath37 largest of the cardinalities @xmath53 as @xmath54 ranges over @xmath46 .
the chromatic symmetric function encodes for each partition of @xmath36 the number of equivalence classes of colorings that give rise to that partition .
it was conjectured in @xcite that nonisomorphic combinatorial trees have distinct chromatic symmetric functions .
it was shown in @xcite that this conjecture is true for caterpillars and that paper also reports on computational results verifying that the conjecture holds for the class of trees with at most @xmath55 vertices .
further work related to the conjecture for the special case of trees with a single centroid is contained in @xcite .
our point of departure in this paper is the well - known fact @xcite that an edge - weighted tree can be reconstructed from its matrix of leaf - to - leaf distances ( see @xcite for an indication of the importance of this observation in the statistical reconstruction of phylogenetic trees ) .
in fact , an edge - weighted tree with @xmath1 leaves can be reconstructed from the collection of total lengths of subtrees spanned by all subsets of @xmath36 leaves provided @xmath56 @xcite .
we remark that the total length of the subtree spanned by a set of leaves is an important quantity in phylogenetics where it is called the phylogenetic diversity of the corresponding set of taxa @xcite . given these results
, one might imagine that the multiset of leaf - to - leaf distances suffices to identify the isomorphism type of an edge - weighted tree .
this is certainly not the case .
for example , consider the two combinatorial caterpillars @xmath14 and @xmath15 with @xmath57 leaves each , where @xmath14 has @xmath58 internal vertices @xmath59 in order along a path that are adjacent respectively to @xmath60 leaves , and @xmath15 has @xmath58 internal vertices @xmath61 in order along a path that are adjacent respectively to @xmath62 leaves .
taking the @xmath63 pairs of distinct leaves in @xmath14 , we see that the distance @xmath64 appears @xmath65 times , the distance @xmath58 appears @xmath66 times , and the distance @xmath67 appears @xmath68 times . similarly , taking the @xmath63 pairs of distinct leaves in @xmath15
, we see that the distance @xmath64 appears @xmath69 times , the distance @xmath58 appears @xmath70 times , and the distance @xmath67 appears @xmath71 times .
probabilistically , we have just shown that if we pick two leaves uniformly at random without replacement from an edge - weighted tree , then the isomorphism type of the tree is not uniquely identified by the probability distribution of the distance between the two leaves . note in this last example that if we looked at the multisets of lengths of subtrees spanned by three leaves , then we would see the length @xmath58 appearing @xmath72 times for @xmath14 and @xmath73 times for @xmath15 , and hence the probability distribution of the length of the subtree spanned by three leaves chosen uniformly at random is not the same for the two trees . in order to proceed further
, we need to introduce some more notation .
write @xmath74 for the set of leaves of an edge - weighted tree @xmath0 . given a subset @xmath75 of @xmath74 ,
let @xmath76 be the length of the subtree spanned by @xmath75 ; that is , @xmath76 is the sum of the lengths of the edges in the smallest connected subgraph of @xmath0 with a vertex set that contains @xmath75 .
it is possible to calculate the total length of @xmath0 , that is , @xmath77 , using the following result from @xcite that extends one for the special case of @xmath58-valent trees in @xcite .
write @xmath78 for the degree of an interior vertex @xmath35 of @xmath0 ( that is , @xmath79 ) .
for distinct leaves @xmath80 denote by @xmath81 the set of interior vertices on the unique path in @xmath0 between @xmath82 and @xmath83 and put @xmath84 let @xmath85 be the sum of the lengths of the edges in the path between @xmath82 and @xmath83 .
then , @xmath86 of course , a similar formula gives @xmath76 for any @xmath87 ; the path between a pair of leaves of the subtree is the same as the path between them in @xmath0 , the length of this path is the same in the subtree as it is in @xmath0 , but the degree of an interior vertex of the subtree can be less than its degree as an interior vertex of @xmath0 .
suppose that @xmath88 and @xmath89 is a uniformly distributed random listing of @xmath74 ; that is , @xmath89 is the result of sampling the leaves of @xmath0 uniformly at random without replacement .
set @xmath90 for @xmath3 ; that is , the random variable @xmath2 is the length of the subtree spanned by the first @xmath4 of the randomly chosen leaves .
we write @xmath91 for the @xmath92-dimensional random vector @xmath5 and call this random vector the _ random length sequence _ of @xmath0 . in this paper
we address the following question . can we reconstruct the edge - weighted tree @xmath0 up to isomorphism from the joint probability distribution of the random length sequence @xmath93 ?
[ q : main ] another way of framing this question is the following .
write @xmath94 for the leaves of @xmath0 and let @xmath95 be the multiset with cardinality @xmath96 that results from listing the @xmath92-dimensional vectors @xmath97 as @xmath98 ranges of the permutations of @xmath99:=\{1,\ldots , n\}$ ] .
we stress that @xmath95 is a multiset ; that is , we do not know which increasing sequences of lengths go with which ordered listings of the leaves .
can we reconstruct the edge - weighted tree @xmath0 up to isomorphism from the multiset of length sequences @xmath95 ?
we end this section with some remarks about the problem of reconstructing trees from various so - called _ decks _ , as this subject has some similarities to the questions we consider . in @xcite
, ulam asked whether it is possible to reconstruct the isomorphism type of a graph with at least @xmath58 vertices from the isomorphism types of the subgraphs obtained by deleting each of the vertices .
this question was resolved in the affirmative for combinatorial trees in @xcite .
moreover , later results established that it is not necessary to know the forests obtained by deleting every vertex .
for example , it was shown in @xcite that it suffices to know the subtrees obtained by deleting leaves .
this latter result was strengthened in @xcite , where it was found that it is only necessary to know which nonisomorphic forests are obtained and not what the multiplicity of each isomorphism type is , and in @xcite , where it was shown that it suffices to take only those leaves @xmath100 that are _ peripheral _ in the sense that @xmath101 along the same lines , it was established in @xcite that it is enough to take only the nonleaf vertices , provided that there are at least three of them .
the line of inquiry in @xcite is the most similar to ours : an example was presented of two trees for which the respective sets of vertices may be paired up in such a way that for each pair the sizes of the trees in the forests produced by removing each element of the pair from its tree are the same , and a necessary and sufficient condition was given for a tree to be uniquely reconstructible from this sort of data , which the authors of @xcite call the _ number deck _ of the tree .
we will answer question [ q : main ] in the affirmative for a few different classes of trees .
some classes will have general edge - weights and some classes will be combinatorial trees .
it is clear that in the case of general edge - weights we must restrict to trees that have no vertices with degree @xmath64 because otherwise we can subdivide any edge into arbitrarily many edges with the same total length and the joint probability distribution of the random length sequence will be unchanged see .
we call such trees _ simple_. the terms irreducible or homomorphically irreducible are also used in the literature .
our first result is for the the class of _ stars _ ; that is , edge - weighted trees with @xmath102 leaves that have a single interior vertex .
note that such trees are simple . for any edge - weighted tree with @xmath1 leaves
, @xmath103 is a constant ( the total length of the tree ) and @xmath104 is a uniformly distributed random pick from the lengths of the @xmath1 edges that are adjacent to one of the leaves .
the following simple result is immediate from this observation .
for @xmath102 the isomorphism type of a star is uniquely determined by the joint probability distribution of its random length sequence .
[ thrm : main star ] the simple trees with two leaves all consist of a single edge and have a random length sequence @xmath105 , where @xmath106 is the length of that edge , and so the isomorphism type of such a tree is uniquely determined by the joint probability distribution of its random length sequence .
the simple trees with three leaves are stars , and it follows from that the isomorphism type of such a tree is uniquely determined by the the joint probability distribution of its random length sequence .
we next consider simple , edge - weighted trees with four leaves . for @xmath107 ,
the isomorphism type of a simple , edge - weighted tree @xmath0 with @xmath1 leaves is uniquely determined by the joint probability distribution of its random length sequence .
[ thrm : main n4 ] the proof of this result is via consideration of possible cases .
similar proofs could be attempted for larger numbers of leaves , but the main reason we include the result is to show how such a proof for even a small number of leaves leads to quite a few cases and because we will need the case of four leaves later .
it is well - known that any simple , combinatorial tree with labeled leaves can be reconstructed from the simple , combinatorial trees spanned by each subset of four leaves ( the so - called quartets ) ( * ? ? ?
* theorem 6.3.7 ) . with this and in mind
, one might imagine that the isomorphism type of simple , edge - weighted tree can be determined from the joint probability distribution of @xmath108 . however , putting such a strategy into practice would seem to be rather complicated because there can be two sets of leaves @xmath109 and @xmath110 such that @xmath111 but @xmath112 , @xmath113 , and @xmath114 .
one way of ruling out such annoying algebraic coincidences is to assume that the edge - weighted tree @xmath0 has _ edge - weights in general position _ , by which we mean that the sums of the lengths of any two different ( not necessarily disjoint ) subsets of edges of @xmath0 are not equal .
the isomorphism type of a simple , edge - weighted tree @xmath0 with edge - weights in general position is uniquely determined by the joint probability distribution of its random length sequence .
[ thrm : main generalposition ] the last family of edge - weighted trees with general edge - weights whose elements we can identify up to isomorphism from the joint probability distributions of their random length sequences is the class of _ ultrametric _ trees . for the sake of completeness
, we now define this class .
recall that for leaves @xmath115 we denote by @xmath116 the distance between them ; that is , @xmath117 is the sum of the lengths of the edges on the unique path between @xmath43 and @xmath54 .
the edge - weighted tree @xmath0 is ultrametric if for any leaves @xmath118 we have @xmath119 from which it follows that for any leaves @xmath118 at least two of the distances @xmath116 , @xmath120 , and @xmath121 are equal while the third is no greater than that common value .
equivalently , an edge - weighted tree @xmath0 is ultrametric if , when it is thought of as a real tree ( that is , a metric space where the edges are treated as real intervals of varying lengths given by their edge - weights see , for example , @xcite ) , then there is a ( unique ) point @xmath122 called the root ( which may be in the interior of an edge ) such that the distance from @xmath122 to a leaf is the same for all leaves .
we will make use of both definitions .
it is immediate from the former definition that the subtree of an ultrametric tree spanned of a subset of leaves is itself ultrametric .
the isomorphism type of an ultrametric , simple , edge - weighted tree @xmath0 is uniquely determined by the joint probability distribution of its random length sequence .
[ thrm : main ultrametric ] the proof of establishes an even stronger result .
namely , the isomorphism type of an ultrametric , simple , edge - weighted tree @xmath0 is uniquely determined by the minimal element of @xmath95 in the lexicographic order .
[ rem : ultrametric min enough ] we call attention to a subtle point in the statements of and .
both results say that if we are given the joint probability distribution of the random length sequence of an edge - weighted tree @xmath0 information that certainly includes the number of leaves of @xmath0 and we know , a priori , that @xmath0 has a certain extra property ( edge - weights in general position or ultrametricity ) , then we can determine the isomorphism type of @xmath0 .
the theorems do not , however , say whether it is possible to determine from the joint probability distribution of its random length sequence whether a simple , edge - weighted tree @xmath0 has its edge - weights in general position or is ultrametric .
we do not have results that settle this question , but we say some more about it in and believe it is an interesting area for future research .
[ rem : ultrametric determination ] observe that if @xmath0 is an edge - weighted tree , @xmath123 is any vertex of @xmath0 , and @xmath38 is a constant such that @xmath124 , then @xmath125 defined by @xmath126 and @xmath127 is an ultrametric on @xmath74 that arises from suitable edge - weights on @xmath0 . the metric @xmath128 is often called the _ farris transform _ of @xmath129
see @xcite for a review of the many appearances of this object in various areas from phylogenetics to metric geometry .
it might be hoped that an affirmative answer to for general edge - weighted trees will follow from .
however , we have been unable to find an argument which shows that the joint probability distribution of the random length sequence of the tree @xmath0 equipped with the new edge - weights is determined by the joint probability distribution of the random length sequence for the original edge - weights .
suppose that @xmath0 is a rooted , simple combinatorial tree with root @xmath122 .
we can define a partial order on @xmath8 by declaring that that @xmath130 if @xmath82 is on the unique path from @xmath122 to @xmath83 .
two vertices @xmath131 have a unique greatest lower bound in this partial order that we write as @xmath132 and call the _ most recent common ancestor _ of @xmath82 and @xmath83 .
the map @xmath133 defined by @xmath134 is an ultrametric on @xmath74 and hence it arises from a collection of edge - weights @xmath135 on @xmath0 .
a directed edge @xmath136 in @xmath0 with @xmath130 is necessarily of the form @xmath137 and @xmath138 for some @xmath139 .
if @xmath140 is the corresponding undirected edge , then @xmath141 therefore , if @xmath14 is a subtree of @xmath0 spanned by some set of leaves @xmath87 and @xmath142 is the set of directed edges of @xmath14 , then we have that the length of @xmath14 is @xmath143 the following result is immediate from and . the isomorphism type of a simple , combinatorial tree @xmath0 is uniquely determined by the minimal element of the set @xmath95 of length sequences obtained after designating a root for @xmath0 and equipping @xmath0 with the edge - weights @xmath135 .
we now turn our focus to combinatorial trees and drop the assumption of simplicity .
that is , all edge - weights are equal to one and there may be vertices with degree two .
we answer question [ q : main ] in the affirmative for two families of combinatorial trees .
( black vertices).,scaledwidth=40.0% ] first , a combinatorial tree @xmath0 is a _ caterpillar _ if the deletion of the leaves along with the edges adjacent to them results in a path with @xmath144 vertices ( and hence @xmath44 edges ) see , for example , . choose some direction for the path and number from @xmath145 to @xmath44 the vertices on the path encountered successively in that direction and write @xmath146 for the number of leaves adjacent to the vertex numbered @xmath43 .
note that @xmath147 and @xmath148 .
two sequences @xmath149 and @xmath150 correspond to isomorphic trees if and only if @xmath151 , say , and either @xmath152 , @xmath153 or @xmath154 , @xmath153 .
the isomorphism type of a caterpillar is uniquely determined by the joint probability distribution of its random length sequence .
furthermore , it is possible to determine from the joint probability distribution of the random length sequence of a combinatorial tree whether the tree is a caterpillar .
[ thrm : main caterpillar ] our final results are for the classes of ( unrooted ) _ @xmath6-valent _ and _ rooted @xmath4-ary _ combinatorial trees . for @xmath155 ,
a @xmath6-valent combinatorial tree is a combinatorial tree for which all vertices have degree either @xmath156 ( the internal vertices ) or @xmath25 ( the leaves ) . for @xmath155 ,
a rooted @xmath4-ary combinatorial tree is a combinatorial tree for which one internal vertex ( the root ) has degree @xmath4 and the remaining internal vertices have degree @xmath156 ; the leaves , of course , have degree @xmath25 .
when @xmath157 we refer to a rooted @xmath64-ary combinatorial tree as a rooted _ binary _ combinatorial tree . attaching an extra vertex via and edge to the root of a rooted @xmath4-ary tree produces a @xmath6-valent combinatorial tree . the isomorphism type of a @xmath6-valent combinatorial tree ( respectively , a @xmath4-ary tree ) is uniquely determined by the joint probability distribution of its random length sequence .
[ thrm : main kary1 ] in fact , our proof of leads us to a stronger conclusion .
fix @xmath158 .
let @xmath159 be a random @xmath6-valent combinatorial tree ( respectively , a random @xmath4-ary combinatorial tree ) with @xmath1 leaves .
then , the probability distribution of the isomorphism type of @xmath159 is uniquely determined by the joint probability distribution of its random length sequence .
[ thrm : main kary2 ] note that in there are two sources of randomness in the construction of the random length sequence : we first choose a realization of the random @xmath159 and then take an independent uniform random listing of the leaves to build the increasing sequence of subtrees and their lengths .
the rest of the paper consists primarily of proofs of the above results in the order we have presented them . in we briefly discuss further open questions related to question [ q : main ] .
we begin by looking at question [ q : main ] for edge - weighted trees with a small number of leaves and give a proof of that answers question [ q : main ] in the affirmative for general , simple edge - weighted trees with @xmath161 or @xmath67 leaves .
the case of for simple trees with @xmath162 leaves is trivial , as all such trees have two leaves and one edge , @xmath163 in this case , and @xmath106 is the length of the edge . the case of @xmath164 leaves is only slightly more complicated ,
as all such trees are star - shaped .
thus , determining @xmath0 from @xmath93 consists of determining its three edge weights .
these can be inferred easily from @xmath93 by looking at the distribution of @xmath165 , which , since @xmath166 is constant ( equal to the total length of @xmath0 ) , is distributed as a uniform random choice from the three edge weights .
finally , we give a proof of in the case when @xmath160 . for @xmath160 leaves ,
there are two possible simple combinatorial trees , and hence two possibilities for the shape of @xmath0 .
the first is the star - shaped tree with four edges and one interior vertex .
the second is the @xmath58-valent tree with two interior vertices and one interior edge .
leaves.,scaledwidth=40.0% ] to determine which possibility @xmath0 is , we first look at the distribution of @xmath167 to find the lengths of the four edges connecting directly to the four leaves . call these edges _
pendent_. if the sum of the four pendent edge lengths equals @xmath168 , then @xmath0 is star shaped and we have determined @xmath0 up to isomorphism . if not , then @xmath0 is @xmath58-valent and the difference between @xmath168 and the sum of the pendent edge lengths is the length @xmath13 of the interior edge .
all that is left to determine @xmath0 up to isomorphism in this second case is determining how the pendent edges pair on each side of the interior edge .
first , if the multiset of the lengths of pendent edges is of the form @xmath169 or @xmath170 , then @xmath0 is already uniquely determined .
next , if the multiset is of the form @xmath171 , then we need to distinguish between the case where the leaves with pendent edges of length @xmath123 are siblings ( and thus so are the leaves with pendent edge length @xmath172 ) and the case where leaves with pendent edge lengths @xmath123 and @xmath172 are paired . in the former case the possible values of @xmath106 are @xmath173 with respective probabilities @xmath174 , whereas in the latter case the possible values of @xmath106 are @xmath175 with respective probabilities @xmath176 , and we can certainly distinguish between the two cases
. if the multiset of pendent edge lengths is of the form @xmath177 , then there are the following two possibilities : * the two leaves with pendent edge length @xmath123 are siblings and the two leaves with pendant edge lengths @xmath172 and @xmath38 are siblings , in which case the possible values of @xmath106 are @xmath178 with respective probabilities @xmath179 ; * a leaf with pendent edge length @xmath123 is the sibling of the one with pendent edge length @xmath172 and the other leaf with pendent edge length @xmath123 is the sibling of the one with pendent edge length @xmath38 , in which case the possible values of @xmath106 are @xmath180 with respective probabilities @xmath181 .
suppose without loss of generality that @xmath182 . if @xmath183 , then @xmath184 is @xmath185 for ( p1 ) and @xmath145 for ( p2 ) . if @xmath186 or @xmath187 , then @xmath188 is @xmath189 for ( p1 ) and @xmath185 for ( p2 ) . in all cases we can distinguish between ( p1 ) and ( p2 ) . finally ,
if the multiset of pendent edge lengths is of the form @xmath190 , then there are the following two possibilities : * the leaf with pendent edge length @xmath123 is paired with the one with pendent edge length @xmath172 and the leaf with pendent edge length @xmath38 is paired with the one with pendent edge length @xmath191 , in which case the possible values of @xmath106 are @xmath192 with common probability @xmath185 ; * the leaf with pendent edge length @xmath123 is paired with the one with pendent edge length @xmath38 and the leaf with pendent edge length @xmath172 is paired with the one with pendent edge length @xmath191 , in which case the possible values of @xmath106 are @xmath193 with common probability @xmath185 ; * the leaf with pendent edge length @xmath123 is paired with the one with pendent edge length @xmath191 and the leaf with pendent edge length @xmath172 is paired with the one with pendent edge length @xmath38 , in which case the possible values of @xmath106 are @xmath194 with common probability @xmath185 ; suppose without loss of generality that @xmath195
. then possibility ( p3 ) holds if and only if @xmath196 and possibility ( p5 ) holds if and only if @xmath197 and @xmath198 , so we can distinguish between ( p3 ) , ( p4 ) and ( p5 )
. the argument in the proof of seems rather _ ad hoc _ and it does not suggest a systematic approach to obtaining the analogous result for trees with an arbitrary numbers of leaves .
the number of simple combinatorial trees with @xmath1 leaves grows so rapidly with @xmath1 ( see , for example , @xcite ) that even for trees with a relatively small fixed number of leaves a case - by - case argument seems rather forbidding .
nonetheless , we do conjecture that an affirmative answer to question [ q : main ] holds more generally .
recall that the edge - weights of a simple , edge - weighted tree @xmath0 are in general position if the sum of the lengths of any two distinct subset of edges of @xmath0 are not equal . by assumption , if @xmath199 and @xmath200 are two subsets of @xmath74 such that @xmath201 , then @xmath202 .
consequently , if @xmath199 and @xmath200 are two subsets of @xmath74 such that @xmath203 for @xmath204 , then @xmath205 and @xmath206 for @xmath207 .
recall that @xmath89 are the successive randomly chosen leaves used in the construction of @xmath208 . because @xmath104 is the length of the pendent edge attaching @xmath209 to the rest of @xmath0
, it follows that the set @xmath210 has @xmath1 elements and @xmath211 for each @xmath212 .
there are at least two leaves of @xmath0 that are siblings , and so there exist @xmath213 such that @xmath214 .
fix such a pair of lengths and write @xmath215 and @xmath216 for the ( unique ) leaves of @xmath0 with pendent edges having respective lengths @xmath217 and @xmath218 .
we have @xmath219 , and the event @xmath220 coincides with the event @xmath221 . by assumption ,
the set @xmath222 has @xmath223 elements and @xmath224 for each @xmath225 .
index the values of @xmath226 as @xmath227 and write @xmath228 , @xmath229 , for the unique leaf of @xmath0 that is distance @xmath230 from the unique vertex of @xmath0 that is adjacent to both of the sibling leaves @xmath215 and @xmath216 .
we will show that it is possible to determine the leaf - to - leaf distances @xmath231 , @xmath232 . as we recalled in the introduction
, this information uniquely identifies the isomorphism type of @xmath0 .
again by assumption , the set @xmath233 has @xmath234 elements and @xmath235 for each @xmath236 . for a given @xmath236
there is a unique ordered pair @xmath237 , @xmath238 , and a unique @xmath239 such that @xmath240 and @xmath241 in which case the two conditional probabilities in question are both @xmath242 .
moreover , every ordered pair @xmath237 , @xmath238 , corresponds to some unique @xmath236 and @xmath239 in this way .
the event @xmath243 coincides with the event @xmath244 and the event @xmath245 coincides with the event @xmath246 .
considering the subtree of @xmath0 spanned by @xmath247 and ignoring the vertices with degree two to produce a simple tree , the leaves @xmath248 and @xmath249 are siblings in this simple tree ( as are @xmath215 and @xmath216 ) , and the quantity @xmath13 is the distance between the vertex in the subtree to which @xmath248 and @xmath249 are adjacent and the vertex to which @xmath215 and @xmath216 are adjacent ; the lengths of the pendent edges connecting @xmath248 and @xmath249 to the rest of the subtree are @xmath250 and @xmath251 .
thus , if the ordered pair @xmath237 corresponds to @xmath236 and @xmath239 , then , recalling the notation @xmath129 for the path length distance in @xmath0 , @xmath252 , @xmath253 , @xmath254 , @xmath255 , @xmath256 , and @xmath257 .
therefore , the joint probability distribution the random length sequence @xmath91 uniquely determines the matrix of leaf - to - leaf distances in @xmath0 and hence the isomorphism type of @xmath0 .
recall that @xmath95 is the set of sequences @xmath258 such that @xmath259 .
write @xmath260 for the usual _ lexicographic _ total order on @xmath95 ( that is @xmath261 if in the first coordinate where the two sequences differ the entry of the @xmath217 is smaller than the entry of @xmath218 ) . equivalently , @xmath261 if either @xmath262 or @xmath263 and for the smallest @xmath7 such that @xmath264 we have @xmath265 . in this section
we prove by showing that that the tree @xmath0 is determined up to isomorphism by the minimal element of @xmath95 .
we use a similar technique ( but with a different total order ) to establish for @xmath156-valent and rooted @xmath4-ary combinatorial trees in .
let @xmath266 be the minimal element of @xmath95 .
write @xmath267 for an ordering of @xmath74 such that @xmath268 for @xmath269
. we will establish by induction that for @xmath270 the ultrametric real tree spanned by the leaves @xmath271 can be reconstructed from @xmath272 and , moreover , if we adopt the convention that we draw ultrametric real trees in the plane with the root at the top and leaves along the bottom , then this particular real tree can be embedded in the plane with the leaves @xmath273 in order from left to right .
the claim is certainly true when @xmath274 .
suppose the claim is true for @xmath275 .
write @xmath276 for the ultrametric real tree spanned by @xmath271 and denote the height of @xmath276 by @xmath277 ; that is , @xmath277 is the common distance from each of the leaves of @xmath276 to the root @xmath278 of @xmath276 .
we can , of course , suppose that @xmath279 .
if @xmath280 , then the ultrametric real tree @xmath281 spanned by @xmath282 must consist of an arc of length @xmath283 from the root @xmath284 of @xmath281 to the leaf @xmath285 and an arc of length @xmath286 from `` new root '' @xmath284 to the `` old root '' @xmath278 . in this case
we can , by the inductive hypothesis , certainly embed @xmath281 in the plane with the leaf @xmath285 to the right of the leaves @xmath273 .
assume , therefore , that @xmath287 .
then the ultrametric real tree @xmath281 must consist of @xmath276 and an arc of length @xmath288 joining @xmath285 to a point @xmath289 .
it will suffice to show that @xmath83 must be on the arc @xmath290 $ ] that connects @xmath278 to @xmath228 because there is a unique ultrametric real tree consisting of @xmath276 and an arc of length @xmath288 joining a new leaf to a point on the arc @xmath290 $ ] ( this tree must have root @xmath278 and the point where the arc of length @xmath288 attaches to @xmath290 $ ] must be at distance @xmath291 from @xmath278 ) and , moreover , such a tree can be embedded in the plane with the new leaf to the right of the leaves @xmath271 .
suppose , then , that @xmath83 is not on the arc @xmath290 $ ] .
let @xmath54 be the maximum of the indices @xmath292 such that @xmath83 is on the arc connecting @xmath248 to @xmath278 .
write @xmath34 for the point that is closest to @xmath293 in the subtree spanned by @xmath294 and @xmath278 .
write @xmath35 for the point that is closest to @xmath293 in the subtree spanned by @xmath294 .
equivalently , @xmath35 is the point in the subtree spanned by @xmath294 that is closest to @xmath34 .
we may , of course , have @xmath295 ( which occurs if and only if @xmath296 ) . by the inductive hypothesis , @xmath34 and @xmath35
are on the arc connecting @xmath249 to @xmath278 and @xmath297 by construction , @xmath83 is the point closest to @xmath285 in the subtree spanned by @xmath294 and @xmath278 .
write @xmath298 for the point closest to @xmath285 in the subtree spanned by @xmath294 .
equivalently , @xmath298 is the point in the subtree spanned by @xmath294 that is closest to @xmath83 .
we have @xmath299 by the definition of @xmath54 , the points @xmath83 and @xmath34 are on the arc connecting @xmath249 to @xmath278 and @xmath300 .
this implies that @xmath301 .
it also implies , by ultrametricity , that @xmath302 consequently , @xmath303 this , however , contradicts the minimality of @xmath258 . as we noted in
, it is interesting to know whether it is possible to determine from the joint probability distribution of the random length sequence whether an edge - weighted tree is ultrametric .
the preceding proof of contains a procedure for reconstructing @xmath0 from the minimal element of @xmath95 in the lexicographic order when @xmath0 is an ultrametric tree .
if @xmath0 is an arbitrary edge - weighted tree and this procedure is applied to the minimal element of @xmath95 in the lexicographic order , then it will still produce an ultrametric tree and so a necessary condition for @xmath0 to be ultrametric is that the joint probability distribution of the random length sequence of this ultrametric tree coincides with the joint probability distribution of @xmath91 . along the same lines ,
suppose that @xmath0 is an arbitrary edge - weighted tree and , thinking of @xmath0 as a real tree , we root it at the unique point @xmath122 such that @xmath304 then @xmath122 will have @xmath4 children for some @xmath4 .
let @xmath305 , @xmath306 , be the number of leaves @xmath35 in the subtree below the @xmath37 child of @xmath122 such that @xmath307 .
it is clear that @xmath0 is ultrametric if and only if @xmath308 .
let @xmath309 be a listing of the nonzero terms in the list @xmath310 .
note for @xmath311 that @xmath312 and @xmath313 thus , the joint probability distribution of @xmath91 determines @xmath44 and the values of the elementary symmetric polynomials of degrees @xmath311 evaluated at @xmath309 , and we want to know whether @xmath314 , the value of the elementary symmetric polynomial of degree @xmath25 evaluated at @xmath309 , is @xmath1 .
the elementary symmetric polynomials of degrees @xmath315 in @xmath44 real variables are algebraically independent over the reals , and so we can not expect to recover @xmath314 from the values of the other elementary symmetric polynomials . however , there are inequalities connecting the values of the various elementary symmetric polynomials that can be used to establish necessary conditions and sufficient conditions for @xmath0 to be ultrametric .
for example , set @xmath316 and @xmath317 if @xmath318 and @xmath319 are positive constants such that @xmath320 and @xmath321 then , by ( * ? ? ? * theorem 77 , chapter ii ) @xmath322 thus , if @xmath323 and @xmath324 satisfy the inequalities and @xmath325 then @xmath326 when @xmath327 , and the opposite inequality hold when @xmath328 .
this observation leads to necessary conditions and sufficient conditions for @xmath0 to be ultrametric .
[ rem : more ultrametric determination ] as we will see in for @xmath156-valent and rooted @xmath4-ary combinatorial trees , a somewhat similar proof argument based on the consideration of length sequences that are minimal with respect to a suitable order leads to a stronger result in that case .
there we can not only determine @xmath0 from the joint probability distribution of its random length sequence , but if we have a random tree @xmath159 with a fixed number of leaves , then it is possible to determine the distribution of @xmath159 from the joint probability distribution of the random length sequence obtained by first picking a realization of @xmath159 and then independently picking a random ordering of the leaves to build a random length sequence .
formally , we have some space @xmath20 of isomorphism types of trees , a corresponding space @xmath329 of possible length sequences , and a probability kernel @xmath330 from @xmath20 to @xmath329 , where , for @xmath331 , @xmath332 is the element of @xmath333 , the space of probability measures on @xmath329 , that is the joint probability distribution of the random length sequence built from @xmath0 .
an affirmative answer to for a particular @xmath20 means that the map @xmath334 from @xmath20 to @xmath333 is injective . given an element @xmath330 of @xmath335 , the space of probability measures on @xmath20 , let @xmath336 be defined as usual by @xmath337 for @xmath338 .
the stronger results obtained in say that , in the situations considered there , the map @xmath339 from @xmath335 to @xmath333 is injective . one can ask if an analogous strengthening is also true for ultrametric trees . a proof along the lines of that given for does nt appear to apply immediately in this situation where the relevant space @xmath20 is uncountable rather than finite .
we leave this as one of many open questions .
recall that a caterpillar is a ( not necessarily simple ) combinatorial tree such that deleting the leaves of the tree results in a path consisting of @xmath144 vertices ( and hence @xmath44 edges of length @xmath25 ) .
choosing one end of the path , we can label the vertices on path consecutively with @xmath340 and denote by @xmath341 the leaves that are attached to vertex @xmath342 on the path . both @xmath343 and @xmath344
are non - zero , but the remaining @xmath146 may be zero . the isomorphism types of caterpillars with @xmath1 leaves are thus seen to be in a bijective correspondence with equivalence classes of nonnegative integer sequences @xmath345 , where @xmath346 and @xmath347 , and we declare that @xmath345 and @xmath348 are equivalent .
the proof of the following , which establishes the first claim in , is straightforward and we omit it .
a combinatorial tree @xmath0 with @xmath1 leaves is a caterpillar with an associated path of length @xmath44 if and only if @xmath349 and @xmath350 almost surely .
[ prop : det if caterpillar ] we now turn to the proof of the main claim in .
consider a box with @xmath1 tickets .
each ticket has a label belonging to @xmath351 and there are @xmath146 tickets with label @xmath43 for @xmath153 .
let @xmath352 be the result of drawing tickets uniformly at random from the box without replacement and noting their labels .
set @xmath353 it is clear that @xmath354 has the same joint probability distribution as @xmath355 , and so it suffices to show that it is possible to determine @xmath356 from a knowledge of the joint probability distribution of @xmath357 ( that is , it is possible to determine up to a reflection the vector that gives the number of tickets with each label ) . to begin with ,
note that , as in , @xmath358 and so we can determine @xmath44 from the joint probability distribution of @xmath359 .
observe next that @xmath360 and @xmath361 we can thus determine the multiset @xmath362 and , in particular , @xmath363 .
for @xmath364 we have @xmath365 and so we can determine @xmath366 . if @xmath44 is even , then @xmath367 and so we can determine @xmath368 . also , @xmath369 and , for @xmath370 , @xmath371 we can therefore determine @xmath372 for @xmath373 .
we claim that we the information we have just derived suffices to determine @xmath356 .
that is , if @xmath374 is a sequence with @xmath375 @xmath376 for @xmath377 , and @xmath378 for @xmath373 , then either @xmath379 for @xmath377 or @xmath380 for @xmath377 . to see that this is so , introduce the fourier transforms @xmath381 and @xmath382 for @xmath383
these are entire functions that uniquely determine @xmath384 and @xmath374 .
note that @xmath385 and a similar formula holds for @xmath386 .
it will thus suffice to show that either @xmath387 or @xmath388 ( equivalently , @xmath389 ) .
it follows from the assumption that @xmath378 for @xmath373 that if we define @xmath390 by @xmath391 and define @xmath392 similarly , then @xmath393 for all @xmath394 and hence @xmath395 for all @xmath383 . by theorem 2.2 in @xcite ,
there exist finitely supported functions @xmath396 and @xmath397 such that if we set @xmath398 and @xmath399 then @xmath400 and @xmath401 it follows from the assumption that @xmath376 for @xmath377 that @xmath402 for all @xmath383 .
therefore , @xmath403 and hence @xmath404 for all @xmath383 .
because the functions @xmath405 and @xmath406 are both entire , we must have either that @xmath407 for all @xmath383 or @xmath408 for all @xmath383 . if @xmath407 for all @xmath383 , then @xmath409 and @xmath410 for @xmath153 .
if @xmath408 for all @xmath383 , then @xmath411 and @xmath379 for @xmath377 .
we now turn our focus to the cases of @xmath6-valent and @xmath4-ary trees . recall that a _ @xmath6-valent tree _ is a tree with all vertices of degree either @xmath156 or @xmath25 . for @xmath155
a _ rooted @xmath4-ary tree _ is a tree with one vertex of degree @xmath4 and the rest of degrees either @xmath412 or @xmath25 .
we refer to the rooted @xmath64-ary tree as a _ rooted binary tree_. note that any @xmath4-ary tree is obtained by removing one leaf of a suitable @xmath6-valent trees .
our general proof methodology for these families of trees is similar to that used in for ultrametric trees .
we first define a particular class of sequences that can appear as elements of @xmath95 ( the down - split sequences ) and a total order on such sequences .
we then show that the minimal down - split sequence in @xmath413 uniquely identifies @xmath0 .
the idea of the proof is the same for all @xmath4 and depends on the following fact .
let @xmath0 be a @xmath6-valent tree or a rooted @xmath4-ary tree and let @xmath414 be a subtree of @xmath0 .
then @xmath414 is a rooted @xmath4-ary tree if and only if @xmath415 [ lem : k subtree size ] because @xmath414 is a subtree of @xmath0 , every interior vertex of @xmath414 has degree at most @xmath156 .
write @xmath416 for the number of vertices of @xmath414 of degrees @xmath417 .
we need to show that @xmath418 for @xmath419 and @xmath420 , or , equivalently , that @xmath420 and @xmath421 .
this is in turn equivalent to showing that @xmath422 which , by the `` handshaking identity '' @xmath423 becomes @xmath424 or , upon rearranging , @xmath425 for simplicity of notation we present the details of the proof for the case of ( unrooted ) @xmath58-valent trees and rooted binary trees ( that is , @xmath274 ) .
we end in with a discussion of the extension to general @xmath4 .
our proof of begins with an analysis of random length sequences for marked ( also known as planted ) @xmath58-valent trees .
a _ marked @xmath58-valent tree _
@xmath426 is an @xmath58-valent tree @xmath0 and a distinguished leaf @xmath35 of @xmath0 .
we define the _ modified random length sequence _
@xmath427 of @xmath426 to be the random length sequence @xmath93 of @xmath0 conditioned on @xmath428 .
we need to distinguish some particular sequences that appear in the support of @xmath429 . as usual , we can define a partial order on @xmath8 by declaring that @xmath82 precedes @xmath83 if @xmath430 is on the path between @xmath35 and @xmath83 , and we can extend this partial order to a total order @xmath431 such that if @xmath432 are such that @xmath298 and @xmath82 are not comparable in the partial order but @xmath433 , @xmath298 precedes @xmath83 in the partial order , and @xmath82 precedes @xmath434 in the partial order , then @xmath435 .
such a total order corresponds to embedding @xmath0 in the plane and listing the elements of @xmath8 in the order they are encountered as one walks around @xmath0 starting from @xmath35 .
suppose that @xmath436 is the ordered listing of @xmath74 .
set @xmath437 , @xmath3 .
if @xmath438 , then the subtree spanned by the @xmath4 leaves @xmath439 has @xmath440 edges and hence , by , this subtree is a binary tree .
if we write @xmath441 for the vertex adjacent to the marked leaf @xmath35 , denote by @xmath442 the other two vertices adjacent to @xmath441 , and suppose that @xmath443 , then it must be the case that @xmath444 .
write @xmath14 ( respectively , @xmath15 ) for the subtree of @xmath0 consisting of @xmath298 and the vertices @xmath34 such that @xmath445 ( respectively , @xmath446 ) is on the path from @xmath441 to @xmath34 .
the sequence @xmath447 satisfies @xmath448 for @xmath449 .
the sequence @xmath450 satisfies @xmath451 for @xmath452 .
[ rem : down - split construction ] [ def : downsplit ] a _ down - split sequence _ is an element of the class of increasing sequences of positive integers defined recursively as follows .
the sequence @xmath453 is a down - split sequence .
a sequence @xmath454 , @xmath455 , is down - split if @xmath456 and , setting @xmath457 * @xmath458 is down - split , * @xmath459 is down - split .
the index @xmath460 is the _ splitting index _ of @xmath461 . for @xmath462 ,
the sequence @xmath463 is a down - split sequence . here
@xmath464 , @xmath465 and @xmath466 .
the following result is immediate from . for every marked @xmath58-valent tree @xmath467
there is at least one down - split sequence @xmath461 with @xmath468 [ lem : exist of split ] we record the following fact for later use . if @xmath469 is a down - split sequence then @xmath470 .
[ lem : size of downsplit ] this follows easily by induction .
if @xmath461 splits at @xmath460 , then , as @xmath471 is a down - split sequence with @xmath472 , we have by the inductive hypothesis that @xmath473 and the claim follows . [ ex : nonunique split seq ] given any down - split sequence @xmath461 , it is possible to reverse the argument in and construct a marked @xmath58-valent tree with a suitable total ordering on its vertices such that @xmath461 is the corresponding down - split sequence .
however , a marked @xmath58-valent tree @xmath467 is not uniquely identified by an arbitrary down - split sequence in the support of @xmath429 , as the example in shows . and @xmath474 with particular realizations of the random selection of leaves.,title="fig:",scaledwidth=40.0% ] and @xmath474 with particular realizations of the random selection of leaves.,title="fig:",scaledwidth=40.0% ]
write @xmath475 and @xmath476 for the random selections of the leaves of @xmath477 and @xmath478 .
suppose that the realizations are such that @xmath479 for @xmath480 and that these leaves of the subtree @xmath481 appear in an order of the type discussed in .
the corresponding realizations for the modified random length sequences are equal .
the common value @xmath482 is a down - split sequence with splitting index @xmath58 .
thus , two non - isomorphic marked @xmath58-valent trees can have a common down - split sequence in the supports of their modified random length sequences .
note that the common down - split sequence results from taking the leaves of @xmath478 according to an order of the type described in , but this is not the case for @xmath477 . with in mind
we see that it would be useful to have a way of recognizing down - split sequences in the support of @xmath429 that result from realizations where the leaves are selected in an order that arises from a suitable total order on the vertices of @xmath0 .
the key is the following total order on down - split sequences .
we re - use the notation @xmath260 that was used in for the lexicographic order .
define a total order @xmath260 on the set of down - split sequences of a given length recursively as follows .
firstly , @xmath483 does not hold .
next , let @xmath484 be down - split sequences indexed by @xmath485 with respective splitting indices @xmath460 and @xmath486 . set @xmath487 and @xmath488 declare that @xmath489 if @xmath490 or @xmath491 or @xmath492 the next result follows easily by induction . the binary relation @xmath260 is a total order on the set of down - split sequences of a given length .
the _ minimal down - split sequence _ for a marked @xmath58-valent tree @xmath467 is the minimal element ( with respect to the total order @xmath260 ) of the set @xmath493 we now proceed to establish some results that culminate in showing that @xmath467 is determined by its minimal down - split sequence .
let @xmath467 be a marked @xmath58-valent tree , with modified random length sequence @xmath494 constructed from the random sequence of leaves @xmath495 with @xmath428 .
denote by @xmath441 the vertex adjacent to the marked leaf @xmath35 and denote by @xmath442 the other two vertices adjacent to @xmath441 .
write @xmath14 ( respectively , @xmath15 ) for the subtree of @xmath0 consisting of @xmath441 and the vertices @xmath34 such that @xmath445 ( respectively , @xmath446 ) is on the path from @xmath441 to @xmath34 . set @xmath496 then @xmath497 if and only if @xmath498 or _ vice versa_. [ lem : split subtrees ] if @xmath499 and @xmath500 , then the subtree spanned by @xmath501 consists of the leaf @xmath35 adjoined to @xmath14 via an edge to the vertex @xmath441 .
this subtree is a rooted binary tree with root @xmath441 .
it follows from that @xmath497 .
for the other direction , assume that @xmath502 .
by , the subtree @xmath414 spanned by @xmath503 is a rooted binary tree with @xmath36 leaves .
we have @xmath504 , @xmath505 , and @xmath506 .
we need to show that @xmath414 consists of the leaf @xmath35 adjoined to either @xmath14 or @xmath15 via an edge to the vertex @xmath441 that is common to both @xmath14 and @xmath15 . by the construction prior to the statement of we know that @xmath507 and so if @xmath508 , then @xmath509 and similarly with the roles of @xmath14 and @xmath15 reversed .
we can rule out the possibility that @xmath510 intersects both @xmath511 and @xmath512 as follows .
if @xmath513 and @xmath514 , then @xmath515 must be a proper subset of @xmath516 and @xmath517 must be a proper subset of @xmath518 .
if @xmath515 is a proper , nonempty subset of @xmath516 , then @xmath414 must have a degree @xmath64 vertex that belongs to @xmath519 , and similarly for @xmath15 .
however , @xmath414 is a rooted binary tree and can not have two or more vertices of degree @xmath64 .
finally , we need to rule out the possibility of @xmath520 is a proper subset of @xmath516 or @xmath518 .
however , if @xmath520 is a proper subset of @xmath516 , then @xmath414 would have at least one degree @xmath64 vertex in that belongs to @xmath519 as well as the degree @xmath64 vertex @xmath441 , which contradicts @xmath461 being a rooted binary tree .
the same argument holds with @xmath15 in place of @xmath14 .
let @xmath467 be a marked @xmath58-valent tree with modified random length sequence @xmath494 .
then @xmath521 is the splitting index for the minimal down - split sequence for @xmath467 .
[ cor : split subtrees ] if @xmath460 is the splitting index of any down - split sequence @xmath461 in the support of @xmath429 , then @xmath522 by definition .
thus @xmath523 and hence @xmath524 . on the other hand ,
let @xmath525 be as in the statement of .
it follows from that result that @xmath526 . by the construction in
if @xmath527 or the analogous one with the roles of @xmath14 and @xmath15 reversed if @xmath528 , we may construct a down - split sequence for @xmath467 that has splitting index @xmath36 . by the definition of the total order @xmath260 , the splitting index for the minimal down - split sequence for @xmath467 is at most @xmath36 .
let @xmath461 be the minimal down - split sequence for a marked @xmath58-valent tree @xmath467 .
there is no other marked @xmath58-valent tree for which @xmath461 is the minimal down - split sequence .
[ prop : det by downsplit ] we will prove this by induction .
the claim is clearly true for the down - split sequence @xmath529 .
let @xmath426 be a marked @xmath58-valent tree and @xmath461 the minimal down - split sequence for @xmath467 .
define @xmath525 as in the statement of .
let @xmath460 be the splitting index of @xmath461 .
let @xmath94 be an ordered listing of @xmath74 such that @xmath530 for @xmath3 . by and we must either have @xmath531 and @xmath532 or the analogous conclusion with the roles of @xmath14 and @xmath15 interchanged holds ( if @xmath533 , then only one alternative is possible )
. we may suppose without loss of generality that the choice of @xmath445 and @xmath446 is such that the first alternative holds .
set @xmath534 by definition , @xmath535 and @xmath536 are down - split sequences . because @xmath537 , we have @xmath538 and @xmath539 we claim that @xmath535 must be the minimal down - split sequence for @xmath540 . to see this ,
note that if there was a down - split sequence @xmath541 with @xmath542 such that @xmath543 then , writing @xmath544 for a positive integer @xmath36 we would have @xmath545 and , by definition of the total order @xmath260 , @xmath546 this , however , contradicts the minimality of @xmath461 .
similarly , @xmath536 is the minimal down - split sequence for @xmath547 . by induction , @xmath540 and @xmath547
are uniquely determined .
since @xmath467 is obtained by gluing @xmath548 and @xmath547 together at the shared vertex @xmath441 and attaching the marked leaf @xmath35 to @xmath441 by an edge , we see that @xmath426 is also determined by @xmath461 .
while the proof of is not in the form of an explicit reconstruction procedure , the argument clearly leads to an algorithm for building a marked @xmath58-valent tree @xmath426 from the corresponding minimal down - split sequence .
namely , @xmath467 is simply the recursion tree that results from parsing @xmath461 as a down - split sequence as in , with leaves corresponding to edges that terminate in the sequence @xmath549 .
-valent tree with its leaves ordered minimally and the corresponding parse tree for the minimal down - split sequence.,title="fig:",scaledwidth=40.0% ] -valent tree with its leaves ordered minimally and the corresponding parse tree for the minimal down - split sequence.,title="fig:",scaledwidth=40.0% ] from we are able to easily prove for ( unmarked ) @xmath58-valent trees .
let @xmath0 be a fixed ( unknown ) @xmath58-valent tree with @xmath1 leaves and let @xmath93 be its random length sequence .
conditional on @xmath550 , @xmath93 is the modified random length sequence of the marked binary tree @xmath551 .
thus , if @xmath552 then there must be some leaf @xmath553 such that @xmath554 let @xmath555 be the minimal element of the set @xmath556 then @xmath555 must be the minimal down - split sequence for @xmath426 for at least one leaf @xmath35 of @xmath0 . by we
can reconstruct @xmath426 and hence @xmath0 from @xmath555 .
the above argument can be pushed further to prove for @xmath159 a random @xmath58-valent tree .
let @xmath159 be a random @xmath58-valent tree with @xmath1 leaves and random length sequence @xmath557
. given a @xmath58-valent tree @xmath0 with @xmath1 leaves , let @xmath558 be the minimal element of the set of down - split sequences of the marked @xmath58-valent trees @xmath467 as @xmath35 ranges over @xmath74 .
we equip the set of @xmath58-valent tree with @xmath1 leaves with a total order that , with a slight abuse of notation , we denote by @xmath260 by declaring that @xmath559 if @xmath560 .
note that if @xmath559 , then @xmath561 and @xmath562 .
now , for each choice of @xmath0 we have @xmath563 and the conclusion that we can recover @xmath564 as @xmath0 ranges over the @xmath58-valent trees with @xmath1 leaves follows simply from the observation that if @xmath172 is a row vector of length @xmath565 and @xmath566 is an @xmath567 matrix that has all entries below the diagonal zero and all entries on the diagonal strictly positive , then there is a unique row vector @xmath82 of length @xmath565 such that @xmath568 .
we now prove and for rooted binary trees .
analogous to the objects we introduced for marked @xmath58-valent trees , we begin with a definition of a class of sequences that will appear in the support of the random length sequence of a rooted binary tree . an _ up - split sequence _ is an element of the class of increasing sequences of nonnegative integers defined recursively as follows .
the sequence @xmath569 is an up - split sequence .
a sequence @xmath570 , @xmath571 , is an up - split sequence if @xmath572 and , setting @xmath573 * @xmath574 is an up - split sequence , * @xmath575 is a down - split sequence .
the index @xmath460 is the _ splitting index _ of @xmath461 .
suppose that @xmath0 is a rooted binary tree with root @xmath441 . in a manner similar to the construction in we
can order a partial order on @xmath8 by declaring that @xmath82 precedes @xmath83 if @xmath430 is on the path between @xmath122 and @xmath83 , and we can extend this partial order to a total order @xmath431 such that if @xmath432 are such that @xmath298 and @xmath82 are not comparable in the partial order but @xmath433 , @xmath298 precedes @xmath83 in the partial order , and @xmath82 precedes @xmath434 in the partial order , then @xmath435 .
suppose that @xmath576 is the ordered listing of @xmath74 .
set @xmath577 and @xmath578 , @xmath3 .
then @xmath579 is an up - split sequences .
the leaves @xmath580 and @xmath581 respectively span the two binary subtrees @xmath14 and @xmath15 that are rooted at the two children of the root @xmath441 .
the subtree spanned by @xmath441 and @xmath581 is a @xmath58-valent tree . and
a marked @xmath58-valent tree @xmath547.,scaledwidth=60.0% ] [ ex : up - split construction ] the following analogue of is clear from . for every rooted binary tree @xmath0
, there is at least one up - split sequence @xmath461 with @xmath582 [ lem : exist upsplit ] the following analogue of can be established using a similar inductive proof .
if @xmath570 is an up - split sequence then @xmath583 .
[ lem : size of upsplit ] define a total order @xmath584 on the set of up - split sequences of a given length recursively as follows .
firstly , @xmath585 does not hold .
next , let @xmath461 and @xmath342 be two up - split sequences indexed by @xmath586 with respective splitting indices @xmath460 and @xmath486 . set @xmath587 and @xmath588 declare that @xmath589 if @xmath590 or @xmath591 or @xmath592 note for up - split sequences @xmath461 and @xmath342 , that @xmath593 implies that the splitting index of @xmath461 is _ greater than or equal _ to the splitting index of @xmath342 . for down - split sequences @xmath34 and @xmath594 , @xmath595 implies that the splitting index of @xmath34 is _ less than or equal _ to the splitting index of @xmath594 .
this change in the direction of the inequalities matches the switch in the definition of the splitting index from an infimum for down - split sequences to a supremum for up - split sequences .
the next result follows easily by induction .
the binary relation @xmath584 is a total order on the set of up - split sequences of a given length .
the _ minimal up - split sequence _ for a rooted binary tree @xmath0 is the minimal element ( with respect to the total order @xmath584 ) of the set @xmath596 the up - split sequence analogues of and are the following and they are proved in essentially the same manner . given a binary tree @xmath0 with root @xmath441 , let @xmath14 and @xmath15 be the binary subtrees rooted at the two children of @xmath441 . set @xmath597 then @xmath497 if and only if @xmath598 and @xmath599 or _ vice versa_. [ lem : upsplit subtrees ] let @xmath0 be a rooted binary tree with random length sequence @xmath600 .
then @xmath601 is the splitting index for the minimal up - split sequence for @xmath0 .
[ cor : upsplit subtrees ] the following analogue of for up - split sequences follows from and in essentially the same manner that followed from and .
let @xmath461 be the minimal up - split sequence for a rooted binary tree @xmath0 .
there is no other rooted binary tree for which @xmath461 is the minimal up - split sequence .
[ prop : det by upsplit ] clearly , completes the proof of . to establish in the case of @xmath159 a random rooted binary tree , we need only repeat the argument of the proof of given in for @xmath58-valent trees .
the proof of the extension and to @xmath6-valent and rooted @xmath4-ary combinatorial trees for @xmath602 is very similar to the @xmath157 case and involves the introduction of suitable notion of down - split and up - split sequences along with appropriate total orders on these sets of sequences .
the only difference is that both types of split sequences are now split into @xmath4 smaller sequences , instead of just two .
we leave the details to the reader .
the original conjecture question [ q : main ] remains open in general , both for simple trees with arbitrary edge weights ( not in general position ) , and for combinatorial trees .
an even more general question is suggested by .
let @xmath159 be a random tree with probability distribution supported either on the set of simple trees with @xmath1 leaves and general edge weights or the set of combinatorial trees with @xmath1 leaves .
can the probability distribution of @xmath159 be determined uniquely from the joint probability distribution of the random length sequence @xmath557 ?
[ q : open general ] even if the answer to question [ q : open general ] is `` no '' , the answer may still be `` yes '' if the probability distribution of @xmath159 is known _ a priori _ to belong to some particular family of probability distributions .
there are , of course , many families of probability models for with random trees with @xmath1 leaves that are described by a small number of parameters ( for example , conditioned galton - watson models , the various preferential attachment models ) , and perhaps the value of these parameters can be determined from the joint probability distribution of the random length sequence of a random tree that is known _ a priori _ to be distributed according to a member of one of these families .
we remarked in the introduction that the focus of this paper is superficially similar to that in @xcite , where the problem of reconstructing a combinatorial tree from its number deck ( the sizes of the subtrees in the forests produced by deleting each vertex ) was studied .
the lists of lists that are the number deck of some combinatorial tree are characterized in @xcite .
are there more parsimonious quantities derived from the joint probability distribution of the random length sequence that still carry a lot of information about @xmath0 ?
for example , how much information about @xmath0 is contained in the expectation @xmath603 , \ldots , \be[w_n])$ ] of the random length sequence and is it possible to characterize those vectors which can arise as the expectation of the random length sequence ? alfred v. aho , john e. hopcroft , and jeffrey d. ullman , _ the design and analysis of computer algorithms _ , addison - wesley publishing co. , reading , mass .- london - amsterdam , 1975 , second printing , addison - wesley series in computer science and information processing .
mr 0413592 ( 54 # 1706 ) to3em , _ a graph reconstructor s manual _ , surveys in combinatorics , 1991 ( guildford , 1991 ) , london math . soc .
lecture note ser .
166 , cambridge univ . press , cambridge , 1991 , pp .
mr 1161466 ( 93e:05071 ) peter buneman , _ the recovery of trees from measures of dissimilarity _ , mathematics in the archaeological and historical sciences ( f.r .
hodson , d.g .
kendall , and p. tautu , eds . ) , edinburgh university press , edinburgh , 1971 , pp .
387395 .
steven n. evans , _ probability and real trees _ , lecture notes in mathematics , vol .
1920 , springer , berlin , 2008 , lectures from the 35th summer school on probability theory held in saint - flour , july 623 , 2005 .
mr mr2351587
gary gordon , eleanor mcdonnell , darren orloff , and nen yung , _ on the tutte polynomial of a tree _ , proceedings of the twenty - sixth southeastern international conference on combinatorics , graph theory and computing ( boca raton , fl , 1995 ) , vol . 108 , 1995 , pp .
mr 1369283 ( 96j:05036 ) i. krasikov and j. schnheim , _ the reconstruction of a tree from its number deck _ , discrete math .
* 53 * ( 1985 ) , 137145 , special volume on ordered sets and their applications ( larbresle , 1982 ) .
mr 786484 ( 86d:05088 ) frederick a matsen and steven n evans , _ ubiquity of synonymity : almost all large binary trees are not uniquely identified by their spectra or their immanantal polynomials . _ ,
algorithms for molecular biology : amb * 7 * ( 2011 ) , no . 1 , 1414 . jeremy l. martin , matthew morin , and jennifer d. wagner , _ on distinguishing trees by their chromatic symmetric functions _ , j. combin .
theory ser .
a * 115 * ( 2008 ) , no . 2 , 237253 .
mr 2382514 ( 2008m:05115 ) | given an edge - weighted tree @xmath0 with @xmath1 leaves , sample the leaves uniformly at random without replacement and let @xmath2 , @xmath3 , be the length of the subtree spanned by the first @xmath4 leaves .
we consider the question , `` can @xmath0 be identified ( up to isomorphism ) by the joint probability distribution of the random vector @xmath5 ? ''
we show that if @xmath0 is known _ a priori _ to belong to one of various families of edge - weighted trees , then the answer is , `` yes . ''
these families include the edge - weighted trees with edge - weights in general position , the ultrametric edge - weighted trees , and certain families with equal weights on all edges such as @xmath6-valent and rooted @xmath4-ary trees for @xmath7 and caterpillars . |
according to quantum chromodynamics , at sufficiently high energy density which can be achieved in ultra - relativistic heavy - ion collisions , a new state of matter , called quark - gluon - plasma ( qgp ) , is created .
one of main features of the qgp is its collective expansion which could be described by relativistic hydrodynamic flows . due to the different pressure gradients in different directions ,
the initial spatial anisotropy converts into momentum anisotropy , observed in the final state as a preferential emission of particles in a certain plane .
the collective behavior of a strongly - coupled hot and dense qgp has been studied using the azimuthal anisotropy of emitted particles detected at experiments at the relativistic heavy ion collider ( rhic ) @xcite .
the studies have been continued also with the experiments @xcite at the large hadron collider ( lhc ) where significantly higher collision energies are achieved .
quantitatively , the anisotropic hydrodynamic flow is described by fourier decomposition of the hadron yield distribution in azimuthal angle , @xmath6 , @xcite @xmath7\ ] ] where fourier coefficients , @xmath8 , characterize magnitude of the azimuthal anisotropy measured with respect to the event plane angle , @xmath9 .
the angle @xmath9 determines the direction of maximum final - state particle density and can be reconstructed from the emitted particles themselves .
the most analyzed anisotropic flow is the second order fourier coefficient , @xmath1 , called elliptic flow. the angle @xmath10 corresponds to the event plane which is correlated with the participant plane spanned over the beam direction and the shorter axis of the approximately elliptical shape of the nucleon overlap region . due to the initial - state fluctuations of the position of nucleons at the moment of impact , higher - order deformations of the initial geometry
are induced , which lead to the appearance of higher - order fourier harmonics ( @xmath8 , @xmath113 in eq .
( [ f1 ] ) ) .
they are measured with respect to the corresponding event plane angles , @xmath9 @xcite . if particles are correlated with the event plane than they are also mutually correlated .
this gives an opportunity to investigate the azimuthal anisotropies by studying the two - particle correlations @xcite .
these correlations can be also fourier decomposed as @xmath12 where @xmath13 is a relative azimuthal angle of a particle pair .
the two - particle fourier coefficient @xmath14 is expected to factorize as @xmath15 into a product of the single - particle anisotropy harmonics @xmath8 .
a key assumption is that the event plane angle @xmath9 in eq .
( [ f1 ] ) is a global quantity for a given collision .
recently , a significant breakdown of the factorization assumption expressed through eq .
( [ fact ] ) has been observed both in the transversal @xmath0 and longitudinal @xmath16 direction is defined as @xmath17 where @xmath18 is the polar angle .
] in symmetric pbpb collisions @xcite as well as in asymmetric ppb collisions @xcite .
the effect has been theoretically predicted in @xcite .
it is shown that even if the hydrodynamic flow is the only source of the two - particle correlations , initial - state fluctuations turns the event plane from a global to both , @xmath0 and @xmath16 dependent quantity .
lumpy hot - spots raised from the initial - state fluctuations can generate a local pressure gradient which makes the corresponding local event plane to be slightly different but still correlated with the global event plane .
this effect of initial - state fluctuations thus breaks the factorization relation of eq .
( [ fact ] ) .
recently , a new approach which employs the principal component analysis ( pca ) to study the flow phenomena is introduced @xcite . using a pca approach , @xmath14 coefficients of the observed two - particle azimuthal correlations as a function of both particles
@xmath0 are represented through the leading and the sub - leading flow mode terms .
the leading flow modes are essentially equivalent to single - particle anisotropy harmonics ( @xmath8 ) extracted from two - particle correlation methods . as a direct consequence of initial - state fluctuations ,
the sub - leading flow modes represent the largest sources of factorization breaking .
the pca study of this effect can give new insights into the expansion dynamics of the strongly coupled qgp , and serves as an excellent tool for testing the hydrodynamical models .
this paper is organized in a following way .
the basic features of hydjet++ model @xcite are described in sect .
[ sec : hydjet ] .
details of the applied construction of the two - particle correlation functions , as well as the pca approach in flow analysis are given in sect .
[ sec : ana ] . using hydjet++ model , approximately 40 m pbpb collisions at @xmath192.76 tev are simulated and analyzed .
the obtained results together with the corresponding discussions are given in sect .
[ sec : res ] .
the results are presented over a wide range of centralities going from ultra central ( 0 - 0.2% centrality ) up to peripheral ( 50 - 60% centrality ) pbpb collisions .
the analyzed @xmath0 interval is restricted to @xmath20 3 gev / c range where hydrodynamics dominates .
a disscussion concerning the results obtained under different hydjet++ model switches is given in sect .
[ sec : dis ] .
conclusions are given in sect .
[ sec : conc ] .
the monte carlo hydjet++ model simulates relativistic heavy ion collisions in an event - by - event manner .
it is made of two components which simulate soft and hard processes .
the soft part provides the hydrodynamical evolution of the system while the hard part describes multiparton fragmentation within the formed medium . within the hard part
, jet quenching effects are also taken into account .
the minimal transverse momentum @xmath21 of hard scattering of the incoming partons regulates does it would contribute to the soft or to the hard part .
the partons which are produced with @xmath22 , or which are quenched below @xmath21 do not contribute to the hard part .
the hard part of the model consists of pythia @xcite and pyquen @xcite event generators .
these generators simulate initial parton - parton collisions , radiative energy loss of partons and parton hadronization . within the soft part of the hydjet++ model , the magnitude of the elliptic flow
is regulated by spatial anisotropy @xmath23 which is the elliptic modulation of the final freeze - out hyper - surface at a given impact parameter vector is a vector which connects centers of the colliding nuclei . ]
magnitude @xmath24 , and by momentum anisotropy @xmath25 which gives the modulation of the flow velocity profile .
additionally introduced triangular modulation of the freeze - out hyper - surface , @xmath26 , determines the @xmath2 magnitude .
the events can be generated under several switches .
the most realistic one , flow+quenched jets , includes both hydrodynamics expansion and quenched jets . in this analysis ,
the pure flow switch is also used .
the details of the model can be found in the hydjet++ manual @xcite .
the construction of the two - dimensional two - particle correlation function follows the definition adopted within the cms experiment .
any charged pion from the @xmath27 2 range can be used as a trigger particle . in order to perform a differential analysis ,
all events are divided into the eight centrality classes , while the analyzed @xmath0 range has seven non - equidistant intervals . since in an event there can be more than one trigger particle from a given @xmath0 interval , the corresponding total number is denoted by @xmath28 . in each event , every trigger particle is paired with all of the remaining charged pions from the @xmath27 2 range within a given @xmath0 interval .
the signal distribution , @xmath29 , is defined as the yield of the per - trigger - particle pairs within the same event , @xmath30 in eq .
( [ s ] ) , @xmath31 denotes the per - trigger - particle pairs yield within a given ( @xmath32 ) bin where @xmath33 and @xmath13 are corresponding differences in pseudorapidity and azimuthal angle between the two charged pions which are forming the pair .
the background distribution , denoted with @xmath34 , is constructed using the technique of mixing topologically similar events which ensure that the pairs are not physically correlated . here
, topological similarity means that events which are mixed have relative difference in multiplicity smaller than 5% .
the trigger particles from one event are combined ( mixed ) with all of the associated particles from a different event . in order to reduce contribution to the statistical uncertainty from the background distribution ,
associated particles from 10 randomly chosen events are used . in the background distribution ,
defined as @xmath35 @xmath36 denotes the number of mixed - event pairs in a given ( @xmath32 ) bin . due to the fact that pairs are formed from uncorrelated particles ,
the background gives a distribution of independent particle emission .
the two - dimensional two - particle differential correlation function is then defined as the normalized ratio of the _ signal _ to the _ background _ distribution @xmath37 the normalization factor , b(0 , 0 ) , is the value of the background distribution at @xmath38 0 and @xmath39 0 . in order to obtain single - particle azimuthal anisotropy harmonics , @xmath8 , the projection of the two - dimensional correlation function given by eq .
( [ 2dcorr ] ) onto @xmath13 axis can be fourier decomposed as given in eq .
( [ f2 ] ) . in order to avoid the short - range correlations arising from jet fragmentation and resonance decays , an averaging over @xmath40 2 is applied .
two - particle fourier coefficients , @xmath14 , as input variables used within the pca method , are calculated as @xmath41 where @xmath42 and @xmath43 stands for the _ signal _ and for the _ background _ respectively .
although the two - particle fourier coefficients , @xmath14 , obtained using the eq .
( [ 2pfc ] ) are essentially equal to those obtained from fourier fit given by eq .
( [ f2 ] ) , the direct calculation given by eq .
( [ 2pfc ] ) has a slight advantage as it is not affected by the bin edge effect .
the _ signal _ is constructed within the same event while the _ background _ is obtained by mixing topologically similar events which ensure that the pairs are not physically correlated .
the @xmath44 denotes averaging over events from the given centrality class . in order to suppress the short - range correlations , which come from fragmentation of jets and from resonance decays , only charged pion pairs with @xmath402 are taken into account in eq .
( [ 2pfc ] ) , where @xmath45 is the relative difference in pseudorapidities between particles @xmath46 and @xmath24 which a forming the pair .
particles @xmath46 and @xmath24 can be taken from the same @xmath0 interval and thus form diagonal @xmath47 elements or from different @xmath0 interval forming non - diagonal @xmath48 elements . by defining @xmath49 differential @xmath0 bins one can construct the corresponding covariance matrix @xmath50_{n_{b}n_{b}}$ ] . in this analysis the @xmath0 range from 0.3 to 3.0 gev / c has been divided into @xmath51 7 non - equidistant @xmath0 bins . by solving the eigenvalue problem of the @xmath50 $ ] matrix , a set of the eigenvalues , @xmath52 , and eigenvectors , @xmath53 , has been obtained . here , @xmath54 . a new @xmath0-dependent variable called flow mode , @xmath55 , is introduced as @xmath56 the corresponding single - particle flow mode , @xmath57 , is then defined as @xmath58 where @xmath59 denotes averaged multiplicity in a given @xmath0 bin .
the mode which corresponds to @xmath601 is called single - particle leading flow mode .
the next one , @xmath602 is called single - particle sub - leading flow mode and so on
. the magnitude of the single - particle leading flow mode , @xmath61 , should be practically equal to the @xmath8 measured using the two - particle correlation method .
the cms collaboration showed in @xcite that the @xmath0 dependence of the single - particle leading elliptic and triangular flow modes for ppb collisions at 5.02 tev and for pbpb collisions at 2.76 tev data are in excellent agreement with corresponding two - particle measurements presented in @xcite and in @xcite , respectively .
in order to check the consistency of extracted single - particle azimuthal anisotropies , @xmath8 , using the method of pca and one of the standard approaches like fourier decomposition given by eq .
( [ f2 ] ) , as well as to perform a pca analysis in order to extract the leading and sub - leading flow modes , the two - particle correlation functions defined by the eq .
( [ 2dcorr ] ) are constructed . for each centrality interval , ranged from the ultra - central 0 - 0.2% till peripheral 50 - 60% , two - particle correlation functions for 7 @xmath62 intervals between 0.3 and 3.0 gev / c are formed .
thus , 7 diagonal and 21 non - diagonal two - particle correlation functions are produced .
as examples , in fig .
[ fig:1 ] are shown two - dimensional , in @xmath33 and @xmath13 , two - particle correlation functions from hydjet++ pbpb simulations at @xmath192.76 tev where both particles belongs to 0.3 @xmath63 0.5 gev / c ( upper raw ) and 1.5 @xmath63 2.0 gev / c ( middle raw ) interval ( diagonal elements ) , while in the bottom raw one particle belongs to 0.3 @xmath63 0.5 gev / c and another one to 1.5 @xmath63 2.0 gev / c ( non - diagonal element ) .
+ + the correlation functions in the left column are constructed from very central 0 - 5% collisions , and those from peripheral 40 - 50% collisions are presented in the right column . as this analysis deals with the long - range collective correlations , the near side peak is truncated .
one can see that , beside the short - range correlated near and the away side peak , hydjet++ model can reproduce well features of the long - range correlated elliptic and triangular flow . = 1 ) and the sub - leading ( @xmath64 = 2 ) flow mode for @xmath65 = 2 harmonic as a function of @xmath0 measured using the pca approach in a wide centrality range of pbpb collisions at 2.76 tev generated within the hydjet++ model .
the @xmath66 results are compared to the @xmath1 measured by the cms @xcite and alice @xcite collaborations , and to the @xmath67 extracted from the same hydjet++ simulation using the two - particle correlation method .
the error bars correspond to statistical uncertainties.,scaledwidth=45.0% ] in fig .
[ fig:2 ] are shown the pca results on the single - particle leading and sub - leading flow modes for the elliptic harmonic in 8 centrality regions ranged from ultra - central ( 0 - 0.2% ) to peripheral ( 50 - 60% ) pbpb collisions at @xmath68 2.76 tev simulated within hydjet++ event generator .
the single - particle leading flow mode , @xmath69 , is dominant and rather well describes the experimentally measured @xmath1 from two - particle correlations taken from @xcite and @xcite .
additionaly , due to consistency , in fig .
[ fig:2 ] are also shown @xmath67 values measured using two - particle correlations constructed from the same hydjet++ generated data . in fig .
[ fig:2 ] these results are depicted with crosses and show an excelent agreement with @xmath69 extracted using the pca method . the extracted @xmath69 has expected centrality behavior : a small magnitude at ultra - central collisions which then gradually increases going to peripheral collisions .
the newly discovered single - particle sub - leading flow mode , @xmath70 , is practicaly equal to zero at small-@xmath0 for all centrality bins . at high-@xmath0 region ,
the sub - leading flow mode has a small positive value and slowly increases going from semi - central to peripheral pbpb collisions .
the cms collaboration recently presented in @xcite experimentally measured the single - particle leading and sub - leading flow mode in pbpb collisions within the same @xmath0 range and for the same centrality bins as it is adopted in this analysis . beside the leading flow mode ,
hydjet++ predictions for the sub - leading flow mode are also in a qualitative agreement with the experimental findings from @xcite . in peripheral collisions
( @xmath71 30% ) the @xmath70 magnitudes predicted by hydjet++ model are slightly larger with respect to the ones observed from the experimental data .
= 1 ) and the sub - leading ( @xmath64 = 2 ) flow mode for @xmath65 = 3 harmonic as a function of @xmath0 measured using the pca approach in a wide centrality range of pbpb collisions at 2.76 tev generated within the hydjet++ model .
the @xmath72 results are compared to the @xmath2 measured by the cms @xcite and the alice @xcite collaborations , and to the @xmath73 extracted from the same hydjet++ simulation using the two - particle correlation method .
the error bars correspond to statistical uncertainties.,scaledwidth=45.0% ] similarly as in fig .
[ fig:2 ] , in fig .
[ fig:3 ] are shown the pca leading and sub - leading flow mode predictions of hydjet++ model for the triangular harmonic . again , the results are extracted from the same 8 centrality regions of pbpb collisions at @xmath68 2.76 tev .
the @xmath72 is in a rather good agreement with the @xmath2 results measured using two - particle correlations taken from @xcite and @xcite . also , the @xmath2 extracted from the two - particle correlations formed from the same hydjet++ generated data are in an excellent agreement with the @xmath72 obtained from the pca method .
the sub - leading mode is , up to 3 gev / c , almost equal to zero .
this supports finding from @xcite that the third harmonic factorizes better than the second one .
also , the small @xmath74 values extracted from hydjet++ simulated pbpb events are in an agreement with those found in @xcite extracted from the experimental pbpb data .
bin , as a function of centrality calculated using the pca method applied to pbpb collisions at @xmath68 2.76 tev simulated with hydjet++ event generator .
the error bars correspond to statistical uncertainties.,scaledwidth=52.0% ] in order to summarize results , in fig .
[ fig:4 ] are depicted ratios which gives the relative strength of the effect . ] between the single - particle sub - leading and leading flow mode .
the ratio is calculated from the values taken from 2.5 @xmath75 3.0 gev / c where the effect is strongest .
the results are presented as a function of centrality .
the results in the top panel of fig .
[ fig:4 ] show that in the case of @xmath76 2 the strength of the relative magnitude @xmath77 is smallest for events with centralities between 10 and 30% , i.e. where the elliptic flow is most pronounced .
going to very central collisions , the magnitude of the effect dramatically increases . also , the effect reaches a significant magnitude going to peripheral collisions .
qualitatively , such behavior is in an agreement with the @xmath78 multiplicity dependence presented in @xcite .
centrality dependence of the ratio which corresponds to the @xmath76 3 case is shown in the bottom panel of fig .
[ fig:4 ] .
the @xmath79 ratio , integrated over all centralities , is 0.095 @xmath80 0.009 .
as the extracted @xmath74 values are small , small fluctuations in their values can easily produce a non - smooth distribution shown in the bottom panel of fig .
[ fig:4 ] .
the overall small @xmath74 values found in this analysis are also in a qualitative agreement with the @xmath81 multiplicity dependence presented in @xcite .
= 1 ) and the sub - leading ( @xmath64 = 2 ) flow mode for @xmath65 = 2 harmonic as a function of @xmath0 measured using the pca approach in a wide centrality range of pbpb collisions at 2.76 tev generated within the hydjet++ model under the pure flow switch ( triangles ) and under the flow + quenched jets switch ( squares ) .
the error bars correspond to statistical uncertainties.,scaledwidth=46.0% ] in order to explore the origin of the sub - leading flow signal observed within the hydjet++ model , beside the analysis of the pbpb data obtained under the flow + quenched jets switch which results are shown in sect .
[ sec : res ] , the pure flow switch has been used for generating pbpb collisions at @xmath19 2.76 tev too .
the comparisons between the pca elliptic and triangular flow results obtained under these two switches are shown in fig .
[ fig:2c ] and fig .
[ fig:3c ] respectively .
as expected , the pure flow hydjet++ switch gives a linearly increasing leading flow mode for both @xmath61 harmonics @xmath76 2 and 3 .
also , as expected , the corresponding magnitude , at a given @xmath0 , is significantly greater with respect to the one extracted from the data obtained under flow + quenched jets switch .
the results for the sub - leading flow mode obtained under pure flow switch , contrary to those shown in sect .
[ sec : res ] are consistent with zero for centralities smaller than 20% .
but , even in the case of the pure flow switch , for centralities above 20% a modest effect of the sub - leading flow starts to appear .
up to the centrality of 40% the magnitude of the effect is still smaller with respect to the both , experimental findings from @xcite and from the results obtained using the flow + quenched jets switch . for the most peripheral , 50 - 60% the @xmath70 magnitude at high enough @xmath0 is greater than the experimental one and the one obtained under the flow + quenched jets switch . at first glance
, it seems that hydjet++ data simulated under the pure flow switch should not show existence of the sub - leading flow modes .
but , resonance decays and fluctuations of particle momenta together with the topology of peripheral events @xcite could imitate hot - spots which at the end could produce a non - zero sub - leading flows . the hydjet++ data simulated under the flow + quenched jets could have charged pions coming from the jet fragmentation , which due to the interaction with the soft medium and because of different path length with respect to the event plane can increase abundance of such high-@xmath0 pions near the event plane .
this also will produce the above mentioned hot - spots and consequently sub - leading flows .
= 1 ) and the sub - leading ( @xmath64 = 2 ) flow mode for @xmath65 = 3 harmonic as a function of @xmath0 measured using the pca approach in a wide centrality range of pbpb collisions at 2.76 tev generated within the hydjet++ model under the pure flow switch ( triangles ) and under the flow + quenched jets switch ( squares ) .
the error bars correspond to statistical uncertainties.,scaledwidth=46.0% ] the results for the sub - leading triangular flow mode are presented in fig .
[ fig:3c ] .
contrary to the results shown in sect .
[ sec : res ] , the @xmath74 values , calculated using the pure flow switch , is equal to zero for all centralities and at all @xmath0 .
this again shows that the assumption of the factorization of the two - particle fourier coefficients into a product of the single - particle anisotropies in the case of the pure flow switch is fully valid .
the pca method for studying flow , by its construction , may fully exploits the information contained in the covariance matrix formed from the two - particle fourier coefficients and thus may provide high sensitivity not only to the standardly defined flow measurements , but also to the influence of the initial - state fluctuations to the hydrodynamic flow . in difference of two - particle correlation method where the information was calculated by integrating over momentum of one of particles which form the pair , within the pca approach , the detailed information depends on the momenta of both particles of the pair . as the leading flow mode represents the hydrodynamic response to the average geometry , it is essentially equal to the single - particle anisotropy measured using the two - particle correlations method .
the sub - leading mode could be understood as the response to the event - by - event initial - state fluctuations which are the main source of the factorization - breaking effect .
the pca analysis of the pbpb collisions simulated by hydjet++ model at @xmath19 2.76 gev shows that the leading flow mode , @xmath61 , for @xmath82 represents dominant mode and qualitatively describes the experimentally measured @xmath8 from two - particle correlations .
additionally , hydjet++ model also shows existence of the sub - leading flow mode @xmath83 which magnitude is in a rather good agreement with the recent experimental finding from the cms collaboration .
this analysis may also provide new insights into the possible influence of the dynamics of the collision onto appearance of the sub - leading flow modes , and help to understand and improve modeling of the evolution of the strongly - coupled quark gluon plasma .
we acknowledge the financial support received from the ministry of education , science and technological development of the republic of serbia throughout the project 171019 .
i. arsene et al .
( brahms collaboration ) , nucl .
phys . a * 757 * , 1 ( 2005 ) b.b .
back et al .
( phobos collaboration ) , nucl .
757 * , 28 ( 2005 ) j. adams et al .
( star collaboration ) , nucl .
phys . a * 757 * , 102 ( 2005 ) k. adcox et al .
( phenix collaboration ) , nucl .
phys . a * 757 * , 184 ( 2005 ) k. aamodt et al .
( alice collaboration ) , phys .
b * 708 * , 249 ( 2012 ) b.b .
abelev et al .
( alice collaboration ) , jhep * 1506 * , 190 ( 2015 ) g. aad et al .
( atlas collaboration ) , phys .
c * 86 * , 014907 ( 2012 ) g. aad et al .
( atlas collaboration ) , jhep * 11 * , 183 ( 2013 ) g. aad et al .
( atlas collaboration ) , phys .
c * 90 * , 024905 ( 2014 ) s. chatrchyan et al .
( cms collaboration ) , eur .
j. c * 72 * , 2012 ( 2012 ) s. chatrchyan et al .
( cms collaboration ) , phys .
c * 87 * , 014902 ( 2013 ) s. chatrchyan et al .
( cms collaboration ) , phys .
c * 89 * , 044906 ( 2014 ) s. chatrchyan et al .
( cms collaboration ) , jhep * 02 * , 088 ( 2014 ) v. khachatryan
( cms collaboration ) , phys .
c * 92 * , 034911 ( 2015 ) j .- y .
ollitrault , phys .
d * 48 * , 1132 ( 1993 ) s. voloshin and y. zhang , z. phys .
c * 70 * , 665 ( 1996 ) a. m. poskanzer and s. a. voloshin , phys .
c * 58 * , 1671 ( 1998 ) b. alver and g. roland , phys . rev .
c * 81 * , 054905 ( 2010 ) p. danielewicz and g. odyniec , phys .
b * 157 * , 146 ( 1985 ) f. g. gardim , f. grassi , m. luzum , and j .- y .
ollitrault , phys .
c * 87 * , 031901 ( 2013 ) u. heinz , z. qiu , and c. shen , phys .
c * 87 * , 034913 ( 2013 ) r. bhalerao , j .-
ollitrault , s. pal , and d. teaney , phys .
* 114 * , 152301 ( 2015 ) a. mazeliauskas and d. teaney , phys .
c * 91 * , 044902 ( 2015 ) i.p .
lokhtin , l.v .
malinina , s.v .
petrushanko , a.m. snigirev , i. arsene , and k. tywoniuk , comput .
phys . commun . *
180 * , 779 ( 2009 ) t. sjostrand , s. mrenna , and p. skands , jhep * 0605 * , 026 ( 2006 ) i.p . lokhtin and a.m. snigirev ,
j. c * 45 * , 211 ( 2006 ) cms collaboration , principal component analysis of two - particle azimuthal correlations in pbpb and ppb collisions at cms , ( cern document server , 2015 ) , http://cds.cern.ch/record/2055291 . accessed 27 september 2015 j. milosevic for the cms collaboration , nucl .
phys . a * 956 * , 308 ( 2016 ) s. chatrchyan et al .
( cms collaboration ) , phys .
b * 724 * , 213 ( 2013 ) a. mazeliauskas , and d. teaney , phys .
c * 91 * , 044902 ( 2015 ) l. v. bravina , e. s. fotina , v. l. korotkikh , i. p. lokhtin , l. v. malinina , e . n. nazarova , s. v. petrushanko , a. m. snigirev , e. e. zabrodin , eur . phys .
j. c**75 * * , 588 ( 2015 ) | recent lhc results on the appearance of sub - leading flow modes in pbpb collisions at 2.76 tev , related to initial - state fluctuations , are analyzed and interpreted within the hydjet++ model . using the newly introduced principal component analysis ( pca ) method applied to two - particle azimuthal correlations extracted from the model data , the leading and
the sub - leading flow modes are studied as a function of the transverse momentum ( @xmath0 ) over a wide centrality range .
the leading modes of the elliptic and triangular flow reproduce rather well the @xmath1 and @xmath2 coefficients experimentally measured using the two - particle correlations . at high-@xmath0 region , where the effect is greatest , the sub - leading flow mode which corresponds to @xmath3 has a small non - zero value and slowly increases from central to peripheral collisions , while the one which corresponds to @xmath4 is even smaller and does not depend on centrality . for @xmath5 2
, the relative magnitude of the effect measured with respect to the leading flow mode shows a shallow minimum for semi - central collisions and increases for very central and for peripheral collisions . for @xmath5 3 case
, there is no centrality dependence .
the obtained results are in a rather good agreement with the experimental findings recently presented by the cms collaboration .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
the formation and evolution of galaxies is one of the outstanding problems in astrophysics , one which can be profitably engaged directly through detailed study of our own galaxy , the milky way ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . in the context of hierarchical structure formation
, galaxies such as the milky way grow by mergers and accretion of smaller systems , perhaps similar to what are now observed as dwarf galaxies .
these satellite galaxies torn apart by the tidal gravitational field of the parent galaxy are progressively disrupted , giving rise to trails of stellar debris streams along their orbits , spatial signatures that eventually disappear due to dynamical mixing .
after the accretion era ends , a spheroidal halo - like component is left from their collective assembly ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
of all the galactic components , it is indeed the stellar halo that offers the best opportunity for probing details of the merging history of the milky way ( see , e.g. , * ? ? ?
past explorations have demonstrated that there is a concrete possibility to identify groups of halo stars that originate from common progenitor satellites @xcite .
simulations show that a milky - way mass galaxy within a @xmath5cdm universe will have halo stars associated with substructures and streams ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these substructures , much like those seen in the halo system of the milky way , are sensitive to recent ( within the last @xmath6 gyr ) merging events , and are more prominent in the outer region of the halo ( galactocentric radii beyond @xmath7 kpc ) , whereas the inner - halo region appears significantly smoother .
based on data from the segue spectroscopic survey , @xcite found that metal - poor main - sequence turnoff stars in the inner - halo region of the milky way ( within @xmath8 kpc from the sun ) exhibit clear evidence for radial velocity clustering on small spatial scales ( they refer to these as echos , for elements of cold halo substructure ) .
they estimated that about @xmath9 of the inner - halo turnoff stars belong to echos , and inferred the existence of about @xmath10 echos in the entire inner halo volume .
@xcite suggest that the most likely progenitors of echos are dwarf spheroidal galaxies with masses on the order of @xmath11 m@xmath12 . in the solar neighbourhood ,
up to @xmath13 kpc of the sun , stellar streams have also been discovered as overdensities in the phase - space distribution of stars , integrals of motion and action - angle variables ( see * ? ? ? * for a review ) .
prominent examples are the two stellar debris streams in the halo population passing close to the sun detected by @xcite when combining high - quality hipparcos proper motions with ground - based observations . formed via destruction of a satellite
whose debris now occupy the inner halo region with no apparent spatial structure , these streamers retain very similar velocities and are seen as clumps in angular momentum space where stars from a common progenitor appear rather confined @xcite .
besides the helmi stream , @xmath14 centauri @xcite , the kapteyn and arcturus ( e.g. , * ? ? ?
* ) streams , @xcite lists a few other halo substructures found in the solar neighborhood : these are still small numbers compared to the few hundred streams expected ( i.e. @xmath15-@xmath16 , * ? ? ?
* ; * ? ? ?
actually , recovering fossil structures in the inner halo is considerably more difficult , as strong phase - mixing takes place .
this degeneracy can only be broken with 6d ( phase - space ) or 7d ( including abundances ) information achievable by integrating astrometry , photometry , and spectroscopy .
the sdss - gsc ii kinematic survey ( from now on sgks ) we exploit here was produced to serve this task ( see * ? ? ?
* ) . in the future ,
new ground- and space - based surveys such as gaia ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , gaia eso survey ( ges ; * ? ? ?
* ) , and lamost @xcite will provide high - precision data that will usher us in a new era of milky way studies . in sect .
[ sec:2 ] , we introduce the data used to isolate a sample of nearby halo subdwarfs from the sgks catalogue .
the kinematic and orbital properties of the local halo subdwarf population are discussed in sect .
[ sec:3 ] , where we present algorithms to search for kinematic substructures , recovering known streams @xcite , as well as new kinematic overdensities . in sect .
[ sec:4 ] , we present the high - resolution n - body numerical simulations of four minor mergers used to study galaxy interactions and the properties of accretion events in the vicinity of the sun . in sect .
[ sec:5 ] we investigate the impact of observational errors resulting from current ground - based data and from high accurate data expected from the gaia satellite . finally , in sect .
[ sec:6 ] , we compare observations to these numerical simulations and infer the nature of the detected fast moving groups .
lcccccccccc @xmath17&@xmath18&@xmath19 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 +
this study is based on a new kinematic catalogue , derived by assembling spectro - photometric stellar data from the seventh data release of the sloan digital sky survey ( sdss dr7 ; * ? ? ?
* ) , which included data from the sloan extension for galactic understanding and exploration ( segue ; * ? ? ?
* ) , supplemented by astrometric parameters extracted from the database used for the construction of the second guide star catalogue ( gsc ii ; * ? ? ?
this sdss - gsc ii catalogue contains positions , proper motions , classification , and @xmath26 photometry for @xmath27 million sources down to @xmath28 , over @xmath29 square degrees .
proper motions are computed by combining multi - epoch positions from sdss dr7 and the gsc ii database ; typically , @xmath30 observations are available for each source , spanning up to @xmath31 years . the typical formal errors on those proper motions are in the range @xmath32 per coordinate for @xmath33 , comparable with the internal precision of the sdss proper motions computed by @xcite .
although much of the photographic material ( schmidt plates ) used to derive first epoch information is in common with @xcite , plate digitisation and measurement processes , and the calibration methods that led to the first epoch positions were somewhat different , with the munn et al. data coming from the usno - b project .
of particular relevance is the minimisation of systematic errors that can affect proper motion accuracy , a true driver in analysis like those conducted in this study .
an accurate validation of our proper motions was discussed in @xcite .
radial velocities and astrophysical parameters are available for about @xmath34 sources cross - matched with the sdss spectroscopic catalogue typical accuracy are of @xmath35 in line of sight velocity , @xmath36 k in effective temperature , @xmath37 , @xmath38 dex in surface gravity , @xmath39 , and @xmath40 dex in metallicity , @xmath41}$ ] , as estimated within the segue spectral parameter pipeline ( sspp ; i. e. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we specify that the sample includes only objects with no problems related to the spectrum , and classified without any cautionary flag by the sspp . in case of multiple spectra
, we take the spectrum with the highest signal - to - noise ratio . from the sdss - gsc ii catalogue , we select as tracers sources with @xmath42 k and @xmath43 , corresponding to fgk dwarfs .
the observed magnitudes are corrected for interstellar absorption via the extinction maps of @xcite based on the @xmath44 resolution cobe / dirbe dust map , that we preferred to the more recent , but of inferior resolution ( @xmath45-@xmath46 ) , reddening maps published by @xcite .
then , we transformed the @xmath47 to the sdss photometric system by adopting the extinction ratio @xmath48 ( from table 1 of * ? ? ? * ) , that is appropriate for our fgk dwarf sample .
photometric distances good to @xmath49 are computed by means of the photometric parallax relation established for fgk main - sequence stars by @xcite . here
, the metallicity - dependent absolute magnitude relations , @xmath50})$ ] , use the spectroscopic @xmath41}$ ] instead of the photometric metallicity adopted by @xcite .
we also apply the additional colour thresholds from @xcite in order to remove turn - off stars , whose estimated @xmath51 may be affected by residual systematic errors .
galactic space - velocity components , @xmath52 , and @xmath53 indicate galactic velocity components relative to the local standard of rest and follow the convention with @xmath54 positive toward the galactic center , @xmath52 positive in the direction of galactic rotation , and @xmath55 towards the north galactic pole . ]
are estimated under the assumption that the sun is at a distance of @xmath6 kpc from the centre of the milky way , the local standard of rest ( lsr ) rotates at @xmath56 about the galactic center , and the peculiar velocity of the sun relative to the lsr is @xmath57 @xcite . finally , in order to minimise the effect of outliers ( e.g. mismatches , blends and sources with low s / n ) and therefore obtain a sample with accurate distance and kinematics suitable for our stellar stream search , we impose a threshold on proper - motion errors ( @xmath58 per component ) , constrain magnitudes to the range @xmath59 , limit the errors on the derived velocity components to better than @xmath60 , and remove total space velocities above @xmath61 .
these are the properties of the @xmath62 stars listed in the sgks catalogue .
among the full sample of fgk dwarfs from the sgks catalogue , we have selected specific sub - samples of tracers of the galactic halo population in the inner - halo region , and analysed their phase - space distribution . here , we focus on a sample of @xmath0 metal - poor stars ( @xmath1<-1.5}$ ] ) outside the galactic plane ( @xmath63 kpc ) , and located within @xmath2 kpc of the sun . within this volume ,
the selected sample has median errors on @xmath54 , @xmath52 , and @xmath53 of @xmath64 , @xmath65 , and @xmath66 , respectively ; this results in errors in the velocity difference between stellar pairs not exceeding @xmath67 .
such a value is suited for careful investigations of substructure , as the kinematic analysis presented below will show . + from the selected sample we measure the mean velocities @xmath68 , the velocity ellipsoid @xmath69 , and the correlations among velocity components ( @xmath70 as reported in table [ table:1 ] .
the kinematic properties of the selected tracers are representative of the halo population in the vicinity of the sun ( e.g. , * ? ? ?
the significant correlation @xmath71 between the radial and vertical velocity components indicates a tilt of the velocity ellipsoid ( fig .
[ prf : fig1 ] , right panel ) . using the tilt formula ( see , e.g. ,
* ) @xmath72 and the values in table [ table:1 ] for the correlation coefficient and velocity dispersions along the @xmath54 and @xmath53 axes , we derive a tilt angle of @xmath73 , revealing that the @xmath74 distribution points toward the galactic center .
in fact , for our halo sample of 2417 fgk subdwarfs with @xmath75 1.2 kpc and @xmath76 kpc , we estimate a mean position angle @xmath77 .
this result is fairly consistent with the tilting effects on the velocity ellipsoids due to the gravitational potential produced by the stellar disk and dark matter halo ( * ? ? ?
* and references therein ) .
we also measure smaller but stastistically significant correlations in @xmath78 and @xmath79 velocity - planes . in the following ,
we look for halo streamers in the high velocity tail of the @xmath80 velocity distribution , where kinematic substructures are more easily detected . in
order to select high velocity stars , we model the velocity distribution as a tilted schwarzschild ellipsoid : @xmath81 where @xmath82 is the velocity function defined by : @xmath83 here , @xmath84 represents the determinant of the symmetrical matrix @xmath85 of the correlation coefficients @xmath86 ( for @xmath87 ) , and @xmath88 designate the cofactor of the corresponding correlation element in @xmath85 ( e.g. , * ? ? ? * ) . figure [ prf : fig1 ] shows the kinematic distribution for the individual components , @xmath89 , of the space - velocity vector for the full sample of @xmath0 selected halo stars ; the @xmath90 objects comprising the sample of the @xmath9 highest velocity tail are represented with crosses .
as expected , the overall velocity distribution is relatively smooth , because of the strong phase - mixing that takes place in the inner - halo region , and slowly prograde ( e.g. , * ? ? ?
+ however , as their motions ( in direction and speed ) are well separated from those of the other nearby subdwarfs , we intend to study the degree of clumpiness of the @xmath9 fastest - moving objects .
the case study is that , of all the objects passing within a few kiloparsecs of the sun , some are part of a diffuse local stellar halo , while some could be debris of accretion events and remnants from the outer - halo population currently in the solar neighbourhood . before starting to look for kinematic substructures
, we check for thick disk stars that possibly contaminate our halo tracers . here
, we applied to kinematic method described in @xcite and estimate the fraction of subdwarfs that is consistent with the 3d velocity distribution of the thick disk population . by assuming a velocity ellipsoid ,
as estimated by @xcite , and a rotation velocity @xmath91 km s@xmath92 , as measured by @xcite for metal - poor thick disk stars with @xmath1}\simeq -1 $ ] dex , we found at a @xmath93 ( i.e. @xmath94 ) confidence level a @xmath95 maximum contamination of thick disk in the whole sample of 2417 halo tracers .
instead , no contaminant is expected among the subsample of the @xmath9 fastest objects .
we use the samples described above to detect and subsequently identify kinematic halo substructures in the solar neighbourhood as groups of stars moving with similar velocities and directions .
detection is accomplished by performing a statistical test based on individual kinematics aimed at quantifying possible deviations from a smooth distribution of the background halo ; cluster analysis in velocity space is then applied for final confirmation of the substructures .
the amount of kinematic substructures that cosmology might leave in the volume is quantified by means of the cumulative two - point correlation function , @xmath96 , on the paired velocity difference @xmath97 that measures the excess in the number of stellar pairs moving within a given velocity difference when compared to a representative random smooth sample ( cfr . * ? ? ?
* for more details ) . here
the random points were drawn from a multivariate distribution obtained from the observed data set by random permutations of the order of the velocity components @xmath98 and @xmath53 , after fixing @xmath54 ; finally , the actual random pairs are obtained after averaging over ten independent realisations . this function
is computed over the full sample of @xmath0 halo stars , and separately for the sub - sample of the @xmath90 fastest - moving stars , corresponding to the @xmath9 high - velocity tail .
a statistical excess of stars with small pairwise velocity differences indicates the presence of likely streamers made of objects with coherent kinematics .
figure [ prf : fig2 ] shows , using bins of @xmath99 width .
] the two - point correlation function @xmath96 for the full sample of @xmath0 halo stars ( dots ) and for the subset of the @xmath9 fastest - moving stars ( diamonds ) .
while weak for the full sample , there is a statistically significant signal ( @xmath100 ) for the subset of the fastest stars , that peaks at @xmath101 : the excess of pairs of stars with similar velocities is very noticeable , and is a direct indication of the presence of kinematical clumps .
fastest - moving subset ( diamonds ) shown in fig .
[ prf : fig1 ] .
the error bars are derived from poisson s statistics of the counts . ] in the following , among the sample of the @xmath90 fastest stars , we focus on the objects with paired velocity differences less than @xmath101 , which yield the statistically significant signal seen in fig .
[ prf : fig2 ] .
in addition , we exclude _ isolated _ pairs , i.e. , groups " with only two objects .
this further selection certainly reduces the number of detected members , however it makes the following analysis more robust by decreasing the contamination of false positives .
the final sample is made of @xmath3 stars .
+ in order to classify these @xmath3 objects , we perform @xmath102medoids clusteringmedoids clustering developed as part of the _ r project for statistical computing _ : www.r-project.org ] in the 3d velocity space that defines the number of kinematic substructures and their members .
this unsupervised learning algorithm is able to group data into a pre - specified number of clusters that minimises the rms of the distance ( in velocity space ) to the center of each cluster .
the original data set is initially partitioned into clusters around @xmath103 data points referred to as the medoids , then an iterative scheme ( pam , for partitioning around medoids ) is applied to locate the medoids that achieve the lowest configuration cost " .
the algorithm employed by pam , similar to the @xmath103-means clustering algorithm , is more robust to outliers and obtains a unique partitioning of the data without the need for explicit multiple starting points for the proposed clusters ( see , e.g. , * ? ? ?
* ; * ? ? ?
there is no general theoretical solution for finding the optimal number of clusters for any given data set . increasing @xmath103 results in the error function values formally much smaller , but this increases the risk of overfitting . in order to keep the final identification safe and simple , we compared the results of runs with different @xmath103 classes , and the best @xmath103 resulted following visual inspection of the generated distribution .
the solution adopted here is with @xmath104 : for @xmath105 the clusters returned by the algorithm would contain a mixture of the natural underlying groups ( e.g. , prograde and retrograde members in the same kinematic clump ) ; for @xmath106 natural groups further partition into artificial " subgroups .
the five kinematic substructures detected are visualised in fig .
[ prf : fig3 ] with different colours .
the individual kinematic properties of the @xmath3 stars belonging to the five kinematics groups are listed in table [ table:2 ] .
other methods have been utilised to isolate groups of stars in the halo like , e.g. , the interesting approach specifically developed for the virgo stellar stream by @xcite . on the other hand , the fact that we have the complete set of 3d kinematical data and that the whole sample is confined within @xmath2 kpc from the sun ( i.e. , the distance segregation is implicitly implemented in our sample ) suggests the direct use of a classical clustering algorithm like pam as the method of choice .
the space of adiabatic invariants allows better identification of the different possible merging events that might have given rise to the observed substructures .
clumping should be even stronger , since all stars originating from the same progenitor should have very similar integrals of motion , resulting in a superposition of the corresponding streams ; that is , the initial clumping of satellites are present even after the system has completely phase - mixed ( e.g. , * ? ? ?
in this study we focus on the plane defined by the components of angular momentum , @xmath107 , @xmath108 , and @xmath109 . here
@xmath110 , @xmath111 , and @xmath112 . ] in and out the plane of the galaxy s disk , i.e. @xmath113 and @xmath114 respectively .
since for a local sample @xmath115 , we remark that @xmath116 is dominated by the velocity perpendicular to the plane and @xmath117 measures the amount of rotation of a given stellar orbit .
essentially , stars with high / low @xmath113 are on high / low - inclination orbits ; stars with @xmath118 are on retrograde orbits and stars with @xmath119 are on prograde orbits .
figure [ prf : fig4 ] shows the distribution of the selected sample within @xmath2 kpc of the sun in the angular momentum diagram @xmath120 versus @xmath113 . as in fig .
[ prf : fig3 ] , the @xmath9 fastest - moving objects are plotted as crosses , and with the star symbols we mark the group members identified in sect .
[ sec:3.2 ] .
different colours indicate the stars associated with the different lumps recovered by the cluster analysis in velocity space .
the solid lines show the loci of the known kinematic structures detected by 1 .
@xcite at + @xmath121 and + @xmath122 , 2 .
@xcite at + @xmath123 and + @xmath124 , 3 .
@xcite at upper part was derived by remapping the @xmath120-@xmath125 region shown in fig .
4 of that paper ; therefore @xmath126 . ]
+ @xmath127 and + @xmath128 llrrrrrr 1@xmath129 / pink & 52209 - 0694 - 596 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath135 + & 52721 - 1050 - 418 & @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 & @xmath141 + & 53315 - 1907 - 393 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & @xmath146 & @xmath147 + & 53349 - 2066 - 511 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 & @xmath153 + & 54175 - 2472 - 370 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath147 + & 54574 - 2904 - 114 & @xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 + & 54577 - 2906 - 088 & @xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 + & 54621 - 2191 - 177 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 + & 54624 - 2189 - 236 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 & @xmath182 + & 54629 - 2902 - 237 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 + 2@xmath129 / blue & 52316 - 0559 - 336 & @xmath189 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 + & 53084 - 1368 - 399 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath170 + & 53242 - 1896 - 109@xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 + & 53262 - 1900 - 359 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 + & 53293 - 1906 - 633@xmath200 & @xmath213 & @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath218 + & 53467 - 2110 - 134 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 + & 53712 - 2314 - 639 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 + & 53726 - 2306 - 188@xmath200 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 + & 53907 - 2209 - 540@xmath200 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 + & 54178 - 2452 - 540 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + & 54380 - 2323 - 448 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 + & 54479 - 2867 - 531 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 + & 54530 - 2889 - 458 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 & @xmath266 + & 54554 - 2918 - 615 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & @xmath272 + & 54580 - 2905 - 169 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 + 3@xmath279 / red & 54179 - 2567 - 458 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath260 + & 54463 - 2856 - 571 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath212 + & 54551 - 2394 - 228 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 + & 54562 - 2920 - 596 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & @xmath301 + & 54569 - 2900 - 046 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + 4@xmath279 / green & 52059 - 0597 - 072 & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath212 + & 52338 - 0788 - 070 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath318 + & 52942 - 1509 - 488 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath260 + & 53710 - 2310 - 141 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 + & 53762 - 2381 - 588 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath153 + & 53800 - 2383 - 625 & @xmath183 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 + & 53823 - 2240 - 213 & @xmath340 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath345 + & 53874 - 2173 - 414 & @xmath346 & @xmath347 & @xmath348 & @xmath349 & @xmath350 & @xmath351 + & 54154 - 2701 - 341 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & @xmath356 & @xmath236 + & 54169 - 2413 - 090 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 + & 54234 - 2663 - 321 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath170 + & 54539 - 2894 - 314 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath372 & @xmath254 + & 54539 - 2894 - 632 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 + & 54544 - 2459 - 072 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 + & 54557 - 2177 - 009 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 + & 54568 - 2899 - 316 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 + & 54595 - 2932 - 091 & @xmath397 & @xmath398 & @xmath399 & @xmath400 & @xmath401 & @xmath402 + & 54597 - 2561 - 326 & @xmath403 & @xmath404 & @xmath405 & @xmath406 & @xmath407 & @xmath408 + & 54616 - 2460 - 420 & @xmath409 & @xmath410 & @xmath410 & @xmath411 & @xmath412 & @xmath413 + & 54616 - 2460 - 616 & @xmath414 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath272 + & 54631 - 2911 - 151 & @xmath419 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath408 + 5@xmath279 / yellow & 53035 - 1433 - 600 & @xmath424 & @xmath425 & @xmath426 & @xmath427 & @xmath428 & @xmath429 + & 53240 - 1894 - 079 & @xmath430 & @xmath431 & @xmath432 & @xmath433 & @xmath434 & @xmath435 + & 53315 - 1907 - 353 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 & @xmath441 + & 53770 - 2387 - 010 & @xmath442 & @xmath443 & @xmath444 & @xmath445 & @xmath446 & @xmath447 + & 53848 - 2437 - 060 & @xmath448 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 + & 53876 - 2134 - 516 & @xmath454 & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 + & 53918 - 2539 - 196 & @xmath460 & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath413 + & 54082 - 2325 - 126 & @xmath465 & @xmath466 & @xmath467 & @xmath468 & @xmath469 & @xmath470 + & 54156 - 2393 - 459 & @xmath471 & @xmath472 & @xmath473 & @xmath474 & @xmath475 & @xmath236 + & 54243 - 2176 - 476 & @xmath476 & @xmath477 & @xmath478 & @xmath479 & @xmath480 & @xmath182 + & 54271 - 2449 - 590 & @xmath481 & @xmath482 & @xmath483 & @xmath484 & @xmath485 & @xmath486 + & 54368 - 2804 - 351 & @xmath487 & @xmath488 & @xmath489 & @xmath490 & @xmath491 & @xmath447 + & 54536 - 2871 - 426 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath496 & @xmath413 + & 54594 - 2965 - 227 & @xmath497 & @xmath498 & @xmath499 & @xmath500 & @xmath501 & @xmath260 + & 54594 - 2965 - 272 & @xmath502 & @xmath503 & @xmath504 & @xmath505 & @xmath506 & @xmath507 + & 54616 - 2929 - 342 & @xmath508 & @xmath509 & @xmath510 & @xmath511 & @xmath512 & @xmath329 + the most noticeable feature in fig .
[ prf : fig4 ] is certainly the kinematic group corresponding to the stream found by @xcite . here , we identify @xmath513 subdwarfs , including @xmath514 stars already detected by @xcite , and @xmath515 new members . by inspection of fig .
[ prf : fig3 ] we notice that the @xmath516 members belonging to group 1 ( pink star symbols ) run along near - parallel orbits and cross the milky way s disk at high speed from south to north , and the @xmath517 objects in group 2 ( blue star symbols ) cross the milky way s disk at similar speed and angle , but from north to south .
the three remaining lumps of fast - moving stars ( red group 3 , green group 4 , and yellow group 5 ) appear on the retrograde side of fig .
[ prf : fig4 ] . the pentagonal box confined by the dashed line includes most of the members of group 3 ( @xmath518 stars ) , group 4 ( @xmath515 stars ) , and group 5 ( @xmath519 stars ) .
these groups , and in particular the small group 3 , do not appear to be easily associated with known streams and , in sect .
[ sec:5 ] , we discuss the possibility that all these stars come from a common progenitor or from three different merging events . anyhow , we note that group 4 might be the parent populations of the counter - rotating `` outliers '' , with @xmath520 , found by @xcite , while the slightly retrograde group 5 ( @xmath521 ) seems to be related to the kinematic structure found by @xcite , and confirmed by both @xcite and @xcite .
these authors have also discussed the possibility that such a stream is formed by the tidal debris of @xmath14 cen in the solar neighborhood , even though @xcite have recently ruled out this hypothesis after detailed analysis of the chemical abundance of this group with respect to the well - known peculiar properties of @xmath14 cen .
we explore a simulated inner halo based on a set of _ four _ high - resolution numerical n - body simulations of minor mergers .
we analyse the kinematics and orbital properties of these simulations in order to investigate and characterise detectable signatures .
it is useful to point out that these simulations are not an attempt to `` fit '' the observations , but they represent four merging events that we assume as representative , in terms of inclination and rotation , of the initial orbits of the satellites . in particular , these choices allow us to analyse the two cases suffering maximum and minimum dynamical friction .
we use a set of high - resolution numerical n - body simulations which simulate minor mergers of prograde and retrograde orbiting satellite halos within a dark matter main halo @xcite .
the main dm halo , which contains a stellar , rotating exponential disk , has a nfw radial density profile @xcite , with a mass ( @xmath522 m@xmath12 ) , radius ( @xmath523 kpc ) , and concentration ( @xmath524 ) , appropriate for a milky way - like dm halo at redshift @xmath525 ; the spin parameter is set and @xmath526 were both studied at lower resolution and the results compared ; the differences are such that they have no bearing on the results presented in this paper . ] to @xmath526 .
the satellite is represented by a secondary dm halo containing a stellar bulge , and a hernquist radial density profile @xcite ; the spin parameter is set to @xmath527 .
the mass ratio , @xmath528 , is similar to the estimated mass ratio of the milky way relative to the large magellanic cloud .
the main physical parameters of our simulated mergers are listed in table [ table:3 ] .
we consider prograde mergers , in which the satellite co - rotates with the spin of the disk , as well as retrograde mergers , with a counter - rotating satellite .
we analyse two orbits : a low - inclination orbit with a @xmath529 tilt with respect to the disk plane , and a high - inclination orbit with a @xmath530 tilt .
initially , the particles of the small system ( satellite galaxy ) orbiting around the ( otherwise static ) disk galaxy are all strongly concentrated in space , and share essentially the same motion .
the initial conditions ( inclination , position , and velocity ) of the main system and the four impacting satellites , cosmologically motivated @xcite , are summarised in table [ table:4 ] . from the grid of simulations by @xcite
, we chose four impactors , all of which having the mass of the large magellanic cloud .
larger masses would affect the stability of the stellar disk , and this is not consistent with a milky way - like galaxy .
conversely , smaller masses would produce minor signatures in our local halo sample .
the four simulations are compiled using the public parallel treecode gadget2 @xcite on the cluster matrix at the caspur ( _ consorzio interuniversitario per le applicazioni del supercalcolo _ ) consortium , rome .
all systems were left to evolve for @xmath531 gyr ( about @xmath519 dynamical timescales of the main halo ) .
after this time , the four satellites have completed their merging with the primary halo .
the final @xmath532 distribution of the inner satellite star particles and host disk , in both the retrograde and prograde cases , as well as for the high and low inclinations , is shown in fig .
[ prf : fig5 ] . lrlrrll[hb ! ] main & @xmath533 & @xmath534&@xmath535 & @xmath535 & 4 & 20 + satellite & @xmath536 & @xmath537&@xmath538 & @xmath539 & 0.709 & ... + lrlrrrrrrr main & @xmath540 & @xmath541 & 0.00 & @xmath542 & @xmath542 & @xmath542 & @xmath542 & @xmath542 & @xmath542 + satellite 1 & @xmath529 & retrograde & @xmath543 & @xmath544 & @xmath545 & @xmath546 & @xmath547 & @xmath548 & @xmath549 + satellite 2 & @xmath529 & prograde & @xmath543 & @xmath544 & @xmath545 & @xmath546 & @xmath550 & @xmath548 & @xmath549 + satellite 3 & @xmath530 & retrograde & @xmath551 & @xmath552 & @xmath553 & @xmath554 & @xmath555 & @xmath556 & @xmath557 + satellite 4 & @xmath530 & prograde & @xmath551 & @xmath552 & @xmath553 & @xmath554 & @xmath558 & @xmath556 & @xmath557 + any satellite can in principal be slowed by dynamical friction exerted on it by disk and halo particles .
it is known that an object , such as a satellite , of mass @xmath559 , moving through a homogeneous background of individually much lighter particles with an isotropic velocity distribution suffers a drag force @xcite : @xmath560 where @xmath561 is the speed of the satellite with respect to the mean velocity of the field , @xmath562 is the total density of the field particles with speeds less than @xmath561 , and @xmath563 is the coulomb logarithm @xcite .
we expect that , the higher the @xmath561 , the weaker is the dynamical friction force .
retrograde satellites are expected to suffer weaker dynamical friction with respect to prograde ones , since in the first case the velocity of the satellite is opposite to that of the disk . as a consequence ,
prograde orbits decay faster .
this effect is even more evident for low - inclination orbits .
another important effect that occurs during mergers is the tidal disruption of satellites .
while tidal disruption is most important near the centre of the main halo , where the gravitational potential is changing more rapidly , dynamical friction is exerted both by the main halo dm particles and by the disk star particles .
we analyse the observational signature left by the satellite stars after selecting particles in a sphere of @xmath2 kpc radius centered at the sun ( @xmath564 kpc from the galaxy center ) , and with @xmath63 kpc .
this last constraint is introduced so that the simulated and observed samples can be compared within a similar volume of the inner halo .
figure [ prf : fig6 ] shows the kinematic distribution ( velocity projections ) of our simulated inner halo .
the different colours indicate the association of the 3902 debris stars with different progenitors : the low / high - inclination retrograde satellites ( 761/616 green / red dots ) , and the high / low - inclination prograde satellites ( 966/1559 blue / yellow dots ) .
+ the angular momentum distribution of the satellite debris is shown in fig .
[ prf : fig7 ] . despite the chaotic build up of the parent halo
, it appears that objects from accreted satellites remain confined in limited portions of the @xmath565 plane .
the satellite on a low - inclination prograde orbit ( yellow particles ) , which suffers more from dynamical friction , quickly loses its orbital energy and proceeds to the inner regions of the main halo @xcite . for this reason
less particles are left in the outer - halo , see top - right panel of fig .
[ prf : fig5 ] .
it is worth noticing that the high inclination prograde satellite suffers the effect of dynamical friction as well , as a result of its co - rotation with the disk .
this effect acts in producing a consistent mass of debris in the solar region having @xmath113 ranging between @xmath16 @xmath566 and @xmath567 @xmath566 ( blue points in fig .
[ prf : fig7 ] ) .
on the other hand , retrograde satellites experience weaker dynamical friction and leave more particles in the outer - halo region , since their orbits have a longer decay time and longer periods .
thus , tidal stripping ( see e.g. , * ? ? ?
* ) can act longer and more efficiently when the satellite is still orbiting at high velocity , and we see that a better populated high - velocity tail results ( compare fig .
[ prf : fig7 ] to fig .
[ prf : fig10 ] ) .
the impact of dynamical friction on the two configurations considered for the retrograde satellites indicates that the high - inclination case is the one less affected by this force , which again results in efficient stripping when the satellite has high orbital velocity .
therefore , such stripping takes place over a large spatial region , and for the conservation of 6d phase - space density , by virtue of liouville s theorem , we expect a small variance in velocity space and in the plane of angular momenta .
this is indeed observed for the red particles with respect to the green ones in figs .
[ prf : fig6 ] and [ prf : fig7 ] .
finally , the effect of both gravitational feedback and dynamical friction on the satellites , which lead to loss of stars at different passages with different energies , is clearly evident for the case of low inclination prograde orbit in fig .
[ prf : fig7 ] at around @xmath568 and @xmath569 .
here we investigate the effects of observational errors on our simulated data , and show how more accurate kinematic data to be provided by future surveys can improve detection and characterisation of halo streams .
moreover , we compare _ actual _ observations with the distributions of debris resulting from the four simulated satellites presented in the previous section , and discuss the orbital properties of the parent dwarf galaxies , possibly responsible for accreting on the milky way halo .
we perturb the original simulations by convolving the true " data with two cases of error distributions .
first , we adopt the accuracy of our sgks catalogue as representative of the quality of current wide - field surveys .
then , we assume the mean accuracy expected from the forthcoming gaia catalogue combined with complementary deep spectroscopic data from on - going and future surveys such as ges @xcite .
the true positions and velocities of each particle are first transformed into their astronomical observables ( @xmath570 , @xmath571 , @xmath572 , or @xmath573 and @xmath574 , @xmath575 , @xmath576 ) ; then the expected observational errors are added to distance modulus ( or directly to parallax , in the case of the gaia - like simulation ) , radial velocity , and proper motion , according to table [ table:5 ] .
the precision in distance is taken to @xmath577 mag ( i.e. , @xmath578 ) for the photometric distances estimated from the sgks catalogue , and to @xmath579 for the final precision on trigonometric parallaxes measured by gaia . in proper motion ,
the precision is assumed to be @xmath580 for ground - based observations , and @xmath581 for gaia .
the precision in the radial velocity is taken to be @xmath582 for the sdss measurements , and @xmath583 for the ges spectroscopic survey .
these quantities are finally transformed back to observed positions vectors and space velocities .
ccccc sgks & @xmath584 mag & 2000 & 10 + gaia / ges & @xmath579 & 20 & 1 + lcrrrrr true " simulation & 2874 & @xmath585 & 262 & 201 & 601 & 39 + observed " sgks & 2874 & @xmath586 & 233 & 170 & 406 & 26 + observed " gaia / ges & 2874 & @xmath587 & 263 & 191 & 570 & 37 + we explore a simulated inner halo based on the superposition of the four simulations of minor mergers and a smooth local component with the same kinematic properties of the observed sample ( table [ table:1 ] ) . consistently with the findings of @xcite and @xcite , we assumed a debris total fraction of @xmath9 within @xmath2 kpc from the sun .
in the true " ( simulated ) catalogue , of the @xmath588 particles with @xmath63 kpc , @xmath589 are part of the local mock halo , while the remaining @xmath590 are debris from the satellites shown in figs .
[ prf : fig6 ] and [ prf : fig7 ] . in the following discussion we focus on the particles of the accreted component that belong to the @xmath9 high velocity tail .
table [ table:6 ] reports the number of particles belonging to each satellite for the pure simulation and the other two cases accounting for observational errors . figure [ prf : fig8 ] shows the region of the @xmath74 plane occupied by the @xmath9 high velocity tail of the resulting simulated milky way inner halo ( right panels ) , as superposition the accreted component ( middle panels ) and the smooth spheroid ( left panels ) . the synthetic
observed " catalogue shown in the top panels represent the current picture , according to the sgks error model .
the bottom panels show the distribution of the high velocity particles as promised by gaia .
the upper panels indicate that distinguishing in velocity space the satellites that gave rise to each of the different moving groups with the extant data is a non - trivial task . on the other hand ,
as the inspection of the lower panels reveals , much of the substructures shown in the middle panels becomes visible again thanks to the superior precision that gaia will achieve . in order to quantify the amount of kinematic substructures present among the 2874 fastest - moving particles ,
we compute the cumulative velocity correlation function described in sect .
[ sec:3.2 ] .
the analysis is performed over three synthetic catalogues : the true " simulation , and two lists derived from the true values after perturbing them with either sgks - like errors or the errors expected for the gaia / ges surveys . .
the filled dots trace the correlation function of the pure simulation , while squares and diamonds depict the correlation after convolution with current ( i.e. , sgks ) or gaia / ges - like errors , respectively .
error bars are derived from poisson s statistics of the counts . ]
figure [ prf : fig9 ] shows , using bins of width @xmath591 up to kinematical separations of @xmath592 , the results for the two - point correlation function @xmath96 for the true " case ( dots ) , the sgks - like ( squares ) and gaia / ges - like ( diamonds ) catalogues .
the clear signal in the first bins , peaking at @xmath593 , evidences an excess of particles moving with similar velocities with respect to what expected from a fully random sample . in the case of the pure simulation
, the two - point velocity correlation function attains a maximum signature of @xmath594 . for the sgks - like catalogue the maximum signal
has an @xmath595% drop to an observed " value of @xmath596 .
the recovery of the intrinsic correlation signal is truly remarkable when looking at the correlation function for the gaia - like case : in fact , we measure @xmath597 , corresponding to @xmath598 of the original signal . figure [ prf : fig8 ] provides a nice visual " confirmation of the recovery in substructure visibility .
as the space of adiabatic invariants is important to gain more insight into the properties of the kinematic substructures detected ( sect . [ sec:3.4 ] ) , we compare the @xmath599 distributions of the observed groups with the results of the simulations , taking into account the effect of the observational errors .
this is shown in fig .
[ prf : fig10 ] , where the top panel corresponds to the sgks - error simulation , while the bottom panel reproduces what will hopefully be seen with the final gaia catalogue . the black star symbols in the upper panel of fig .
[ prf : fig10 ] represent the @xmath3 high velocity objects we found from our statistical analysis in the same volume and shown in fig .
[ prf : fig4 ] as colored stars . with current data ,
different satellites mix over some regions so that a discrete classification is not always straightforward .
the bottom panel of fig .
[ prf : fig10 ] clearly shows that this situation is highly improved with gaia - like data .
+ we see that our groups 1 and 2 , corresponding to the stream of @xcite , are consistently associated with the high inclination prograde satellite ( blue dots ) . because of dynamical friction ( cfr .
[ sec:4.3 ] ) , this satellite includes a low @xmath113 component shown in the full sample ( fig .
[ prf : fig7 ] ) that is not part of the high velocity tail ( fig .
[ prf : fig10 ] ) .
thus , these simulated observations " suggest the possible presence in the @xcite stream of debris with lower @xmath113 yet to be discovered . of particular interest is the case of the retrograde kinematic groups .
in fact , neither the high - inclination simulated satellite nor the one at low - inclination appear to fairly match the observed groups 3 , 4 , and 5 , i.e. the black stars with @xmath600 in fig .
[ prf : fig10 ] . actually , these three groups appear to occupy an _ intermediate _ region between the debris of the two simulated retrograde satellites .
furthermore , in sect . [ sec:3.4 ]
we remark that the observed groups 3 , 4 , and 5 do not well match the streams detected by @xcite and @xcite . for this reason , we suggest that these three groups may represent the debris of an unique progenitor accreted along an initial retrograde orbit having an intermediate inclination in the range comprised between @xmath529 and @xmath530 . in alternative
, these groups could belong up to three different impacting satellites on retrograde orbits with inclinations in that same range .
the results presented in this section show that the methodology proposed is certainly capable of detecting fossil signatures as kinematic substructures among high - velocity stars . from the data at our disposal , there is clear indication that more debris are found from dwarf galaxies on high - inclination prograde and retrograde orbits , as well as on low - inclination retrograde ones .
we have not identified any debris coming from low - inclination prograde satellites and this might be a limitation intrinsic to the methodology of looking at structures in the space motions of very high velocity stars .
future work will have to investigate these issues .
we have explored the solar neighborhood of the milky way through the use of spectro - photometric data from the sloan digital sky survey and high - quality proper motions derived from multi - epoch positions extracted from the guide star catalogue ii database . a sample with accurate distances , space velocities , and metallicities is selected as a tracer of the inner - halo population resulting in @xmath0 subdwarfs with @xmath1}<-1.5 $ ] and @xmath601 kpc within @xmath2 kpc of the sun .
this set is then analysed to identify and characterise kinematic streams possibly arising from merging events . in angular momenta space , the two prograde groups we have identified ( groups 1 and 2 in table [ table:2 ] ) appear confined in the region encompassing the stream first identified by @xcite among red giants and rr lyrae within @xmath602 kpc of the sun .
our analysis found @xmath513 additional subdwarf members belonging to that same stream : @xmath514 are in common with those found by @xcite , while the other @xmath515 are newly discovered members . of the remaining groups , the most counter - rotating one ( group 4 ) partially overlaps with the region of retrograde outliers " found by @xcite , while a dozen stars belonging to group 4 and group 5 fall in the region of the mildly retrograde stream detected by @xcite .
comparison to our high resolution n - body simulations confirms that the two groups associated with the helmi stream are likely fossil remnants of a dwarf galaxy which co - rotates with the disk of the galaxy and moves on a high - inclination orbit . as for the three retrograde groups ( 3 , 4 and 5 in table [ table:2 ] )
, they may be debris of an unique progenitor accreted along an initial retrograde orbit having an intermediate inclination in the range @xmath4 .
however , we can not exclude that these groups belong to different impacting satellites on retrograde orbits with inclinations within that same range . a more detailed analysis of the chemical abundances of the three detected groups , as well as more quantitative comparisons to extended simulations , are necessary to resolve this issue . in any event ,
the fastest objects appear with positive and negative @xmath120 values ( i.e. prograde and retrograde motions , respectively ) in the angular momentum @xmath565 regions for high - inclination orbits ( @xmath603 ) . on the other hand , for low - inclination both observations and simulations
show that the fastest objects appear only on _ retrograde _ orbits ( e.g. , fig .
[ prf : fig10 ] , top panel ) .
this asymmetric distribution is suggestive of the role played by dynamical friction during accretion . in anticipation of the much improved data expected over the coming years ,
in particular the gaia catalogue and the new ground - based spectroscopic surveys , we also investigated the impact of observational errors in our dynamical simulations .
the analysis indicates that ( see the relevant panels of figs .
[ prf : fig8 ] and [ prf : fig10 ] ) gaia will greatly influence these studies : for , velocity and angular momentum distribution will be almost completely dominated by the physics we are trying to recover , i.e. , the dynamical history of the merging events . at that point ,
full grids ( in , e.g. , inclination and amount of rotation ) of prograde and retrograde high resolution satellite simulations will be required to precisely characterise the debris detected . then , we will be able to number the merging events for direct comparison with the predictions of the ( @xmath604)cdm theory and its associated merging paradigm . in conclusion , the results shown might lead us to claim that the inner halo might have seen " only two events ; however the large uncertainties in the extant data , mostly observational , do not exclude the possibility that the events might be as many as four , and perhaps more given the intrinsic difficulty of our technique to deal with low inclination prograde mergers .
we are grateful to the referee for his / her comments that helped us improve the original manuscript .
this work has been partially funded by asi , under contract to inaf i/058/10/0 gaia mission - the italian participation to dpac " , and by miur , through prin 2012 grant no 1.05.01.97.02 chemical and dynamical evolution of our galaxy and of the galaxies of the local group " .
m.g.l . acknowledges support from the chinese academy of sciences through 2015 cas president s international fellowship initiative ( pifi ) for visiting scientists .
abadi , m. g. , navarro , j. f. , & steinmetz , m. 2006 , mnras , 365 , 747 abazajian , k. n. , adelman - mccarthy , j. k. , ageros , m. a. , et al .
2010 , apjs , 182 , 543 allende prieto , c. , sivarani , t. , beers , t. c. , et al .
2008 , aj , 136 , 2070 bullock , j. s. , & johnston k. v. 2005 , apj , 635 , 931 binney , j. , & merrifield , m. 1998 , galactic astronomy ( princeton univ .
press , princeton ) binney , j. , & tremaine , s. 1987 , galactic dynamics ( princeton univ . press , princeton ) bond , n. a. , ivezi , . ,
sesar , b. , et al .
2010 , apj , 716 , 1 chandrasekar , s. 1943 , apj , 97 , 255 chiba , m. , & beers , t. c. 2000 , aj , 119 , 2843 colpi , m. , mayer , l. , & governato , f. 1999 , apj , 525 , 720 dehnen , w. , & binney , j. j. 1998 , mnras , 298 , 387 de lucia , g. 2012 , astron .
nachr . , 333 , 460 dinescu , d. i. 2002 , in : f. van leeuwen , j. hughes , & g. piotto ( eds . ) , @xmath605 centauri : a unique window into astrophysics , ( san francisco : asp ) , vol .
265 , p. 365
duffau , s. , vivas , k. a. , zinn , r. , mendez , r. a. , & ruiz , m. t. 2014 , a&a , 566 , 118 eggen , o. j. 1977 , apj , 215 , 812 eggen , o. j. 1971 , pasp , 83 , 285 freeman , k. , & bland - hawthorn , j. 2002 , ara&a , 40 , 487 gilmore , g. , randich , s. , asplund , m. , et al .
2012 , the messenger , 147 , 25 girardi , l. , grebel , e. k. , odenkirken , m. , & chiosi , c. 2004 , a&a , 422 , 205 gomez , f. a. , helmi , a. , cooper , a. p. , et al .
2013 , mnras , 436 , 3602 gould , a. 2003 , apj , 592 , l63 harding , p. , morrison , h. l. , olszewski , e. w. , et al .
2001 , aj , 122 , 1397 hastie , t. , tibshirani , r. , & friedman , j. 2001 , the elements of statistical learning ( springer ) helmi , a. , cooper , a. p. , white , s. d. m. , et al .
2011 , apj , 733 , l7 helmi , a. 2008 , a&ar , 15 , 145 helmi , a. , & de zeeuw , p. t. 2000 , mnras , 319 , 657 helmi , a. , white , s. d. m. , de zeeuw , p. t. , & zhao , h. s. 1999 , nature , 402 , 53 helmi , a. , & white , s. d. m. 1999 , mnras , 307 , 495 hernquist , l. 1993 , apjs , 86 , 389 ibata , r. a. , irwin , m. j. , lewis , g. f. , ferguson , a. m. n. , & tanvir , n. 2003 , mnras , 340 , 21 ibata , r. a. , gilmore , g. , & irwin , m. j. 1994 , nature , 370 , 194 ivezi , . , sesar , b. , juri , m. , et al . 2008 , apj , 684 , 287 johnston , k. v. 1998 , apj , 495 , 297 kaufmann , l. , & rousseeuw , p. j. 1990 , finding groups in data ( john wiley & sons , inc . ) kepley , a. a. , morrison , h. l. , helmi , a. , et al .
2007 , aj , 134 , 1579 klement , r. j. 2010 , a&ar , 18 , 567 klement , r. , rix , h. -w . , flynn , c. , et al .
2009 , apj , 698 , 865 lasker , b. m. , lattanzi , m. g. , mclean , b. j. , et al .
2008 , aj , 139 , 735 lee , y. s. , beers , t. c. , sivarani , t. , et al .
2008a , aj , 136 , 2022 lee , y. s. , beers , t. c. , sivarani , t. , et al .
2008b , aj , 136 , 2050 majewski , s. r. , nidever , d. l. , smith , v. v. , et al .
2012 , apj , 747 , l37 majewski , s. r. , munn , j. e. , & hawley , s. l. 1996 , apj , 459 , 73 meza , a. , navarro , j. f. , abadi , m. g. , & steinmetz , m. 2005 , mnras , 359 , 93 moore , b. , diemand , j. , madau , p. , zemp , m. , & stadel , j. 2006 , mnras , 368 , 563 morrison , h. l. , helmi , a. , sun , j. , et al .
2009 , apj , 694 , 130 munn , j. a. , monet , d. g. , levine , s. e. , et al .
2008 , aj , 136 , 895 murante , g. , poglio , e. , curir , a. , & villalobos , a. 2010 , apj , 716 , l115 navarrete , c. , chaname , j. , meza , a. , et al .
2015 , to appear in apj navarro , j. f. , frenk , c. s. , & white , s. d. m. 1997 , apj , 490 , 493 pasetto , s. , grebel , e. k. , zwitter , t. , et al .
2012 , a&a , 547 , a70 perryman , m. a. c. , de boer , k. s. , gilmore , g. , et al .
2001 , a&a , 369 , 339 read , j. i. , lake , g. , agertz , o. , & debattista , v. p. 2008 , mnras , 389 , 1041 re fiorentin , p. , bailer - jones , c. a. l. , lee , y. s. , et al .
2007 , a&a , 467 , 1373 re fiorentin , p. , helmi , a. , lattanzi , m. g. , & spagna , a. 2005 , a&a , 439 , 551 sales , l. v. , navarro , j. f. , abadi , m. g. , & steinmetz , m. 2007 , mnras , 379 , 1464 schlaufman , k.c . ,
rockosi , c. m. , allende prieto , c. , beers , t. c. , bizyaev , d. , et al .
2009 , apj , 703 , 2177 schlaufman , k.c .
, rockosi , c. m. , lee , y.s .
, beers , t.c .
, & allende prieto , c. 2011 , apj , 734 , 49 schlafly , e. f. , green , g. , finkbeiner , d. p. , et al .
2014 , apj , 789 , 15 schlafly , e. f. , & finkbeiner , d. p. 2011 , apj , 737 , 103 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , apj , 500 , 525 searle , l. , & zinn , r. 1978 , apj , 225 , 357 smith , m. c. , evans , n. w. , belokurov , v. , et al .
2009 , mnras , 399 , 1223 spagna , a. , lattanzi , m. g. , re fiorentin , p. , & smart , r. l. 2010a , a&a , 510 , l4 spagna , a. , bucciarelli , b. , lattanzi , m. g. , re fiorentin , p. , & smart , r. l. 2010b , memsait , 14 , 67 spagna , a. , carollo , d. , lattanzi , m. g. , & bucciarelli , b. 2004 , a&a , 428 , 451 springel , v. 2005 , mnras , 364 , 1105 starkenburg , e. , helmi , a. , morrison , h. l. , et al .
2009 , apj , 698 , 567 trumpler r. j. , & weaver h. f. 1953 , statistical astronomy ( dover pubblications , inc . , new york )
turon , c. , oflaherty , k. s. , & perryman , m. a. c. eds .
2005 , the three - dimensional universe with gaia , esa spec .
576 villalobos , a. , & helmi , a. 2008 , mnras , 391 , 1806 yanny , b. , rockosi , c. , newberg , h. j. , et al .
2009 , aj , 137 , 4377 zhao , g. , zhao , y .- h . ,
chu , y .- q .
2012 , raa , 12 , 723 | we explore the vicinity of the milky way through the use of spectro - photometric data from the sloan digital sky survey and high - quality proper motions derived from multi - epoch positions extracted from the guide star catalogue ii database . in order to identify and characterise streams as relics of the milky way formation ,
we start with classifying , select , and study @xmath0 subdwarfs with @xmath1 < -1.5}$ ] up to @xmath2 kpc away from the sun as tracers of the local halo system .
then , through phase - space analysis , we find statistical evidence of five discrete kinematic overdensities among @xmath3 of the fastest - moving stars , and compare them to high - resolution n - body simulations of the interaction between a milky - way like galaxy and orbiting dwarf galaxies with four representative cases of merging histories .
+ the observed overdensities can be interpreted as fossil substructures consisting of streamers torn from their progenitors ; such progenitors appear to be satellites on prograde and retrograde orbits on different inclinations .
in particular , of the five detected overdensities , two appear to be associated , yelding twenty - one additional main - sequence members , with the stream of helmi et al .
( 1999 ) that our analysis confirms on a high inclination prograde orbit .
the three newly identified kinematic groups could be associated with the retrograde streams detected by dinescu ( 2002 ) and kepley et al .
( 2007 ) ; whatever their origin , the progenitor(s ) would be on retrograde orbit(s ) and inclination(s ) within the range @xmath4 .
+ finally , we use our simulations to investigate the impact of observational errors and compare the current picture to the promising prospect of highly improved data expected from the gaia mission . |
impressive progress is being achieved in deriving properties and interactions of hadrons using lattice qcd . in several instances ,
lattice methods are making predictions of hadronic quantities at the several - percent level @xcite . despite remarkable technical advances ,
current computational limitations continue to necessitate the use of quark masses , @xmath1 , that are significantly larger than the physical values , lattice spacings , @xmath2 , that are not significantly smaller than the physical scales of interest , and lattice sizes , @xmath0 , that are not significantly larger than the pion compton wavelength @xcite . therefore , lattice qcd simulations of hadronic physics require extrapolations in the quark masses , lattice spacing and lattice size , and ultimately it is confidence in these extrapolations that will allow a confrontation between lattice qcd and experiment . fortunately ,
in many cases , the dependence of hadronic physics on these parameters can be calculated analytically in the low - energy effective field theory ( eft ) .
calculability requires maintaining the hierarchy of mass scales , @xmath3 where @xmath4 is a typical momentum in the system of interest , @xmath5 is the pion mass and @xmath6 is the scale of chiral symmetry breaking ( @xmath7 is the pion decay constant ) . in a spatial box of size @xmath0 , momenta
are quantized such that @xmath8 with @xmath9 .
the hierarchy of eq .
( [ eq : hierarchy ] ) then requires maintenance of the additional inequality @xmath10 .
this bound ensures that ( non - pionic ) hadronic physics lives inside the box .
in addition , the bound @xmath11 ensures that the box size has no effect on spontaneous chiral symmetry breaking @xcite .
these two bounds , taken together , then imply that we must have @xmath12 .
when @xmath13 , and therefore @xmath14 , momentum zero - modes must be treated nonperturbatively @xcite and one is in the so - called @xmath15-regime . here
we will consider the range of pion masses , @xmath16 , and therefore we will take @xmath17 , keeping in mind that the eft may be reaching the limits of its validity when this bound on @xmath0 is saturated , particularly when the pions are light . for the observables considered here , finite - volume effects tend to be small for @xmath18 .
it is therefore of interest to have control over the finite - size dependence of hadronic observables in the range @xmath19 .
chiral perturbation theory ( @xmath20pt ) , which provides a systematic description of low - energy qcd near the chiral limit , is the appropriate eft to exploit the hierarchy of eq .
( [ eq : hierarchy ] ) and to describe the dependence of hadronic observables on @xmath0 @xcite .
recent work has investigated the finite - volume dependence in the meson @xcite sector and in the baryon @xcite sector . in this paper
we compute the leading finite - volume dependence of the nucleon masses and magnetic moments in baryon @xmath20pt , including the @xmath21 as an explicit degree of freedom .
the finite - size dependence of the nucleon mass was first studied in ref .
@xcite , and has recently been computed to @xmath22 in baryon @xmath20pt ( without including the @xmath21 as an explicit degree of freedom ) in ref .
( some discussion of the effects of the @xmath21 on the finite - size dependence of the nucleon mass appears in ref .
@xcite . )
we also give expressions for the finite - size dependence of the nucleon masses and magnetic moments in partially - quenched qcd ( pqqcd ) , including strong isospin breaking .
the cost of simulating dynamical quarks with light masses suggests separately varying the sea and valence quark masses in the lattice qcd partition function , a procedure known as partial - quenching .
this procedure has important advantages beyond issues of cost ; by increasing the dimensionality of the parameter space that is explored , lattice qcd simulations can provide additional `` data '' , which can significantly improve the quality of extrapolations .
@xmath20pt has been extended to describe both quenched qcd ( qqcd ) @xcite with quenched chiral perturbation theory ( q@xmath20pt ) and pqqcd @xcite with partially - quenched chiral perturbation theory ( pq@xmath20pt ) .
recently , meson and baryon properties have been studied extensively in both q@xmath20pt
@xcite and pq@xmath20pt @xcite . the effective field theory ( eft ) describing the low - energy dynamics of two - nucleon systems and nucleon - hyperon systems in pqqcd
has also been explored @xcite .
this paper is organized as follows . in section [ sec : masses ] , the leading finite - size corrections to the nucleon masses are computed .
the same is done for the nucleon magnetic moments in section [ sec : magmoms ] .
we conclude in section [ sec : conc ] .
mathematical details and the partially - quenched extensions of the qcd results ( including strong isospin breaking ) are left to appendices .
for purposes of setting notation , we will begin by reviewing the derivation of leading terms in the chiral expansion of the nucleon mass .
the relevant leading baryon mass operators in two - flavor @xmath20pt are @xmath23 - i \overline { t}^\mu v\cdot { \cal d } t_\mu \ + \ \delta\ \overline{t}^\mu t_\mu \ \ , \label{eq : freeqcd}\end{aligned}\ ] ] where the chirally - invariant mass operator is @xmath24 , with @xmath25 , and @xmath26 is the usual two - flavor goldstone matrix .
the relevant leading axial operators are @xmath27 \ .
\label{eq : intsqcd}\end{aligned}\ ] ] the mass of the @xmath28-th nucleon has a chiral expansion of the form @xmath29 where a term @xmath30 denotes a contribution of order @xmath31 , and @xmath32 . the nucleon mass is dominated by a term in the @xmath20pt lagrangian , @xmath33 , that is independent of @xmath1 . here
@xmath21 , the @xmath21-nucleon mass splitting , is assumed to be of the same chiral order as the pion mass @xcite .
each of the contributions in eq .
( [ eq : massexp ] ) depends upon the scale chosen to renormalize the theory . while to @xmath34 , the objects @xmath33 and @xmath35 are scale independent , at one - loop level , @xmath36 , they are scale dependent .
the leading dependence upon @xmath1 , occurring at @xmath34 , is due to the operators in eq .
( [ eq : freeqcd ] ) with coefficients @xmath37 and @xmath38 .
the leading non - analytic dependence upon @xmath1 arises from the one - loop diagrams shown in fig .
[ fig : masses ] .
= 4.0 in 0.15 in .2 in in isospin - symmetric qcd , with @xmath39 , one finds the nucleon mass at one - loop order in the chiral expansion @xcite , @xmath40 \ \ \ , \label{eq : massesqcdisolimit}\end{aligned}\ ] ] where @xmath41 and @xmath21 is the @xmath21-nucleon mass splitting .
here we have used dimensional regularization ( dim reg ) with @xmath42 to define the divergent loop integrals and @xmath43 and @xmath44 . in the infinite - volume limit
, the nucleon mass may be written as @xmath45 where @xmath46 here @xmath47 denotes a choice of ultraviolet regulator and a renormalization scheme , @xmath48 where @xmath49 , which recovers the results of eq .
( [ eq : massesqcdisolimit ] ) . ] . in a spatial box of size @xmath0
, @xmath50 generalizes to @xmath51 where we have rotated the integral to euclidean space and accounted for the quantization of the momentum levels due to the periodic boundary conditions .
feynman parameterizing and explicitly evaluating the @xmath52 integration lead to @xmath53^{3/2}}}\ \right)\ , \end{aligned}\ ] ] where @xmath54 .
we can now write the finite - size corrections to @xmath50 as @xmath55^{1/2 } } } \right ) - \beta_\delta^2 \delta_l \left ( { { 1}\over{[{\vec k}^2+\beta_\delta^2]^{3/2 } } }
\right ) \right ] \ , \label{eq : deltaisublgeneral}\end{aligned}\ ] ] where @xmath56 is defined in eq .
( [ eq : app1 ] ) of the appendix .
notice that @xmath57 is a purely infrared quantity and , as such , is independent of @xmath47 , which appears implicitly as the ultraviolet cutoff @xmath58 in all sums and integrals . ] .
using eqs .
( [ eq : massgeneral ] ) and ( [ eq : deltaisublgeneral ] ) , the finite - size corrections to the nucleon mass may then be expressed as @xmath59 using the `` master '' formula , eq .
( [ eq : masterformula ] ) , derived in the appendix , we find @xmath60 where @xmath61 \ .
\label{eq : defined } \end{aligned}\ ] ] 0.4 in = 3.2 in 0.2cm=3.2 in here @xmath62 is a modified bessel function of the second kind .
with @xmath63 the integral over @xmath64 can be carried out explicitly ( see appendix ) and one has @xmath65 notice that this function contains no power - law corrections . finally , we have @xmath66 this is the exact formula for the finite - size corrections to the nucleon mass at leading order in baryon @xmath20pt . in fig . [
fig : asymratio ] ( right panel ) , the ratio of the nucleon mass size dependence to the ( infinite volume ) nucleon mass has been plotted against the box size @xmath0 for various pion masses .
the solid and dashed lines correspond to the qcd formula of eq .
( [ eq : exactmassformula2compact ] ) with @xmath67 and @xmath68 , respectively .
the dotted line corresponds to pqqcd in the isospin limit taken from eq .
( [ eq : nucleonmasspqisolimit ] ) with @xmath67 and @xmath69 .
we use the parameter set : @xmath7 , @xmath70 and @xmath71 . using eqs .
( [ eq : delzeronew ] ) and ( [ eq : kdelasym ] ) , in the large-@xmath0 expansion one has @xmath72 where @xmath73 . in the @xmath74 limit ,
the leading term in the large-@xmath0 expansion is in agreement with ref .
@xcite and in disagreement with ref .
@xcite . fig .
[ fig : asymratio ] ( left panel ) plots the ratio @xmath75 as a function of @xmath0 for various pion masses . clearly the utility of eq .
( [ eq : lowenergytheoremasym ] ) is purely aesthetic ; even with heavy pions , the asymptotic formula , eq .
( [ eq : lowenergytheoremasym ] ) , is not accurate for @xmath76 .
this points to the importance of exponential corrections ; for @xmath77 convergence of the momentum sums requires keeping terms with @xmath78 , _
i.e. _ one must include corrections of @xmath79 .
with the finite - size corrections for the masses , it is straightforward to get the magnetic moments .
the leading operators contributing to the nucleon magnetic moments are @xmath80 where @xmath81 is the electromagnetic field - strength tensor , @xmath82 is the physical value of the nucleon mass , @xmath83 is the isoscalar nucleon magnetic moment , @xmath84 is the isovector nucleon magnetic moment and @xmath85
. in isospin - symmetric qcd one finds the nucleon magnetic moment matrix at one - loop order in the = 4.0 in 0.15 in .2 in chiral expansion @xcite @xmath86\ { \hat\tau}_3 \ .
\label{eq : magmomsqcd}\end{aligned}\ ] ] the scale dependence is left implicit .
the proton and the neutron magnetic moments are the diagonal elements of @xmath87 .
the first term within the brackets is from fig .
[ fig : masses](a ) while the second term is from fig .
[ fig : masses](b ) .
the function @xmath88 is @xmath89 here again we have used dim reg with @xmath42 . in the limit @xmath90 , @xmath91 . in the infinite - volume limit
, the nucleon magnetic moments may be written as @xmath92\ { \hat\tau}_3 \ , \label{eq : magmomsqcdwithjs}\end{aligned}\ ] ] where @xmath93 therefore , the finite - size corrections to @xmath87 are @xmath94\ { \hat\tau}_3 \ . \label{eq : magmomsfinitesize}\end{aligned}\ ] ] using eqs .
( [ eq : delilfinalform ] ) , ( [ eq : defined ] ) and ( [ eq : jdefined ] ) , and the properties of modified bessel functions , it is straightforward to find @xmath95 where @xmath96 \ .
\label{eq : ydef}\end{aligned}\ ] ] with @xmath63 one has @xmath97 finally one arrives at @xmath98\ { \hat\tau}_3 \ .
\label{eq : magmomsfinitesizecompact}\end{aligned}\ ] ] this is the exact formula for the finite - size corrections to the nucleon magnetic moments at leading order in baryon @xmath20pt . in fig .
[ fig : asymratio2 ] ( right panel ) , the ratio of the proton magnetic moment size dependence to the ( infinite volume ) magnetic moment has been plotted against the box size @xmath0 .
the solid and dashed lines correspond to the qcd formula of eq .
( [ eq : magmomsfinitesizecompact ] ) with @xmath67 and @xmath68 , respectively
. the dotted line corresponds to pqqcd in the isospin limit taken from eq .
( [ eq : nucleonmagpqisolimit ] ) with @xmath67 and @xmath69 . 0.4 in = 3.2 in 0.2cm=3.2 in again using eqs .
( [ eq : delilfinalform ] ) , ( [ eq : defined ] ) and ( [ eq : jdefined ] ) , together with eqs .
( [ eq : delzeronewj ] ) and ( [ eq : kdelasym ] ) , in the large-@xmath0 expansion one has @xmath99 where the dots denote contributions of @xmath100 .
[ fig : asymratio2 ] ( left panel ) plots the ratio @xmath101 as a function of @xmath0 for various pion masses .
the curves are similar to those of the nucleon mass in fig .
[ fig : asymratio ] , as is the conclusion about the practical utility of eq .
( [ eq : magmomsfinitesizewithdeltaasymptotic ] ) .
it is hoped that in the near future lattice ( pq)qcd simulations of baryon properties will encounter the chiral regime , where the quark masses are sufficiently small to allow a meaningful chiral expansion in quark masses , box size and lattice spacing .
it is likely that this regime has been encountered in recent work on heavy - meson systems @xcite . the results of this paper , together with the results of refs .
@xcite and @xcite , give the dependence of the nucleon masses and magnetic moments on the sea and valence quark masses _ and _ on the lattice spacing , @xmath2 , and size , @xmath0 , to leading order in the chiral expansion .
we eagerly await lattice ( pq)qcd simulations within the chiral regime where this parameter space may be fruitfully explored .
i would like to thank david lin and martin savage for helpful discussions .
this work was partly supported by doe contract de - ac05 - 84er40150 , under which the southeastern universities research association ( sura ) operates the thomas jefferson national accelerator facility .
we wish to evaluate @xmath102^{m}}}\ \right ) \ & \equiv & \ { 1\over l^3 } \sum_{\vec l}\ { { 1}\over{({\vec l}^2+{\cal m}^2)^m}}- \int \frac{d^3l}{(2\pi ) ^3}\ { { 1}\over{({\vec l}^2+{\cal m}^2)^m}}\ \ .
\label{eq : app1}\end{aligned}\ ] ]
as this difference is ultraviolet finite , we omit the label denoting the scheme dependence of the individual sum and integral .
this expression has been evaluated in many places . using the identity @xmath103 one finds @xmath104^{m}}}\ \right ) = \frac{1}{(4\pi)^{3/2}\gamma(m ) } \int_0^\infty d\eta \eta^{m-5/2 } e^{-\eta { { \cal m}^2 } } \left [ \frac{(4\pi\eta)^{3/2}}{l^3 } \sum_{\vec l } e^{-\eta{\vec l}^{\;2}}- 1 \right ] .
\label{eq : app3}\end{aligned}\ ] ] expressing the momentum as @xmath105 and using the jacobi identity @xcite , @xmath106 leads to @xmath102^{m}}}\ \right ) \ & = & \frac{1}{(4\pi)^{3/2}\gamma(m ) } \sum_{{\vec n}\neq 0}\int_0^\infty d\eta \eta^{m-5/2 } e^{-\eta { { \cal m}^2 } } e^{-{l^2{\vec n}^2}/{4\eta } } \ .
\label{eq : app6}\end{aligned}\ ] ] performing the integral over @xmath107 then gives the `` master '' formula @xmath102^{m}}}\ \right ) \ & = & \
\frac{2^{-\frac{1}{2}-m}\,{\cal m}^{3 - 2\,m}}{{\pi } ^{3/2}\ , \gamma(m ) } \sum_{{\vec n}\neq 0}\ { \left ( l\,{\cal m}\,|{\vec n}| \right ) } ^{-\frac{3}{2}+m}\ , k_{\frac{3}{2 } - m}(l\,{\cal m}\,|{\vec n}| ) \ , \label{eq : masterformula}\end{aligned}\ ] ] where @xmath62 is a modified bessel function of the second kind .
a well - chosen change of integration variable and the properties of modified bessel functions allow one to write eq .
( [ eq : defined ] ) as @xmath108 we find no useful simplification of this formula in the general case . with @xmath63 one directly finds @xmath109 which gives eq .
( [ eq : delzeronew ] ) .
in the large-@xmath0 limit , using the expansion of the modified bessel function for large argument , one finds from eq .
( [ eq : usefulformula ] ) , @xmath110 where the dots denote contributions of @xmath111 .
observe that one can expand the integrand in powers of @xmath112 , @xmath113 where the dots denote contributions of @xmath114 .
( similar technology has been developed in ref .
@xcite in the context of the heavy - meson systems . ) plugging this back into eq .
( [ eq : usefulformula2 ] ) one finds , in the asymptotic limit , @xmath115 where the dots denote contributions of @xmath111 .
we work in pqqcd including isospin breaking , but with electromagnetism turned off .
the lagrangian describing the interactions of @xmath116 ( containing the nucleon ) and @xmath117 ( containing @xmath21 ) , which transform in the @xmath118 and @xmath119 of @xmath120 , respectively , with the pseudo - goldstone bosons at leading - order in the chiral expansion is @xcite @xmath121 \ .
\label{eq : pqints}\end{aligned}\ ] ] here the axial - vector field @xmath122 is a six - by - six matrix . matching to the qcd effective lagrangian of eq .
( [ eq : intsqcd ] ) and to the additional operator @xmath123 \ , \label{eq : intsqcdextra}\end{aligned}\ ] ] one finds that at tree - level , @xmath124 the finite - size corrections to the proton mass are given by @xmath125 \right.\nonumber\\ & & \left . \quad \ + \
{ g_1 ^ 2\over { 6\pi}}\left[\ { \cal
k}(m_{uu},0 ) - 5{\cal k}(m_{ud},0 ) + 3{\cal k}(m_{jd},0 ) + 2{\cal k}(m_{ju},0 ) + 3{\cal k}(m_{ld},0 ) + 2{\cal k}(m_{lu},0 ) \right.\right . \nonumber\\ & & \left.\left .
\quad + 3 { \cal g}_{\eta_u , \eta_u}(0 ) + 6 { \cal g}_{\eta_u , \eta_d}(0 ) + 3 { \cal g}_{\eta_d , \eta_d}(0 ) \ \right ] \right.\nonumber\\ & & \left . \quad \ + \ { 2g_a g_1\over{3\pi}}\left[\ { \cal k}(m_{ju},0 ) + { \cal k}(m_{lu},0 ) - { \cal k}(m_{ud},0 ) + 2{\cal k}(m_{uu},0 ) + 3 { \cal g}_{\eta_u , \eta_d}(0 ) + 3 { \cal g}_{\eta_u , \eta_u}(0 ) \ \right ] \right.\nonumber\\ & & \left .
+ { 2g_{\delta n}^2\over 9\pi}\left[\ 5{\cal k}(m_{ud},\delta ) + { \cal k}(m_{uu},\delta ) + { \cal k}(m_{ju},\delta ) + { \cal k}(m_{lu},\delta ) + 2{\cal k}(m_{jd},\delta ) + 2{\cal k}(m_{ld},\delta ) \right.\right .
\nonumber\\ & & \left.\left .
\ + 2 { \cal g}_{\eta_d , \eta_d}(\delta ) + 2 { \cal g}_{\eta_u , \eta_u}(\delta ) - 4 { \cal g}_{\eta_u , \eta_d}(\delta ) \ \right]\ \right ) \ \ \ , \label{eq : pmass}\end{aligned}\ ] ] where @xmath126 , @xmath127 is given by @xmath128 \
\ \ , \label{eq : hpsdef}\end{aligned}\ ] ] the mass , @xmath129 , is given by @xmath130 , and @xmath131 is defined in eq .
( [ eq : defined ] ) ( where now the @xmath5 dependence is made explicit ) .
note that @xmath132 refers to the goldstone - boson mass with quark content @xmath2 and @xmath133 ( hence @xmath134 , _ etc .
_ ) ; @xmath135 and @xmath136 label the sea quark masses .
the finite - size corrections to the neutron mass are given by @xmath137 \right.\nonumber\\ & & \left . \quad \ + \
{ g_1 ^ 2\over { 6\pi}}\left[\ { \cal
k}(m_{dd},0 ) - 5{\cal k}(m_{ud},0 ) + 2{\cal k}(m_{jd},0 ) + 3{\cal k}(m_{ju},0 ) + 2{\cal k}(m_{ld},0 ) + 3{\cal k}(m_{lu},0 ) \right.\right . \nonumber\\ & & \left.\left .
\quad + 3 { \cal g}_{\eta_u , \eta_u}(0 ) + 6 { \cal g}_{\eta_u , \eta_d}(0 ) + 3 { \cal g}_{\eta_d , \eta_d}(0 ) \ \right ] \right.\nonumber\\ & & \left . \quad \ + \ { 2g_a g_1\over{3\pi}}\left[\ 2{\cal k}(m_{dd},0 ) + { \cal k}(m_{jd},0 ) + { \cal k}(m_{ld},0 ) - { \cal k}(m_{ud},0 ) + 3 { \cal g}_{\eta_d , \eta_d}(0 ) + 3 { \cal g}_{\eta_u , \eta_d}(0 ) \ \right ] \right.\nonumber\\ & & \left . + { 2g_{\delta n}^2\over 9\pi}\left[\ 5{\cal k}(m_{ud},\delta ) + { \cal k}(m_{dd},\delta ) + { \cal k}(m_{jd},\delta ) + { \cal k}(m_{ld},\delta ) + 2{\cal k}(m_{ju},\delta ) + 2{\cal k}(m_{lu},\delta ) \right.\right . \nonumber\\
& & \left.\left .
\quad + 2 { \cal g}_{\eta_d , \eta_d}(\delta ) + 2 { \cal g}_{\eta_u , \eta_u}(\delta ) - 4 { \cal g}_{\eta_u , \eta_d}(\delta ) \ \right]\ \right ) \ \ \ .
\label{eq : nmass}\end{aligned}\ ] ] in the isospin limit , one has @xmath138\ -\
{ g_{\delta n}^2\over 6\pi^2f^2}\left[\ { \cal k}(m_\pi,\delta ) + { \cal k}(m^s_\pi,\delta)\ \right ] \nonumber\\ & & \qquad\qquad + \ { g_1\over { 24 \pi^2 f^2}}\ \left ( 5g_1 + 4g_a \right)\ \left[\ { \cal k}(m_\pi,0)-{\cal k}(m^s_\pi,0)\ \right ] \ \ \ , \label{eq : nucleonmasspqisolimit}\end{aligned}\ ] ] where we have used the fact that @xmath139 in the isospin limit . here
@xmath140 denotes the mass of a pion made of one valence quark and one sea quark .
these expressions further collapse down to isospin - symmetric qcd in the limit @xmath141 .
the finite - size corrections to the proton magnetic moment in pqqcd are @xmath143 \right . \nonumber\\ & & \left . \ + \ { 2 g_1 g_a\over 9}\left[\ \;{\cal y}(m_{ud},0 ) + \;{\cal y}(m_{uu},0 ) - \;{\cal y}(m_{ju},0 ) - \;{\cal y}(m_{lu},0)\ \right ] \right .
\nonumber\\ & & \left .
\ + \ { g_1 ^ 2\over 36}\left[\ \;{\cal
y}(m_{ud},0 ) + 4 \;{\cal y}(m_{uu},0 ) - 3 \;{\cal y}(m_{dd},0 ) + 3 \;{\cal y}(m_{jd},0 ) \right.\right . \nonumber\\ & & \left.\left .
\qquad\qquad - 4 \;{\cal y}(m_{ju},0 ) + 3 \;{\cal y}(m_{ld},0 ) - 4 \;{\cal y}(m_{lu},0)\ \right ] \right .
\nonumber\\ & & \left .
+ q_j\left({2 g_a^2\over 3}\left[\;{\cal y}(m_{ju},0)-\;{\cal y}(m_{uu},0)\right ] + { g_1 g_a\over 3}\left[\;{\cal y}(m_{ju},0)-\;{\cal y}(m_{uu},0)\right ] \right.\right .
\nonumber\\ & & \left.\left .
\qquad\qquad + { g_1 ^ 2\over 6}\left[\;{\cal y}(m_{ju},0 ) - \;{\cal y}(m_{uu},0 ) + { 3\over 2 } \;{\cal y}(m_{jd},0 ) - { 3\over 2 } \;{\cal y}(m_{ud},0)\right ] \right ) \right .
\nonumber\\ & & \left
. + q_l\left({2 g_a^2\over 3}\left[\;{\cal y}(m_{lu},0)-\;{\cal y}(m_{ud},0)\right ] + { g_1 g_a\over 3}\left[\;{\cal y}(m_{lu},0)-\;{\cal y}(m_{ud},0)\right ] \right.\right .
\nonumber\\ & & \left.\left .
\qquad\qquad + { g_1 ^ 2\over 6}\left[\;{\cal y}(m_{lu},0 ) - \;{\cal y}(m_{ud},0 ) + { 3\over 2 } \;{\cal y}(m_{ld},0 ) - { 3\over 2 } \;{\cal y}(m_{dd},0)\right ] \right ) \right .
\nonumber\\ & & \left
. + \ { g_{\delta n}^2\over 27}\ \left[\ \;{\cal
y}(m_{dd},\delta ) - \;{\cal y}(m_{uu},\delta ) - 6 \;{\cal y}(m_{ud},\delta ) - \;{\cal y}(m_{jd},\delta ) \right.\right .
\nonumber\\ & & \left.\left.\qquad\qquad\qquad + \;{\cal y}(m_{ju},\delta ) - \;{\cal y}(m_{ld},\delta ) + \;{\cal y}(m_{lu},\delta ) \right .
\right.\nonumber\\ & & \left .
\left.\qquad + { 3\over 2}\ q_j\ \left ( \;{\cal y}(m_{uu},\delta ) + 2 \;{\cal y}(m_{ud},\delta ) - \;{\cal y}(m_{ju},\delta ) - 2 \;{\cal y}(m_{jd},\delta)\right ) \right.\right .
\nonumber\\ & & \left.\left.\qquad + { 3\over 2 } \ q_l\ \left ( \;{\cal y}(m_{ud},\delta ) + 2 \;{\cal y}(m_{dd},\delta ) - \;{\cal y}(m_{lu},\delta ) - 2 \;{\cal y}(m_{ld},\delta)\right ) \ \right ] \
\right ) , \label{eq : pmag}\end{aligned}\ ] ] where @xmath144 is defined in eq . ( [ eq : ydef ] ) ( where now the @xmath5 dependence is made explicit ) .
the finite - size corrections to the neutron magnetic moment are @xmath145 \right . \nonumber\\ & & \left . \ + \ { g_1 g_a\over 9}\left[\ \;{\cal y}(m_{jd},0 ) + \;{\cal y}(m_{ld},0 ) - \;{\cal y}(m_{ud},0 ) - \;{\cal y}(m_{dd},0)\ \right ] \right .
\nonumber\\ & & \left .
\ + \ { g_1 ^ 2\over 18}\left[\ 3\;{\cal
y}(m_{uu},0 ) + 2 \;{\cal y}(m_{ud},0 ) - \;{\cal y}(m_{dd},0 ) + \;{\cal y}(m_{jd},0 ) \right.\right . \nonumber\\ & & \left.\left .
\qquad\qquad - 3 \;{\cal y}(m_{ju},0 ) + \;{\cal y}(m_{ld},0 ) - 3 \;{\cal y}(m_{lu},0)\ \right ] \right .
\nonumber\\ & & \left .
+ q_j\left({2 g_a^2\over 3}\left[\;{\cal y}(m_{jd},0)-\;{\cal y}(m_{ud},0)\right ] + { g_1 g_a\over 3}\left[\;{\cal y}(m_{jd},0)-\;{\cal y}(m_{ud},0)\right ] \right.\right .
\nonumber\\ & & \left.\left .
\qquad\qquad + { g_1 ^ 2\over 6}\left[\;{\cal y}(m_{jd},0 ) - \;{\cal y}(m_{ud},0 ) + { 3\over 2 } \;{\cal y}(m_{ju},0 ) - { 3\over 2 } \;{\cal y}(m_{uu},0)\right ] \right ) \right .
\nonumber\\ & & \left .
+ q_l\left({2 g_a^2\over 3}\left[\;{\cal y}(m_{ld},0)-\;{\cal y}(m_{dd},0)\right ] + { g_1 g_a\over 3}\left[\;{\cal y}(m_{ld},0)-\;{\cal y}(m_{dd},0)\right ] \right.\right .
\nonumber\\ & & \left.\left .
\qquad\qquad + { g_1 ^ 2\over 6}\left[\;{\cal y}(m_{ld},0 ) - \;{\cal y}(m_{dd},0 ) + { 3\over 2 } \;{\cal y}(m_{lu},0 ) - { 3\over 2 } \;{\cal y}(m_{ud},0)\right ] \right ) \right .
\nonumber\\ & & \left
. + \ { g_{\delta n}^2\over 54}\ \left[\ \;{\cal
y}(m_{dd},\delta ) - 4\;{\cal y}(m_{uu},\delta ) + 9 \;{\cal y}(m_{ud},\delta ) - \;{\cal y}(m_{jd},\delta ) \right.\right .
\nonumber\\ & & \left.\left.\qquad\qquad\qquad + 4\;{\cal y}(m_{ju},\delta ) - \;{\cal y}(m_{ld},\delta ) + 4\;{\cal y}(m_{lu},\delta ) \right .
\right.\nonumber\\ & & \left .
\left.\quad + { 3}\ q_j\ \left ( \;{\cal y}(m_{ud},\delta ) + 2 \;{\cal y}(m_{uu},\delta ) - \;{\cal y}(m_{jd},\delta ) - 2 \;{\cal y}(m_{ju},\delta)\right ) \right.\right . \nonumber\\ & & \left.\left.\quad + { 3 } \
q_l\ \left ( \;{\cal y}(m_{dd},\delta ) + 2 \;{\cal y}(m_{ud},\delta ) - \;{\cal y}(m_{ld},\delta ) - 2 \;{\cal y}(m_{lu},\delta)\right ) \ \right ] \
\label{eq : nmag}\end{aligned}\ ] ] in the isospin limit ( with @xmath146 ) , one has @xmath147\ + \ {
2g_{\delta n}^2\over 9}\ { \cal y}(m_\pi,\delta ) \right.\nonumber\\ & & \left . \qquad\qquad -\ { g_1\over { 18}}\ \left ( g_1 + 8g_a \right)\ \left[\ { \cal y}(m_\pi,0)-{\cal y}(m^s_\pi,0)\ \right ] \right)\ { \hat\tau}_3 \ .
\label{eq : nucleonmagpqisolimit}\end{aligned}\ ] ] these expressions further collapse down to isospin - symmetric qcd in the limit @xmath141 . | we compute finite - size corrections to nucleon masses and magnetic moments in a periodic , spatial box of size @xmath0 , both in qcd and in partially - quenched qcd
. 0.5 cm 0.5 cm |
in model theory , and in s. shelah s classification theory in particular , one central program is the search for robust dividing lines among complete theories dividing lines between intelligibility and non - structure . for a dividing line to be sufficiently interesting
, one often desires that both sides of the line have interesting mathematical content .
moreover , if such a dividing line has multiple characterizations coming from seemingly different contexts , this lends credence to the notion that the line is indeed important .
the exemplar of this in model theory is the notion of stability , first introduced by s. shelah @xcite .
both stable and unstable theories are inherently interesting , and stability enjoys many different characterizations , from cardinalities of stone spaces , to coding orders , to collapse of indiscernible sequences to indiscernible sets .
another example of such a robust dividing line is nip ; like stability , there are important structures with theories on both sides of the line , and the model theory on both sides can be very rich . in this paper , we focus on the notion of collapsing generalized indiscernibles as a means of characterizating / defining various model - theoretic dividing lines that are already well - established in the literature .
it can be said that this work started with shelah s original characterization of stable theories , theorem ii.2.13 of @xcite .
there he shows that a theory is stable if and only if every indiscernible sequence is an indiscernible set ( i.e. , the order of the sequence `` does not matter '' or is invisible to models of the theory ) . in a similar fashion , it can be shown that other well - known model - theoretic dividing lines can be characterized similarly by such `` collapse '' statements , which we formalize here . in order to carry this out
, we must expand our definition of `` indiscernibility . '' a quintessential example of this phenomenon beyond stability is the third author s characterization of nip theories by ordered - graph indiscernibles . in @xcite
, she shows that a theory is nip if and only if any indiscernible `` picture '' of the generic ordered graph is actually indiscernible _ without _ the graph structure ( i.e. , it collapses to order in the sense the the graph relation is invisible to models of the theory ) . the project in this paper was also alluded to in the work of the first and second authors on op - dimension @xcite , and this will be discussed further in section [ sect_opdim ] below .
generalized indiscernibles were first introduced by shelah ( see section vii.2 of @xcite ) , where they were used in the context of tree - indexed indiscernibles in the hope of understanding the tree property ( see , for example , theorem iii.7.11 of @xcite ) , but propose a slightly different formulation suggested by the second author in presently - unpublished work
see @xcite which slightly simplifies some aspects of section [ sect_rosiness ] . for more on tree indiscernibles in particular ,
see @xcite .
* outline of the paper*. in section [ sect_indisc ] , we introduce all the relevant notation and notions for generalized indiscernibility , the ramsey property , and the modeling property .
as noted above , this will mostly follow @xcite . in the remaining sections
, we will apply this to specific cases , exhibiting the `` collapse '' characterizations of various dividing lines . in section [ sect_opdim ] ,
we begin with the example of @xmath0-multi - order indiscernibles , first considered in @xcite , and show that collapse down to @xmath0 orders characterizes op - dimension @xmath0 ( theorem [ thm_opdimindisc ] below ) . in section [ sect_rosiness ] , we consider function - space indiscernibles , and show that collapse down to two indiscernible sequences characterizes rosiness ( theorem [ thm : collapse - result - rosiness ] below ) . in section [ sect_ntp2 ] , we consider convex equivalence relation indiscernibles , showing that ntp2 is equivalent to a dichotomy between collapsing down to an indiscernible sequence or having dividing witnessed by a particular formula ( theorem [ thm_ntp2 ] below ) .
in this section , we review the definitions associated with structural ramsey theory : -like theories , the ramsey property , and the modeling property , and theories of indiscernibles .
the development ( taking up subsections 2.0.1 through 2.2 ) is just a summary of that in @xcite , which in turn is based in large part on @xcite . in subsection 2.4 ,
we formalize the notion of `` collapse of indiscernibles '' as it will be used in this article . mirroring the notation used in @xcite
, we adopt a somewhat eccentric streamlined notation for finite structures . in order to save ourselves from writing `` ... is a finite structure '' and `` is a finite substructure of '' _ ad nauseum _ , we establish a slightly eccentric convention . with the one exception of @xmath1-embeddings ( see the next subsection ) , upper - case calligraphic letters @xmath2 , @xmath3 and so forth always denote _ infinite _ structures . on the other hand , lower - case gothic letters
@xmath4 and so forth always denote _ finite _ structures .
thus , @xmath5 , for example , means that @xmath6 is a finite ( induced ) substructure of @xmath2 .
moreover , given a 1-sorted structure @xmath6 , @xmath7 is its universe ( a finite set ) , and @xmath8 is the cardinality of @xmath6 .
we write @xmath9 , @xmath10 , @xmath11 for the sets of all embeddings @xmath12 , @xmath13 and @xmath14 , respectively . for an embedding @xmath15 in one of these sets , @xmath16 ( or @xmath17 )
is the substructure of the codomain induced on the image of @xmath15 .
although the machinery of theories of indiscernibles accomodates countably infinite relational languages ( even allowing finitely many constant symbols ) , in this article , the signature @xmath18 of every language of a -like theory @xmath19 will be finite relational . since our languages will never have function symbols ,
it is necessary to accommodate `` base theories '' that are not universal . to do this
, we define the notion of a fragment a surrogate for the quantifier - free formulas and the notion of a -like theory , which is analogous to the universal theory of a genuine class . a theory in a language with a function symbol that is known to be a theory of indiscernibles can be converted to a purely relational theory over a base theory where the base theory states that the new relation symbol behaves as the old function symbol did .
given a language @xmath20 , we define a _ fragment _ of @xmath20 to be a subset @xmath21 of @xmath20-formulas with the following closure properties : * @xmath21 contains every quantifier - free @xmath20-formula .
* @xmath21 is closed under taking boolean combinations and under changes of free variables .
* @xmath21 is closed under taking subformulas .
we write @xmath22 for the set of formulas of the form @xmath23 where @xmath24 . abusing the notation somewhat , we also write @xmath22 for the closure of this set under conjunctions and disjunctions .
we say that @xmath21 is the _ trivial fragment _ if it is the set of quantifier - free formulas of @xmath20 . given a language @xmath20 , a _
-like theory _ is a pair @xmath25 , where @xmath21 is a fragment of @xmath20 and @xmath26 is a set of @xmath22-sentences , satisfying the following conditions : * for all @xmath27 , there is a @xmath28 such that @xmath29 ( @xmath30 ) . * for @xmath31 and @xmath32 ( @xmath30 )
, there are @xmath33 and @xmath34 ( @xmath30 ) such that @xmath35 .
* there is a function @xmath36 such that for any @xmath37 and @xmath38 , there is an @xmath39 such that @xmath40 and @xmath41 .
as with classes and -limits , -like theories support countably infinite generic models .
the following proposition summarizes this fact .
[ prop_comp ] let @xmath25 be a -like theory in @xmath20 .
uniquely up to isomorphism , there is a countable @xmath20-structure @xmath2 such that : * for all @xmath42 , @xmath43 .
* for any @xmath44 with @xmath45 , for any @xmath46 , there is an @xmath47 such that @xmath48 .
moreover , @xmath49 is @xmath50-categorical and model - complete relative to @xmath21 .
[ defn_comp ] in the notation of proposition [ prop_comp ] , for a -like theory @xmath25 , @xmath51 is the _ @xmath21-model - completion of @xmath26_. the structure @xmath52 is called the _ generic model _ of @xmath25 . in the case that @xmath21 is the trivial fragment
, @xmath51 is the usual model - completion of @xmath26 .
here we give a list of all -like theories @xmath25 discussed in the paper .
[ defn_index ] * @xmath26 : stands for an arbitrary index theory * @xmath53 , @xmath54 : *
* let @xmath55 be the @xmath0-sorted language with sorts @xmath56 and for each @xmath57 , a binary relation symbol @xmath58 on @xmath59 .
let @xmath53 be the @xmath55-theory asserting that each @xmath58 is a linear ordering of @xmath59 .
* * let @xmath60 be the trivial fragment .
* @xmath61 : this is the theory of linear order ( see above ) * @xmath62 , @xmath63 : * * let @xmath64 be the @xmath65-sorted language with @xmath0 binary relations .
let @xmath66 be the @xmath67-theory which asserts that @xmath58 is a linear order for each @xmath68 .
* * let @xmath69 to be the trivial fragment .
* @xmath70 , @xmath71 , @xmath63 : * * let @xmath72 be the two - sorted language with sorts @xmath73 and @xmath74 , two binary relation symbols @xmath75 and @xmath76 on @xmath73 and @xmath74 , respectively , and one more relation symbol @xmath77 . define @xmath70 to contain the following : * * *
@xmath78 `` @xmath79 is an equivalence relation on @xmath74 . '' * * * @xmath80 . * * * `` @xmath75 is a linear order of @xmath73 '' and `` @xmath76 is a linear order of @xmath74 . ''
* * * for all @xmath0 , @xmath81 * * let @xmath82 be the smallest fragment of @xmath72 containing the formula @xmath83 .
* * for each @xmath54 , let @xmath84 be the the @xmath82-universal theory obtained from @xmath70 by adding the sentence @xmath85 * @xmath86 : * * let @xmath87 be a language with two binary relations .
define the theory of _ convexly - ordered equivalence relations _
, @xmath86 , to be the theory in @xmath88 given by defining @xmath89 to be an equivalence relation and @xmath90 to be a linear ordering such that equivalence classes are convexly ordered , i.e. the following holds in any model of @xmath86 : @xmath91 * * let @xmath21 denote the trivial fragment . to conclude this subsection ,
we introduce the notion of an indiscernible picture of a model of @xmath51 for a -like theory @xmath25 . of course , not all -like theories admit indiscernible pictures , but the definition is sensible even so .
by @xmath92 we mean the formulas in the fragment @xmath21 satisfied by @xmath93 in @xmath2 .
let @xmath25 be a -like theory , and let @xmath94 ( possibly uncountable ) .
let @xmath95 be an infinite structure .
an _ indiscernible picture _
@xmath96 ( of @xmath2 ) is an injection @xmath97 such that for all @xmath54 and @xmath98 , @xmath99 by an indiscernible @xmath26-picture in @xmath95 , we mean an indiscernible picture @xmath100 of some @xmath94 . in this subsection
, we establish the twin properties that identify generalized indiscernibles ( theories of indiscernibles ) among all -like theories .
the first of these the finitary ramsey property is precisely the hypothesis that allows the classical proof of the existence of indiscernible sequences ( via ramsey theorem and compactness ) to go through largely unchanged .
let @xmath25 be a -like theory .
* for @xmath42 , we say that @xmath25 has the @xmath6-ramsey property if for all @xmath101 and all @xmath28 , there is a @xmath33 such that for any coloring @xmath102 , there is an embedding @xmath103 such that for all @xmath104 , @xmath105 .
( this condition is denoted by the rado partition arrow : @xmath106 . )
* @xmath25 has the ramsey property if it has the @xmath6-ramsey property for every @xmath42 .
the second batch of definitions is more clearly identifiable as characterizing a `` theory of indiscernibles .
'' the reader may find it helpful to recall that for any infinite structure @xmath95 and any sequence @xmath107 , there is an indiscernible sequence @xmath108 ( in an elementary extension of @xmath95 ) `` patterned '' on @xmath107 .
the first two definitions below ( of @xmath1-embeddings and -templates ) formalize the idea of patterning .
the modeling property , then , uses these notions to describe what is , naturally , a `` theory of indiscernibles . ''
let @xmath25 be a -like theory , and let @xmath109 . by an `` @xmath1-embedding @xmath110 , ''
we mean a family @xmath111 of embeddings , where for each @xmath112 , @xmath113 is a finite model of @xmath26 containing @xmath114 .
implicitly , there is a uniform bounding function @xmath115 such that @xmath116 for all @xmath112 .
let @xmath25 be a -like theory , and let @xmath94 . in an arbitrary language @xmath117 ,
let @xmath95 be some @xmath117-structure .
we define an _ -template _ to be a family @xmath118 of maps @xmath119 ( @xmath54 ) satisfying the following : * for all @xmath54 , @xmath120 , @xmath121 and @xmath122 , @xmath123 * for all @xmath124 , @xmath125 and @xmath126 , @xmath127 now , suppose @xmath128 , @xmath129 , @xmath130 is an indiscernible picture of @xmath3 in @xmath131 .
we say that _ @xmath100 is patterned on @xmath118 _ if there is a family @xmath132 of patterning maps such that for all @xmath133 , @xmath134 and @xmath135 , @xmath136 .
a -like theory @xmath25 with generic model @xmath94 is said to have the _ modeling property _ if for every infinite structure @xmath95 and every -template @xmath137 , there are @xmath128 and an indiscernible picture @xmath138 patterned on @xmath118 .
finally , we can state the main characterization theorem connecting the ramsey property and the modeling property .
a very - slightly weaker form of this theorem was proved in @xcite , and in @xcite , it was proved as one of a somewhat longer list of equivalent conditions .
let @xmath25 be a -like theory
. then @xmath25 has the finitary ramsey property if and only if it has the modeling property .
a _ theory of indiscernibles _ is precisely a -like theory @xmath25 that has the modeling property ( equivalently , the finitary ramsey property ) .
we hope that the following examples will be helpful to the reader in understanding the definitions that we have just laid down .
[ example_mutuallyindiscernible ] @xmath139 is a -like theory , and @xmath140 is the theory asserting just that each @xmath141 is a dense linear order without endpoints . if @xmath142 is countable , then @xmath143 , which is extremely amenable .
so by lemma 6.7 of @xcite we can conclude that @xmath139 is a theory of indiscernibles .
now , consider an indiscernible picture @xmath96 , where @xmath95 is some infinite @xmath117-structure .
we note that for @xmath144 and @xmath145 , @xmath146 , @xmath147 and @xmath148 may be of different sorts .
however , each sequence @xmath149 ( @xmath150 ) is an indiscernible sequence in the usual sense , and moreover , the family @xmath151 comprises mutually indiscernible sequences in the sense that for each @xmath150 , @xmath152 is indiscernible over @xmath153 .
[ defn_nmultiorder ] @xmath154 is a -like theory , and @xmath155 is the theory where each @xmath58 is a dense linear order and these orders are independent
. then @xmath155 is @xmath50-categorical and admits elimination of quantifiers .
moreover , @xmath66 is a theory of indiscernibles . for more details ,
see definition 1.2 of @xcite , where this is called @xmath156 . also see corollary 1.4 of @xcite and theorem 4 of @xcite .
@xmath70 is a @xmath82-universal theory .
it is easy to see that @xmath157 is a -like theory .
in fact , the model companion shares an automorphism group with that of the shelah tree @xmath158 from @xcite where @xmath159 is the meet in the partial tree order @xmath160 , @xmath161 is the lexicographical order on sequences , and @xmath162 is a unary predicate picking out the @xmath0-th level of the tree .
identify leaves of @xmath163 with the one - point classes @xmath164 / \bigcap_{1 \leq 1 \leq n } e_{x_i}$ ] .
@xmath165 is a -like theory .
a countable model of the model companion @xmath166 is a collection of densely ordered convex equivalence classes , each of which is a dense linear order .
it probably is not surprising ( given an understanding of stable theories ) that for a given theory @xmath167 , not every indiscernible picture of a model of @xmath51 `` survives '' sometimes , the indiscernible picture is already indiscernible relative to a reduct of @xmath25 .
this phenomenon is the basis for our approach to dividing lines in this article , so obviously , we must formalize exactly what we mean .
let @xmath19 and @xmath168 be -like theories in @xmath20 and @xmath169 , respectively , and let @xmath170 and @xmath171 be their respective generic models .
* as usual , we say that @xmath172 is a reduct of @xmath173 , if there is a bijection @xmath174 such that for every formula @xmath175 of @xmath20 , there is a formula @xmath176 of @xmath169 such that @xmath177 for all @xmath178 .
as all of the theories in question are @xmath50-categorical , this is equivalent to asserting that , up to the identification provided by @xmath179 , every complete type of @xmath51 over @xmath180 is equal to a ( finite ) union of complete types of @xmath181 over @xmath180 .
we will say that @xmath19 is a reduct of @xmath168 if the generic model of the former is a reduct of the generic model of the latter .
let @xmath19 and @xmath168 be theories of indiscernibles , and let @xmath167 be an arbitrary complete theory .
assume @xmath19 is a reduct of @xmath168 so fix @xmath174 witnessing this , where @xmath172 and @xmath173 are the generic models of the two theories , respectively . * let @xmath182 be an indiscernible picture of @xmath171 in a model @xmath95 of @xmath167 .
we say that _
@xmath100 collapses to @xmath19 _ if the composite @xmath183 is an indiscernible picture of @xmath172 in @xmath95 . *
we say that _
@xmath167 collapses @xmath168 to @xmath19 _ if every indiscernible picture @xmath182 of @xmath171 in a model @xmath95 of @xmath167 collapse to @xmath19 .
we note that our formalization of collapse requires that a specific reduct is isolated before hand .
although it is also possible ( presumably ) to speak of `` collapse '' without requiring that the remainder is anything in particular , this is _ not _ the approach we adopt here .
although we can get similar collapse results for op - dimension and rosiness , for ntp2 , we will need to consider a slightly different notion .
this is primarily because of the fact that , while stability , nip , op - dimension , and rosiness are characterized by the non - existence of a pattern with only existential criteria , ntp2 ( and likewise , ntp1 for that matter ) is characterized by the non - existence of a pattern with both existential and universal requirements .
so we will need a notion of dividing across a fixed sequence @xmath100 to deal with the `` universal part '' of the pattern .
[ defn_divideacross ] let @xmath184 be an indiscernible picture of @xmath94 , and let @xmath185 be an @xmath117-formula .
we say that @xmath185 _ divides across @xmath100 _ if there is an indiscernible sequence @xmath186 in @xmath2 such that : @xmath187 notice that saying @xmath185 divides across @xmath100 , witnessed by @xmath186 , shows that @xmath188 divides over @xmath180 , since @xmath189 is an indiscernible sequence in @xmath95 . as usual , by compactness , this is equivalent to demanding that @xmath190 is @xmath191-inconsistent for some @xmath192 .
we will use this notion in section [ sect_ntp2 ] below .
we will introduce a specialized versions of dividing , namely `` dividing vertically . '' to connect diving to indiscernible collapse , we will also have a notion of _ coding _ formulas . though this definition depends on the specific index theory @xmath26 ,
the basic idea is to have a single @xmath117-formula witnessing the collapse of a particular relation symbol in @xmath20 for a indiscernible picture @xmath100 .
let @xmath167 be a complete first - order @xmath117-theory and let @xmath194 be a monster model for @xmath167 .
there are many equivalent definitions for op - dimension given in @xcite .
the main one of interest here will be the following .
we say that a partial type @xmath195 has _ op - dimension _
@xmath196 if there exists formulas @xmath197 and sequences @xmath198 for @xmath68 such that , for all functions @xmath199 , the following partial type is consistent @xmath200 we will call such @xmath201 an _ ird - pattern _ of length @xmath0 and depth @xmath202 in @xmath195 .
we will also denote this by @xmath203 .
we will write @xmath204 for @xmath205 .
if the op - dimension of @xmath206 is @xmath196 for all @xmath0 , we say it is infinite and write @xmath207 .
notice that by the modeling property in example [ example_mutuallyindiscernible ] , we may assume that the @xmath208 in the definition of op - dimension are such that @xmath209 is mutually indiscernible ( and indexed by @xmath210 or , in fact , any infinite order @xmath211 ) .
notice that if @xmath212 has finite op - dimension , then in particular @xmath167 has nip in the sort of @xmath213
. the converse does not hold .
consider @xmath214 ( see definition [ defn_index ] ) .
a partial type @xmath206 is stable if and only if @xmath215 .
it is easy to check that the op - dimension of a partial type @xmath206 is bounded above by the dp - rank of @xmath206 ( any ird - pattern can be converted to an ict - pattern ) . in general , however , op - dimension can be much smaller ( e.g. , there are stable , non - strongly dependent theories ) . on the other hand , op - dimension enjoys many of the same properties as dp - rank , including subadditivity ( see theorem 2.2 of @xcite ) .
that is , @xmath216 if @xmath167 is o - minimal ( or simply distal ) , then dp - rank and op - dimension coincide ( theorem 3.1 of @xcite ) . for a partial type @xmath195 over a set @xmath217 ,
the following are equivalent 1 .
@xmath218 , 2 . for all formulas
@xmath219 , all sequences @xmath220 indiscernible over @xmath217 , and for all @xmath221 , there exists a partition of @xmath210 into convex sets @xmath222 such that , for all @xmath223 , the set @xmath224 is either finite or cofinite in @xmath225 .
3 . for all @xmath226 almost mutually indiscernible over @xmath217 and all @xmath221 ,
there exists @xmath227 such that @xmath228 is almost indiscernible over @xmath229 .
[ prop_118guinhillop ] for a partial type @xmath195 and @xmath230 , the following are equivalent : 1 .
@xmath203 , 2 .
there exists an indiscernible @xmath66-picture @xmath231 and a formula @xmath219 such that , for all @xmath68 and for all @xmath232 a cut of @xmath233 , there exists @xmath234 such that @xmath235 . along similar lines , we can use collapse of indiscernible @xmath66-pictures to characterize op - dimension , supposing the ambient theory has nip .
the following is the main result of this section .
[ thm_opdimindisc ] suppose @xmath167 has nip .
for a partial type @xmath195 and @xmath230 , the following are equivalent : 1 .
@xmath236 , 2 . for all indiscernible @xmath66-pictures @xmath231 ,
there exists @xmath237 such that @xmath100 is an indiscernible picture for @xmath238 \(1 ) @xmath239 ( 2 ) : suppose ( 2 ) fails .
that is , there exists an indiscernible @xmath66-picture @xmath231 such that , for all @xmath237 , @xmath100 is not an indiscernible picture for @xmath238 .
that is , there exists @xmath192 and @xmath240 such that @xmath241 we may assume the same @xmath191 works for all @xmath237 ( take the maximal such ) and the coordinates of @xmath242 are distinct and the coordinates of @xmath243 are distinct . without loss of generality , by reordering , suppose @xmath244 since transpositions generate the symmetric group on @xmath191 , we may assume that the @xmath245-order type of @xmath242 and @xmath243 differ by transposing two elements .
that is , for some @xmath246 , @xmath247 since @xmath248 , there exists @xmath249 such that @xmath250 fix @xmath251 .
we use this as a template to construct an ird - pattern of depth @xmath252 and length @xmath0 in @xmath206 , showing that @xmath203 ( and hence ( 1 ) fails ) . for each @xmath237
, choose @xmath253 such that , for all @xmath68 , * @xmath254 if @xmath255 for all @xmath256 , * @xmath257 if @xmath258 for all @xmath259 , * @xmath260 if @xmath261 for all @xmath259 .
thereby , for all @xmath262 , for all @xmath263 , @xmath264 has the same @xmath58-order type as @xmath93 . in other words , for all @xmath262 , @xmath265 for each @xmath237 and @xmath266 , define @xmath267 as follows : for each @xmath68 , * @xmath268 for all @xmath269 , * @xmath270 .
so , for @xmath271 , we are merely taking @xmath272 and stretching it , in the @xmath273th coordinate , across @xmath274 .
now , for any @xmath275 , for any @xmath237 and any @xmath266 , we have * @xmath276 for @xmath277 and , * @xmath278 for @xmath279 . in other words , for any @xmath275 , @xmath266 , and @xmath237 , * @xmath280 for @xmath277 and , * @xmath281 for @xmath279 .
after embedding into @xmath2 ( by first extending each @xmath58 arbitrarily to be a linear order ) , we see that , for all @xmath275 , @xmath266 , and @xmath237 , @xmath282 let @xmath283 , let @xmath284 for all @xmath275 , and let @xmath285 for all @xmath237 and @xmath266 . then , for all @xmath237 , @xmath266 , and @xmath275 , @xmath286 hence , @xmath287 and @xmath288 form an ird - pattern of depth @xmath252 and length @xmath0 in @xmath206 .
since @xmath252 was arbitrary , by compactness , we get that @xmath203 , as desired .
\(2 ) @xmath239 ( 1 ) : suppose ( 1 ) fails , so @xmath203 . by proposition [ prop_118guinhillop
] , there exists an indiscernible @xmath66-picture @xmath231 and a formula @xmath219 such that , for all @xmath237 and for all @xmath232 a cut of @xmath289 , there exists @xmath234 such that @xmath235 .
since @xmath167 has nip , @xmath219 has nip , so fix @xmath251 such that @xmath290 has vc - dimension @xmath291 .
that is , @xmath292 we show that ( 2 ) fails on @xmath100 .
fix @xmath237 and suppose , by means of contradiction , that @xmath100 is an indiscernible picture for @xmath238 .
fix any @xmath293 . by choice of @xmath252 ( i.e. , that @xmath252 is greater than the vc - dimension of @xmath290 ) , there exists @xmath294 such that @xmath295 let @xmath296 and @xmath297 .
choose @xmath298 such that * @xmath299 , and * for all @xmath300 and @xmath301 , @xmath302 .
by indiscernibility with respect to @xmath303 , @xmath304 without loss of generality , by reordering , suppose @xmath305 , so we have @xmath306 where @xmath307 . now choose any @xmath232 a cut of @xmath289 with @xmath308 if and only if @xmath309 .
therefore , there exists @xmath234 such that @xmath310 .
in particular , @xmath311 this is a contradiction . hence ( 2 ) fails and we are done .
note that we only use nip in the proof of theorem [ thm_opdimindisc ] , ( 2 ) @xmath239 ( 1 ) .
however , this use is necessary . consider the theory @xmath167 of the random graph ( where @xmath312 ) and take @xmath313 $ ] in the homesort
then , any indiscernible sequence is an indiscernible set ( either everything is @xmath314-connected or @xmath314-disconnected ) .
that is , theorem [ thm_opdimindisc ] ( 2 ) holds for @xmath315 . on the other hand ,
theorem [ thm_opdimindisc ] ( 1 ) fails for all @xmath0 ( i.e. , @xmath207 ) . to see this , notice that @xmath316 together with any distinct @xmath317 form an ird - pattern of length @xmath0 .
indeed , for any @xmath199 , the type @xmath318 is consistent by saturation .
if we assume that @xmath167 is o - minimal , since dp - rank and op - dimension coincide , then this collapse of multi - order indiscernibles characterizes dp - rank .
there are several equivalent ways to define rosy theories , but for the sake of brevity , we present the one that is closest to the work we will carry in this article . this definition is based on the following notion of `` equivalence relation ranks '' as proposed in @xcite .
we remark that when defined in terms of local -ranks or the behavior of -independence , it is very important to work with a theory that eliminates imaginaries ; that is , the class of rosy theories is strictly contained the class of `` real - rosy theories . ''
however , the characterization in terms of equivalence relation ranks eliminates this discrepancy a theory @xmath167 has finite equivalence relation ranks if and only if @xmath319 has finite equivalence relation ranks .
let @xmath195 be a partial type , and let @xmath320 be a finite set of formulas @xmath321 such that for every @xmath322 , @xmath323 defines an equivalence relation .
( we call such a @xmath320 a _ finite set of families of equivalence relations_. ) * @xmath324 * for a limit ordinal @xmath325 , @xmath326 if @xmath327 for all @xmath328 . *
@xmath329 if there are a formula @xmath330 and a parameter @xmath322 such that , with @xmath331 , the set , @xmath332 is infinite . * @xmath333 if @xmath327 for every ordinal @xmath334 .
[ fact : omega - is - infty ] let @xmath195 be a partial type , and let @xmath320 be a finite set of families of equivalence relations . by compactness , if @xmath335 , then @xmath333 .
the key theorem of @xcite ( for our purposes ) is the following .
let @xmath167 be a theory with big model @xmath194 .
@xmath167 is rosy if and only if @xmath336 for every partial type @xmath195 and every finite set @xmath320 of families of equivalence relations . in the remainder of this subsection
, we define a variant of the equivalence relation ranks as a coding property , and we show that non - rosiness can also be characterized by `` coding equivalence relations . ''
let @xmath167 be a theory with big model @xmath194 .
for a formula @xmath337 , we say that @xmath338 _ codes equivalence relations _ if for every @xmath54 , there are @xmath339 of sort @xmath213 and @xmath340 of sort @xmath341 such that for all @xmath57 and @xmath342 , @xmath343 [ lemma : non - rosy - equals - coding ] let @xmath167 be a theory with big model @xmath194 , and assume @xmath167 eliminates imaginaries .
@xmath167 is non - rosy if and only if there is a partitioned formula @xmath337 that codes equivalence relations in @xmath194 .
first , suppose @xmath337 codes equivalence relations .
then , let @xmath344 by compactness and the fact that @xmath337 codes equivalence relations , there are @xmath345 and @xmath346 such that for all @xmath347 and @xmath348 , @xmath343 now , we make a simple observation : let @xmath349 , and for each @xmath350 , let @xmath351 ; then , @xmath352 for every @xmath350 . from this
, it follows immediately that @xmath353 , and this shows that @xmath167 is not rosy .
conversely , suppose @xmath167 is not rosy . through simple coding tricks
, we may assume that there is a single family of equivalence relations @xmath354 such that @xmath355 .
from this , we recover @xmath356 and @xmath357 such that , for any @xmath358 and @xmath359 ( @xmath350 ) @xmath360 by the pigeonhole principle , we may assume that @xmath361 whenever @xmath362 .
now , let @xmath54 be given . applying several automorphisms
, we may assume that @xmath363 whenever @xmath364 satisfy @xmath362 ; then , for @xmath365 , we choose any @xmath366 .
as @xmath0 was arbitrary , we have shown that @xmath367 codes equivalence relations .
we have established that non - rosiness is equivalent to `` coding equivalence relations . '' in this subsection , develop of theory of indiscernibles that corresponds to the latter notion .
@xmath157 is a theory of indiscernibles .
first we prove that @xmath368 is a theory of indiscernibles .
the base case follows since @xmath369 is interdefinable with the theory of linear order .
now suppose the generic model of @xmath84 has extremely amenable automorphism group @xmath370 , and consider @xmath371 .
call the @xmath372 simply @xmath373 for @xmath374 .
the @xmath114 sort is finite and rigid .
thus the automorphism group acts on the @xmath375 classes ( and permutes any point(s ) in the classes . ) adding a new equivalence relation cross - cuts each @xmath375 class .
thus the automorphism group of the generic model of @xmath376 is @xmath377 , which we know to be extremely amenable . to complete the argument for @xmath378
, there are two ways to proceed .
firstly , one can observe that the automorphism groups of the generic models of the @xmath368 s form a directed family of subgroups of the automorphism group of the generic model of @xmath378 , and that union of this family is dense the latter ; by lemma 6.7 of @xcite , it follows that @xmath378 is a theory of indiscernibles .
a second approach would be to use compactness and the modeling properties of the @xmath368 s to prove that @xmath378 also has the modeling property . now that we know what theory of indiscernibles to work with , we can state the main characterization theorem of this section .
its proof comprises all of the next subsection .
[ thm : collapse - result - rosiness ] let @xmath167 be a theory with big model @xmath194 , and assume @xmath167 eliminates imaginaries .
@xmath167 is rosy if and only if every indiscernible @xmath70-picture in @xmath194 is an indiscernible @xmath379-picture via @xmath380 , @xmath381 .
the proof of theorem [ thm : collapse - result - rosiness ] amounts to the conjunction of propositions [ prop : char - rosy - necessity ] and [ prop : char - rosy - sufficiency ] , which we prove in turn . throughout this subsection , assume that @xmath167 eliminates imaginaries .
[ prop : char - rosy - necessity ] let @xmath167 be a theory with big model @xmath194 .
if @xmath167 is rosy , then every indiscernible @xmath70-picture in @xmath194 is an indiscernible @xmath379-picture via @xmath380 , @xmath381 . before proving this proposition
, we will need to establish the following lemma .
[ lemma : indisc - to - fs ] let @xmath167 be a theory with big model @xmath194 . if there is an indiscernible @xmath70-picture in @xmath194 that is not an indiscernible @xmath379-picture via @xmath380 , @xmath381 , then there is a formula @xmath382 that codes equivalence relations in @xmath194 .
let @xmath383 be an indiscernible picture of the countable model @xmath384 , and suppose @xmath100 is not an indiscernible @xmath379-picture .
in particular , there are @xmath385 , @xmath386 and @xmath387 in @xmath388 , @xmath389 and @xmath390 in @xmath391 , and a formula @xmath185 in the language of @xmath167 such that : * @xmath392 and @xmath393 * @xmath394 * @xmath395 and @xmath396 by model - completeness and the indiscernibility of @xmath100 , we may assume that @xmath397 .
we will also take the lemma as obvious in the case that @xmath398 ( using @xmath399 to code equivalence relations ) , so in all that follows , we assume @xmath400 .
we may assume that @xmath401 and @xmath402 differ only in the following way : there are @xmath403 and @xmath404 such that for all @xmath405 , @xmath406 , @xmath407 let @xmath408 be the set of all functions @xmath409 satisfying the following conditions : 1 . for all @xmath405 and @xmath410 , @xmath411 @xmath412 2 .
for all @xmath413 and @xmath414 , for all @xmath404 , @xmath415 is a convex subset of @xmath416 then @xmath408 is precisely the set of functions @xmath417 that correspond naturally to quantifier - free - complete types @xmath418 extending @xmath419 ; we write @xmath420 and @xmath421 for the two directions of the bijection .
let s also name the function @xmath422 given by , @xmath423it s clear that @xmath424 . for any @xmath425 , we define @xmath426 as follows : * for @xmath427 , we say that @xmath428 is _ @xmath429-isolated _ if for all @xmath405 , @xmath430 . *
set @xmath431 * finally , define @xmath432 by @xmath433 it is not hard to verify that @xmath432 remains in @xmath408 , and it is equally clear that for each @xmath425 , there is a number @xmath434 such that if @xmath435 , and @xmath436 for each @xmath437 , then @xmath438 .
consequently , if @xmath439 and @xmath440 are the members of @xmath408 corresponding to @xmath441 and @xmath442 , then there is a `` path '' @xmath443 such that for each @xmath444 , either @xmath445 or @xmath446 .
since @xmath447 and @xmath448 , by indiscernibility , there is an @xmath444 such that @xmath449 then , replacing @xmath450 with @xmath451 and @xmath452 with @xmath453 is sufficient for the claim . by indiscernibility and model - completeness , we may further assume that @xmath454 whenever @xmath455 . finally , taking @xmath456 and @xmath457 @xmath458 ( when @xmath459 ) it is not difficult to verify ( following @xcite ) that @xmath460 codes equivalence relations .
( when @xmath461 , we use a similar definition of @xmath462 . )
proving the contrapositive , if there is an indiscernible @xmath70-picture in @xmath194 that is not an indiscernible @xmath379-picture via @xmath380 , @xmath381 , then by lemma [ lemma : indisc - to - fs ] , there is a formula @xmath382 that codes function spaces in @xmath194 .
then , by lemma [ lemma : non - rosy - equals - coding ] , the existence of a formula that codes equivalence relations in @xmath194 implies that @xmath167 is not rosy .
[ prop : char - rosy - sufficiency ] let @xmath167 be a theory with big model @xmath194 .
if every indiscernible @xmath70-picture in @xmath194 is an indiscernible @xmath379-picture via @xmath380 , @xmath381 , then @xmath167 is rosy .
for the contrapositive , suppose @xmath167 is not rosy .
by lemma [ lemma : non - rosy - equals - coding ] , there is a formula @xmath382 that codes function spaces in @xmath194 .
we will use this fact to construct an indiscernible @xmath70-picture in @xmath194 that is not an indiscernible @xmath379-picture via @xmath380 , @xmath381 . in this paragraph
, we collect the data necessary to devise an -template . by compactness , there are @xmath463 and @xmath464 in @xmath194 such that @xmath465 for all @xmath347 and @xmath466 .
let @xmath384 be the generic model , and let @xmath467 and @xmath468 be injective functions such that @xmath469 for all @xmath470 and @xmath471 . from these
, we define our -template by setting @xmath472 whenever @xmath473 and @xmath474 . by the modeling property of @xmath378
, we recover an indiscernible picture @xmath475 patterned on @xmath118 , and we take it as obvious that @xmath100 is not an indiscernible @xmath379-picture via @xmath380 , @xmath381 .
let @xmath167 be a theory with big model @xmath194 , and let @xmath185 be a partitioned formula of the language @xmath117 of @xmath167 . for @xmath101 ,
@xmath185 is said to have the _
@xmath191-tree property of the second kind _
( @xmath191-tp2 ) in @xmath167 if there is a family @xmath476 in @xmath194 such that : we will argue independently for why @xmath166 is a theory of indiscernibles .
by @xcite , the finite models of @xmath86 form a ramsey class , and so by @xcite , @xmath166 is a theory of indiscernibles .
these indiscernibles have already been widely used in the tp2 context , typically as _ array indiscernibles _ ( as described in @xcite ) where the language is composed of two linear orderings .
we follow claim 2.19 in @xcite in the sense that the absence of tp2 in a theory is characterized by dynamics in the em - type of indiscernible sequences .
namely , if the indiscernible sequence allows for a certain kind of consistency , it must then allow for more .
since we are tracking the dynamics of formulas asserting consistency , as they show up in the em - type , it is helpful to make use of a notation from @xcite : let @xmath483 be a countable model , and let @xmath484 be an indiscernible picture . for a formula @xmath485
, we say that @xmath338 _ codes equivalence on @xmath100 _ if , for all @xmath486 with @xmath487 , @xmath488 [ lemma_strictpat_ntp2 ] let @xmath167 be a theory with big model @xmath194 .
let @xmath475 be an indiscernible picture of @xmath483 .
then @xmath100 does not collapse to order if and only if there exist a formula @xmath485 and an embedding @xmath489 such that @xmath338 codes equivalence on @xmath490 . as before ,
one direction is clear , so we show the other direction .
suppose @xmath100 does not collapse to order .
then , there are @xmath491-increasing @xmath0-tuples @xmath492 from @xmath100 such that @xmath493 but @xmath494 do not have the same complete type in @xmath194 .
thus for some formula @xmath495 of the language of @xmath117 of @xmath167 , @xmath496 and @xmath497 .
let @xmath498 , @xmath499 .
in general , let @xmath500 be the set of all quantifier - free @xmath0-types @xmath501 where * @xmath502 for all @xmath503 , * @xmath504 for all @xmath68 , * if @xmath505 , then @xmath506 for all @xmath503 , * if @xmath507 for some @xmath508 , then @xmath509 , and * if @xmath505 for @xmath510 , then @xmath511 .
that is , @xmath500 are the set of all quantifier - free @xmath0-types realized by some @xmath491-ascending sequence of @xmath0-tuples in @xmath2 . hence @xmath512 .
now form a graph @xmath513 with @xmath514 and , for @xmath515 , @xmath516 if there exists @xmath517 such that that is , @xmath524 thinks @xmath525 is @xmath89-related to either @xmath526 or @xmath527 ( but not both , as they are @xmath89-unrelated ) and @xmath528 thinks that @xmath525 is @xmath89-unrelated to @xmath526 and @xmath527 .
to each edge @xmath516 , we can associate a number @xmath529 of the size of the @xmath89-class in @xmath528 to which @xmath524 thinks @xmath525 is @xmath89-related , namely @xmath530 if @xmath521 and @xmath531 if @xmath522 .
call this @xmath532 . the reader should check that @xmath533 is connected ( one can get between any two configurations by pulling apart or adding on a single element to an @xmath89-class iteratively ) . since @xmath100 is an indiscernible picture
, it induces a vertex @xmath534-coloring of @xmath533 , namely @xmath535 where @xmath536 if and only if , for some ( equivalently any ) realization @xmath537 ( with @xmath538 ) , @xmath539 .
hence , @xmath540 and @xmath541 .
in particular , @xmath542 is non - constant .
now choose an edge @xmath516 such that since @xmath533 is connected and @xmath542 is non - constant , such an edge exists .
let @xmath273 witness that @xmath545 is an edge , without loss suppose @xmath521 ( the argument for @xmath522 is similar ) , and let @xmath546 . without loss of generality ,
suppose @xmath547 and @xmath548 ( by possibly swapping @xmath549 and @xmath550 ) .
notice that , by minimal choice of @xmath544 , we have in fact that , for all @xmath551 , if @xmath552 then @xmath553 ( otherwise , we could find a path from @xmath554 to @xmath528 maintaining this condition of equality , hence contradicting the minimality of @xmath252 ) .
similarly , for all @xmath551 , if @xmath555 then @xmath556 .
now , for some realization @xmath557 , let @xmath558 by ultrahomogeneity , there exists an embedding @xmath559 such that @xmath560 and each element of @xmath561 is @xmath562 unrelated to both @xmath563 and @xmath564 .
we claim @xmath338 codes equivalence on @xmath490 , as desired .
recall the general definition of dividing across , given in definition [ defn_divideacross ] . for the specific case of convex equivalence relation indiscernibles
, we define dividing vertically .
one could also define the corresponding notion of dividing laterally .
let @xmath483 be a countable model , and let @xmath484 be an indiscernible picture .
for a formula @xmath485 , we say that @xmath338 _ divides vertically across @xmath100 _ if there exists an indiscernible sequence @xmath186 on @xmath2 with @xmath575 such that @xmath576 fix @xmath577 .
if @xmath338 codes equivalence on @xmath100 , then @xmath338 does not divide vertically across @xmath100 .
this is because , for any finite set @xmath578 of pairwise @xmath562-unrelated elements , there exists an element @xmath579 such that @xmath580 and @xmath581 for all @xmath582 .
therefore , @xmath583 for all @xmath582 . by compactness , for any @xmath584 of pairwise @xmath562-unrelated elements ,
@xmath585 is consistent . 1 .
@xmath167 is ntp2 .
2 . for every indiscernible @xmath86-picture @xmath475 , for every formula
@xmath485 , if @xmath586 does not divide vertically across @xmath100 , then @xmath185 does not divide across @xmath100 .
let @xmath475 be an indiscernible picture of @xmath483 and suppose @xmath577 is such that @xmath338 does not divide vertically across @xmath100 .
suppose @xmath338 divides across @xmath100 that is , we have an indiscernible sequence @xmath186 in @xmath2 such that @xmath190 is inconsistent . without loss ,
suppose @xmath587 for all @xmath588 .
the only two possibilities for @xmath186 are then since @xmath338 does not divide vertically across @xmath100 , the second case can not hold .
therefore , the first case holds .
however , by compactness , @xmath591 is inconsistent for some @xmath192 .
since @xmath100 is an indiscernible picture , for any @xmath562-related @xmath578 with @xmath592 , @xmath593 is inconsistent .
therefore , for any @xmath594 so that @xmath595 is isomorphic to @xmath596 ( where @xmath597 iff @xmath598 ) , we have that @xmath599 is a witness to the fact that @xmath185 has @xmath191-tp2 in @xmath167
. therefore , @xmath167 has tp2 .
this completes the proof of 1@xmath6002 .
let @xmath606 be the expansion of @xmath117 ( the language of @xmath167 ) with one new sort for @xmath2 and a new function symbol @xmath607 , and let @xmath118 be the set of sentences of @xmath608 that includes the diagram @xmath609 of @xmath2 , the elementary diagram @xmath610 of a model @xmath611 containing @xmath217 , and the sentences , @xmath612 for all @xmath613 . using @xmath217 to verify finite - satisfiability ,
it s clear that @xmath118 has a model , and from this model , we recover a picture @xmath614 ( as a restriction of the interpretation of @xmath615 ) such that @xmath616 for all @xmath613 . using @xmath211 to make an -template @xmath617 , and
with the modeling property of @xmath86 , we obatin an indiscernible picture @xmath618 patterned on @xmath619 so also satisfying statements ( [ array : thing1],[array : thing2 ] ) with @xmath100 in place of @xmath211 . in particular , @xmath620 thus the indiscernible picture given by @xmath100 does not collapse to order .
moreover , by compactness and , for any set of pairwise @xmath562-unrelated elements
@xmath584 , @xmath621 is consistent .
thus , @xmath338 does not divide vertically across @xmath100 . on the other hand , for any set of pairwise @xmath562-related elements @xmath584 ,
@xmath621 is inconsistent .
hence @xmath338 divides across @xmath100 .
therefore , we get a dichotomy for theories with ntp2 : either @xmath100 collapses to order or there exists @xmath577 and an embedding @xmath559 such that @xmath338 codes equivalence on @xmath490 . hence @xmath338 does not divide vertically across @xmath490 .
then , the theorem says that @xmath338 does not divide across @xmath490 . in summary : 1 .
@xmath167 is ntp2 .
2 . for every indiscernible @xmath86-picture @xmath475 ,
either @xmath100 is an indiscernible @xmath61-picture or , for every formula @xmath485 that does not divide vertically across @xmath100 , @xmath185 does not divide across @xmath100 .
note that the first condition in ( 2 ) above that @xmath100 be an indiscernible @xmath61-picture is a special case of the second condition , that for every formula @xmath622 that does not divide vertically across @xmath100 , @xmath219 does not divide across @xmath100 .
this is because if @xmath100 collapses to order , then the type of @xmath623 in @xmath194 is constant for @xmath491-increasing tuples @xmath93 from @xmath2 , and so there is no distinction between dividing vertically across @xmath100 and dividing across @xmath100 .
also note that the first condition in ( 2 ) on its own would not be good enough to imply ( 1 ) from the corollary ; look at the case where @xmath194 is a big model of @xmath166 .
we can argue that the theory is nip , because basic relations are nip , and therefore the theory is ntp2 .
however , it will not be the case that every indiscernible @xmath166-picture is an indiscernible @xmath61-picture .
consider the example where the identity @xmath624 is used to make the -template on which indiscernible @xmath625 is patterned .
clearly @xmath100 will not be an indiscernible @xmath61-picture , as the equivalence relation is present .
one way to think of the ( 1 ) @xmath239 ( 2 ) direction of the above corollary is that , in a tp2 theory , the specific formula witnessing non - collapse to order may not express a distinction between consistency and inconsistency for @xmath290-types , for some @xmath290 , as that would create a witness to tp2 .
thus we have a partial notion of collapse : in tp2 theories and for indiscernible @xmath86-pictures @xmath100 , there is some restricted set of formulas @xmath626 such that if @xmath491-increasing @xmath627 give @xmath628 disagreeing on their complete type in @xmath194 , it must be disagreement on some @xmath549 from this restricted set . | we use the notion of collapse of generalized indiscernible sequences to classify various model theoretic dividing lines .
in particular , we use collapse of @xmath0-multi - order indiscernibles to characterize op - dimension @xmath0 ; collapse of function - space indiscernibles ( i.e. parameterized equivalence relations ) to characterize rosy theories ; and finally , convex equivalence relation indiscernibles to characterize ntp2 theories . |
the data of neutrino oscillation experiments revealed a very noticeable form of a mixing matrix with @xmath5 and @xmath6 at the @xmath7 confidence level(cl ) .
the current data also include the mass squared differences which are accompanied by solar and atmospheric neutrino oscillations , @xmath8 and @xmath9 , respectively@xcite@xcite .
the central values of the elements of the pontecorvo - maki - nakagawa - sakata(pmns ) matrix are pointing a unique form of the matrix , which consists of @xmath10 and @xmath11 .
the three - flavor neutrino mixing matrix specified by those angle sizes is called tri - bi - maximal(tbm ) mixing@xcite , based on the idea that the neutrino mass matrix in a lepton flavor basis is an @xmath12 group matrix and a canonical subgroup @xmath13 plays a role of a @xmath14 interchange@xcite .
therefore , its elegant form suggests that the mass model has originated from a flavor symmetry . models with various types of discrete flavor symmetries have been constructed @xcite@xcite@xcite@xcite .
a previous study considered a minimal seesaw model with only two righthanded neutrinos , where texture zeros and equalities in the dirac mass matrix constrain the number of parameters and all the elements of the @xmath15 dirac matrix can be expressed in terms of low energy physical parameters@xcite . in this study
, @xmath0 symmetry is adopted to construct a minimal seesaw model , which is established with a small content of additional fields beyond the standard model(sm ) .
the higgs contents are extended to include three @xmath16 doublet scalars in one- or two- dimensional representations of @xmath4 addition to the standard model higgs , and then the model has a relatively simple higgs potential , while the models with three righthanded neutrinos constructed in a group framework of the same kind should have included a rather complicated potential@xcite@xcite .
the model naturally represents @xmath14 symmetry and the properties of group @xmath0 lead the basis to the one in which the charged lepton mass matrix is diagonal . since the exact @xmath0 symmetry for a given field content and given charge assignments is necessary , a meaningful issue can be the deviation from tbm mixing as a consequence of the symmetry breaking by a few vevs .
the outline is as follows .
section ii introduces the framework within the flavor symmetry and the lagrangian .
the aspects that may arise during the symmetry breaking processes and their low energy effects since the seesaw mechanism are reported in section iii .
some extended effects of symmetry breaking , which are plausible for the deviation from tbm , as well as the experimental accessibility of the model are considered in section iv . in the conclusion ,
the theoretical remarks on the model and an analysis on aspects of deviation from tbm mixing are summarized .
the appendix contains the potential of all scalar particles along with their vacuum expectation values .
the 5 irreducible representations of @xmath4 are @xmath17 while the tensor products of @xmath18 , one of the 1-dim representations , are so trivial that they are simply @xmath19 the tensor products of @xmath20 , the other 1-dim representation , are as follows : @xmath21 in general , the elements of @xmath3 are denoted by @xmath22 or @xmath23 , for @xmath24 and @xmath25 1 to 2 . according to the group theory details for the @xmath4-ii version in ref.@xcite , the product of @xmath3 with another @xmath3 is @xmath26 the su(2 ) lepton doublets @xmath27 , the righthanded charged lepton singlets @xmath28 , and the higgs scalar doublet @xmath29 in the standard model are assigned into such representations of @xmath0 as follows : @xmath30 in the above table , the su(2 ) representation and hypercharge of a field are denoted by the subscription ` @xmath31 ' of the gauge symmetry , and @xmath4 representation and @xmath32 charge of the field are denoted by the subscription ` @xmath33 ' . under flavor symmetry , the charge assignment in eq.([charge_lepton ] ) does not distinguish the @xmath34 flavor from the @xmath35 flavor , because @xmath1 and @xmath2 as well as @xmath36 and @xmath37 are in doublets .
therefore , this @xmath0 symmetry appears as @xmath14 symmetry . to explain the light neutrino masses , additional higgs scalars , @xmath38 and @xmath39 , and right - handed neutrinos , @xmath40 , are introduced .
their representations under gauge symmetry and flavor symmetry are given in the following table .
@xmath41 the flavor charges of @xmath42 and @xmath43 , @xmath44 and @xmath45 will be assigned according to the mass type , normal hierarchy or inverted hierarchy . in general , the lagrangian constructed with the leptons and higgs scalars in eq.([charge_lepton ] ) and eq.([charge_nh ] ) can be written as @xmath46 where @xmath47 .
each @xmath24 and @xmath48 runs 1 to 3 , whereas each @xmath49 and @xmath50 runs 1 to 2 .
the @xmath0 invariant higgs potential is @xmath51 where @xmath52 and @xmath53 since the product @xmath54 can be any of the following representations , @xmath55 or @xmath56 of @xmath0 . according to the product rules in eqs .
( [ prod1 ] ) - ( [ prod3 ] ) , @xmath57 , and @xmath58 . the potential in eq .
( [ potentialnh ] ) is symmetric under the interchange of @xmath59 and @xmath60 .
therefore , their vacuum expectation values ( vevs ) are the same , @xmath61 . the vevs @xmath62 and @xmath63 , which are denoted as @xmath64 and @xmath65 , respectively , will be defined to minimize the potential in eq .
( [ potentialnh ] ) in appendix.a . after spontaneous symmetry breaking
, the dirac mass matrix of the charged leptons from the yukawa couplings , @xmath66 , becomes @xmath67 thus , this flavor model generates a basis where the matrix of charged lepton masses is diagonal .
all the sm particles are @xmath32-even as shown in eq.([charge_lepton ] ) , so that their yukawa couplings are protected from the yukawa couplings with a @xmath32-odd higgs scalar , @xmath68 or @xmath39 .
* minimal model for normal hierarchy * : in a minimal model where only two sm singlet righthanded neutrinos @xmath42 and @xmath43 are added to the sm fermions , one zero mass eigenvalue is assigned to three active neutrinos . if the charges of right - handed neutrinos @xmath42 and @xmath43 are given as @xmath69 and @xmath70 , then the @xmath0 invariant yukawa couplings give rise to the matrix form : @xmath71 the yukawa coupling @xmath72 is obtained via the tensor product @xmath73 , whereas the coupling @xmath74 is obtained via the tensor product @xmath75 .
thus , the different product rules of the same couplings can cause different sizes in the coupling constants , between a pair of @xmath76 and a pair of @xmath77 .
this model assumes that @xmath78 . when @xmath79 and @xmath39 obtain their vacuum expectation values to minimize the potential @xmath80 in eq.([potentialnh ] ) , breaking the symmetry @xmath0 , the process to acquire the vevs @xmath81 and @xmath82 are described in eq.([vev_nh ] ) . with the flavor charges of @xmath42 and @xmath43 , the mass matrix of the right - handed neutrinos is diagonal , so that @xmath83 . according to the seesaw mechanism @xmath84 ,
the mass matrix of light neutrinos @xmath85 is @xmath86 where the overall minus sign has been removed by a phase transformation of neutrino fields .
if the masses of light neutrinos have a normal hierarchy , then @xmath87 . if @xmath88 , and the yukawa couplings @xmath89 s are constrained such that @xmath90 and @xmath91 , the mixing matrix of @xmath85 is exactly tbm .
besides , there are a priori conditions for tbm mixing that should be considered . in eq.([md_nh ] ) , the zero element is a key to @xmath92 in pmns , and the opposite sign between the 22 element and 32 element is necessary for @xmath93 .
the structure of the dirac matrix with @xmath94 , @xmath95 , and @xmath96 is derived from @xmath4 symmetry . in other words ,
@xmath92 and @xmath93 are the consequences of @xmath4 symmetry .
however , it is unlikely for a dirac matrix to have exact zero and exact equalities as a coincidence . it will be shown that slight discrepancies between those elements cause physically significant deviation from the tbm mixing matrix .
the non - zero eigenvalues of the light neutrino mass matrix in eq.([mass_nh ] ) are given by @xmath97 and @xmath98 , where @xmath99 and @xmath100 are the factors determined in terms of yukawa coupling constants @xmath101 s .
* minimal model for inverted hierarchy * : when the charges of righthanded neutrinos @xmath42 and @xmath43 are given as @xmath102 and @xmath70 , then the @xmath0 invariant yukawa couplings give rise to @xmath103 with such charge assignment , the mass matrix of righthanded neutrinos is diagonal ; @xmath104 .
according to the seesaw mechanism @xmath84 , the mass matrix of light neutrinos @xmath85 is @xmath105 where @xmath106 , and @xmath81 .
when taking the seesaw mechanism , the two zero elements in eq.([md_ih ] ) are necessary for @xmath92 , and @xmath107 is necessary for @xmath108 . in other words , the @xmath0 flavor symmetry results in the texture zeros and the equalities , which are the necessary conditions for tbm mixing in pmns .
non - zero eigenvalues of light neutrino mass matrix in eq.([mass_ih ] ) , @xmath109 and @xmath110 , are given by @xmath111 where @xmath99 and @xmath100 can be determined in terms of yukawa coupling constants @xmath101 s .
in addition , possible deviations from those specific elements in dirac mass matrix can cause a phenomenological deviation from tbm mixing in the pmns matrix .
it is necessary to examine the possible aspects that can appear when the zeros and equalities generated by the flavor symmetry are collapsed during symmetry breaking .
therefore , some simple cases will be discussed . in a trial model , it can be assumed to have a right - handed mass matrix as follows : @xmath112 and @xmath99 and @xmath100 , in normal hierarch case or in inverted hierarch case eq.([eigen_ih ] ) , are assumed to be one , for simplicity .
it is unlikely to have an exact equality @xmath113 , or to have an exact zero element , as shown in eq.([md_nh ] ) and eq.([md_ih ] ) , after a series of spontaneous symmetry breaking mechanisms . in order to examine the effects of symmetry breaking to a deviation from low energy tbm mixing matrix ,
the dirac mass matrix for normal hierarchy is assumed to have the following form : @xmath114 where @xmath115 or @xmath116 , and @xmath117 is a small value that can occur from @xmath0-violating yukawa coupling .
the factor @xmath118 is a yukawa coupling ratio that affects the mass ratio at low energy , which can arise from a @xmath119 product weaker than @xmath120 .
the curve ( a ) in fig.1 describes the change in @xmath121 as the ratio @xmath122 varies from 0 to 1.5 for @xmath123 , even though the change in @xmath121 turns out to be independent of @xmath117 . on the other hand ,
the variation in @xmath122 does not cause any change in @xmath124 or @xmath125 .
the curves ( b ) and ( c ) in fig.1 describe the changes in @xmath124 and the change in @xmath125 , respectively , as the ratio @xmath126 varies from 0 to 1 , while keeping @xmath127 . under the given conditions , a range of @xmath122 , @xmath128 ,
can lead to @xmath121 at the 90% cl .
the range of @xmath124 at the 90% cl matches with @xmath129 , while the range of @xmath130 at 90% cl matches with @xmath131 .
the size of @xmath118 does not affect the mixing angles in pmns , but does affect the mass ratios . for inverted hierarchy , the dirac mass matrix can be supposed to have the following form : @xmath132 where @xmath133 or @xmath116 .
yukawa couplings @xmath134 and @xmath135 are assumed to be very small so that @xmath136 and @xmath137 are much smaller than @xmath65 .
the curve ( a ) in fig.2 describes the change in @xmath121 as the ratio @xmath138 varies from 0 to 1.5 for @xmath113 and @xmath123 . under the given conditions , a range of @xmath138 , @xmath139 ,
can lead to @xmath121 at the 90% cl .
since @xmath65 is the vacuum expectation value of the higgs scalar in the standard model , this model does not cause any new physics below approximately 1 tev .
the curves ( b ) and ( c ) in fig.1 describe the changes in @xmath124 and @xmath125 , respectively , when the 2 - 1 and 3 - 1 elements obtain a non - zero value due to @xmath0-violating coupling , and their relative ratio @xmath140 varies from 0 to 10 . the range of @xmath124 at the 90% cl matches with @xmath141 , whereas the range of @xmath125 at the 90% cl matches with @xmath142 .
a model of lepton masses was constructed in the framework of a discrete flavor symmetry @xmath0 .
each family of sm leptons or each of the right - handed neutrinos , is distinguished from each other by their flavor charges .
however , a pair of @xmath1 and @xmath2 are assigned to the representation @xmath3 of @xmath4 so that they can not be distinguished under the yukawa interaction , and so are a pair of @xmath36 and @xmath37 , indicating @xmath14 symmetry .
when the symmetry is accompanied with @xmath32 symmetry , we can obtain the basis where the mass matrix of charged leptons is diagonal , as shown in eq.([smlepton ] ) .
since the sm particles are all @xmath32-even and the additional scalar particles are @xmath32-odd , there are no extra yukawa couplings of charged leptons found .
the model contains only two righthanded neutrinos so that the couplings with three scalar fields and three lefthanded neutrinos generate a @xmath15 yukawa matrix .
this model can be characterized by a few noteworthy points .
first , the group theoretical properties of @xmath4 gives rise to @xmath14 symmetry , accordingly tbm mixing , and the diagonal mass matrix of charged leptons .
second , the type of neutrino mass spectrum , normal or inverted hierarchy , is resulted in by the flavor symmetry .
the flavor charges of righthanded neutrinos play a key role in determining the order of light neutrino masses .
third , it is possible to build the mass matrices with a significantly small content of scalar particles under a non - abelian discrete flavor symmetry @xmath0 .
the texture zeros or the equalities due to the symmetry in a yukawa matrix appear to be barely protected , once the symmetry undergoes a chain of breakdown .
it is natural for pmns to obtain a deviation from tbm mixing when the symmetry of the yukawa matrix is broken .
fig.[fig : nh3 ] and fig.[fig : ih3 ] show a schematic diagram of some aspects that can appear in the correlation between the order of symmetry breaking and the deviation from tbm .
ks wishes to thank the members in the physics department at chonnam national university for their warm hospitality .
the @xmath0 invariant higgs potential in eq.([potentialnh ] ) can be rephrased in terms of @xmath143 with @xmath144 and 2 , @xmath39 , and @xmath29 .
@xmath145 when the higgs particles obtain their vacuum expectation values such that @xmath146 , @xmath147 , and @xmath148 where @xmath149 and @xmath116 are real , the minimal potential can be expressed as follow : @xmath150 which satisfies @xmath151 where @xmath152 , @xmath153 and @xmath154 .
therefore , the vacuum expectation values @xmath155 and @xmath116 are k. nakamura _ et al . _
[ particle data group ] , j. phys .
g * 37 * , 075021 ( 2010 ) .
m. c. gonzalez - garcia and m. maltoni , phys .
* 460 * , 1 ( 2008 ) p. f. harrison , d. h. perkins and w. g. scott , phys .
b * 530 * , 167 ( 2002 ) p. f. harrison and w. g. scott , phys .
b * 557 * , 76 ( 2003 ) a. zee , phys .
b * 630 * , 58 ( 2005 ) e. ma , phys .
d * 72 * , 037301 ( 2005 ) k. s. babu and x. g. he , arxiv : hep - ph/0507217 .
e. ma , mod .
a * 20 * , 2601 ( 2005 ) g. altarelli and f. feruglio , nucl .
b * 741 * , 215 ( 2006 ) x. g. he , y. y. keum and r. r. volkas , jhep * 0604 * , 039 ( 2006 ) x. g. he and a. zee , phys .
b * 645 * , 427 ( 2007 ) r. r. volkas , arxiv : hep - ph/0612296 . f. feruglio , c. hagedorn , y. lin and l. merlo , nucl .
b * 775 * , 120 ( 2007 ) [ erratum - ibid . * 836 * , 127 ( 2010 ) ] w. grimus , l. lavoura and p. o. ludl , j. phys .
g * 36 * , 115007 ( 2009 ) t. araki , j. mei and z. z. xing , phys .
b * 695 * , 165 ( 2011 ) c. s. lam , phys .
* 101 * , 121602 ( 2008 ) h. zhang , phys .
b * 655 * , 132 ( 2007 ) f. bazzocchi , l. merlo and s. morisi , nucl .
b * 816 * , 204 ( 2009 ) f. bazzocchi , l. merlo and s. morisi , phys .
d * 80 * , 053003 ( 2009 ) s. chang , s. k. kang and k. siyeon , phys .
b * 597 * , 78 ( 2004 ) l. merlo , arxiv:1004.2211 [ hep - ph ] . | this work proposes a neutrino mass model that is derived using the minimal seesaw mechanism which contains only two right - handed neutrinos , under the non - abelian discrete flavor symmetry @xmath0 .
two standard model doublets , @xmath1 and @xmath2 , are assigned simultaneously to a @xmath3 representation of @xmath4 . when the scalar fields introduced in this model , addition to the standard model higgs , and the leptons are coupled within the symmetry , the seesaw mechanism results in the tri - bi - maximal neutrino mixing .
this study examined the possible deviations from tbm mixing related to the experimental data . |
topological insulators ( tis ) represent new type of material which has emerged in the last few years as one of the most actively research subjects in condensed matter physics .
they are characterized by a full insulating gap in the bulk and gapless edge or surface states , which are protected by the time - reversal symmetry @xmath3 .
regarding bernevig and hughes prediction @xmath4 , tis have been experimentally observed with such properties that host bound states on their surface , e.g. in 3-dimensional topological insulators ( 3dti ) @xmath5 , @xmath6 , @xmath7 and @xmath8 alloy , and also in the @xmath9 quantum well heterostructure @xmath10 .
these states form a band - gap closing dirac cone on each surface , and lead to a conducting state with properties unlike any other known electronic systems .
in particular , conformity of the conduction and valence bands to each other in and around dirac points in the first brillouin zone , possessing an odd number of dirac points , description of fermionic excitations as massless two - dimensional chiral dirac fermions , depending chirality on the spin of electron , having the significant electron - phonon scattering on the surface , owning very low room - temperature electron mobility are the peculiar properties of electronic structure of tis .
interestingly , the charge carriers in the surface states can behave as massive dirac fermions @xmath11 due to its proximity to a ferromagnetic material , that the vertical component of the magnetic vector potential may be proportional to the effective mass of dirac fermion .
the experimentally observed proximity - induced superconductivity on the surface state is another interesting dynamical feature occuring in 3dti , see refs .
@xmath12 . more importantly
, the coexistence of superconductivity and ferromagnetism as one of potential interests for spintronics and high magnetic field applications has firstly been predicted by fulde and ferrel @xmath13 , and larkin and ovchinnikov @xmath14 as _ fflo state_. this effect can be in compliance with standard bcs theory for phonon - mediated @xmath15-wave superconductivity , because the ferromagnetic exchange field is expected to prevent spin - singlet cooper pairing , ( see , ref .
@xmath16 as a prior work )
. the magnetic polarization of a pair electron caused by a ferromagnetic material can lead to the different momentum of cooper pair occurring in a ferromagnetic superconducting ( fs ) segment .
it seems to be in contrast to the formation of a typical cooper pair , where two electrons may be in opposite spin direction with the same momentum .
however , bergeret et.al . @xmath17 and li et.al .
@xmath18 have studied the effect of superconductor / ferromagnetic bilayer on the critical josephson current , where the orientation of ferromagnetic exchange field strongly affects the critical current .
also , the effect of superconductivity in coexistence with ferromagnetism has been studied on the superconducting gap equation for two case of singlet @xmath15-wave and triplet @xmath19-wave symmetries @xmath20 .
the authors have reconsidered the clogston - chandrasekhar limiting @xmath21 . according to the clogston criterion in the conventional fs mixture
, the normal state is regained as soon as the ferromagnetic exchange field exceeds @xmath22 at zero temperature . to be empirically
, the @xmath23 @xmath24 has been discovered to be the first ferromagnetic superconductor , which superconductivity is found to occur in a small temperature interval with adjusted ferromagnetic phase .
also , superconductivity is detected in itinerant ferromagnetic @xmath25 in a limited range of pressure and temperature @xmath26 .
regarding several works in the recent few years concerning with the topological insulator - based junctions @xmath27 , which are related to the andreev process and resulting subgap conductance we proceed , in this paper , to theoretically study the dynamical properties of dirac - like charge carriers in the surface states of 3dti under influence of both superconducting and ferromagnetic orders via the introducing the proper form of corresponding dirac spinors , which are principally distinct from those given in ref .
the magnetization induction opens a gap at the dirac point ( no inducing any finite center of mass momentum to the cooper pair ) , whereas the superconducting correlations causes an energy gap at the fermi level in the 3dti .
it will be particularly interesting to investigate the topological insulator superconducting electron - hole excitations in the presence of a exchange field .
we assume that the fermi level is close to the dirac point , and the ferromagnet has a magnetization @xmath29 .
the chirality conservation of charge carriers on the surface states in the presence of magnetization ( due to opening the band gap ) allows to use a finite magnitude of @xmath30 . in the absence of topological insulator ,
the spin - splitting caused by magnetization gives rise to limiting the magnitude of @xmath30 in a fs structure .
these excitations , therefore , are found to play a crucial role in andreev reflection ( ar ) process leading to the tunneling conductance below the renormalized superconducting gap .
particularly , we pay attention to the formation of majorana bound energy mode , as an interesting feature in topological insulator ferromagnet / superconductor interface , depending on the magnetization of fs hybrid structure .
we present , in section 2 , the explicit signature of magnetization in low - energy effective dirac - bogogliubov - de gennes ( dbdg ) hamiltonian .
the electron(hole ) quasiparticle dispersion energy is analytically calculated , which seems to exhibit qualitatively distinct behavior in hole excitations ( @xmath31 ) by varying the magnitude of magnetization . by considering
the magnetization is ever less than chemical potential in fs region , the superconducting wavevector and corresponding eigenstates are derived analytically .
section 3 is devoted to unveil the above key point of fs energy excitation , majorana mode energy , andreev process and resulting tunneling conductance in n / f / fs junction and respective discussions . in the last section ,
the main characteristics of proposed structure are summarized .
in order to investigate how both superconductivity and ferromagnetism induction to the surface state affects the electron - hole excitations in a 3dti hybrid structure , we consider magnetization contribution to the dbdg equation .
let us focus first on the hubbard model hamiltonian @xmath32 that is included the effective exchange field @xmath33 follows from : @xmath34 where @xmath35 denotes the effective attractive interaction between arbitrary electrons , labeled by the integer @xmath36 and @xmath37 with spins @xmath15 and @xmath38 . the matrices @xmath39 are responsible for the hopping between different neighboring sites , and @xmath40 and @xmath41 indicate the second quantized fermion and number operators , respectively . here
, @xmath42 is the vector of pauli matrix . using the hartree - fock - gorkov approximation and bogoliubov - valatin transformation @xmath43 , the bogoliubov - de gennes hamiltonian describing dynamics of bogoliubov quasiparticles
is found . in nambu basis , that electron(hole ) state is given by @xmath44 , the bdg hamiltonian for a @xmath15-wave spin singlet superconducting gap in the presence of an exchange splitting can be written as : @xmath45 where @xmath46 denotes the non - superconducting schrodinger - type part , and @xmath47 is superconducting order parameter . in the simplest model
, @xmath47 can be chosen to be real to describe time - reversed states . the effective exchange field by rotating our spin reference frame can be gain as @xmath48 .
the four corresponding levels of a singlet superconductor in a spin magnetic field is obtained @xmath49 with @xmath50 , where @xmath51 is the normal state energy for @xmath46 .
however , dependence of superconducting order parameter on the exchange energy can be exactly derived from self - consistency condition @xmath20 : @xmath52 where @xmath53 is the conventional order parameter in absence of ferromagnetic effect , @xmath54 and @xmath55 are the boltzmann constant and temperature , respectively .
the exchange splitting dependence of superconducting gap indicates that equation has no functionality of @xmath33 at zero temperature .
this takes place under an important condition known as clogston - chandrasekhar limiting @xmath21 .
according to this condition , if the exchange splitting becomes greater than a critical value @xmath56 , then the normal state has a lower energy than the superconducting state .
this means that a phase transition from the superconducting to normal states is possible when the exchange splitting is increased at zero temperature .
we now proceed to treat such a ferromagnetic superconductivity coexistence at the dirac point of a 3dti .
it should be stressed that the dressed dirac fermions with an exchange field in topologically conserved surface state have to be in superconducting state . here
, the influence of exchange field interacts in a fundamentally different way comparing to the conventional topologically trivial system , where the exchange field splits the energy bands of the majority and minority spins .
a strong ti is a material that the conducting surface states at an odd number of dirac points in the brillouin zone close the insulating bulk gap unless time - reversal symmetry is broken .
candidate dirac - type materials include the semiconducting alloy @xmath57 , as well as hgte and @xmath58 under uniaxial strain @xmath59 . in the simplest case
, there is a single dirac point in the surface fermi circle and general effective hamiltonian is modeled as @xmath60 , where @xmath61 indicates the surface fermi velocity , and @xmath62 is the chemical potential . under the influence of a ferromagnetic proximity effect , the hamiltonian for the two - dimensional surface states of a 3dti reads as : @xmath63 where the ferromagnetic contribution corresponds to an exchange field @xmath64 .
it has been shown @xmath11 that transverse components of the magnetization on the surface @xmath65 are responsible to shift the position of the fermi surface of band dispersion , while its perpendicular component to the surface induces an energy gap between conduction and valence bands . in what follows
, we will employ the relativistic generalization of bdg hamiltonian , which is interacted by the effective exchange field to obtain the dispersion relation of fs dressed dirac electrons in a topological insulator : @xmath66 the superconducting order parameter now depends on both spin and momentum symmetry of the cooper pair , that the gap matrix for spin - singlet can be given as @xmath67 , where @xmath68 is the uniform amplitude of the superconducting gap and phase @xmath69 guarantees the globally broken @xmath70 symmetry . by diagonalizing this hamiltonian
we arrive at an energy - momentum quartic equation . without lose of essential physics
, we suppose the component of magnetization vector along the transport direction to be zero @xmath71 for simplicity . also , we set @xmath72 , since the analytical calculations become unwieldy otherwise .
the dispersion relation resulted from eq . for electron - hole excitations
is found to be of the form : @xmath73 where , the parameter @xmath74 denotes the electron - like and hole - like excitations , while @xmath75 distinguishes the conduction and valence bands .
we might expect several anomalous properties from the above superconducting excitations , which is investigated in detail in the next section .
equation is clearly reduced to the standard eigenvalues for superconductor topological insulator in the absence of exchange field as @xmath76 ( see ref .
@xmath11 ) , @xmath77 .
the mean - field conditions are satisfied as long as @xmath78 . in this condition ,
the exact form of superconducting wavevector of charge carriers can be acquired from the eigenstates @xmath79 .
the hamiltonian eq . can be solved to obtain the electron ( hole ) eigenstates for fs topological insulator .
the wavefunctions including a contribution of both electron - like and hole - like quasiparticles are analytically found as : @xmath80 where we define @xmath81 note that , the solution is allowed as long as the zeeman field being lower than chemical potential @xmath82 .
we consider andreev reflection in a hybrid n / f / fs structure formed on the surface of a 3dti which coexistence between ferromagnet and superconductor is assumed to be induced by means of the proximity effect .
the wide topological insulator junction is taken along the @xmath83-axis with the fs region for @xmath84 , f region for @xmath85 and n region for @xmath86 .
the superconducting order parameter vanishes identically in n and f regions , and we can neglect its spatial variation in the fs region close to the interface .
the magnetization vectors of both sections is taken , in general , @xmath87 , which can be at the parallel or antiparallel configuration , as shown in fig .
1 . in the scattering process follows from the blonder - tinkham - klapwijk ( btk ) formula @xmath88 , we find the reflection amplitudes from the boundary condition at the interface . in ferromagnetic case , right- and left- moving electrons ( holes ) with energy excitation @xmath89 below the superconducting gap , transmitted ( normal reflected ) from the n region and reflected ( andreev reflected ) at the fs interface
thus , the leftover @xmath90 charge is transferred into the fs region as a cooper pair at fermi level . at energy excitation above
the normalized superconducting gap resulted from eq .
( 5 ) ( see , in particular , fig .
2 ) quasiparticle states can directly tunnel into the superconducting section .
the reflected hole leading to ar can be actually controlled by the doping level in order to take place possible specular andreev reflection .
particularly , we have to determine ( via the dynamical features of system ) the allowed values of fermi energy in three regions .
we set the fermi energy to zero in f region .
the electron(or hole ) transmitted to the fs region angle may be accordingly obtained from the fact of conservation of transverse wavevector under quasiparticle scattering at the interface : @xmath91 where @xmath92 and @xmath93 are the chemical potential and incidence angle in n region , respectively . as an important point ,
the electron(hole ) angle of incidence in all regions may be span the range from @xmath94 to @xmath95 around the normal axis . regarding the eqs .
, the angle @xmath96 needs to be meaningful when the chemical potential of fs region takes a magnitude greater than its value in n region ( @xmath97 ) . on the other hand
, we previously applied the condition @xmath98 , as an experimentally used manner to calculate the wave functions eq .
( 6 ) . by introducing the normal @xmath99 and andreev @xmath100 reflection coefficients and the scattering coefficients in f region ,
the total wave function inside the n and f region can be written as : @xmath101 @xmath102 where the eigenvectors @xmath103 can be found in appendix a. the probability amplitude of reflections in eq .
are calculated from the continuity of the wavefunctions at the interface .
the wave function in fs region is defined as @xmath104 .
finally , we find the following analytical expressions for the reflection coefficients , that the auxiliary quantities is described in appendix a : @xmath105(i\sin{(k^{xe}_{f}l}))+\ ] ] @xmath106\cos{(k^{xe}_{f}l})-1,\ ] ] @xmath107(i\sin{(k^{xe}_{f}l}))-\ ] ] @xmath108\cos{(k^{xe}_{f}l)}. \label{r}\ ] ] the reflection amplitudes measurements under the btk formalism enables us to capture the tunneling conductance through the junction : @xmath109 where the critical angle of incidence @xmath110 is determined depending on the doping of f region . the quantity @xmath111 is a renormalization factor corresponding to the ballistic conductance of normal metallic junction .
in this section , we proceed to analyze in detail the dynamical features of dirac - like charge carriers in 3dti with ferromagnetic and superconducting orders deposited on top of it .
we assume that the fermi level controlled by the chemical potential @xmath62 is close to the dirac point . in this case
, it is expected the signature of @xmath98 to be significant . in fig .
2 , we demonstrate the fs 3dti electron - hole excitations .
a net superconducting gap @xmath112 is obtained in dirac points ( for @xmath113 , where @xmath114 is fermi wavevector ) when we set @xmath115 .
increasing @xmath0 up to its possible maximum value results in three outcomes : i ) the superconducting excitations , which is renormalized by a factor @xmath116 , disappear in hole branch ( @xmath31 ) .
it means that for the greater magnetizations , if we consent the superconductivity in fs 3dti still exists , there is almost vanishing quantum state for reflected hole by andreev process in the valence band , ii ) dirac point is shifted towards smaller fs quasiparticle electron - hole wavevectors , iii ) the superconducting gap decreases slowly , where the variation of net gap is very low @xmath117 .
the andreev process , therefore , is believed to inconsiderably supress .
the signature of these valence band excitations can be clearly shown in ar , where the majorana mode may also be formed at the 3dti f / fs interface @xmath118 . as a verified result , considering the topological insulator interface between the ferromagnetic insulator and conventional superconductor leads to the appearance of the chiral majorana mode as an andreev bound state .
in other words , the majorana mode and andreev reflection are strongly related to each other .
the latter can be realized by the fact of looking for bound energies produced by the perfect ar , which yields the following solution : @xmath119,\ ] ] where we define @xmath120 we have checked numerically that sign of @xmath121 is changed by @xmath122 .
thus , the sign of andreev resonance states may be changed by reversing the direction of @xmath2 , and it corresponds to the chirality of majorana mode energies . as shown in fig .
3 , the slope of the energy curves of @xmath123 around @xmath124 become steep and show no change with the increase of @xmath125 for fixed @xmath2 , while it exhibits significantly decreasing behavior with the increase of @xmath126 for fixed @xmath0 .
the dispersion of majorana modes along the interface ( @xmath127 ) decreases with the increase of both magnetizations of fs and f regions .
note that , due to the presence of @xmath0 it needs to consider the fermi level mismatch between normal and fs sections , i.e. @xmath128 . then , the above contributions can be considerable in andreev process and resulting subgap tunneling conductance . from the angle - resolved andreev and normal reflection probabilities using eq .
, we see from fig .
4(a ) the main contribution of ar belongs to the angle of incidence @xmath129 in zero bias .
it , therefore , is expected to achieve the lower zero bias conductance , as shown in fig .
4(b ) and ( c ) .
furthermore , varying @xmath0 has no significant influence on ar in zero bias @xmath130 owing to the very small decrease of the renormalized superconducting gap with the increase of the @xmath0 , while the increasing @xmath2 results in more suppression of ar .
the latter can be understood by the increase of band gap in dirac point in f region .
the resulting normalized angle - averaged tunneling conductance curves are reported in figs .
4(b ) and ( c ) for two parallel and antiparallel configurations of magnetizations in fs and f regions .
zero bias conductance peak disappears with the decrease of the @xmath2 , and instead of it a high conductance peak appears in bias @xmath131 .
this result should be compared to that is obtained in ref .
interestingly , by increasing the @xmath0 the magnitude of subgap bias @xmath133 , for which the new peak takes place , is limited , as seen from fig .
thus , parameter @xmath134 can be considered as a _ bias - limitation coefficient _ " .
these features have been obtained when the direction of magnetizations in f and fs regions are at the parallel configuration .
the fundamentally distinct scenarios we find for the case of antiparallel configuration of magnetizations . in this case , first , the tunneling subgap conductance is enhanced , secondly , the zero bias conductance peak presented in parallel case is replaced by a deep , see , in detail , fig .
dynamically description , when the direction of @xmath2 is inverted we , indeed , meet with an inverse energy gap in dirac point of 3dti giving rise to enhancing the conductance peak respective bias energy @xmath135 in low values of @xmath2 .
one can express that the zero bias conductance originates from the chiral majorana mode , which significantly depends on the @xmath2 .
the chirality actually corresponds to the sign of @xmath2 , while the magnitudes of zero bias conductance at the parallel and antiparallel configurations are the same .
hence , the both deep and peak of conductance curves in antiparallel case are significance being influenced by the inverted gap caused by the @xmath136 .
remarkably , the importance of above findings can be featured by the capture of magnetoresistance ( mr ) of the topological insulator junction .
the magnetization ( specialy in fs region ) dependence of mr is presented in fig . 5 , where we observe a considerable mr peak for extra values of @xmath2 ( e.g. @xmath137 in figure ) .
importantly , increasing the @xmath0 weakens the mr peak , since , regarding the superconducting excitations in fig .
2 , the ar is more or less suppressed in the presence of @xmath0 and fermi wavevector mismatch also causes to decrease the @xmath138-bias conductance peak at the antiparallel configuration . according to the conductance curves ,
there is no mr in zero bias .
in summary , we have investigated the influence of ferromagnetic superconducting orders coexistence in the surface state of topological insulator .
the topological insulator superconducting electron - hole excitations in the presence of magnetization have led to achieve qualitatively distinct transport properties in tunneling n / f / fs junction .
one of key findings of the present work is that the resulting subgap conductance has been found to be strongly sensitive to the parallel or antiparallel configuration of magnetization directions in fs and f regions .
thus , this feature has actually led to present the magnetoresistance peak for bias energy close to the renormalized superconducting gap @xmath139 , which the bias limitation coefficient @xmath140 includes the magnetization of fs region @xmath0 .
particularly , we have found the presence of majorana mode at the f / fs interface to be controlled by the tuning of magnetizations magnitude .
however , these results have been obtained in the case of @xmath141 and @xmath142 , which is relevant to the experimental regime .
to complete calculation of probability of reflections in n / f / fs junction , we write down right and left moving electron and hole spinors in f and
n region : @xmath143^{t},\ \ \psi^{e-}_{n}=\left[1,-e^{-i\theta},0,0\right]^{t},\ \ \psi^{h-}_{n}=\left[0,0,1,-e^{-i\theta}\right]^{t},\ ] ] @xmath144^{t},\ \psi^{e-}_{f}=\left[1,-\alpha e^{-i\theta_f},0,0\right]^{t } , \
\psi^{h+}_{f}=\left[0,0,1,\alpha e^{i\theta_f}\right]^{t } , \
\psi^{h-}_{f}=\left[0,0,1,-\alpha e^{-i\theta_f}\right]^{t}\ ] ] where we define @xmath145 . by matching boundary conditions on @xmath146 and @xmath147 at @xmath148 and @xmath147 and @xmath149 at @xmath150 ,
the reflection amplitudes are obtained .
we introduce auxiliary quantities in eq . as : @xmath151 with @xmath152,\ ] ] @xmath153 @xmath154 * figure captions * + * figure 1 * ( color online ) sketch of the topological insulator - based n / f / fs junction .
the magnetization vectors in f and fs regions can be at the parallel or antiparallel configuration .
+ * figure 2 * ( color online ) the ferromagnetic superconducting excitation spectra on the surface state of 3dti for several values of @xmath157 , calculated from eq .
we set the net value of superconducting gap @xmath158 ( this value of pair potential is taken only to more clarify the behavior of spectra in dirac point , although it does not further need to use it in our calculations , since @xmath159 is supposed .
+ * figure 3 * ( color online ) the dispersion of majorana modes as a function of the electron incident angle for several values of magnetizations in fs and f regions .
the solid lines correspond to @xmath160 and the dashed lines to @xmath161 . + * figure 4(a ) , ( b ) , ( c ) , ( d ) * ( color online ) ( a ) probability of the normal and andreev reflections as a function of electron incidence angle at the interface in zero bias @xmath162 with @xmath163 and @xmath164 .
the plots show the results for different values of @xmath2 .
( b ) normalized tunneling conductance versus bias voltage @xmath165 and magnetization of f region .
we set @xmath163 ( c ) normalized tunneling conductance versus bias voltage and magnetization of fs region .
we set @xmath160 ( d ) the tunneling conductance as a function of bias voltage for two @xmath166 signs of @xmath2 , corresponding to the parallel and antiparallel configurations in f and fs regions .
the solid lines correspond to @xmath167 and marker dashed lines correspond to @xmath136 .
we set @xmath160 .
+ * figure 5 * ( color online ) the magnetoresistance spectra as function of bias voltage , where the influence of @xmath0 and @xmath2 is indicated , separately .
we have set @xmath168 in the resulting conductance and magnetoresistance spectra . | among the potential applications of topological insulators , we investigate theoretically the effect of coexistence of proximity - induced ferromagnetism and superconductivity on the surface states of 3-dimensional topological insulator , where the superconducting electron - hole excitations can be significantly affected by the magnetization of ferromagnetic order .
we find that , majorana mode energy , as a verified feature of ti f / s structure , along the interface sensitively depends on the magnitude of magnetization @xmath0 in fs region , while its slope in perpendicular incidence presents steep and no change .
since the superconducting gap is renormalized by a factor @xmath1 , hence andreev reflection is more or less suppressed , and , in particular , resulting subgap tunneling conductance is more sensitive to the magnitude of magnetizations in fs and f regions
. furthermore , an interesting scenario happens at the antiparallel configuration of magnetizations @xmath2 and @xmath0 resulting in magnetoresistance in n / f / fs junction , which can be controlled and decreased by tuning the magnetization magnitude in fs region .
* pacs * : 74.45.+c ; 85.75.-d ; 73.20.-r + * keywords * : topological insulator ; ferromagnetic superconductivity ; andreev reflection ; majorana mode ; tunneling conductance |
in 1989 , weiss , von klitzing , ploog , and weimann @xcite discovered the striking effect of a periodic potential varying in one direction on the magnetoresistivity of a two - dimensional electron gas ( 2deg ) .
they observed that even a weak potential modulation ( grating ) with a wave vector @xmath0 induces very strong oscillations of the magnetoresistivity @xmath1 , while showing almost no effect on @xmath2 and @xmath3 .
the oscillation period corresponds to a change of the ratio @xmath4 by unity , where @xmath5 is the cyclotron radius , and @xmath6 the grating wave length .
these findings were explained by gerhardts _
et al _ @xcite and winkler _ et al _ @xcite in terms of the quantum - mechanical band structure induced by the modulation ( see also @xcite for explanation of the weak oscillations induced by the grating in @xmath2 and @xmath3 and @xcite for discussion of the temperature dependence of the magnetoresistivity ) .
beenakker @xcite presented a semiclassical theory of the effect .
he showed that the geometric resonance of the cyclotron motion in the grating induces an extra contribution to the drift velocity of the guiding center , with a r.m.s .
amplitude oscillating with period @xmath4 .
these arguments were supported by an exact solution of the boltzmann equation , where isotropic scattering with a relaxation time @xmath7 was assumed . however , while accounting for the above experimental features , the theoretical results @xcite disagree strongly with the experiment as far as the damping of the oscillations with decreasing magnetic field is concerned . specifically , in the regime @xmath8 ( which is valid for the experimental value of the transport time @xmath7 in the whole range of the magnetic fields where the oscillations are observed ) the theory @xcite predicts a linear decrease of the oscillation amplitude , @xmath9 : @xmath10 here @xmath11 is the relative amplitude of the potential modulation @xmath12 , @xmath13 is the mean free path , @xmath14 the cyclotron frequency , @xmath15 the fermi energy , and @xmath16 the fermi velocity .
in contrast with this result , the experimental damping is much stronger , so that only few oscillations ( @xmath17 in @xcite ) are observed .
the problem of the damping of the weiss oscillations was addressed in a number of publications . in ref.@xcite
an empirical observation was made that inclusion of the factor @xmath18 in eq.([e1 ] ) with @xmath19 being the total relaxation rate ( as opposed to the momentum relaxation rate @xmath20 ) yields a reasonable description of the experimentally observed damping .
the authors of ref.@xcite argued quite oppositely that only the transport time should enter the expression for the damping factor , and proposed an empirical form @xmath21 $ ] , with a fitting parameter @xmath22 .
monte carlo simulations @xcite , as well as numerical solution of the boltzmann equation @xcite , showed that allowing for anisotropic scattering leads to a stronger damping of the oscillations .
however , despite considerable interest in this problem , no quantitative theory of the weiss oscillations for the experimentally relevant situation , when small angle scattering is dominant , has been available so far .
the purpose of the present paper is to fill in this gap .
we begin by reminding the reader briefly about the boltzmann equation approach to the problem @xcite .
one starts from the kinetic equation for the distribution function @xmath23 of electrons on the fermi surface , @xmath24 where @xmath25 is the direction of the electron velocity , @xmath26^{1/2}$ ] is the magnitude of the velocity , @xmath27 is the electron charge , and @xmath28 is the collision integral ( to be discussed below ) . in the absence of the grating , the solution is given by @xmath29 where @xmath30 is the drude conductivity tensor , @xmath31 @xmath32 , and @xmath33 is the density of states at the fermi level . using the ansatz @xmath34f^{(0)}+f^{(1)}$ ]
, one may transform eq.([e4 ] ) into the form @xmath35 this implies that the grating induced correction to the conductivity tensor , @xmath36 , is given by @xmath37 where @xmath38 is the solution to the equation @xmath39 and the angular brackets @xmath40 denote the averaging over both the velocity direction and a period of the modulation , @xmath41 calculating now the resistivity tensor @xmath42 , one finds that @xmath43 , where @xmath44^{-1}$ ] and the only non - zero component of @xmath45 is @xmath46 given by @xmath47 for a weak modulation , @xmath48 , the corrections @xmath45 and @xmath49 are proportional to @xmath50 .
therefore , to calculate the average @xmath51 in the leading order , one would have to solve the boltzmann equation ( [ e7 ] ) up to the second order in @xmath11 .
however , it was observed by beenakker @xcite that the following identity holds : @xmath52 since @xmath53 , it is now sufficient to calculate @xmath38 to first order in @xmath11 , in order to find @xmath49 or @xmath45 to the order @xmath50 .
we note that , while only presented for the isotropic scattering case in ref.@xcite , the identity ( [ e10 ] ) holds for a general collision integral @xmath28 as well , since its derivation uses only the fact that @xmath54 , @xmath55 , and @xmath56 .
to proceed further , one has to solve the kinetic equation ( [ e7 ] ) .
this requires first specifying the collision integral .
its general form reads : @xmath57\ , \label{e11}\ ] ] where @xmath58 is the scattering cross - section .
the collision integral can be most conveniently characterized by its eigenvalues @xmath59 note that @xmath60 reflects the particle number conservation ; @xmath61 is the transport relaxation rate , while @xmath62 is the total relaxation rate .
expanding the solution of eq.([e7 ] ) in the angular harmonics , @xmath63 , we get to the leading order in @xmath11 the following coupled system of equations @xcite : @xmath64 according to eqs.([e9 ] ) , ( [ e10 ] ) , the correction to the resistivity tensor is determined by @xmath65 , since @xmath66 in turn , @xmath65 is found from the system ( [ e13 ] ) as @xmath67 where @xmath68 is determined by the set of ( homogeneous ) equations ( [ e13 ] ) with @xmath69 , and @xmath70 . for a generic set of eigenvalues @xmath71 , the solution @xmath72 can be presented as a continuous fraction , @xmath73 the result can be expressed in closed form in the case when @xmath71 depends linearly on @xmath74 , @xmath75 $ ] with an arbitrary @xmath76 , yielding @xcite @xmath77 where @xmath78 .
the two most important particular cases are @xmath79 , corresponding to a white - noise random potential ( isotropic scattering ) , and @xmath80 , describing the scattering by a white - noise random magnetic field with the scattering rate @xmath81 . in the next section
we investigate @xmath46 in these two situations , exploiting the existence of the exact analytical solution ( [ e16 ] ) .
this will give us a first example of how the small angle scattering ( dominant in the random magnetic field case , when @xmath82 ) modifies the magnetic field dependence of the amplitude of the weiss oscillations .
using eqs . ( [ e13a ] ) , ( [ e14 ] ) , and ( [ e16 ] ) , we find for arbitrary values of the parameter @xmath76 @xmath83 for the particular case of isotropic scattering , @xmath79 , the following identity can be used to simplify eq.([e18 ] ) : @xmath84\ .
\label{e19}\ ] ] this allows us to reduce eq.([e18 ] ) to the form @xmath85 substituting ( [ e20 ] ) in ( [ e9 ] ) and assuming that @xmath11 is small enough so that @xmath86 ( this will be valid for all @xmath87 , if @xmath88 ) , we finally get @xmath89 where @xmath90 is the drude resistivity . eq.([e21 ] ) is completely equivalent to beenakker s result ( eq .
( 5 ) of ref.@xcite ) , which is , however , presented in @xcite in the somewhat different form of an infinite series .
the equivalence follows from the bessel function identity @xcite @xmath91 we will assume that @xmath92 , which is a necessary condition for the existence of oscillations . at @xmath93 ( i.e. in the oscillation region ) ,
the asymptotic behavior of @xmath94 reads according to eq .
( [ e21 ] ) @xmath95\ , \qquad \pi / s\gg 1 \label{e24}\end{aligned}\ ] ] thus , the amplitude of oscillations depends linearly on magnetic field @xmath87 in the high - field regime @xmath96 , and gets exponentially damped in @xmath97 , @xmath98 , in the lowest fields , @xmath99 .
note that for a typical value of the transport time @xmath7 in the gaas heterostructures , @xmath100 , the condition @xmath101 corresponds to very low fields , @xmath102 .
we turn now to the random magnetic field scattering , where @xmath80 . analyzing eq .
( [ e18 ] ) in this case , we find the following two asymptotic regimes : @xmath103\ , \qquad 4q / s\gg 1 \label{e26}\end{aligned}\ ] ] again eq.([e25 ] ) corresponds to a power - law behavior of the oscillation amplitude in the range of relatively high magnetic fields @xmath87 , while eq.([e26 ] ) describes the exponential damping of the oscillations in low fields .
there is however a number of important differences between the results for the isotropic scattering [ eqs.([e23 ] ) , ( [ e24 ] ) ] and those for the random magnetic field scattering [ eqs.([e25 ] ) , ( [ e26 ] ) ] : * the `` high '' field behavior ( [ e25 ] ) is @xmath104 , rather than @xmath105 as in eq.([e23 ] ) ; * the value of the magnetic field , at which the exponential damping starts , is much larger in the random magnetic field case ( @xmath106 ) , than for the isotropic scattering ( @xmath107 ) ; * the low - field damping is of a gaussian form ( with respect to @xmath97 ) in the random magnetic field case , @xmath108 $ ] , rather than of a simple exponential , @xmath109 , as in eq.([e24 ] ) .
all these features are consequences of the small - angle scattering , dominant in the random magnetic field case . in the next two sections we will consider the weiss oscillations in the long - range correlated random potential , which also leads to small - angle scattering . since in this case the exact solution can not be expressed in a simple analytic form , we will use different methods to find the shape of the weiss oscillations .
the random potential experienced by electrons of a 2deg in the gaas heterostructures is very smooth , since the impurities are separated from the 2deg layer by a relatively large distance ( spacer ) @xmath110 ( @xmath111 being the fermi wave vector ) . assuming that the positions of ionized donors are uncorrelated , while their density is equal to that of the electrons , @xmath112 , and treating the screening in the random phase approximation , one finds for the random potential correlation function @xmath113 in momentum space @xmath114 where @xmath115 is the density of states per spin orientation .
the scattering cross - section in the born approximation is thus given by @xmath116 the total ( @xmath19 ) and the momentum ( @xmath20 ) relaxation rates are defined as @xmath117 so that their ratio is @xmath118 . both @xmath7 and @xmath119
are experimentally measurable : the former is found from the _ dc _ conductivity at @xmath120 , while the latter follows from the dingle analysis of the damping of the shubnikov de haas oscillations . for the structures of the type used in the weiss oscillation experiment @xcite ( @xmath121 , @xmath122 ) , their typical values are @xmath123 , @xmath124 , so that the ratio @xmath125 is large , in agreement with the theory .
note , however , that the experimental values of @xmath7 and @xmath119 are usually somewhat different from the theoretical estimates ( [ e29 ] ) , ( [ e30 ] ) .
this was attributed to correlations between the positions of the scatterers @xcite .
below we will express our results for the damping of the weiss oscillations in terms of @xmath7 and @xmath119 , so that for comparison with experiment one should simply use the experimental values of these quantities . as was shown in @xcite , the physical origin of the oscillations
can be traced to the additional contribution to the drift velocity of the guiding center , @xmath126 where @xmath127 is the @xmath128-coordinate of the guiding center and @xmath129 is the electric field induced by the grating .
averaging @xmath130 over @xmath127 , one finds @xcite : @xmath131 for @xmath132 , the oscillations are determined by the interference of the contributions from the vicinities of the saddle points @xmath133 , @xmath134 ( and vice versa ) , leading to @xmath135 ( in this section we do not keep power - law prefactors , since we calculate the exponential damping factor only ) .
impurity scattering leads to a change @xmath136 of the guiding center position @xmath127 during the time an electron moves along the cyclotron orbit from @xmath137 to @xmath138 .
this introduces in eq.([e33 ] ) a damping factor @xmath139 for isotropic scattering , any scattering event kicks an electron far away from its original cyclotron orbit , so that @xmath140 is determined simply by the probability for an electron to remain unscattered during the half cyclotron period time @xmath141 , yielding @xmath142 , in agreement with eq.([e24 ] ) .
the situation is , however , different when the scattering is of small - angle character , so that even after one or several scattering events an electron remains close to its original cyclotron orbit and contributes to the oscillations .
we first consider the random magnetic field scattering .
since in this case the exact solution has been found ( eq.([e26 ] ) , this will be an additional check for the correctness of our method of calculation in this section . in this case
, the probability for an electron to be scattered by an angle @xmath143 in a small time interval @xmath144 is given by @xcite @xmath145 scattering by the angle @xmath146 changes the position @xmath127 of the guiding center by @xmath147 , so that @xmath148 where @xmath149 ; @xmath150 ; @xmath151 ; @xmath152 .
according to eq.([e35 ] ) , @xmath153 therefore , the damping factor @xmath140 is equal to @xmath154 in precise agreement with the exact solution , eq.([e26 ] ) .
we turn now to the long - range random potential . in this case
, the scattering cross - section ( [ e28 ] ) implies that the distribution function of the scattering angles @xmath146 in a time interval @xmath144 is given by @xmath155 therefore , @xmath156 this leads to the following result for the damping factor @xmath157\right\}\ .
\label{e41}\end{aligned}\ ] ] as is seen from eq.([e41 ] ) , the region of exponentially damped oscillations can be further subdivided into two regimes : @xmath158 at @xmath159 and @xmath160 at @xmath161 . in the former the leading contribution to @xmath140
comes from multiply scattered carriers , while in the latter @xmath140 is governed by those electrons , which remain unscattered within the time @xmath141 .
the exponential damping found in the above is efficient at low enough magnetic fields , @xmath162^{1/3}$ ] . at stronger magnetic fields ,
the @xmath87-dependence of @xmath46 is of a power - law nature .
this will be considered in the next section using a different method .
to find the grating - induced correction to the resistivity at relatively strong fields , we return to the boltzmann equation approach of sec.[s2 ] . for the scattering cross - section ( [ e28 ] )
the eigenvalues @xmath163 of the collision integral are equal to @xmath164 the approximation we will make at this point will be to consider the limit @xmath165 , which reduces eq.([e42 ] ) to @xmath166 .
this approximation is valid for not too small magnetic fields , such that @xmath167 , where the total relaxation rate @xmath119 is irrelevant , and only the transport one , @xmath7 , is important .
this will allow us to find @xmath46 in the high - field regime ( where the @xmath87-dependence of the oscillation amplitude is of the power - law nature ) , which was outside the regime of validity of the consideration in sec.[s4 ] , where only the exponential damping factor was studied .
furthermore , we will also reproduce the result of sec.[s4 ] for the intermediate regime @xmath168 , where the damping is already exponential , but @xmath119 is still irrelevant . within the above approximation , the collision integral @xmath169 is replaced by the differential operator @xmath170 , and eq.([e7 ] ) with @xmath171 takes the form @xmath172 for @xmath8 , @xmath173 , we can solve this equation iteratively considering the @xmath174 term as a perturbation . dropping this term in ( [ e43 ] ) and differentiating with respect to @xmath175
we find @xmath176 substituting this back to eq.([e43 ] ) , we get a first order differential equation , the solution of which reads @xmath177\ .
\label{e45}\end{aligned}\ ] ] here we introduced again the compact notations @xmath78 , @xmath32 .
substituting now the solution ( [ e45 ] ) into eqs .
( [ e9 ] ) , ( [ e14 ] ) , we finally get @xmath178 combining the results of this and the preceding section , we get a complete description of @xmath46 in the full range of magnetic fields @xmath87 , which will be analyzed in sec.[s6 ] .
according to the results of secs .
[ s4 ] , [ s5 ] , we find in general the following three regions of the magnetic field @xmath87 with different behavior of the oscillation amplitude : * high field domain , @xmath179 , with @xmath180 . in this region @xmath181 , and eq.([e46 ] ) yields @xmath182 the amplitude of the oscillations is proportional to @xmath183 in this regime , in contrast to the linear behavior , eq.([e23 ] ) , for the isotropic random potential , and to the @xmath184 behavior , eq.([e25 ] ) in the case of random magnetic field scattering . *
intermediate domain @xmath185 , with @xmath186 .
the oscillations get exponentially damped : @xmath187\ .
\label{e48}\ ] ] * low - field region , @xmath188 .
the exponential damping factor changes its form : @xmath189\ , \label{e49}\ ] ] where @xmath190 .
note that eq.([e49 ] ) differs from the low - field behavior for the isotropic scattering case , eq.([e24 ] ) , only in that @xmath7 is replaced by @xmath119 in the exponential damping factor .
the two regimes ( i ) and ( ii ) are jointly described by @xmath191 with @xmath192 , while both the regimes ( ii ) and ( iii ) are given by @xmath193\right\ } \sin 2q\right]\ .
\label{e51}\end{aligned}\ ] ] finally all the three regimes can be described by a single formula of the form ( [ e50 ] ) with @xmath194 \label{e52}\ ] ] let us also note that the behavior of @xmath46 at @xmath195 ( i.e for magnetic fields stronger than those generating the oscillations ) is the same for all the types of scattering , @xmath196 as is easily found from eqs .
( [ e18 ] ) , ( [ e21 ] ) , ( [ e46 ] ) . in reality , in most of the weiss oscillation
experiments the magnetoresistivity at such strong fields is strongly affected by the shubnikov - de haas oscillations and by the quantum hall effect .
however , the giant quadratic magnetoresistance ( [ e53 ] ) was observed in the experiment @xcite , where the measurements were performed at somewhat higher temperatures .
now we evaluate the boundaries @xmath197 , @xmath198 of the regions with different behavior of the oscillation amplitude in a typical experimental situation .
we find from the parameters of the sample of ref.@xcite ( @xmath199 , @xmath200 , @xmath201 ) the transport time @xmath202 .
furthermore , a typical value of the total relaxation rate for such structures is @xcite @xmath203 ( as we will see below , with this value of @xmath119 our results perfectly describe the experimentally observed damping of the oscillations in low fields ) .
this gives the following crossover values : @xmath204 , @xmath205 .
for a higher mobility sample studied in ref.@xcite ( the weiss oscillation data are presented there in the inset of fig .
1 ) these values are somewhat larger : @xmath206 , @xmath207 .
we conclude that although parametrically @xmath208 is large at @xmath209 , for realistic parameters @xmath198 and @xmath197 are numerically close to each other .
therefore , the regime ( ii ) corresponds in reality to a rather narrow intermediate domain between the regime ( i ) describing the fast ( @xmath210 ) drop of @xmath46 in high fields , and the regime ( iii ) corresponding to the @xmath18 damping of the oscillatory part of @xmath46 in low fields . in fig.1
we compare our theoretical results with experimental data of weiss _ et al _ @xcite .
the sample parameters are @xcite @xmath199 , @xmath200 , @xmath202 , the total relaxation rate is taken to be @xmath211 .
as is seen from the figure , with the modulation strength @xmath212 we get a very good description of the experimentally observed magnetoresistivity @xcite . to illustrate the difference between the long - range potential scattering and the isotropic scattering , we plotted in fig.2 the corresponding results for the modulation - induced resistivity correction at the same value of the transport time @xmath7 .
it is seen that the results for these two different scattering mechanism differ drastically not only in low magnetic fields ( where the small - angle scattering leads to the exponential damping of oscillations ) , but also in high fields ( due to the different power - law dependence of @xmath46 on @xmath87 ) . in particular , already at the first maximum of @xmath46 ,
the value of the correction is smaller by factor of a @xmath213 for the long - range random potential scattering , as compared to the isotropic scattering case , at the same value of the grating strength @xmath11 . for the second maximum
this ratio is already a factor of @xmath214 , etc . as a consequence ,
one generates a considerable error when trying to evaluate the modulation strength by fitting the amplitude of one of the first maxima by the white - noise random potential results .
for example , beenakker @xcite underestimated @xmath11 by factor of 3 @xcite .
in this paper , we have presented a theory of the weiss oscillations of the magnetoresistivity in a 2deg , for the case when the impurity scattering is predominantly of the small - angle nature .
while this is precisely the experimental situation in currently fabricated semiconductor heterostructures , no quantitative theoretical description of this case has been available so far .
previous theories @xcite assumed the scattering to be isotropic .
we have shown that the small - angle scattering strongly affects the shape and the damping of the weiss oscillations .
this was demonstrated both for the scattering by a white - noise random magnetic field ( linearly increasing eigenvalues @xmath215 of the collision integral ) and for long - range random potential scattering ( @xmath216 $ ] ) . in the latter case , which was of primary interest for us , the shape of the magnetoresistivity
is determined by both the momentum and the total relaxation rates ( @xmath20 and @xmath19 , respectively ) , and has the following features . in relatively strong magnetic fields
@xmath87 , the amplitude of the modulation induced correction @xmath46 falls of as @xmath183 with decreasing @xmath87 , i.e. much more rapidly , than in a white - noise random potential , where @xmath217 .
this rapid fall - off is determined by the quadratic increase of the eigenvalues @xmath218 for not too large @xmath219 . in low magnetic fields ,
the oscillatory part of the magnetoresistivity is found to be exponentially damped , much more efficiently than in the case of isotropic scattering with the same @xmath7 .
the low - field damping factor is of the form @xmath220 , familiar from the theory of the shubnikov - de haas oscillations .
let us stress , however , that while the shubnikov - de haas oscillations are of purely quantum nature , the weiss oscillations represent a classical effect .
nevertheless , due to the interplay of the finite modulation wave vector @xmath221 and the cyclotron radius @xmath5 , the resistivity correction @xmath46 is determined by the whole spectrum of eigenvalues @xmath163 of the collision integral ( and , in particular , by @xmath62 ) , rather than by @xmath61 only ( governing conventional transport in the absence of modulation ) .
finally , our results describe nicely the shape of the experimentally observed magnetoresistance and the dependence of the oscillation amplitude on the magnetic field .
we are grateful to d. e. khmelnitskii for attracting our attention to the problem of weiss oscillations and to y. levinson for useful discussions at various stages of this work .
this work was supported by the sfb195 der deutschen forschungsgemeinschaft . also at petersburg nuclear physics institute , 188350 st.petersburg , russia .
d. weiss , k. v. klitzing , k. ploog , and g. weimann , europhys .
* 8 * , 179 ( 1989 ) .
gerhardts , d. weiss , and k. von klitzing , phys .
. lett . * 62 * , 1173 ( 1989 ) .
r. w. winkler , j. p. kotthaus , and k. ploog , phys .
* 62 * , 1177 ( 1989 ) .
p. vasilopoulos and f.m .
peeters , phys .
lett . * 63 * , 2120 ( 1989 ) ; f.m .
peeters and p. vasilopoulos , phys .
b * 46 * , 4667 ( 1992 ) . p. h. beton , p. c. main , m. davison , m. dellow , r. p. taylor , e. s. alves , l. eaves , s. p. beaumont , and c. d. w. wilkinson , phys .
b * 42 * , 9689 ( 1990 ) . c. w. j. beenakker , phys . rev .
lett . * 62 * , 2020 ( 1989 ) .
y. paltiel , u. meirav , d. mahalu , and h. shtrikman , phys .
b * 56 * , 6416 ( 1997 ) . p. boggild , a. boisen , k. birkelund , c. b. sorensen , r. taboryski , and p. e. lindelof , phys .
rev . b*51 * , 7333 ( 1995 ) .
r. menne and r. r. gerhardts , phys .
b * 57 * , 1707 ( 1998 ) .
a. d. mirlin and p. wlfle , phys .
* 78 * , 3717 ( 1997 ) .
a. p. prudnikov , yu .
a. brychkov , o. i. marichev , integrals and series , vol . 2 : special functions ( gordon and breach , 1986 ) , p. 670 .
t. coleridge , phys .
b * 44 * , 3793 ( 1991 ) .
e. buks , m. heiblum , y. levinson , and h. shtrikman , semicond .
* 9 * , 2031 ( 1994 ) .
a. d. mirlin , e. altshuler , and p. wlfle , ann .
physik * 5 * , 281 ( 1996 ) .
geim , r. taboryski , a. kristensen , s.v .
dubonos , and p.e .
lindelof , phys .
b * 46 * , 4324 ( 1992 ) . j.h .
smet , k. von klitzing , d. weiss , and w. wegscheider , phys .
* 80 * , 4538 ( 1998 ) .
note that the positive magnetoresistance in the very low fields ( below @xmath222 ) is related to the existence of open ( channeled ) orbits in this regime , see p. h. beton , e. s. alves , p. c. main , l. eaves , m. w. dellow , m. henini , o. h. hughes , s. p. beaumont , and c. d. w. wilkinson , phys .
b * 42 * , 9229 ( 1990 ) and ref.@xcite .
this effect is beyond the perturbative - in-@xmath11 calculation employed in the present paper . theoretically , the condition for the existence of the channeled orbits is @xmath223 , which defines a very narrow region of magnetic fields for small @xmath11 .
this regime is well separated from the weiss oscillation regime if @xmath224 . for the experiment of ref.@xcite
we find @xmath225 , and the condition is well satisfied , in agreement with experimental data of @xcite presented in fig . 1 .
at high fields , @xmath226 the experimental data show the onset of the shubnikov - de haas oscillations , which are also not included in our ( essentially classical ) theory .
note that the r.m.s .
amplitude @xmath227 of the density modulation used by beenakker @xcite is related to our parameter @xmath11 as @xmath228 . | we calculate the magnetoresistivity of a two - dimensional electron gas in the presence of a periodic potential within classical transport theory , using realistic models of impurity scattering .
the magnetooscillations induced by geometric resonance of the cyclotron orbits in the periodic grating , known as weiss oscillations , are shown to be affected strongly by the small - angle scattering processes dominant in conventional semiconductor heterostructures .
our results are in full agreement with experimental findings .
psfig 2 |
an interesting and sensitive test for models of baryon structure will soon be performed at the recently completed jefferson lab facility when the radiative decay widths of the first two excited states of the @xmath4 hyperon , @xmath5 and @xmath6 , will be determined directly in a quadrupole coincidence measurement experiment @xcite .
there are four such widths to be measured , two for each excited hyperon radiatively decaying into ground states @xmath7 and @xmath8 , and so far none of the predictions of various models of baryons inspired by quantum chromodynamics ( qcd ) agree with each other @xcite .
calculation of these widths is extremely model dependent involving the wave functions of excited hyperons as well as transition operators both of which must be determined in a self consistent manner .
it is therefore a challenge for _ any _ model of baryon structure to predict correctly all four widths .
discrepancies among existing model predictions are large enough to be able to distiguish experimentally among various results .
but unfortunately , presently available experimental data on the four radiative decay widths either disagree by an order of magnitude @xcite or have been extracted in a model dependent manner @xcite .
phenomenologically this test is particularly welcome since almost all the models of baryon structure can describe the static properties of the ground state octet and decuplet baryons rather well , and therefore it is very difficult to judge the success of any given model based on its description of the ground states alone .
in addition , it is often the case that a given model which can reproduce the properties of ground state baryons will run into difficulties when extended to describe the excited states .
an example of this difficulty is the removal of the center of mass motion in relativistic quark models where no rigorous method to project out the spurious degrees of freedom exist .
thus a successful description of ground state baryons in any model is a necessary but not a sufficient condition for the model to be realistic , and the upcoming direct precision measurements of the excited hyperon radiative decay widths at jefferson lab present an excellent testing ground for any model of baryons under development .
one such model is the chiral confining model ( ccm ) of baryons @xcite where a baryon is described as a color singlet composite object consisting of interacting valence quarks , mesons and a color and chiral singlet hybrid field @xmath9 .
the model produces absolute quark confinment using a mechanism introduced by nielsen and ptkos @xcite to dynamically generate a confining region , or a `` bag '' , inside which the above constituents coexist to form a baryon .
although the model has not been directly derived from qcd , it is at least consistent with the essential properties of the underlying theory of strong interactions such as chiral invariance and the behaviour in the limit of large number of colors , @xmath10 .
the initial version of the ccm@xcite revealed two major problems .
the model prediction for the product of the nucleon mass and the quark rms radius was too large , while that for @xmath11 mass splitting was too small .
mcgovern , _ et al _ @xcite showed there was an approximate scaling law for the ccm which stated that if either the mass or the size of the nucleon was kept fixed at some particular value all other predictions of the ccm were also more or less fixed .
the problems were resolved @xcite by extending the model to include the t hooft interaction @xcite . by fitting the @xmath12 mass difference to determine the strength of the effective t hooft interaction and using simple estimates to correct for the center of mass motion ,
the prediction for the product of the nucleon mass and the quark rms radius were reduced to an acceptable value in the range of 3.8 to 4.4 , the empirical value being about 3.5 . in this two
part series the ccm is extended to the strangeness sector and is used to examine the properties of low lying negative parity @xmath0 , @xmath1 , @xmath2 and @xmath3 baryons in the mean field approximation .
excited baryons are regarded as rpa excitations of the ground state baryons and includes both particle hole excitations of the valence quarks as well as creation of pseudoscalar mesons .
the first paper considers excited baryons built with valence quarks in @xmath13 , @xmath14 and @xmath15 states and examines the contribution of such states to the masses , wave functions and hadronic form factors using a spin isospin dependent instanton induced interaction @xcite .
baryons obtained with only the valence quarks are refered to as bare baryons .
mesonic corrections to properties of bare excited baryon will be presented in a followinging paper .
the role of gluons in the ccm appears in three distinct ways : * first , the gluon condensate of the vacuum manifests itself through the appearance of the color dielectric function @xcite .
the fluctuation of the gluon condensate appears in the form a @xmath16 hybrid field @xcite .
* second , the instanton , a feature of _ classical _ euclidean qcd , appears through the @xmath17 ( @xmath18 being the number of flavors ) fermion interaction introduced by t hooft @xcite to represent approximately the role of instantons on the quark contribution to the partition function . in the absence of this idea
it would not be possible to include the role of instantons in a mean field theory in the minkowski world .
* third , the role of the exchange of _ quantum _ gluons is described through meson exchanges@xcite .
note that perturbative one gluon exchange ( oge ) interaction can not be included in this model as that would lead to double counting . since oge interaction is widely used in quark models by many authors , the differences between three
quark wave functions obtained with the instanton induced and oge interactions are examined . in order to diagonalize the bare baryon hamiltonian , @xmath19 , it is necessary to remove the spurious components in the excited baryon wave functions corresponding to the translation of the center of mass of the composite baryon . for the ccm , where a baryon is a composite object consisting of relativistic quarks , mesons and the @xmath9 field , seeking an exact solution to this problem is an impractical task .
thus in this work a prescription originally used in the study of excited baryons in the mit bag model by degrand @xcite is used to approximately eliminate the unphysical states .
the method is able to predict the observed number of low lying negative parity @xmath0 and @xmath1 states but is strictly valid only in the non relativistic and su(6 ) spin flavor limits .
a simple test of this prescription is presented in the appendix b in order to estimate the relative amount of spurious component present in the excited hyperon wave function for different strange quark masses . after eliminating the spurious states
@xmath19 is diagonalized by assuming equal masses for up , down and strange quarks to be consistent with the center of mass removal prescription .
corrections from flavor symmetry breaking to the hyperon masses are given to first order by modifying the single particle energies of strange quarks after diagonalization .
the three
quark wave functions for excited baryons are used to calculate meson bare excited baryon vertex functions to lowest order in meson - quark coupling .
when the baryons are on mass shell each of these vertex functions is a product of a coupling constant and a form factor .
bare baryon form factors are extracted from the vertex functions and used to determine the quark contribution to the axial coupling constants of negative parity nucleons .
it is pointed out that these hadronic form factors do not have the correct analytical behaviour in the complex @xmath20plane . in the second paper
these vertex functions are used to determine meson cloud contribution to physical baryon properties in the heavy baryon and one loop approximations .
essential features of the ccm are presented in the following section together with single particle energies and spinors of confined quarks in the ground and relevant excited states .
the model hamiltonian for bare baryons is then defined by using the t hooft interaction as the residual interaction between quarks . in section
iii , this hamiltonian is diagonalized and bare masses and wave functions of negative parity baryons @xmath0 , @xmath1 , @xmath2 and @xmath3 are presented and compared to other model calculations .
these three quark wave functions are used in section iv to calculate meson bare excited baryon vertex functions .
corresponding form factors for negative parity @xmath0 resonances and their axial coupling constants are also presented as examples and problems with the results discussed .
finally , the paper is summarized in section v.
the ccm @xcite is based on the color dielectric theory proposed by nielsen and ptkos @xcite .
these authors introduced two collective variables , @xmath21 , a color singlet , charge
conjugation even object , and @xmath22 , a coarse grained gluon field expressed in terms of averages of link operators for loops contained in a hypercube of side @xmath23 . upon integrating out the qcd gluon fields in favor of these new collective variables nielsen and ptkos obtained the following lagrangian in the form of a derivative expansion , @xmath24 \psi(x ) -\frac{k(x)^4}{4}g^a_{\mu\nu}g^{a\,\mu\nu}+\ldots .
\label{npl1}\ ] ] in terms of the gauge field , @xmath25 , the coarse grained field tensor is @xmath26 from the gluonic term one identifies @xmath27 as the color dielectric function .
nielsen and ptkos conjectured that @xmath28 the ccm adds four new features to the nielsen and ptkos model .
first , the conjecture shown in eq .
( [ kvac1 ] ) , crucial for the ccm , was justified from the lattice gauge point of view by lee _
et al . _
it is customary to introduce the field @xmath29 as @xmath30 whereupon @xmath31 has the dimensions of @xmath23 .
the second feature in ccm is the demonstration , through large @xmath10 analysis @xcite , that @xmath29 is a hybrid and not a glueball field .
this is crucial for proper large @xmath10 behavior of ccm itself .
the third feature in ccm is the conjecture that there exists a suitable gauge in which the coarse grained gluon field @xmath22 may be integrated out in favor of meson fields .
the meson fields were described with the gell
mann levy linear @xmath32model @xcite .
this allowed the introduction , through @xmath33 mev , of an ever present mass like term for the quarks in a perfectly chiral invariant manner .
thus field variables in the ccm are quarks , mesons and the hybrid field @xmath9 .
the vanishing of @xmath21 in the vacuum means that @xmath34 the meson
quark coupling may involve a factor @xmath35 .
the fourth feature of ccm is the fixing of the exponent @xmath36 .
this was found by comparing the @xmath37dependence of the four fermion interaction obtained by integrating out the @xmath38 fields of the nielsen and ptkos lagrangian , eq .
( [ npl1 ] ) , with that obtained by integrating out the mesons in the ccm lagrangian .
it was found that @xmath39 .
finally , canonical quark fields were introduced using the transformation @xmath40 . in terms of the canonical quark fields the effective ccm lagrangian , extended to su(3 )
flavor group is @xmath41 \psi + { \cal l}_\chi + { \cal l}_{\rm meson}. % \label{eq : ccm}\ ] ] where @xmath42 are the pseudoscalar octet mesons with flavor index @xmath43 , @xmath44 , and the ellipses refer to other mesons such as @xmath45 or @xmath46 .
mann matrices @xmath47 are normalized to @xmath48 .
the inclusion of vector mesons requires that ccm lagrangian use the lee and nieh s form for @xmath49 @xcite .
two simple forms for the @xmath9 potential , pure mass and quartic , have been considered .
this paper will use only the pure mass form given by @xmath50 extensive su(2 ) calculations have shown that the meson fields produced by the valence quarks themselves play relatively less important role in the baryon structure .
hence a reasonable approximation is to set the mesons fields at their vacuum values .
thus @xmath51 and all other meson fields are fixed at the value zero .
in addition , the parameter @xmath52 in eq .
( [ eq : ccm ] ) is set to @xmath53 .
the resulting model is called the toy model .
it must be understood that it is used solely to obtain reasonable approximations to the valence spinors and the @xmath9 field .
for example , in order to calculate meson
baryon coupling constants the full ccm lagrangian must be used to obtain the meson source currents .
the toy model lagrangian is @xmath54 the bag formation mechanism depends on two critical ingredients specified by eqs .
( [ chivac ] ) , @xmath55 , and ( [ sigvac ] ) , @xmath56 and incorporated in the toy model .
the euler lagrange equation for the quark field from the toy model is @xmath57 thus a quark can not exist in the region where @xmath58 .
however because of the coupling between quarks and the @xmath9 field , valence quarks change the value of the @xmath9 field away from its vacuum value and make it nonzero .
this region is the dynamically created bag where a quark can exist .
there are four adjustable parameters in @xmath59 , @xmath60 , @xmath61 , @xmath31 and @xmath53 .
the first is left fixed at the su(3 ) flavor limit value of @xmath62 mev .
corrections to hyperon masses from su(3 ) breaking are discussed below .
@xcite showed that the toy model possesses approximate scaling and , as a consequence , the combination @xmath63^\frac15 $ ] , which has the dimensions of @xmath64 , is the only operative parameter , not the individual values of @xmath65 and @xmath53 . in the description of ground state properties of @xmath66 and @xmath67 @xcite these parameters
are fixed to be @xmath61 = 1200 mev , @xmath31 = 0.08 mev@xmath68 and @xmath69 .
the t hooft interaction introduces two additional parameters , the interaction strength @xmath70 and the regulator parameter @xmath71 , which are also fixed in @xcite to be @xmath72 and @xmath73 .
these values of model parameters are used in the present work .
the toy ccm hamiltonian corresponding to @xmath59 of eq .
( [ toylag ] ) is given by @xmath74\right ] , % \label{eq : toy}\ ] ] and the corresponding energy function when the three valence quarks occupy the same spinor state @xmath75 is @xmath76 where @xmath77 is the _ classical _ @xmath9 field energy given by @xmath78 .
% \ ] ] in the mean field treatment of the ccm the ground state problem is solved by making @xmath79 of eq .
( [ etoy ] ) stationary with respect to the quark spinor @xmath75 and the @xmath9 field , subject to the normalization @xmath80 .
let @xmath81 be the @xmath9 field resulting from the mean field treatment of the ground state and @xmath82 the corresponding classical @xmath9 field energy . then from eq .
( [ eq : toy ] ) , the ccm hamiltonian describing non interacting quarks moving in a fixed mean @xmath9 field is @xmath83 @xmath84 describes the foundation of the quark model used in this work .
the lowest energy eigenspinor of this hamiltonian is the one which makes the energy function given by eq .
( [ etoy ] ) stationary .
the numerical value of @xmath82 is found to be about 300 mev @xcite .
being the same for all baryon states it plays no role in determining the spin flavor compositions of excited baryons . in the present quark shell model description of excited baryons ,
the low lying negative parity states are constructed by exciting a single quark from the ground @xmath85 state to either the @xmath86 or @xmath87 state .
adapting the notation @xmath88 , @xmath89 and @xmath90 , quark spinors for @xmath91 , @xmath92 and @xmath46 states are defined as @xmath93 these spinors satisfy the eigenvalue equation @xmath94 where @xmath95 or @xmath46 .
figure 1 shows the upper , @xmath96 , and the lower , @xmath97 , radial functions in the quark spinor for @xmath91 , @xmath92 and @xmath46 states defined in eqs .
( [ eq : swave ] ) to ( [ eq : awave ] ) for @xmath98 mev and @xmath99 mev .
also shown in figure 1a is the radial profile for the @xmath81 field which appears in @xmath84 .
the eigenvalues of @xmath91 , @xmath92 and @xmath46 spinors of @xmath84 as defined in eq .
( [ eq : eigen ] ) for @xmath100 and @xmath101 are presented in table [ h0energies ] .
in addition , table [ stable ] shows the values of the quark scalar charge @xmath102 defined as , @xmath103 with the normalization @xmath104 as expected the higher the current quark mass @xmath60 smaller is the lower components of the quark spinors and higher are the scalar charges .
note , however , that changes in @xmath102 are @xmath105% and larger . in the present work the effect of quark mass difference @xmath106
is treated perturbatively . only the spinors with @xmath107 mev
are used to evaluated matrix elements .
su(3 ) breaking effects on hyperon masses are estimated using the quark scalar charge given in eq .
( [ eq : scalar ] ) as follows .
quark wave functions for excited hyperons are first obtained by diagonalizing the mass matrix including the t hooft interaction .
in general these wave functions are linear combinations of states with the excited quark in either the @xmath92 or the @xmath46 state .
after diagonalization , single particle energies for strange quarks are modified from @xmath108 to @xmath109 where @xmath110 using the scalar charges given in table [ stable ] for @xmath62 mev , the numerical values for @xmath111 , are @xmath112 for example , consider a state where the strange quark is excited to the @xmath46 state while the remaining two non strange quarks are in their ground states .
the sum of single quark energies for this state is modified from @xmath113 mev to @xmath114 mev where @xmath108 s are the single quark energy eigenvalues obtained with @xmath60 = 7.5 mev .
phenomenological consequences of the t hooft interaction in baryon structure have been investigated using various quark models by kochelev @xcite , shuryak and rosner @xcite , oka and takeuchi @xcite , blask _ et.al _
@xcite , kim and banerjee@xcite , klabuar @xcite and most recently by takeuchi @xcite .
the role of the t hooft interaction in the excited baryon mass spectra has previously been examined by blask _
_ @xcite and takeuchi @xcite using variants of the non relativistic quark model . the model used by blask _
_ consists of a linear confining potential simulating a string like confinement and the t hooft interaction which was used as the residual interaction without any oge interactions . using this model they managed to reproduce the qualitative features of meson and baryon mass spectra up to masses of order 2 gev including the @xmath115 splitting .
shuryak and rosner @xcite also used a similar constituent quark model to parametrize the ground state baryon mass spectrum using only the two body effective instanton induced force .
thus the constituent quark model seems to give a reasonable description of the low lying baryon and meson mass spectra with either the oge or the t hooft interaction as the residual interaction between a pair of quarks .
the hamiltonian for the constituent quark model used by oka and takeuchi @xcite and later by takeuchi @xcite includes both the oge and t hooft interactions in the follwoing way , @xmath116 where @xmath37 and @xmath117 are quark kinetic energy and confining potentials , respectively . the parameter @xmath118 is introduced to adjust the strength of the oge potential , @xmath119 , relative to that of the t hooft interaction potential , @xmath120 .
this parameter is in the range of 0.3 to 0.4 when the model is used to fit the @xmath115 mass splitting .
takeuchi used this hamiltonian to calculate the masses of excited nucleons up to the @xmath121 harmonic oscillator level and the nucelon
nucleon force in the quark cluster model @xcite .
it was realized early in the development of the nrqm by isgur and karl @xcite that the low lying negative and positive parity non strange baryon mass spectra can be well reproduced if the spin
orbit force from the oge interaction is somehow suppressed . in calculating the excited nucleon masses takeuchi found that the spin
orbit force from the t hooft interaction mostly cancels the spin orbit force originating in the oge interaction .
however , this cancellation does not occur in the nucleon
nucleon force where the combination of the oge and t hooft interactions produces a strong spin orbit component . in the same work takeuchi also examined the low lying negative parity @xmath0 and @xmath1 mass spectra using the mit bag model with both oge and t hooft interactions and found that the cancellation seem to persist for bag models as well @xcite .
unfortunately , however , both blask _ et al . _ and takeuchi focused only on the masses of the excited states and did not investigate the spin flavor compositions of the baryon wave functions . in this work close attention
is paid to the spin flavor contents of excited baryons since they determine the meson
baryon vertex functions as well as radiative transition amplitudes to the ground states .
kim and banerjee developed a program to incorporate the t hooft interaction into the ccm @xcite and their method is used in this work .
they pointed out that although the strength of the t hooft interaction in qcd can be estimated using shuryak s instanton liquid model of the qcd vacuum @xcite , the result can not be used directly in the ccm where the block spinning procedure of nielsen and ptkos @xcite modifies the effective t hooft interaction lagrangian .
because of this reason the strength parameter @xmath70 was fixed by fitting the @xmath122 mass splitting including both the t hooft and perturbative one pion exchange ( ope ) interactions .
the results for the mass and the size of the nucleon showed remarkable improvement relative to earlier ccm results as mentioned in the introduction .
however , the neutron charge radius turned out to be too small .
et al . _
@xcite had pointed out that non perturbative effects of oge can explain the size of @xmath123 .
similar effects exist in the ccm when the t hooft interaction is used in conjunction with perturbative ope interaction as the source of spin and isospin dependent forces .
the nucleon problem was redone in the ccm in @xcite by mixing all seven possible configurations made out of @xmath124 , @xmath125 and @xmath126 single quark states .
in addition , two configurations of the type @xmath127 and one of the type @xmath128 have also been used .
the calculation yielded a very good fit for @xmath123 .
potential used in this work , the coupling constant @xmath129 was largest for the lowest @xmath0 state and sharply smaller for the remaining ones . ] for @xmath18 flavors the t hooft interaction consists of a sum of one , two and up to @xmath18body terms@xcite . for 3 flavors the interaction
is given by @xmath130 \nonumber \\ & & + bd^{abc } \biggl[\bar{\psi}_{1r}\sigma_{\mu\nu}\lambda^a \psi_{1l}\bar{\psi}_{2r } \sigma_{\mu\nu}\lambda^b\psi_{2l}\bar{\psi}_{3r}\lambda^c\psi_{3l } + { \rm permut . } \biggr ] \nonumber \\ & & + cd^{abc}\bar{\psi}_{1r}\lambda^a\psi_{1l}\bar{\psi}_{2r } \lambda^b\psi_{2l } \bar{\psi}_{3r}\lambda^c\psi_{3l } \nonumber \\ & & + df^{abc}\bar{\psi}_{1r}\sigma_{\mu\nu}\lambda^a\psi_{1l } \bar{\psi}_{2r } \sigma_{\nu\gamma}\lambda^b\psi_{2l}\bar{\psi}_{3r}\sigma_{\gamma\mu } \lambda^c\psi_{3l}. % \label{thooft3 } \end{aligned}\ ] ] where @xmath131 refer to the three flavors and @xmath132 is the right / left handed quark spinor with flavor @xmath133 and current quark mass @xmath134 .
the constants , @xmath43 , @xmath135 , @xmath136 , @xmath46 , @xmath137 , @xmath138 , @xmath139 and @xmath140 depend upon the distribution of instanton size in the instanton liquid vacuum . when applied to the ccm the one body term induces a change in the current quark mass of about 7% . in practice
, this small modification is ignored and kim and banerjee retained only the dominant two
body interaction for the description of the @xmath11 system @xcite . the same apporach was used by kochelev @xcite and takeuchi @xcite , all of whom used bag models to calculate the effects of the t hooft interaction on baryon masses . using fierz transformations the two body t
hooft interaction in the ccm may be written in the su(3 ) limit as @xcite , @xmath141 .
% \label{eq : hooft}\end{aligned}\ ] ] as mentioned earlier the parameter determining the interaction strength @xmath70 is taken to be 1.6 @xmath142 . in the mean field approximation the two
body attractive t hooft interaction is ultraviolet divergent due to the fact that quark fields of different flavors are at the same space time point , and leads to a collapsed baryon . in this work the same regulator as in kim and banerjee @xcite
is used to control this divergence with the same parameter of @xmath71 = 0.25 fm . note
that this two
body interaction acts only on flavor anti
symmetric quark pairs .
being a contact interaction the spatial state must be a @xmath91 state to receive any contribution .
then total spin
space symmetry requires that the quark pairs be in spin anti
symmetric state . in a flavor decuplet
every quark pair is in spin symmetric state and the t hooft interaction makes no contribution .
however , in an octet quark pairs are in spin anti
symmetric state half the time and the t hooft interaction contributes to its mass .
also contributes to flavor singlet states as well .
] this contribution is attractive , thus lowering the mass of the nucleon relative to that of the @xmath67 , helping to explain the observed @xmath11 mass splitting .
it should be noted that the pattern of the mass splitting induced by the t hooft interaction is different from the one obtained by using the traditional oge interaction .
the reason is that while the t hooft interaction does not contribute when the quark pair is in spin symmetric state , oge produces a repulsive force 1/9 in magnitude of the attractive force in the spin anti
symmetric state . for @xmath143 , a three
body interaction among the differently flavored quarks are also possible in addition to the one and two body interactions . however , for baryons such a three body force is absent for the following reason @xcite .
the t hooft interaction effectively describes the _ change _ of the eigenvalue of the zero mode due to the interaction between the quarks and the instanton liquid vacuum . in order to put three quarks in the initial zero mode they must be in a totally anti
symmetric state in flavor and for these quarks to feel the three body contact interaction the spatial wave function of the baryon must be totally symmetric .
thus the spin wave function for the baryon must be totally anti symmetric according to total spin flavor space symmetry .
however this is not possible with spin 1/2 quarks and the three body t hooft interaction does not contribute for hyperons .
baryon @xcite . ]
the hamiltonian used to evaluate the bare baryon masses and three quark wave functions in this work is given by , @xmath144 the matrix elements of @xmath19 are evaluated in the @xmath145 coupled basis constructed by coupling the total spin @xmath146 of individual quarks . , orbital angular momentum @xmath147 and intrinsic spin @xmath148 are used to label quarks while upper case letters are used for baryons . ] for negative parity baryons they are labeled as @xmath149 where @xmath91 is a quark in the ground @xmath85 state and @xmath150 can be either in the @xmath151 or the @xmath152 state . before proceeding to the evaluation of the matrix elements of @xmath19 ,
it is necessary to identify and project out as much as possible the spurious components in the excited baryon wave functions corresponding to center of mass excitations .
the prescription used in this work to do this has been used previously @xcite and is strictly valid only in the non relativistic and su(6 ) limits . in the non relativistic limit the spin
flavor wave functions are constructed in the @xmath153 basis by coupling the relative orbital angular momentum of the three quark system , @xmath23 , to its total intrinsic spin @xmath91 .
the @xmath153 coupled basis , @xmath154 , are labeled by the su(6 ) spin flavor multiplet , the su(3 ) flavor multiplet , total angular momentum @xmath155 and the spin multiplet @xmath156 of the three quark system . the resulting spin
flavor wave functions of low lying negative parity baryons belong either to the 56 or the 70 representation of su(6 ) spin flavor .
56plet wave functions share the same permutational symmetry as the ground state spin
flavor wave functions . in the non relativistic quark model with harmonic oscillator confining potential the negative parity 56plet correspond to the translational mode of the three quark system .
the genuine excitations are represented by permutationally mixed symmetric 70plet wave functions . motivated by this example in the non relativistic limit , the projection prescription assumes that those linear combinations in the @xmath145 basis corresponding to the 56plet @xmath153 coupled wave functions are the translational modes of the baryon and is thus spurious
. therefore , @xmath19 is diagonalized by using only those linear combinations of the @xmath145 basis corresponding to the 70plet in the @xmath153 basis . consequently ,
the spin flavor compositions of negative parity baryons are described by using the @xmath153 coupled basis of the form @xmath157 as shown in the fourth column of table iii . also shown in the table are the spin parity and masses of well established low lying negative parity baryons examined in this and the following paper .
appendix a presents the result of @xmath153 to @xmath145 basis recoupling which is used to express the 56 and 70 components of the @xmath153 basis in terms of the @xmath145 basis .
supurious states are mixed into hyperon wave functions when su(6 ) symmetry is broken by @xmath158 mass difference . the relative amount of spurious component contained in hyperon wave functions for different values of strange quark masses is estimated in appendix b. it is found that the spurious components in hyperon wave functions increases with the strange quark mass indicating that this prescription is inapplicable for descriptions of baryons containing a heavy quark @xcite . before presenting the results for bare masses and three quark wave functions it is instructive to study the matrix elements of @xmath84 and @xmath159 . since the calculation itself is straightforward , but quite lengthy and not very informative , only the results are discussed in detail .
the matrix elements of the hamiltonian in the @xmath153 coupled basis are labeled as @xmath160 henceforth the su(6 ) spin flavor label in the @xmath153 basis is dropped and will always be understood to be the 70plet .
@xmath161 can be either a flavor singlet for the @xmath2 or a decuplet for the @xmath3 states .
the mass matrices for @xmath162 and @xmath163 @xmath0 states are given by the lower right @xmath164 submatrix of eq .
( [ eq : matrix ] ) involving only the flavor octet matrix elements .
table [ melements ] shows the matrix elements of @xmath84 and @xmath159 for negative parity nucleon and hyperon resonances . because the t hooft interaction does not contribute to the @xmath1 states their bare masses are simply given by the sum of single quark energies plus the constant contribution from the classical @xmath9 field energy
. states . ] analytical expressions for the matrix elements of @xmath159 are given in appendix c. the fourth column of the table shows the corresponding matrix elements of the oge interaction evaluated in the mit bag model with vanishing quark masses for comparison .
note that in the su(6 ) limit simple patterns emerge among the matrix elements : * in the hyperon sector the model hamiltonian does not connect flavor singlet or decuplet states with flavor octet states . as a result there are two pure flavor singlet @xmath2 states with @xmath165 and @xmath163 .
correspondingly , there are two @xmath3 states which are purely flavor decuplets with @xmath165 and @xmath163 .
since the t hooft interaction does not act on flavor decuplets , the bare masses of these two @xmath3 states are given by the diagonal matrix elements of @xmath84 of 1497 mev and 1477 mev , respectively .
* these values of decuplet @xmath3 masses also correspond to that of @xmath162 and @xmath163 @xmath1 baryons resulting in two sets degenerate @xmath3 and @xmath1 states . however , there are no degenerate partners in the non strange sector for the purely flavor singlet @xmath2 states since no permutationally anti
quark flavor wave function can be constructed with two flavors . *
the matrix elements of @xmath84 and @xmath159 involving flavor octet states @xmath166 and @xmath167 are the same for the @xmath0 , @xmath2 and @xmath3 baryons sharing the same spin and parity .
thus the bare ccm hamiltonian produces two sets of degenerate @xmath0 , @xmath2 and @xmath3 states with @xmath162 and another two degenerate sets with @xmath168 in the su(6 ) limit .
these mass degeneracies are lifted when single particle energies of strange quarks are corrected for flavor symeetry breaking as discussed in section [ toymodel ] .
however since this correction is applied perturbatively it does not affect the hyperon wave functions . note that the off
diagonal matrix elements of @xmath159 for @xmath165 states are neglegibly small while those for @xmath169 states are quite large inducing large mass splittings between the flavor octet states .
when compared to the matrix elements of oge interaction in the mit bag model one apparent difference can be seen in the decuplet matrix elements involving the two @xmath3 states .
morever , for the flavor singlet matrix elements the t hooft interaction gives a much larger and negative value of @xmath170 mev for the @xmath165 state compared with @xmath171 mev for the @xmath169 matrix element .
this pattern is reversed for the oge interaction where the singlet matrix element of the @xmath169 state of @xmath172 mev is more negative than the @xmath165 matrix element of @xmath173 mev .
bare masses and three quark wave functions of negative parity baryons @xmath0 , @xmath1 , @xmath2 and @xmath3 in the ccm are given in table [ results ] which lists the results before and after including the two body t
hooft interaction in the hamiltonian . with the present set of model parameters
the t hooft interaction affects the masses of baryon resonances more than their spin flavor compositions , the latter being determined mostly by single particle energies of valence quarks .
the fourth column also shows the bare masses of hyperons after being corrected for flavor symmetry breaking . the matrix elements of @xmath84 for @xmath174 @xmath0 , @xmath2 and @xmath3 states are all equal and has the value of 1470 mev .
these baryons are purely flavor octet states . in the present quark shell model the three
quark wave functions for @xmath175 baryons are constructed by exciting one of the ground state quarks from the @xmath91 to the excited @xmath46 state .
thus the contribution from single quark energies to @xmath84 is simply @xmath176 = 1170 mev while the classical @xmath81 field energy contributes the remaining 300 mev .
there is no flavor decuplet @xmath177 state which is interpreted as a spurious excitation according to the center of mass projection prescription as shown in eq .
( [ eq : jjlsa ] ) .
the absence of this @xmath1 state is in agreement with experiment ( see table iii ) .
note that numerically , the matrix elements of @xmath84 and @xmath19 are practically equal implying that the effects of t hooft interaction on @xmath174 states are neglegiblely small .
this is because the spin
isospin dependent t hooft interaction acts on quark pairs where one of them is in the @xmath91 and the other in the @xmath46 state . being a contact interaction , the resulting interaction energy is expected to be small as verifed numerically in this model calculation .
corrections from flavor symmetry breaking lifts the mass degeneracy between the well established @xmath178 and @xmath179 states @xcite .
however , the resulting mass splitting is only 14 mev in contrast to the empirical values of 55 mev .
thus , a further mass splitting of about 40 mev is necessary from meson exchange corrections . for @xmath165 baryons
there are three sets of degenerate eigenstates of @xmath19 .
* purely flavor decuplet @xmath180 and @xmath181 with @xmath182 = 1497 mev . * @xmath183 and @xmath184 with @xmath182 = 1485 mev .
the spin flavor composition of these states is approximately 72 % @xmath185 and 28 % @xmath186 . using the results of @xmath153 to @xmath145 recoupling given in appendix a
, this corresponds to 8% @xmath187 , 14% @xmath188 and 78% @xmath189 in terms of @xmath145 coupled basis .
thus the excited valence quark in these baryons is mostly in the @xmath92 orbit with about 20% probability of being in the @xmath46 orbit . *
@xmath190 and @xmath191 with @xmath182 = 1439 mev . in this set of degenerate states ,
the excited quark is almost always in the @xmath46 state with a vanishingly small @xmath189 component .
recall that the single quark energy for the @xmath46 state ( @xmath192 = 490 mev ) is smaller than that of the @xmath92 state ( @xmath193 = 550 mev ) .
this explains qualitatively why this set of degenerate states has a lower mass than the above set where the excited quark is dominantly in the @xmath92 orbit . finally , the purely flavor singlet @xmath194 with @xmath182 = 1380 mev has no mass degenerate partners .
the values for @xmath182 for the octet and singlet states are all lower than the values obtained by using @xmath195 alone indicating that the t hooft interaction is attractive for the @xmath165 baryons . as in the @xmath165 sector
there are three sets of degenerate eigenstates of @xmath19 for @xmath169 baryons .
* purely flavor decuplet @xmath196 and @xmath197 with @xmath182 = 1477 mev . * @xmath198 and @xmath199 with @xmath79 = 1529 mev and @xmath182 = 1742 mev .
the spin flavor composition of these states is about 86% @xmath200 and 14% @xmath201 which translates into 92% @xmath202 and 8% combination of @xmath203 and @xmath204 states .
thus , the excited quark is dominantly in the @xmath92 orbit . *
@xmath205 and @xmath206 with @xmath79 = 1496 mev and @xmath182 = 1450 mev . in this set the excited quark is about 45% in the @xmath202 state and 55% in the @xmath46 state leading to a lower value of energy eigenvalue than the above degenerate set .
note that for the latter two sets @xmath182 = 1742 and 1450 mev , respectively , implying that the t hooft interaction induces a large mass splitting among the @xmath168 baryons .
this result can be anticipated by observing that the diagonal matrix elements of @xmath159 for the octet @xmath169 states differ by an order of magnitude as well as by sign ( 178 and -12 mev ) as shown in table iv .
finally , the pure flavor singlet @xmath207 state with @xmath182 = 1520 mev has no degenerate partners as in the @xmath162 case .
it is interesting to compare the ccm results for the three
quark wave functions with the results obtained in the mit bag model using the oge interaction .
table [ percent ] presents the masses and relative percentages of spin flavor contents of low
lying @xmath165 and @xmath163 @xmath2 states in the ccm and in the mit bag model using respectively the t hooft and oge matrix elements of table iv .
= 145 mev , @xmath208 = 0.25 , @xmath209 = 1.5 and @xmath210 = 0 mev . ] the same center of mass projection prescription is used for both calculations .
the table also presents the results obtained in the nrqm @xcite which includes explicit flavor symmetry breaking . these negative parity @xmath2 states are chosen since
radiative decay widths of the lightest @xmath165 @xmath6 and @xmath163 @xmath5 hyperons are examined in the following paper .
the results for the wave functions of @xmath211 states are qualitatively similar for both the ccm and the mit bag model , with the lightest @xmath165 hyperon being a flavor singlet state .
both models predict similar spin
flavor compositions for @xmath212 and @xmath213 .
a similar result is also given in the nrqm with the lightest hyperon state having a 81% flavor
singlet component .
it should be noted that in the bound state approach to the skyrme model no bound state solution corresponding to this state exists with physically acceptable values of model parameters @xcite .
thus these models suggest that the well - established @xmath6 is a genuine three quark excited state with a dominant spin doublet flavor singlet component .
differences in model predictions for hyperon wave functions can clearly be seen in the @xmath169 sector where the only agreement among the three models is the heaviest @xmath214 , which is mostly a @xmath200 state . in the ccm
@xmath207 is a pure flavor singlet state while it is a spin doublet flavor
octet state in the mit bag model .
the nrqm finds @xmath207 to be a mixture of 56% @xmath215 , 34% @xmath216 and 10% @xmath217 .
the lightest @xmath169 hyperon is found to be dominantly a @xmath201 state in the ccm , while it is a pure flavor singlet state in the mit bag model in the limit of vanishing quark masses .
however , when flavor symmetry is broken in the mit bag model a non
negligible octet component is mixed into the @xmath218 wave function @xcite . in the nrqm
this hyperon resonance is predominantly a flavor singlet state . in previous works on the @xmath219 atom and @xmath220 scattering in the cloudy bag model @xcite ,
it has been assumed that the @xmath5 is a pure flavor singlet .
the present example indicates that the cloudy bag model hamiltonian must be properly diagonalized before neglecting the flavor octet component of the @xmath5 wave function .
in this section three quark wave functions obtained by diagonalizing @xmath19 are used to calculate meson
excited bare baryon vertex functions to lowest order in meson quark coupling .
each of these vertex functions is a product of a coupling constant and a form factor when the baryons are on their mass shell .
bare baryon form factors for negative parity resonances are extracted by applying a method originally introduced for ground states @xcite . it will be shown that model predictions for the form factors have problems with their analytical structures and , consequently , only their values at zero three momentum transfer are reliable . as examples , axial coupling constants for excited nucleons
are determined using the goldberger trieman relation .
the equation of motion for the pseudoscalar meson fields @xmath221 corresponding to the su(3 ) ccm lagrangian given in eq .
( [ eq : ccm ] ) is , @xmath222 in the present work , where the excited quark is either in the @xmath92 or the @xmath46 state , the meson quark source current is given by @xmath223 % \label{eq : qmsource } % \end{aligned}\ ] ] where the quark spinors for states @xmath91 , @xmath92 and @xmath46 are given in eqs .
( [ eq : swave ] ) to ( [ eq : awave ] ) , respectively .
to lowest order in meson quark coupling , the meson excited bare baryon vertex function is given by the matrix element of the fourier transform of eq .
( [ eq : qmsource ] ) in the static limit , @xmath224 where @xmath225 is a three quark state corresponding to baryon @xmath138 . on the right side of eq .
( [ eq : ffactor ] ) , summation over the quark index @xmath226 is shown explicitly for the single quark operator @xmath227 .
in general , the initial and final baryons , @xmath138 and @xmath228 , may belong to different spin parity states .
however , in this work only those vertex functions involving initial and final states with the same @xmath229 value are considered .
this means that when using @xmath230 to calculate one meson loop contributions to baryon observables , the intermediate baryon state is restricted to share the same spin parity as the initial and final baryons .
the fourier transformed meson quark source current operator may be written as , @xmath231 here @xmath232 is the @xmath147th multipole single quark operator acting on the spin flavor space of the @xmath226th quark .
the subscripts @xmath233 identifies various contributions from the meson quark source current , eq ( [ eq : qmsource ] ) , in an obvious manner .
these operators and their corresponding functions @xmath234 are defined in appendix d. if the initial and final baryon states are restricted to share the same @xmath229 value , the operators shown in eq .
( [ eq : vfunction ] ) are the only posssible multipole contributions in the present model with a quark shell model space consisting only of @xmath91 , @xmath92 and @xmath46 states .
meson baryon form factors are determined by defining a meson baryon interaction at the _ hadronic _ level and equating the fourier transformed meson quark and meson baryon source currents@xcite .
for spin 1/2 baryons assume a meson baryon interaction of the form @xmath235 where @xmath236 is a single baryon field with total spin @xmath237 and parity @xmath92 .
the corresponding meson
baryon source current @xmath238 is @xmath239 for example , the @xmath240 form factor @xmath241 for ground state octet baryons @xmath138 and @xmath228 with @xmath242 is determined as follows : @xmath243 the superscript in @xmath244 indicates that it is a dipole form factor .
note that by definition @xmath244 is divided by twice the nucleon mass @xmath245 regardless of the types of initial and final baryons @xmath138 and @xmath228 .
this convention , which differs from the one given in @xcite for the @xmath246 form factor , will be extended to extract excited baryon form factors . for excited baryons with @xmath247 , and @xmath248 , it is found that only dipole form factors are numerically significant .
they are given by @xmath249 figure 2 shows the normalized form factors @xmath250 for negative parity nucleons under consideration .
it can be seen from the figure that all the @xmath251 form factors are qualitatively similar to lowest order in meson quark coupling . in particular , the three momentum cut off of each of the form factors is around @xmath252 = 1200 mev . in the hyperon sector , a similar @xmath252 dependence can also be seen in the normalized dipole form factors involving the @xmath165 flavor singlet and @xmath0 states , @xmath253 , as shown in figure 3a
. however , the corresponding form factors involving the @xmath169 states , shown in figure 3b , are quite different . in this case
the @xmath254 form factors change sign for the first time around @xmath255 mev and encounter a second node at about @xmath252 = 1200 mev . as the figure indicates
these two form factors are proportional to each other since the fourier trasformed meson
quark source current only connects the flavor singlet @xmath214 to the spin doublet flavor
octet ( @xmath201 ) component of the excited nucleon @xmath256 .
having extracted the meson
baryon form factors from meson quark vertex functions it is tempting to calculate the meson baryon coupling constants
. however it should be emphasized that the model form factors presented in this section do not have the correct behaviour when analytically continued to the complex @xmath252plane .
the analytically continued ccm form factors are entire functions on the _ whole complex plane _ and not the cut plane with a branch point at the normal threshold of @xmath257 @xcite .
in particular , there are no dynamical singularities in the ccm form factors which arise from summing over intermediate states in the dispersion representation .
thus the @xmath252 dependencies of these form factors are questionable and the only reliable quatity to extract from these form factors are their values at @xmath258 , such as @xmath259 to be discussed below . in order to evaluate the meson baryon coupling constants the value of the form factor at @xmath260 is required . for pion
baryon coupling constants it might be reasonable to taylor expand the form factor around @xmath258 since the pion mass is small , @xmath261 however this expansion would certainly not be justified for kaons and other heavier mesons .
it should be remarked that this problem is shared by other relativistic quark models with confinement and meson quark coupling , such as the chiral bag model @xcite .
when @xmath262 and @xmath263 , the value of the ccm hadronic form factor at @xmath258 is given by @xmath264 for excited nucleons the values of dipole form factors @xmath265 at zero three momentum transfers are found to be @xmath266 @xmath267 @xmath268 the reason for large values of @xmath269 and @xmath270 are due to the spin matrix elements of ( 1/17 ) and ( 1/5 ) appearing on the left sides of eqs .
( [ eq : dff52 ] ) and ( [ eq : dff32 ] ) , respectively . if the goldberger
trieman relation @xmath271 is applied to the ground state form factor @xmath272 , the resulting axial coupling constant is @xmath273 which is close to the empirical value of @xmath274 .
note that in the derivation of eq .
( [ eq : gtr ] ) , the free dirac equation is used to introduce the nucleon mass @xmath275 . when this relation is extended to excited states , the nucleon mass @xmath275 in eq .
( [ eq : gtr ] ) should be replaced with the excited nucleon masses @xmath276 , _ i.e. _ @xmath277 thus to lowest order in meson quark coupling , the axial coupling constants for the excited nucleons are @xmath278 @xmath279 @xmath280
in this paper a toy model version of ccm is used together with the classical t hooft interaction to calculate the quark contribution to the masses and wave functions of low lying negative parity baryons @xmath281 and @xmath3 .
a baryon in the ccm is a composite object consisting of interacting quarks , mesons and a chiral singlet @xmath9 field confined inside a dynamically generated region .
the meson fields in this model are effective degrees of freedom describing the exchange of a tower of gluons between quarks . in the mean field approximation which is used in this work ,
meson exchanges between quarks are treated perturbatively at the quantum level .
hence in the mean field theory there is a bare " baryon with its wave function determined by the valence quark dynamics which couples to mesons yielding a physical baryon state .
mesonic exchange corrections to bare baryon properties will be examined in a following paper .
masses and three quark wave functions of bare excited baryon states are calculated by diagonalizing @xmath19 defined in eq .
( [ eq : bare ] ) using @xmath210 = 7.5 mev .
a prescription , strictly valid only in the su(6 ) and non relativistic limits , is used to approximately eliminate the spurious degrees of freedom associated with the center of mass motion of the baryon .
a simple test showed that this prescription becomes less reliable with increasing strange quark mass .
it is found that the t hooft interaction affects the bare masses of negative parity baryons more than their wave functions which are determined mostly by single quark energies .
the ccm results for the bare masses indicate that mesonic corrections to excited baryon masses must play an important role in order to reproduce the observed mass splittings .
it should be remarked that corrections to bare masses and wave functions due to flavor symmetry breaking should be investigated . using the bare baryon wave functions meson
excited baryon vertex functions have been calculated to lowest order in meson quark coupling .
baryon form factors for negative parity baryons were extracted from vertex functions by extending the conventional definition used for ground states .
however , it is emphasized that because the model form factors do not have the correct behaviour when analytically continued to the complex plane , their momentum dependencies are questionable and only their values at zero momentum transfer can be trusted . by applying the goldberger trieman relationship the quark contribution to weak axial charges for negative parity nucleons have been determined . in a following paper
the vertex functions are used to regulated mesonic corrections to bare baryon properties in the one loop and heavy baryon approximations .
the second paper also examines the four radiative decays of the two lightest @xmath2 hyperons which will be measured in an upcoming jefferson lab experiment @xcite , and will serve as an excellent testing ground for _ all _ models of baryon structure .
this work is supported by the us department of energy under grant no .
de fg0293er40762 .
as discussed in section [ diagonal ] , the center of mass projection prescription used in this work assumes that the spurious components in negative parity baryon wave functions are described by those linear combinations in the @xmath145 basis corresponding to the 56plet in the @xmath153 basis @xcite .
the remaining linear combinations corresponding to the 70plet are used to diagonalize the model hamiltonian .
this appendix presents the results of @xmath153 to @xmath145 recoupling using the notations for @xmath153 and @xmath145 coupled basis defined in section [ diagonal ] .
because the projection prescription assumes su(6 ) symmetry , the matrix elements of @xmath19 , eq .
( [ eq : bare ] ) , are evaluated using @xmath283 .
corrections from flavor symmetry breaking ( @xmath284 ) to neative parity hyperon masses are given to first order as discussed in section [ toymodel ] .
however , the breaking of flavor symmetry introduces spurious components in excited hyperon wave functions .
this appendix presents an estimate of the relative amount of spurious component introduced in excited hyperon wave functions for different values of strange quark mass using the well established @xmath174 @xmath285 resonance as an example .
according to the projection prescription the wave function of this hyperon resonance , @xmath286 , is a pure flavor octet state identified with the 70plet in the @xmath153 coupled basis as shown in eq .
( [ eq : jjlsb ] ) . the flavor decuplet state given in eq .
( [ eq : jjlsa ] ) is assumed to be a spurious state , @xmath287 , corresponding to the translational mode of the baryon . in the @xmath145 coupled basis
these two orthogonal states are given by @xmath288 where , @xmath289 the spin states are defined as @xmath290 and @xmath291 .
note that in eq .
( [ eq : sigma ] ) the strange quark is in the excited @xmath46 state with a probability of two thirds , while this number is lowered to one third for the spurious @xmath287 state shown in eq .
( [ eq : spurious ] ) .
the su(6 ) spin flavor wave function for the ground state hyperon @xmath292 , denoted as @xmath293 , is @xmath294 where , @xmath295 with @xmath296 .
let @xmath297 be some operator exciting the internal degrees of freedom of the three quark system so that @xmath298 is proportional to @xmath299 .
then the following relations hold between the ground and excited state wave functions in the exact su(6 ) limit @xmath300 the deviation from the vanishing value of eq .
( [ eq : testb ] ) is used to estimate the amount of spurious component introduced due to flavor symmetry breaking as follows .
the excitation operator @xmath297 may be constructed by considering two single particle operators @xmath301 and @xmath302 defined as @xmath303 where @xmath304 and @xmath305 are the quark spinors for states @xmath91 and @xmath46 shown in eqs .
( [ eq : swave ] ) and ( [ eq : awave ] ) , respectively .
let @xmath306 and @xmath307 .
the operator @xmath308 acts on the spin
flavor space of a quark in the following way , @xmath309 and @xmath310 in eq .
( [ eq : lsb ] ) @xmath311 , @xmath312 , @xmath313 and @xmath314 are defined as @xmath315 the value of the up and down quark masses used to obtain the radial functions @xmath316 and @xmath317 for @xmath311 and @xmath312 is 7.5 mev , while the strange quark mass has been left as a parameter for @xmath313 and @xmath314 .
thus the operator @xmath318 raises the total angular momentum @xmath226 of a quark of given flavor from @xmath319 to @xmath320 while annihilating a single quark state with @xmath226 = @xmath321 . using @xmath318 the excitation operator @xmath297 is constructed as @xmath322 where @xmath323 is the quark index . when @xmath297 acts on the ground state wave function @xmath324 it excites quarks with @xmath226 = @xmath319 to @xmath320 thus creating either the @xmath174 @xmath325 or @xmath287 state where the excited quark must be in the @xmath46 state . taking @xmath158 mass difference into account , the matrix elements corresponding to eqs .
( [ eq : testa ] ) and ( [ eq : testb ] ) are found to be @xmath326 when @xmath283 , @xmath327 and @xmath328 .
consequently , eqs .
( [ eq : testa ] ) and ( [ eq : testb ] ) hold exactly in the su(6 ) limit and @xmath325 is interpreted as the genuinly excited @xmath329 state .
however when @xmath284 , @xmath330 and eq .
( [ eq : testbb ] ) no longer vanishes indicating the presence of spurious components in the excited @xmath298 state . the difference @xmath331 , which may be calculated in any relativistic quark model , is used as an estimate of theoretical uncertainty arsing from @xmath158 mass difference .
the values for @xmath332 are found to be 0.151 and 0.206 for @xmath333 = 200 and 300 mev , respectively , indicating that the projection prescription becomes less applicable with increasing strange quark mass .
similar conclusion is presented in @xcite where it is remarked that mit bag model and its chirally extended versions are not suitable for describing a baryon with one heavy quark .
it should be remarked that the projection prescription assumes that the angular momentum of a quark spinor is given by its upper component and this assumption contributes to the model uncertainty even in the su(6 ) limit .
this appendix presents the matrix elements of @xmath159 , eq .
( [ eq : hooft ] ) , used in calculating the bare masses of low lying negative parity baryons .
the matrix elements have been evaluated assuming @xmath283 and are to be multiplied by an universal factor of @xmath334 with @xmath335 @xcite .
as discussed in section [ hooft ] , the t hooft interaction needs to be regularized .
the regulator used in this work is @xmath336 which is the same as in kim and banerjee @xcite with the same parameter @xmath71 = 0.25 fm .
the notation for the @xmath145 coupled basis are defined in section [ diagonal ] .
recall that since the two
hooft interaction operates on flavor anti
symmetric quark pairs , the @xmath1 states and flavor decuplet components of @xmath3 are unaffected by this interaction . also because of assumed flavor symmetry , the matrix elements involving the octet components of @xmath0 , @xmath2 and @xmath3 are equal for given spin and parity .
@xmath337 @xmath338 @xmath339 @xmath340 @xmath341 @xmath342 the functions f1 to f15 are given by @xmath343 \\
% { \rm f3 } & = & \frac{2}{3}\int\!\ !
dr r\:r^2 \biggl [ \biggl(g_0 ^ 2-f_0 ^ 2\biggr)\biggl(g_3 ^ 2-f_3 ^ 2\biggr ) + \frac{4}{5}\biggl ( g_0f_0g_3f_3\biggr ) + \frac{4}{5}\biggl(g_0f_3+g_3f_0\biggr)^2 \biggr ] \\ % { \rm f4 } & = & \frac{2}{3}\int\!\ ! dr r\:r^2 \biggl [ \biggl(g_0 ^ 2-f_0 ^ 2\biggr)\biggl(g_3 ^ 2-f_3 ^ 2\biggr ) - \frac{4}{15}\biggl(g_0f_0g_3f_3\biggr ) \nonumber \\ & & \hspace{5 cm } - \frac{2}{3}\biggl(g_0g_3-f_0f_3\biggr)^2 + \frac{1}{5}\biggl(g_0f_3+g_3f_0\biggr)^2 \biggr ] \\ % { \rm f5 } & = & \int\!\ ! dr r\:r^2
\biggl(g_0 ^ 2+f_0 ^ 2\biggr)^2 \\ % { \rm f6 } & = & 2\int\!\ ! dr r\:r^2 \biggl
[ \biggl(g_0 ^ 2-f_0 ^ 2\biggr)\biggl(g_2 ^ 2-f_2 ^ 2\biggr ) - \frac{4}{3}\biggl(g_0f_0g_2f_2\biggr ) - \frac{1}{3}\biggl(g_0g_2-f_0f_2\biggr)^2 \nonumber \\ & & \hspace{5 cm } + \biggl(f_0g_2+g_0f_2\biggr)^2 \biggr ] \\ % { \rm f7 } & = & \frac{2\sqrt{2}}{3}\int\!\ ! dr r\:r^2 \biggl(f_0f_3+g_0g_3\biggr)\biggl(g_0g_2+f_0f_2\biggr)\\ % { \rm f8 } & = & \sqrt{\frac{8}{27}}\int\!\ ! dr r\:r^2 \biggl [ -2\biggl(g_0f_0\biggr)\biggl(g_2f_3+g_3f_2\biggr ) \nonumber \\ & & \hspace{5 cm } + \biggl(g_0g_3-f_0f_3\biggr)\biggl(g_0g_2-f_0f_2\biggr ) \biggr ] \\ % { \rm f9 } & = & \frac{2}{3}\int\!\ ! dr r\:r^2 \biggl [ \biggl(g_0 ^ 2-f_0 ^ 2\biggr)\biggl(g_3 ^ 2-f_3 ^ 2\biggr ) + \frac{4}{15}\biggl(g_0g_3f_0f_3\biggr ) \nonumber \\ & & \hspace{5 cm } -\frac{1}{3}\biggl(g_0g_3-f_0f_3\biggr)^2 + \frac{3}{5}(g_0f_3+g_3f_0\biggr)^2 \biggr ] \\ % { \rm f10 } & = & \frac{2}{3 } \int\!\ ! dr r\:r^2 \biggl [ -\frac{16}{15}\biggl(g_0f_0g_3f_3\biggr ) - \frac{2}{3}\biggl(g_0g_3-f_0f_3\biggr)^2 + \frac{2}{5}\biggl(g_0f_3+g_3f_0\biggr)^2 \biggr ] \\ % { \rm f11 } & = & 2 \int\!\ ! dr rr^2
\biggl [ \biggl(g_0 ^ 2-f_0 ^ 2\biggr)\biggl(g_2 ^ 2-f_2 ^ 2\biggr ) + \frac{4}{3}\biggl(g_0f_0g_2f_2\biggr ) - \frac{2}{3}\biggl(g_0g_2-f_0f_2\biggr)^2 \biggr ] \\ % { \rm f12 } & = & 2\int\!\ ! dr rr^2 \biggl [ -\frac{8}{3}\biggl(g_0f_0g_2f_2\biggr ) + \frac{1}{3}\biggl(g_0g_2-f_0f_2\biggr)^2 + \biggl(f_0g_2+g_0f_2\biggr)^2 \biggr ] \\ % { \rm f13 } & = & 2\int\!\ ! dr r\:r^2 \biggl [ \frac{1}{3}\biggl(g_0g_2+f_0f_2\biggr)^2-\biggl(g_0f_2-g_2f_0\biggr)^2 \biggr ] \\ % { \rm f14 } & = & \sqrt{\frac{8}{27}}\int\!\ ! dr r\:r^2 \biggl [ 2\biggl(g_0f_0\biggr)\biggl(g_2f_3+g_3f_2\biggr ) \nonumber \\ & & \hspace{5 cm } + \biggl(g_0g_3-f_0f_3\biggr)\biggl(g_0g_2-f_0f_2\biggr ) \biggr ] \\ % { \rm f15 } & = & -\sqrt{\frac{32}{27}}\int\!\ ! dr r\:r^2 \biggl(g_0f_0\biggr)\biggl(g_2f_3+g_3f_2\biggr ) % \end{aligned}\ ] ] the radial functions @xmath344 and @xmath345 are defined in eqs .
( [ eq : swave ] ) to ( [ eq : awave ] ) and are shown in figure 1 for @xmath60 = 7.5 and 300 mev .
definitions of operators @xmath346 and corresponding functions @xmath347 appearing in the fourier transformed meson quark source function , eq .
( [ eq : vfunction ] ) , are presented in this appendix .
the functions @xmath347 in eq .
( [ eq : vfunction ] ) are defined as @xmath348 where @xmath349 is the usual spherical bessel function of order @xmath147 .
the corresponding single quark operators @xmath346 are @xmath350 } \hat{q } \cdot \vec{\sigma } \sigma_b^{[1/2,3/2 ] } \left ( \chi_m^{3/2 } \right ) \\ % { \cal o}^{a , l=1}_{ap}(\hat{q } ) & = & \lambda^a \left ( \chi_m^{3/2 } \right)^\dagger \hat{q } \cdot \vec{\sigma}^{[3/2,1/2 ] } \left ( \chi_m^{1/2 } \right)\\ % { \cal o}^{a , l=3}_{aa}(\hat{q } ) & = & \lambda^a \left ( \chi_m^{3/2 } \right)^\dagger \bigg [ \hat{q } \cdot \vec{\sigma}^{[3/2,1/2 ] } \hat{q } \cdot \vec{\sigma } \hat{q } \cdot \vec{\sigma}^{[1/2,3/2 ] } - \frac{1}{5 } \sigma_b^{[3/2,1/2 ] } \hat{q } \cdot \vec{\sigma } \sigma_b^{[1/2,3/2 ] } \bigg ] \left ( \chi_m^{3/2 } \right ) % \end{aligned}\ ] ] in the above definitions @xmath351 is a @xmath352 component spinor analogous to the two component pauli spinors for @xmath226=1/2 case .
the operators @xmath353 with @xmath354 are the usual pauli matrices , while @xmath355}$ ] are @xmath356 multipole transition matrices given by @xcite , @xmath357 } & = & \frac{1}{\sqrt{6 } } \left ( \begin{array}{cccc } -\sqrt{3 } & 0 & 1 & 0 \\ 0 & -1 & 0 & \sqrt{3 } \end{array } \right ) , \\ % \sigma_2^{[1/2,3/2 ] } & = & \frac{-i}{\sqrt{6 } } \left ( \begin{array}{cccc } \sqrt{3 } & 0 & 1 & 0 \\ 0 & 1 & 0 & \sqrt{3 } \end{array } \right ) , \\ % \sigma_3^{[1/2,3/2 ] } & = & \sqrt{\frac{2}{3 } } \left ( \begin{array}{cccc } 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \end{array } \right ) , % \label{eq : mtm}\end{aligned}\ ] ] with the property , @xmath358}_a \right)^\dagger = \sigma^{[3/2,1/2]}_a .
% \ ] ] cebaf expt89024 , radiative decays of low - lying hyperons , clas collaboration , g.s . mutchler spokesman .
darewych , m. horbatsch and r. koniuk , phys .
d * 28 * , 1125 .
e. kaxiras , e.j .
moniz and m. soyeur , phys .
d * 32 * , 1125 ( 1985 ) .
y. umino and f. myhrer , nucl . phys . *
a529 * , 713 ( 1991 ) ; * a554 * , 593 ( 1993 ) . c.l .
schat , n.n .
scoccola and c. gobbi , nucl . phys . *
a585 * , 627 ( 1995 ) .
mast , m. alston
garnjost , r.o .
bangerter , a. barbaro galtieri , l.k .
gershwn , f.t .
solmitz and r.d .
tripp , phys .
* 21 * , 1715 ( 1968 ) .
r. bertini _
et.al._ , contribution no .
m18 to the 10th particles and nuclei international conference ( panic ) , heidelberg ( 1984 ) ( unpublished ) .
h. burkhardt and j. lowe , phys .
c * 44 * , 607 ( 1991 ) . m.k .
banerjee , prog .
nucl . phys . * 33 * , 77 ( 1993 ) and references therein .
m. k. banerjee and j. a. tjon , phys .
c * 56 * , 497 ( 1997 ) .
nielsen and a. ptkos , nucl .
phys . * b195 * , 137 ( 1982 ) .
ren , thesis , university of maryland , 1990 ( unpublished ) .
j. mcgovern , m. c. birse and d. spanos , j. phys .
* g16 * , 1561 ( 1990 ) .
see also m.k .
banerjee , phys .
c * 45 * , 1359 ( 1992 ) .
m. kim and m.k .
banerjee , phys . rev .
c * 48 * , 2035 ( 1993 ) .
g. t hooft , phys .
* 37 * , 8 ( 1976 ) ; phys .
d * 14 * , 3432 ( 1976 ) .
shifman , a.i .
vainshtein and v.i .
zakharov , nucl .
b163 * , 46 ( 1980 ) .
w. broniowski , ph .
d. thesis , u. of maryland preprint pp#89 - 113 ( 1986 ) .
m. k. banerjee , w. broniowski and t. d. cohen , _ chiral solitons _
( keh - fei liu , ed . ) , world scientific , singapore ( 1987 ) .
r. d. ball , _ proc .
workshop on skyrmions and anomalies , krakow _ ( 1987 ) .
t. degrand , ann .
of phys . * 101 * , 496 ( 1976 ) .
lee , t.d .
cohen and m.k .
banerjee , phys .
d * 40 * , 3060 ( 1989 ) .
m. gell - mann and m. lvy , nuovo cimento , * 16 * , 705 ( 1960 ) .
b. w. lee and h. t. nieh , phys .
166 * , 1507 ( 1968 ) .
m. kim , phd .
thesis , university of maryland , 1994 ( unpublished ) .
kochelev , yad .
* 41 * , 456 ( 9185 ) [ sov . j. nucl . phys . * 41 * , 291 ( 1985 ) ] .
e. shuryak and j. rosner , phys .
218b * , 72 ( 1989 ) .
m. oka and s. takeuchi , phys .
letts . * 63 * , 1780 ( 1989 ) ; nucl . phys .
* a524 * , 649 ( 1991 ) .
blask , u. bohn , m.g .
huber , b.ch .
metsch and h.r .
petry , z. phys * a337 * , 327 ( 1990 ) .
d. klabuar , phys .
d * 49 * , 1506 ( 1994 ) .
s. takeuchi , phys .
letts . * 73 * , 2173 ( 1994 ) .
s. takeuchi , phys .
d * 53 * , 6619 ( 1996 ) .
n. isgur and g. karl , phys .
72b * , ( 1977 ) 109 ; phys .
d * 18 * , ( 1978 ) 4187 .
n. isgur and g. karl , phys .
d * 19 * , ( 1979 ) 2653 .
for example see e. shuryak , _ the qcd vacuum , hadrons and the superdense matter _ ( world scientific , singapore , 1988 ) .
n. isgur , g. karl and r. koniuk , phys .
letts . * 41 * , 1269 ( 1978 ) , ibid .
* 45 * , 1738 ( e ) ( 1980 ) .
n. isgur , g. karl and d.w.l .
sprung , phys .
d * 23 * , 163 ( 1981 ) .
f. myhrer and j. wroldsen , z. phys .
* c25 * , 281 ( 1984 ) . y. umino and f. myhrer , phys .
d * 39 * , 3391 ( 1989 ) .
particle data group , phys .
d * 50 * , 1173 ( 1994 ) .
n. isgur and g. karl , phys .
74b * , ( 1978 ) 353 .
riska , private communication . e.a .
veit , a.w .
thomas and b.k .
jennings , phys .
d * 31 * , ( 1985 ) 2242 .
zhong , a.w .
thomas , b.k .
jennings and r.c .
barrett , phys .
d * 38 * , ( 1988 ) 837 .
w. weise , _ quarks and nuclei _ ( world scientific , singapore , 1984 ) , p. 58 .
brown and w. weise , phys . rep .
* 22 * , ( 1975 ) 279 .
g. barton , _ introduction to dispersion techniques in field theory _
benjamin , new york , 1965 ) .
eeg and j. wroldsen , physica scripta * 28 * , ( 1983 ) , 577 ; physica scripta * 30 * , ( 1984 ) , 234 .
table i. eigenvalues @xmath108 of @xmath84 as defined in eq ( [ eq : eigen ] ) for @xmath359 and @xmath46 quark spinor states with masses @xmath60 = 7.5 and 300 mev .
0.75 cm table ii .
values of scalar charge @xmath360 defined in eq .
( [ eq : scalar ] ) for @xmath361 and @xmath46 with quark masses @xmath362 mev and 300 mev .
0.75 cm table iii .
well established low lying negative parity baryon states considered in this work @xcite .
parity assignments and experimental mass ranges of the resonances are shown in columns two and three , respectively . in column four the @xmath153 coupled basis with the notation @xmath157 allowed by the projection prescription discussed in section [ diagonal ]
are shown for each state .
0.75 cm table iv .
matrix elements of @xmath84 , eq . ( [ h0toy ] ) , and @xmath159 , eq . ( [ eq : hooft ] ) , in mev for negative parity baryons .
these matrix elements are obtained using @xmath283 = 7.5 mev and using the projection prescription discussed in section [ diagonal ] .
the values of model parameters used in the calculation are given in section [ ccm ] .
the fourth column shows the matrix elements of oge interaction obtained in the mit bag model evaluated in the su(3 ) limit with vanishing current quark masses .
the oge matrix elements are multiplied by the strong coupling constant @xmath209 .
0.75 cm table v. bare masses and su(6 ) spin flavor wave functions of negative parity baryons @xmath0 , @xmath1 , @xmath2 and @xmath3 in the ccm .
masses and wave functions obtained by using the toy model hamiltonian , @xmath84 , alone are shown in the second column while the entries in the third column are calculated using the bare ccm hamiltonian , @xmath363 .
baryon masses in mev are indicated within parenthesis and the three
quark wave functions are expressed in the @xmath153 coupled basis .
column four shows the bare masses of hyperons in mev after being corrected for @xmath158 mass difference as explained in section [ toymodel ] .
0.75 cm table 6 .
relative percentages of spin flavor contents in the @xmath153 coupled basis for low lying negative parity @xmath165 and @xmath163 @xmath2 hyperons in the ccm of this work , the mit bag model with massless quarks and the nrqm @xcite . fig 1 .
the upper , @xmath364 , and the lower , @xmath365 , radial functions for quark spinors in states @xmath85 , @xmath86 and @xmath87 as defined in eqs .
( [ eq : swave ] ) to ( [ eq : awave ] ) .
the solid and dashed lines indicate radial functions obtained with quark masses of @xmath60 = 7.5 mev and @xmath60 = 300 mev , respecitively .
a ) @xmath85 radial functions @xmath366 and @xmath367 . the radial profile of the @xmath81 field is also shown for comaprison using the dot dashed line .
b ) @xmath86 radial functions @xmath368 and @xmath369 .
c ) @xmath87 radial functions @xmath370 and @xmath345 .
note that the distance shown in the horizontal axis is given in mesic units of @xmath371 .
0.75 cm fig 2 .
normalized @xmath372 dipole form factor @xmath373 for a ) @xmath374 , b ) @xmath375 ( dashed line ) , @xmath376 ( solid line ) and c ) @xmath377 ( dashed line ) , @xmath378 ( solid line ) .
0.75 cm fig 3 .
normalized @xmath379 dipole form factor @xmath380 for a ) @xmath375 ( dashed line ) , @xmath376 ( solid line ) and b ) @xmath377 and @xmath378 ( solid line ) .
note that @xmath381 involving @xmath256 states are identical . | in this two part series a chiral confining model of baryons is used to describe low lying negative parity resonances @xmath0 , @xmath1 , @xmath2 and @xmath3 in the mean field approximation .
a physical baryon in this model consists of interacting valence quarks , mesons and a color and chiral singlet hybrid field coexisting inside a dynamically generated confining region .
this first paper presents the quark contribution to the masses and wave functions of negative parity baryons calculated with an effective spin
isospin dependent instanton induced interaction . it does not include meson exchanges between quarks .
the three
quark wave functions are used to calculate meson excited baryon vertex functions to lowest order in meson quark coupling .
when the baryons are on mass shell each of these vertex functions is a product of a coupling constant and a form factor . as examples ,
quark contributions to @xmath0 hadronic form factors as well as axial coupling constants are extracted from the vertex functions and problems with the analytical behaviour of the model form factors are discussed .
the second paper will examine the mesonic corrections to excited baryon properties in the heavy baryon and one loop approximations .
pacs number(s ) : 12.40.yx , 14.20.gk , 14.20.jn |
the study of entanglement in quantum systems has been a major field of research in the last two decades .
motivations for this are various and important since entanglement provides not only a characterization of quantum states @xcite and a pathway on how to simulate them with numerical tools such as dmrg @xcite and tensor network states @xcite , but it helps also to characterize quantum phase transitions @xcite and to detect novel quantum phases , including topological phases .
non - trivial , explicit examples of the use of entanglement - related quantities , such as entanglement spectrum and negativity , range from quantum hall states @xcite to bose - hubbard @xcite and kondo models @xcite .
in addition the study of entanglement allows to characterize the computational power of quantum phases @xcite .
a key quantity entering the characterization of entanglement is provided by the entanglement entropy ( ee ) . for its
definition one takes a partition of a given system in two subsystems @xmath0 and @xmath1 ( the complement of @xmath0 ) , determines the reduced density matrix of a subsystem ( say , of @xmath0 ) @xmath7 by tracing out the degrees of freedom in @xmath1 and then computes its entropy : @xmath8 @xcite . the celebrated _ area law _
@xcite for the ee refers to the fact that typically the ee grows as the boundary of the subsystem @xmath0 : i.e. , for a system in @xmath9 dimensions and a subsystem of size @xmath10 having volume @xmath11 and area @xmath12 , @xmath13 according of the area law @xcite . the study of ee in various models has been recently a subject of intense research : ee can be explicitly computed in non - interacting systems of bosons and fermions @xcite , including trapped fermions @xcite , in integrable @xcite and 1d critical models @xcite , and in spin chains with long - range interactions @xcite .
an important result is that , for gapless @xmath14 integrable systems , the ee grows as @xmath15 , and that the prefactor is proportional to the central charge of the model .
the next - leading term of the ee has been studied as well : e.g. , for @xmath16 systems the size - independent constant entering @xmath17 is the so - called topological ee @xcite .
furthermore , the ee of a subsystem made of two disjoint intervals has been also intensively studied @xcite : for these issues we refer the interested readers to @xcite .
a first possible deviation from the area law is provided by logarithmic corrections : as shown in @xcite , for ( critical ) fermionic systems of dimension @xmath9 , the ee of a subsystem of size @xmath10 typically grows as @xmath18 ( this result does not hold for bosonic systems @xcite ) .
an explicit expression for the prefactor entering @xmath2 in dimension larger than @xmath19 may be given using the widom conjecture @xcite , and is found in remarkable agreement with numerical results @xcite .
entropy bounds for reduced density matrices of fermionic states were given in @xcite ; the role of disorder was also investigated @xcite , and it was shown that the momentum space entanglement spectrum reveals the location of delocalized states in the energy spectrum @xcite and that the entanglement structure depends only on the probability distribution of the length of the effective bonds @xcite .
furthermore , allowing for long - range interactions leads to a logarithmically diverging ee in gapped noncritical models @xcite , in spin chains @xcite and bose - einstein condensates @xcite .
non - local exponentially decaying couplings were considered in @xcite : at intermediate distances a volume law is observed , but as soon as @xmath10 becomes of the order of the length scale of the decay of the couplings the area law is recovered .
pertinent inhomogeneous couplings in simple spin chain hamiltonians with only nearest - neighbor interactions have been shown to induce a volume law in absence of translational invariance @xcite .
a violation of the area law for bosonic systems with bose surfaces was analyzed in @xcite .
fermi liquids are expected to obey the area law while non - fermi liquids in @xmath16 , although have been shown to satisfy the area law , are at the border between area law and non - area law ee @xcite .
a construction of a translation invariant fermionic state violating the area law was explicitly given in @xcite , when the fermi surface is a cantor - like set .
non - logarthmic deviations from the area law were also observed in @xcite , where two different kinds of disordered fermionic chains were considered .
an analysis of ee in spin chains having long - range interactions and a fractal fermi surface for the associated jordan - wigner fermions has been presented in @xcite : in particular in @xcite it was shown that ee for all translational - invariant pure states is at least logarithmic and there is an arbitrary fast sublinear entropy growth .
states supporting an area law for the ee are not maximally entangled and yet maximally entangled states have been deeply studied both for their intrinsic interest and in connection to quantum information protocols : maximally multiqubit entangled states up to eight qubits in qubit registers were reported in @xcite , while applications of absolutely maximally entangled states ( i.e. , multipartite quantum states maximally entangled with respect to any possible bipartition ) to a variety of quantum information protocols , including quantum secret sharing schemes @xcite and open - destination teleportation protocols @xcite . in this paper we show that a volume law for the ee of the ground state may emerge in fermionic lattices : to avoid ambiguities , we say that the a violation of the area law is obtained if , apart from logarithmic corrections , the ee scales as @xmath20 , with @xmath21 . in particular , for @xmath14 chains , a violation of the area law corresponds to @xmath22 and for @xmath6 we have the volume law . to set the notation
, we write the hamiltonian of ( spinless ) free fermions hopping on a generic lattice as @xmath23 the lattice has @xmath24 sites and its connectivity is characterized by the hopping matrix @xmath25 with @xmath26 ; of course @xmath27 and @xmath28 are the annihilation and creation fermionic operators on the site @xmath29 .
the number of fermions is @xmath30 , and the filling is then @xmath31 ( @xmath32 ) .
the sites of the lattice are denoted by capital letters @xmath33 while the sites of a generic subsystem having @xmath10 sites are denoted by ( small ) letters @xmath34 .
if the system is in the pure state @xmath35 , the ee of the subsystem @xmath0 is given by @xmath36 , \label{res}\ ] ] where @xmath37 is one of the @xmath10 eigenvalues of the correlation matrix @xmath38 in appendix [ app : a ] , following @xcite , we provide an explicit expression of the correlation matrix @xmath39 if @xmath35 is the ground - state of the hamiltonian , which is the situation we are going to mostly consider . from eq .
( [ res ] ) , one easily sees that for the ee to satisfy the volume law , one may construct a state for which each @xmath37 is equal to @xmath40 since for this state @xmath41 . in the following , we shall provide a method for the construction of such states and give some examples of hamiltonians supporting as a ground state a state with maximal ee . we get the remarkable result that , associated with these states , fermi surfaces with nontrivial topology naturally emerge . at first sight
one might think that a volume law could emerge only as a result of introducing a long - range hopping matrix @xmath25 in the hamiltonian ( [ hamiltonian ] ) .
our analyses shows that this is not the case since , at least for translational invariant systems , it is rather the topology of the fermi surface which really matters , as also pointed out in previous analysis @xcite .
indeed we demonstrate that , give a partition of the single - particle hilbert space to orthogonal subspaces @xmath42 and @xmath43 the state yielding maximal ee may be factorized into bell - pairs formed by two states belonging to @xmath42 and @xmath43 .
we call such state bell - paired state ( bp - state ) .
as we shall see explicitly , for translational - invariant hamiltonians , the bp - states are highly nonlocal in the space and the fermi surfaces have nontrivial topology . the paper is organized as follows : in sec .
[ sec1 ] we analyze a free fermion model with hoppings decaying as a power law with exponent @xmath44 .
we find that the ee obeys to the area law for each finite and positive @xmath44 , even tough , for @xmath45 the energy is not extensive @xcite ; only for @xmath46 , i.e. for the fully connected lattice , one has a volume law for the ee since @xmath47 .
unfortunately , the fully connected lattice is pathological in many respects , since the fermi level is infinitely degenerate and the number of sites of a given subsystem @xmath0 is at the same time its volume and its surface ( defining on a graph the volume of a subgraph as the number of vertices on it and the surface as the number of vertices on it linked to vertices outside the subgraph itself @xcite ) . in sec .
[ topol ] we explicitly construct , for any given lattice and arbitrary filling @xmath48 , the general form of the states rendering the ee and all the rnyi entropies maximal .
this construction allows for to have an explicit momentum representation of the state with maximal entanglement entropy . in sec .
[ hamiltonians ] we provide explicit examples of hamiltonians supporting a bp - state as the ground state and analyze the topology of their fermi surface .
[ concl ] is devoted to our concluding remarks , and in the appendices [ app : a ] [ app : lrphase ] [ app : fcn ] we provide the reader with technical details about the models described in the main text . in appendix
[ sec : random ] we analyze the violation of the area law of the ground state for a free fermionic model which is not translational invariant .
in this section we focus our attention on a translationally invariant one - dimensional chain , with non - local hopping matrix @xmath49 given by @xmath50 where the distance @xmath51 , due to periodic boundary conditions , is defined as @xmath52 the hopping matrix , being translationally invariant , is readily diagonalized : its eigenstates are given by plane waves and , in the thermodynamic limit , the energy spectrum is @xmath53 where @xmath54 belongs to the first brillouin zone ( @xmath55 for even @xmath24 ) and @xmath56 ( with @xmath57 ) . as usual , even if for @xmath58 the ground - state energy in the thermodynamic limit diverges , one can make the energy extensive by the so - called kac rescaling @xcite . the function @xmath59 is plotted for two values of @xmath57 in fig.[fig1 ] .
one sees that the spectra are monotonic for @xmath60 and @xmath61 , and , thus , the filling of the momentum eigenstates leading to the occupation of the fermi sea is the same for each value of @xmath57 .
the same result holds also for @xmath62 for any finite number of sites ( appendix [ app : lrphase ] ) . as a consequence ,
the ee does not change since the correlation matrix ( [ correl_intr ] ) depends only on the ground state ( and not on the spectrum ) ; in the thermodynamic limit , @xmath63 for each @xmath64 for any filling @xmath48 , just as it happens if the hopping @xmath25 was short range @xcite . a similar analysis , yielding the same results ,
may be carried out also for @xmath65 and @xmath66 .
( in units of @xmath67 ) of a model with nearest - neighbor hopping corresponding to @xmath68 ( solid line ) and non - local hoppings with @xmath69 ( dashed line ) in the thermodynamic limit .
the horizontal lines represent the fermi energy at half - filling for both models .
the regions of @xmath70-space comprised within the points @xmath71 and @xmath72 are the occupied wave vectors , indicated in the whole diagram with a ( pale blue ) shading . in the white ( top and bottom left )
the hopping structures of both models are represented - in the bottom right inset the energy spectrum for a next - nearest - neighbor model having four fermi points ( denoted by @xmath71 , @xmath72 , @xmath73 and @xmath9 ) is plotted ( @xmath74 , @xmath75).,title="fig : " ] things change for @xmath46 . here , it is easy to verify that the single - particle energy spectrum is made of a non - degenerate ground - state and of an @xmath76-fold degenerate excited state , implying that the many - body ground - state is highly degenerate .
in addition , the fermi surface passes from a two - point set ( as it happens for @xmath64 ) to a continuous set . in appendix
[ app : fcn ] we show that , for @xmath77 , @xmath78 ; \label{s_fully}\ ] ] in particular , @xmath79 for @xmath80 .
this corresponds to an equal _ a priori _ probability of occupation of all degenerate states by the available particles .
the fully connected hopping model does not have a specific dimensionality @xmath9 : however , if we think it as the @xmath81 limit of a @xmath9-dimensional long - range hopping model and @xmath0 is a cubic subsystem with size @xmath82 , then the number @xmath10 of sites of @xmath0 is given by @xmath83 .
it appears as we already obtained a volume law for the ee .
unfortunately , for the fully connected hopping model , the number of sites @xmath10 of the subsystem @xmath0 is at the same time the volume and the surface of @xmath0 in the sense that all the @xmath10 sites of @xmath0 are linked with the other sites of the rest of the system @xmath1 .
in addition the mutual information between @xmath0 and @xmath1 is vanishing for @xmath77 and , thus , the emergence of a volume law corresponds here to nothing but the transition to a classical state .
the analysis carried out in this section shows that , for translational invariant systems , long - range hoppings alone are not enough to guarantee the emergence of the volume law for the ee and it appears that the structure of the fermi surface is bound to play a key role in the behavior of the ee . in particular , for a translational invariant chain and a quadratic hamiltonian of the form , all the models having fermi wave vectors @xmath84 and @xmath85 at the points @xmath86 of fig.[fig1 ] have the same correlation matrix , which , in the continuum limit , reads @xmath87 leading to the same ee ( we remind that our hamiltonian does not include any `` superconducting '' @xmath88 terms , that would change the correlation matrix ) . if @xmath89 then the well known result @xmath90 } / \left [ \pi \left ( i - j \right)\right]$ ] of the nearest - neighbor free fermionic chain is recovered @xcite . to better clarify the role played by the fermi surface
let us consider the energy spectrum represented in the bottom right inset of fig.[fig1 ] having @xmath91 fermi wave vectors @xmath92 at the points @xmath93 : the same argument leading to yields @xmath94 ( this formula can be easily generalized to fermi surfaces with wave vectors @xmath95 ) .
as a result , the ee depends only on @xmath92 and not on other details of the energy spectrum .
in order to elucidate the role played by the fermi surface for constructing a maximal entangled state in a translationally invariant free fermionic lattice , it is instructive to look at _ all _ the possible fermi surfaces arising in 1-dimensional systems of small size @xmath24 .
this task is simplified if one notices that circular shifts ( @xmath96 ) and reflections ( @xmath97 ) of the fermi surface do not alter the ee . as a result
, the number of fermi surfaces yielding different values of the ee is much reduced ; from combinatorics this number is obtained by counting all the distinct reversible bracelets @xcite .
the result is shown in fig.[fig2 ] for a system with @xmath98 sites at half - filling .
as one can see , the fermi surfaces exhibiting the maximal ee are realized with an alternating filling of the wave vectors periodic in @xmath70-space with period @xmath99 , while fermi surfaces exhibiting higher periodicity in @xmath70-space , have a piecewise linear behavior .
the minimal ee for a system of @xmath100 sites at half filling is achieved with a fermi surface made up of two points , i.e. made occupying @xmath101 states with adjacent wave vectors .
similar findings are obtained for different system sizes and fillings .
for all the combinatorially distinct fermi surfaces for a system of @xmath98 sites at half - filling . in the right legend a filled ( empty ) dot denote a filled ( unoccupied ) momentum eigenstate .
the states are ordered with respect to increasing @xmath70 ( modulo @xmath102 ) .
the states having a periodic structure in @xmath70-space are highlighted in the legend with a color red and are plotted in the left figure with thick continuous lines : the lower thick blue line refers to momenta occupied up to @xmath103 ( resulting in logarithmic ee ) , the upper thick red line refers to the state with alternating filled momenta ( having linear ee ) , the central thick scarlet and green lines to the states having sequence of two ( three ) filled momenta and two ( three ) holes.,title="fig : " ] this procedure allows to explicitly construct a state of maximal ee for small fermionic systems through a pertinent filling of the fermi sea .
in addition , once the sistem size and the filling are given , it selects the fermi surface for which a volume law emerges . in the following
, we shall generalize the above result to systems of finite , but arbitrarily large , size @xmath24 . as shown in sec .
[ intro ] , the ee between @xmath0 and @xmath1 for a non - interacting fermionic system in its ground - state is determined by the correlation matrix , which may be usefully rewritten as @xmath104 where @xmath105 labels the set of single particle states entering in ( [ eqn : corrbis ] ) . to compute the ee one needs to find the eigenvalues @xmath37 of the matrix @xmath106 . in the following ,
we denote with @xmath42 and @xmath107 two nonorthogonal subspaces of the single - particle state space @xmath108 such that @xmath109 and @xmath110 .
we then define @xmath111 and @xmath112 as the projection operators over the subspaces @xmath42 and @xmath107 , respectively . upon introducing the operator @xmath113
one has that @xmath114 with @xmath115 . as a result
the ee can be written as @xmath116.\ ] ] with the notation used in ( [ eqn : sgamma ] ) , the symmetries of @xmath2 are made manifest since @xmath117 and @xmath118 ( where @xmath119 is the orthogonal complement of @xmath107 ) . in the following
we shall determine , for a given @xmath42 , the vector space @xmath107 for which the ee is maximal . as we shall see , for translational invariant free fermionic lattices , this amounts to determine the topology of the fermi surface maximizing the ee .
the ee is strictly upper bounded by the dimension of the smallest space between @xmath42 and @xmath43 times @xmath120 , since in eq .
( [ eqn : sgamma ] ) natural logarithms have been used .
of course , @xmath121 and @xmath122 , where @xmath123 ( @xmath124 ) is the cardinality of the set @xmath0 ( @xmath1 ) . as a result @xmath125
we shall now explicitly construct the states satisfying this upper bound .
we observe that , in the construction of maximal ee states , we do not need to fix the dimension of @xmath107 , i.e. the number of fermions ( @xmath126 ) . indeed ,
if @xmath127 , we shall show that the maximal ee @xmath128 is obtained when the filling fraction @xmath48 is such that @xmath129 .
it follows that , for fixed @xmath24 , the maximal ee is obtained for @xmath130 and @xmath131 . for simplicity we set latexmath:[$|a|
\leq worked out by exchanging @xmath0 and @xmath1 .
since the number of nonzero eigenvalues of the hermitean operator @xmath133 is smaller or equal than @xmath134 and the maximum contribution of each of these eigenvalues to the total ee is @xmath135 , we conclude that , in order to obtain the maximum ee , @xmath133 should have @xmath134 eigenvectors @xmath136 , @xmath137 with eigenvalue @xmath40 .
namely , one should have @xmath138 from the definition of @xmath133 , one easily sees that , in order to have a maximum ee state , @xmath139 should be at least equal to @xmath134 . as a consequence ,
if the subspace @xmath107 is spanned by the orthonormal vectors @xmath140 , ( so that latexmath:[$p_{\mathcal{b}}=\sum_{i=1}^{|b| } choose the first @xmath134 vectors @xmath142 to have a nonzero projection on @xmath42 .
one has then that @xmath143 where the complex coefficients @xmath144 are yet to be determined , the @xmath145 are an orthonormal basis for @xmath42 and the @xmath146 are orthonormal vectors in @xmath43 .
of course , additional vectors will not give rise to nonzero eigenvalues of @xmath133 thus we can limit ourselves to @xmath147 i.e. we are determining @xmath107 up to vectors orthogonal to @xmath42 @xcite .
the above decomposition indeed is similar to the one obtained in @xcite where it was used in order to obtain _ lower bounds _ for the entanglement entropy in a fermionic systems . by requiring that ( [ eq : eigengonehalf ] )
is satisfied , one gets that @xmath148 ( @xmath149 ) . without loss of generality
one may choose @xmath150 , thus @xmath107 is spanned by : @xmath151 this determines the form of the maximal ee state , and explicitly shows that , given @xmath152 , the space @xmath107 maximizing the ee is made out of @xmath10 bell - paired states among @xmath0 and @xmath1 . in the following
we shall refer to these states as bp - states .
a bp - state is pictorially represented in fig .
[ fig3 ] .
it should be stressed that * in our construction the nature of the set @xmath107 is left unspecified , the only natural requirement being that it is a set of allowed single particle states . only for translational invariant systems
the set @xmath105 may coincide with the set of single particle momentum states ; * the set @xmath0 does not need to be simply connected : in particular , if @xmath0 it is not simply connected the bp - states provided by the indicated construction are not localized around the sites ; * if , instead , the set @xmath105 is fixed , our construction allows to determine the set @xmath0 yielding the state with maximal ee ; * if one wishes to find maximal ee states as the size of @xmath0 is enlarged , the problem to be considered is the following : given a sequence of sets @xmath153 ( with @xmath154 if @xmath155 ) to which corresponds a set of linear spaces @xmath156 one should determine a subspace @xmath107 for which the ee is maximal for every @xmath157 .
for this purpose it is enough to construct the basis when @xmath158 for @xmath24 even or @xmath159 for @xmath24 odd : in other words the maximal ee states are obtained at half - filling and have @xmath160 for @xmath161 as plotted in fig.[fig2 ] , red line ; * the bp - state also maximizes the rnyi entropy of order @xmath162 since , for @xmath163 , @xmath164.\ ] ] as one can easily verify that the rnyi entropy of order @xmath162 is bounded from above by @xmath165 and attains a maximum when all the @xmath37 are equal to @xmath40 : therefore , for a bp - state , all the rnyi entropies are maximal and equal to the maximum value of the ee @xmath166 .
let us consider now a spatial partition in which @xmath0 is a simply connected subsystem of the lattice : then the states in @xmath0 sharing the bell - pairs with @xmath1 are localized around sites .
one then expects that , plotting @xmath106 ( @xmath115 ) for a maximal ee state , yield @xmath167 .
a useful quantity to visualize the correlations emerging between @xmath42 and @xmath43 is the correlator @xmath168 which equals @xmath169 if @xmath29 and @xmath170 belong to @xmath0 .
we plot @xmath171 for various states in fig.[fig4 ] , where the correlation matrix @xmath172 of @xmath0 is the one on the top left part of @xmath173 .
the translationally invariant states obtained by occupying contiguous momentum eigenvectors up to the fermi wave vector are characterized by an alternating pattern of zero and nonzero diagonals as we move away from the main diagonal ( which is by construction equal to one ) , as shown in panel ( a ) of fig.[fig4 ] .
the maximal ee state for translational invariant states is plotted in panel ( b ) and it is made out of four identity submatrices .
it should be noticed that the bp- state is made out of bell pairs connecting sites at distance @xmath174 .
the panel ( c ) instead refers to a state where the occupation in @xmath70-space alternates in the momentum space sequences of two filled states and two holes ( in fig.[fig2 ] its ee is represented with the continuous scarlet line and it does not have maximal ee ) .
finally panel ( d ) shows the general structure of a bp - state having maximal ee obtained coupling @xmath174 randomly chosen orthogonal states @xmath175 in subspace @xmath42 and @xmath174 randomly chosen orthogonal states @xmath176 in subspace @xmath177 . notice that , by a change of basis , the maximal ee state plotted in panel ( d ) may be represented as in panel ( b ) .
one concludes that the condition that the top left matrix of the matrix @xmath173 is diagonal is sufficient to have a maximal ee state . for four different kind of entangled states at half - filling in a system with @xmath178 .
the four matrices @xmath71 , @xmath72 , @xmath73 , and @xmath9 refer to a state with ferromagnetic hopping , zigzag state , a state with two filled states and two holes alternatively and a generic bp - state respectively.,title="fig : " ]
we shall exhibit here a few one - dimensional models and supporting ground - states leading to a violation of the area law , and we shall look at the nontrivial topology of the fermi surface . for this purpose we firstly notice that , for a translationally invariant chain and for @xmath0 a simply connected domain such that @xmath179 , the maximum ee state may be obtained by occupying the even ( or odd ) momentum eigenvectors .
indeed , a basis of @xmath42 is given by vectors @xmath180 such that @xmath181 while a basis of @xmath43 is given by vectors @xmath182 such that @xmath183 in eqs .
( [ alpha1 ] ) and ( [ alpha ] ) @xmath184 with @xmath185 and the @xmath186 accounts for the filling of the even and odd frequencies respectively .
one sees that the subspace spanned by is the state with only the even ( odd ) frequencies occupied .
this is a bp - state maximizing the ee .
notice that , if we choose a state with an alternating sequence of two filled momenta and two holes , then the ee is linear up to @xmath187 : for @xmath98 this corresponds to the ee thick scarlet line of fig .
, i.e. , the @xmath188th state from top in the right part of the legend of the same figure .
similarly , for the state with an alternating sequence of three filled momenta and three holes , the ee is linear up to @xmath189 : for @xmath98 this corresponds to the ee thick green line of fig .
, i.e. , the bottom state in the left part of the legend . for general @xmath24 at half - filling , a state having a sequence of @xmath190 filled momenta and @xmath190 holes ( with @xmath24 multiple of @xmath191 ) will have linear ee up to @xmath192 .
a simple hamiltonian supporting a bp - state as a ground state , thus yielding the volume law for the ee , has the form with a hopping matrix @xmath49 given by ( with even @xmath24 ) @xmath193 with @xmath194 and periodic boundary conditions .
notice that , in , only hoppings between sites distant @xmath174 are allowed . at half - filling ,
the ground - state is constructed by occupying only the states with @xmath195 even ( occupation of the states with @xmath195 odd is obtained for @xmath65 ) . as a result
, the fermi surface has a fractal topology and its counting box dimension @xmath196 @xcite equals to @xmath19 .
this example provides an explicit and simple setting where the emergence of the volume law is associated to a non - trivial topology of the fermi surface : this sheds light on the results of previous investigations @xcite .
slight modifications of the hoppings can be built to have states having sequences of @xmath190 filled momenta and @xmath190 holes .
fractal fermi surfaces may be realized as pertinent limits of other model hamiltonians . in the following
we analyze two specific models where this happens .
a possible way to obtain a fractal fermi surface is to consider the effect of a phase in a model with long - range hoppings : @xmath197 where @xmath198 , being @xmath199 a constant and @xmath200 the oriented distance between the sites @xmath29 and @xmath170 , whose definition is given in ( [ eqn : distphase ] ) .
the spectrum of the ensuing hopping hamiltonian is analyzed in appendix [ app : lrphase ] . for odd @xmath24 the eigenvalues
are given by @xmath201 where @xmath202}{m^\alpha } ; \label{elle}\ ] ] as usual , @xmath54 with @xmath203 .
a similar formula is obtained for even @xmath24 . for @xmath204 ,
the spectrum is always monotonous in the interval @xmath205 $ ] , while for @xmath206 the spectrum is monotonous for @xmath207 . more precisely , at fixed @xmath208 and @xmath77
, there is a critical value of @xmath209 , depending both on @xmath210 and @xmath208 , such that , for @xmath211 , at half - filling , all the momenta @xmath70 are occupied in an alternating way , as shown in fig.[fig5 ] .
thus , for @xmath211 and at half - filling , the ground - state is a bp - state , ee is linear with slope @xmath135 and the fermi surface has a fractal topology with @xmath212 : this is shown in fig.[fig6 ] . for @xmath213 one
has that @xmath214 : this happens since , in the thermodynamic limit ( @xmath215 ) , it is not possible to define and occupy only the even momenta ; however , for each @xmath24 arbitrarily large , @xmath209 is strictly positive .
when @xmath216 only a fraction of the momenta are occupied in an alternating way , since the `` zig - zag '' structure of the dispersion relation is partially lost . as a result , the slope of the ee decreases , as shown in the inset of fig.[fig6 ] . , @xmath217 , filling factor @xmath218 and @xmath219 .
right inset : detail of the main plot showing the alternating occupation of the modes @xmath70 , the fermi energy corresponding to the dashed line . left inset : loss of the alternating occupation increasing @xmath44 , with @xmath220 , @xmath221 , @xmath218 and @xmath219.,title="fig : " ] , @xmath220 .
different total number of sites are considered : @xmath219 ( full circles ) , @xmath222 ( triangles ) , @xmath223 ( empty circles ) .
inset : slope of the entanglement entropy fitted with a linear function @xmath224 for different values of @xmath44 : @xmath225 ( triangles ) , @xmath220 ( full circles ) , @xmath226 ( empty circles ) and @xmath227 .
, title="fig : " ] we consider here a translational invariant chain of @xmath24 sites with periodic boundary conditions and with eigenfunctions given by plane waves @xmath228 .
we assume that the hamiltonian is such that the single - particle energy spectrum has the form @xmath229 where @xmath44 is a positive odd integer .
the spectrum is plotted in fig.[fig7 ] for @xmath230 .
the fermi surface , in a pertinent range of fillings , has a fractal topology and , at half - filling , the fermi energy is zero so that fermi surface is simply given by the set of points @xmath231 .
the point @xmath232 is an accumulation point for this set with box counting dimension @xcite @xmath233 so that @xmath234 for @xmath230 . , with the thick dashed line corresponding to the fermi energy for half - filling.,title="fig : " ]
we numerically determined the ee for the above model for increasing values of @xmath24 as the parameter @xmath44 takes the values @xmath235 .
the ee has a well defined thermodynamic limit and , for small size of the subsystem @xmath0 , is well described by a power - law . in order to give a reliable estimate of this power - law growth one needs to compute the ee for @xmath236 and fit the obtained values with the function @xmath237 for different system sizes .
one needs this procedure to get rid of finite - size effects since , even in the the short - range model , the ee shows finite - size effects when @xmath10 is comparable to the system size @xcite .
thus , to recover the expected logarithmic growth one has to fix @xmath10 and vary @xmath24 .
the results of this fit are reported in fig .
[ fig8 ] . here
we plot , on the left panel , the ee for different values of @xmath44 . the results for the ee obtained for the short - range model and the bp - state are also shown for comparison . in the right panel of fig .
[ fig8 ] we plot the exponent @xmath4 as a function of @xmath24 .
we see that , as @xmath24 is increased , @xmath4 approaches @xmath196 .
since this feature is shared also by the previous models , one is tempted to conjecture that this is a general feature of the scaling of the ee in translational invariant chains .
our findings agree with the results presented in the appendix of @xcite , where a construction of a translation invariant fermionic state violating the area law was explicitly given with cantor - like fermi surfaces , and with the results of @xcite where , for infinitely many intervals in a spin chain , an ee of the form @xmath238 was found with @xmath44 possibly taking any value between @xmath239 and @xmath19 .
we observe that in @xcite a formal criterion for the growth of the ee in presence of fractal fermi surfaces has been discussed : in particular , it was shown in that if @xmath240 for a small set @xmath241 and @xmath242 ( with @xmath243 and @xmath244 real constants and @xmath133 the fermi surface ) , then there is a deviation from the area law with exponent @xmath245 ( see also @xcite ) .
our results imply that such coefficient @xmath246 for this class of hamiltonians is related to the box counting dimensions through @xmath247 : an interesting problem for future research is the study of such relation in the general case .
so far we analyzed only models in which a fermi surface can be defined : it is natural to expect that violations of the area law may emerge also in situations where it is impossible to define a fermi surface .
when disorder is present such situation arises naturally in single realizations of disorder . in appendix
[ sec : random ] we shall analyze a model with random long - range hoppings and we shall see how deviations from the area law may appear also in absence of the translational invariance .
we investigated how the area law for the entanglement entropy ( ee ) may be violated in non - interacting fermionic lattices and provided a method enabling to construct the states with maximal ee exhibiting a volume law .
we called these states bp - states . for these states
the ee is linear in the size of the subsystem @xmath0 and the fermi surface has fractal topology . for translational invariant
free fermionic hamiltonians , bp - states may be obtained , at half filling , by occupying , according to fermi statistics , even or odd momentum eigenvectors providing an explicit momentum representation of the state with maximal entanglement entropy . by this procedure ,
one originates a `` zig - zag '' structure of the dispersion relation leading for fermionic chains to the emergency of a fractal fermi surface with box - counting dimension @xmath19 . by means of this procedure
one can construct an explicit hamiltonian whose ground - state supports exactly the volume law .
we then provided some examples of fermionic models for which the ground state may have an ee @xmath2 between the area and the volume law , and we gave an explicit example of a one - dimensional free fermion model where the ee is such that @xmath248 with @xmath4 being intermediate between @xmath5 ( area law ) and @xmath6 ( bp - state ) .
we saw that , also for this model , the dispersion relation has a `` zig - zag '' structure leading to a fractal fermi surface whose counting box dimension equals , for large lattices , @xmath4 .
it is attractive to speculate that there may be a general relation between the fractal dimension of the fermi surface , measured by the box counting dimension , and the exponent @xmath4 measuring the amount of violation of the area law for one - dimensional translational invariant free fermion lattices . here , we only report the fact that , in all the one - dimensional examples analyzed in this paper , this relation holds true . as a by product , our analysis shows that , a volume law for the ee can not emerge in free fermion lattices as a result of long range hopping alone .
indeed , our analysis shows that , at least for translational invariant systems , a fractal structure of the fermi surface is needed to establish a volume law for the ee .
although we studied only non - interacting fermions on the lattice , our analysis is relevant also for spin models admitting a fermionic representation .
indeed , it has been recently exhibited a spin chain model supporting a volume law for the ee @xcite : in its fermionic representation , the hamiltonian is highly non - local in agreement with the scenarios presented in this paper .
_ acknowledgements : _ we would like to thank a. bayat , f. buccheri , j. magan and e. tonni for useful discussions . during the completion of this work , we became aware of results on volumetric law in fermionic lattices with inhomogeneous nearest - neighbor hoppings by g. ramrez , j. rodrguez - laguna , and g. sierra @xcite : it is a pleasure to thank them for discussions and useful correspondence .
thanks the ministry of science , technology and innovation of brazil for financial support and cnpq for granting a `` bolsa de produtividade em pesquisa '' . p.s . and s.p .
acknowledge partial support from mcti and ufrn / mec ( brazil ) .
a.s . acknowledges support from cnpq , and from the center for nanoscience and nanotechnology at tel aviv university and the pbc indo - israeli fellowship . a.t . acknowledges support from the italian prin `` fenomeni quantistici collettivi : dai sistemi fortemente correlati ai simulatori quantistici '' ( prin 2010_2010llkjbx ) .
is supported by a rita levi - montalcini fellowship of miur .
acknowledge hospitality from international institute of physics ( natal ) where part of this work has been carried out .
g.g . and a.t .
acknowledge support from the european project matterwave .
_ note added : _ after the submission of this paper we noticed on the ` arxiv ` a very interesting paper on the power law violation of the area law in quantum spin chains @xcite .
here we derive the correlation matrix and the ee of the subsystem @xmath0 for a generic graph @xmath249 with @xmath24 sites and @xmath30 fermions hopping on it .
the model is described by the hamiltonian and the filling is @xmath31 with @xmath32 . as in sec .
[ sec1 ] , the sites of the lattice @xmath249 are denoted by capital letters and the eigenvalues equations of read as @xmath250 where @xmath251 are @xmath24 single - particle energy eigenvalues ordered so that @xmath252 and @xmath253 are the corresponding @xmath24 orthonormal eigenfunctions . upon defining @xmath254
one immediately sees that the operators @xmath255 obey to the canonical fermionic anticommutation relations and that the hamiltonian may be rewritten as@xmath256 so that the ground - state @xmath35 can be written as @xmath257 given a set @xmath0 , whose sites are labelled by @xmath34 , one may define the correlation matrix @xmath258 as the matrix whose entries are given by @xmath259 using and , one finds @xmath260 if one denotes with @xmath37 ( @xmath261 ) the @xmath10 eigenvalues of the matrix @xmath39 one gets @xcite @xmath36 . \label{res_app}\ ] ]
we analyze here the spectrum of the model , introduced in sec .
[ long_phi ] .
the hopping matrix reads @xmath262 where @xmath263 due to the translational invariance , the eigenstates are plane waves , while , for finite @xmath264 , the spectrum is given by @xmath265 with @xmath266 . even if , for finite @xmath24 , @xmath59 forms a discrete set corresponding to integer values of @xmath195 , it is most useful to provide an expression of ( [ lerch ] ) valid for all values of @xmath70 .
for this purpose ,
eq . may be rewritten using lerch transcendent functions @xcite : @xmath267 yielding @xmath268 & for\ : odd\ : n_{s}\\ \re\left[z^{\frac{n}{2}}\phi(z,\alpha,\frac{n}{2})-z\phi(z,\alpha,1)\right ] -\frac{\cos\left(\pi n_{k}\right)}{2\left(\frac{n}{2}\right)^{\alpha } } & for\ : even\ : n_{s } , \end{cases}\ ] ] with @xmath269 .
let us start by analyzing the spectrum when @xmath204 ; this corresponds to power - law decaying hoppings . since @xmath270
one may consider only the interval of the brillouin zone corresponding to @xmath271 .
for @xmath272 , @xmath273 is a monotonically increasing function of @xmath70 , so that the many - body ground - state is filled following an ascending order of @xmath274 , just as in the short - range tight - binding model .
thus , for every value of @xmath275 , the ee is the same of the tight binding model and , thus , follows the usual area law for the ee . for @xmath45
, @xmath273 is an oscillating function with @xmath276 maxima ( and minima ) almost equidistant in the interval @xmath277 .
in addition , the set @xmath278 is still monotonous in @xmath271 so that every wave number @xmath70 lies between a different pairs of local maxima an minima . as a result ,
eve for @xmath45 the ee follows an area law . if @xmath279 the function @xmath273 shifts , losing its parity . for @xmath45 , if one considers two consecutive @xmath70s of the discrete set @xmath278 one sees that the shift introduced by a small @xmath208 increases the energy of one of them and decreases the energy of the other .
it follows that the set @xmath278 is no more monotone and takes a zigzag shape . as shown in fig.[fig5 ] , the energies corresponding to @xmath280 s of different parity arrange themselves on two different branches .
the maximum spacing between the two branches is bounded by the amplitude of the oscillation of @xmath273 . to give an estimate of that , one may , for @xmath204 , approximate @xmath273 around @xmath281 with a cosine function @xcite @xmath282 where @xmath283 the study of the amplitude @xmath105 shows a weak polynomial dependence on the number of sites , as plotted in fig.[fig9 ] , so that a zigzag behavior of the spectrum is expected for every finite @xmath264 .
the lower branch has always negative concavity in @xmath281 , while the top branch s concavity can be positive or negative , depending on @xmath44 . in the first case and for half - filling , there appears an alternation in the occupancy of @xmath70 , i.e the fermi energy is lying between the two branches , giving rise to the bp - state described in section [ topol ] .
the concavity remains positive for @xmath284 and , in this regime , the ee is always maximal showing a volume law behavior with fixed slope ( see inset in fig.[fig5 ] ) . for @xmath285 ,
as the concavity of the upper branch becomes negative , some consecutive @xmath70s close to @xmath232 are occupied while close , to @xmath281 the fermi energy goes below the two branches .
this breaks the complete alternating configuration but , for @xmath44 larger but close to @xmath286 the ee has still depends linearly on @xmath10 but with a lower slope .
the crossover from volume to area law occurs smoothly as @xmath44 is increased .
for @xmath46 the long - range hopping model becomes simply @xmath287 where we divided the hopping coefficient @xmath67 by @xmath24 to keep the single particle spectrum lower bounded .
the spectrum of ( [ fully_conn ] ) is made of two eigenvalues : @xmath288 and @xmath289 corresponding , respectively , to a non degenerate ground state and a @xmath290 degenerate excited state .
the entries @xmath291 of the correlation matrix are given by @xmath292 where @xmath31 is the filling and @xmath72 has to be determined .
since the ground - state energy ( more precisely , the free energy for @xmath293 ) is @xmath294 it follows that @xmath295 the correlation matrix has , the , the form @xmath296 with eigenvalues @xmath297 @xmath298 inserting ( [ eqn : eta0 ] ) and ( [ eqn : eta1 ] ) into ( [ res ] ) one gets @xmath299 which , for @xmath300 , yields @xmath301.\ ] ] from ( [ c9 ] ) a volume law for the ee is attained .
however , @xmath2 is not a true measure of entanglement since the initial state is mixed and , in addition , the mutual information turns out to be zero . as a result
, the emergence of a volume law does not lead to nonlocal correlations for this model . in the following ,
we show that the mutual information is indeed zero .
if @xmath0 and @xmath1 are two complementary sets covering the full lattice , the mutual information is defined as @xmath302 where @xmath303 is the total entropy . for a mixture of @xmath304 degenerate states ,
the total entropy is given by @xmath303 @xmath305 here , the degeneracy of the many - body ground state is given by @xmath306 so that @xmath307.\ ] ] in the limit of large @xmath24 , at fixed filling , one easily obtains @xmath308.\ ] ] the entropy of the set @xmath1 has an analogous expression to ( [ eqn : sa ] ) @xmath309 where @xmath310 and @xmath311 . for large @xmath24 ( [ eqn : sbara ] ) becomes @xmath312.\ ] ] finally , putting together ( [ eqn : sa ] ) , ( [ eqn : sbara ] ) and ( [ eqn : st ] ) one gets @xmath313
here , we present some preliminary results on a model where long - range randomness is included in the hopping matrix .
ee has been studied for different disordered models @xcite ; in particular , a violation of the area law has been found for free fermionic models in their metallic phase @xcite .
the model is defined by hamiltonian with a random long - range hopping matrix given by @xmath314 in eq .
( [ random ] ) @xmath315 is a random variable assuming the values @xmath316 with equal probability .
this model breaks the translational symmetry ; as a result , one can not analyze the states using momentum eigenvectors .
a direct diagonalization of the matrix @xmath49 is needed to compute the correlation matrix , whose eigenvalues are then used to compute the ee of the ground state for various sizes , @xmath10 , of the subsystem @xmath0 .
our findings are summarized in fig .
[ fig10 ] and are the following : for @xmath317 , the logarithmic behavior of the random short - range model is recovered @xcite .
when @xmath44 decreases , at fixed size @xmath24 , the ee clearly drops off and then , at a value of @xmath44 of order @xmath19 , the ee grows back again , as shown in the left part of fig .
[ fig10 ] .
we observe that , for @xmath318 , the ee is larger than the one of the random short - range model and it appears to be approximately linear . to look for the asymptotic behavior of @xmath2 ,
it is most convenient to allow @xmath10 to vary in a fixed interval ( for example between @xmath19 and @xmath319 ) and vary @xmath24 . it turns out that , for @xmath320 , a reasonable fit function has the form @xmath248 .
varying @xmath24 we plot , in the right part of fig .
[ fig10 ] , @xmath4 as a function of @xmath24 for three different values of @xmath45 . for @xmath321
the ee decreases as @xmath24 increases and there is an interval of values of @xmath44 where the ee becomes even smaller than the one of the clean short - range model @xcite . for @xmath322 ,
one finds a power - law behavior of @xmath17 , and @xmath4 appears to grow as @xmath24 increases .
indeed , considering the dimension of @xmath107 beyond @xmath323 , proves to be detrimental for getting higher ee states .
in fact if we switch @xmath107 and @xmath324 due to particle - hole simmetry we see that the number of nonzero eigenvalues of @xmath258 gets smaller than @xmath325 . | we show how the area law for the entanglement entropy may be violated by free fermions on a lattice and look for conditions leading to the emergence of a volume law .
we give an explicit construction of the states with maximal entanglement entropy based on the fact that , once a bipartition of the lattice in two complementary sets @xmath0 and @xmath1 is given , the states with maximal entanglement entropy ( volume law ) may be factored into bell - pairs ( bp ) formed by two states with support on @xmath0 and @xmath1 .
we then exhibit , for translational invariant fermionic systems on a lattice , an hamiltonian whose ground state is such to yield an exact volume law . as expected , the corresponding fermi surface has a fractal topology .
we also provide some examples of fermionic models for which the ground state may have an entanglement entropy @xmath2 between the area and the volume law , building an explicit example of a one - dimensional free fermion model where @xmath3 with @xmath4 being intermediate between @xmath5 ( area law ) and @xmath6 ( bp - state inducing volume law ) . for this model ,
the dispersion relation has a `` zig - zag '' structure leading to a fractal fermi surface whose counting box dimension equals , for large lattices , @xmath4 .
our analysis clearly relates the violation of the area law for the entanglement entropy of the ground state to the emergence of a non - trivial topology of the fermi surface . |
an important next step " in the field of extrasolar planet research will be to characterize their atmospheres .
the close - in extrasolar giant planets ( cegps ) will be the first targets ; indeed , the tentative detection of na i in hd 209458b has been reported ( charbonneau et al . 2002 ) .
old cool stars ( like most known cegp hosts ) all have very small fractional linear polarization ( lp ; e.g. , leroy 1993 ) . at the same time , lp measurements are quite sensitive to the properties of scatterers in a planetary atmosphere , and the lp ( as a fraction of the planet s total reflected light ) can be significant ( e.g. , seager , whitney & sasselov 2000 [ = sws ] ) .
thus , lp observations of exoplanets can potentially greatly reduce the star - planet contrast and yield useful data on cegp atmospheres .
since observations will be of the exoplanet - parent star system , it is important to estimate the lp from the host star as well . in the absence of dust disks , lp from cool older stars
is dominated by the
magnetic intensification " effect from optically thick lines in a magnetic field ( leroy 1962 ) .
this magnetic lp has distinctive wavelength and rotational phase properties ( huovelin & saar 1991 ; saar & huovelin 1993 ) .
we use the models of saar & huovelin ( 1993 ) together with active region area estimates from saar ( 1996 ) to estimate the maximum broadband lp ( blp ) for selected cegp host stars .
results are in agreement ( within the large observational errors ) with observed values from leroy ( 1993 ; fig .
1 ) ; typically , blp from the host star is on the order of 10@xmath3 .
we model the planetary blp as in sws .
our fiducial model cegp is @xmath4=0.05 au from its star , with a radius of @xmath5 , and a gravity @xmath6 = 3.2 .
it is covered by a uniform mgsio@xmath7 cloud , two pressure scale heights thick , with a log - normal particle size distribution having mean radii of @xmath8 m and @xmath9 m .
the associated scattering phase and lp probability functions are given in figure 2 .
for comparison , we also computed the same functions for fe and al@xmath10o@xmath7 particles ( fig .
3 ) and a rayleigh scattering model ( fig .
4 ) . figure 2 also shows results of the monte - carlo scattering simulation ( see sws for details ) , the light , and the lp phase curves .
these results show lp measurements can better discriminate atmospheric properties than the light curve for several reasons : * the light curve largely reflects geometric ( % illumination ) effects ( fig .
2b ) , while for fractional lp ( the ratio of polarized to white light ) , these effects cancel out ( fig .
* multiple scattering smooths the light curve , blurring features distinctive to particular particles ( fig .
the lp curve , which is dominated by single scattering , is much less affected in this regard ( fig .
* lp measurements are inherently more sensitive to the composition ( e.g. , fig .
3 ) , size , and shape of particles than the light curve .
lp observations of cegps will be quite difficult ; our fiducial model suggests lp of @xmath1120 - 40% of the reflected cegp light at best ( fig .
but considerable diagnostic information is gained for this extra detection difficulty , and the contrast between star and planet is improved ( in the @xmath12 band ) from @xmath13(cegp)/@xmath13(star ) @xmath14 ( fig .
2b ) to @xmath15(cegp)/@xmath15(star ) @xmath16 ( few@xmath17)/(few @xmath18 , i.e. , a factor of @xmath19 .
the contrast improvement , and differential nature of the lp measurement , should aid detection at these low flux levels .
detection could be carried out by blp polarimetry : searching for a weak ( few % of the total ) @xmath15 signal phased to the planet s orbital period .
this method would work best in systems which are not tidally locked ( e.g. , @xmath20 and ) .
another method would be to obtain lp spectra and use some multi - line method to detect the reflected , doppler shifted spectrum ( e.g. , collier - cameron et al .
the two spectra will be quite distinctive : the star will display an lp spectrum with lines characteristic of those in strong magnetic fields ( i.e. , stokes q and u profiles ) , while the planet will display a reflected ( and linearly polarized in the process ) spectrum of the star , and thus appear as a copy of the _ unpolarized _ stellar spectrum ( stokes i ) . the striking difference between the two superposed spectra
should aid the identification of the exoplanet s signature .
sensitive polarimeters and lp - capable spectrographs on large telescopes may make such observations feasible in the near future . | we point out some advantages of making observations of extrasolar planets in linearly polarized ( lp ) light .
older cool stars have quite low levels ( @xmath0 to @xmath1 ) of fractional lp , while extrasolar planets can have relatively high fractional lp ( @xmath2 ) .
observations in lp light can therefore significantly enhance contrast between the planet and its parent star .
data on lp as a function of planetary orbital phase can be used to diagnose the properties ( e.g. , composition , size , and shape ) of the scatterers in the planetary atmosphere .
we discuss the feasibility of lp observations of extrasolar planets .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
quantum entanglement is the key element in many applications of quantum information processing ( qip ) ranging from quantum computation @xcite to communication @xcite .
the studies of entanglement represent an active line of research in modern quantum physics which so far has found only partial answers . on the practical side , entanglement is the core of a new _ paradigm _ for quantum computation @xcite and the catalyst for the performance of tasks which are impossible within the classical domain @xcite .
more recently , the relation between quantum entanglement and thermodynamical properties of macroscopic objects has become the center of an extensive study @xcite .
the existence of long - range quantum correlations between the parties of a complex many - body system is allegedly the reason for the peculiar behavior of macroscopic properties of solids @xcite . in this context
, it is intellectually stimulating and pragmatically very important to investigate and understand the existence of entanglement between macroscopic physical systems and the influences of temperature on it .
theoretically , some steps in this direction have been performed with the study of radiation pressure - induced entanglement between a movable mirror and the electromagnetic field of a cavity @xcite .
very often , the quantitative analysis of even simplified models of interaction faces important practical difficulties related to the necessity of treating non - gaussian states of continuous - variable ( cv ) systems . for non - gaussian states ,
there is a lack of objective criteria to determine whether or not entanglement is present . in these years , there have been a few proposals designed to bypass this problem , including @xcite .
unfortunately , so far a totally satisfactory answer has not been provided .
this serious limitation largely affects the extent to which an analysis of thermally entangled systems can be conducted . a way to bypass the problem
is given by the formal restriction of the hilbert space of the composite cv system to an effective discrete one @xcite , where necessary and sufficient conditions for the existence of bipartite entanglement can be used @xcite .
however , even though this strategy is theoretically exploitable , its experimental realization is extremely hard . in this paper , we investigate the entanglement properties of a class of states which have been very recently introduced by jeong and ralph @xcite in order to provide a reasonable analogy of the schrdinger s cat paradox @xcite .
a class of entangled state in @xcite can be considered as a generalization of the two - mode schrdinger cat - like state @xcite , which corresponds to an entangled state between microscopic and macroscopic systems , to a thermal mixture .
macroscopic _ part of this cat - like state is represented by the displaced thermal state of a harmonic oscillator while the microscopic part is an atomic ( or single photon ) qubit .
the other type of entangled state studied in @xcite corresponds to an entangled state between two macroscopic systems and can be generated _ through an additional conditional measurement_. such a state can be understood as a generalization of an entangled coherent state @xcite , which has been found useful for qip @xcite , to a thermal mixture . for the second type of the macroscopic entangled state , it was found that bell s inequality can be violated up to the maximum bound even though the thermal temperature becomes extremely large @xcite . on the other hand ,
the existence and degree of entanglement in the first type of the microscopic - macroscopic entangled state , which we shall call generalized cat - like state " , remains a question to be answered .
in other words , the first question that should be answered in this paper is `` how much entanglement can be generated by an interaction between microscopic quantum state ( qubit ) and macroscopic classical system ( thermal state ) without any additional process when the thermal temperature becomes extremely large ? ''
however , unfortunately , the states which we are interested in are non - gaussian mixed cv states for which the known entanglement criterions for cv states @xcite fail , leaving a degree of ambiguity which prevents any firm statement on the presence of quantum correlations . to bypass this problem , we describe an alternative theoretical method which is then corroborated by testing the _ entangling power _ of the generalized cat - like state @xcite . the test is based on the capability of a state to induce entanglement , by means of only bilocal interactions , between two initially uncorrelated qubits . as local unitary operations alone can not create entanglement , the entangling power provides a sufficient condition for the inseparability of the cv state being investigated .
therefore , another question that we naturally address in this paper is how can we transfer the entanglement generated by a microscopic - macroscopic interaction to the initially separable bipartite system of two non - interacting qubits ? " differently from the above mentioned effective - projection technique @xcite , our procedure is operative as it is immediate to design the general scenario for an experimental investigation .
this approach can be extended to other non - trivial class of non - gaussian cv states , for example a set of states which can be generated from the generalized cat - like state by measuring the microscopic part of the superposition and subsequently unitarily manipulating the remaining macroscopic part .
this results in the coherent superposition of thermally averaged two - mode states . again , the application of the entangling power test shows ability to induce entanglement in a two - qubit system .
our analysis reinforces the ideas related to the possibility of macroscopic entanglement at non - zero temperature .
we thus provide a new tool in the non - trivial problem of the quantitative analysis of the properties of non - gaussian states which is flexible enough to reveal temperature - resilient macroscopic entanglement .
the remainder of this paper is organized as follows . in section [ gattogenerale ]
, we introduce a generalized cat - like state , which is a macroscopic non - local state .
we first address the behavior of the state s variance matrix to show that the simon s cv - state entanglement criterion fails in revealing any quantum correlation in this state . on the other hand , by changing picture and restricting the attention to bidimensional subsectors of the infinite dimensional hilbert space of the cv subsystem , it is possible to infer the entanglement properties of this class of generalized cat - like states .
this approach allows us to highlight the striking effects of temperature and displacement ( in the phase space ) over the entanglement of the state , thus providing an important quantitative insight into its properties . in section [ trasferisco ] , by adopting the entangling power viewpoint , we show that entanglement can be reliably transferred to two independent qubits which have interacted with the state under investigation , giving us the possibility of constructing a highly entangled quantum channel of static qubits . on the other hand
, this approach provides us with a way to detect entanglement in the cv state simply by looking at the state of two qubits .
section [ experiment ] describes exactly this experimental protocol and a modified one which , with limited modifications to the original idea , turns out to be realistic .
finally , in section [ conclusioni ] we resume the central results of our analysis and complement it with a brief study the entanglement of thermally weighted entangled coherent states .
we show that the entangling power is an exploitable tool in this case as well and reveals the presence of entanglement within a significant range of temperature .
let us consider the interaction of a qubit of its logical states @xmath0 , initially prepared in the balanced linear superposition @xmath1 , with the macroscopic part of the generalized cat - like state here represented by a physical system @xmath2 in an initial displaced thermal state @xmath3 .
we have introduced the coherent state @xmath4 of its amplitude @xmath5 and the thermal probability distribution @xmath6 where @xmath7 is the variance of the distribution , related to the temperature of the thermal distribution by the relation @xmath8 with @xmath9 . here
, @xmath10 is the boltzmann constant , @xmath11 is the radiation temperature and @xmath12 its frequency .
we have indicated with @xmath13 the displacement of the state from the origin of the phase - space .
the relation between the variance and the mean photon number @xmath14 of the field is @xmath15 .
the interaction is ruled by the cross - phase modulation hamiltonian @xmath16 ( @xmath17 is assumed throughout the paper ) .
here @xmath18 and @xmath19 ( @xmath20 and @xmath21 ) are the annihilation and creation operators of the microscopic ( macroscopic ) system respectively and @xmath22 is the rate of non - linearity of the interaction @xcite . after an interaction time @xmath23 ,
the dynamical evolution corresponding to the hamiltonian @xmath24 gives rise to the state @xmath25 with @xmath26 .
here , @xmath27 is the cross - phase shift induced by the model @xmath24 . in order to pursue our investigation
, we find it convenient to preliminarily consider just the pure superposition @xmath28 , whose entanglement properties we want to quantitatively address .
this state has been considered as a reasonable example of schrdinger cat state and has been experimentally produced as an entangled state of internal and external degrees of freedom of a single trapped - ion @xcite .
clearly , for @xmath29 , this state carries almost one ebit of entanglement .
however , the asymmetric non - gaussian nature of this superposition makes the entanglement analysis of this state quite non - trivial .
for instance , the entanglement does not emerge at the level of the variance matrix .
the variance matrix of a two - mode cv state is defined as @xmath30 .
the variance matrix is in one - to - one correspondence with the characteristic functions of a gaussian cv state which , in turns , gives us information about the actual state of the system @xcite .
however , this is not true for the non - gaussian state @xmath28 , whose variance matrix reads @xmath31 with @xmath32 , @xmath33 and @xmath34 here we have defined @xmath35 , @xmath36 and @xmath37 . we have found that simon s separability criterion @xcite which is the most successful criterion for cv entanglement , is not able to show quantum correlations in this state . interestingly enough
, the criterion fails for @xmath38 which is the value at which one would expect the largest degree of entanglement to be found between the subsystems .
we have also checked that the conditions for inseparability established in @xcite is not verified by the state we are considering so that other ways have to be researched .
the above intuitive expectation about the entanglement at @xmath38 is confirmed by the following simple analysis .
the state @xmath28 can be written in its schmidt decomposition as @xmath39 where @xmath40 and @xmath41 with the normalization factors @xmath42)}}.\ ] ] the coefficients of the superposition ( [ schmidt ] ) are defined in terms of the @xmath43 factors as @xmath44 . in order to quantify the entanglement within a state of its density matrix @xmath45
, we use the _ negativity of partial transposition _ ( npt ) @xcite .
npt is a necessary and sufficient condition for entanglement of any bipartite qubit state @xcite .
the corresponding entanglement measure is defined as @xmath46 with @xmath47 the negative eigenvalue of the partial transposition of @xmath45 with respect to one of its parties .
the entanglement in @xmath28 can be expressed as @xmath48 in the range @xmath49 $ ] and regardless of @xmath50 , eq .
( [ gattononintegrato ] ) is a monotonously increasing function of the phase @xmath51 saturating at @xmath52 .
the saturation value of this function depends on @xmath50 and , as a function of the interaction phase @xmath51 , is reached faster as @xmath50 is increased . 0.5cm*(a ) * the schmidt decomposition of @xmath28 is a useful tool in order to test in a simple way the entanglement present in the thermally weighted superposition @xmath53 . by applying the npt criterion
, we find that @xmath54 which , involving the double gaussian integration of the square root of an exponential function is , in general , hard to compute analytically .
nevertheless , it has been possible to numerically sample the behavior of this function against the interaction phase @xmath51 , for different values of the variance @xmath7 and the displacement @xmath13 .
the results are shown in figs .
[ gattointegrato ] * ( a ) * and * ( b ) * , where @xmath55 and @xmath56 have been considered , respectively .
counterintuitively , the thermal weighting of the superposition does not smear out the entanglement in the bimodal state .
significant entanglement can be found for any value of the variance @xmath7 ( _ i.e. _ at any value of the temperature considered ) . by looking at fig .
[ gattointegrato ] * ( a ) * we see that , as the variance @xmath7 increases , more entanglement is found in the state in eq .
( [ eq3 ] ) .
moreover , as soon as a large displacement @xmath13 is considered , the differences between the results corresponding to distributions characterized by increasing variances becomes irrelevant .
the saturation value of entanglement is reached regardless of the actual value of @xmath7 ( see fig .
[ gattointegrato ] * ( b ) * ) , quicker than the case of a small @xmath13 .
this matches the results found in terms of the negativity of the wigner function associated with the state @xmath53 which , for moderate values of @xmath7 , deepens its negative part for larger values of @xmath13 .
on the other hand , this effect represents a huge practical advantage in the production of the entangled state we are studying .
indeed , it is well - known that the currently achievable rates of non - linearity are not sufficiently large to guarantee an interaction phase @xmath38 @xcite .
this would signify that , for a small @xmath7 , the produceable entanglement would be very faint ( see fig .
[ gattointegrato ] * ( a ) * ) .
however , it is sufficient to displace the thermal field more in order to achieve a significant improvement in the rate of entanglement generation .
this is in line with the idea of using weak nonlinearities for various applications @xcite .
thus , our scheme seems to be experimentally accessible even in the more realistic situation of a small kerr non - linearity .
we believe that this result is important in the context of entanglement in mesoscopic systems at finite temperatures @xcite as we have shown a model where the quantum nature of a microscopic resource is enough in order to induce intrinsically quantum features in a macroscopic object ( as our displaced thermal field ) .
these features , which are temperature - resilient , can be highlighted by partitioning the cv hilbert space into quadruplets at fixed @xmath57 and averaging over the temperature - dependent probability distribution .
in this section we address a relevant question related to the possibility of transferring the quantum correlations established in the generalized cat - like state @xmath53 to the initially separable bipartite system of two non - interacting qubits .
this problem is worth addressing under many viewpoints . _ in primis _ , would the entanglement set in the generalized cat - like state be useful if a quantum channel has to be realized ? in the context of distributed qip , this is a relevant question as it has been shown that reliable channels , exhibiting genuine quantum features , are an irremissible resource for the performances of quantum computation @xcite . on the other hand ,
the interface between hetero - dimensional systems is _ per se _ a hot topic which has attracted a considerable attention , recently , especially focused onto the transfer of entanglement from a cv system to one which
_ lives _ in a discrete hilbert space @xcite .
finally , while the tomographic reconstruction of the properties of a cv system is a hard task to perform , its discrete - variable counterpart may be accomplished with much easier experimental protocols @xcite .
it would be thus desirable to design a protocol which allows one to infer the entanglement within a cv state without relying on its direct tomography . in a discrete - variable system
this is possible through , for instance , entanglement witnesses detected with a minimal number of measurements @xcite .
such a protocol is provided by the research for the entanglement generated in a discrete - variable bipartite subsystem by the cv one via some _ local _ qubit - cv interaction .
this gives a sufficient criterion for the entanglement between the field modes @xcite .
indeed , if the two cv modes are separable , there is no way that , through simple local interactions , quantum correlations could be established .
this strategy has been already proven to be efficient in revealing the entanglement between two cv modes which have been fed into two spatially separated cavities and interacted with a pair of qubits .
the question of there being any entanglement left between the cavity field modes after the interaction with the qubits has found a positive quantitative answer through the analysis of the _ entangling power_. this is the strategy we would like to use here . of course , the quantitative results will depend on the model chosen for the local mode - qubit interaction . however , the implementation of an arbitrary model for light - matter interaction is not a trivial point . specifically , it is a physical setup - dependent issue ( meaning that certain physical setups enable the implementation of certain interaction models more straightforwardly than other ) .
it is the physical system that one has in mind which dictates the most suitable form of the local interaction .
nevertheless , recently it has been shown that the standard jaynes - cummings ( jc ) model for resonant ( as well as dispersive ) qubit - boson interaction is common to many different physical realization of a quantum device and in a range of frequency from microwave to optical @xcite .
this dresses with physical significance the choice of the interaction @xmath58 [ @xmath59 for the first field mode - qubit system and @xmath60 for the second ] to govern the local dynamics of each qubit - field mode subsystem . here
, @xmath61 is the raising operator of the @xmath62-th qubit ( ordered logical basis @xmath63 ) and @xmath64 is a coupling strength . in the qubit computational basis , this interaction is the generator of the propagator @xmath65 with @xmath66 being a rescaled interaction time . it is convenient to proceed with the entangling power test by previously considering the initial separable state @xmath67 and their joint dynamics @xmath68 .
tracing out the cv degrees of freedom , we are left with the two - qubit density matrix @xmath69\\ & = e^{-{{\left\vert\alpha\right\vert}^2}{}}\sum^{\infty}_{n=0}\frac{{\left\vert\alpha\right\vert}^{2n}}{n ! } \begin{pmatrix } a&ig_{1}&ig_{2}&-g_{3}\\ -ig^{}_{1}&b_{1}&f_{1}&if_{2}\\ -ig^{}_{2}&f^{}_{1}&b_{2}&ih_{1}\\ -g^{}_{3}&-if^{}_{2}&-ih^{}_{1}&b_{3 } \end{pmatrix } \end{aligned}\ ] ] with the explicit expressions for the density matrix elements @xmath70{\mbox c}^{2}_{n}(\tau),{\hskip0.2cm}g_{1 } = \frac{1+e^{-i\varphi}{\mbox c}^{2}_{1}(\tau)}{\sqrt{n+1}}\alpha^ { * } { \mbox c}_{n}(\tau){\mbox s}_{n}(\tau),\\ & g_{2}=e^{-i\varphi{n}}{\mbox s}_{0}(\tau ) { \mbox c}^{2}_{n}(\tau),\hskip0.2cm{g_{3}}=\frac{e^{-i\varphi(n+1 ) } { \mbox s}_{0}(\tau)}{\sqrt{n+1}}\alpha^{*}{\mbox s}_{n}(\tau){\mbox c}_{n}(\tau),\\ & b_{1}=\frac{1+\cos^{2}\tau}{n+1}{\left\vert\alpha\right\vert}^{2}{\mbox s}^{2}_{n } ( \tau),\hskip0.2cm{f_{1}}=\frac{e^{-i\varphi{n}}{\mbox s}_{0}(\tau ) } { \sqrt{n+1}}\alpha{\mbox c}_{n}(\tau){\mbox s}_{n}(\tau),\\ & { b_{2}}={{\mbox s}^{2}_{0}(\tau)}{\mbox c}^{2}_{n}(\tau ) , \hskip0.2cm{f}_{2}=\frac{e^{-i\varphi(n+1)}{\mbox s}_{0}(\tau ) } { n+1}{\left\vert\alpha\right\vert}^2{\mbox s}^{2}_{n}(\tau)\\ & b_{3}=\frac{{\mbox s}^{2}_{0}(\tau)}{n+1}{\left\vert\alpha\right\vert}^2 { \mbox s}^{2}_{n}(\tau),\hskip0.2cm{h}_{1}=\frac{e^{-i\varphi } { \mbox s}^{2}_{0}(\tau)}{\sqrt{n+1}}\alpha^ { * } { \mbox c}_{n}(\tau){\mbox s}_{n}(\tau ) .
\end{split}\ ] ] here we have used the notation @xmath71 .
( [ nonintegratedelements ] ) have now to be integrated over the thermal weighting function in order to gather the reduce density matrix of the qubits after the interaction with @xmath53 . by exchanging the summation in eq .
( [ matricedensita ] ) and the gaussian integrals and using the binomial formula , evaluating the integral over @xmath72 results in the introduction of the functions @xmath73 where @xmath74 is the symbol for the binomial coefficient and @xmath75 is the gamma function of its argument @xmath76 but in a density matrix of the same form of eq .
( [ matricedensita ] ) .
the explicit eigenvalues of the integrated density matrix are not very informative as their form is quite complicated @xcite . nevertheless , it is possible to evaluate them numerically for a considerable range of @xmath7 s and @xmath13 s .
some striking results are presented in fig .
[ trasferito ] for @xmath77 ( corresponding to an average photon number @xmath78 ) and displacement @xmath56 ( @xmath79 ) and @xmath80 ( @xmath81 ) .
it is evident that the same effect shown in fig .
[ gattointegrato ] * ( b ) * , regarding the increase in @xmath13 , is achieved in the amount of entanglement transferred to the initially separable qubits .
it is particularly interesting to stress the considerably large amount of transferred entanglement . in this situation a significantly entangled two - qubit channel
can thus be constructed by exploiting the cv entangler represented by @xmath53 .
the usefulness of such a channel has to be quantified by including , in this analysis , the mixedness properties of the two - qubit state .
indeed , it is known that , for instance , a bipartite mixed state becomes useless for quantum teleportation whenever its linearized entropy @xmath82 $ ] exceeds @xmath83 @xcite .
@xmath84 is a good measure for mixedness which ranges from 0 ( for pure states ) to 1 ( for maximally mixed ones ) .
the calculation of the linearized entropies corresponding to the examples reported in fig .
[ gattointegrato ] and their comparison to the threshold for quantum teleportation is presented in fig .
[ purezza ] , showing the high - entanglement bumps which corresponds to a sufficiently pure state .
the channel , in this case , can be faithfully used for quantum teleportation protocols .
it is worth stressing the non - monotonic behavior of the entangling power against the entanglement initially present in the cv state .
this is a well - known feature of this entanglement test @xcite and is a result of the interference of the rabi flopping induced by the distribution of photons characterizing each state having a specific value of @xmath7 and @xmath13 . as a specific instance of this peculiar non - linear relation between the amount of entanglement initially contained in the cv entangler and what is finally found in the two - qubit reduced state
, we consider the case of @xmath85 with @xmath86 .
the analysis conducted in section [ gattogenerale ] results in an entanglement which rapidly approaches @xmath87 , as shown in the inset of fig .
[ insieme ] * ( a ) * , a behavior quite consistent with the trend shown in fig .
[ gattointegrato ] * ( b)*. on the other hand , the entanglement transferred to the two - qubit state is never larger that @xmath88 , as evidenced by fig .
[ insieme ] * ( a ) * , which is smaller than the entanglement transferred to the qubits for @xmath77 , @xmath56 .
we believe this result is still extremely significant as the high value of @xmath7 considered here ( corresponding to @xmath89 photons ) shows that quite a considerable entangling power is in a high - temperature generalized cat - like state . coming back to the example of teleportation we have addressed
, the corresponding two - qubit channel is still useful as its mixedness , at @xmath90 , can be well - below the threshold ( see fig .
[ insieme ] * ( b ) * ) .
in the previous section we have theoretically addressed the question of how to infer the entanglement between the subsystems of a generalized cat - like state . on a practical side , however , if our attention is restricted to the case of travelling - wave fields , some problems have to be faced . in the first place ,
we have already stressed the difficulties related to the achievement of large - enough rates of non - linearity . in section [ gattogenerale ]
we have outlined a strategy to highlight the quantum correlations in generalized cat - like states generated with smaller interaction phases @xmath51 s . nevertheless , the realization of even small @xmath51 s would pass , for instance , through the use of long optical fibers where the effect of dephasing channel is still an unknown issue @xcite . in the second place , the realization of the entangling power test would require a demanding ( even if foreseeable ) experimental setup . in fig . [ esperimento ]
we sketch the scheme of the idea put forward in section [ trasferisco ] .
the generalized cat - like state being embodied by two travelling field modes following the general scheme of fig .
[ schema ] , would feed two optical cavities , each crossed by a two - level atom ( or containing an integrated quantum dot , as an alternative ) .
the passage of the atoms would set each cavity field mode in resonance with the external driving mode , which will penetrate the cavity and interact with the two - level system through the dynamics described in eq .
( [ jcrisolto ] ) .
each step required by this experimental setup has been independently demonstrated ( see ref .
@xcite and references within for a detailed discussion ) and we will not comment further on them . on the other hand , the main point of this section would be the introduction of a simplified scheme which is still able to highlight the important features of our study with a much more realistic physical setup .
in our idea , the most striking difference with respect to the previous configuration is the presence of a single cavity and no travelling light field .
indeed , it is well - known @xcite ( see also paternostro _ et al . _
@xcite ) that the dispersive interaction between a single two - level system and a cavity mode would lead to an effective hamiltonian reading @xmath91 where the microscopic subsystem @xmath92 has been identified with the qubit @xmath93 whose spectrum has been rescaled so that its ground state has zero energy .
we have indicated @xmath94 with @xmath95 the large atom - field detuning .
this model can be achieved by assuming a qubit @xmath87-cavity mode jc interaction in the presence of a static electric field .
the induced stark shift on the atomic levels creates a detuning @xmath96 such that eq .
( [ hamiltonianaeffettiva ] ) holds .
the propagator generated by this hamiltonian , starting from the state @xmath97 , creates a generalized cat - like state between a thermal displaced state of the cavity field mode and a two - level atom .
the displacement of the thermal field can be effectively performed via the modification to the cavity refractive index induced by the presence of the two - level atom .
this acts as an effective non - linear intra - cavity medium which shifts the resonance frequency of the cavity and can inject an external coherent state of its amplitude @xmath98 .
the effective displacement would result in @xmath99 with @xmath100 the time - of - flight for atom @xmath93 crossing the cavity @xcite .
we are thus considering a hybrid cat - like state where the cv and the qubit parts have clearly distinguished physical embodiments .
as soon as qubit @xmath93 leaves the cavity , a second qubit @xmath101 , identical to the first , crosses the cavity in absence of the stark electric field ( _ i.e. _ the interaction model is the standard jc one ) . by interacting with the cavity field
, this process leads to an entangled state of subsystems @xmath102 and @xmath101 .
tracing out the cavity field , we are left with a reduced two - qubit state whose entanglement represents , again , a sufficient condition for the entanglement in the generalized cat - like state .
we are thus describing a unilateral entanglement - transfer process for a modified entangling power test @xcite .
obviously , the quantitative analysis reported in section [ trasferisco ] are not applicable to the present modified protocol and the effectiveness of the entanglement transfer process should be re - tested . however , this is a straightforward process which may be derived directly from what is presented in eqs .
( [ matricedensita ] ) and ( [ nonintegratedelements ] ) .
we find that the structure of the reduced density matrix @xmath103 is similar to eq .
( [ matricedensita ] ) .
the calculation of the negativity of partial transposition gives us , once again , useful information about the entanglement within this two - qubit state , transferred after the @xmath104 interaction .
the results are shown in fig .
[ insieme2 ] * ( a ) * for the same parameters as in fig .
[ trasferito ] . at the same time , in fig .
[ insieme2 ] * ( b ) * we reprise our teleportation example and present the analysis in terms of linearized entropy .
it is evident that a good degree of entanglement can still be found ( despite the high peaks of fig .
[ trasferito ] have disappeared ) , with a quite low degree of mixedness of the associated quantum channel also for this modified entangling test protocol . obviously , this represents a huge advantage in terms of experimental feasibility as all the ingredients for the entangling power test are within the current state of the art .
in investigating the entangled properties of mesoscopic systems at finite temperature , in the quantum optics domain , we often face some difficulties related to a certain lack of analytical tools .
this has strongly limited the insight we could gain about the specific behavior , in terms of quantum correlations , of particular classes of states ( especially non - gaussian states ) .
these difficulties have forced the researchers to look for ways to bypass the problem and infer the quantumness of a state .
it appears that such the alternative tests are based on the restriction of the analysis from infinite to finite dimensional hilbert spaces .
this can be done by effective ( formal ) projection of a cv state onto bipartite hilbert sectors ( as done in refs .
@xcite ) as well as looking at the amount of non - classical correlations that a given state under investigation can transfer to mutually non - interacting qubits @xcite . in this paper
we have performed some interesting steps along these directions , investigating the entanglement properties of a class of recently introduced ( highly non - gaussian ) generalized cat - like states @xcite .
our approach has been twofold . in the first place
, we have calculated the entanglement of the cat - like state by partitioning the hilbert space of the combined system into manifolds spanned by the states of an effective bipartite two - level system .
the entanglement of the whole state , then , is the result of a thermal average of the quantum correlations within each manifold .
the qualitative features highlighted by this approach have been confirmed , in the second place , by testing the entangling power of the generalized cat - like state with respect to two initially separable qubits .
this second approach is particularly useful under a practical as well as a theoretical viewpoint .
it has revealed that powerful experimental criteria exist , for the entanglement investigation , beyond the cv criteria for inseparability @xcite .
our study is , thus , a particularly illuminating case of this second , more operative scenario . as an additional example of the effectiveness of the entangling power test , here we briefly assess the problem of the entanglement in the class of states generated by the entire protocol described by the setup in fig .
[ schema ] ( dashed - box included ) .
the detection of mode @xmath92 is performed onto the basis @xmath105 ( with @xmath106 ) . conditioned on finding @xmath107 , the action of the beam - splitter which mixes mode @xmath2 to vacuum gives rise to the state ( we assume @xmath38 ) @xmath108 where @xmath109 and @xmath110 are proper normalization factors . in order to fix the ideas ,
let us focus on the case of @xmath111 being found with @xmath112 , for which the expression of the variance matrix is particularly straightforward , reading @xmath113 it is easy to check that simon s separability condition is not violated by eq .
( [ doposplitter ] ) so that no firm statement about the entanglement properties of this state can be made .
however , following the lines depicted in ref .
@xcite , it is straightforward to show that eq .
( [ doposplitter ] ) violates a bell - chsh inequality @xcite regardless of @xmath7 and @xmath13 , thus revealing an inherent quantum entanglement of the state .
this result is confirmed by the application of the entangling power test which shows that entanglement ( as large as @xmath114 for v around @xmath115 and @xmath116 ) can be transferred to two mutually non - interacting qubits using the symmetric scheme of section [ trasferisco ] .
this second example completes and complements our investigation which , we believe , represents an interesting example of the possibilities of revealing quantum non - locality in non - trivial situations of mesoscopic superpositions at high temperature .
+ we acknowledge useful discussions with prof . p. l. knight and dr .
j. fiurasek and communication with a. ferreira .
this work has been supported by the leverhulme trust ( ecf/40157 ) , the australian research council , the uk epsrc and the korea research foundation ( 2003 - 070-c00024 ) .
+ d. deutsch , proc .
a * 400 * , 97 ( 1985 ) ; d. gottesman and i.l .
chuang , _ nature _ * 402 * , 390 ( 1999 ) ; j. eisert , k. jacobs , p. papadopoulos , and m. b. plenio , _ phys .
a _ * 62 * , 052317 ( 2000 ) . j. eisert , m. wilkens , and m. lewenstein , _ phys .
_ * 83 * , 3077 ( 1999 ) ; l. vaidman , y. aharonov , and d. z. albert , phys . rev
58 , 1385 ( 1987 ) ; m. paternosto , m. s. tame , and m. s. kim , _ quant - ph/0509066 _ ( 2005 ) .
a. peres , _ phys
lett . _ * 77 * , 1413 ( 1996 ) ; m. horodecki , p. horodecki , and r. horodecki , _ phys .
lett . a _ * 223 * , 1 ( 1996 ) ; j. lee , m. s. kim , y. j. park , and s. lee,_j .
_ * 47 * , 2151 ( 2000 ) .
van enk and o. hirota , phys .
a * 64 * , 022313 ( 2001 ) ; h. jeong , m.s .
kim , and j. lee , phys .
a * 64 * , 052308 ( 2001 ) ; x. wang , phys .
a 64 , 022302 ( 2001 ) ; t.c .
ralph , w.j .
munro , and g. j. milburn , proceedings of spie * 4917 * , 1 ( 2002 ) ; h. jeong and m. s. kim , _ phys . rev .
a _ * 65 * 042305 ( 2002 ) ; t.c .
et al . _ ,
a 68 , 042319 ( 2003 ) ; nguyen ba an , phys .
a 68 , 022321 ( 2003 ) ; m. paternostro , m. s. kim , and p. l. knight , _ phys . rev .
a _ * 71 * , 022311 ( 2004 ) .
m. paternostro , g. falci , m. s. kim , and g.m .
palma , phys .
b * 69 * , 214502 ( 2004 ) ; i. wilson - rae and a. imamoglu , _ phys .
b _ * 65 * , 235311 ( 2002 ) ; f. marquardt and c. bruder , _ phys . rev .
b _ * 63 * , 054514 ( 2001 ) ; j. m. taylor , g. giedke , h. christ , b. paredes , j. i. cirac , p. zoller , m. d. lukin , and a. imamoglu , _ cond - mat/0407640 _ ( 2004 ) .
properly arranged matter - light interactions , in the regime of electromagnetically - induced - transparency , can produce a cross - phase modulation with huge non - linear rates ( see m. paternostro , m. s. kim , and b. s. ham , _ phys .
a _ * 67 * , 023811 ( 2003 ) for a protocol which generates entangled coherent states ) .
however , as far as the authors are concerned , there has not been a satisfactory experimental implementation yet .
j. fiurek l. mita , jr . and r. filip , phys . rev .
a * 67 * , 022304 ( 2003 ) ; h. jeong , phd thesis , queen s university belfast , uk ( 2003 ) ; s. d. barrett _
_ , phys . rev .
a * 71 * , 060302 ( 2005 ) ; k. nemoto and w. j. munro phys .
93 * , 250502 ( 2004 ) .
h. jeong , phys .
a * 72 * , 034305 ( 2005 ) ; in this reference , it has been numerically shown that , under proper conditions , decoherence may be overcome in optical fibers with small non - linearities for generation of macroscopic entanglement .
j. eisert , k. jacobs , p. papapdopoulos , and m. b. plenio , _ phys .
a _ * 62 * , 52317 ( 2000 ) ; d. collins , n. linden , and s. popescu , _ phys .
rev . a _ * 64 * , 032302 ( 2001 ) ; m. paternostro , m. s. kim , and g. m. palma , _ j. mod
. opt . _ * 50 * , 2075 ( 2003 ) .
m. paternostro , w. son , and m. s. kim , _ phys .
lett . _ * 92 * , 197901 ( 2004 ) ; b. kraus and j. i. cirac , _ phys .
lett . _ * 91 * , 013602 ( 2004 ) ; m. paternostro , m. s. kim , and w. son , _ int . j. quant .
inf . _ * 3 * , 213 ( 2005 ) .
englert , n. sterpi , and h. walther , _ opt .
commun . _ * 100 * , 526 ( 1993 ) ; i. g. luttenbach and l. davidovich , _ phys .
lett . _ * 78 * , 2547 ( 1997 ) ; m. s. kim and j. lee , _ phys .
a _ * 61 * , 042102 ( 2000 ) . | we investigate entanglement properties of a recently introduced class of macroscopic quantum superpositions in two - mode mixed states .
one of the tools we use in order to infer the entanglement in this non - gaussian class of states is the power to entangle a qubit system .
our study reveals features which are hidden in a standard approach to entanglement investigation based on the uncertainty principle of the quadrature variables .
we briefly describe the experimental setup corresponding to our theoretical scenario and a suitable modification of the protocol which makes our proposal realizable within the current experimental capabilities . |
in a classic paper@xcite on the statistical mechanical theory of solutions in equilibrium mcmillan and mayer ( hereafter mm ) derived the following two results ( or theorems ) : * mmi * : for a solution with any number of components , the solvent variables in the grand canonical partition function can be formally `` integrated out '' leaving an effective partition function for the solutes only . the total effective solute potential entering the effective solute partition function
is the true vacuum solute interaction potential plus a contribution induced by the integrated - out solvent .
the total effective solute potential can be decomposed into a sum of one - body , two - body , etc , effective solute potentials .
these effective solute potentials are potentials of mean force ( pmfs ) for the solute molecules .
* mmii * : using the results of mmi , the osmotic pressure @xmath9 of a dilute solution can be developed in a virial series in the solute concentration , analogous to the gas virial expansion for the pressure @xmath7 in powers of the gas density at low density .
the statistical mechanical expressions for the osmotic pressure virial coefficients have exactly the same form as the expressions for the gas pressure virial coefficients , with solute pmfs replacing gas vacuum potentials .
a few years after the mm paper appeared , another classic paper @xcite on the statistical mechanical theory of solutions in equilibrium was published by kirkwood and buff ( hereafter kb ) .
this paper gives molecular expressions for specific solution macroscopic properties , i.e. , the isothermal compressibility , the partial molecular volumes , and the inverse osmotic susceptibilities , in terms of what are now called kb integrals .
the latter are the integrals @xmath10 over the solution pair correlation functions @xmath11 , where @xmath12 and @xmath13 label the molecular species , e.g. , @xmath0 and @xmath1 for a two - component solution , and @xmath14 is the separation between the centers of an @xmath15 pair of molecules .
the pair correlation functions themselves can in principle be determined @xcite both theoretically , from liquid state simulations and analytic methods using model intermolecular potentials , and experimentally , from x - ray and neutron diffraction experiments .
kb theory has been simplified , extended , reviewed extensively @xcite , and widely applied @xcite .
one leading practitioner has expressed the opinion `` i believe that in all regards the kb theory is immensely superior to the mm theory , as i hope to convince the reader ... '' .
we hope to convince the reader that , with current and future developments , mm theory could be far more useful than is now the case , although there is no doubt that at the present time , particularly since the introduction of the ben naim kb inversion technique@xcite , which expresses the kb integrals in terms of the specific macroscopic properties mentioned above , kb theory is used much more .
in contrast to kb theory , mm theory has not been greatly simplified or extended , and has been reviewed @xcite infrequently .
it also has not been widely applied , with one exception : the osmotic pressure virial expansion .
for a two - component solution with dilute solute ( species a ) in a solvent ( species b ) , the mm result for the osmotic pressure @xmath9 as a virial series in solute number density @xmath16 , where @xmath17 is the number of solute molecules and @xmath6 is the system volume , is @xmath18~ , \ ] ] where @xmath19 is boltzmann s constant , @xmath8 the temperature , and @xmath20 the solvent chemical potential .
the first virial term @xmath21 is the famous vant hoff solute one - body term which is universal , i.e. , independent of all intermolecular interactions , and has the same form as the ideal gas pressure .
solution properties whose first virial term depends only on the solute number density and is independent of solute molecular details and interactions are called colligative .
the second term , with coefficient @xmath22 , is @xmath23 and takes into account the contribution from pairs of interacting solute molecules in the solvent , and the mm molecular expression for this osmotic second virial coefficient is @xmath24 where @xmath25 , and @xmath26 is the pmf for a pair of solute molecules at infinite dilution in the solvent , which is written as an explicit function of solute - solute separation @xmath14 and which also depends implictly on solvent chemical potential and temperature .
note that ( [ eq1.2 ] ) has the same form as the expression for the gas second virial coefficient , with pair pmf @xmath26 in place of the vacuum pair potential @xmath27 which occurs in the gas second virial coefficient expression .
this is an example of mmii above .
equation ( [ eq1.1 ] ) has long been used @xcite by polymer , biopolymer and colloid solution researchers to obtain molecular weights of macromolecules ( from the first virial term @xmath21 ) and to study macromolecule solute - solute interactions ( using the second virial term ) .
the osmotic second virial coefficient has also been used in discussions of the hydrophobic interaction of small nonpolar solutes in water@xcite , and we give an example in the concluding section .
unfortunately the mm paper @xcite is very hard to read and this is likely the main reason for the relative neglect of mm theory compared to kb theory . indeed , as another leading practitioner @xcite has put it `` ... somewhere , hidden in an impenetrable jungle of notation in the 1945 paper of mcmillan and mayer , must be the potential distribution theorem ! '' as it turns out the widom potential distribution theorem@xcite ( also called the particle insertion theorem ) is not explicitly written down by mm ( hill@xcite comes close in his reformulation of mm theory using the semi - grand canonical ensemble - see remarks following ) .
we find that the widom expression for the excess chemical potential of a solute molecule in an otherwise pure solvent is the first virial term in the virial expansion of the solute chemical potential @xmath28 ; the latter expansion is obtained by differentiating the semi - grand potential virial series with respect to solute number @xmath17 .
the semi - grand potential is the thermodynamic potential generated by the semi - grand canonical ensemble ( grand with respect to solvent only ) , the ensemble we employ in this paper .
mm did not obtain the semi - grand potential series because they worked with the full grand canonical ensemble which generates the grand potential . because the grand potential is proportional to the pressure
, mm obtained the virial series ( [ eq1.1 ] ) for the osmotic pressure directly , and because this series has a trivial first virial term , a nontrivial first virial term which occurs for non - colligative properties such as the solute chemical potential is not visible explicitly .
some of the difficulty of the mm paper has been reduced by hill @xcite who separates better the quantum and classical discussions , does not include solid solutions in the discussion , discusses explicitly a two - component classical liquid solution , and introduces a semi - grand canonical ensemble to derive mmi . in this paper
we greatly simplify the presentation further by working with the semi - grand canonical ensemble in a new way to derive mmi as described in the next paragraph , and by using a simple and nongraphical virial expansion method @xcite to derive mmii from mmi .
we also extend mm theory by deriving virial expansions for other properties such as the enthalpy of mixing , which are needed to analyze recent isothermal titration calorimetry ( itc ) experiments @xcite on dilute peptide aqueous solutions . in the next section , for a classical two - component liquid solution
we introduce a semi - grand canonical ensemble which is grand canonical with respect to the solvent(b ) and canonical with respect to the solute(a ) ; the complete set of independent variables is @xmath2 .
these state conditions are the natural ones for analysis of osmotic pressure experiments where the solvent and solution are in contact through a semi - permeable membrane ( so that @xmath20 is fixed ) and the solution pressure is measured as a function of the controlled solute number @xmath17 at fixed solution volume @xmath6 and temperature @xmath8 .
various semi - grand ensembles have been discussed in the literature , starting with stockmayer @xcite ( see also @xcite ) who , in order to discuss light scattering from multicomponent dilute solutions , introduced a constant pressure ensemble which is grand with respect to the solutes but not with respect to the solvent .
the semi - grand canonical ensemble we employ was introduced by hill @xcite .
this ensemble was also used to derive mmi by kodytek@xcite , who , however , needlessly complicated the derivation by introducing the total solute pmf in the canonical ensemble , and by dijkstra et al @xcite .
our discussion of mmi is similar in spirit to those of hill@xcite and dijkstra et al@xcite , but simpler , because rather than focussing on the semi - grand potential for the solution , we focus instead on the difference between the solution and solvent semi - grand potentials , which can be expressed in terms of the logarithm of the ratio of the solution and solvent semi - grand partition functions . as with other such problems in statistical mechanics involving the difference in free energies for two systems with some state conditions in common , such as zwanzig s thermodynamic perturbation theory@xcite , van kampen s gas virial expansion method@xcite , and widom s particle insertion theorem@xcite , the ratio of partition functions generates automatically the required solvent - to - solution change in the form of a normalized average , and this average can be expressed in terms of the total solute pmf and its one - body , two - body , etc , components from simple physical considerations .
thus , for example , in contrast to reference @xcite , in discussing mmi we do not need graphical methods , and a zero - body term in our total solute pmf does not arise .
this will become clearer in section iii .
a similar semi - grand ensemble but with fixed pressure rather than fixed volume has been introduced for computer simulations of multicomponent solutions @xcite . in section iii
we introduce the solute pmfs , first @xmath29 for the complete set of @xmath17 solute molecules as a function of their configuration @xmath30 , and then reduced pmfs for isolated groups of one , two , three , etc , solute molecules alone in the solvent , i.e. , @xmath31 , etc .
, respectively , where @xmath32 is the position of solute molecule @xmath33 , @xmath34 the separation of solute molecules @xmath33 and @xmath35 , etc .
these are related in the usual way to the corresponding solute reduced distribution and correlation functions , and to mean forces on solute molecules under the proper constraints @xcite .
the semi - grand canonical ensemble partition function introduced in section ii is re - expressed in terms of an effective solute - only canonical ensemble partition function involving @xmath29 .
the total solute pmf or effective potential @xmath29 is then decomposed into a sum of its one - body , two - body , etc , components , involving @xmath36 , etc .
this is the semi - grand canonical ensemble version of mmi . for the reasons discussed
, we believe this derivation of mmi to be much simpler than those presented previously .
the virial expansions are introduced in section iv .
starting with the result of section iii expressing the semi - grand canonical partition function in terms of the effective solute - only canonical ensemble partition function involving the effective solute potential @xmath29 and its decomposition into one - body , two - body , etc , components , we assume a dilute solution and derive a virial series in @xmath37 for the difference between the solution and solvent semi - grand potentials , @xmath38 , from which we immediately obtain the virial series for the osmotic pressure @xmath39 , where @xmath7 and @xmath40 are the solution and solvent pressures , both at chemical potential @xmath20 .
this is the semi - grand canonical version of mmii .
our derivation is much simplified compared to that of mm by using a simple nongraphical canonical ensemble virial expansion method developed for gases @xcite .
another advantage , mentioned above , is that the virial series for @xmath41 , unlike that for @xmath9 , has a nontrivial first virial coefficient , which is needed for other non - colligative properties such as solute chemical potential and solubility , and we extend mm theory and derive virial expansions for various such properties including the enthalpy of mixing .
for some of the extensions , we also change independent variables from semi - grand canonical variables @xmath2 to those more relevant to the particular experiment , e.g. , @xmath3 for the enthalpy of mixing experiment . in this section
we also clarify the relation between mm theory and widom s potential distribution theory .
the paper concludes with a simple example , and some remarks on ongoing work by our group .
we first consider the classical statistical mechanics of a binary liquid solution of @xmath17 identical solute molecules and @xmath42 identical solvent molecules in equilibrium at temperature @xmath8 in a volume @xmath6 .
for notational simplicity we consider monatomic molecules but , as we indicate later , the final results are immediately generalizable to polyatomic molecules which may be nonrigid .
we write the hamiltonian @xmath43 as @xmath44 where @xmath45 denotes the momentum of molecule @xmath46 , @xmath47 the mass of an @xmath0-molecule , @xmath48 the configuration of the @xmath0-molecules , and assuming pairwise additive vacuum potentials , the total potential energy @xmath49 is given by @xmath50 here @xmath51 and @xmath52 denote the total solute - solute , solvent - solvent and solute - solvent potentials , respectively , and @xmath53 the separation between molecules @xmath46 and @xmath54 . the assumption that @xmath55 , etc , are pairwise additive is again for notational convenience , and without this assumption the derivation to follow goes through with no substantive change . with canonical ensemble conditions of fixed @xmath56 , the canonical ensemble partition function @xmath57 for the solution is @xmath58 where @xmath59 is planck s constant , and here and throughout the paper the configurational integrals are over the volume @xmath6 .
note that @xmath57 is dimensionless . because the momentum integrals in ( [ eq2.3 ] ) are gaussian , the kinetic part of the partition function can be evaluated analytically giving @xmath60 where @xmath61 is the thermal wavelengh of an a - molecule , @xmath62 the helmholtz free energy @xmath63 of the solution is calculated from the canonical partition function using @xmath64 as described earlier , mm worked in the grand canonical ensemble @xcite .
the grand canonical partition function @xmath65 is related to the canonical partition function @xmath66 by the double discrete laplace transform @xmath67 and the corresponding grand potential @xmath68 is related to the helmholtz free energy @xmath69 by the double legendre transform @xmath70 ( in thermodynamic relations like ( [ eq2.8 ] ) we should strictly use the averages @xmath71 and @xmath72 , but since the relative fluctuations in @xmath17 and @xmath42 are negligible for macroscopic systems no harm will come from using the simpler notation @xmath17 and @xmath42 for the average numbers . ) following hill@xcite we find it simpler to work in a semi - grand canonical ensemble ( grand with respect to solvent only ) , with semi - grand canonical partition function @xmath73 defined by the single laplace transform of the canonical partition function @xmath66 , @xmath74 and corresponding semi - grand potential @xmath75 related to the helmholtz free energy @xmath69 by the single legendre transform @xmath76 as discussed in the introduction , @xmath77 are the natural control variables for analysis of osmotic pressure experiments . from ( [ eq2.10 ] ) and the standard expression for the differential @xmath78 , @xmath79 where @xmath80 is the entropy , we find for @xmath81 @xmath82 from ( [ eq2.12 ] ) we obtain the pressure @xmath7 and solute chemical potential @xmath28 using @xmath83 we consider the ratio of solution / pure solvent semi - grand canonical partition funtions at the same @xmath84 state conditions , i.e. @xmath85 , where the pure solvent semi - grand partition function @xmath86 is also a grand partition function and could just as well be denoted @xmath87 . from ( [ eq2.9 ] ) and ( [ eq2.4 ] )
we can express these two partition functions as @xmath88 @xmath89 where @xmath90 . from the last two equations we see that the ratio @xmath85 can be written as @xmath91 where the solvent grand - canonical average @xmath92 of a quantity which depends on @xmath1-coordinates but not @xmath1-momenta is @xmath93 in the averaging @xmath94 in ( [ eq2.16 ] ) using ( [ eq2.17 ] ) , the solute @xmath0-coordinates are held fixed , and the solvent @xmath1-coordinates are averaged over with normalized boltzmann weight @xmath95 . in the final step we multiply and divide ( [ eq2.16 ] ) by @xmath96 and
write it as @xmath97 where @xmath98 denotes an unweighted average over the @xmath0-coordinates , @xmath99 the solution - solvent difference of semi - grand potentials is @xmath100 . from ( [ eq2.18 ] ) , using the standard stirling approximation for large numbers @xmath101 , we find for the dimensionless difference @xmath102 the expression @xmath103 where the ideal gas value of the solute helmholtz free energy is given by @xmath104 with @xmath16 the solute number density .
the inner solvent average @xmath105 in ( [ eq2.20 ] ) is a dimensionless positive function of the fixed @xmath0-configuration @xmath30 , which we write as @xmath106 which defines a solvent - induced effective @xmath0-potential @xmath107 .
we form the total effective @xmath0-potential @xmath29 by adding the vacuum solute potential @xmath108 , @xmath109 where the total vacuum @xmath0-potential @xmath55 is defined in ( [ eq2.2 ] ) for our model .
the complete average in ( [ eq2.20 ] ) then has the form @xmath110 the right side of ( [ eq3.3 ] ) has the form of a solute - only canonical ensemble configurational partition function , since the solvent @xmath1-coordinates have been integrated out in defining the effective solute potential @xmath29 .
this is the semi - grand canonical ensemble version of the mmi theorem discussed in the introduction . in terms of the total effective solute potential @xmath111 , using ( [ eq3.3 ] ) and ( [ eq2.19 ] ) we re - write the semi - grand potential ( [ eq2.20 ] ) in terms of the solute average @xmath112 , @xmath113 the total solute effective potential @xmath29 is the pmf for the whole set of solute a molecules ( see discussion later in this section ) . in our model , for simplicity the total solute vacuum potential @xmath55 has been assumed to be pairwise additive , @xmath114 . in contrast , the total solute effective potential @xmath29 will necessarily contain one - body , two - body , three - body , etc , terms : @xmath115 this is a cluster decomposition of the total effective potential @xmath116 .
we give next an intuitive derivation of ( [ eq3.4 ] ) , with explicit expressions for the one- and two - body terms , and a few more details are given in appendix a. at very low solute concentrations ( infinite dilution ) , we neglect pairs , triplets , etc , of @xmath0 molecules . in the solvent average @xmath94 with fixed @xmath0-configuration
, we assume the @xmath0-molecules are all isolated from each other and write the average as @xmath117 where the contibutions @xmath118 , etc , to @xmath119 from molecules @xmath120 , etc , in the average ( [ eq3.5 ] ) have been assumed uncorrelated due to the assumption of far - apart @xmath0-molecules ; in the pairwise additive vacuum potential model these contributions are given explicitly by @xmath121 , etc .
the last step in ( [ eq3.5 ] ) simply defines the effective solute one - body potentials @xmath122 , etc , by @xmath123 the average in ( [ eq3.6 ] ) involves the interaction with the solvent of just one solute molecule @xmath33 fixed at position @xmath124 , averaged over the solvent configurations . for @xmath17 identical solute molecules in a uniform solvent the effective one - body potentials @xmath125 are all equal and independent of position , so that the last expression in ( [ eq3.5 ] ) is the product of @xmath126 identical constant factors . the total one - body pmf term in ( [ eq3.4 ] ) , @xmath127 , is thus given by @xmath128 where @xmath129 is the constant one - body potential defined by ( [ eq3.6 ] ) . equation ( [ eq3.6 ] ) defining @xmath129 was first derived by widom @xcite with potential distribution theory using the canonical ensemble , and its physical significance is discussed further later .
we now consider a slightly increased solute concentration so that solute - solute pair interactions , but no triplets etc , also need to be considered .
suppose the fixed @xmath0-configuration has the @xmath130 pair and a few other solute pairs close together , and the other solute molecules isolated .
just as ( [ eq3.5 ] ) factored into a product of uncorrelated one - body terms at infinite dilution , now the average @xmath131 will factor into a product of uncorrelated two - body terms such as @xmath132 and other such terms arising from the close @xmath0-pairs , and constant factors coming from the uncorrelated and isolated @xmath0-molecules .
we denote the @xmath130 pair term by @xmath133 according to ( [ eq3.8 ] ) we now fix two solute molecules @xmath33 and @xmath35 at positions @xmath124 and @xmath134 in the otherwise pure solvent , and average the exponential of @xmath135 over solvent configurations . if @xmath33 and @xmath35 are far apart as assumed previously , the result is @xmath136 as before . when @xmath33 and @xmath35 are close together , the result is modified and denoted by the right side of ( [ eq3.8 ] ) , thereby defining a solvent - induced contribution @xmath137 to the total effective @xmath138 pair potential @xmath139 , where @xmath140 is the vacuum pair potential .
thus @xmath141 is the extra contribution , beyond @xmath142 , to the solvent - induced effective potential of solute molecules @xmath33 and @xmath35 , and is therefore normalized such that @xmath143 for @xmath144 , the same normalization as the vacuum pair potential @xmath145 .
the total effective @xmath146 pair potential @xmath147 thus has the same normalization .
the total effective pairwise solute potential is a sum of terms like @xmath148 as written in ( [ eq3.4 ] )
. equation ( [ eq3.8 ] ) is easily related to one derived in potential disribution theory @xcite for the solute - solute pair correlation function at infinite dilution @xmath149 , using ( see below ) @xmath150 .
this brief derivation of the effective pairwise term is elaborated in appendix a. in the next section we will consider virial expansions to @xmath23 so that we will not need to consider the triplet @xmath151 , etc , effective solute potentials in ( [ eq3.4 ] ) in this paper , but it is clear how they can be generated by continuing the pattern @xmath152 , ... .
the total effective solute potential @xmath153 , and its component @xmath154-body effective potentials @xmath155 , @xmath156 , are solute pmfs .
these are reversible works , or changes in constrained free energy ( here the semi - grand potential ) , when , starting from rest and far apart in the vacuum , we add solute molecules at fixed positions to the pure solvent at a given chemical potential and temperature .
thus @xmath157 is the reversible work to add a solute molecule at position @xmath124 in the pure solvent , with fixed state conditions @xmath158 . for this case , where we add just one solute molecule to the pure solvent , the work is a solvent intensive property and we can just as well regard @xmath159 or @xmath160 as fixed , so that this reversible work is also equal to the changes in gibbs and helmholtz free energies .
like all pmfs @xcite , the gradient gives the mean force on a molecule ; here @xmath161 gives the solvent - averaged mean force on a solute molecule fixed at @xmath124 . in this case
the mean force vanishes , as it should for a uniform solvent , as @xmath157 is independent of @xmath124 .
similarly , the total reversible work to add a pair of solute molecules @xmath33 and @xmath35 to the pure solvent at positions @xmath124 and @xmath134 , starting with @xmath33 and @xmath35 at rest and far apart in the vacuum , is @xmath162 .
the terms @xmath163 give the reversible work to add @xmath33 and @xmath35 to the solvent keeping them far apart , and @xmath164 is the reversible work to bring them to finite separation @xmath34 in the solvent , starting far apart .
for the pair pmf @xmath148 , @xmath165 gives the solvent - averaged mean force on solute molecule @xmath33 fixed at @xmath124 , with @xmath35 fixed at @xmath134 .
this mean force will be nonvanishing if the separation @xmath34 between @xmath33 and @xmath35 is not too large .
part of the mean force is the vacuum force @xmath166 and the other part @xmath167 is induced by the solvent .
pmfs can be expressed in terms of the logarithm of the corresponding correlation functions@xcite .
thus for @xmath148 we have @xmath168 , where @xmath149 is the solute - solute pair correlation function for an infinitely dilute solution . for notational simplicity
we have assumed monatomic solute and solvent molecules . for polyatomic molecules , which may be nonrigid , besides the obvious change@xcite required for the solute ideal gas free energy ( [ eq2.21 ] ) ,
one need only include the internal coordinates of the solute and solvent molecules in the averaging over the @xmath1-coordinates when forming the total effective solute potential for the solute molecular centers @xmath29 , and its one - body and two - body components @xmath157 and @xmath148 calculated using ( [ eq3.6 ] ) and ( [ eq3.8 ] ) .
the internal degrees of freedom effectively all become part of the solvent to be integrated out , in forming the effective solute potentials which depend only on the coordinates of the fixed solute molecular centers ( usually chosen to be the centers of mass ) .
here we assume a dilute solution and derive the virial expansions for various properties .
we first employ a simple canonical ensemble technique , developed for gas virial expansions by van kampen@xcite ( see also landau and lifshitz@xcite ) , to derive the virial series in solute concentration @xmath169 for the change of semi - grand potential upon adding @xmath17 solute molecules to pure solvent at fixed solvent chemical potential , volume , and temperature , @xmath170 .
because our effective solute partition function is canonical , we can use the van kampen canonical ensemble method and thereby avoid the complications of the standard grand canonical ensemble method ; the concentration expansion is derived directly with no need for the introduction and subsequent elimination of an activity expansion , and no graphs at all are needed , which avoids the problem of defining the graphs , with their various types such as reducible and irreducible , and then showing that the net result is due only to the irreducible ones . from this basic virial series ,
we then use thermodynamic arguments to derive virial series for the solute and solvent chemical potentials , @xmath171 , @xmath172 and @xmath173 , the osmotic pressure @xmath174 , and the changes of the gibbs free energy and enthalpy , @xmath175 and @xmath176 , respectively .
as is clear from the notation , some of these derived series involve changing independent variables as well as the dependent one .
the semi - grand potential is an extensive quantity , and we now assume low solute concentration and derive a virial series for @xmath178 to @xmath179 , i.e. , @xmath180 where @xmath181 is defined in ( [ eq2.21 ] ) , and @xmath182 and @xmath183 are the first and second virial coefficients for the quantity @xmath184 .
note that @xmath182 is dimensionless and @xmath183 has dimensions of volume .
the virial coefficients to be defined later for other properties will also have these dimensions for the most part . from the virial series ( [ eq4.1 ] ) for @xmath41 we immediately get one for the solute chemical potential @xmath28 by differentiating with respect to @xmath17 ( see the second thermodynamic relation in ( [ eq2.13 ] ) ) , @xmath185 where @xmath186 is the ideal monatomic gas value ( which will be different for polyatomic solutes ) .
the solute chemical potential is an intensive property and for a binary solution can be expressed in terms of three independent intensive variables , here chosen to be @xmath187 .
the virial series ( [ eq4.1 ] ) will now be derived using ( [ eq4.3 ] ) from an analysis of the low order solute clusters which dominate the average @xmath188 at low solute concentrations . at very low solute concentration ( infinite dilution ) , essentially only isolated solute molecules occur and thus only the one - body term @xmath189 in @xmath111 will matter , as pair , etc , interactions are negligible . because the one - body term is independent of the @xmath0-coordinates ( see ( [ eq3.7 ] ) ) the averaging in ( [ eq4.3 ] ) is redundant and a contribution to @xmath41 linear in @xmath17 results , with first virial cofficient @xmath182 given by @xmath190 where @xmath129 is defined by ( [ eq3.6 ] ) .
the physical significance of @xmath129 is now apparent from ( [ eq4.2 ] ) and ( [ eq4.4 ] ) : @xmath129 defined by ( [ eq3.6 ] ) is the solute excess chemical potential at infinite dilution , @xmath191 for @xmath192 , as first shown by widom@xcite with potential distribution theory using the canonical ensemble .
we now consider slightly increased solute concentrations such that essentially only solute singlets and pairs occur , and thus one - body and two - body , but not three - body , etc , solute interaction terms in @xmath29 need to be considered , i.e. , the first two terms in ( [ eq3.4 ] ) .
for these one - body and two - body terms the total effective potential is @xmath193 , where @xmath194 and @xmath195 , and the average in ( [ eq4.3 ] ) can be written @xmath196 since @xmath197 is independent of the @xmath0-coordinates .
when we take the logarithm of the expression ( [ eq4.5 ] ) , @xmath197 will contribute the first virial term to ( [ eq4.3 ] ) as before , and we focus now on the pairs average @xmath198 in ( [ eq4.5 ] ) , which will be seen to generate the second virial term . written out in full , this average is @xmath199 because of the neglect of triplet , etc , solute configurations , the pair terms @xmath200 , etc .
, are uncorrelated in the average ( [ eq4.6 ] ) , so that we have approximately @xmath201 since all @xmath202 pairs are equivalent on average by symmetry , we have @xmath203^{n_a(n_a-1)/2}~.\ ] ] for a uniform fluid the averaging on the right side can be taken over the relative position @xmath204 of molecule @xmath35 with respect to @xmath33 , and because the fluid is isotropic the effective pair potential @xmath205 depends only on the magnitude @xmath34 of @xmath204 .
we assume @xmath206 is short ranged so that the quantity on the right side being averaged , @xmath207 , differs from unity only in a small part of the complete averaging region @xmath6 , and we therefore introduce the two - body mayer cluster function @xmath208 which differs from zero only in the small region , and write ( [ eq4.8 ] ) as @xmath209^{n_a(n_a-1)/2}~.\ ] ] we need @xmath210 and by using @xmath211 for @xmath212 ( here @xmath213 is @xmath214 ) , and @xmath215 for @xmath216 , we have @xmath217 where @xmath218 using ( [ eq4.10 ] ) and ( [ eq4.5 ] ) in ( [ eq4.3 ] ) now gives the second virial term in ( [ eq4.1 ] ) , with osmotic second virial coefficient given by ( [ eq4.11 ] ) .
we will not be needing them , but molecular expressions for the third and higher osmotic virial coefficients @xmath219 , etc , can also be derived by this simple method , parallelling the results derived for the higher gas virial coefficients by this method @xcite .
we find the virial series for the osmotic pressure @xmath174 by differentiating the series ( [ eq4.1 ] ) for the semi - grand potential @xmath41 with respect to volume ( see the first thermodynamic relation in ( [ eq2.13 ] ) ) .
for @xmath7 we thus have @xmath220 from which we immediately get the series ( [ eq1.1 ] ) for @xmath9 .
equation ( [ eq4.12 ] ) can also be derived from the virial series ( [ eq4.2 ] ) for @xmath28 and the gibbs - duhem relation which we will need later , @xmath221 from which we get at fixed @xmath20 and @xmath8 , @xmath222 it is easy to check that the series ( [ eq4.12 ] ) and ( [ eq4.2 ] ) are consistent with ( [ eq4.14 ] ) .
there has been some earlier work on extending mm theory to obtain virial expansions for other solution properties .
@xcite uses the stockmayer semi - grand ensemble with fixed @xmath223 to obtain virial series for the chemical potentials , gibbs free energy and enthalpy .
friedman @xcite derives the relation between intensive properties expressed as a function of mm variables @xmath224 and the more experimentally accessible variables @xmath225 .
kozak et al @xcite discuss the solvent chemical potential @xmath20 and the enthalpy of mixing within the mm framework , and rossky and friedman @xcite derive the second virial correction to the henry law `` constant '' @xmath226 , which is a measure of ( in)solubility of the @xmath0 species .
using our semi - grand canonical ensemble with fixed @xmath2 we have obtained above the virial series for semi - grand potential @xmath41 , solute chemical potential @xmath28 , and pressure @xmath7 , and we now derive virial series for other properties from these using purely thermodynamic arguments .
we will also change independent variables in some cases to match the most convenient experimental variables .
for example , we will discuss the enthalpy @xmath43 as a function of the independent or control variables @xmath227 used in itc experiments .
the enthalpy @xmath43 as a function of @xmath3 can be obtained from the gibbs free energy @xmath228 using the gibbs - helmholtz relation @xmath229 where we find @xmath230 from @xmath231 the extensivity of @xmath230 is manifested in ( [ eq4.16 ] ) by the factors @xmath17 and @xmath42 and we focus on the intensive chemical potentials @xmath28 and @xmath20 , which are functions of three independent intensive variables . in ( [ eq4.2 ] )
we have a virial series in @xmath37 for @xmath28 at constant @xmath20 and @xmath8 .
we first use thermodynamics to convert this @xmath28 series to one at constant @xmath7 and @xmath8 , and from this new @xmath28 series we will use thermodynamics to obtain a @xmath20 series at constant @xmath7 and @xmath8 .
figure 1 shows what is to be done to change control variable from @xmath20 to @xmath7 in the @xmath28 series .
the series ( [ eq4.2 ] ) for @xmath171 is an expansion along the curve @xmath232 in fig.1 , starting from the initial point ( pure solvent ) with @xmath233 .
we now transform this expansion to one along the curve @xmath234 , and , to @xmath235 , it will take the form @xmath236 the new virial coefficients @xmath237 and @xmath238 are taylor series expansion coefficients defined as usual by @xmath239 where @xmath240 is the solute excess chemical potential
. we will relate these new coefficients to the old ones defined by taylor series ( [ eq4.2 ] ) , @xmath241 the quantities @xmath237 and @xmath182 are equal since they both are equal to the solute excess chemical potential in the infinite dilution limit @xmath242 ; in this limit @xmath243 depends on just the pure solvent independent intensive conditions and these can be chosen as @xmath159 or @xmath244 , or any other suitable pair such as @xmath160 .
thus our first relation is simply @xmath245 to derive the relation between @xmath238 and @xmath183 we use the rule for partial derivatives to re - write the definition in ( [ eq4.18 ] ) in terms of a derivative at constant @xmath20 in place of one at constant @xmath7 : @xmath246 where it is to be understood that @xmath8 is constant and the limit @xmath247 is to be taken in each partial derivative here .
the first partial derivative on the right side is @xmath248 as seen from ( [ eq4.19 ] ) . the second partial derivative in ( [ eq4.21 ] )
can be written as @xmath249 since @xmath250 does not depend on @xmath20 , and @xmath251 is evaluated in appendix b ( see ( [ eqc.1d ] ) ) , giving @xmath252 where @xmath253 is the solute partial molecular volume in the infinite dilution limit , @xmath254 is the pure solvent density , and @xmath255 is the pure solvent isothermal compressibility .
the third partial derivative in ( [ eq4.21 ] ) can be transformed to the following form with the help of the gibbs - duhem relation ( [ eq4.13 ] ) with fixed @xmath7 and @xmath8 : @xmath256 in the limit @xmath247 , @xmath257 on the right side is dominated by the ideal term @xmath258 of the virial series ( [ eq4.17 ] ) , so that @xmath259 , and hence @xmath260 equation ( [ eq4.24 ] ) is a version of raoult s law , and can also be derived from the usual formulation of that law .
gregorio and widom @xcite derive ( [ eq4.24 ] ) from potential distribution theory .
putting it all together gives the relation we seek : @xmath261 using ( [ eq4.25 ] ) and ( [ eq4.20 ] ) we write the first two terms of the constant pressure virial series ( [ eq4.17 ] ) for @xmath28 in terms of the original virial coefficients as @xmath262 where the original virial coefficients @xmath182 and @xmath183 , which are intensive solvent properties defined in equations ( [ eq4.4 ] ) and ( [ eq4.11 ] ) and expressed there as functions of chemical potential @xmath20 , are now to be re - expressed as functions of pressure @xmath7 .
we see that the first virial coefficient for @xmath28 in the series at constant pressure is the same as the first virial coefficient @xmath182 in the @xmath28 series at constant solvent chemical potential , but the new second virial coefficient differs from @xmath263 which occurs in the original series . in general
we expect the magnitudes of @xmath183 and @xmath253 to be of the same order and @xmath264 to be negligible ( e.g. , it is of order @xmath265 for water at normal state conditions ) and thus , for example , in cases where @xmath266 and @xmath267 , the second virial coefficient of @xmath28 at constant pressure , @xmath268 , will be negative and larger in magnitude than the second virial coefficient of @xmath28 at constant solvent chemical potential , @xmath248 .
exceptional situations can arise near a solvent theta temperature , where @xmath269 , and near a solvent liquid - gas critical point , where @xmath270 , and @xmath255 can all become large . and examine the parameter space formed by two other intensive variables @xmath37 and @xmath271 , showing schematic contours of the two intensive properties @xmath20 and @xmath7 .
starting at the initial point ( pure solvent , with density @xmath254 ) with @xmath233 , the virial series for property @xmath28 along the curve @xmath272 will be converted to one along the curve @xmath273 . ]
the corresponding virial series for the solvent chemical potential @xmath20 is found from the @xmath28 series and the thermodynamic relation ( [ eq4.23 ] ) . substituting the virial series ( [ eq4.17 ] ) for @xmath274 on the right side of ( [ eq4.23 ] ) gives @xmath275 we now integrate ( [ eq4.27 ] ) with respect to @xmath37 , bearing in mind we require terms in @xmath20 only up to @xmath23 . in terms of the independent variables
@xmath225 , we therefore expand @xmath276 to first order in @xmath37 , @xmath277 where @xmath278 we use ( [ eq4.28 ] ) in the first @xmath279 term in ( [ eq4.27 ] ) and expand @xmath280 as @xmath281 to first order . we can use @xmath282 in the second term , as corrections would generate terms of higher order in @xmath37 than we need . with these substitutions
the integrals are elementary and give , to @xmath23 , @xmath283 where @xmath284 is the pure solvent chemical potential .
some thermodynamic manipulations will turn the expression ( [ eq4.29 ] ) for @xmath285 into something more familiar . with constant @xmath8 to be understood in all derivatives to follow , and remembering that it is the limiting quantity with @xmath242 that we need
, we first re - write the derivative @xmath286 occuring in the definition ( [ eq4.29 ] ) of @xmath285 using the anti - chain rule as @xmath287 the first partial derivative factor on the right side in ( [ eq4.30a ] ) becomes a pure solvent quantitity in the limit @xmath242 , and from the definition of the isothermal compressibility is given by @xmath288 the second partial derivative factor on the right side in ( [ eq4.30a ] ) can be re - written using the rule for relating a partial derivative at fixed @xmath271 to one at fixed @xmath20 , @xmath289 the first partial derivative on the right side of the last equation , in the limit @xmath242 , is @xmath290 , which we find from the pressure virial series ( [ eq4.12 ] ) .
the second partial derivative on the right side of ( [ eq4.32a ] ) , in the limit @xmath242 , becomes a pure solvent quantity which can be found from ( [ eq2.12 ] ) , or from the gibb - duhem relation ( [ eq4.13 ] ) , to be @xmath291 again in the limit of infinite dilution , the third partial derivative on the right side in ( [ eq4.32a ] ) can be derived using the raoult relation ( [ eq4.24 ] ) , and in appendix b ( see ( [ eqc.15 ] ) and discussion ) is shown to be @xmath292 combining the last three relations we find @xmath293 substituting ( [ eq4.35a ] ) and ( [ eq4.31a ] ) into ( [ eq4.30a ] ) we get @xmath294 and then from ( [ eq4.29 ] ) @xmath295 thus the coefficient @xmath296 in ( [ eq4.30 ] ) is @xmath297 , and since from ( [ eq4.25 ] ) we have @xmath298 , we find @xmath299
. substituting this last result into ( [ eq4.30 ] ) gives the virial series for @xmath20 at constant pressure expressed in terms of the original second virial coefficient @xmath183 , @xmath300 where @xmath301 .
we see that the first virial coefficient here , @xmath302 , is trivial in the sense used earlier , so that the first - order change in the solvent chemical potential upon addition of solute , unlike the corresponding change in solute chemical potential , is a colligative property like the osmotic pressure ( raoult law ) .
the second virial coefficient for @xmath303 , i.e. , @xmath304 , will be positive at the temperatures where @xmath305 , in the usual situations where @xmath306 is negligible . from the constant pressure virial series ( [ eq4.26 ] ) and ( [ eq4.31 ] ) for @xmath28 and @xmath20 we obtain the first two terms in the constant pressure virial series for the gibbs free energy @xmath230 using ( [ eq4.16 ] ) , @xmath307 where @xmath308 is the pure solvent volume , and @xmath309 is the solute ideal gas gibbs free energy , which will be different for polyatomic solutes . in deriving ( [ eq4.32 ] ) from ( [ eq4.26 ] ) and ( [ eq4.31 ] )
we express everything in terms of the independent variables @xmath3 , and we retain terms in @xmath17 only of orders @xmath17 and @xmath310 .
thus we write @xmath16 , where @xmath311 is expanded to first order in @xmath17 as @xmath312 where @xmath313 is the infinite dilution limit ( @xmath314 ) of the solute partial molecular volume @xmath315 . using ( [ eq4.32x ] )
we express @xmath37 to @xmath316 as @xmath317 from the last relation and @xmath318 we readily find ( [ eq4.32 ] ) from ( [ eq4.26 ] ) and ( [ eq4.31 ] ) . from ( [ eq4.32 ] )
we obtain the virial series at constant pressure for the enthalpy @xmath43 using ( [ eq4.15 ] ) , @xmath319 where @xmath320 is the solute ideal gas value and the first two virial coefficients in the enthalpy series are given by @xmath321 @xmath322 where @xmath323 is the pure solvent thermal expansion coefficient .
because @xmath324 is linear in @xmath17 we could instead include a term @xmath325 in @xmath326 but choose not to do so as a reminder that this term will change for polyatomic solutes , and because the solute ideal gas term does not contribute to the enthalpy of mixing , which is often the quantity of interest and which is determined by the last two terms in ( [ eq4.33 ] ) .
equations ( [ eq4.34 ] ) and ( [ eq4.35 ] ) will be useful in analyzing itc experiments as discussed briefly in the next section . from the two virial series for @xmath228 and @xmath327 we can get one for the entropy @xmath328 using @xmath329 .
a simple example of a solute pair pmf @xmath26 determined from computer simulation using a nonequilibrium work method@xcite for benzene in water at temperature @xmath330 is shown in fig.2 .
benzene is a classic hydrophobe with very low solubility in water .
the experimental value @xmath331 for the osmotic second virial coefficient at @xmath332 was obtained from heroic vapor pressure measurements@xcite using the concentration dependence of the henry law constant @xmath226 .
the negative sign found for @xmath183 is that expected for attractive hydrophobic interaction .
the theoretical value we obtain from the expression ( [ eq4.11 ] ) and the data of fig.2 is @xmath333 , in reasonable agreement with the experimental value ; the parameters for the vacuum pair potentials of the benzene and water models were taken from the literature ( see references in the caption of fig.2 ) , and data for @xmath183 is not used in determining these parameters .
the pmf for benzene - benzene in water and the osmotic second virial coefficient have been determined by a number of other workers@xcite from various simulation algorithms and vacuum pair potential models and a variety of state conditions , and by rossky and friedman@xcite with a simplified model for the benzene - benzene pmf .
there are also some calorimetry data@xcite for the benzene - water system and if enthalpy osmotic virial coefficients can be extracted from them the equations ( [ eq4.34 ] ) and ( [ eq4.35 ] ) , together with simulation results over a range of temperature for the benzene excess chemical potential and the benzene - benzene pmf , will provide the basis for an analysis . for benzene - benzene in water at t
= 303k obtained from md simulation , with rigid molecular models for tip3p water and for 12-site benzene constructed in vmd 1.9/molefacture 1.3 environment .
force - field parameters from charmm27 .
pmf constructed from a 12ns @xmath334 simulation with 2fs time step using nonequilibrium work ofr algorithm@xcite with ofr - at analysis code@xcite running on namd 2.8 . ]
we are currently analyzing itc experimental data for the enthalpy of mixing of antimicrobial peptides at low concentrations in water @xcite and extracting enthalpy second virial coefficients . using the new expression ( [ eq4.35 ] ) for the enthalpy second virial coefficient at constant pressure , with the temperature
derivative being obtained from simulation results for the peptide - peptide pmf @xmath26 taken at three temperatures , we will compare theory and experiment for this more complicated system in a future publication .
we are grateful to nserc for financial support , compute canada for computing support , and mostsfa nategholeslam for help with latex .
we give a few more details on the derivation of the cluster decompostion ( [ eq3.4 ] ) of the total effective solute potential @xmath29 up to the pairs term . in terms of the solvent - induced part @xmath335 of @xmath111 defined by ( [ eq3.1 ] ) ,
we want to show @xmath336 where @xmath337 and @xmath338 are defined by ( [ eq3.6 ] ) and ( [ eq3.8 ] ) respectively .
we write the total solute - solvent potential @xmath52 as a sum of its contributions @xmath339 , etc , from solute molecules @xmath340 , etc , so that @xmath341 where the solute @xmath0-coordinates are fixed and the solvent @xmath1-coordinates are averaged . the solution is assumed dilute such that solute configurations with only singlets and pairs of @xmath0-molecules need be considered . as an example , consider an @xmath0-configuration where @xmath130 are a close solute pair , @xmath342 are a close pair , and the remaining solute molecules @xmath343 , etc , are singlets . for this @xmath0-configuration we group the terms on the right side of ( [ eqa.2 ] )
accordingly as @xmath344 in the averaging in ( [ eqa.3 ] ) the exponential factors can be taken to be uncorrelated because of the assumed isolation of the indicated pairs and singlets so that we have @xmath345 when we examine the definitions of @xmath337 and @xmath338 in ( [ eq3.6 ] ) and ( [ eq3.8 ] ) , we see that in ( [ eqa.4 ] ) a factor of @xmath346 arises for every @xmath0-molecule @xmath347 ; the first average factor @xmath348 on the right side of ( [ eqa.4 ] ) generates @xmath349 , the second average generates @xmath350 , the next average generates @xmath351 , the next @xmath352 , etc .
a pairs average factor @xmath353 arises from just the close pairs average factors in ( [ eqa.4 ] ) , i.e. , @xmath354 from the first average and @xmath355 from the second .
the general pattern should now be clear : the averaging in ( [ eqa.2 ] ) generates a factor of @xmath346 for every @xmath0-molecule and a factor of @xmath353 for every close pair of @xmath0-molecules . but because @xmath338 vanishes for large separation @xmath53 , we can include a factor @xmath353 for every solute pair , and thus replace @xmath356 with @xmath357 .
we thus see how ( [ eqa.2 ] ) has the form ( [ eqa.1 ] ) .
in section iv.c we used some thermodynamic relations which we now derive . for the first one , equation ( [ eq4.22 ] ) , we write the quantity needed @xmath251 using the chain rule as where here and in the rest of this appendix , if not indicated explicitly , all partial derivatives are understood to be at fixed @xmath8 , and we recall we need the various quantities in the dilute limit @xmath242 . for the first partial derivative on the right side of ( [ eqc.1 ] ) we use the reciprocal relation where @xmath313 is the solute partial molecular volume at infinite dilution , and @xmath254 and @xmath255 are the pure solvent density and isothermal compressibility , respectively . the second partial derivative on the right side of ( [ eqc.1 ] ) becomes a pure solvent quantity in the limit @xmath247 and using the chain rule @xmath361 , the definition of isothermal compressibility for @xmath362 , and ( [ eq4.33a ] ) for @xmath363 , is readily shown to be the next relation we need to derive is ( [ eqc.1b ] ) above , which is also ( [ eq4.34a ] ) of the text .
this relation can be derived using statistical mechanics / potential distribution theory@xcite , and we give here an alternative derivation from thermodynamics .
the first term on the right side of ( [ eqc.1b ] ) is the limiting value of @xmath366 as seen from the raoult relation ( [ eq4.24 ] ) .
the general relation between the derivatives @xmath367 and @xmath368 is in this equation , because @xmath20 is intensive we can write the first factor on the right side as @xmath366 , and using @xmath16 we find @xmath372 for the second factor on the right side , where @xmath373 is the solute partial molecular volume , so that we have on the right side of the last equation , the first factor is given by the definition @xmath378 where @xmath379 is the solution isothermal compressibility , and the second factor is , by definition , the solute partial molecular volume @xmath373 , so that we have where constant @xmath8 is to be understood in the two partial derivatives here . in this relation ,
when we take the limit @xmath242 for an infinitely dilute solution , the left side is what we need in ( [ eqc.1b ] ) , the partial derivative on the right side becomes @xmath383 by the raoult relation ( [ eq4.24 ] ) , the factor @xmath384 becomes unity , and the last term becomes @xmath385 .
this completes the derivation of ( [ eqc.1b ] ) ( and ( [ eq4.34a ] ) ) , subject to proving ( [ eqc.12a ] ) .
m. nichols , m. kuljanin , m. nategholeslam , t. hoang , s. vafaei , l.debruin , b. tomberli , c.g .
gray , and m. jelokhani - niaraki .
, 117:14697 , 2013 .
work to be published will deal with the enthalpy second virial coefficient for peptide - peptide interaction . | mcmillan and mayer ( mm ) proved two remarkable theorems in their paper on the equilibrium statistical mechanics of liquid solutions .
they first showed that the grand canonical partition function for a solution can be reduced to a one with an effectively solute - only form , by integrating out the solvent degrees of freedom .
the total effective solute potential in the effective solute grand partition function can be decomposed into components which are potentials of mean force for isolated groups of one , two , three , etc , solute molecules . secondly , from the first result , now assuming low solute concentration , mm derived an expansion for the osmotic pressure in powers of the solute concentration , in complete analogy with the virial expansion of gas pressure in powers of the density at low density .
the molecular expressions found for the osmotic virial coefficients have exactly the same form as the corresponding gas virial coefficients , with potentials of mean force replacing vacuum potentials . in this paper
we restrict ourselves to binary liquid solutions with solute species @xmath0 and solvent species @xmath1 and do three things : ( a ) by working with a semi - grand canonical ensemble ( grand with respect to solvent only ) instead of the grand canonical ensemble used by mm , and avoiding graphical methods , we have greatly simplified the derivation of the first mm result , ( b ) by using a simple nongraphical method developed by van kampen for gases , we have greatly simplified the derivation of the second mm result , i.e. , the osmotic pressure virial expansion ; as a by - product , we show the precise relation between mm theory and widom potential distribution theory , and ( c ) we have extended mm theory by deriving virial expansions for other solution properties such as the enthalpy of mixing .
the latter expansion is proving useful in analyzing ongoing itc experiments with which we are involved .
for the enthalpy virial expansion we have also changed independent variables from semi - grand grand canonical , i.e. , fixed @xmath2 , to those relevant to the experiment , i.e. , fixed @xmath3 , where @xmath4 denotes chemical potential , @xmath5 the number of molecules , @xmath6 the volume , @xmath7 the pressure , and @xmath8 the temperature . |
in recent years , the quantum computing community has demonstrated the basic building blocks of a scalable ion trap quantum computer @xcite .
the design uses the electronic states of ions trapped in a radiofrequency ( rf ) paul trap as qubits and accomplishes logic gates and state readout by laser - ion interactions .
ions are shuttled between zones in an array of ion traps to perform two qubit gates between arbitrary pairs of qubits .
recent experiments have accomplished state readout @xcite , one and two qubit gates @xcite , and ion shuttling in straight lines and through tees @xcite . scaling this architecture up to the many thousands of qubits necessary for
a useful computation , however , will involve significant physics and engineering challenges .
one of these challenges is to build a many zone ion trap array capable of holding several thousands of ions in memory zones and moving arbitrary pairs of them together in interaction zones .
the ion traps used for quantum computation are based on the linear rf paul trap @xcite .
current experiments typically use gold electrodes deposited on two or more alumina substrates with a geometry similar to that shown in fig .
[ fig : trap - schematics](a ) @xcite .
ions are confined radially to the null axis of an rf quadrupole electric field , while axial confinement and ion movement operations are accomplished by dc electric fields .
while these multi - level traps can be adapted to microfabrication techniques @xcite , it is not clear whether they can be scaled to many zone traps because they require slots through the trap structure .
this will make trap topologies that include loops difficult , as there will be islands of electrodes that will have to be mechanically supported and electrically connected .
chiaverini et al .
@xcite proposed using a planar rf paul trap geometry for ion trap quantum computing which is easy to scale up to many zone traps and amenable to modern microfabrication techniques .
the electrodes all lie in a plane and ions are trapped above the plane of the electrodes @xcite . in the five electrode planar trap design shown in fig .
[ fig : trap - schematics](b ) , the center and outermost electrodes are held at rf ground while the remaining two electrodes are biased with an rf potential for radial confinement . either the center electrode or the outermost two electrodes can be segmented and dc biased for axial confinement .
such planar ion traps can be built using silicon vlsi technology and thus have the capability to scale to arbitrarily large and complex trap arrays @xcite .
of 200400 @xmath0m.,width=309 ] several challenges will have to be addressed , however , before planar ion traps can be used for quantum computing .
perhaps the most significant challenge is that planar ion traps have trap depths that are only of order 1% that of a multi - level trap of comparable dimensions @xcite .
while this is not a problem once the ions are loaded and laser cooled to near the ground state of the motion , it makes loading from a thermal ion source at room temperature or above difficult .
a second challenge is to show that it is possible to perform the three basic ion movement operations required to implement the kielpinski et al .
@xcite ion trap quantum computer architecture in a planar trap : ballistic ion transport in straight lines , splitting and joining pairs of ions in a single trap , and shuttling ions around corners in intersections .
further challenges are that optical access for laser cooling is blocked along the axis orthogonal to the trap substrate , and that ion heating is enhanced first by the close proximity of the electrodes that is required to achieve a reasonable trap depth , and second by the more resistive semiconductor materials used for vlsi @xcite . in this work
we address the first and second challenges .
the latter challenges are addressed elsewhere @xcite .
we investigate planar ion traps using printed circuit board ( pcb ) electrodes to trap macroscopic charged particles in a rough vacuum environment .
we do not cool our ions , and imaging is accomplished by classical ( off resonant ) scattering of laser light . because these experiments do not require an ultra high vacuum environment , we are able to achieve a fast cycle time for testing different electrode layouts .
we optimize the electrode geometry to maximize the trap depth , and further demonstrate loading from a conventional four rod linear paul trap acting as a reservoir of ions .
atomic ions could be laser cooled in the four rod trap before being transferred to the planar trap to circumvent the problem of low trap depth .
we test trap layouts with segmented center electrodes and segmented outer electrodes , and have accomplished all three of the movement operations required for the kielpinski et al .
@xcite architecture .
while transport , splitting and joining @xcite , and movement through intersections @xcite have already been achieved in multi - level traps , this work represents the first realization of these operations in planar ion traps .
this paper proceeds as follows .
section [ sec : design ] offers a general discussion of planar ion trap design and a detailed description of our experimental setup . in sec .
[ sec : characterization ] , we characterize the macroscopic charged particles we use to test the traps . section [ sec : performance ] presents the results of our investigation of planar ion traps including secular motion and ion movement experiments . in sec .
[ sec : variations ] we discuss some of the various alternative trap geometries we have tested and propose a novel planar ion trap geometry that forms a lattice of point rf paul traps . finally , in sec .
[ sec : conclusions ] we summarize and suggest how to proceed in experimentally demonstrating atomic planar ion traps .
the design of a planar trap for trapping macroscopic particles involves three main considerations : the ion source , the ion loading scheme , and the pcb design .
particle ion sources in common use include electrospray ionization @xcite , laser ablation @xcite , and piezoelectric particle generation @xcite .
the careful choice of an ion loading mechanism is important because of the low trap depth of planar traps .
the planar trap described in this paper is loaded using a four rod trap with a higher trap depth . designing
the pcb involves choosing the electrode dimensions to get the desired trap depth , ion height above the substrate , and secular frequencies ; choosing the layout of the control electrodes to allow for performing the desired movement operations ; and choosing a pcb substrate that has a high voltage breakdown threshold , does not accumulate charge easily , and is vacuum compatible in the case of a trap in a uhv environment .
this section begins by presenting some general considerations for planar trap design and proceeds to describe the experimental setup .
the dynamics of an ion in a linear rf paul trap are determined by solving the classical equations of motion @xcite . for an rf quadrupole electric potential @xmath2 the @xmath3 and @xmath4 equations of motion
take the form of mathieu equations : @xmath5 and @xmath6 here @xmath7 is a dimensionless time , the mathieu parameters @xmath8 and @xmath9 are dimensionless rf and dc voltages , and @xmath10 and @xmath11 are the ion charge and mass .
the motion is stable ( i.e. , the components of the ion position vector @xmath3 and @xmath4 are bounded in time ) in regions of @xmath12 parameter space . in particular , for @xmath13
the motion is stable for @xmath14 . in the pseudopotential approximation , where @xmath15 , the ion motion along axis @xmath16 can be decomposed into slow , large amplitude secular motion at the secular frequency @xmath17 and fast , small amplitude micromotion at the rf drive frequency @xmath18 @xcite . for an arbitrary electric potential of the form @xmath19 the secular motion is determined by a secular potential @xmath20 using the secular potential we can calculate the height @xmath21 of the ion above the plane of the electrodes , the secular frequencies @xmath17 in the harmonic region of the potential , and the trap depth @xmath22 .
the trap depth is defined in the usual way as the energy an ion would need to escape the trap . the secular potential for the planar trap used in these experiments is shown in fig .
[ fig : secular - potential ] . secular potential for a planar ion trap .
the particular geometry depicted is that of the cross trap described in sec .
[ sec : expt - apparatus ] .
the secular potential is plotted on a linear color scale as a function of @xmath3 and @xmath4 , with blue representing the lowest and red representing the highest secular potential values .
secular potential values above twice the trap depth are truncated for clarity .
the ion is located at the origin of the coordinate system .
an rf potential @xmath23 is applied to the red electrodes , a dc potential @xmath24 is applied to the blue top plate electrode ( @xmath25 in this figure ) , and the black electrodes are grounded.,width=309 ] in order to determine the optimum relative sizes of the trap electrodes , we have calculated the secular potential numerically for the planar trap geometry defined in fig .
[ fig : secular - potential ] over a range of dimensions using a two dimensional finite element electrostatic package named bela @xcite . here
@xmath26 is the width of the center electrode , @xmath27 is the width of the rf electrodes , @xmath28 is the width of the outer electrodes , @xmath29 is the width of the gap between the center and the rf electrodes , @xmath30 is the width of the gap between the rf and outer electrodes , @xmath31 is the height of a planar electrode which is parallel to and above the trap electrodes , @xmath21 is the height of the ion above the trap electrodes , and @xmath32 is a measure of the trap size .
we set @xmath33 and @xmath34 , and varied @xmath26 , @xmath27 , and @xmath31 .
note that @xmath35 for a true planar trap .
figures [ fig : trap - depth - vs - geometry ] and [ fig : frequency - vs - geometry ] show the normalized trap depth @xmath36 defined by @xmath37 and the normalized secular frequency @xmath38 defined by @xmath39 as functions of @xmath40 and @xmath41 for @xmath42 .
note that @xmath43 for a perfect quadrupole trap and that we consider the secular frequency in the harmonic region of the potential where it is the same along both axes ( this is appropriate when the ions are at low temperature ) .
trap depth versus electrode widths .
the normalized trap depth @xmath36 is plotted in color as a function of @xmath44 and @xmath40 for @xmath42 .
the dashed lines are contours of constant @xmath45 .
note that the maximum trap depth is obtained at @xmath46 and @xmath41 as large as possible.,width=309 ] secular frequency versus electrode widths .
the secular frequency at zero ion temperature @xmath38 is plotted in color as a function of @xmath44 and @xmath40 for @xmath42 .
the dashed lines are contours of constant @xmath45.,width=309 ] using figs .
[ fig : trap - depth - vs - geometry ] and [ fig : frequency - vs - geometry ] , we can design a planar trap to maximize the trap depth subject to the experimental constraints .
as an example , consider a planar ion trap for @xmath47sr@xmath48 .
the ion height @xmath21 is set by either the acceptable ion heating rate or by how tightly the cooling and other lasers can be focused such that they do not scatter off the surface of the trap .
suppose we pick @xmath49 m .
the electrode widths @xmath26 and @xmath27 should be chosen to maximize the trap depth using fig .
[ fig : trap - depth - vs - geometry ] , subject to the constraint that the gap between them @xmath29 must be large enough that the rf voltage does not induce electrical breakdown . in general , the optimum center electrode width is around @xmath50 , and the optimum rf electrode width is as wide as possible given practical considerations such as the capacitance of a large electrode . for our example ,
suppose further that @xmath51 m and @xmath52 m .
[ fig : trap - depth - vs - geometry ] and [ fig : frequency - vs - geometry ] , then , give @xmath53 m , @xmath54 , and @xmath55 . at @xmath56 v and @xmath57 mhz , this corresponds to @xmath58 ev and @xmath59 mhz .
we have found that by adding a planar top electrode ( i.e. , choosing @xmath31 to be finite ) biased at a positive dc potential @xmath24 we can increase the trap depth by up to a factor of a 40 for a given rf amplitude and frequency .
figure [ fig : trap - depth - vs - top - plate ] shows the normalized trap depth defined by @xmath60 as a function of the normalized voltage on the top plate , @xmath61 for @xmath62 and several values of @xmath63 .
note that the trap depth is normalized differently for this plot than it was for fig .
[ fig : trap - depth - vs - geometry ] . this is because @xmath21 is a logical starting point for selecting the electrode dimensions , but @xmath64 is the fixed length scale once the trap is built . at @xmath65 , the smallest barrier in the secular potential over which the ion can escape is in the positive @xmath4 direction . as @xmath66 increases , the minimum energy escape path shifts first to the sides and then to the negative @xmath4 direction . for a given geometry ,
the trap depth increases with @xmath66 until the minimum energy escape path is straight down to the center electrode , then decreases with a further increase in @xmath66 .
the maximum trap depth occurs at @xmath67 and is given by @xmath68 , or a little more than one fourth the trap depth of a standard four rod ion trap with similar dimensions .
the secular potential depends only weakly on @xmath63 because the effect of nonzero @xmath66 is well approximated by a constant electric field in the regime @xmath69 that we consider .
note that because the dc potential pushes the ion off of the rf quadrupole null , there will be increased micromotion .
to first order in a harmonic potential , the micromotion amplitude is @xmath70 where @xmath71 is the displacement of the ion from the rf null . in practice
we envision turning the dc potential on for loading , then gradually turning it off as the ions are laser cooled .
continuing our previous example , with a top plate electrode @xmath72 cm above the trap biased at @xmath73 v , we get a trap depth @xmath74 ev and a micromotion amplitude of 15 @xmath0 m using eq .
( [ eq : micromotion ] ) . a direct numerical calculation of the classical ion trajectory along the @xmath4-axis predicts a micromotion amplitude of 20 @xmath0 m , confirming that eq .
( [ eq : micromotion ] ) is a reasonable approximation despite the rather non - harmonic nature of the potential .
trap depth versus top plate location and potential .
the normalized trap depth @xmath75 is plotted as a function of @xmath66 for @xmath62 and several values of @xmath63 .
for @xmath76 , the ion escapes straight up in the positive @xmath4 direction ; for @xmath77 , the ion escapes out to the side ; and for @xmath78 the ion escapes straight down in the negative @xmath4 direction .
note that the maximum trap depth @xmath68 is obtained at @xmath67.,width=309 ] in this section , we discuss the specifics of the trap s design and construction , our source of ions , the method for loading the planar trap , the environment used for performing our experiments , and our methods for measurement and control of ions .
the planar trap we investigated is made up of four straight arms joined at a cross intersection ( fig .
[ fig : cross ] ) .
it is a printed circuit board ( pcb ) , made using standard techniques .
the electrodes are tin coated copper , and the substrate is gml-1000 , a microwave laminate .
the trap is a five electrode design , with outer electrodes segmented to control the axial potential .
connections to these electrodes are made on the bottom layer of the pcb using surface mount headers .
most opposing pairs of control electrodes are electrically connected , but the two pairs nearest the intersection on each arm are electrically independent for finer control .
the center electrodes are also segmented at the intersection to provide additional control in that region .
the middle three electrodes are all 1.27 mm wide , the electrode spacing is 0.89 mm , and the outer control electrodes are 2.5 mm long .
we do not use the optimal electrode sizes as discussed in sec .
[ sec : trap - design ] due to the practical considerations of printed circuit board manufacture , specifically minimum feature and drill sizes .
electrospray ionization ( esi ) provides ions to the trap .
this technique , commonly used with linear quadrupole filters for mass spectroscopy , applies high voltage to a liquid solution at a sharp tip .
strong electric fields at the tip blow off fine droplets of solution , and as solvent evaporates from these charged droplets , self - repulsion breaks them up into smaller particles , eventually producing individual charged solute particles . following @xcite ,
we begin with a 5% suspension of 0.44 @xmath0 m diameter aminopolystyrene spheres .
we then prepare a solution buffered to ph = 3.9 , add methanol to produce a 4:1 methanol / solution mixture , and add the spheres to produce a 0.05% suspension . the suspension
is then placed in a sealed bottle , and pressurized to 50 - 100 kpa .
this forces the liquid through a 0.45 @xmath0 m filter to block any clusters of microspheres , and out the electrospray tip .
the tip itself is made from fused silica capillary tubing , which is heated , stretched to produce a neck , and cleaved to produce a 100 @xmath0 m opening .
we apply 4 kv directly to the liquid with a copper wire inserted in the fluid near the tip ( fig .
[ fig : electrospray ] ) . the spray plume from the tip
is directed through a grounded mask and into the end of a four rod trap .
m filter and past a copper wire with + 4 kv applied voltage , and out the capillary tip . the charged spray passes through a mask and into a four rod trap.,scaledwidth=45.0% ] by arranging the rods so that they match the strips of a planar trap , we interface the four rod trap directly to the planar trap .
a slot cut in the side of the pcb allows us to lower the four rod trap to a point where its trap axis coincides with that of the planar trap ( fig .
[ fig:4rodtoplanar ] ) .
mutual repulsion of ions loaded in the far end of the four - rod trap forces them along , eventually pushing them into the planar trap .
when completely full , the planar trap can hold @xmath7950 ions .
electrospray ionization must be performed at or near atmospheric pressure , and experiments on free trapped particles require a vacuum , so we enclose both four rod and planar traps in a custom built clear acrylic box ( fig .
[ fig : acrylicbox ] ) , and load through an open side of the box .
we perform all loading at atmospheric pressure , then seal the box shut and carefully pump down .
we use a needle valve to control flow rate during pumping , since any significant air flow can push ions out of the trap .
the lowest attainable pressure in this enclosure is about 15 pa ( @xmath1 torr ) .
green ( 532 nm ) laser light directed along the trap axes illuminates ions by classical ( off resonant ) scattering .
the ions are clearly visible by eye , and we photograph them using an ordinary camera lens mounted on a ccd camera . also shown in fig .
[ fig : acrylicbox ] is a metal plate installed above the trap to increase its trap depth and shield from stray static charge .
machined slots in this plate allow us to view ions from above , but we can also see them from the side .
this slotted top plate design has the advantage of masking laser scatter , but for more complex trap topologies it might be easier to use a transparent conductor such as indium tin oxide or a thin film of gold @xcite .
this could be deposited on a glass plate or directly on the vacuum window or an imaging optic inside the vacuum chamber .
the electronics used to generate the rf and dc potentials for the trap are shown in fig .
[ fig : electronics ] . frequencies for which the microspheres are stable range from a few hundred hertz to several kilohertz , depending on the pressure .
this is in the audio frequency ( af ) range , but we will continue to refer to it as rf .
since these frequencies are easily synthesized , we use a function generator as the source , then use an active high voltage amplifier to reach the target voltage range .
a small tickle signal can be added to the main rf signal before amplification to probe for resonances in ion motion , allowing us to determine secular frequencies ( sec .
[ sec : secular ] ) . in this design , the upper limit to rf voltage
is set by arcing between rf and ground electrodes , which occurs at about 400 v amplitude .
typical operating conditions are 250 v and 1.5 khz . to provide dc control of the trap s 48 independent electrodes
, we use a board with 24 digital output channels to address six serial octal digital to analog converter ( dac ) integrated circuits .
the outputs of these dac chips are then amplified to @xmath8020 v , sufficient to axially confine or to move these particles .
our goal in this work is to study the properties of planar traps rather than the ions loaded into them . to be able to draw conclusions about atomic ion traps from work on macroscopic ions ,
we must ( i ) scale away the charge - to - mass dependence from measured parameters and ( ii ) understand the effect of damping on macroscopic ions since it is negligible for atomic ions .
section [ sec : eoverm ] discusses measurement of the @xmath81 spectrum of the ions .
section [ sec : damping ] discusses the effects of air drag on the ions stability and motion .
the principal drawback of using macroscopic charged particles to investigate a trap design is that , unlike atomic ions , these particles are not identical .
electrospray ionization ( esi ) is a soft ionization technique that produces multiply charged ions with a significant spread in the charge state .
the variation in the mass of the microspheres ( @xmath82 10% ) also contributes to broadening the charge - to - mass distribution .
nevertheless , we are able to keep the @xmath81 spectrum constant by maintaining constant electrospray pressure , voltage , capillary diameter and loading parameters . to measure the @xmath81 spectrum of the ions , groups of about five ions at a time
were loaded into the four rod trap with @xmath83 khz and @xmath84 v. with the chamber pressure at 70 pa , the rf frequency was lowered while observing the ejection frequency @xmath85 of each ion .
the charge - to - mass ratio was calculated using the formula @xmath86 , where @xmath87 , the value of the mathieu @xmath88 parameter at the boundary of the first stability region of a quadrupole trap .
figure [ fig : eoverm ] shows a spectrum obtained for 178 ions with mean @xmath89/amu ( electronic charges per nucleon mass ) . is @xmath90 @xmath91/amu and the standard deviation is @xmath92 @xmath91/amu.,width=309 ] when trapping macroscopic ions at atmospheric pressure or in rough vacuum , the background gas exerts a drag force on the ions that both stabilizes the radial motion and slows down the axial ion movement operations .
it is important to verify that the background pressure in our experiments is low enough to make accurate assessments of the trap performance . for small reynolds number @xmath93 and small knudsen number @xmath94 ,
the drag on a sphere is given by stokes law @xmath95 here @xmath96 is the density of the gas , @xmath97 is the radius of the sphere , @xmath0 is the dynamic viscosity , @xmath98 is the mean free path , and @xmath99 is the ion velocity along the @xmath3 axis .
when the knudsen number is large , stokes law overestimates the drag because the flow is no longer continuous .
we divide eq .
( [ eq : stokes ] ) by an empirical slip correction factor @xmath100 to obtain an expression for the drag force which is valid at large knudsen number @xcite .
we determine the effect of drag on stability by considering a modified dimensionless equation of motion @xmath101 where the dimensionless drag coefficient @xmath102 a similar equation holds for the @xmath4 motion . for @xmath103 ,
this reduces to eq .
( [ eq : mathieu - x ] ) .
numerical computations verify that the region of stable trapping in @xmath12 parameter space grows with increasing @xmath104 @xcite .
in particular , the maximum stable mathieu @xmath88 parameter at @xmath13 goes from @xmath87 at @xmath103 to @xmath105 at @xmath106 .
the drag parameter @xmath104 is plotted as a function of vacuum pressure @xmath107 in fig .
[ fig : drag - parameters ] . at 70 pa ( 0.5 torr ) ,
@xmath108 is already less than 1 so air drag has only a small effect on the trap stability in these experiments .
drag coefficients describing ion stability and the speed of linear ion shuttling versus background pressure .
the dimensionless drag coefficients @xmath104 and @xmath109 for stability and ion shuttling , respectively , are plotted as functions of the vacuum chamber pressure on a log - log scale .
the drag coefficients are computed using @xmath97 = 0.22 @xmath0 m , @xmath110 c , @xmath111 kg , @xmath112 kg m@xmath113 s@xmath113 , @xmath114 , @xmath115 khz , @xmath116 mm , and @xmath117 ; which is appropriate for the cross trap assuming that we are trapping single microspheres .
the circles indicate the values of the drag coefficients at the pressure used in our experiment.,width=309 ] the speed of ion motion in our shuttling experiments , however , is significantly affected by drag .
the physical reason for this is that the characteristic time for shuttling experiments is much longer than the characteristic time for stability ( @xmath118 ) . as a simple model ,
consider an ion which starts from rest at @xmath119 in an axial potential @xmath120 and suppose we measure the ion velocity when @xmath121 . in this model
the length scale @xmath122 is of the order of the axial dimension of a control electrode , and the electric field @xmath123 is of the order of the electric potential applied to a control electrode divided by its axial dimension .
drag is important when the dimensionless drag parameter @xmath124 where @xmath125 and @xmath126 is the time to reach @xmath121 .
here @xmath126 is the solution to the transcendental equation @xmath127 the dimensionless drag coefficient @xmath109 is plotted as a function of pressure in fig .
[ fig : drag - parameters ] . at 70 pa ,
@xmath128 so drag plays an important role in ion shuttling experiments in the cross trap despite the fact that the trap stability is not affected by damping . for large damping ( @xmath129 ) , @xmath130 approaches the terminal velocity @xmath131 .
the drag coefficient @xmath132 , so @xmath133 .
this is experimentally verified in sec .
[ sec : movement ] .
when damping forces can be neglected , the mathieu equations of motion of a trapped ion depend only on dimensionless parameters .
it follows that the dynamics of macroscopic charged particles , viewed in the appropriate time scale ( a micromotion period ) , are identical to that of atomic ions , and can be fully explored without the much more demanding experimental requirements of trapping atomic ions @xcite .
we have found macroscopic charged particle planar traps to be a rapid and accurate test bed for investigating traps of different geometries ( e.g. three electrode , five electrode ) and control electrode layouts ( e.g. segmented center , segmented outer electrodes ) . in this section
, we present the results of an experimental investigation of planar trap performance using this test bed .
a planar trap with a segmented center electrode can be used to control the height of individual ions above the substrate by applying dc voltages to the electrode segments .
one advantage of such control in a quantum computer is the ability to perform single qubit gates on different ions using a single stationary laser beam .
the laser beam would have to be parallel to the ion chain but slightly raised or lowered . by changing the height of individual ions
, they can be brought into the path of the beam to perform single qubit gates .
we measured the height of an ion above the trap substrate using a ccd camera mounted on a calibrated translation stage .
the micromotion of the ion in the @xmath4 direction causes the images ( which have an exposure time much longer than the period of the rf drive ) to show a streak where the ion is located , so we actually measured the positions of the top and bottom of the ion motion .
figure [ fig : ion - height - vs - center - potential ] shows the average of the positions of the top and bottom of the ion motion as a function of the dc potential on the center electrode .
we measured @xmath81 for the ion by lowering the rf drive frequency until the ion became unstable as described in sec .
[ sec : eoverm ] . the expected ion height shown in fig .
[ fig : ion - height - vs - center - potential ] is calculated using this value of @xmath81 and a numerical computation of the secular potential .
ion height above the trap substrate , @xmath21 , is plotted as a function of the dc potential on the center electrode .
the solid line is the expected ion height from numerical computations of the secular potential as described in sec .
[ sec : trap - design ] using the measured charge - to - mass ratio of the ion.,width=309 ] figure [ fig : ion - size - vs - center - potential ] shows the amplitude of the ion motion as a function of the dc potential on the center electrode .
the micromotion is minimized when the ion is located at the null of the rf quadrupole electric field , which occurs when the potential on the center electrode is zero . even without a bias on the center electrode ,
however , there is still a substantial amplitude of ion motion because we do not cool the secular motion of our ions .
ion motion amplitude .
the peak to peak amplitude of the ion motion in the @xmath4 direction is plotted as a function of the dc potential on the center electrode.,width=309 ] we determine the secular frequency of an ion by adding a fourier component of frequency @xmath134 and amplitude @xmath135 to the rf signal .
typically , @xmath135 is between 0.5 v and 2 v. by sweeping @xmath136 , we are able to observe a sharp increase in the size of the ion trajectory along the direction of the secular mode excited by the drive signal as seen in fig . [
fig : excitingsecularmodes ] .
although the drive signal is applied to the rf rods , indirect coupling of the drive to the axial motion of the ion is sufficient to excite axial modes .
the exact axial secular frequency depends on which trap electrodes are used as endcaps and what voltage is applied to them , but it is typically between 15 hz and 30 hz .
the transverse secular frequencies are of more interest because they depend directly on the trap geometry : the geometric factor @xmath137 in eq .
( [ eq : secularfreq ] ) depends only on electrode dimensions and their spacing .
we were not able to resolve separate resonant frequencies for @xmath138 and @xmath139 , which indicates that the ion energy is much less than the trap depth .
figure [ fig : secdependencies ] shows the dependence of the @xmath3 secular frequency on @xmath140 and @xmath18 .
both dependences behave as expected from the pseudopotential approximation .
v and @xmath141 khz .
only the center ground electrode is seen in these pictures because the grounded top plate electrode masks the other electrodes from view .
the width of the ground electrode is 1.27 mm .
( b ) transverse mode excited at a drive frequency of 288 hz .
( c ) axial mode excited at 25 hz.,title="fig : " ] ( a ) v and @xmath141 khz .
only the center ground electrode is seen in these pictures because the grounded top plate electrode masks the other electrodes from view .
the width of the ground electrode is 1.27 mm .
( b ) transverse mode excited at a drive frequency of 288 hz .
( c ) axial mode excited at 25 hz.,title="fig : " ] ( b ) v and @xmath141 khz .
only the center ground electrode is seen in these pictures because the grounded top plate electrode masks the other electrodes from view .
the width of the ground electrode is 1.27 mm .
( b ) transverse mode excited at a drive frequency of 288 hz .
( c ) axial mode excited at 25 hz.,title="fig : " ] ( c ) ) with @xmath142 as a fit parameter.,title="fig : " ] ( a ) ) with @xmath142 as a fit parameter.,title="fig : " ] ( b ) to confirm the accuracy of our simulations of pseudopotentials in planar traps , we measured the ratio of the geometric factor @xmath143 in the planar trap to that of the loading four rod trap .
the large spread in the @xmath81 distribution does not allow for the determination of the geometric factor for each trap separately .
however , we can measure the ratio of the geometric factors of the traps by first measuring @xmath138 of an ion in the planar trap , then applying a sequence of voltages on the control electrodes to move that ion into the four rod trap , where @xmath138 is measured again . by doing several experiments of this kind , we determined the geometric factor ratio to be @xmath144 .
numerical simulations of this setup were also performed using bela , giving for this ratio the value @xmath145 , which is within the error bar of the experimental result .
three important ion movement primitives for a multiplexed ion trap quantum computer are shuttling , turning corners and splitting / joining two ions @xcite .
we have demonstrated these three operations in planar traps , using our macroscopic charged particle system .
figure [ fig : shuttlestrobe ] shows a strobe image of shuttling a microsphere in the planar trap at 40 pa ( 0.34 torr ) .
the shuttling shown was performed by raising the potential of the nearest control electrode to the ion . in this case , the speed of the ion and its final position are determined both by the electrode potential and the drag force on the ion . figure [ fig : ionspeed ] shows how the maximum speed the ion achieves during its motion depends on these two parameters .
( a ) ( b ) we have also shuttled ions using another scheme where the nearest control electrodes on both sides of an ion are used to create a confining well along the axial direction . by moving this well along the trap , we had more control over the acceleration , speed , and final position of the ion .
an ion attains during a shuttling operation versus ( a ) the potential @xmath146 applied to the nearest control electrode at constant pressure ( 40 pa ) fit to a 3/4 power law ( @xmath147 where @xmath148 cm s@xmath113v@xmath149 ) ( b ) the chamber pressure at constant control electrode voltage ( 5 v ) .
ion speed is found to be inversely proportional to the chamber pressure .
this behavior is expected when ( 1 ) the mean free path is much greater than the radius of the particles and ( 2 ) @xmath150 where @xmath151 the damping factor and @xmath152 is the acceleration time of the ion . when the damping is low enough such that ( 2 ) no longer holds , the ion speed is expected to reach a pressure - independent value . ]
an ion attains during a shuttling operation versus ( a ) the potential @xmath146 applied to the nearest control electrode at constant pressure ( 40 pa ) fit to a 3/4 power law ( @xmath147 where @xmath148 cm s@xmath113v@xmath149 ) ( b ) the chamber pressure at constant control electrode voltage ( 5 v ) .
ion speed is found to be inversely proportional to the chamber pressure .
this behavior is expected when ( 1 ) the mean free path is much greater than the radius of the particles and ( 2 ) @xmath150 where @xmath151 the damping factor and @xmath152 is the acceleration time of the ion .
when the damping is low enough such that ( 2 ) no longer holds , the ion speed is expected to reach a pressure - independent value . ] splitting a pair of ions was performed by using the control electrodes to introduce a potential hill that pushes the ions apart .
the ion spacing is approximately equal to the axial dimension of one control electrode , so this technique is satisfactory .
typically , an electrode between the ions is raised to 5 v for separation and lowered back to 0 v to join them again . for performing two qubit gates with atomic ions , it is desirable to bring the ions as close to each other as possible , because gates are fastest when the frequencies of the two axial normal modes of the ions are well separated @xcite . in that case ,
the ion spacing is likely to be smaller than the electrode spacing , and it has been shown that it is advantageous to use the dc electrodes to produce an electric octopole moment for fast separation @xcite .
( a ) ( b ) figure [ fig : turningcorner ] shows an ion turning a corner in the planar trap . simulations of the pseudopotential indicate that there are potential hills in the rf nodal lines in each of the four arms close to the intersection , with a point node right at the intersection . to turn a corner , it is necessary for an ion to overcome the hills in both the source and destination arms .
we have used the following pulse sequence to do that : ( 1 ) lower the center electrode of the destination arm to -2 v , then ( 2 ) raise a control electrode in the source arm to 5 v to push the ion over the hill in the source arm .
we find that the ion has enough kinetic energy after traversing the first potential barrier to also overcome the second one .
macroscopic charged particles and pcb traps allow us to rapidly test a variety of planar trap layouts .
we have focused on the loading and transport properties of multi - zone linear traps that are joined at intersections .
in addition to the cross trap described in this paper , designs under investigation have as many as 100 zones and include corners , three way junctions , and four way junctions .
these elements are necessary to generate the ion trap geometries that arise in complex quantum computation geometries @xcite , particularly those involved in realizing scalable , fault - tolerant quantum computation circuits @xcite .
in addition to fault - tolerant quantum computation , ion traps are promising candidates for quantum simulations .
porras and cirac have proposed using the motional modes of coupled ions to simulate bose - hubbard models @xcite .
the planar version of a point paul trap is a natural way to implement this scheme for a two dimensional system . starting with a plane of rf voltage
one can place dc electrodes at arbitrary positions defining a planar point trap at each position .
the distance between trapping points controls the strength of vibrational coupling and the layout of the electrodes can be used to create both ordered and disordered systems . using this approach ,
we have implemented a hexagonal lattice of traps shown in fig . [
fig : matrix ] ; this lattice successfully trapped macroscopic ions . for experiments with single atom ions , the lattice spacing needs to be @xmath7950 @xmath0 m to be near interesting phase transitions in condensed matter systems such as that studied by porras and cirac .
such a trap can be operated as described above with the ions trapped at the lattice points as shown in fig .
[ fig : matrix ] . if the rf and ground electrodes are switched , the ions are trapped along the honeycomb lines , and applying an offset voltage to the rf on the dots moves the ions through the lattice .
this increased coordination between ion positions may improve fault - tolerant thresholds .
we have experimentally demonstrated the planar ion trap design using macroscopic charged particles as a rapid development test bed .
we have addressed the challenges of ion loading and ion motion given the smaller relative trap depth of planar traps .
the challenge of loading ions into the planar trap was solved by coupling a traditional four rod trap to a planar trap .
ions at high kinetic energy were first loaded into the linear trap and then offloaded onto the planar trap .
this method allows us to load the planar trap with the efficiency of loading a traditional linear trap .
furthermore , we have found that the addition of a charged conductive plane above the trap can be used to increase the trap depth . in the atomic case the ions will be initally loaded at a position of high micromotion limiting the effect of laser cooling . in our design example
this micromotion has an amplitude of @xmath153 when the potential on the top plane is such that the trap depth is increased by a factor of 10 .
however , after initial trapping the micromotion can be minimized and the cooling maximized simply by grounding the conductive plane . controlled movement of the ions in the planar trap was accomplished despite the small trap depth . using increasingly complex traps ,
we have been able to perform all the fundamental movement operations required in a multi - zone architecture : splitting and joining ion chains and moving ions around corners and through four way intersections . before this work ,
ion movement through a four way intersection was predicted to not be possible .
questions remain about the control of ions at low vacuum .
our experiments were performed at a vacuum where the background pressure no longer effects trap stability .
however , the pressure still contributed a significant drag term to the linear motion .
additionally , we have demonstrated that planar traps can be used to produce a wide range of geometries . as an example , we have trapped in a hexagonal lattice .
a similar trap for atomic ions could be used to simulate two dimensional quantum simulations .
one can then controllably introduce disorder into the system by application of voltages or changes in trap fabrication .
planar ion traps are a general tool with many applicatuions in quantum simulation , quantum computation , and mass spectrometry .
using modern two dimensional fabrication techniques , the traps described here can be reconstructed to be compatible with uhv and atomic ions .
open questions about the effect of the surface on ion heating can then be addressed .
planar traps or `` ion chips '' offer an exciting route to study new atomic ion physics . | chiaverini et al . [ quant .
inf . comput .
* 5 * , 419 ( 2005 ) ] recently suggested a linear paul trap geometry for ion trap quantum computation that places all of the electrodes in a plane .
such planar ion traps are compatible with modern semiconductor fabrication techniques and can be scaled to make compact , many zone traps . in this paper
we present an experimental realization of planar ion traps using electrodes on a printed circuit board to trap linear chains of tens of 0.44 @xmath0 m diameter charged particles in a vacuum of 15 pa ( @xmath1 torr ) . with these traps
we address concerns about the low trap depth of planar ion traps and develop control electrode layouts for moving ions between trap zones without facing some of the technical difficulties involved in an atomic ion trap experiment .
specifically , we use a trap with 36 zones ( 77 electrodes ) arranged in a cross to demonstrate loading from a traditional four rod linear paul trap , linear ion movement , splitting and joining of ion chains , and movement of ions through intersections .
we further propose an additional dc biased electrode above the trap which increases the trap depth dramatically , and a novel planar ion trap geometry that generates a two dimensional lattice of point paul traps . |
since the experimental realization of weakly interacting atomic bose - einstein condensates confined in magnetic traps @xcite much work was done to study the interaction of such a system with light .
examples include the scattering of light by a condensate @xcite , its refraction index @xcite , nonlinear atom optics @xcite , the optical detection of the relative phase between two condensates @xcite , and the spontaneous emission in the presence of two condensates @xcite .
in this paper we are concerned with the question of how a condensate of two - level atoms interacts with a single resonant quantized cavity mode of the electromagnetic field .
the system can be realized by placing the trap containing the condensate in a microwave cavity .
we consider the limit of vanishing temperature so that essentially all atoms are in the condensate .
the two states @xmath1 and @xmath2 are both members of the internal ground - state manifold of the condensed atoms but have a different hyperfine magnetic quantum number @xmath3 . for both states
@xmath3 should be chosen in such a way that the atoms remain trapped after the transition from @xmath1 to @xmath4 .
the internal energies corresponding to these states are denoted by @xmath5 and @xmath6 , respectively .
although for the existing condensates the trapping potential is different for states having a different magnetic quantum number we assume for simplicity that the trapping potential is the same for @xmath1 and @xmath2 .
this situation may be realized by using non - magnetic traps or by exposing the atoms to a highly detuned laser beam . if the shape of the beam is appropriately chosen the induced dipole - potential can compensate for the difference between the trap potentials . under these conditions
the hamiltonian in its second quantized form is given by @xmath7 \psi_i ( \vec { x } ) \big \ } \nonumber \\ & & + \hbar \omega_c a^\dagger a + h_{int } + h_{n.l .
} \label{start } \end{aligned}\ ] ] where @xmath8 , @xmath9 is the field operator for atoms in the state @xmath10 and @xmath4 , respectively .
@xmath11 is the annihilation operator for photons in the microwave cavity .
the center - of - mass hamiltonian @xmath12 is the same for both internal states .
@xmath13 denotes the atomic mass , and @xmath14 and @xmath15 are the trap frequencies in the @xmath16 plane and in the z - direction , respectively .
the nonlinear hamiltonian @xmath17 describes the interaction between the atoms .
the quantities @xmath18 are connected to the scattering lengths for the scattering between atoms in the internal states @xmath19 and @xmath20 .
we will make the hartree approximation and consider the case that all atoms are in the same state @xmath21 regarding the center - of - mass motion .
this state fulfills the stationary nonlinear schrdinger equation ( see , e.g. , ref .
@xcite ) with energy eigenvalue @xmath22 and is the same for both internal states .
we model the interaction of the atoms with the microwave photons by the usual magnetic dipole coupling @xmath23 \cdot \int d^3 x \ { \vec{m}_{ge } \psi_e^\dagger ( \vec { x})\psi_g ( \vec{x } ) + h.c . \}\ ] ] where @xmath24 is the magnetic dipole moment of the atoms and @xmath25 the magnetic field of the cavity mode at the position of the trap .
as we consider the case that the trap size is much smaller than the wavelength of the microwave photons the interaction does not affect the center - of - mass motion of the condensed atoms .
this justifies the assumption that the atoms will always remain in the spatial mode @xmath26 . under these assumptions
we can neglect all spatial modes beside @xmath21 and can replace the atomic field operators @xmath27 by @xmath28 , @xmath9 , where the operators @xmath29 are the annihilation operators for atoms in the internal state @xmath30 . substituting this in the hamiltonian ( [ start ] ) the latter can be simplified to @xmath31 \cdot \ { b_g^\dagger b_e \vec{m}_{ge } + b_g b_e^\dagger \vec{m}_{ge}^*\ } \ ; .\end{aligned}\ ] ] in the interaction picture and after the rotating wave approximation the hamiltonian reduces in resonance ( @xmath32 ) to @xmath33 the rabi frequency @xmath34 is given by @xmath35 .
the hamiltonian ( [ htilde ] ) is formally equivalent to that used to describe parametric down conversion in nonlinear optics ( see , e.g. , ref .
@xcite ) . in this system @xmath36 destroys a photon of the pumping beam and @xmath37 and @xmath38 create a signal and an idler photon , respectively .
the interaction between the different modes is caused by a nonlinear medium and describes the process that a pumping photon is converted into two other photons under conservation of energy .
theoretically the pumping beam is often considered as a classical field . to be short
we will focus in the following on the interpretation of the hamiltonian ( [ htilde ] ) as an interaction between two - level atoms in a cavity . though the physical interpretation is different our results
can immediately be applied to parametric down conversion if one considers @xmath39 as the number operator for pumping photons .
a trilinear hamiltonian has been examined numerically by walls and barakat @xcite .
kumar and mehta @xcite derived analytical expressions for the time evolution under the condition that the number of one of the three particle species involved in eq .
( [ htilde ] ) remains strongly populated . that the degenerate parametric oscillator may show collapse and
revival has been predicted numerically by jyotsna and agarwal @xcite .
these results are similar to the numerical results of drobn and jex @xcite .
an exact solution for a fixed number of atoms was given in a non - closed form by tavis and cummings @xcite .
our aim is to present a new analytical method for solving the underlying operator equations . in particular this
will then enable us to describe for the system given above the long - time behaviour of the mean occupation @xmath40 of the excited atomic mode for the case that initially all atoms are excited and there are zero photons in the cavity and to confront it with the time dependence of its variance . to do
so we want to give closed - form approximate analytical solutions for @xmath40 for different initial states of the condensate like atomic number state , coherent state , or mixture .
this will allow us to demonstrate that @xmath40 shows a periodic behaviour or the existence of collapse and revival , respectively .
the hamiltonian ( [ htilde ] ) leads to two conserved quantities : the total number of atoms @xmath41 and , because of the rotating - wave approximation , the number of excitations @xmath42 .
because any two of the number operators @xmath43 can be expressed by the third and the two conserved quantities , it suffices to solve the heisenberg equation for only one of the three number operators .
we will consider the number operator @xmath44 of excited atoms .
introducing the new time variable @xmath45 and considering the heisenberg equation for @xmath44 as well as for @xmath46 where the derivative is taken with respect to @xmath47 , one easily finds ( see , e.g. , ref .
@xcite ) @xmath48 for notational convenience we have introduced the operators @xmath49 and @xmath50 .
note that while @xmath51 and @xmath52 commute with every operator occurring in the problem , @xmath53 does not commute with @xmath44 .
this fact seems to prevent the integration of the second order differential equation to an equation of first order as it is usually done with differential equations for ordinary functions . a common way to circumvent
this difficulty is the _ vanishing variance approximation _
, where a differential equation for @xmath54 is derived under the assumption that the variance @xmath55 defined by @xmath56 is negligible . here and in the following the mean value of an operator @xmath57 will be represented by a bar , @xmath58 .
the complete state of the system is denoted by @xmath59 . for the system under consideration
this approximation was studied by kumar and mehta @xcite starting from eq .
( [ 2ndorder ] ) .
they showed that it is valid for short times only and can be applied only to a very limited class of physical problems .
our objective is to improve this scheme : we want to go a step further and derive differential equations for the mean value @xmath60 as well as for the variance @xmath55 and study the respective long time behaviour for different states of the condensate .
because of @xmath61 , the discussion of @xmath62 needs the knowledge of @xmath63 and therefore the exact integration of the operator equation ( [ 2ndorder ] ) .
this can indeed be done by observing that the commutator between @xmath44 and @xmath53 is given by @xmath64 = \tilde{h}/(i\hbar \omega)$ ] so that @xmath65 follows , what can be used to integrate eq .
( [ 2ndorder ] ) to @xmath66 here @xmath67 is a constant of motion which after some algebra can be written as @xmath68 .
it is therefore related to the conservation of energy and particle numbers .
note that no approximation has been made up to this point .
although eq .
( [ 1stdgl ] ) can not be exactly integrated because @xmath67 does not commute with @xmath44 , we are now able to treat @xmath54 together with @xmath55 if we make the two following assumptions of a _ vanishing asymmetry approximation _ : ( i ) the conserved particle numbers are always uncorrelated to the number of excited atoms so that @xmath69 holds for @xmath70 and any integer @xmath71 .
this equality is exactly fulfilled for all times if the initial state is an eigenstate of @xmath72 , e.g. , a number state .
( ii ) we can always neglect @xmath73 so that @xmath74 can be approximated by @xmath75 .
accordingly , in contrast to the vanishing variance approximation , we only assume the vanishing of the higher moment @xmath73 in the hierarchy of moments @xmath76 of the probability distribution .
the case @xmath77 describes its degree of asymmetry . based on this
we derive from eqs .
( [ 2ndorder ] ) and ( [ 1stdgl ] ) the system of ordinary differential equations @xmath78eq .
( [ sys1 ] ) may be considered as an algebraic equation for @xmath79 which can be inserted into eq .
( [ sys2 ] ) to decouple the system .
this results finally in @xmath80 @xmath55 can then be obtained from eq .
( [ sys1 ] ) .
we are now able to treat the envisaged physical situation : all atoms are initially for @xmath81 excited and no photons are present .
it is desirable to obtain the development in time of @xmath40 for the case of the atomic state being a number state with @xmath82 atoms , the corresponding complete state of the system being denoted by @xmath83 .
we will demonstrate below that @xmath40 for other physical situations can be reduced to this .
the operator equations ( [ 2ndorder ] ) and ( [ 1stdgl ] ) show that the initial conditions for @xmath40 which are to be fulfilled for a number state are : @xmath84 , @xmath85 , and @xmath86 .
we have in fact been able to find an exact solution of eq .
( [ nosys ] ) fulfilling the correct number state initial conditions for @xmath87 , and , to leading order in @xmath82 , for @xmath88 .
it is given by @xmath89 where the parameters @xmath90 and @xmath91 are solutions of the algebraic equations @xmath92
-2 \omega^2(2m-1)\ } \nonumber \\ & & - 120 ( \bar{n}_e^{\mbox{\scriptsize num}}(0))^3 + 60 \bar{a } ( \bar{n}_e^{\mbox{\scriptsize num}}(0))^2 + 8 \bar{a } \bar{b}\nonumber + 6 \bar{c } \nonumber \\ & & + 4 \bar{n}_e^{\mbox{\scriptsize num}}(0)\{3 - 2 \bar{a}^2 - 6\bar{b}\ } \nonumber \\ 0 & = & \omega^4 [ 4 + 19m(m\ ! -\!1 ) ] + 10 \omega^2 ( 2m-1)[\bar{a}-6 \bar{n}_e^{\mbox{\scriptsize num}}(0 ) ] \nonumber \\ & & + 180(\bar{n}_e^{\mbox{\scriptsize num}}(0))^2 -60\bar{a } \bar{n}_e^{\mbox{\scriptsize num}}(0 ) -6 + 4\bar{a}^2 + 12\bar{b}\ ; .
\nonumber\end{aligned}\ ] ] the function cn@xmath93 is one of the jacobian elliptic functions , and @xmath94 is the complete elliptic integral @xcite .
if @xmath90 is close to one cn@xmath93 becomes almost identical to the expression @xmath95 except that cn@xmath96 is periodic in @xmath97 with period @xmath98 $ ] . for a number state @xmath99 we have @xmath100 , @xmath101 , and @xmath102 .
the corresponding solution ( [ lsg ] ) is , to the first two leading orders in @xmath103 , characterized by @xmath104 and @xmath105 .
a second exact solution of eq .
( [ nosys ] ) is structurally similar to that of eq .
( [ lsg ] ) if cn@xmath93 is replaced by the elliptic function dn@xmath93 . in this case
the parameters are determined by different algebraic equations which we will omit in this paper .
= 5.8 cm our improved mean value solution @xmath106 of eq .
( [ lsg ] ) is plotted in fig . 1 together with the ordinary mean value solution obtained in the vanishing variance approximation ( @xmath107 ) .
as one can see , the fluctuations become very large when the dip in the number of excited atoms occurs .
as compared with the vanishing - variance approximation the dip is reduced by a factor of 1/2 .
this is in qualitative agreement with the results of ref .
@xcite where a similar physical system was examined under the assumption that the photons escape quickly so that the number of excitations is not conserved .
note that the frequency of the occurrence of the dip is modified , too .
the period @xmath108 is determined by @xmath109 leading to @xmath110 $ ] .
it diminishes with increasing number of atoms .
based on the solution ( [ lsg ] ) we are now able to calculate the mean number of excited atoms if initially no photons and only excited atoms are present which form a coherent state with mean number @xmath111 of atoms .
we expand this state in the number states @xmath112 .
since each of the states @xmath113 @xmath71 they are orthogonal for different @xmath71 .
this reduces the mean value of @xmath44 to the expression @xmath114 because the r.h.s .
is only a function of the mean values @xmath115 of the number state case , it may directly be evaluated with the help of eq .
( [ lsg ] ) .
the physical reason for this simplification is that the interaction conserves the total number of atoms .
thus , no interferences between number states of different total atom number can occur .
note that the expression ( [ coh ] ) is identical to the mean value for a statistical mixture of number states with poissonian statistics . whereas @xmath106 for a number state shows a periodic behaviour with period @xmath108 it is an important consequence of eq .
( [ coh ] ) that a coherent atomic state as initial state leads to collapse and revival in the number @xmath116 of excited atoms ( see fig . 2 ) .
this is also the case for a statistical mixture . leaving open the question if a condensate can at all be described by a coherent state ( compare , e.g. , ref .
@xcite ) an experimental verification of collapse and revival can not be used to discriminate between coherent state and mixture .
we turn to the details of the time development of @xmath117 .
the time @xmath118 when the first ( large ) revival occurs can be determined by the criterion given in ref .
@xcite : the expression @xmath119 must simultaneously have a maximum for the two neighboring values @xmath120 and @xmath121 .
this guarantees that many terms in the sum in eq .
( [ coh ] ) become simultaneously large .
since cn@xmath93 is periodic this condition becomes a condition on the argument of the jacobian elliptic function , namely @xmath122 and @xmath123 with an integer @xmath124 .
this can be solved for @xmath124 and @xmath118 and results to leading order in @xmath111 in @xmath125 according to its derivation the revival time @xmath118 is independent of the probability weights of the mixture or state when it is expanded in @xmath126 .
it is therefore the same for all mixtures with the same initial mean number of excited atoms .
in addition to the large revival at time @xmath118 there are many smaller revivals at earlier times which arise at fractions of @xmath118 ( see fig .
these fractional revivals have been predicted numerically for the degenerate parametric oscillator in refs .
the fractional revivals arise when only a part of the sum terms in eq .
( [ coh ] ) have simultaneously a maximum .
their revival times can be calculated by assuming that the terms for @xmath127 and @xmath128 are simultaneously maximal .
this leads to @xmath129 . since the revivals reappear
periodically one generally finds them at fractions of small integers of the time @xmath118 .
an estimation of the almost constant value @xmath130 taken by @xmath131 between the collapse and the first revival can be derived under the assumption that the arguments of cn@xmath93 of eq .
( [ lsg ] ) that appear in the sum in eq .
( [ coh ] ) are statistically distributed over the whole period of cn@xmath93 ( i.e. , @xmath132 $ ] ) .
@xmath133 then is simply the average of eq .
( [ lsg ] ) over one period @xmath108 .
the corresponding integration @xmath134 can be performed if cn@xmath135 is approximated by @xmath136 and results to leading order in @xmath111 in @xmath137 the two analytical expressions for @xmath118 and @xmath133 are in good agreement with the numerical evaluation of eq .
( [ coh ] ) .
in conclusion we have presented analytical results for the collapse and revival of the mean occupation number and the variance of excited atoms in a microwave cavity .
the phenomenon of a revival in the context of bose - einstein condensation was also examined for different physical systems .
zhang and walls @xcite discovered it numerically for atoms passing through a standing light wave .
@xcite showed that it occurs if the bose condensate is described by a superposition of states with different total number of atoms .
our results differ from these approaches in that a third mode ( the microwave photons ) is fully incorporated and that the new theoretical method applied here allows to derive closed expressions for the interesting physical quantities . * acknowledgement * : we thank markus holzmann for discussions and the optik - zentrum konstanz for financial support .
et al . _ ,
science * 269 * , 198 ( 1995 ) ; c.c .
lett . * 75 * , 1687 ( 1995 ) ; k.b .
et al . _ , phys .
* 75 * , 3969 ( 1995 ) .
l. you , m. lewenstein , r.j .
glauber , and j. cooper , phys .
rev . a * 53 * , 329 ( 1996 ) .
o. morice , y. castin , and j. dalibard , phys .
a * 51 * , 3896 ( 1995 ) .
g. lenz , p. meystre , and e.m .
wright , phys .
a * 50 * , 1681 ( 1994 ) .
w. zhang and d.f .
walls , phys .
a * 49 * , 3799 ( 1994 ) .
j. javanainen , phys .
a * 54 * r4629 ( 1996 ) .
savage , j. ruostekoski , and d.f .
walls , preprint cond - mat/9612174 ( 1996 ) .
g. baym and c.j .
pethick , phys .
lett . * 76 * , 6 ( 1996 ) ; a.l .
fetter , phys .
a * 53 * , 4245 ( 1996 ) .
l. mandel and e. wolf , _ optical coherence and quantum optics _ , cambridge university press , cambridge 1995 .
walls and r. barakat , phys .
rev . a * 1 * , 446 ( 1970 ) .
s. kumar and c.l .
mehta , phys .
a * 21 * , 1573 ( 1980 ) .
i.v . jyotsna and g.s .
agarwal , journ .
mod . opt . * 44 * , 305 ( 1997 ) .
g. drobn and i. jex , phys . rev .
a * 45 * , 1816 ( 1992 ) .
m. tavis and f.w .
cummings , phys . rev . *
170 * , 379 ( 1968 ) .
m. abramowitz and i.a .
stegun ( eds . ) , _ handbook of mathematical functions _
, dover , new york 1972 .
r. bonifacio , p. schwendimann , and f. haake , phys .
rev . a * 4 * , 302 , 854 ( 1971 ) .
cirac , c.w .
gardiner , m. naraschewski , and p. zoller , phys .
a * 54 * , r3714 ( 1996 ) .
narozhny , j.j .
sanchez - mondragon , and j.h .
eberly , phys .
a * 23 * , 236 ( 1981 ) .
w. zhang and d.f .
walls , phys . rev .
a * 52 * , 4696 ( 1995 ) .
wright , d.f .
walls , and j.c .
garrison , phys .
lett . * 77 * , 2158 ( 1996 ) . | we present a new theoretical method to study a trapped gas of bosonic two - level atoms interacting with a single mode of a microwave cavity .
this interaction is described by a trilinear hamiltonian which is formally completely equivalent to the one describing parametric down - conversion in quantum optics .
a system of differential equations describing the evolution , including the long - time behaviour , of not only the mean value but also the variance of the number of excited atoms is derived and solved analytically . for different initial states
the mean number of excited atoms exhibits periodically reappearing dips , with an accompanying peak in the variance , or fractional collapses and revivals .
closed expressions for the period and the revival time are obtained .
@xmath0 + |
motivated by scenarios of large extra dimensions @xcite , the study of black holes in higher dimensions has attracted attention . in these scenarios ,
extradimensional effects play an important role for the properties of black holes with radius smaller than the size of extra dimensions .
an interesting phenomena is mini
black hole production in planned accelerators . if the planck energy is of @xmath0(tev ) as suggested in scenarios of @xcite , phenomena associated with quantum gravity will be observed at the cern large hadron collider ( lhc ) through black hole production @xcite . at the lhc
, the produced black hole is expected to settle down to a higher - dimensional kerr black hole after emission of gravitational waves .
the formed black hole will be subsequently evaporated by the hawking radiation , for which high - energy particles are emitted and may be detected . to know the parameters of the formed kerr black hole which determines the feature of the hawking radiation observed , studies for gravitational radiation are an important issue . in this paper , we study axisymmetric collision of two black holes in higher dimensions as a first step for the investigations on the black holes formed in particle colliders .
s but only one black hole is produced per 1 s. ] this is an extension of our previous study @xcite of the higher - dimensional brill - lindquist initial data and its temporal evolution by the close - limit approximation . in four dimensions ,
there is the well - known bowen - york method @xcite for generating initial data of several moving black holes .
we generalize this method for higher dimensions . similarly to the four - dimensional case , the conformally transformed extrinsic curvature is given analytically and the conformal factor should be calculated numerically .
we perform numerical calculation for the conformal factor in the case of the head - on collision of two equal - mass black holes and then determine the location of apparent horizon ( ah ) that encloses two black holes ( the common ah ) .
we clarify the parameter space for common ah formation .
then , the initial data is evolved adopting the close - slow approximation . in this approximation ,
the distance between two black holes and the linear momentum of each black hole are regarded as small parameters compared to the black hole size and the gravitational mass of the system , respectively . in this case
, the system may be considered to be a perturbed single schwarzschild black hole . in the four - dimensional case ,
the results in the close - slow approximation agree with those in the fully nonlinear analysis ( see @xcite and references in @xcite ) , and hence , it is natural to expect that this is also the case in the higher - dimensional case .
we provide a formula for the radiation efficiency ( i.e. , the ratio of the radiated energy to the system energy ) and discuss their dependence on the dimensionality .
the purpose of this paper is twofold .
one is to present a formulation for providing the initial condition of colliding black holes in higher dimensions .
gravitational waves at the collision of two black holes are accurately computed only by a fully nonlinear numerical relativity simulation .
although such a simulation has not been done yet for higher - dimensional spacetime , several groups have developed robust techniques for numerically computing merger of binary black holes in four dimensions @xcite .
the methods used in four dimensions can be applied for higher - dimensional problems and hence , the simulation will be done in the near future . for such future simulations , a method for providing initial condition we present here will be useful .
the second purpose is to approximately evaluate gravitational radiation in the collision of two black holes in the higher - dimensional case .
since we adopt a linear perturbative method , a reliable result is derived only for a small parameter space . however , our result for such parameters will be useful for calibrating fully nonlinear results in the near future
. this paper is organized as follows . in the next section ,
we derive a generalized bowen - york formulation @xcite for higher - dimensional space . applying the formulation ,
we construct the initial data for head - on collision of two equal - mass black holes in sec .
the adm mass and the common ah are analyzed . in sec .
iv , we evolve the initial data adopting the close - slow approximation . the master variable of the linear perturbation around a @xmath1-dimensional schwarzschild black hole is calculated numerically and the formula for radiation efficiency is derived .
the radiation efficiency is shown to depend weakly on the dimensionality .
v is devoted to a discussion on possible implication of our results for the radiation efficiency in the particle collision . in appendix a , we present some solutions of the extrinsic curvature in the generalized bowen - york formulation that were not introduced in sec .
ii . in appendix
b , the method for imposing the initial condition for the master variable is explained .
in this section we generalize the bowen - york formulation for higher dimension space and present a method for generating initial condition of @xmath2 black holes with linear momenta .
let @xmath3 be a @xmath4-dimensional spacelike hypersurface @xmath5 with the metric @xmath6 and the extrinsic curvature @xmath7 in a @xmath1-dimensional spacetime .
we introduce a number @xmath8 .
the hamiltonian and momentum constraints are @xmath9 @xmath10 where @xmath11 denotes the ricci scalar of @xmath5 and @xmath12 is the covariant derivative with respect to @xmath6 . introducing @xmath13 defined by @xmath14 the hamiltonian and momentum constraints
are rewritten as @xmath15,\ ] ] @xmath16 where @xmath17 and @xmath18 denote the ricci scalar and the covariant derivative with respect to @xmath13 , respectively . here raising and lowering the index are done with @xmath6 . in the following ,
we assume the conformal flatness on @xmath5 , @xmath19 , and impose the maximal slicing condition , @xmath20 .
a weighted extrinsic curvature is defined by @xmath21 and hereafter its index is raised and lowered by @xmath22 ( i.e. , @xmath23 and @xmath24 ) .
then the hamiltonian and momentum constraints become @xmath25 @xmath26 where @xmath27 denotes the ordinary derivative with respect to the cartesian coordinate @xmath28 and @xmath29 . following bowen and york @xcite , we assume that @xmath30 does not have the tensor mode and thus takes the following form : @xmath31 substituting this formula into the momentum constraint , we obtain @xmath32 introducing auxiliary functions @xmath33 and @xmath34 , we decompose @xmath35 as @xmath36 then the equation becomes @xmath37 hence the momentum constraint is satisfied if @xmath38 since the solutions of eqs .
( [ eq - ba ] ) and ( [ eq - chi ] ) are analytically given , solutions for @xmath39 are easily provided .
to give linear momenta of black holes , we choose the solution @xmath40 where @xmath41 is a constant vector , @xmath42 the gravitational constant , @xmath43 the @xmath44-dimensional area of a unit sphere , and @xmath45 .
then , we obtain @xmath46 \right\},\ ] ] where @xmath47 .
this solution provides the extrinsic curvature for one boosted black hole located at @xmath48 . actually , @xmath41 agrees with the adm momentum : @xmath49 since the momentum constraint is a linear equation , we can superpose @xmath2 solutions . denoting the locations of @xmath2 black holes as @xmath50
, a solution of the extrinsic curvature is written as @xmath51 \right\},\ ] ] where @xmath52 , @xmath53 , and @xmath54 denotes the momentum of the @xmath55th black hole .
the conformal factor @xmath56 is obtained by solving the hamiltonian constraint . following @xcite , we assume that @xmath56 has the following form @xmath57 where @xmath58 and @xmath59 denotes the mass parameter of @xmath55th black hole . then , the equation for @xmath60 becomes @xmath61 the solution in this procedure represents the so - called `` puncture '' space with @xmath2 einstein - rosen bridges and @xmath62 asymptotically flat regions ( say , one upper sheet and @xmath2 lower sheets ) . since the right hand side of eq .
( [ eq20 ] ) behaves like @xmath63 for @xmath64 , there is a regular solution for @xmath60 , which can be solved numerically ( this fact was first pointed out in @xcite for the four - dimensional case ) .
the adm mass @xmath65 is given by @xmath66 using the gauss law , we find @xmath67 the first and second terms could be interpreted as the sum of the mass of @xmath2 black holes and the kinetic energy of the black holes , respectively .
in this section , we present initial data of the axisymmetric two - black - hole system following the formalism described in the previous section . we introduce the cartesian coordinate @xmath68 for @xmath69-dimensional space and write the solution of @xmath30 as @xmath70 where two black holes @xmath71 are located at @xmath72 and have momenta @xmath73 .
we assume that the two black holes have the same mass @xmath74 .
the gravitational radius of a black hole of mass @xmath75 is defined by @xmath76 in the case @xmath77 , the common apparent horizon ( ah ) that encloses two black holes is located at @xmath78 where @xmath79 using @xmath80 the conformal factor in the puncture framework @xcite is given by @xmath57 where @xmath81^{n-1}\left(\frac{1}{r_+^{n-1 } } + \frac{1}{r_-^{n-1}}\right).\ ] ] for a numerical solution of @xmath60 , we introduce the cylindrical coordinate @xmath82 where @xmath83 . in this coordinate , the equation for @xmath60 becomes @xmath84 since the system is axisymmetric and equatorial - plane symmetric , it is sufficient to solve the equation for the domain @xmath85 and @xmath86 with the boundary conditions @xmath87 at the @xmath88-axis and @xmath89 at the equatorial plane . for @xmath90 , @xmath60
asymptotically behaves as @xmath91 thus , we impose the so - called robin condition , @xmath92 , at the outer boundary . in the cylindrical coordinate , it is rewritten as @xmath93 we put the outer boundary at @xmath94 and solve the equation for @xmath60 using a finite difference method with the grid number @xmath95 .
numerical computation was carried out for the parameter space @xmath96 and @xmath97 with the @xmath98 interval or the 0.01 interval .
there are two sources for numerical error of the conformal factor : one is associated with the finiteness of the grid spacing and the other with the finiteness of the outer boundary location . to evaluate the numerical error by the grid spacing , we took reference data with sufficiently large grid number ( @xmath99 grids ) and evaluated following three characteristic numerical error values : @xmath100 where @xmath2 stands for the label of the grids . changing the grid spacing , we confirm that @xmath101 decreases with improving the grid resolution at second order .
all the three error estimates were found to be small . with our standard choice ,
@xmath102 is @xmath103@xmath104% for @xmath105@xmath106 .
computation was also performed changing the location of the outer boundaries from @xmath94 to @xmath107 and the numerical solution converges with increasing the radius of the outer boundaries . in this case , @xmath108 is found to be relatively large , about 917% for @xmath105@xmath106 .
such large differences occur at outer boundaries and come from the fact that the robin condition is an approximate boundary condition . however , because @xmath60 is small at the outer boundary , the error of the conformal factor @xmath56 is small . in fact
, we found that @xmath102 is about @xmath109 for @xmath105 and becomes smaller as @xmath1 is increased ( less than @xmath110 in the case @xmath111 ) .
the error @xmath101 is smaller than @xmath108 by a factor for @xmath105 and is comparable to @xmath102 for @xmath111 . because @xmath56 is used in the calculations of the adm mass or the ah , the error for these calculations is expected to have the order of @xmath102 .
hence , we consider that sufficient accuracy is obtained in our calculation .
figure [ contour - adm ] shows the contours of @xmath112 on the @xmath113-plane for @xmath105@xmath106 . the difference between @xmath114 and @xmath75 indicates the strength of nonlinearity due to the right hand side of the hamiltonian constraint .
we find that the nonlinearity of the system becomes large as @xmath1 is increased for fixed values of @xmath115 and @xmath116 .
a common ah that covers two black holes is searched for introducing the spherical coordinate @xmath117 by @xmath118 and @xmath119 and assuming that the location is denoted by @xmath120 . then the equation for the ah is given by @xmath121 where @xmath122 is the unit normal to the surface : @xmath123 then , the equation for the ah reduces to an ordinary differential equation for @xmath124 : @xmath125 this equation is solved under the boundary conditions @xmath126 at @xmath127 and @xmath128 .
figure [ critical - ah ] shows the critical line for the common ah formation .
irrespective of the dimensionality , the common ah is formed if @xmath115 is sufficiently small .
it is interesting to point out that , for @xmath105 , the ah formation is more subject for larger values of @xmath129 while for @xmath130 , the presence of @xmath129 tends to prevent formation of the common ah .
this might seem strange because the kinetic energy is naturally expected to help the ah formation .
in order to obtain a corroboration for this result , we direct attention to the ah ( denoted by @xmath131 ) at the critical separation @xmath132 in the case @xmath133 .
then we add small @xmath134 without changing the value of @xmath135 . because the change in the conformal factor is @xmath136 and
can be ignored , the surface @xmath131 is interpreted as a minimal surface on which @xmath137 holds .
now let us consider the expansion on this minimal surface which is given by @xmath138 .
figure [ expansion ] shows the behavior of @xmath139 in the four- and six - dimensional cases . in the case
@xmath105 , we find that @xmath140 is negative and thus the minimal surface is a trapped surface .
this guarantees the existence of an ah outside of the minimal surface .
hence the motion helps ah formation in the four - dimensional case . on the other hand , in the case @xmath141
, @xmath140 becomes positive on some part of the minimal surface .
the similar behavior was found for all @xmath142 .
therefore , the minimal surface is not a trapped surface and the motion does not necessarily help the ah formation in the higher - dimensional cases .
we found numerically that the term @xmath143 mainly contributes to the positivity of @xmath140 .
this is because for high @xmath1 , the ah at the critical separation is hourglass shaped in the neighborhood of the equatorial plane and the value of @xmath144 is quite large around @xmath145 .
it enhances the value of @xmath143 which is proportional to @xmath146 .
.the values of @xmath147 evaluated on the ah critical line for @xmath148 .
the unit is % . [
cols="^,^,^,^,^,^,^,^,^",options="header " , ] [ compare - erad ] table [ compare - erad ] shows the values of @xmath149 evaluated in this procedure . in the higher - dimensional cases , it is @xmath1503% for @xmath151 and @xmath152 for @xmath153 .
these values do not depend significantly on @xmath1 in the higher - dimensional cases .
note that the values in table [ compare - erad ] does not contradict the bounds derived from the ah area shown in table [ lower - bound ] .
the difference between @xmath75 and @xmath114 increases with increase of @xmath116 ( cf .
[ contour - adm ] ) .
this difference comes from time asymmetry of the initial data and has the magnitude of order @xmath154 .
thus for a large difference with @xmath155 , the linear approximation breaks down . in the four - dimensional case , we know that the nonlinear effect suppresses @xmath156 @xcite .
if this is also the case for @xmath157 , the value of @xmath158 for @xmath116 would be an overestimate and would indicate that gravitational radiation do not significantly contribute to the loss of the system energy in the head - on collision of two black holes .
in this paper , we presented a generalized bowen - york formulation for the initial value problem of multi black holes in higher - dimensional space . using this formulation , we derived analytic solutions of the momentum constraint equation for extrinsic curvature of moving black holes in the conformally flat higher - dimensional space .
as an application , initial data for head - on collision of two equal - mass black holes were computed numerically solving the hamiltonian constraint for the conformal factor .
the properties of the obtained space such as the adm mass and the common ah were analyzed .
we determined the critical line for formation of a common ah in the parameter space of @xmath159 .
then , we evolved this system using the close - slow approximation and analyzed gravitational waves .
we derived the formula for the radiated energy for a wide variety of @xmath1 [ see eq . together with table [ c1c2c3 ] ] and clarified that it depends weakly on @xmath1 ; irrespective of @xmath1 the efficiency is less than @xmath152 for @xmath160 and @xmath161 . as we mentioned in sec .
i , one of the motivations of this study is to clarify the total radiated energy of gravitational waves in the black hole formation through particle collisions in accelerators . however , there are two shortcomings in this study .
one is that we use the linear perturbation method which provides a reliable result only for small values of @xmath135 and @xmath129 .
the other is that the bowen - york black holes system does not accurately model the high - energy particles system . as for the first problem
, we can learn from the four - dimensional study . in @xcite ,
the results of the close - slow approximation and of the fully nonlinear numerical simulation are compared .
it shows that the results in the close - slow approximation agrees approximately with the fully nonlinear results for small separation and small momentum .
however , the difference becomes larger as separation or momentum is increased . for @xmath162 , the nonlinear numerical results indicate that the radiation efficiency increases as @xmath129 is increased and approaches to an asymptotic value , while extrapolation of the linear perturbation results gives infinity . to know the results for @xmath162 in higher - dimensional spacetime ,
obviously , fully nonlinear numerical simulation is necessary .
the difference between the high - energy particles and the bowen - york black holes is the other problem . as the model of a high - energy particle
, the metric of aichelburg - sexl @xcite is often used .
it is obtained by boosting the schwarzschild black hole to the speed of light with fixed energy @xmath163 and taking lightlike limit @xmath164 .
the gravitational field is localized in the transverse plane to the direction of motion and forms a gravitational shock wave , which is a reminiscent of the infinite lorentz contraction of the isotropic gravitational field of the original schwarzschild black hole .
the spacetime is flat except at the shock wave and this enables us to write down the metric of two high - energy particles outside of the lightcone of shock collision .
hence , the ah formation at the instant of collision has been investigated so far @xcite .
obviously , gravitational field ( the conformal factor or the extrinsic curvature ) is not localized in the case of bowen - york black holes .
this indicates that the bowen - york black holes system can not be a precise model for the high - energy particles system even in the case @xmath165 and @xmath166 .
actually , there exist several discussions @xcite about the fact that the bowen - york black hole system contains unphysical gravitational waves , i.e. , so - called junk energy .
thus , in the fully nonlinear numerical calculation with the bowen - york initial data , one should take into account the possibility that junk energy changes the estimate of radiation efficiency . to avoid this
, one should make the junk energy radiate away before the collision setting the initial separation sufficiently large . finally , we comment on the importance of the study of the grazing collision . in the head - on collision case ,
the upper bound on the radiation efficiency ( @xmath167 ) in the head - on collision of aichelburg - sexl particles was obtained @xcite by studying the ah area and applying the area theorem .
although the same discussion was done in the grazing collision with nonzero impact parameter @xmath168 @xcite , the upper bound on the radiation efficiency is very weak for large @xmath168 .
hence , the direct calculation of gravitational radiation in the grazing collision is more important than the head - on collision case .
also , it is required to calculate the kerr parameter of the resultant black hole . in order to solve these problems , the off - axis collision of the bowen - york black holes
could be an approximation and provide several implications .
we plan to study this process using the close - slow approximation as a first step .
the work of h. y. was partially supported by a grant for the 21st century coe program ( holistic research and education center for physics self - organization systems ) at waseda university .
h. y. also thanks the killam trust for the financial support .
the work of t. s. was supported by grant - in - aid for scientific research from ministry of education , science , sports and culture of japan ( no . 13135208 , no .
14102004 , no . 17740136 and no .
17340075 ) , the japan - u.k . and japan - france research cooperative program
we briefly comment on other solutions of @xmath30 for the momentum constraint , which were not introduced in sec .
since the equations for @xmath33 and @xmath34 are pure laplace equations , we can make infinite number of solutions for the extrinsic curvature . here , we present solution of low - multipole momenta . to make a spinning black hole , we write @xmath169 where @xmath170 is an anti - symmetric tensor .
then , @xmath30 is written as @xmath171 @xmath170 denotes an angular momentum tensor expressed by @xmath172 in the three - dimensional case , we can define the angular momentum vector by the formula @xmath173 the conformal factor @xmath56 can be determined by solving the hamiltonian constraint in the same procedure as in sec .
it is possible to derive other solutions , e.g. , by setting @xmath174 or @xmath175,~~\chi=0,\ ] ] where @xmath176 and @xmath177 are symmetric and anti - symmetric tensors , respectively .
however the physical meaning of these solutions is unclear .
in this appendix , we explain how to set the initial condition of the master variable @xmath178 .
we begin by studying the initial condition of the following perturbative metric @xmath179 where @xmath180 denotes the hyper - spherical harmonics @xmath181}_\ell \equiv k^{[n]}_\ell c^{[(n-1)/2]}_\ell(\cos\theta),\ ] ] @xmath182}_{\ell}=\left[\frac{4\pi^{(n+1)/2}\gamma(n+\ell-1 ) } { ( n+2\ell-1)\gamma(\ell+1)\gamma((n-1)/2)\gamma(n-1)}\right]^{-1/2 } , \label{xkln}\ ] ] on the unit sphere of which metric is @xmath183 @xmath184 and @xmath185 are given by @xmath186 @xmath187 where @xmath188 denotes the covariant derivative on the unit sphere and @xmath189 is defined in eq . .
since the initial metric is the brill - lindquist one , the values of @xmath190 are same as the ones that we have derived in @xcite : @xmath191}\right)^{-1 } , \label{chi}\ ] ] @xmath192 here our purpose is to compute the time derivative of these variables .
the nonzero components of @xmath30 shown in eqs .
are rewritten as @xmath193}}{k_2^{[n]}}+n\frac{s_0^{[n]}}{k_0^{[n]}}\right]r^{-(n+1)},\ ] ] @xmath194}}{k_2^{[n]}}\gamma_{ij } + \frac{2(n+1)}{n-1}\frac{s_{2~ij}^{[n]}}{k_2^{[n ] } } + \frac{s_0^{[n]}}{k_0^{[n]}}\gamma_{ij } \right]r^{-(n-1)}.\ ] ] thus @xmath7 is composed of @xmath195 and @xmath196 modes . since @xmath195 mode is absorbed in a coordinate transformation , we omit it hereafter . the extrinsic curvature @xmath197 [ where @xmath198 is defined in eq . ] in the @xmath199 coordinate is @xmath200}}{k_2^{[n]}}\right ] f^{-1}\varpsi_0^{-2(n+1)/(n-1)}r^{-(n+1)}. \label{krr}\ ] ] @xmath201}}{k_2^{[n]}}\gamma_{ij } + \frac{2(n+1)}{n-1}\frac{s_{2~ij}^{[n]}}{k_2^{[n ] } } \right]\varpsi_0^{-2}r^{-(n-1)}. \label{kij}\ ] ] @xmath7 is written as @xmath202 in terms of the lapse function @xmath203 and the shift vector @xmath204 . although @xmath204 can be freely chosen provided that @xmath205 , we impose @xmath206 for simplicity .
this brings us @xmath207 and @xmath208 . using eqs . and
, we have the explicit formulas of @xmath209 and @xmath210 and they are compared with eq . : @xmath211 @xmath212 @xmath213 @xmath214 where @xmath215 } } \frac{\sqrt{f}}{r^{n+1}}. \label{eq : eta}\ ] ] now we are in a position to derive the equation for @xmath178 and @xmath216 .
setting @xmath217 the gauge invariant quantities derived in @xcite are @xmath218 @xmath219 where @xmath220 and @xmath221 denotes the covariant derivative with respect to the metric @xmath222 . in our case , @xmath223 and @xmath224 @xmath225
is related to @xmath178 and @xmath226 as shown in @xcite and we find the same equation for @xmath178 as that in our previous paper @xcite .
hence , the initial value of @xmath178 is unchanged . calculating @xmath227
, we obtain @xmath228,\ ] ] where we have substituted eq . .
since @xmath227 is related to @xmath216 and @xmath229 as @xmath230 we find the equation @xmath231 where @xmath232 is a function of @xmath233 given in @xcite and we used eq . .
the solution is given by eq . .
n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 429 * , 263 ( 1998 ) [ arxiv : hep - ph/9803315 ] ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , _ ibid .
_ * 436 * , 257 ( 1998 ) [ arxiv : hep - ph/9804398 ] ; l. randall and r. sundrum , phys . rev . lett . * 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ]
. l. randall and r. sundrum , phys .
lett . *
83 * , 4690 ( 1999 ) [ arxiv : hep - th/9906064 ] .
t. banks and w. fischler , arxiv : hep - th/9906038 . s. b. giddings and s. thomas , phys . rev .
d * 65 * , 056010 ( 2002 ) [ arxiv : hep - ph/0106219 ] ; s. dimopoulos and g. landsberg , phys .
lett . * 87 * , 161602 ( 2001 ) [ arxiv : hep - ph/0106295 ] , g. landsberg , arxiv : hep - ph/0211043 ; m. cavaglia , int .
j. mod .
a * 18 * , 1843 ( 2003 ) [ arxiv : hep - ph/0210296 ] ; p. kanti , int .
j. mod .
a * 19 * , 4899 ( 2004 ) [ arxiv : hep - ph/0402168 ] ; s. hossenfelder , arxiv : hep - ph/0412265 .
f. pretorius , phys .
rev . lett . * 95 * , 121101 ( 2005 ) [ arxiv : gr - qc/0507014 ] ; m. campanelli , c. o. lousto , p. marronetti and y. zlochower , phys .
* 96 * , 111101 ( 2006 ) [ arxiv : gr - qc/0511048 ] .
j. g. baker , j. centrella , d. i. choi , m. koppitz and j. van meter , phys .
lett . * 96 * , 111102 ( 2006 ) [ arxiv : gr - qc/0511103 ] ; p. diener _
et al . _ ,
* 96 * , 121101 ( 2006 ) [ arxiv : gr - qc/0512108 ] ; f. herrmann , d. shoemaker and p. laguna , arxiv : gr - qc/0601026 . h. yoshino and y. nambu , phys
. rev .
d * 67 * , 024009 ( 2003 ) [ arxiv : gr - qc/0209003 ] ; h. yoshino and v. s. rychkov , phys .
d * 71 * , 104028 ( 2005 ) [ arxiv : hep - th/0503171 ] .
n. jansen , p. diener , a. khokhlov and i. novikov , class .
grav . * 20 * , 51 ( 2003 ) [ arxiv : gr - qc/0103109 ] . | scenarios of large extra dimensions have enhanced the importance for the study of black holes in higher dimensions . in this paper , we analyze an axisymmetric system of two black holes . specifically , the bowen - york method is generalized for higher dimensions in order to calculate the initial data for head - on collision of two equal - mass black holes .
then , the initial data are evolved adopting the close - slow approximation to study gravitational waves emitted during the collision .
we derive an empirical formula for radiation efficiency , which depends weakly on the dimensionality .
possible implications of our results for the black hole formation in particle colliders are discussed . |
the interplay of superfluidity ( sf ) and localization in disordered bose systems has been the subject of intense study for two decades @xcite .
most of the theoretical investigative effort has been focused on the quantum phase transition between a superfluid and insulating phase .
for example , it has been established that disorder leads to the appearance of an ( insulating ) bose glass , sandwiched between the superfluid and mott insulating phases .
this topic is enjoying continued interest , especially since cold atom physicists have recently produced controllable disorder using laser speckles @xcite and looked at such phenomena as anderson localization of a one - dimensional condensate @xcite and the suppression of the condensate fraction in three dimensions.@xcite the idea of superglass " has come to the forefront in the context of the investigation of the ( super)solid @xmath0he . the superglass phase was initially observed in quantum monte carlo simulations , in which the superfluid phase had an inhomogeneous condensate map on a microscopic scale.@xcite biroli _
_ proved that such a superglass phase does exist ( at least as a metastable phase ) by introducing an ( artificial ) model which could be mapped to a classical system of hard spheres and studied in a controlled fashion.@xcite recent experiments on solid @xmath1he have confirmed the strong interplay between a superfluid component and a slow ( glassy ) dissipative component.@xcite however , little is known yet about the superglass phase , and specifically about the actual role of disorder in promoting or enhancing superfluidity .
give the current controversies and puzzles surrounding the interpretation of experiments on the possible supersolid phase of helium , further investigations of superfluid glassy phases are thus warranted . in this work ,
we provide strong numerical evidence for disorder - induced superfluidity in a lattice realization of hard - core bosons with a strong nearest - neighbor attraction , in the presence of external disorder .
in particular , we show that at low temperature and in a small range of attractive interactions , disorder of sufficient strength stabilizes a glassy " superfluid phase .
the superfluid density reaches a maximum and then decays as the strength of the disorder increases , as an insulating glassy phase intervenes .
in contrast to the case of repulsive bosons where disorder reduces the size of the superfluid phase , we see that strongly attracting hard - core bosons can be stabilized and made superfluid by disorder . in other words ,
disorder _ induces _ superflow in an otherwise insulating phase . aside from supersolid @xmath1he ,
such a scenario is possibly relevant to other condensed matter systems , e.g. , high - temperature superconductors,@xcite as well as to the elusive superfluid phase of molecular hydrogen,@xcite and to the role of substrate disorder in the superfluidity of ( sub)monolayer helium films.@xcite
we describe a disordered bose system by means of the following hamiltonian : @xmath2 we consider here a square lattice of @xmath3 sites , with periodic boundary conditions .
the sums @xmath4 run over all pairs of nearest - neighboring lattice sites , @xmath5 @xmath6 is the bose creation ( annihilation ) operator for a particle at site @xmath7 , @xmath8 is the local density operator .
the first term of ( [ ham ] ) represents the hopping of particles to nearest - neighboring sites . henceforth , we choose the hopping integral @xmath9 as our energy unit .
the second term represents the interaction among bosons .
hard - core _ on - site repulsion is assumed , limiting the occupation of every site to no more than one particle .
for the nearest - neighbor interaction , essentially all previous work based on ( [ ham ] ) has focused on the _ repulsive _ case , i.e. @xmath10 , chiefly to elucidate the nature of the disorder - driven superfluid to insulator transition.@xcite an _ enhancement _ of superfluidity by disorder has been predicted in some cases.@xcite here , on the other hand , we consider the case of _ attractive _ nearest - neighbor interaction ( i.e. , _ negative _
@xmath11 ) , and neglect interactions among particles lying at distances greater than nearest neighbors . with this choice , the hamiltonian ( [ ham ] ) , which is essentially a lattice model of quantum sticky " spheres , is isomorphic to that of a spin-1/2 xxz quantum ferromagnet .
we model disorder by means of a random on - site potential @xmath12 , uniformly distributed in the interval @xmath13 .
other theoretical representations of a disordered environment could be considered , e.g. , one in which the hopping matrix element @xmath9 randomly varied from site to site , but in this work we restricted ourselves to the above , widely adopted _ diagonal _ model of disorder.@xcite in the spin language , the disordering potential is equivalent to a random on - site magnetic field along the @xmath14 axis . in the absence of disorder , the ground state of ( [ ham ] ) is a superfluid for @xmath15 , whereas for @xmath16 only a mott insulating phase exists , with exactly one particle per site , regardless of lattice geometry and dimensionality .
this is simply because the system can maximally lower its energy by having each particle surrounded by as many nearest - neighboring particles as possible , trumping any contribution from the hopping term .
the regime of interest in this work is the latter , i.e. , that in which no superfluid phase exists in the absence of disorder .
we perform grand - canonical quantum monte carlo simulations to study the ground state properties of ( [ ham ] ) , using the worm algorithm in the lattice path - integral representation.@xcite as the details of this computational method are extensively described elsewhere , and because the calculations performed here are standards , we shall not review it here , and simply refer interested readers to the original references . the results shown here correspond to a temperature @xmath17 sufficiently low ( typically @xmath18 ) , to be regarded as essentially ground state estimates .
simulations are carried out over square lattice of size varying from @xmath19=12 to @xmath19=96 , and estimates are averaged over a number @xmath20 of independent realizations of the random disordering potential , typically @xmath20=100 ( 20 ) for @xmath19=12 ( 96 ) .
versus chemical potential @xmath21 for @xmath22=2.3 , for weak ( @xmath23=0.5 , squares ) , intermediate ( @xmath23=3.0 , circles ) and strong ( @xmath23=9.0 , triangles ) disorder .
results shown are for a lattice of size @xmath19=96 , and are obtained by averaging over 20 independent realizations of the disorder .
statistical errors are smaller than symbol sizes . ]
[ f1 ] shows the average particle density @xmath24 as a function of the chemical potential @xmath21 , for a particular value of @xmath22 greater than 2 ( @xmath22=2.3 ) . for weak disorder ( i.e. , small @xmath23 )
, the ground state of the system has exactly one particle per site , with an abrupt density jump at @xmath25 , when the lattice turns from empty to fully filled .
however , for disorder of sufficient strength ( figure shows results for @xmath26 ) , the density jump disappears , being replaced by a smooth curve , signaling continuous dependence of density on chemical potential . in other words ,
the disorder stabilizes phases at intermediate densities , consisting of interconnected clusters " of particles , pinned by local fluctuations of the disordering potential .
in this situation , the value @xmath25 corresponds to a particle density @xmath27 . versus particle density @xmath24 for @xmath22=2.3 and disorder strength @xmath28 .
statistical errors are smaller than symbol sizes .
results shown are for a square lattice with @xmath29 , and @xmath30=@xmath19 , and are obtained by averaging over 20 independent realizations of the disorder . the solid line is a guide to the eye .
_ inset _ : superfluid density for a fixed particle density @xmath24=0.5 , computed on square lattices of varying size @xmath19 .
extrapolation to infinite system size still gives a finite superfluid density . ] for @xmath22=2.3 and disorder strength @xmath28 .
statistical errors are smaller than symbol sizes .
results shown are for a square lattice with @xmath29 , and @xmath30=@xmath19 , and are obtained by averaging over 20 independent realizations of the disorder .
data show a weak power law decay of @xmath31 at long distances . ] clearly , the issue immediately arises of whether such disordered phases may turn superfluid at low @xmath17 , and what the nature would be of such a disordered superfluid phase , simultaneously featuring broken translational invariance .
we investigated the occurrence of superfluid behavior by directly calculating the superfluid density @xmath32 ( using the standard winding number estimator ) .
[ f2 ] shows @xmath32 as a function of particle density @xmath24 , in the limit @xmath17@xmath33 @xmath34 , for the one of the choices of model parameters of fig .
[ f1 ] , namely @xmath28 and @xmath35 .
the superfluid density increases from zero and reaches a maximum value at half filling , where approximately 12% of the system is superfluid .
obviously , numerical data such as those shown in fig . [ f2 ] must be extrapolated to the @xmath36 , in order for us to be able to male confidently the statement that superfluidity observed in these systems is not merely a finite - size effect but survives in the thermodynamic limit .
the inset of fig .
[ f2 ] shows a typical extrapolation ; estimates are shown for the superfluid density obtained for a fixed particle density @xmath27 , on square lattices of different sizes ( 12 , 24 , 48 and 96 ) , for @xmath22=2.3 and @xmath23=3 .
it is worth restating that these estimates are obtained by averaging results corresponding to several independent realizations of the disordering potential
. based on results such as those shown in the inset of fig .
[ f2 ] , we conclude that the superfluid signal remains finite in the thermodynamic limit . in general , we have observed that results obtained on a lattice with @xmath19=96 offer a close representation of the physics of the thermodynamic limit , at least in the range of parameters discussed here
. additional numerical evidence of superfluidity is extracted from the behavior of the one - particle density matrix @xmath31 , shown in fig .
[ f3 ] for the case @xmath37 and @xmath28 .
the data are consistent with quasi long - range order , as expected in two dimensions , namely a power law decay of @xmath31 at long distances .
( attained for @xmath27 ) versus disorder strength @xmath23 for different attractive interactions @xmath38 .
statistical errors are smaller than symbol sizes .
solid lines are only meant to guide the eye . ]
the observed superfluid phase is ostensibly induced by disorder , which stabilizes uniform phases of filling intermediate between zero and one . in order to gain further insight and in - depth understanding of the role of disorder in actually promoting superfluidity , it is of interest to study the competition between the strength of disorder ( @xmath23 ) and that of the attractive boson interaction ( @xmath22 ) . for definiteness ,
we consider the case of half filling , corresponding to a maximum in the superfluid density ( for those systems for which superfluidity is observed ) .
the same trends are also observed away from half filling . when the disorder is weak ( @xmath39 )
, it can not break apart clusters of particles , hence the system remains insulating , as shown in fig .
[ f4 ] or by the vanishing compressibility @xmath40 for @xmath41 in fig .
there are thus macroscopic domains ( empty or fully filled ) with hidden long - range order in the system.@xcite when the disorder becomes of the order of the attraction ( @xmath42 ) , sites and regions begin to appear throughout the system where the chemical potential is low enough to rip particles off the cluster , which then breaks down into large grains .
these particles , however , are still largely localized in the vicinity of the energetically favorable sites created by disorder , as the curve for @xmath43 in fig . [ f5 ] shows .
if we further increase the disorder strength , the grain size decreases to a microscopic scale , a ( relatively ) large fraction of the particles are delocalized , and superfluidity along interfaces ( ridges ) becomes possible , as also shown in previous numerical studies.@xcite this effect takes place essentially due to percolation .
we can thus say that the disorder counters the insulating trend caused by the attractions , and actually makes the system superfluid .
this is shown in fig .
[ f4 ] where we see a rather large superfluid fraction as a function of disorder . naturally , as the disorder strength
is increased even further , insulating glassy behavior re - appears , because the disorder is now so strong that it can block any superfluid path and localize particles , much as in the case of repulsive interactions .
a similar scenario takes place on increasing the interaction strength @xmath44 at constant disorder bound @xmath23 , as shown in fig .
we have already explained the steep decay of the curve corresponding to @xmath45 above , due to the lack of carriers .
when @xmath23 is greater than @xmath44 , the disorder is sufficiently strong to destroy all macroscopic domains , and superfluidity can occur all over the sample .
but as the disorder becomes stronger , it prevents particle world lines from winding around the lattice . in the regime of strong disorder , both the disorder and the attractive interactions contribute to suppress superfluidity , as regions with nearly uniform chemical potential will be insulating due to the strong attraction , which pulls particles together in such regions .
it is worth noting that the above scenario is quite different from that of the repulsive disordered bose - hubbard model , where regions of uniform chemical potential are crucial for stabilizing locally a liquid phase , and thus the superfluid properties and the compressibility of the whole system.@xcite we also note here that this insulating phase is compressible , as shown in fig .
[ f1 ] for @xmath46 , which justifies the nomenclature
bose glass".@xcite the compressibility at half filling goes from zero in the phase segregated regime ( no disorder ) , to a large value in the superfluid phase , and decreases then monotonically over the bose glass phase , when increasing the disorder bound at constant interaction strength .
( attained for @xmath24=0.5 ) versus absolute value of attractive interaction @xmath22 for different diagonal disorder @xmath23 at inverse temperature @xmath47 .
statistical errors are smaller than symbol sizes .
results shown are for a square lattice with @xmath29 .
the solid lines are a guide to the eye . ]
= 96 , @xmath48 and @xmath28 . the total superfluid density
@xmath32 equals 0.068(1 ) for this run .
the white areas are small insulating grains , connected by superfluid interfaces . ]
this superglass " phase can be visualized through local superfluid density maps , shown in fig . [ f6 ] , for a particular realization of disorder .
the local value is obtained by statistically averaging local contributions to the total superfluid density ( i.e. , to the square of the winding number ) , which , in the case shown in fig .
[ f6 ] for @xmath28 and @xmath49 , amounts to slightly less than 7% .
we found that the covariance between the superfluid density and the disordering potential is virtually zero , i.e. , for these values of the parameters the physics is mostly driven by the attraction between bosons , consistent with the picture given above in the case of strong disorder and strong attraction .
we have shown that disorder induces a superfluid phase in a lattice system of hard - core bosons with a strong nearest - neighbor attraction .
while the system without disorder is an insulator of the ferromagnetic ising type , the disorder can induce an inhomogeneous superfluid ( or superglass ) phase ( corresponding to in plane order in the spin parlance ) in a window of interaction and disorder strengths . for stronger disorder bounds , the disorder and
the attractive interactions work together to localize the particles .
the coherence induced by disorder might easily be observable in time - of - flight images for ultracold atoms or molecules . in the absence of disorder ,
the physics of our model is reminiscent of that of molecular _ para_-hydrogen , long speculated to be a potential second superfluid " , due to the light mass of its constituents ( one half of that of helium atoms ) . on the other hand
, superfluidity is not observed in _
para_-hydrogen due to the strength of the intermolecular potential , which causes the system to crystalize at temperatures significantly above that at which bose condensation is expected to take place.@xcite recent numerical studies@xcite have shown that disorder ought _ not _ give rise to a superfluid phase of _ para_-hydrogen .
based on the results obtained in this work , we may argue that _
para_-hydrogen may be a system too deep " into the insulating regime ( i.e. , the effective value of @xmath44 is too large ) for disorder to stabilize a superfluid phase .
on the other hand , the results obtained here suggest that disorder may be responsible for the observation of superfluidity is helium films at coverages corresponding to less than a full monolayer
. it should be noted , though , that helium films are expected to be superfluid at negative " pressure , in fact all the way down to the spinodal density,@xcite rendering them essentially different than what is discussed here .
this work was supported in part by the natural science and engineering research council of canada under research grant g121210893 , by the alberta informatics circle of research excellence and by the swiss national science foundation .
one of us ( mb ) gratefully acknowledges the hospitality of the institute of theoretical physics , university of innsbruck .
99 m. p. a. fisher , p. b. weichman , g. grinstein , d. s. fisher , phys .
b * 40 * , 546 ( 1989 ) .
j. billy , v. josse , z. zuo , a. bernard , b. hambrecht , p. lugan , d. cl ' ement , l. sanchez - palencia , p. bouyer , and a. aspect , nature * 453 * , 891 ( 2008 ) ; l. sanchez - palencia , d. cl ' ement , p. lugan , p. bouyer , and a. aspect , new j. phys . , * 10 * , 045019 ( 2008 ) . m. white , m. pasienski , d. mckay , s. zhou , d. ceperley , and b. demarco , phys . rev .
lett . * 102 * , 055301 ( 2009 ) .
g. roati , c. derrico , l. fallani , m. fattori , c. fort , m. zaccanti .
g. modugno , m. modugno and m. inguscio , nature * 453 * , 895 ( 2008 ) .
m. boninsegni , n. prokofev and b. svistunov , phys .
lett . * 96 * , 105301 ( 2006 ) .
g. biroli , c. chamon and f. zamponi , phys .
b * 78 * , 224306 ( 2008 ) . j. c. davis , private communication ( 2008 ) .
see , for instance , j. c. phillips , pnas * 105 * , 9917 ( 2008 ) .
j. turnbull and m. boninsegni , phys .
b * 78 * , 144509 ( 2008 ) .
h. cho and g. a. williams , phys .
* 75 * , 1562 ( 1995 ) ; g. a. csthy , j. d. reppy , and m. h. w. chan , phys .
lett * 91 * , 235301 ( 2003 ) .
w. krauth , n. trivedi and d. m. ceperley , phys .
lett . * 67 * , 2307 ( 1991 ) .
m. makivic , n. trivedi and s. ullah , phys .
lett . * 71 * , 2307 ( 1993 ) ; m. wallin , e. s. sorensen , s. m. girvin and a. p. young , phys .
b * 49 * , 12115 ( 1994 ) ; f. alet and e. s. sorensen , phys .
e * 67 * , 015701 ( 2003 ) ; n. prokofev and b. svistunov , phys .
* 92 * , 015703 ( 2004 ) .
n. v. prokofev , b. v. svistunov , and i. s. tupitsyn , phys .
a * 238 * , 253 ( 1998 ) ; sov .
jetp * 87 * , 310 ( 1998 ) .
k. v. houcke , l. pollet and s. m. a. rombouts , j. comp . phys .
* 225 * , 2249 ( 2007 ) .
e. t. seppl , v. petj , and m. j. alava , phys . rev .
e * 58 * , r5217 ( 1998 ) .
y. imry and s. ma , phys .
* 35 * , 1399 ( 1975 ) .
e. burovski , e. kozik , a. kuklov , n. prokofev , and b. svistunov , phys .
lett . * 94 * , 165301 ( 2005 ) .
v. l. ginzburg and a. a. sobyanin , jetp lett .
* 15 * , 242 ( 1972 ) .
m. boninsegni , m. w. cole and f. toigo , phys . rev
. lett . * 83 * , 2002 ( 1999 ) . | we use quantum monte carlo simulations to study the phase diagram of hard - core bosons with short - ranged _ attractive _ interactions , in the presence of uniform diagonal disorder .
it is shown that moderate disorder stabilizes a glassy superfluid phase in a range of values of the attractive interaction for which the system is a mott insulator , in the absence of disorder .
a transition to an insulating bose glass phase occurs as the strength of the disorder or interactions increases . |
each of the three allotropes of iron possess interesting magnetic properties which have a profound influence on their stability.@xcite the body - centered cubic ( bcc ) form is ferromagnetic;@xcite the face - centered cubic ( fcc ) fe at low temperatures is antiferromagnetic,@xcite but its higher energy ferromagnetic state can be thermally excited with complex variations in magnetic structure as a function of temperature.@xcite ferromagnetism is eliminated when bcc iron transforms into the hexagonal close - packed ( hcp ) crystal structure at high pressures @xmath3gpa.@xcite density - functional theory has been proven to be reliable in estimating the magnetic properties of iron - based transition - metal alloys , for example the pioneering work by akai _ _ et al.__@xcite who used the korringar - kohn - rostoker ( kkr ) green s function method based on the local spin density approximation ( lsda ) , and anismov _ et al._,@xcite who used the linear muffin - tin orbital ( lmto ) green s function method to obtain a consistent picture of the electronic and magnetic structures of @xmath0 impurities in bcc fe . however , the lsda is known to underestimate exchange effects,@xcite which are crucial for determining magnetism .
furthermore , lattice optimization , which may affect the electronic and magnetic structures of the impurity and host , was not taken into account in these previous calculations .
lattice optimization may not change the general conclusions drawn previously , but can give more quantitative information .
the relative stability of @xmath0 impurities in bcc fe is also important for practical implementations in the steel industry which can be precisely calculated at the optimized lattice parameters and for this purpose , the lattice parameters of the @xmath0 impurities are also optimized .
we focus also on the magnetic interactions between impurity and the host element which can give a detailed picture of the magnetism of impurities in bcc fe .
the purpose of the present work was to use the highly precise all - electron full - potential linearized augmented plane - wave ( flapw ) method@xcite based on the generalized gradient approximation ( gga)@xcite to study the magnetic and electronic structures and hyperfine fields of fe where it is substitutionally alloyed with the @xmath0 transition metals .
the longer term aim is to be able to contribute to the development of the so - called electrical steels which are used in the manufacture of motors and transformers .
supercell of the primitive cell of bcc fe , which contains 27 atoms .
the fe1 , fe2 , and fe3 atoms are represented by black , white , and grey spheres , respectively , while the impurity atom ( corner atom ) is marked with x. ] a @xmath4 supercell of the primitive bcc cell containing 27 fe atoms ( fig .
[ supercell ] ) was used with @xmath0 solute atoms ( @xmath5 ) substituting a body - centered iron atom , giving a composition @xmath6fe@xmath7 equivalent to a concentration of @xmath8 , which is consistent with the levels of solute added to steels .
the impurity atom @xmath1 has first , second and third nearest neighbors designated fe1 , fe2 , and fe3 respectively .
the kohn - sham equations@xcite were solved in the framework of the flapw@xcite method under the gga.@xcite an energy cutoff at @xmath9@xmath10 , where @xmath11 is the lattice parameter of each calculation , employed for expanding the linearized augmented plane wave ( lapw ) basis set , corresponding to @xmath12lapws per * k*-point and spin .
a @xmath13@xmath10 cutoff was used for the star functions depicting the charge density and potential in the interstitial regions .
lattice harmonics with @xmath14 were employed to expand the charge density , potential , and wave - functions inside each muffin - tin ( mt ) sphere of radius @xmath15a.u .
for all the atoms .
integrations inside the brillouin zone ( bz ) were performed using the improved tetrahedron method@xcite over a @xmath16 mesh within the three dimensional ( 3d ) bz , corresponding to @xmath17*k * points inside the irreducible wedge of the 3d - bz .
all core electrons were at first treated fully relativistically and valence states scalar relativistically , _
i.e. _ , without spin - orbit coupling ( soc).@xcite for spin - orbit coupling on valence states , we employed the second variation method@xcite with the spin diagonal parts of the density subjected to a self - consistency loop . during the second variation procedure
, integrations inside the 3d - bz were done in the full - bz , _
i.e. _ , 1099 .
the explicit orthogonalization ( xo ) scheme was employed to ensure the orthogonality between the core and valence states.@xcite all atoms were fully relaxed at each lattice volume until the atomic forces on each atom were less than @xmath18 the equilibrium lattice constants and bulk moduli @xmath19 were determined by fitting the total energy and volume to the birch - murnaghan equation of states.@xcite using the optimized lattice constants , further calculations were carried out in the spin - unpolarized and spin - polarized states with and without soc .
self - consistency was assumed when the difference between input and output charge densities became less than @xmath20electrons / a.u.@xmath21 note that all the computational parameters used in the present calculations satisfy the convergence test.@xcite
table [ lattice ] and fig . [ latticestructure ] show the optimized lattice parameters @xmath11 ( in units of ) and bulk moduli @xmath19 ( in units of gpa ) of @xmath6fe@xmath7 supercell at zero kelvin .
the optimized lattice constant of pure bcc fe is @xmath22 , which is only @xmath23% smaller than that of the finite temperature experimental value @xmath24.@xcite it is interesting to find out that the all the substitutional @xmath0 impurities in bcc fe increases the lattice parameter .
the calculated bulk modulus of bcc fe , @xmath25gpa is also comparable with the experimentally observed value.@xcite we can see @xmath19 has minimum when the @xmath26 band is half filled,_i.e . _ , mn .
we also analyzed the local lattice expansion / contraction around the @xmath1 impurities and we arrive at a conclusion that the lattice distortions around the @xmath1 impurities are negligible .
however , one can easily calculate the bond lengths of fe-@xmath1 , using the lattice parameters given in table [ lattice ] .
.calculated lattice parameters @xmath11 ( in units of ) , bulk moduli @xmath19 ( in units of gpa ) , and formation enthalpy per atom ( @xmath27 in units of ev / atom ) of @xmath6fe@xmath7 .
numbers in the parentheses are the formation enthalpy in units of kj / atom - mol . [
cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] figure [ deltah ] illustrates the changes , @xmath28 , @xmath29 , and @xmath30 in the fermi contact hyperfine fields of the fe1 , fe2 , and fe3 atoms , respectively , with the normalization by the host fermi contact hyperfine field , @xmath31 .
positive values imply that the hyperfine field of fe@xmath32 neighbor is smaller than that of the host and vice versa .
the calculated @xmath28 for ni and co are negative , in agreement with experiments,@xcite whereas @xmath28 is inconsistent with the calculations done by dederichs _
et al._,@xcite who found a zero value for @xmath28 .
this discrepancy is due to partly the different computational method .
our calculated gga trend of @xmath33 can be compared with the experimental observations and the other previous calculations.@xcite the @xmath34 results in table [ table - b ] show that @xmath35 of fe1 is negative for all the @xmath0 impurities except for cr , and these positive quantities increase the transfer field of the fe1 atoms through its proportionality to the magnetic moment of the surrounding atoms , _
i.e. _ , the fe1 , fe2 , and fe3 atoms .
hence , the transfer field of the fe2 atom is smaller than that of the fe1 atom and this supports the notion that the reduced @xmath2 of bcc fe alloyed with cr and mn is caused mainly by the fe2 atoms .
reported @xmath34 values for pure bcc fe calculated using lsda and gga are , without soc , @xmath36 and @xmath37 kg , respectively , and , with soc , @xmath38 kg and @xmath39 kg , respectively.@xcite our gga result is @xmath40 kg ( and @xmath41 kg with soc ) , which is close to the experimental value of @xmath42kg.@xcite the somewhat larger values obtained here are consistent with flapw - lsda value of @xmath43 kg reported by ohnishi.@xcite table [ table - b ] shows again that soc does not significantly affect the calculations because of the localization of the impurity .
finally , some test calculations were also carried out for big and small supercells of fe@xmath44ni@xmath45 and fe@xmath46ni and the local properties around ni were similar to fe@xmath7ni .
this indicateds that @xmath4 supercell of the primitive bcc cell as discussed in sec .
[ sec : method ] is sufficient for @xmath8at.% impurities in bcc fe , but may not be sufficient for the other host , _
e.g. _ bcc cr.@xcite
the magnetism and electronic structures of @xmath0 impurities in body - centered iron have been investigated in terms of the flapw method based on the generalized gradient approximation .
the results compare favorably with existing experimental data , and have been contrasted with published calculations where available .
the lattice parameters were optimized and it is shown that the lattice constant decreases with the atomic number of the impurity @xmath1 ( sc fe ) , whereas it increases for co zn .
the calculated formation energy results are good agreement with the alloying experiences in steels design .
it is found that iron induces a magnetic moment at the sc , ti , v , cu , and zn atoms substituted into the iron lattice ; sc , ti , v , cr , and zn couple antiferromagnetically with iron , resulting in an overall reduction in the saturation magnetization relative to pure iron .
in contrast , co , and ni couple ferromagnetically with iron and resulting in increase in saturation magnetization .
the trend of the atom averaged magnetic moment versus effective atomic number is shown to lie on the left branch of the well - known slater - pauling curve , but it is not simply linear the deviation from the standard slater - pauling curve is attributed by the nature of the weak ferromagnetism of bcc iron .
these results have been interpreted by examining in detail the local density of states , near neighbor interactions , and other features of the electronic structure of the alloys resulting from the substitution of impurity atoms into iron .
an important outcome is that calculations of the type reported here are not significantly influenced by spin - orbit coupling . the calculated fermi contact hyperfine fields also support this conclusion . in the light of previous theoretical calculations , where lsda was used our calculations
showed that although lsda can not describe the correct magnetic properties of bcc fe , but can describe the correct magnetic properties ( qualitatively ) of bcc fe because our gga results are close to the previous lsda results . comparing the lsda and gga , we showed that the gga results are closer to the experimental data .
this work was supported by the steel innovation program by posco , the wcu ( world class university ) program ( project no .
r32 - 2008 - 000 - 10147 - 0 ) , the basic science research program ( grant no .
2009 - 0088216 ) through the national research foundation funded by the ministry of education , science and technology of republic of korea .
the present work was also funded by office of scientific research ( grant no . fa9550 - 07 - 1 - 0013 ) and in part by the u.s . air force . | the magnetic and electronic structures of @xmath0 impurity atoms from sc to zn in ferromagnetic body - centered cubic iron are investigated using the all - electron full - potential linearized augmented plane - wave method based on the generalized gradient approximation ( gga ) .
we found that in general , the gga results are closer to the experimental values than those of the local spin density approximation .
the calculated formation enthalpy data indicate the importance of a systematic study on the ternary fe - c-@xmath1 systems rather than the binary fe-@xmath1 systems , in steel design .
the lattice parameters are optimized and the conditions for spin polarization at the impurity sites are discussed in terms of the local stoner model .
our calculations , which are consistent with previous work , imply that the local spin - polarizations at sc , ti , v , cu , and zn are induced by the host fe atoms .
the early transition - metal atoms couple antiferromagnetically , while the late transition - metal atoms couple ferromagnetically , to the host fe atoms .
the calculated total magnetization ( @xmath2 ) of bcc fe is reduced by impurity elements from sc to cr as a result of the antiferromagnetic interaction , with the opposite effect for solutes which couple ferromagnetically .
the changes in @xmath2 are attributed to nearest neighbor interactions , mostly between the impurity and host atoms .
the atom averaged magnetic moment is shown to follow generally the well - known slater - pauling curve , but our results do not follow the linearity of the slater - pauling curve .
we attribute this discrepancy to the weak ferromagnetic nature of bcc fe .
the calculated fermi contact hyperfine fields follow the trend of the local magnetic moments .
the effect of spin - orbit coupling is found not to be significant although it comes into prominence at locations far from the impurity sites . *
physical review b * 81 * , 184423 ( 2010 ) * |
in just over 5 years , the ibis instrument on board the integral gamma - ray satellite has revolutionised our classical view of high mass x - ray binaries ( hmxb ) . specifically , ibis has discovered many new hmxbs hosting massive ob supergiant stars ( the so - called supergiant high mass x - ray binaries , sgxb ) ; in just a few years their population has been almost tripled thanks to the ibis identification of two previously unrecognized subclasses of sgxbs : the highly obscured sgxbs @xcite and the supergiant fast x - ray transients ( sfxts ; @xcite , @xcite ) . around 9 sfxts are known to date .
they spend most of the time in a low level of x - ray activity characterized by x - ray luminosity values in the range 10@xmath210@xmath3 erg s@xmath4 , well below the persistent bright state of other classical sgxbs ( @xmath510@xmath6 erg s@xmath4 ) . only occasionally do they display bright and fast x - ray outbursts with a duration no longer than a few days ( typically only a few hours ) and a dynamical range in flux of 10@xmath710@xmath8 .
a number of hypotheses have been proposed to explain the peculiar flaring behaviour of sfxts , both as a class of objects on their own , and in the context of persistently accreting supergiant x - ray binaries .
some of the models proposed invoke structure in the wind of the supergiant companion .
this structure could be either in the form of clumping of spherically symmetric outflow from the supergiant donor , or in the form of an equatorially density enhanced wind from the supergiant , inclined at some angle to the orbit of the neutron star .
futhermore , variations in the orbital eccentricity @xcite can be used to explain the difference between supergiant x - ray binary systems showing transient as opposed to persistent x - ray emission .
a different model proposed by @xcite makes use of transitions between accretion gating mechanisms , such as centrifugal and magnetic barriers , brought about by variations in the stellar wind , to explain the large dynamic range in flux observed in these systems .
fast outbursts from sfxts are relatively rare occurances ( i.e. a few are detected per year ) , and long periods of inactivity are interspersed here and there with short flares spaced by irregular intervals of time . because of this , the identification of periodicity is a very challenging task which can be fullfilled only through long - term and continous monitoring . to date , while recurrent ( but non - periodic ) outbursts have been observed in many sfxts , only one , igr j11215@xmath05952 has been shown to exhibit periodic flaring activity separated by regular intervals of @xmath5 165 days .
the identification of periodicity in the outburst behaviour of sfxts , most likely represents the orbital period of the binary system , and is a key diagnostic for studying the geometry of the system ( i.e. orbital radius , eccentricity ) and hence for understanding the physical reasons behind their very unusual x - ray behaviour . in this letter
we report a comprehensive temporal study of the sfxt sax j1818.6@xmath01703 , leading to the discovery of a @xmath5 30 day periodicity , making this the second confirmed periodic sfxt .
sax j1818.6@xmath01703 was reported for the first time by bepposax / wfc on march 11 , 1998 during one of the galactic bulge observation campaigns @xcite .
the peak flux was @xmath5400 mcrab ( 9 - 25 kev ) and the outburst lasted for 0.1 days and was characterised by a fast decay . on september 9 , 2003 ,
the ibis telescope on integral detected 2 short and intense outbursts with fluxes reaching 400 mcrab in the 18 - 45 kev energy band .
this source activity was reported by @xcite as showing a complex time profile lasting for about 1 day , while the main event duration was 2.7 hours .
spectra at different evolutionary phases extending up to @xmath5100 kev were fitted with thermal bremsstrahlung models .
the nature of the source was still unclear but rxte and integral data exhibiting such outbursts were considered indicative of a possible sfxt @xcite , a picture later strengthened by @xcite on the basis of an optical / nir counterpart with an i - band spectrum typical of an ob supergiant .
this result was confirmed soon afterwards by @xcite , with a chandra observation revealing only one bright and variable source within the error region of sax j1818.6@xmath01703 , and 0.21 and 0.47 from the 2mass and usno - b1.0 positions of the proposed counterpart .
the authors reported a strongly absorbed ( @xmath9 ) power law spectrum with a photon index of 1.9 and a 0.5 - 10 kev flux of @xmath10 and also suggested the source may not be a true transient but a persistent hmxb with luminous flares such as 4u 1907 + 097 or vela x-1 . confirmed the ob supergiant nature of the counterpart and inferred a distance of 2.5kpc .
the source was not detected by xmm during a 13 ks observation on october 8 , 2006 .
following the new report of activity by @xcite , a swift too was activated on april 18 , 2008 and the flux value recorded was consistent with the chandra observation , while the xmm upper limit is a factor of 80 below these detections . with these new findings
, @xcite concluded that sax j1818.6@xmath01703 shows an x - ray flux dynamic range similar to that observed for other sfxts .
ibis / isgri data is organised and analysed in short pointings ( science windows , scw ) of typically 2000s duration .
we analysed ibis / isgri data for all available science windows covering the date range from march 12 , 2003 to march 12 , 2008 ( mjd range 52710 - 54537 ) in the 18 - 60 kev energy band using the standard osa 7 analysis tools to provide intensity and significance images for each science window .
we excluded data flagged as taken during bad time intervals , and additionally removed science windows with high image rms .
thereafter , we extracted fluxes from each science window at the position of sax j1818.6@xmath01703 to create a light curve at science window resolution . for periodicity analysis , science windows where sax j1818.6@xmath01703 was more than 12 degrees off - axis
were excluded to avoid any effects due to incomplete off - axis response correction .
after filters were applied , the ibis dataset consisted of 3756 scw for a total effective exposure of 6.1 ms .
the swift / bat 15 - 50 kev light curve for sax j1818.6@xmath01703 was obtained from the bat transient monitor programme @xcite , and covered the date range from may 10 , 2005 to august 23 , 2008 ( mjd range 53500 - 54705 ) .
data were filtered to include only pointings with data quality flag 0 ( good quality ) . after filtering , 10975 individual pointings with a total effective exposure of 5.6 ms
were retained .
a standard test for periodicity was carried out using the lomb - scargle method on the ibis combined light curves in the 18 - 60 kev band .
an indication of periodic signal at around 30 days was found , but with a significant degree of aliasing and a strong first harmonic ( figure [ fig : ls ] , upper panel ) . ]
some aliasing is not unusual in ibis light curves , because of the window function imposed by the approximately 6-monthly observation pattern for sources near the galactic bulge .
furthermore , significant leakage of power into the harmonic is usually considered an indicator of non - sinusoidal modulation , which the lomb - scargle algorithm must compensate for by the addition of higher frequency components . to provide an independent verification of the suspected periodicity , the same analysis was performed on the swift / bat 15 - 50 kev light curve ( figure [ fig : ls ] , lower panel ) . in this case
, the aliasing is much reduced due to the more even temporal coverage provided by the large bat field of view , but the power of the periodicity is substantially reduced due to the lower sensitivity per pointing compared to ibis / isgri .
nevertheless , the periodicity at @xmath530 days is still detected at above 99.9% significance in both light curves . in order to investigate further ,
the detected periodicity was verified by the use of an epoch - folding method on the ibis / isgri light curve , this method being less dependent on the periodic flux profile .
this search again resulted in a clear maximum at around 30 days ( figure [ fig : efsearch ] ) .
fitting the peak yields an estimate of both the period and its associated error , specifically @xmath11 ( 1@xmath12 ) days . ]
the ibis light curve folded on the 30 day period ( figure [ fig : folded ] ) shows a very significant modulation , and also shows that the source has a duty cycle of around 4 - 6 days during each 30 day period .
simulations performed on data with the same duty cycle verified that the strong harmonics observed in the lomb - scargle analysis are due to this profile . in order to ensure that the large outburst at around mjd 52891 was not dominating the overall folded light curve , the analysis was repeated removing the largest outburst ( specifically data for mjds 52889 - 52893 ) . figure [ fig : folded ] also shows the folded ibis light curve with the outburst removed , and there is no substantial difference in shape with or without the large outburst .
indeed , the ibis light curve analysed represents some 70 cycles of the @xmath530d period , and should be relatively robust against any such dominance by a single outburst .
furthermore , the analysis in section [ section : recurrence ] shows that sax j1818.6@xmath01703 has a high degree of recurrence , indicating that this folded light curve should be a reasonable representation of the average outburst . .
knowledge of the period and phasing of the known outbursts is of great benefit when examining the light curves for events of lower significance .
a search was carried out for other outbursts with similar characteristics to the known ones in both the ibis and bat light curves .
a significance was determined for a 4-day period around each of the expected outburst times where sufficient data were available .
table [ table : newbursts ] reports detections above 7 sigma in either instrument , and the corresponding significance in the other instrument lightcurve . [ cols="<,^,^,^,^,^,^ " , ] the results of the x - ray observations carried out to date are consistent with this orbital period , and for reference the phase ( arbitrarily referenced to the bat trigger at mjd 54540.659 ) of the various x - ray observations are indicated in figure [ fig : folded ] .
it can be seen that the xmm observation , which reported a firm upper limit on the flux , occurred at a phase of 0.51 , i.e. apastron .
thus this measurement can be interpreted as setting an upper limit on the quiescent flux from the system . by contrast , the chandra observation was performed at phase 0.89 , and the reported monotonic rising flux appears to show the onset of emission , but whether this is a slow increase as the compact object approached periastron , or the onset of an outburst , we can not say .
the xrt observation was performed at a phase of 0.11 and indicates some low - level activity , compatible with the chandra flux , and an order of magnitude above the xmm upper limit .
we performed a similar analysis using the rxte / asm light curve .
unfortunately , while we can conclude that rxte / asm is detecting x - ray emission from the source , it does not provide any further diagnostic information due to its limited instantaneous sensitivity to a source that is only rarely above 20mcrab , and time coverage that does not really manage to capture the detail of the flaring time structure .
a systematic search of the ibis data for both the large outburst and newly discovered outbursts has been performed to look for pulsations in the range 0.2 - 5000s .
no statistically significant pulse period has been detected by this analysis and indeed no pulsations from sax j1818.6@xmath01703 have been reported in the literature .
we conclude that although sax j1818.6@xmath01703 was originally detected as a transient , and later categorised as a sfxt , knowledge of the orbital solution enables us to see a rather different picture .
sax j1818.6@xmath01703 is a supergiant system with a compact object in a 30 day orbit , and shows regular emission on at least 50% of the periastron passages .
moreover , some periastron passages give rise to larger outbursts , during which fast flaring structure is evident in the light curve - it is this structure which has led to the identification as a sfxt .
sax j1818.6@xmath01703 shows an extreme dynamic range of emission , from @xmath13 ( 0.5 - 10 kev ) at apastron @xcite , to @xmath14 by extrapolation of the best - fit spectrum during the largest recorded ibis outburst @xcite to 0.5 - 10kev , which again supports the sfxt categorisation .
the simplest interpretation of this behaviour is of a compact object in an eccentric orbit , periodically accreting from the wind of a donor supergiant . as the second sfxt to be shown to have a periodic outbursting behaviour
, sax j1818.6@xmath01703 adds to our overall picture of the supergiant binary systems , placed between the persistent systems with circular orbits , and the long - period system igr j11215@xmath05952 which shows outbursts on a 165 day period .
this indicates that orbital characteristics may play an important role in the accretion mechanisms of these systems , much as they do in be transient systems
. the shorter period of sax j1818.6@xmath01703 has enabled us to determine that the system shows a high degree of recurrence .
the system appears similar to igr j18483@xmath00311 , which has an orbital period of 18.52 days , spin period of 21.0526 s , and has recently been confirmed as a supergiant system @xcite .
other sfxt systems may eventually reveal similar behaviour , but with more eccentric orbits and/or lower recurrence rates which have so far prevented the determination of their orbital periods .
we encourage detailed follow - up observations based on this orbital period , such as have proved so effective for igr j11215@xmath05952 .
a detailed picture of the temporal structures seen during both faint and bright periastron passages should prove an invaluable probe into the environment from which the compact object is accreting .
based on observations with integral , an esa project with instruments and science data centre funded by esa member states ( especially the pi countries : denmark , france , germany , italy , switzerland , spain ) , czech republic and poland , and with the participation of russia and the usa .
hjw thanks the school of physics and astronomy , university of southampton for financial support .
we acknowledge the following funding : in italy , asi - inaf contract i/023/05/0 and i/088/06/0 ; in the uk , stfc grant gr/2002/00446 ; a.b . and v.s .
acknowledge the support of nature ( 455 , 835 - 836 ) and thank the editors for increasing the international awareness of the current critical situation in italian research . | sax j1818.6@xmath01703 has been characterised as a supergiant fast x - ray transient system on the basis of several integral / ibis detections since the original bepposax wide field camera detection . using ibis / isgri , swift / bat and archival observations ,
we show that in fact sax j1818.6@xmath01703 exhibits emission on a period of 30 @xmath1 0.1 days , with a high degree of recurrence .
sax j1818.6@xmath01703 is therefore the second sfxt shown to exhibit periodic outbursts , but with a considerably shorter period than the other known system , igr j11215@xmath05952 .
[ firstpage ] gamma - rays : observations x - rays : binaries x - rays : individual : sax j1818.6 - 1703 |
difficult combinatorial landscapes are found in many important problems in physics , computing , and in common everyday life activities such as resource allocation and scheduling .
for example , spin - glass systems give rise to such energy landscapes which are characterized by many local minima and high energy barriers between them .
these landscapes generally show frustration , i.e. frozen disorder where the system is unable to relax into a state in which all constraints are satisfied . in completely different fields , such as combinatorial optimization ,
similar hard problems also arise , for example the well - known _ traveling salesman problem _ and many others . in order to understand the reasons that make these problems difficult to optimize ,
a number of model landscapes have been proposed .
one of the simplest yet representative example is kauffman s family of @xmath0 landsdcapes @xcite .
the @xmath0 family of landscapes is a problem - independent model for constructing multimodal landscapes that can gradually be tuned from smooth to rugged , where the term `` rugged '' is intuitively related to the degree of variability in the objective function value in neighboring positions in configuration space .
the more rugged the landscape , the higher the number of local optima , and the landscape becomes correspondingly more difficult to search for the global optimum .
the idea of an @xmath0 landscape is to have @xmath2 spins " or `` genes '' , each with two possible values , @xmath3 or @xmath4 .
the model is a real stochastic function @xmath5 defined on binary strings @xmath6 of length @xmath2 , @xmath7 .
the value of @xmath1 determines how many other spin values in the string influence a given spin @xmath8 .
the value of @xmath5 is the average of the contributions @xmath9 of all the spins : @xmath10 by increasing the value of @xmath1 from 0 to @xmath11 , @xmath0 landscapes can be tuned from smooth to rugged . for @xmath12
all contributions can be optimized independently which makes @xmath5 a simple additive function with a single maximum . at the other extreme when @xmath13 the landscape becomes completely random ,
the probability of any given configuration of being the optimum is @xmath14 , and the expected number of local optima is @xmath15 .
intermediate values of @xmath1 interpolate between these two cases and have a variable degree of `` epistasis '' , i.e. of spin ( or gene ) interaction @xcite .
the @xmath1 variables that form the context of the fitness contribution of gene @xmath16 can be chosen according to different models .
the two most widely studied models are the _ random neighborhood _ model , where the @xmath1 variables are chosen randomly according to a uniform distribution among the @xmath11 variables other than @xmath16 , and the _ adjacent neighborhood _ model , in which the @xmath1 variables are those closest to @xmath16 in a total ordering @xmath17 ( using periodic boundaries ) .
no significant differences between the two models were found in terms of global properties of the respective families of landscapes , such as mean number of local optima or autocorrelation length @xcite .
similarly , our preliminary studies on the characteristics of the @xmath0 landscape optima networks did not show noticeable differences between the two neighborhood models .
therefore , we conducted our full study on the more general random model .
the @xmath0 model is related to spin glasses , and more precisely to @xmath18-spin models @xcite , where @xmath18 plays a role similar to @xmath1 . in spin glasses
the function analogous to @xmath5 is the energy @xmath19 and the stable states are the minima of the energy hypersurface . in this study we seek to provide fundamental new insights into the structural organization of the local optima in combinatorial landscapes , particularly into the connectivity of their basins of attraction .
combinatorial landscapes can be seen as a graph whose vertices are the possible configurations .
if two configurations can be transformed into each other by a suitable operator move , then we can trace an edge between them . the resulting graph , with
an indication of the fitness at each vertex , is a representation of the given problem fitness landscape .
a useful a simplification of the graphs for the energy landscapes of atomic clusters was introduced in @xcite .
the idea consists in taking as vertices of the graph not all the possible configurations , but only those that correspond to energy minima . for atomic clusters these are well - known , at least for relatively small assemblages .
two minima are considered connected , and thus an edge is traced between them , if the energy barrier separating them is sufficiently low . in this case
there is a transition state , meaning that the system can jump from one minimum to the other by thermal fluctuations going through a saddle point in the energy hyper - surface .
the values of these activation energies are mostly known experimentally or can be determined by simulation . in this way
, a network can be built which is called the `` inherent structure '' or `` inherent network '' in @xcite .
we propose a network characterization of combinatorial fitness landscapes by adapting the notion of _ inherent networks _ described above .
we use the family of @xmath0 landscapes as an example . in our case
the inherent network is the graph where the vertices are all the local maxima and the edges account for transition probabilities between maxima .
we exhaustively extract such networks on representative small @xmath0 landscape instances , and perform a statistical characterisation of their properties .
our analysis is inspired , in particular , by the work of @xcite on energy landscapes , and in general , by the field of complex networks @xcite .
a related work can be found in @xcite , where the case of lattice polymer chains is studied . however , the notion of an edge there is very different , being related to moves that bring a given conformation into an allowed neighboring one .
similar ideas have been put forward in physical chemistry to understand the thermodynamics and kinetics of complex biomolecules through the network study of their free - energy landscapes @xcite .
it should also be noted that our approach is different from the barrier - tree representations of landscapes proposed by stadler et al .
( see , for example , @xcite ) .
the next section describes how combinatorial landscapes are mapped onto networks , and includes the relevant definitions and algorithms used in our study .
the empirical analysis of our selected @xmath0 landscape instances is presented in the following two sections ; one devoted to the study of basins , and the other to the network statistical features .
finally , we present our conclusions and ideas for future work .
many natural and artificial systems can be modeled as networks .
typical examples are communication systems ( the internet , the web , telephonic networks ) , transportation lines ( railway , airline routes ) , biological systems ( gene and protein interactions ) , and social interactions ( scientific co - authorship , friendships ) .
it has been shown in recent years that many of these networks exhibit was has been called a _ small - world _
topology @xcite , in which nodes are highly clustered yet the path length between them is small
. additionally , in several of these networks the distribution of the number of neighbours ( the degree distribution ) is typically right - skewed with a heavy tail " , meaning that most of the nodes have less - than - average degree whilst a small fractions of hubs have a large number of connections .
these topological features are very relevant since they impact strongly on networks properties such as their robustness and searchability . to model a physical energy landscape as a network , doye @xcite needed to decide first on a definition both of a state of the system and how two states were connected
. the states and their connections will then provide the nodes and edges of the network . for systems with continuous degrees of freedom , this was achieved through the ` inherent structure ' mapping . in this mapping
each point in configuration space is associated with the minimum ( or ` inherent structure ' ) reached by following a steepest - descent path from that point .
this mapping divides configurations into basins of attraction surrounding each minimum on the energy landscape .
our goal is to adapt this idea to the context of combinatorial optimization . in our case
, the vertexes of the graph can be straightforwardly defined as the local maxima of the landscape .
these maxima are obtained exhaustively by running a best - improvement local search algorithm ( see fig . [ algohc ] ) from every configuration of the search space . the definition of the edges , however , is a much more delicate matter . in our initial attempt
@xcite we considered that two maxima @xmath20 and @xmath21 were connected ( with an undirected and unweighed edge ) , if there exists at least one pair of direct neighbors solutions @xmath16 and @xmath22 , one in each basin of attraction ( @xmath23 and @xmath24 ) .
we found empirically on small instances of @xmath0 landscapes , that such definition produced densely connected graphs , with very low ( @xmath25 ) average path length between nodes for all @xmath1 .
therefore , apart from the already known increase in the number of optima with increasing @xmath1 , no other network property accounted for the increase in search difficulty .
furthermore , a single pair of neighbours between adjacent basins , may not realistically account for actual basin transitions occurring when using common heuristic search algorithms .
these considerations , motivated us to search for alternative definitions of the edges connecting local optima .
in particular , we decided to associate weights to the edges that account for the transition probabilities between the nodes ( local optima ) .
more details on the relevant algorithms and formal definitions are given below .
* definition : * fitness landscape @xcite . +
a landscape is a triplet @xmath26 where @xmath27 is a set of potential solutions i.e. a search space , @xmath28 , a neighborhood structure , is a function that assigns to every @xmath29 a set of neighbors @xmath30 , and @xmath31 is a fitness function that can be pictured as the _ height _ of the corresponding solutions . in our study , the search space is composed by binary strings of length @xmath2 , therefore its size is @xmath32 .
the neighborhood is defined by the minimum possible move on a binary search space , that is , the 1-move or bit - flip operation . in consequence , for any given string @xmath33 of length @xmath2 , the neighborhood size is @xmath34 . the @xmath35 algorithm to determine the local optima and
therefore define the basins of attraction , is given below : choose initial solution @xmath36 choose @xmath37 such that @xmath38 @xmath39 * definition : * local optimum . + a local optimum is a solution @xmath40 such that @xmath41 , @xmath42 .
the @xmath35 algorithm defines a mapping from the search space @xmath27 to the set of locally optimal solutions @xmath43 .
+ * definition : * basin of attraction .
+ the basin of attraction of a local optimum @xmath44 is the set @xmath45 .
the size of the basin of attraction of a local optimum @xmath20 is the cardinality of @xmath23 .
notice that for non - neutral fitness landscapes , as are @xmath0 landscapes , the basins of attraction as defined above , produce a partition of the configuration space @xmath27 .
therefore , @xmath46 and @xmath47 @xmath48 , @xmath49 + * definition : * edge weight .
+ for each pair of solutions @xmath33 and @xmath50 , let us define @xmath51 as the probability to pass from @xmath33 to @xmath50 with the given neighborhood structure . in the case of binary strings of size @xmath2 , and the neighborhood defined by the single bit - flip operation , there are @xmath2 neighbors for each solution , therefore : if @xmath37 , @xmath52 and + if @xmath53 , @xmath54 .
we can now define the probability to pass from a solution @xmath55 to a solution belonging to the basin @xmath24 , as : @xmath56 notice that @xmath57 .
thus , the total probability of going from basin @xmath23 to basin @xmath24 is the average over all @xmath58 of the transition probabilities to solutions @xmath59 : @xmath60 @xmath61 is the size of the basin @xmath23 .
we are now prepared to define our ` inherent ' network or network of local optima . + * definition : * local optima network .
+ the local optima network @xmath62 is the graph where the nodes are the local optima , and there is an edge @xmath63 with weight @xmath64 between two nodes @xmath20 and @xmath21 if @xmath65 .
notice that since each maximum has its associated basin , @xmath66 also describes the interconnection of basins . according to our definition of edge weights
, @xmath67 may be different than @xmath68 .
thus , two weights are needed in general , and we have an oriented transition graph .
the following two definitions are relevant to the discussion of the boundary of basins . *
definition : * boundary of a basin of attraction .
+ the _ boundary _ @xmath69 of a basin of attraction @xmath70 can be defined as the set of configurations within a basin that have at least one neighbor s solution in another basin @xmath71 . *
definition : * interior of a basin of attraction .
+ conversely , the _ interior _ @xmath72 of a basin is composed by the configurations that have all their neighbors in the same basin . formally , @xmath73 @xmath74
in order to avoid sampling problems that could bias the results , we used the largest values of @xmath2 that can still be analyzed exhaustively with reasonable computational resources .
we thus extracted the local optima networks of landscape instances with @xmath75 , and @xmath76 . for each pair of @xmath2 and @xmath1 values ,
30 randomly generated instances were explored .
therefore , the networks statistics reported below represent the average behaviour of 30 independent instances . besides the maxima network
, it is useful to describe the associated basins of attraction as these play a key role in search algorithms .
furthermore , some characteristics of the basins can be related to the optima network features .
the notion of the basin of attraction of a local maximum has been presented before .
we have exhaustively computed the size and number of all the basins of attraction for @xmath77 and @xmath78 and for all even @xmath1 values plus @xmath13 . in this section ,
we analyze the basins of attraction from several points of view as described below . in fig .
[ nbasins ] we plot the average size of the basin corresponding to the global maximum for @xmath77 and @xmath78 , and all values of @xmath1 studied .
the trend is clear : the basin shrinks very quickly with increasing @xmath1 .
this confirms that the higher the @xmath1 value , the more difficult for an stochastic search algorithm to locate the basin of attraction of the global optimum [ !
ht ] , title="fig:",width=302 ] + [ !
ht ] + [ tab : cumulsize ] @xmath1 & @xmath79 & @xmath80 & @xmath81 + 2 & @xmath82 & @xmath83 & @xmath84 + 4 & @xmath85 & @xmath86 & @xmath87 + 6 & @xmath88 & @xmath89 & @xmath90 + 8 & @xmath91 & @xmath92 & @xmath93 + 10 & @xmath94 & @xmath95 & @xmath96 + 12 & @xmath97 & @xmath98 & @xmath99 + 14 & @xmath100 & @xmath101 & @xmath102 + 15 & @xmath103 & @xmath104 & @xmath105 + + 2 & @xmath106 & @xmath107 & @xmath108 + 4 & @xmath109 & @xmath110 & @xmath111 + 6 & @xmath112 & @xmath113 & @xmath114 + 8 & @xmath115 & @xmath116 & @xmath117 + 10 & @xmath118 & @xmath119 & @xmath120 + 12 & @xmath121 & @xmath122 & @xmath123 + 14 & @xmath124 & @xmath125 & @xmath126 + 16 & @xmath127 & @xmath128 & @xmath129 + 17 & @xmath130 & @xmath131 & @xmath132 + fig .
[ bas-18-size ] shows the cumulative distribution of the number of basins of a given size ( with regression line ) for a representative instances with @xmath78 , @xmath133 .
table [ tab : cumulsize ] shows the average ( of 30 independent landscapes ) correlation coefficients and linear regression coefficients ( intercept ( @xmath134 and slope ( @xmath81 ) ) between the number of nodes and the basin sizes for instances with @xmath135 and all for all the studied values of @xmath1 .
notice that distribution decays exponentially or faster for the lower @xmath1 and it is closer to exponential for the higher @xmath1 .
this observation is relevant to theoretical studies that estimate the size of attraction basins ( see for example @xcite ) .
these studies often assume that the basin sizes are uniformly distributed , which is not the case for the @xmath0 landscapes studied here . from the slopes @xmath136 of the regression lines ( table [ tab : cumulsize ] ) one can see that high values of @xmath1 give rise to steeper distributions ( higher @xmath137 values ) .
this indicates that there are less basins of large size for large values of @xmath1 . in consequence ,
basins are broader for low values of @xmath1 , which is consistent with the fact that those landscapes are smoother . [ !
ht ] + + the scatter - plots in fig .
[ fig : cor_fit - size ] illustrate the correlation between the basin sizes of local maxima ( in logarithmic scale ) and their fitness values .
two representative instances for @xmath2 = 18 and @xmath1 = 4 , 8 are shown .
notice that there is a clear positive correlation between the fitness values of maxima and their basins sizes .
in other words , the higher the peak the wider tend to be its basin of attraction .
therefore , on average , with a stochastic local search algorithm , the global optimum would be easier to find than any other local optimum .
this may seem surprising .
but , we have to keep in mind that as the number of local optima increases ( with increasing @xmath1 ) , the global optimum basin is more difficult to reach by a stochastic local search algorithm ( see fig .
[ nbasins ] ) .
this observation offers a mental picture of @xmath0 landscapes : we can consider the landscape as composed of a large number of mountains ( each corresponding to a basin of attraction ) , and those mountains are wider the taller the hilltops .
moreover , the size of a mountain basin grows exponentially with its hight .
we now briefly describe the statistical measures used for our analysis of maxima networks .
the standard clustering coefficient @xcite does not consider weighted edges .
we thus use the _ weighted clustering _ measure proposed by @xcite , which combines the topological information with the weight distribution of the network : @xmath138 where @xmath139 , @xmath140 if @xmath141 , @xmath142 if @xmath143 and @xmath144 . for each triple formed in the neighborhood of the vertex @xmath20 , @xmath145 counts the weight of the two participating edges of the vertex @xmath20 .
@xmath146 is defined as the weighted clustering coefficient averaged over all vertices of the network .
the standard topological characterization of networks is obtained by the analysis of the probability distribution @xmath147 that a randomly chosen vertex has degree @xmath148 . for our weighted networks , a characterization of weights
is obtained by the _ connectivity and weight distributions _
@xmath149 and @xmath150 that any given edge has incoming or outgoing weight @xmath151 . in our study
, for each node @xmath20 , the sum of outgoing edge weights is equal to @xmath4 as they represent transition probabilities .
so , an important measure is the weight @xmath152 of self - connecting edges ( remaining in the same node ) .
we have the relation : @xmath153 . the vertex _ strength _ , @xmath16 ,
is defined as @xmath154 , where the sum is over the set @xmath155 of neighbors of @xmath20 @xcite .
the strength of a node is a generalization of the node s connectivity giving information about the number and importance of the edges .
another network measure we report here is _ disparity _
@xcite @xmath156 , which measures how heterogeneous are the contributions of the edges of node @xmath20 to the total weight ( strength ) : @xmath157 the disparity could be averaged over the node with the same degree @xmath148 .
if all weights are nearby of @xmath158 , the disparity for nodes of degree @xmath148 is nearby @xmath159 . finally , in order to compute the average shortest path between two nodes on the optima network of a given landscape , we considered the expected number of bit - flip mutations to pass from one basin to the other .
this expected number can be computed by considering the inverse of the transition probabilities between basins . in other words , if we attach to the edges the inverse of the transition probabilities , this value would represent the average number of random mutations to pass from one basin to the other . more formally , the distance ( expected number of bit - flip mutations ) between two nodes is defined by @xmath160 where @xmath161 .
now , we can define the length of a path between two nodes as being the sum of these distances along the edges that connect the respective basins .
@xmath162 & @xmath163 & @xmath164 & @xmath165 + + 2 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 + 4 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 + 6 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + 8 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 + 10 & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath190 + 12 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 + 13 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath190 + + 2 & @xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 + 4 & @xmath205 & @xmath206 & @xmath207 & @xmath208 & @xmath209 + 6 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 + 8 & @xmath215 & @xmath216 & @xmath217 & @xmath218 & @xmath219 + 10 & @xmath220 & @xmath221 & @xmath222 & @xmath199 & @xmath223 + 12 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 + 14 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 + 15 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + + 2 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 + 4 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + 6 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 + 8 & @xmath254 & @xmath255 & @xmath256 & @xmath257 & @xmath258 + 10 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 + 12 & @xmath264 & @xmath265 & @xmath266 & @xmath267 & @xmath268 + 14 & @xmath269 & @xmath270 & @xmath271 & @xmath272 & @xmath273 + 16 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 + 17 & @xmath279 & @xmath280 & @xmath281 & @xmath277 & @xmath282 + in this section we study in more depth some network features which can be related to stochastic local search difficulty on the underlying fitness landscapes .
table [ tab : statistics ] reports the average ( over 30 independent instances for each @xmath2 and @xmath1 ) of the network properties described .
@xmath283 and @xmath162 are , respectively , the mean number of vertices and the mean number of edges of the graph for a given @xmath1 rounded to the next integer .
@xmath163 is the mean weighted clustering coefficient .
@xmath164 is the mean disparity , and @xmath165 is the mean path length .
the fourth column of table [ tab : statistics ] lists the average values of the weighted clustering coefficients for all @xmath2 and @xmath1 .
it is apparent that the clustering coefficients decrease regularly with increasing @xmath1 for all @xmath2 . for the standard unweighed clustering
, this would mean that the larger @xmath1 is , the less likely that two maxima which are connected to a third one are themselves connected .
taking weights , i.e. transition probabilities into account this means that either there are less transitions between neighboring basins for high @xmath1 , and/or the transitions are less likely to occur .
this confirms from a network point of view the common knowledge that search difficulty increases with @xmath1 . [ !
ht ] + +
the average shortest path lengths @xmath165 are listed in the sixth column of table [ tab : statistics ] . fig .
[ fig : distances ] ( top ) is a graphical illustration of the average shortest path length between optima for all the studied @xmath0 landscapes .
notice that the shortest path increases with @xmath2 , this is to be expected since the number of optima increases exponentially with @xmath2 .
more interestingly , for a given @xmath2 the shortest path increases with @xmath1 , up to @xmath284 , and then it stagnates and even decreases slightly for the @xmath285 . this is consistent with the well known fact that the search difficulty in @xmath0 landscapes increases with @xmath1 .
however , some paths are more relevant from the point of view of a stochastic local search algorithm following a trajectory over the maxima network . in order to better illustrate the relationship of this network property with the search difficulty by heuristic local search algorithms , fig .
[ fig : distances ] ( bottom ) shows the shortest path length to the global optimum from all the other optima in the landscape .
the trend is clear , the path lengths to the optimum increase steadily with increasing @xmath1 . [
cols="^ " , ]
we have proposed a new characterization of combinatorial fitness landscapes using the well - known family of @xmath0 landscapes as an example .
we have used an extension of the concept of inherent networks proposed for energy surfaces @xcite in order to abstract and simplify the landscape description . in our case
the inherent network is the graph where the nodes are all the local maxima and the edges accounts for transition probabilities ( using the 1-flip operator ) between the local maxima basins of attraction .
this mapping leads to oriented weighted graphs , instead of the more commonly used unordered and unweighed ones @xcite .
we believe that the present representation is closer to the view offered by a monte carlo search heuristic such as simulated annealing which produces a trajectory on the configuration space based on transition probabilities according to a boltzmann equilibrium distribution@xcite . we have exhaustively obtained these graphs for @xmath286 , and for all even values of @xmath1 , plus @xmath13 , and conducted a network analysis on them . the network representation of the @xmath0 fitness landscapes has proved useful in characterizing the topological features of the landscapes and
gives important information on the structure of their basins of attraction .
in fact , our guiding motivation has been to relate the statistical properties of these networks , to the search difficulty of the underlying combinatorial landscapes when using stochastic local search algorithms ( based on the bit - flip operator ) to optimize them .
we have found clear indications of such relationships , in particular : the clustering coefficients : : : suggest that , for high values of @xmath1 , the transition between a given pair of neighboring basins is less likely to occur .
the shortest paths to the global optimum : : : become longer with increasing @xmath2 , and for a given @xmath2 , they clearly increase with higher @xmath1 . the outgoing weight distribution
: : : indicate that , on average , the transition probabilities from a given node to neighbor nodes are higher for low @xmath1 . the incoming weight distribution
: : : indicate that , on average , the transition probabilities from the neighborhood of a node become lower with increasing @xmath1 .
the disparity coefficients : : : reflect that for high @xmath1 the transitions to other basins tend to become equally likely , which is an indication of the randomness of the landscape .
the previous results clearly confirm and justify from a novel network point of view the empirically known fact that @xmath0 landscapes become harder to search as they become more and more random with increasing @xmath1 .
the construction of the maxima networks requires the determination of the basins of attraction of the corresponding landscapes .
we have thus also described the nature of the basins , and found that the size of the basin corresponding to the global maximum becomes smaller with increasing @xmath1 .
the distribution of the basin sizes is approximately exponential for all @xmath2 and @xmath1 , but the basin sizes are larger for low @xmath1 , another indirect indication of the increasing randomness and difficulty of the landscapes when @xmath1 becomes large .
furthermore , there is a strong positive correlation between the basin size and the degree of the corresponding maximum , which confirms that the synthetic view provided by the maxima graph is a useful .
finally , we found that the size of the basins boundaries is roughly the same as the size of basins themselves .
therefore , nearly all the configurations in a given basin have a neighbor solution in another basin .
this observation suggests a different landscape picture than the smooth standard representation of 2d landscapes where the basins of attraction are visualized as real mountains .
some of these results on basins in @xmath0 landscapes were previously unknown @xcite .
this study represents our first attempt towards a topological and statistical characterization of easy and hard combinatorial landscapes , from the point of view of complex networks analysis .
much remains to be done .
first of all , the results found should be confirmed for larger instances of @xmath0 landscapes .
this will require good sampling techniques , or theoretical studies since exhaustive sampling becomes quickly impractical .
other landscape types should also be examined , such as those containing neutrality , which are very common in real - world applications , and especially the landscapes generated by important hard combinatorial problems such as the traveling salesman problem and other resource allocation problems .
work is in progress for neutral versions of @xmath0 landscapes and for knapsack problems .
finally , the landscape statistical characterization is only a step towards implementing good methods for searching it .
we thus hope that our results will help in designing or estimating efficient search techniques and operators .
p. f. stadler .
fitness landscapes . in m.
lssig and valleriani , editors , _ biological evolution and statistical physics _ , volume 585 of _ lecture notes physics _ , pages 187207 , heidelberg , 2002 .
springer - verlag . | we propose a network characterization of combinatorial fitness landscapes by adapting the notion of inherent networks proposed for energy surfaces .
we use the well - known family of @xmath0 landscapes as an example . in our case
the inherent network is the graph whose vertices represent the local maxima in the landscape , and the edges account for the transition probabilities between their corresponding basins of attraction .
we exhaustively extracted such networks on representative @xmath0 landscape instances , and performed a statistical characterization of their properties .
we found that most of these network properties are related to the search difficulty on the underlying @xmath0 landscapes with varying values of @xmath1 . |
= knuth and yao @xcite showed that the expected number of independent bernoulli @xmath1 random bits needed to generate an integer - valued random variable @xmath2 whose distribution is given by , where @xmath3 , is at least equal to the binary entropy of @xmath2 : @xmath4 they also exhibited an algorithm dubbed the ddg tree algorithm for which the expected number of random bernoulli @xmath5 bits is not more than @xmath6 . by grouping
, one can thus develop algorithms for generating batches of @xmath7 independent copies of @xmath2 such that the expected number of bernoulli @xmath1 random bits per random variable does not exceed @xmath8 .
while these results settle the discrete random variate case quite satisfactorily , the generation of continuous or mixed random variables has not been treated satisfactorily in the literature .
the objective of this note is to study the number of bernoulli @xmath1 random bits to generate a random variable @xmath9 with a given precision @xmath10 , provided that we can define `` precision '' in a satisfactory manner .
note that any algorithm that takes as input the accuracy parameter @xmath10 , returns a random variable @xmath11 where @xmath12,@xmath13,
,@xmath14 are independent identically distributed ( or i.i.d . )
bernoulli @xmath1 bits , @xmath15 is the number of bits needed , and @xmath16,@xmath17 , are given sequences of functions . for a vector @xmath18 ,
let @xmath19 denote the @xmath20-norm of @xmath21 for @xmath22 : @xmath23 .
for @xmath24 , the @xmath25-norm is @xmath26 . with @xmath27 ,
all @xmath28-norms are the same for @xmath29 $ ] . + an algorithm with accuracy @xmath30 is such that for some coupling of @xmath2 ( the target random variable ) and @xmath31 , @xmath32 where @xmath28 is usually @xmath33 or @xmath25 .
this natural notion of accuracy corresponds to the wasserstein @xmath0 metric between two probability measures . for a random variable @xmath2 ,
we denote by @xmath34 the distribution of @xmath2 .
let @xmath35 denote the space of all distributions of pairs of random variables with fixed marginal distributions @xmath36 and @xmath37 , respectively .
then the wasserstein @xmath0 distance between @xmath2 and @xmath38 , or between @xmath36 and @xmath37 , is @xmath39 where @xmath40 denotes the essential supremum .
this is a distance metric : @xmath41 if @xmath42 then there exists a random variable @xmath38 ( output ) coupled with @xmath2 ( target ) such that , i.e. , with probability one , @xmath43 .
this definition of distance should satisfy simulation professionals in the sense that if their calculations require the evaluation of @xmath44 where @xmath45, ,@xmath46
are given independent random variables and @xmath47 is a real - valued functions , then , with probability one , @xmath48 which is usually a quantity that can easily be controlled .
we believe that software packages should have the capability of accepting @xmath30 as input parameter in random variate generation .
it is interesting that @xmath49 , the expected value of @xmath15 , can be related to the entropy almost in the way knuth and yao did for discrete random variables in @xcite .
our note provides the foundational background for such a study in terms of universal lower bounds and various useful upper bounds for particular algorithms .
we include examples for the main distributions .
several authors have adressed the problem of arbitrary precision in sampling algorithms .
these include flajolet and saheb @xcite , who explain how to generate the first @xmath50 bits of an exponential random variable for an integer @xmath51 .
karney @xcite describes an algorithm for the standard normal distribution .
thesis @xcite also discusses arbitrary precision sampling .
the quantity that appears in our lower bound is the partition entropy .
more precisely , let @xmath52 be a partition of @xmath53 and let @xmath10 be a fixed parameter for the precision .
the partition entropy of @xmath2 with respect to @xmath52 is the quantity @xmath54 while our results apply to all distributions , we will mainly focus on absolutely continuous distributions , i.e. , random variables @xmath2 with density @xmath55 .
we recall the definition of differential entropy : @xmath56 the differential entropy can be ill - defined , @xmath57 , finite or @xmath58 . for more information on differential entropy and entropy in general , one can read cover and thomas @xcite .
when @xmath2 has compact support , then the case @xmath58 can not occur .
when @xmath55 is bounded , then the case @xmath57 is excluded .
when @xmath59 , it can be shown that @xmath60 is well - defined and either finite or @xmath57 ; see rnyi @xcite , csiszr @xcite for a proof .
our main result is the following : let @xmath61 be a target random vector with density @xmath55 , and assume that . consider any algorithm that for given @xmath62 outputs @xmath63 using @xmath15 random fair coins , such that @xmath64 .
then @xmath65 for @xmath66 and @xmath24 , the third term in the lower bound is @xmath67 and @xmath68 , respectively . for @xmath27 , it is @xmath69 .
the second part of this note describes several algorithms that come within a constant term of this lower bound , and therefore , are basically optimal if grouping is used for generation . before tackling all
that , we introduce a brief section in which we recall the main exact sampling algorithms for discrete random variables , and their properties , as these will be essential for the understanding of the main algorithms .
in this section , we give simple proofs of two important results for generating discrete random variables the optimal algorithm of knuth and yao @xcite and the more practical but slightly suboptimal algorithm of han and hoshi @xcite .
we recall that we want to sample @xmath2 with probability vector @xmath70 .
every @xmath71 is decomposed into its binary representation @xmath72 consider the new random variable @xmath73 with probability vector @xmath74 with @xmath75 .
any algorithm for generating a discrete random variable using a source of i.i.d .
fair coins @xmath76 , @xmath77 , and that is based on a stopping time @xmath15when it returns an output can be viewed as a binary tree , where @xmath76 , @xmath77 , uniquely determines an infinite path in the tree by the rule `` @xmath78 '' is left and `` @xmath69 '' is right .
we refer to this general class of algorithm as tree - based algorithms
they include all possible practical algorithms .
leaves in this tree correspond to outputs .
the algorithm of knuth and yao can be implemented by a binary tree , a ddg tree , in which each leaf at level @xmath79 corresponds to a `` @xmath69 '' bit @xmath80 in ( 2 ) .
one randomly walks down this tree starting at the root and reaches the leaf for @xmath81 with probability . at that point ,
the value `` @xmath82 '' is returned , and indeed , @xmath83 as required .
a tree - based algorithm is optimal if it minimizes @xmath49 for a given probability vector .
[ kythm ] the expected number of bits of an optimum tree - based algorithm for sampling @xmath84 is bounded from below by @xmath85 and from above by @xmath86 . given a probability vector @xmath87 with @xmath88 possibly infinite , let the binary expansion of @xmath71 for @xmath89 be @xmath90 if @xmath15 denotes the number of bits required by an optimal algorithm for sampling @xmath87 , then for @xmath91 @xmath92 therefore , @xmath93 we now show that the quantity within parentheses of line ( [ crosseteton1 ] ) is lower bounded by @xmath94 and upper bounded by @xmath95 and then the result follows . for convenience ,
let @xmath96 $ ] and its binary expansion be @xmath97 to complete the proof , it remains to prove that @xmath98 if @xmath7 is the first non - zero coefficient of the binary expansion of @xmath99 , then there are two cases : either ( 1 ) @xmath100 or ( 2 ) @xmath101 .
the inequalities are strict for case 2 since @xmath102 .
for the first case , @xmath103 , and line ( [ crosseteton2 ] ) is obviously true . for the second case , @xmath104 then for the upper bound , @xmath105 for the lower bound , @xmath106 we now recall the han - hoshi algorithm published in @xcite that implements the inversion method . given @xmath107 with @xmath88 countably finite or infinite , the algorithm partitions the interval @xmath108 $ ] into a countable collection of disjoint subintervals @xmath109 where @xmath110 the idea behind the algorithm is to iteratively refine a random interval @xmath111 and to stop when @xmath112 for a certain @xmath89 .
the inversion principle says that if @xmath113 uniformly distributed on @xmath108 $ ] , then the unique @xmath89 such that @xmath114 is distributed according to @xmath115 . for a binary random source of unbiased i.i.d .
bits , their algorithm is as follow : @xmath116 @xmath117 @xmath118 @xmath119 @xmath120 @xmath121 @xmath122 @xmath123 return @xmath82 .
the following two figures [ exec_fig_hanhoshi1 ] and [ exec_fig_hanhoshi2 ] are examples that show the underlying ddg tree during the execution of the han and hoshi algorithm . . the cumulative values are @xmath124 , @xmath125 , @xmath126 , @xmath127 , @xmath128 , @xmath129 , and @xmath130 . ] [ exec_fig_hanhoshi1 ] such that @xmath131 , @xmath132 , and @xmath133 . ]
let @xmath15 be the number of random coins needed and also the number of iteration of the `` repeat '' loop .
for @xmath134 , @xmath135 . to every node ( internal or external ) corresponds an interval @xmath136 .
the root corresponding to the interval @xmath137 . for each internal node corresponds an interval @xmath136 that is not contained in one of the interval @xmath138 , and , if the source produces @xmath139 , then the left child corresponds to the interval @xmath140 and , if @xmath141 , then the right child corresponds to @xmath142 .
each leaf ( external node ) corresponds to an interval @xmath136 entirely contained in @xmath138 upon which the integer @xmath82 is returned with probability @xmath143 .
the following theorem was proved by han and hoshi @xcite : [ hhthm ] for the han - hoshi algorithm , @xmath144 our ( new ) proof partitions the leaves @xmath145 for symbol @xmath82 in the ddg tree arbitrarily into two sets , @xmath146 and @xmath147 , such that @xmath146 and @xmath147 each possess at least one leaf per level .
let @xmath148 , @xmath149 , where @xmath150 is the probability attached to leaf @xmath151 , i.e. .
we have @xmath152 . by nesting and elementary calculations ,
@xmath153 let @xmath154 be the @xmath79-th bit in the binary expansion of @xmath155 , and let @xmath156 be the @xmath79-th bit for @xmath157 .
then @xmath158 as in the proof of theorem [ kythm ] , we have @xmath159 so that , using ( [ adenozine ] ) , @xmath160
in this section , we give a lower bound for the complexity of sampling any continuous distribution to an arbitrary precision .
let @xmath52 be a countable partition of @xmath53 , and let @xmath10 be a fixed precision parameter .
consider the infinite graph @xmath37 with as vertices the sets @xmath161 , and as edges all pairs @xmath162 such that @xmath163 therefore , if @xmath164 is not an edge of @xmath37 , then @xmath165 for all @xmath166 , @xmath167 .
let @xmath168 be the maximal degree of any vertex of @xmath37 .
we now state a lemma that we shall use in conjunction with the knuth - yao result , mentioned and reproved in the previous section , in order to prove our main theorem mentioned in the introduction .
[ lowerbndlem ] let @xmath2 be a target random vector of @xmath53 .
let @xmath38 be an output with the property that , with probability one , .
let @xmath15 be the number of bits used to generate @xmath38 by an algorithm .
then @xmath169 where @xmath52 and @xmath168 are as above .
we can maximize the bound from lemma [ lowerbndlem ] , of course , by selecting the most advantageous partition @xmath52 and @xmath168 combination .
the bound from lemma [ lowerbndlem ] coincides with the bound in shannon @xcite when the distribution @xmath2 is discrete with a finite number of atoms since , in that case , @xmath170 by choosing @xmath30 sufficiently small .
let @xmath2 and @xmath38 be two ( dependent ) random variables of @xmath53 , and denote by . note that @xmath171 if @xmath164 is not an edge of @xmath37 .
thus @xmath172
if @xmath15 is the random number of bits needed to generate a discrete random variable @xmath38 that outputs a vertex @xmath173 of @xmath37 with probability @xmath174 , then @xmath175 and therefore @xmath176 it is interesting to recall a general result from csiszr @xcite about the hypercube partition entropy of an absolutely continuous random vector @xmath2 that will become useful later .
of particular interest to us is the cubic partition @xmath177 partitioned by @xmath178 .
the cells of this partition are of the form @xmath179 we recall that if @xmath180 has finite entropy a condition we refer to as rnyi s condition then if @xmath2 has a density @xmath55 , @xmath181 is well - defined , i.e. , it is either finite or @xmath57 .
we have [ absctsrenyithm ] under rnyi s condition , for general partition @xmath52 , and random variable @xmath9 with density @xmath55 , @xmath182 where @xmath183 denotes the lebesgue measure .
in particular , @xmath184 fix @xmath161 .
if @xmath73 is uniform on @xmath173 and @xmath185 , then @xmath186 thus @xmath187 and , by jensen s inequality and the concavity of , @xmath188 the inequality follows by summing over @xmath161 .
[ csiszar_lemma ] let @xmath9 have density @xmath55 , and let rnyi s condition be satisfied .
if @xmath189 , then as @xmath190 , @xmath191 if @xmath192 , then as @xmath190 , @xmath193 the fifth theorem of csiszr @xcite stipulates that if @xmath59 and @xmath55 is not absolutely continuous , then , as @xmath194 , @xmath193 for more information about the asymptotic theory for the entropy of partitioned distributions as the partitions become finer , one can consult rnyi @xcite , csiszr @xcite , csiszr @xcite , and linder and zeger @xcite .
[ lowerbndthm ] let @xmath9 have density @xmath55 .
let @xmath38 be an output with the property that with probability one , . then , under rnyi s condition , @xmath195 where @xmath196 is the volume of the unit ball in @xmath53 , and @xmath15 is the number of random bits needed to generate to @xmath38 .
let @xmath177 be a cubic partition .
then @xmath197 where @xmath198 is the maximal degree in the graph on defined by connecting @xmath199 with @xmath200 if .
we set @xmath201 and use @xmath202 observe that if @xmath203 denotes the @xmath20-ball of radius @xmath204 centered at @xmath78 , then by elementary geometric considerations , @xmath205 so that as @xmath206 , @xmath207 also , @xmath208 so that @xmath209
consider a random variable @xmath9 .
we call a partition @xmath210 a @xmath30-partition if for every set @xmath211 , there exists @xmath212 ( called the center ) such that @xmath213 then any algorithm that selects @xmath211 with probability @xmath214 can be used to generate a random variable @xmath38 that approximate @xmath2 to within @xmath30 . after generating @xmath173 , set @xmath215 then , necessarily , there is a coupling @xmath216 with . if the selection of @xmath173 is done with the help of the method of knuth and yao , then , if @xmath15 still denotes the number of random bits required , @xmath217 for @xmath24 , we can take @xmath218 , the cubic partition with sides @xmath219 . for @xmath27 ,
a simple partition into intervals of length @xmath219 can be used for all values of @xmath28 .
if @xmath2 has a density @xmath55 and @xmath24 or @xmath27 , then the procedure suggested above has , as @xmath220 , @xmath221 where in the last step we assume rnyi s condition and .
compare with the lower bound @xmath222 and note that the difference is @xmath223 . for later reference ,
we recall these values of @xmath60 for the main distributions : @xmath224\hspace{-1mm}:\phantom{12 } & \mathcal{e}(f)=0,\\ \text{exponential}(1)\hspace{-1mm}:\phantom{12 } & \mathcal{e}(f)=\log_{2}(e),\\ \text{normal}(0,1)\hspace{-1mm}:\phantom{12 } & \mathcal{e}(f)=\log_{2}\sqrt{2\pi{}e}.\end{aligned}\ ] ] recall that for @xmath225 , @xmath226 , a scale factor @xmath227 shows up as @xmath228 in the upper and lower bounds because @xmath229 for general @xmath230 , we can take @xmath231 the cubic partition with sides @xmath232 . under rnyi s condition and @xmath189
, we have @xmath233 the difference with the lower bound is @xmath234 using @xmath235 , @xmath236 , we obtain @xmath237 which unfortunately increases linearly with @xmath68 . to avoid this growing differential
which we did not have for @xmath24it seems necessary to consider partitions that better approximate @xmath20-balls .
the inversion method for generating a random variable @xmath2 with distribution function @xmath36 uses the property that @xmath238 has distribution function @xmath36 , where @xmath239 denotes the inverse , and @xmath113 is uniform @xmath108 $ ] .
one can use this method as a basis for generating an approximation using only a few random bits .
in particular , if @xmath240 and @xmath241 , @xmath242 , are independent bernoulli@xmath1 random variables , then setting @xmath243 then @xmath244 note that @xmath245 , @xmath246 .
graphically , we have the number of random coins is @xmath247 if we define @xmath248 then @xmath2 and @xmath38 are coupled in such a way that @xmath249 the @xmath15 defined above is also the number of bits needed to generate @xmath38 .
observe that the inversion method requires @xmath239 in a `` black box '' , also called an oracle . on the other hand
, it avoids the cumbersome calculation of the cell probabilities and the set - up of the knuth - yao ddg tree , and thus shines by its simplicity . in spirit
, the inversion method mimics the method of han and hoshi , and indeed , this observation leads to a simple bound .
let @xmath250 be a cubic partition of @xmath251 into intervals of equal length @xmath219 .
denote the probabilities of these intervals by @xmath252 assume that we select an interval from @xmath250 following this law by the method of han and hoshi using random bits @xmath241 , @xmath242 , @xmath253 , also used in the inversion method .
it is easy to see that the number of bits needed before halting in the inversion method is smaller .
therefore , for the inversion method , we have @xmath254 from this and lemma [ csiszar_lemma ] , we conclude [ inversion_borne_sup_thm ] if @xmath2 has a density @xmath55 satisfying rnyi s condition and if @xmath255 , then as @xmath256 , @xmath257 comparing theorem [ inversion_borne_sup_thm ] with the universal lower bound @xmath258 we see that the difference is @xmath259 .
the partition - based method has un upper for @xmath49 that is one less .
moreover , the simplicity of the inversion method can not be underestimated .
in addition , one can tighten the analysis under additional conditions on @xmath55 such unimodality , monotonicity , or for specific forms .
[ thm_inv_mono_bornee ] assume that @xmath2 has a bounded nonincreasing density @xmath55 on @xmath260 .
then for the inversion method , as @xmath256 , @xmath261 define @xmath262 , @xmath263 as in figure [ fig_inversion ] .
then @xmath264 now , @xmath265 ( even if the latter integral is infinite ) .
the theorem follows if we can show that @xmath266 . to this end , note that @xmath267 for a fixed value of @xmath268 , we see that @xmath269 only if @xmath2 falls in the interval that `` captures '' the value @xmath270 , if such an interval exists . but the probability of each interval is precisely @xmath271 .
however , if @xmath272 , then no such interval exists .
thus , @xmath273 for the uniform @xmath108 $ ] density , we have @xmath274 , and so the bound of theorem [ thm_inv_mono_bornee ] is @xmath275 however the `` o(1 ) '' can be omitted in this case as the following simple calculation shows : @xmath276 theorem [ thm_inv_mono_bornee ] improves over the bound for the partition method ( for @xmath27 ) by @xmath277 . under other regularity conditions
, one can hope to obtain similar bounds that beat the partition bound . for the exponential density , the inversion method yields @xmath278 where @xmath279 flajolet and saheb @xcite proposed a method for the exponential law that has @xmath280 where @xmath281 as @xmath256 . for the normal law ,
karney @xcite proposes a method that addresses the variable approximation issue but does not offer explicit bounds .
inversion would yield @xmath282 but the drawback is that this requires the presence of ( an oracle for ) @xmath239 , the inverse gaussian distribution function .
even the partition method requires a nontrivial oracle , namely @xmath36 . to sidestep this , one can use a slightly more expensive method based on the box - mller @xcite , which states that the pair of random variables @xmath283 with @xmath284 exponential and @xmath285 uniform on the unit circle , provides a standard gaussian in @xmath286 of zero mean and unit covariance matrix .
the random variable @xmath287 is maxwell , i.e. , it has density @xmath288 , @xmath289 , and its differential entropy is @xmath290 where @xmath291 is the euler - mascheroni constant .
we sketch the procedure , which also serves as an example for more complicated random variate generation problems .
assume that the two normals are required with @xmath30-accuracy each ( this corresponds to the choice of @xmath292 and @xmath24 ) .
then we first generate a maxwell random variable @xmath293 by inversion , noting that @xmath294 the maxwell random variable @xmath293 is needed with @xmath295-accuracy .
the maxwell law is unimodal with mode at @xmath296 .
its left piece has probability @xmath297 .
so we first pick a piece randomly using on average no more than two bits .
the we apply inversion on the appropriate piece . by theorem [ thm_inv_mono_bornee ] , we use @xmath298 random bits where @xmath299 the generated approximation is called @xmath300 . next we generate a uniform random variable @xmath301 with accuracy @xmath302 the generated value @xmath303 has @xmath304 . since @xmath113 has differential entropy @xmath305 , we see that the expected number of bits , @xmath306 , needed is bounded by @xmath307 then we return @xmath308 and claim that jointly , @xmath309 and similarly for the cosine .
next , @xmath310 putting everything together , we see that the total expected number of bits is not more than @xmath311 the lower bound for generating two independent gaussians is @xmath312 less , i.e. , @xmath313
turning a sequence of i.i.d .
random variables @xmath45 , @xmath314 , into a sequence of i.i.d .
bernoulli @xmath1 bits has been the subject of many papers .
the setting of interest to us is the following .
let @xmath315 , @xmath316 , be a ( possibly infinite ) number of cumulative distributions supported on the positive integers .
let @xmath317 be a fixed probability vector .
let @xmath2 , @xmath45 , @xmath314 , be i.i.d .
random integers drawn from @xmath317 . given @xmath318 for @xmath319 ,
draw @xmath320 independently from the distributions @xmath321 , @xmath322 ,
, @xmath323 . as a special case
, we have the classical setting when @xmath324 and then @xmath325 , @xmath326 , , @xmath327 are i.i.d . according to @xmath315 .
let @xmath315 , @xmath316 , have binary entropies given by @xmath328 , @xmath329 , , all assumed to be finite .
in other words the entropy of @xmath330 is denoted by @xmath331 .
assume also that @xmath332 [ quibivilortrinuloxifene ] there exits an algorithm ( described below ) that , upon input @xmath333 , @xmath334 ,
, @xmath335 outputs a sequence of @xmath336 i.i.d .
bernoulli @xmath1 bits where @xmath337 furthermore , these bits are independent of @xmath338 .
theorem [ quibivilortrinuloxifene ] describes how many perfect random bits we can extract from @xmath325 , @xmath326 , , @xmath327 , i.e. , @xmath336 should be near the information content , the entropy of @xmath339 . not surprisingly then
, the way to achieve this can be inspired by the optimal or near - optimal methods of compression , and , in particular , arithmetic coding . note that one can assume @xmath340 for where @xmath341 are i.i.d .
uniform random variables on @xmath108 $ ] . for all @xmath342 ,
let the binary expansions of @xmath71 be @xmath343 for convenience , @xmath344 for all @xmath345 .
also , for all @xmath342 , let @xmath346 following the methodology of arithmetic coding , associate a uniform@xmath108 $ ] random variable @xmath113 ( with binary expansion @xmath347 ) with the ( infinite ) data sequence @xmath348 .
the bits @xmath349 , @xmath350 , are i.i.d .
bernoulli@xmath1 and independent of @xmath45 , @xmath314 , @xmath351 , to be more precise , consider this algorithm . a sequence of pairs @xmath333 , ,
@xmath335 with @xmath352 and @xmath353 as described previously for @xmath354 .
@xmath355 @xmath356 @xmath357 @xmath358 @xmath359 @xmath360 to verify the correctness of algorithm [ algo_extractx ] , the intervals @xmath361 $ ] are nested .
more precisely , for all @xmath362 , @xmath363 \supseteq [ u_{\ell+1}^{-},u_{\ell+1}^{+}].\ ] ] define @xmath364 , and note @xmath365=\bigcap_{\ell=1}^{n}{[u_{\ell}^{-},u_{\ell}^{+}]}.\ ] ] for every iteration , we have @xmath366.\ ] ] since @xmath367 , @xmath368 the bits @xmath369 , @xmath242 , ,
@xmath370 are clearly i.i.d .
let @xmath371 and consider the two cases @xmath372 and @xmath373 .
we show that @xmath268 is concentrated around @xmath374 . before considering the two cases , we compute the useful quantity @xmath375 note that @xmath376 the pairs @xmath377 are i.i.d .
and therefore , by the previous calculation , @xmath378 by the law of large numbers , for all @xmath10 , .
we have @xmath379 by the law of large numbers , for all @xmath10 , if @xmath380 , @xmath381 for an arbitrary fixed integer @xmath382 , @xmath383 which is as small as desired by the choice of @xmath50 .
the @xmath384 term is due to the fact that the event + occurs only if and @xmath385 .
assume given a random variable @xmath386 with fixed probability vector @xmath70 of finite entropy denoted by @xmath387 .
assume that we employ a given ddg tree based algorithm for the generation of @xmath2 .
in this tree , let @xmath388 be the set of leaves and let @xmath389 be the label of leaf @xmath390 .
define @xmath391 let @xmath392 be the depth of @xmath390 .
then we have @xmath393 if the algorithm returns a variable @xmath2 , then we know that we must have exited via a leaf in @xmath394 .
given @xmath395 , each exit - leaf has a given probability : @xmath396 let us call the ( random ) exit - leaf @xmath38 .
the ddg tree algorithm thus returns a pair @xmath216 .
we have @xmath397 for example , for the knuth - yao ddg algorithm , we have @xmath398 , while for the han - hoshi algorithm , we have @xmath399 .
our method of batch generation will be valid for any ddg tree with finite @xmath400 .
the algorithm for batch generating i.i.d .
random variables @xmath45 , @xmath314 , , @xmath401 uses an atomic operation ` fetchbit ` that first gets a bit from a queue @xmath402 if the this queue is not empty , and otherwise it gets a bit from a bernoulli @xmath1 generator .
it is understood that that ` fetchbit ` drives the ddg tree algorithm .
@xmath403 @xmath404 generate @xmath405 by a ddg tree algorithm . @xmath406 feed @xmath405 to the retrieval algorithm ( randomness extraction procedure ) , and recover @xmath407 bits which are added to @xmath402 .
[ thm_batch_algo ] the batch generation algorithm uses @xmath408 random bits , where @xmath409 whenever @xmath410 . by the knuth - yao lower bound , @xmath411 and therefore , the procedure is asymptotically optimal to within @xmath412 bits .
the symbol @xmath412 means it is @xmath413 in probability as @xmath414 .
we choose a large integer constant @xmath50 and look at @xmath415 .
let @xmath416 be the size of the queue at time @xmath268 , and set @xmath417 .
for @xmath418 , let @xmath419 be the number of bits needed for generating @xmath420 without extraction .
the @xmath421 are i.i.d
. random variables .
then we have the following simple identity : @xmath422 by the law of large numbers , @xmath423 we note that @xmath424 because random variables @xmath421 are i.i.d . by theorem [ quibivilortrinuloxifene ]
, we also have that @xmath425 as @xmath426 .
therefore , @xmath427 the result follows if @xmath428 as @xmath414 . for this , we need only consider an upper bound , since @xmath429 , and then @xmath430 since @xmath431 , we have @xmath432 and @xmath433 fix @xmath10 , and let @xmath173 be the event that both lefthand sides in ( [ grumblegrumble1 ] ) and ( [ grumblegrumble2 ] ) are less than @xmath30 , so that @xmath434 .
the critical observation is that on @xmath173 , @xmath435 and therefore , @xmath436 and @xmath437 if we choose @xmath50 large enough such that , then @xmath438 if batch generation is applied to the partition method for continuous distributions , and the knuth - yao or han - hoshi method are used for the discrete part of that method , then the expected number of random bits needed per random variable , under rnyi s condition and , is bounded from above by @xmath439 thus matching the lower bound for the case @xmath24 , and all dimensions .
using the notion of maximal coupling between a generated random variable and a theoretical target random variable , we were able to lay the groundwork for a theory of generating with universal lower bound in terms of the number of random bernoulli@xmath1 bits needed . in order to grace the world
s software libraries with variable accuracy generators , much more work is needed , both algorithmic and theoretic .
we have shown that the well - known inversion method is nearly optimal in an information - theoretic sense , and will submit further work and other paradigms in the near future .
the method given in this appendix is based on taking the sum of independent random variables .
using convolution does not require an oracle for @xmath36 or @xmath239 which were required by the algorithms given in this paper .
its range of applicability seems however restricted to the uniform and the exponential as suggested by kakutani s result @xcite that can be stated as follows : for all @xmath342 , let @xmath440 $ ] and let @xmath441 be independent bernoulli random variables such that @xmath442 . if @xmath443 , then @xmath444 first of all , as shown in @xcite , if @xmath2 is an exponential mean @xmath69 random variable , then @xmath445 is distributed as a geometric random variable with parameter @xmath446 , and @xmath447 , the fractional part of @xmath2 , is distributed as a truncated exponential random variable on the interval @xmath137 , and @xmath445 and @xmath447 are independent .
we concentrate on the fractional part therefore .
the following theorem tells us that the fractional part is the convolution of independent bernoulli random variables .
[ gravelconvolbernoexpo ] let @xmath448 be a sequence of independent bernoulli distributed random variables with @xmath449\text { , and } \\
\mathbf{p}\{x_j=0\}&=1-p_j\text { for all $ j\in\mathbb{n}$.}\end{aligned}\ ] ] let @xmath450 . if @xmath451 then @xmath2 is a truncated exponential random variable , i.e. , the p.d.f
. of @xmath2 is @xmath452$.}\ ] ] since @xmath453 the fourier transform of @xmath454 is @xmath455 since @xmath2 is the sum of the independent @xmath454 s , we have @xmath456 which is the fourier transform of @xmath457
. we can thus generate @xmath447 with precision @xmath30 if we set , and let @xmath458 since a raw bernoulli random variable requires @xmath33 bits on average , this simple method , which has accuracy guarantee , uses an average not more than bits . on the other hand ,
the lower bound for @xmath447 is @xmath459 where @xmath460 the factor `` 2 '' in @xmath461 can be avoided when batch generation is used .
it can also be eliminated at a tremendous storage cost if the vector @xmath462 is generated using the algorithm of knuth - yao since we know the individual probabilities , i.e. , @xmath463 for all @xmath464 .
one can show that @xmath465 thus , for the knuth - yao method for this vector , @xmath466 again using knuth - yao , a geometric@xmath467 random variable can be generated exactly using not more than @xmath468 random bits .
an exponential random variable generated by this method has an overall expected complexity @xmath469 as @xmath256 .
both authors thank tamas linder for his help . claude gravel wants to thank gilles brassard from universit de montral for his financial support .
20 d. e. knuth and a. c .- c .
yao , `` the complexity of nonuniform random number generation , '' in _ algorithms and complexity : new directions and recent results .
_ , j. f. traub , ed.1em plus 0.5em minus 0.4emnew york : carnegie - mellon university , computer science department , academic press , 1976 , pp .
357428 , reprinted in knuth s _ selected papers on analysis of algorithms _
( csli , 2000 ) . | we study the problem of the generation of a continuous random variable when a source of independent fair coins is available .
we first motivate the choice of a natural criterion for measuring accuracy , the wasserstein @xmath0 metric , and then show a universal lower bound for the expected number of required fair coins as a function of the accuracy . in the case of an absolutely continuous random variable with finite differential entropy , several algorithms are presented that match the lower bound up to a constant , which can be eliminated by generating random variables in batches . +
* keywords : * random number generation , random bit model , differential entropy , partition entropy , inversion , probability integral transform , tree - based algorithms , random sampling |
rapid growth in the use of wireless services coupled with inefficient utilization of scarce spectrum resources has led to much interest in the analysis and development of cognitive radio systems .
hence , performance analysis of cognitive radio systems is conducted in numerous studies to gain more insights into their potential uses . in most of the previous work ,
transmission rate is considered as the main performance metric .
for instance , the secondary user mean capacity was studied in @xcite by imposing a constraint on the signal - to - interference - noise ratio ( sinr ) of the primary receiver and considering different channel side information ( csi ) levels .
the authors in @xcite determined the optimal power allocation strategies that achieve the ergodic capacity and the outage capacity of the cognitive radio channel under various power and interference constraints . in @xcite ,
the authors studied the optimal sensing time and power allocation strategy to maximize the average throughput in a multiband cognitive radio network .
recently , the work in @xcite proposed generic expressions for the optimal power allocation scheme and the ergodic capacity of a spectrum sharing cognitive radio under different levels of knowledge on the channel between the secondary transmitter and the secondary receiver and the channel between the secondary transmitter and the primary receiver subject to average / peak transmit power constraints and the interference outage constraint .
although transmission rate is a common performance metric considered for secondary users , error rate is another key performance measure to quantify the reliability of cognitive radio transmissions . in this regard ,
several recent studies incorporate error rates in cognitive radio analysis @xcite@xcite .
for instance , the authors in @xcite characterized the optimal constellation size of @xmath0-qam and the optimal power allocation scheme that maximize the channel capacity of secondary users for a given target bit error rate ( ber ) , interference and peak power constraints .
the work in @xcite mainly focused on the power allocation scheme minimizing the upper bound on the symbol error probability of phase shift keying ( psk ) in multiple antenna transmissions of secondary users .
the authors in @xcite proposed a channel switching algorithm for secondary users by exploiting the multichannel diversity to maximize the received snr at the secondary receiver and evaluated the transmission performance in terms of average symbol error probability .
the optimal antenna selection that minimizes the symbol error probability in underlay cognitive radio systems was investigated in @xcite .
moreover , the recent work in @xcite analyzed the minimum ber of a cognitive transmission subject to both average transmit power and interference power constraints . in their model , the secondary transmitter is equipped with multiple antennas among which only one antenna that maximizes the weighted difference between the channel gains of transmission link from the secondary transmitter to the secondary receiver and interference link from the secondary transmitter to the primary receiver is selected for transmission .
the authors in @xcite obtained a closed - form ber expression under the assumption that the interference limit of the primary receiver is very high . also , the work in @xcite focused on the optimal power allocation that minimizes the average ber subject to peak / average transmit power and peak / average interference power constraints while the interference on the secondary users caused by primary users is omitted .
moreover , in @xcite , the opportunistic scheduling in multiuser underlay cognitive radio systems was studied in terms of link reliability . in the error rate analysis of the above - mentioned studies , channel sensing errors
are not taken into consideration .
practical cognitive radio systems , which employ spectrum sensing mechanisms to learn the channel occupancy by primary users , generally operate under sensing uncertainty arising due to false alarms and miss - detections . for instance , different spectrum sensing methods for gaussian @xcite , @xcite and non - gaussian environments @xcite , @xcite and dynamic spectrum access strategies @xcite have extensively been studied recently in the literature , and as common to all schemes , channel sensing is generally performed with errors and such errors can lead to degradation in the performance . with this motivation , we in this paper study the symbol error rate performance of cognitive radio transmissions in the presence of imperfect channel sensing decisions .
we assume that secondary users first sense the channel in order to detect the primary user activity before initiating their own transmissions . following channel sensing , secondary users employ two different transmission schemes depending on how they access the licensed channel : sensing - based spectrum sharing ( sss ) and opportunistic spectrum access ( osa ) . in the sss scheme @xcite
, cognitive users are allowed to coexist with primary users in the channel as long as they control the interference by adapting the transmission power according to the channel sensing results .
more specifically , secondary users transmit at two different power levels depending on whether the channel is detected as busy or idle . in the osa scheme @xcite ,
cognitive users are allowed to transmit data only when the channel is detected as idle , and hence secondary users exploit only the silent periods in the transmissions of primary users , called as spectrum opportunities . due to the assumption of imperfect channel sensing , two types of sensing errors , namely false alarms and
miss detections , are experienced .
false alarms result in inefficient utilization of the idle channel while miss - detections lead to cognitive users transmission interfering with primary user s signal .
such interference can be limited by imposing interference power constraints . in our error rate analysis
, we initially formulate the optimal decision rule and error rates for an arbitrary digital modulation scheme .
subsequently , motivated by the requirements to efficiently use the limited spectrum in cognitive radio settings , we concentrate on quadrature amplitude modulation ( qam ) as it is a bandwidth - efficient modulation format . more specifically , in our analysis , we assume that the cognitive users employ rectangular qam for data transmission , analysis of which , as another benefit , can easily be specialized to obtain results for square qam , pulse amplitude modulation ( pam ) , quadrature phase - shift keying ( qpsk ) , and binary phase - shift keying ( bpsk ) signaling .
in addition to the consideration of sensing errors and relatively general modulation formats , another contribution of this work is the adoption of a gaussian mixture model for the primary user s received faded signals in the error - rate analysis .
the closed - form error rate expressions in aforementioned works @xcite@xcite are obtained when primary user s faded signal at the secondary receiver is assumed to be gaussian distributed . however , in practice , cognitive radio transmissions can be impaired by different types of non - gaussian noise and interference , e.g. , man - made impulsive noise @xcite , narrowband interference caused by the primary user @xcite , primary user s modulated signal @xcite , and co - channel interference from other cognitive radios @xcite .
therefore , it is of interest to investigate the error rate performance of cognitive radio transmissions in the presence of primary user s interference which is modeled to have a gaussian mixture probability density function ( pdf ) ( which includes pure gaussian distribution as a special case ) @xcite .
main contributions of this paper can be summarized as follows . under the above - mentioned assumptions ,
we first derive , for both sss and osa schemes , the optimal detector structure , and then we present a closed - form expression of the average symbol error probability under constraints on the transmit power and interference . through this analysis , we investigate the impact of imperfect channel sensing ( i.e. , the probabilities of detection and false alarm ) , interference from the primary user , and both transmit power and interference constraints on the error rate performance of cognitive transmissions . also , the performances of sss and osa transmission schemes are compared when primary user s faded signal is modeled to have either a gaussian mixture or a purely gaussian density .
the remainder of this paper is organized as follows : section [ sec : system_model ] introduces the system model . in section [ sec : general_formulation ] , general formulations for the optimal detection rule and average symbol error probability in the presence of channel sensing errors are provided for sss and osa schemes . in section [ sec : error_rate_qam_pam ] , closed - form average symbol error probability expressions for specific modulation types , i.e. , arbitrary rectangular qam and pam are derived subject to both transmit power and interference constraints under the assumptions of gaussian - mixture - distributed primary user faded signal and imperfect channel sensing decisions .
numerical results are provided and discussed in section [ sec : num_results ] .
finally , section [ sec : conclusion ] concludes the paper .
we consider a cognitive radio system consisting of a pair of secondary transmitter - receiver and a pair of primary transmitter - receiver .
the secondary user initially performs channel sensing , which can be modeled as a hypothesis testing problem .
assume that @xmath1 denotes the hypothesis that the primary users are inactive in the channel , and @xmath2 denotes the hypothesis that the primary users are active .
various channel sensing methods , including energy detection , cyclostationary detection , and matched filtering , have been proposed and analyzed in the literature .
regardless of which method is used , one common feature is that errors in the form of miss - detections and false - alarms occur in channel sensing .
the ensuing analysis takes such errors into account and depends on the sensing scheme only through the detection and false - alarm probabilities .
assume that @xmath3 and @xmath4 denote the sensing decisions that the primary users are inactive and active , respectively .
then , the detection and false - alarm probabilities can be expressed respectively as the following conditional probabilities : @xmath5 following channel sensing , the secondary transmitter performs data transmission over a flat - fading channel . in the sss scheme ,
the average transmission power is selected depending on the channel sensing decision .
more specifically , the average transmission power is @xmath6 if primary user activity is detected in the channel ( denoted by the event @xmath4 ) whereas the average power is @xmath7 if no primary user transmissions are sensed ( denoted by the event @xmath3 ) .
we assume that there is a peak constraint on these average power levels , i.e. , we have @xmath8 where @xmath9 denotes the peak transmit power limit .
we further impose an average interference constraint in the following form : @xmath10 where @xmath11 is the detection probability and @xmath12 is the channel fading coefficient between the secondary transmitter and the primary receiver .
note that with probability @xmath11 , primary user activity is correctly detected and primary receiver experiences average interference proportional to @xmath13 . on the other hand , with probability @xmath14 ,
miss - detections occur , secondary user transmits with power @xmath7 , and primary receiver experiences average interference proportional to @xmath15 .
therefore , @xmath16 can be regarded as a constraint on the average interference inflicted on the primary user primary receivers by replacing the constraint in ( [ eq : interference_power_constraint_sss ] ) with @xmath17 , where @xmath18 is the channel fading coefficient between the secondary transmitter and the @xmath19 primary receiver . in this
setting , @xmath16 effectively becomes a constraint on the worst - case average interference . ] . in the osa scheme ,
no transmission is allowed when the channel is detected as busy and hence , we set @xmath6 = 0 . now with the peak power and average interference constraints , we have @xmath20 which can be combined to write @xmath21 above , we have introduced the average interference constraint . however ,
if the instantaneous value of the fading coefficient @xmath12 is known at the secondary transmitter , then a peak interference constraint in the form @xmath22 for @xmath23 can be imposed .
note that transmission power @xmath7 in an idle - sensed channel is also required to satisfy the interference constraint due to sensing uncertainty ( i.e. , due to the consideration of miss - detection events ) .
hence , a rather strict form of interference control is being addressed under these limitations .
now , including the peak power constraint , we have @xmath24 for @xmath25 ( while keeping @xmath26 in the osa scheme ) . above
, @xmath27 denotes the peak received power limit at the primary receiver . as a result of channel sensing decisions and the true nature of primary user activity
, we have four possible cases which are described below together with corresponding input - output relationships : * _ case ( slowromancap1@ ) _ : a busy channel is sensed as busy , denoted by the joint event @xmath28 .
@xmath29 * _ case ( slowromancap2@ ) _ : a busy channel is sensed as idle , denoted by the joint event @xmath30 . @xmath31 * _ case ( slowromancap3@ ) _ : an idle channel is sensed as busy , denoted by the joint event @xmath32 .
@xmath33 * _ case ( slowromancap4@ ) _ : an idle channel is sensed as idle , denoted by the joint event @xmath34 .
@xmath35 in the above expressions , @xmath36 is the transmitted signal , @xmath37 is the received signal , and @xmath38 denotes zero - mean , circularly - symmetric complex fading coefficient between the secondary transmitter and the secondary receiver with variance @xmath39 .
@xmath40 is the circularly - symmetric complex gaussian noise with mean zero and variance @xmath41 , and hence has the pdf @xmath42 the active primary user s received faded signal at the secondary receiver is denoted by @xmath43 .
notice that if the primary users are active and hence the hypothesis @xmath2 is true as in cases ( slowromancap1@ ) and ( slowromancap2@ ) , the secondary receiver experiences interference from the primary user s transmission in the form of @xmath43 .
we assume that @xmath43 has a gaussian mixture distribution , i.e. , its pdf is a weighted sum of @xmath44 complex gaussian distributions with zero mean and variance @xmath45 for @xmath46 : @xmath47 where the weights @xmath48 satisfy @xmath49 with @xmath50 for all @xmath51 .
primary user s received faded signal has a gaussian mixture distribution , if we , for instance , have @xmath52 where @xmath53 , which is the channel fading coefficient between the primary transmitter and the secondary receiver , is a circularly symmetric , complex , zero - mean , gaussian random variable , and @xmath54 is the primary user s modulated digital signal .
note that @xmath43 is conditionally gaussian given @xmath54 .
now , assuming that the modulated signal @xmath54 can take @xmath44 different values with prior probabilities given by @xmath48 for @xmath46 , @xmath43 has a gaussian mixture distribution as in ( [ eq : gaussian_mix ] ) . in the case of multiple primary transmitters for which we have
@xmath55 the above argument can easily be extended if all channel fading coefficients @xmath56 are zero - mean gaussian distributed .
gaussian mixture model is generally rich enough to accurately approximate a wide variety of density functions ( * ? ? ?
* section 3.2 ) .
this fact indicates that the applicability of our results can be extended to various other settings in which @xmath43 has a distribution included in this class of densities .
additionally , in the special case of @xmath57 , the gaussian mixture distribution becomes the pure complex gaussian distribution .
hence , the results obtained for the gaussian mixture distribution can readily be specialized to derive those for the gaussian distributed @xmath43 as well . as observed from the input - output relationships in ( [ eq : received_signal1])([eq : received_signal4 ] ) , when the true state of the primary users is idle , corresponding to the cases ( slowromancap3@ ) and ( slowromancap4@ ) , the additive disturbance is simply the background noise @xmath40 . on the other hand , in cases ( slowromancap1@ ) and ( slowromancap2@ )
in which the channel is actually busy , the additive disturbance becomes @xmath58 whose distribution can be obtained through the convolution of density functions of the background gaussian noise @xmath40 and the primary user s received faded signal @xmath43 . using the result of gaussian convolution of gaussian mixture given by @xcite , the distribution of @xmath59 can be obtained as @xmath60 note that @xmath59 also has a gaussian mixture distribution .
note further that the pdf of @xmath59 can be expressed in terms of its real and imaginary components as @xmath61 moreover , the marginal distributions of each component are given by @xmath62 it is easily seen that the pdf of @xmath59 in ( [ eq : gaussian_mix_conv ] ) can not be factorized into the product of the marginal pdf s of its real and imaginary parts @xmath63 , given in ( [ eq : marginal_inphase ] ) and ( [ eq : marginal_quad ] ) , respectively .
therefore , the real and imaginary parts of the additive disturbance @xmath59 are dependent .
when @xmath57 , i.e. , in the case of a pure gaussian distribution , the joint distribution can written as a product of its real and imaginary components since they are independent .
in this section , we present the optimal decision rule and the average symbol error probability for the cognitive radio system in the presence of channel sensing errors .
we provide general formulations applicable to any modulation type under sss and osa schemes .
more specific analysis for arbitrary rectangular qam and pam is conducted in section [ sec : error_rate_qam_pam ] . in the cognitive radio setting considered in this paper , the optimal maximum _ a posteriori _ probability ( map ) decision rule given the sensing decision @xmath64 and the channel fading coefficient @xmath38 can be formulated for any arbitrary @xmath0-ary digital modulation as follows : @xmath65 where @xmath66 is the prior probability of signal constellation point @xmath67 and @xmath68 @xmath69 .
above , ( [ eq : map_decision_rule-2 ] ) follows from bayes rule and ( [ eq : cond - prob - yr ] ) is obtained by conditioning the density function on the hypotheses @xmath1 and @xmath2 .
note that @xmath70 in ( [ eq : cond - prob - yr ] ) is the conditional distribution of the received real signal @xmath37 given the transmitted signal @xmath67 , channel fading coefficient @xmath38 , channel sensing decision @xmath64 , and true state of the channel @xmath71 , and can be expressed as @xmath72 for @xmath73 .
note that the sensing decision @xmath64 affects the density function through the power of the transmitted signal @xmath67 .
moreover , the conditional probabilities in ( [ eq : cond - prob - yr ] ) can be expressed as @xmath74 where @xmath75 and @xmath76 are the prior probabilities of the channel being idle and busy , respectively , and the conditional probabilities in the form @xmath77 depend on the channel sensing performance . as discussed in section [ subsec : sensing ] , @xmath78 is the detection probability and @xmath79 is the false alarm probability . from these formulations , we see that the optimal decision rule in general depends on the sensing reliability .
the average symbol error probability ( sep ) for the map decision rule in ( [ eq : map_decision_rule ] ) in the sss scheme can be computed as @xmath80 the above average symbol error probability can further be expressed as in ( [ eq : pe_intdecisionregions ] ) @xmath81 \end{split}\end{aligned}\ ] ] ' '' '' where @xmath82 and @xmath83 are the decision regions of each signal constellation point @xmath67 for @xmath84 when the channel is sensed to be idle and busy , respectively .
if cognitive user transmission is not allowed in the case of the channel being sensed as occupied by the primary users , we have the osa scheme for which the average probability of error can be expressed as @xmath85
in this section , we conduct a more detailed analysis by considering rectangular qam transmissions to demonstrate the key tradeoffs in a lucid setting . correspondingly , we determine the optimal decision regions by taking channel sensing errors into consideration and identify the error rates for sss and osa schemes .
we derive closed - form minimum average symbol error probability expressions under the transmit power and interference constraints .
note that the results for qam can readily be specialized for pam , qpsk , and bpsk transmissions .
the signal constellation point @xmath86 in @xmath87 rectangular qam signaling can be expressed in terms of its real and imaginary parts , respectively , as @xmath88 where the amplitude level of each component is given by @xmath89 above , @xmath90 and @xmath91 are the modulation size on the in - phase and quadrature components , respectively , and @xmath92 denotes the minimum distance between the signal constellation points and is given by @xmath93 where @xmath94 is the transmission power under sensing decision @xmath95 .
it is assumed that the fading realizations are perfectly known at the receiver . in this case ,
phase rotations caused by the fading can be offset by multiplying the channel output @xmath37 with @xmath96 where @xmath97 is the phase of the fading coefficient @xmath38 .
hence , the modified received signal can be written in terms of its real and imaginary parts as follows : @xmath98 where the subscripts @xmath99 and @xmath100 are used to denote the real and imaginary components of the signal , respectively . note that @xmath101 and @xmath102 have the same statistics as @xmath40 and @xmath43 , respectively , due to their property of being circularly symmetric .
hence , given the transmitted signal constellation point @xmath103 , the distribution of the modified received signal @xmath104 is given by @xmath105 moreover , the real and imaginary parts of noise , i.e. , @xmath106 and @xmath107 are independent zero - mean gaussian random variables , and the real and imaginary parts of primary users faded signal , i.e. , @xmath108 and @xmath109 , are gaussian mixture distributed random variables . in the following ,
we characterize the decision regions of the optimal detection rule for equiprobable qam signaling in the presence of sensing uncertainty . [
prop : decisionrule ] for cognitive radio transmissions with _
equiprobable _ rectangular @xmath0-qam modulation ( with constellation points expressed as in ( [ eq : complexqampoint])([eq : imaginaryqampoint ] ) ) under channel sensing uncertainty in both sss and osa schemes , the optimal detection thresholds under any channel sensing decision are located midway between the received signal points .
hence , the optimal detector structure does not depend on the sensing decision .
_ proof _ : see appendix [ app : proof - prop ] . in this subsection , we present closed - form average symbol error probability expressions under both transmit power and interference constraints for sss and osa schemes .
initially , we express the error probabilities for a given value of the fading coefficient @xmath38 .
subsequently , we address averaging over fading and also incorporate power and interference constraints .
we note that in the presence of peak interference constraints , the transmitted power level depends on the fading coefficient @xmath12 experienced in the channel between the secondary and primary users as seen in ( [ eq : instantaneous - power - g ] ) .
therefore , we in this case consider an additional averaging of the error rates with respect to @xmath12 . under the optimal decision rule given in the previous subsection , the average symbol error probability of _ equiprobable _ signals for a given fading coefficient @xmath38 can be expressed as @xmath110 where @xmath111 denotes the conditional error probability given the transmitted signal @xmath103 , channel fading @xmath38 , sensing decision @xmath64 , and true state of the channel @xmath71 .
we can group the error patterns of rectangular @xmath0-qam modulation into three categories . specifically , the probability of error for the signal constellation points on the corners is equal due to the symmetry in signaling and detection .
the same is also true for the points on the sides and the inner points .
the symbol error probability for the four corner points is given by @xmath112 where @xmath113 and @xmath114 .
the distributions of the gaussian noise @xmath115 and the primary user s interference signal plus noise @xmath116 are given in ( [ eq : gaussian_noise ] ) and ( [ eq : gaussian_mix_conv ] ) , respectively .
after evaluating the integrals , the above expression becomes @xmath117 where @xmath118 is the gaussian @xmath119-function . for the @xmath120 points on the sides , except the corner points ,
the symbol error probability is @xmath121 where @xmath122 .
after performing the integrations , we can express @xmath123 as @xmath124 finally , the symbol error probability for @xmath125 inner points is obtained from @xmath126 which can be evaluated to obtain @xmath127 overall , we can express @xmath128 in ( [ eq : pe_mpam_h ] ) by combining @xmath129 , @xmath123 and @xmath130 as follows @xmath131 after rearranging the terms , the final expression for the average symbol error probability @xmath128 is given by ( [ eq : pe_h_qam ] ) shown at the top of next page .
@xmath132\\ + \pr\{{\mathcal{h}}_1|{\hat{\mathcal{h}}}_k\}\bigg[2\bigg(2-\frac{1}{m_i}-\frac{1}{m_q}\bigg)\sum_{l=1}^p\lambda_l q\bigg(\sqrt{\frac{d_{min , k}^2|h|^2}{4(\sigma_{l}^2+\sigma_{n}^2 ) } } \bigg)-4\bigg(1-\frac{1}{m_i}\bigg)\bigg(1-\frac{1}{m_q}\bigg)\sum_{l=1}^p\lambda_l q^2\bigg(\sqrt{\frac{d_{min , k}^2|h|^2}{4(\sigma_{l}^2+\sigma_{n}^2 ) } } \bigg)\bigg]\bigg\}. \end{split}\ ] ] ' '' '' this expression can be specialized to square qam signaling by setting @xmath133 .
we observe above that while the optimal decision rule does not depend on the sensing decisions , the error rates are functions of detection and false alarm probabilities .
note also that the expressions above are for a given value of fading .
the unconditional symbol error probability averaged over fading can be evaluated from @xmath134 in the special case of a rayleigh fading model for which the fading power has an exponential distribution with unit mean , i.e. , @xmath135 , the above integral can be evaluated by adopting the same approach as in @xcite and using the indefinite integral form of the gaussian @xmath119-function @xcite and square of the gaussian @xmath119-function @xcite , given , respectively , by @xmath136 the resulting unconditional average symbol error probability over rayleigh fading is given by ( [ eq : pe_sss_without_power_constraints ] ) at the top of next page @xmath137\\ & + \pr\{{\mathcal{h}}_1|{\hat{\mathcal{h}}}_k\}\bigg[\bigg(2-\frac{1}{m_i}-\frac{1}{m_q}\bigg)\sum_{l=1}^p\lambda_l\bigg(1-\frac{1}{\beta_{1,k}}\bigg)-2\bigg(1-\frac{1}{m_i}\bigg)\bigg(1-\frac{1}{m_q}\bigg)\sum_{l=1}^p\lambda_l\bigg(\frac{2}{\pi}\frac{1}{\beta_{1,k}}\tan^{-1}\left(\frac{1}{\beta_{1,k}}\right)-\frac{1}{\beta_{1,k}}+\frac{1}{2}\bigg)\bigg]\bigg\}. \end{split}\ ] ] ' '' '' where @xmath138 and @xmath139 for @xmath46 .
the average symbol error probability for rectangular qam signaling in the presence of gaussian - distributed @xmath43 can readily be obtained by letting @xmath140 in ( [ eq : pe_sss_without_power_constraints ] ) .
although the sep expression in ( [ eq : pe_sss_without_power_constraints ] ) seems complicated , it is in fact very simple to evaluate .
furthermore , this @xmath141 can be upper bounded as @xmath142 this upper bound follows by removing the negative terms that include @xmath143 on the right - hand side of ( [ eq : pe_h_qam ] ) and then integrating with respect to fading distribution .
note also that the upper bound in ( [ eq : pe_sss_tight ] ) with @xmath144 is the exact symbol error probability for pam modulation .
note further that the sep expression in ( [ eq : pe_sss_without_power_constraints ] ) is a function of the transmission powers @xmath7 and @xmath6 . the optimal choice of the power levels under peak power and average interference constraints given in ( [ eq : peak_power_constraint ] ) and ( [ eq : interference_power_constraint_sss ] ) and the resulting error rates can be determined by solving @xmath145 as discussed in section [ subsec : power - interference ] , if the fading coefficient @xmath12 between the secondary transmitter and the primary receiver is known and peak interference constraints are imposed , then the maximum transmission power is given by @xmath146 after inserting these @xmath147 and @xmath148 into the @xmath149 upper bound in ( [ eq : pe_sss_tight ] ) and evaluating the expectation over the fading coefficient @xmath12 , we obtain @xmath150 where @xmath151 and @xmath152 denotes the upper bound in ( [ eq : pe_sss_tight ] ) . if @xmath153 is exponentially distributed with unit mean , then by using the identity in ( * ? ?
3.362.2 ) , we can evaluate the second integral on the right - hand side of ( [ eq : pe_sss_tight - avg ] ) and express the upper bound as in ( [ eq : pe_sss_min ] ) given on the next page @xmath154\\&+\pr\{{\mathcal{h}}_1|{\hat{\mathcal{h}}}_k\}\sum_{l=1}^p\lambda_l\bigg[{{\mathrm{e}}}^{b_1}-2\sqrt{\gamma_{1}\pi}{{\mathrm{e}}}^{\gamma_{1}}q\big(\sqrt{2}(b_1+\gamma_{1})\big)\bigg]\bigg\}\\ \end{split}\ ] ] ' '' '' where @xmath155 , @xmath156 .
it should be noted that we can easily obtain the _ exact _ symbol error probability expressions for pam modulation by replacing @xmath157 and @xmath144 in ( [ eq : pe_h_qam ] ) , ( [ eq : pe_sss_without_power_constraints ] ) , ( [ eq : pe_sss_min ] ) . in the osa scheme ,
secondary users are not allowed to transmit when the primary user activity is sensed in the channel .
therefore , we only consider error patterns under @xmath3 given in ( [ eq : sep1 ] ) , ( [ eq : sep2 ] ) , ( [ eq : sep3 ] ) . hence , following the same approach adopted in section [ subsubsec : error_rate_sss ] , the average symbol error probability under the osa scheme is obtained as in ( [ eq : pe_osa_without_power_constraints ] ) given on the next page .
@xmath158\\ & + \pr\{{\mathcal{h}}_1|{\hat{\mathcal{h}}}_0\}\bigg[\bigg(2-\frac{1}{m_i}-\frac{1}{m_q}\bigg)\sum_{l=1}^p\lambda_l\bigg(1-\frac{1}{\beta_{1,0}}\bigg)-2\bigg(1-\frac{1}{m_i}\bigg)\bigg(1-\frac{1}{m_q}\bigg)\sum_{l=1}^p\lambda_l\bigg(\frac{2}{\pi}\frac{1}{\beta_{1,0}}\tan^{-1}\left(\frac{1}{\beta_{1,0}}\right)-\frac{1}{\beta_{1,0}}+\frac{1}{2}\bigg)\bigg ] .
\end{split}\ ] ] ' '' '' similarly , the sep upper bound becomes @xmath159 note that under average interference constraints , the maximum allowed transmission power in an idle - sensed channel is given by @xmath160 on the other hand , if the peak interference power constraint is imposed , the maximum allowed transmission power is @xmath161 after inserting this @xmath147 into ( [ eq : pe_osa_tight ] ) , assuming again that @xmath153 is exponentially distributed with unit mean , and evaluating the integration in a similar fashion as in section [ subsubsec : error_rate_sss ] , an upper bound on the average symbol error probability can be obtained as in ( [ eq : pe_osa_min ] ) on the next page where @xmath162 and @xmath163 .
@xmath164\\&+\pr\{{\mathcal{h}}_1|{\hat{\mathcal{h}}}_0\}\sum_{l=1}^p\lambda_l\bigg[{{\mathrm{e}}}^{\frac{q_{{\rm{pk}}}}{p_{{\rm{pk}}}}}-2\sqrt{\psi_{1}\pi}{{\mathrm{e}}}^{\psi_{1}}q\bigg(\sqrt{2}\bigg(\frac{q_{{\rm{pk}}}}{p_{{\rm{pk}}}}+\psi_{1}\bigg)\bigg)\bigg]\bigg\ } \end{split}\ ] ] ' '' ''
in this section , we present numerical results to demonstrate the error performance of a cognitive radio system in the presence of channel sensing uncertainty for both sss and osa schemes .
more specifically , we numerically investigate the impact of sensing performance ( e.g. , detection and false - alarm probabilities ) , different levels of peak transmission power and average and peak interference constraints on cognitive transmissions in terms of symbol error probability .
theoretical results are validated through monte carlo simulations . unless mentioned explicitly ,
the following parameters are employed in the numerical computations .
it is assumed that the variance of the background noise is @xmath165 . when the primary user signal is assumed to be gaussian , its variance ,
@xmath166 is set to 0.5 . on the other hand ,
in the case of primary user s received signal @xmath43 distributed according to the gaussian mixture model , we assume that @xmath167 , i.e. , there are four components in the mixture , @xmath168 for all @xmath169 , and the variance is still @xmath170 . also , the primary user is active over the channel with a probability of @xmath171 , hence @xmath172 and @xmath173 . finally , we consider a rayleigh fading channel between the secondary users with fading power pdf given by @xmath135 for @xmath174 , and also assume that the fading power @xmath153 in the channel between the secondary transmitter and primary receiver is exponentially distributed with @xmath175
. 0.45 0.45 0.45 we initially consider peak transmit and average interference constraints as given in ( [ eq : peak_power_constraint ] ) and ( [ eq : interference_power_constraint_sss ] ) , respectively . in the following numerical results , for the sss scheme , we plot the error probabilities and optimal transmission power levels obtained by solving ( [ eq : sss_expectation_pe ] ) . in the case of osa , we plot the average error probability expressed in ( [ eq : pe_osa_without_power_constraints ] ) with maximum allowed power @xmath147 given in ( [ eq : maxpower - averageinterference - osa ] ) . in fig .
[ fig : sep_p0_p1_qavg_sss ] , we display the average symbol error probability ( @xmath149 ) and optimal transmission powers @xmath7 and @xmath6 as a function of the average interference constraint , @xmath16 , in the sss scheme . @xmath11 and @xmath176
are set to @xmath177 and @xmath178 , respectively .
the peak transmission power is @xmath179 db . we assume that the secondary users employ 2-pam , 4-qam , 8-pam and @xmath180-qam modulation schemes for data transmission .
we have considered both gaussian and gaussian - mixture distributed @xmath43 .
in addition to the analytical results obtained by using ( [ eq : pe_sss_without_power_constraints ] ) and solving ( [ eq : sss_expectation_pe ] ) , we performed monte carlo simulations to determine the sep .
we notice in the figure that analytical and simulation results agree perfectly .
additionally , it is seen that for all modulation types , error rate performance of secondary users improves as average interference constraint becomes looser ( i.e. , as @xmath16 increases ) , allowing transmission power levels @xmath7 and @xmath6 to become higher as illustrated in the lower subfigures .
saturation seen in the plot of @xmath7 is due to the peak constraint @xmath9 .
other observations are as follows . as the modulation size increases , @xmath149 increases as expected .
it is also interesting to note that lower @xmath149 is attained in the presence of gaussian - mixture distributed @xmath43 when compared with the performance when @xmath43 has a pure gaussian density with the same variance . 0.45 0.45 in fig .
[ fig : sep_p0_qavg_osa ] , average @xmath149 and transmission power @xmath7 are plotted as a function of the average interference constraint , @xmath16 , for the osa scheme .
we again set @xmath179 db , @xmath181 and @xmath182 , and consider 2-pam , 4-qam , 8-pam and @xmath180-qam schemes .
it is observed from the figure that as @xmath16 increases , error probabilities initially decrease and then remain constant due to the fact that the secondary users can initially afford to transmit with higher transmission power @xmath7 as the interference constraint becomes less strict , but then get limited by the peak transmission power constraint and send data at the fixed power level of @xmath9 .
again , we observe that lower error probabilities are attained when the primary user s received signal @xmath43 follows a gaussian mixture distribution . 0.45 0.45 in fig .
[ fig : sep_pd_avg ] , the average @xmath149s of 4-qam ( in the upper subfigure ) and 8-pam signaling ( in the lower subfigure ) are plotted as a function of the detection probability @xmath11 .
@xmath176 is set to @xmath178 .
we consider both sss and osa schemes .
it is assumed that @xmath183 db and @xmath184 db . we observe that @xmath149 for both modulation types in both sss and osa schemes decreases as @xmath11 increases .
hence , performance improves with more reliable sensing . in this case
, the primary reason is that more reliable detection enables the secondary users transmit with higher power in an idle - sensed channel .
for instance , if @xmath185 , then the transmission power @xmath7 is only limited by @xmath9 in both sss and osa . in the figure
, we also notice that lower @xmath149 is achieved in the osa scheme , when compared with the sss scheme , due to the fact that osa avoids transmission over a busy - channel in which interference from the primary user s transmission results in a more noisy channel and consequently higher error rates are experienced . at the same time
, it is important to note that not transmitting in a busy - sensed channel as in osa potentially reduces data rates . 0.45 0.4 in fig .
[ fig : sep_pf_avg ] , the average @xmath149s of 4-qam and 8-pam signaling are plotted as a function of the false alarm probability @xmath176 for both sss and osa . it is assumed that @xmath186 . it is further assumed that @xmath179 db and @xmath187 db , again corresponding to the case in which average interference power constraint is dominant compared to the peak transmit power constraint . in both schemes
, @xmath149 increases with increasing false alarm probability @xmath176 .
hence , degradation in sensing reliability leads to performance loss in terms of error probabilities . in osa ,
the transmission power @xmath188 does not depend on @xmath176 and hence is fixed in the figure .
the increase in the error rates can be attributed to the fact that secondary users more frequently experience interference from primary user s transmissions due to sensing uncertainty .
for instance , in the extreme case in which @xmath189 , the probability terms in ( [ eq : pe_osa_without_power_constraints ] ) become @xmath190 and @xmath191 , indicating that although the channel is sensed as idle , it is actually busy with probability one and the additive disturbance in osa transmissions always includes @xmath43 . in the sss scheme ,
higher @xmath176 leads to more frequent transmissions with power @xmath6 which is generally smaller than @xmath7 in order to limit the interference on the primary users .
transmission with smaller power expectedly increases the error probabilities . on the other hand
, we interestingly note that as @xmath176 approaches 1 , @xmath6 becomes higher than @xmath7 when ( [ eq : sss_expectation_pe ] ) is solved , resulting in a slight decrease in @xmath149 when @xmath176 exceeds around 0.9 .
we now address the peak interference constraints by assuming that the transmission powers are limited as in ( [ eq : instantaneous - power - g ] ) . in this section ,
analytical error probability curves are plotted using the upper bounds in ( [ eq : pe_sss_min ] ) in the case of sss and in ( [ eq : pe_osa_min ] ) in the case of osa since we only have closed - form expressions for the error probability upper bounds when we need to evaluate an additional expectation with respect to @xmath153 .
note that these upper bounds provide exact error probabilities when pam is considered . additionally , the discrepancy in qam is generally small as demonstrated through comparisons with monte carlo simulations which provide the exact error rates in the figures . 0.45 0.45 in fig .
[ fig : sep_ppk_sss ] , we plot the average @xmath149 as a function of the peak transmission power , @xmath9 , for the sss scheme in the presence of gaussian distributed and gaussian - mixture distributed primary user s received faded signal @xmath43 in the upper and lower subfigures , respectively .
the secondary users again employ 2-pam , 4-qam , 8-pam and @xmath180-qam schemes . the peak interference power constraint ,
@xmath27 is set to 4 db .
it is seen that monte carlo simulations match with the analytical results for pam and are slightly lower than the analytical upper bounds for qam .
as expected , the average @xmath149 initially decreases with increasing @xmath9 and a higher modulation size leads to higher error rates at the same transmission power level .
we again observe that lower error rates are experienced when @xmath43 has a gaussian mixture distribution rather than a gaussian distribution with the same variance .
it is also seen that as @xmath9 increases , the @xmath149 curves in all cases approach some error floor at which point interference constraints become the limiting factor . another interesting observation is the following . in fig .
[ fig : sep_ppk_sss ] , @xmath149s are plotted for two different pairs of detection and false alarm probabilities . in the first scenario ,
channel sensing is perfect , i.e. , @xmath185 and @xmath192 .
in the second scenario , we have @xmath186 and @xmath193 . in both scenarios ,
we observe the same error rate performance .
this is because the same transmission power is used regardless of whether the channel is detected as idle or busy , i.e. , @xmath194 for both @xmath23 .
the interference constraints are very strict as noted in section [ subsec : power - interference ] .
hence , averaging over channel sensing decisions becomes averaging over the prior probabilities of channel occupancy , which does not depend on the probabilities of detection and false alarm . indeed ,
spectrum sensing can be altogether omitted under these constraints . 0.45 0.45 in fig .
[ fig : sep_ppk_osa ] , we plot the average @xmath149 as a function of @xmath9 for the osa scheme . as before , 2-pam , 4-qam , 8-pam and @xmath180-qam are considered .
imperfect sensing with @xmath186 and @xmath193 is considered in the upper subfigure whereas perfect sensing ( i.e. , @xmath185 and @xmath192 ) is assumed in the lower subfigure . in both subfigures
, it is seen that increasing @xmath9 initially reduces @xmath149 which then hits an error floor as the interference constraints start to dominate .
it is also observed that perfect channel sensing improves the error rate performance of cognitive users .
note that if sensing is perfect , secondary users transmit only if the channel is actually idle and experience only the background noise @xmath40 . on the other hand , under imperfect
sensing , secondary users transmit in miss - detection scenarios as well , in which they are affected by both the background noise and primary user interference @xmath43 , leading to higher error rates .
cognitive radio transmission impaired by gaussian mixture distributed @xmath43 again results in lower @xmath149 compared to gaussian distributed @xmath43 .
but , of course , this distinction disappears with perfect sensing in the lower subfigure since the secondary users experience only the gaussian background noise @xmath40 as noted above . finally , the gap between the analytical and simulation results for qam is due to the use of upper bounds in the analytical error curves as discussed before . 0.45 0.45 in fig .
[ fig : sep_pd_qpk ] , we display the average @xmath149 of 4-qam and 8-pam signaling as a function of the detection probability @xmath11 .
@xmath176 is set to @xmath178 . both sss and osa schemes are considered . here , we also assume that @xmath195 db , @xmath196 db . it is seen that error rate performances for sss scheme for both modulation types do not depend on detection probability because of the same reasoning explained in the discussion of fig .
[ fig : sep_ppk_sss ] . on the other hand
, the error rate performance for the osa scheme improves with increasing detection probability since the secondary user experiences less interference from the primary user activity .
it is also seen that osa scheme outperforms sss scheme .
0.45 0.45 in fig .
[ fig : sep_pf_qpk ] , we analyze the average @xmath149 of 4-qam and 8-pam signaling as a function of the false alarm probability @xmath176 .
detection probability is @xmath186 .
similarly as before , @xmath179 db and @xmath197 db . again , error rate performance does not depend on @xmath176 in the sss scheme .
it is observed that @xmath149 in osa scheme increases with increasing false alarm probability .
hence , degradation in the sensing performance in terms of increased false alarm probabilities leads to degradation in the error performance . as discussed in section [ subsec : averageinterference ] regarding the error rates in fig .
[ fig : sep_pf_avg ] , deterioration in the performance is due to more frequent exposure to interference from primary user s transmissions in the form of @xmath43 .
one additional remark from the figure is that sss scheme gives better error rate performance compared to osa scheme at higher values of @xmath176 .
we have studied the error rate performance of cognitive radio transmissions in both sss and osa schemes in the presence of transmit and interference power constraints , sensing uncertainty , and gaussian mixture distributed interference from primary user transmissions . in this setting , we have proved that the midpoints between the signals are optimal thresholds for the detection of equiprobable rectangular qam signals .
we have first obtained exact @xmath149 expressions for given fading realizations and then derived closed - form average sep expressions for the rayleigh fading channel .
we have further provided upper bounds on the error probabilities averaged over the fading between the secondary transmitter and primary receiver under the peak interference constraint .
the analytical @xmath149 expressions have been validated through monte - carlo simulations . in the numerical results
, we have had several interesting observations .
we have seen that , when compared to sss , lower error rates are generally attained in the osa scheme .
also , better error performance is achieved in the presence of gaussian - mixture distributed @xmath43 in comparison to that achieved when @xmath43 is gaussian distributed with the same variance .
we have also addressed how the error rates and transmission powers vary as a function of the power and interference constraints .
finally , we have demonstrated that symbol error probabilities are in general dependent on sensing performance through the detection and false alarm probabilities .
for instance , we have observed that as the detection probability increases , the error rate performance under both schemes improves in interference - limited environments .
similarly , sep is shown to decrease with decreasing false - alarm probability .
hence , we conclude that sensing performance is tightly linked to error performance and improved sensing leads to lower error rates . since the signals are _
equiprobable _ , the maximum likelihood ( ml ) decision rule is optimal in the sense that it minimizes the average probability of error @xcite .
since cognitive transmission is allowed only when the channel is sensed as idle under osa scheme , it is enough to evaluate the ml decision rule under sensing decision @xmath3 , which can be expressed as @xmath198 the above decision rule can further be expressed as @xmath199 above maximization simply becomes the comparison of the likelihood functions of the received signals given the transmitted signals @xmath103 . without loss of generality , we consider the signal constellation point @xmath103 .
then , the decision region for the in - phase component of this signal constellation point is given by ( [ eq : comparision_ml1 ] ) .
the right - side boundary of the corresponding decision region , which can be found by equating the likelihood functions in ( [ eq : comparision_ml1])([eq : comparision_ml2 ] ) shown at the top of next page . gathering the common terms together
, the expression in ( [ eq : comparision_ml1 ] ) can further be written as in ( [ eq : comparision_ml3 ] ) given on the next page .
we note that all the terms on the left - hand side of ( [ eq : comparision_ml3 ] ) other than the terms inside the parentheses are nonnegative .
let us now consider these difference terms .
inside the first set of parentheses , we have @xmath202 which can easily be seen to be greater than zero if @xmath203 and is zero if @xmath204 .
the same is also true for the term inside the second set of parentheses given by @xmath205 therefore , the inequality in ( [ eq : comparision_ml3 ] ) can be reduced to @xmath206 similarly , ( [ eq : comparision_ml2 ] ) simplifies to @xmath207 from these observations , we immediately conclude that the decision rule to detect @xmath208 involves comparing @xmath209 with thresholds located at midpoints between the received neighboring signals . following the same approach
, we can determine the decision region for the quadrature component of the signal constellation point @xmath103 by comparing the likelihood functions in ( [ eq : comparision_ml1_im ] ) ( [ eq : comparision_ml2_im ] ) shown on the next page , under the sss scheme , the secondary users are allowed to transmit in busy - sensed channel ( i.e. , under sensing decision @xmath4 ) as well . since the simplified decision rules in ( [ eq : comparision_ml3-simplified ] ) and ( [ eq : comparision_ml2-simplified ] ) do not depend on the sensing decision @xmath213 , the same set of inequalities are obtained for the ml detection rule under sensing decision @xmath4 , leading to the same conclusion regarding the decision rule and thresholds .
@xmath214 p. j. smith , p. a. dmochowski , h. a. suraweera , and m. shafi , the effects of limited channel knowledge on cognitive radio system capacity , " _ ieee trans .
927 - 933 , feb . 2013 .
x. kang , y .- c .
liang , a. nallanathan , h. k. garg , and r. zhang , optimal power allocation for fading channels in cognitive radio networks : ergodic capacity and outage capacity , " _ ieee trans .
wireless commun .
_ , vol . 8 , no .
940 - 950 , feb . 2009 .
r. sarvendranath and n. b. mehta ,
antenna selection in interference - constrained underlay cognitive radios : sep - optimal rule and performance benchmarking , " _ ieee trans .
496 - 506 , feb . 2013 .
suraweera , j. p. smith , and m. shafi , capacity limits and performance analysis of cognitive radio with imperfect channel knowledge , " _ ieee trans .
4 , pp . 1811 - 1822 , may 2010 .
i. f. akyildiz , w .- y .
lee , m. c. vuran , and s. mohanty , next generation / dynamic spectrum access / cognitive radio wireless networks : a survey , " _ comput .
13 , pp . 2127 - 2159 , sep . 2006 .
alouini and a. j. goldsmith , a unified approach for calculating the error rates of linearly modulated signals over generalized fading channels , " in _ ieee trans .
1324 - 1334 , sep . 1999
j. w. craig , a new , simple , and exact result for calculating the probability of error for two - dimensional signal constellations , " in _ proc .
ieee military communications conf .
( milcom91 ) _ , mclean , va , pp .
571 - 575 , oct . 1991 . | this paper studies the symbol error rate performance of cognitive radio transmissions in the presence of imperfect sensing decisions .
two different transmission schemes , namely sensing - based spectrum sharing ( sss ) and opportunistic spectrum access ( osa ) , are considered . in both schemes ,
secondary users first perform channel sensing , albeit with possible errors . in sss , depending on the sensing decisions , they adapt the transmission power level and coexist with primary users in the channel . on the other hand , in osa ,
secondary users are allowed to transmit only when the primary user activity is not detected .
initially , for both transmission schemes , general formulations for the optimal decision rule and error probabilities are provided for arbitrary modulation schemes under the assumptions that the receiver is equipped with the sensing decision and perfect knowledge of the channel fading , and the primary user s received faded signals at the secondary receiver has a gaussian mixture distribution .
subsequently , the general approach is specialized to rectangular quadrature amplitude modulation ( qam ) .
more specifically , the optimal decision rule is characterized for rectangular qam , and closed - form expressions for the average symbol error probability attained with the optimal detector are derived under both transmit power and interference constraints .
the effects of imperfect channel sensing decisions , interference from the primary user and its gaussian mixture model , and the transmit power and interference constraints on the error rate performance of cognitive transmissions are analyzed .
cognitive radio , channel sensing , fading channel , gaussian mixture noise , interference power constraint , pam , probability of detection , probability of false alarm , qam , symbol error probability . |
the reliable calculation of physical properties of interacting quantum mechanical systems presents a formidable task . typically , one has to cope with the interplay of different energy - scales possibly covering several orders of magnitude even for simple situations .
approximate tools like perturbation theory , but even numerically exact techniques can usually handle only a restricted window of energy scales and are furthermore limited in their applicability by the approximations involved or the computational resources available .
in addition due to the divergence of certain classes of feynman diagrams some of the interesting many - particle problems can not be tackled by straightforward perturbation theory .
a general scheme that is designed to handle such multitude of energy scales and competition of divergences is the renormalization group @xcite .
the idea of this approach is to start from high energy scales , leaving out possible infrared divergences and work one s way down to the desired low - energy region in a systematic way .
the precise definition of `` systematic way '' does in general depend on the problem studied .
in particular for interacting quantum mechanical many - particle systems , two different schemes attempting a unique , problem independent prescription have emerged during the past decade .
one is wegner s flow equation technique @xcite , where a given hamiltonian is diagonalized by continuous unitary transformation . from the final result
one can extract detailed information about the structure of the ground state and low - lying excitations .
this technique has been applied successfully to both fermionic and bosonic problems @xcite . however ,
in general it becomes a rather cumbersome task to really calculate physical quantities , especially dynamics from correlation functions . here
, one typically has to introduce further approximations @xcite , again tightly tailored for the problem under investigation .
the second field theoretical approach , which we want to focus on in the following , is based on a functional representation of the partition function of the system under consideration .
it has become known as functional renormalization group ( frg ) @xcite .
a detailed description of the frg will be given in the next section ; here we make some principle remarks and discuss previous applications .
different versions of the frg have been developed over the last few years .
one either generates an exact infinite hierarchy of coupled differential equations for the amputated connected @xmath0-particle green functions of the many - body system @xcite , or the one - particle irreducible @xmath0-particle vertex functions @xcite respectively . for explicit calculations one has to truncate the set of equations which is the major approximation involved . at what level this truncation
is performed to obtain a tractable set of equations depends on the complexity of the problem .
we here exclusively study the one - particle irreducible version of the frg .
it has the advantage of including self - energy corrections already in low truncation orders and being formally easily extendable to higher orders @xcite . up to now
, most applications of the frg in many - body physics concentrate on low - dimensional , interacting fermion systems where it provides a possibility to sum diverging classes of diagrams . the homogeneous two - dimensional hubbard model @xcite , the homogeneous one - dimensional tomonaga - luttinger model @xcite , and the one - dimensional lattice model of spinless fermions with nearest neighbor interaction and local impurities @xcite were investigated .
the focus was put on properties of the system close to the fermi surface , for example on the hierarchy of interactions to identify possible instabilities @xcite or on tomonaga - luttinger liquid exponents @xcite .
the frequency dependence of the vertex functions was mostly neglected @xcite . in this paper
we investigate the frequency dependence of the self - energy and the effective interaction . for this purpose , we study two different zero - dimensional ( local ) models : the quantum anharmonic oscillator and a well - known problem of solid state physics , the single impurity anderson model ( siam ) .
the former has quite often been used as a `` toy model '' to investigate the performance of different approximation schemes of many - particle physics @xcite . for this problem conventional perturbation theory
is regular although it generates a generic example of an asymptotic series @xcite and one expects that compared to perturbation theory the frg leads to a better agreement with the exact solution at larger coupling constants ( `` renormalization group enhanced perturbation theory '' ) .
we calculate the ground state energy and the energy of the first excited state as well as the spectral function of the propagator .
exact results for these observables can numerically be obtained quite easily .
the siam has a known hierarchy of energy scales , and presents a true challenge to any many - body tool due to the generation of an exponentially small energy scale , the kondo scale , leading for example to a sharp resonance in the single - particle spectrum .
no exact solutions for dynamical quantities of the model are known .
we present frg results for the one - particle spectral function of the impurity site and compare them to conventional second order perturbation theory ( in the interaction @xmath1 of the impurity site ) and wilson s numerical renormalization group ( nrg ) . in both models
the truncated frg scheme , which is correct at least to second order perturbation theory in the interaction , leads to a considerable improvement compared to plain second order perturbation theory .
the paper is organized as follows . in the next section we present a detailed discussion of the frg .
the third part contains the application to the quantum anharmonic oscillator , while in the forth section we discuss the frg for the single impurity anderson model .
a summary and outlook concludes the paper .
expressed as a functional integral the grand canonical partition function of a system of quantum mechanical particles ( either fermions or bosons ) interacting via a two - particle potential can be written as @xcite @xmath2^{-1}\psi ) - s_{\rm int}\left(\ { \bar \psi \ } , \ { \psi\ } \right ) \right\ } } \ ; , \end{aligned}\ ] ] with @xmath3 and @xmath4 being the non - interacting partition function . here
@xmath5 and @xmath6 denote either grassmann ( fermions ) or complex ( bosons ) fields .
the multi - indices @xmath7 stand for the quantum numbers of the one - particle basis in which the problem is considered ( e.g. momenta and spin directions ) and matsubara frequencies @xmath8 .
we have introduced the short hand notation @xmath9^{-1 } \psi \right ) = \sum_{k , k ' } \bar \psi_{k } \left[{\mathcal g}^0 \right]^{-1}_{k , k ' } \psi_{k ' } \ ; , \end{aligned}\ ] ] with the propagator @xmath10 of the related non - interacting problem given as a matrix .
the anti - symmetrized ( fermions ) or symmetrized ( bosons ) matrix elements of the two - particle interaction are denoted by @xmath11 .
they contain the energy conserving factor @xmath12 and the factor @xmath13 , with @xmath14 being the inverse temperature .
we consider units such that @xmath15 and @xmath16 .
the generating functional of the @xmath0-particle green function is given by @xmath17^{-1 } \psi \right ) - s_{\rm int}(\{\bar \psi \ } , \
{ \psi \ } ) \nonumber \right .
\\ & & \left
. - \left ( \bar \psi , \eta\right ) - \left(\bar \eta , \psi \right ) \right\ } \ ; , \end{aligned}\ ] ] with @xmath18 and the external source fields @xmath19 and @xmath20 . from this
the generating functional of the connected @xmath0-particle green function follows as @xmath21 } \ ; .\end{aligned}\ ] ] the ( connected ) @xmath0-particle green function @xmath22 can be obtained by taking functional derivatives @xmath23 by a legendre transformation @xmath24^{-1 } \phi\right ) \ ; , \end{aligned}\ ] ] the generating functional of the one - particle irreducible vertex functions @xmath25 , with external source fields @xmath26 and @xmath27 and @xmath28 is obtained .
note that in contrast to the usual definition @xcite of @xmath29 we have added a term @xmath30^{-1 } \phi\right)$ ] in eq .
( [ gammadef ] ) for convenience ( see below ) .
the relation between the @xmath22 and @xmath25 can be found in text books @xcite .
the 0-particle vertex @xmath31 provides the interacting part of the grand canonical potential @xmath32 @xmath33 for the 1-particle green function we obtain @xmath34^{-1 } \right]^{-1}_{k',k } \ ; , \end{aligned}\ ] ] where @xmath35^{-1 } - \sigma \right]^{-1}_{k',k } \ ; , \end{aligned}\ ] ] with the self - energy @xmath36 , and @xmath37 for fermions or @xmath38 for bosons , respectively .
this implies the relation @xmath39 .
the 2-particle vertex @xmath40 is usually referred to as the effective interaction . for @xmath41 ,
the @xmath0-particle interaction @xmath25 has diagrammatic contributions starting at @xmath0-th order in the two - particle interaction . in eqs .
( [ partitionfunction ] ) and ( [ genfunct ] ) we replace the non - interacting propagator by a propagator @xmath42 depending on a cutoff @xmath43 and @xmath4 by @xmath44 determined using @xmath45 .
the boundary conditions for the cutoff @xmath46 $ ] are taken as @xmath47 i.e. at the starting point @xmath48 no degrees of freedom are `` turned on '' while at @xmath49 the cutoff independent problem is recovered . in our applications
we use a sharp matsubara frequency cutoff with @xmath50 and consider @xmath51 @xcite . through @xmath45 the quantities defined in eqs.([partitionfunction ] ) to ( [ vertexfundef ] ) acquire a @xmath43-dependence .
one now derives a functional differential equation for @xmath52 . from this , by expanding in powers of the external sources , a set of coupled differential equations for the @xmath53 is obtained . as a first step we differentiate @xmath54 with respect to @xmath43 , which after straightforward algebra leads to @xmath55 with @xmath56^{-1 } \ ; .\end{aligned}\ ] ]
considering @xmath26 and @xmath57 as the fundamental variables we obtain from eq .
( [ gammadef ] ) @xmath58 applying the chain rule and using eq .
( [ flusswc ] ) this leads to @xmath59 the last term in eq .
( [ gammadef ] ) is exactly cancelled , which a posteriori justifies its inclusion . using the well known relation @xcite between the second functional derivatives of @xmath60 and @xmath61
we obtain the functional differential equation @xmath62 where @xmath63 stand for the upper left block of the matrix @xmath64^{-1 } & \frac{\delta^2 \gamma^{\lambda}}{\delta \bar \phi \delta \bar \phi } \\
\frac{\delta^2 \gamma^{\lambda}}{\delta \phi \delta \phi } & \frac{\delta^2 \gamma^{\lambda}}{\delta \phi \delta \bar \phi } - \left [ \left[{\mathcal g}^{0,\lambda } \right]^{-1}\right]^t \end{array } \right)^{-1 } \end{aligned}\ ] ] and the upper index @xmath65 denotes the transposed matrix . to obtain differential equations for the @xmath66 which include self - energy corrections we express @xmath67 in terms of @xmath68 instead of @xmath69 .
this is achieved by defining @xmath70 and using @xmath71^{-1 } - \zeta \gamma_1^{\lambda } \right]^{-1 } \end{aligned}\ ] ] which leads to @xmath72 with @xmath73^t \end{array }
\right ) \left ( \begin{array}{cc } { \mathcal u}_{\bar \phi , \phi } & \frac{\delta^2 \gamma^{\lambda}}{\delta \bar \phi \delta \bar \phi } \\
\frac{\delta^2 \gamma^{\lambda}}{\delta \phi \delta \phi } & \zeta { \mathcal u}^{t}_{\bar \phi , \phi } \end{array } \right ) \right]^{-1 } \ ; .\end{aligned}\ ] ] for later applications it is important to note that @xmath74 as well as @xmath75 and @xmath76 are at least quadratic in the external sources .
the initial condition for the exact functional differential equation ( [ gammaflussalt ] ) can either be obtained by lengthy but straightforward algebra not presented here , or by the following simple argument : at @xmath77 , @xmath78 ( no degrees of freedom are `` turned on '' ) and in a perturbative expansion of the @xmath79 the only term which does not vanish is the bare 2-particle vertex .
we thus find @xmath80 an exact infinite hierarchy of flow equations for the @xmath66 can be obtained by expanding eq .
( [ vtildedef ] ) in a geometric series and @xmath52 in the external sources @xmath81 the equation for @xmath82 reads @xmath83 via @xmath68 the derivative of @xmath82 couples to the one - particle self - energy .
for the flow of the self - energy we obtain @xmath84 with the so - called single scale propagator @xmath85 here @xmath86 is a matrix in the variables not explicitly written , i.e. @xmath87_{q',q } = \gamma_2^{\lambda}(k',q ' ; k , q ) $ ] . diagrammatically eq . ( [ gamma1fl ] ) is shown in fig .
the derivative of @xmath88 is determined by @xmath88 and the 2-particle vertex @xmath89 .
thus an equation for @xmath90 is required @xmath91^{t } \gamma_2^{\lambda}(k_1 ' , k_2 ' ; \ldots , \ldots ) \big ) \nonumber \\ & & + \zeta \mbox { tr}\ , \big ( { \mathcal s}^{\lambda } \gamma_2^{\lambda}(k_1 ' , \ldots ; k_1,\ldots ) { \mathcal g}^{\lambda } \gamma_2^{\lambda}(k_2 ' , \ldots ; k_2,\ldots ) \nonumber \\ & & + \zeta \left [ k_1 ' \leftrightarrow k_2 ' \right ] + \zeta \left [ k_1 \leftrightarrow k_2 \right ] + \left [ k_1 ' \leftrightarrow k_2 ' , k_1 \leftrightarrow k_2 \right ] \big ) \ ; .\end{aligned}\ ] ] the corresponding diagrammatic representation is shown in fig.[fig2 ] . the right hand side ( rhs ) of eq .
( [ gamma2fl ] ) contains @xmath92 , @xmath93 , and the 3-particle vertex @xmath94 . for @xmath95
the equation for @xmath96 contains @xmath97 with @xmath98 .
the initial condition for the @xmath79 can be obtained from eq.([anfanggamma ] ) and is given by @xmath99 we here refrain from explicitly presenting equations for @xmath100 with @xmath41 since later on the set of differential equations is truncated by setting @xmath101 , which implies that vertices with @xmath41 do not contribute . .
the slashed line stands for the single scale propagator @xmath102 .
[ fig1],scaledwidth=28.0% ] .
the slashed line stands for the single scale propagator @xmath102 the unslashed line for @xmath68 .
[ fig2],scaledwidth=60.0% ] following the above systematics , a truncation scheme emerges quite naturally .
if , for @xmath103 , the vertex @xmath104 on the rhs of the coupled flow equations is replaced by its initial condition @xmath105 , a closed set of equations for @xmath106 with @xmath107 follows .
this set of differential equations can then be integrated over @xmath43 starting at @xmath108 down to @xmath49 providing approximate expressions for the @xmath109 of the original ( cutoff free ) problem with @xmath107 .
expanding @xmath106 in terms of the bare interaction , conventional perturbation theory for the grand canonical potential , the self - energy , the effective interaction and higher order vertex functions can be recovered from an iterative treatment of the flow equations .
it is easy to show that the vertex functions obtained from the truncated equations are at least correct up to order @xmath110 in the bare interaction . to obtain a manageable set of equations in applications of the frg to one- and two - dimensional quantum many - body problems
@xcite , further approximations in addition to the above truncation scheme were necessary . in the following two sections we will avoid such additional approximations and solve the truncated frg equations to order @xmath111 for two zero - dimensional ( local ) interacting many - particle problems : the quantum harmonic oscillator with a quartic perturbation and the siam .
very recently a modified version of the flow equation for @xmath93 was suggested . as we will consider the truncation order @xmath111 we here only describe this modified scheme for this order .
guided by the idea of an improved fulfillment ( compared to the scheme described above ) of a certain ward identity katanin replaced the combined propagator @xmath112 in the last five terms of eq .
( [ gamma2fl ] ) ( the first term does not contribute to order @xmath111 ) by @xmath113 @xcite . using @xmath114 obtained from eq .
( [ gg ] ) it is obvious that the terms added to eq .
( [ gamma2fl ] ) are at least of third order in the bare interaction . using this modification in the order @xmath111 equations for @xmath92 and @xmath93 leads to the exact solution for certain exactly solvable models ( reduced bcs model , interacting fermions with forward scattering only ) .
this suggests that the replacement possibly improves the results of the truncated frg also for other models . in the next section
we show that this is indeed the case for the harmonic oscillator with a quartic perturbation .
the same holds for the siam as discussed in sect .
in appropriate units the hamiltonian of the harmonic oscillator with a quartic perturbation is given by @xmath115 with the position operator @xmath116 , the momentum operator @xmath117 , and the coupling constant @xmath118 .
we here focus on @xmath119 and are interested in low - lying eigenenergies @xmath120 as well as the ( imaginary ) time - ordered propagator @xmath121 \left| e_0 \right > \ ; , \ ] ] with @xmath122 being the eigenstates of the hamiltonian in eq .
( [ anhaos ] ) .
the fourier transform of the propagator can be written as @xmath123^{-1}- \sigma(i\omega ) } \ ; , \end{aligned}\ ] ] where we have introduced the self - energy @xmath36 and the non - interacting propagator @xmath124 in contrast to the more general notation used in the last section , propagators and the self - energy only depend on a single frequency and do no longer contain the energy conserving @xmath125-function ( @xmath119 ) here .
the propagator has the lehmann representation @xmath126 \left| \left < e_0 \right| ( a + a^{\dag } ) \left| e_n \right > \right|^2 \ ; , \end{aligned}\ ] ] where @xmath127 , @xmath128 denote the usual raising and lowering operators .
the spectral weights @xmath129 and energies @xmath120 fulfill the f - sum rule @xmath130 it turns out that for coupling constants @xmath131 considered here the sums in eqs .
( [ lehmann ] ) and ( [ sumrule ] ) are dominated by the first term .
for this reason only the first few eigenstates and eigenenergies are required to obtain accurate ( `` numerically exact '' ) results for @xmath132 .
these can quite easily be obtained by expressing @xmath133 in the basis of eigenstates @xmath134 of the unperturbed harmonic oscillator and numerically diagonalizing the upper left corner of the ( infinite ) matrix @xmath135 with @xmath136 and a sufficiently large @xmath137 . for @xmath131
, @xmath138 turns out to be large enough to fulfill the sum rule eq .
( [ sumrule ] ) to very high precision .
second order perturbation theory for the @xmath118-dependent part of the ground state energy yields @xmath139 and for the self - energy one obtains @xmath140 within the frg approximate expressions for @xmath141 and @xmath142 can be obtained from the poles and residues of the propagator @xmath143 .
furthermore , since eqs .
( [ lehmann ] ) and ( [ sumrule ] ) are dominated by the first terms we only consider @xmath144 and @xmath145 @xcite .
second order approximations for these quantities are given by the smallest pole of @xmath146^{-1}$ ] and the related residue .
it is important to note that this approximation @xmath147 agrees with @xmath148 , where @xmath149 is determined directly from rayleigh - schrdinger perturbation theory , only up to second order in @xmath118 , but is closer to the exact @xmath144 . within mean field theory , @xmath150 and a frequency independent @xmath151
are given by @xmath152 and @xmath153 is the solution of the self - consistency equation @xmath154 from the mean field propagator one obtains @xmath155 the functional integral representation of the grand canonical partition function of the hamiltonian eq .
( [ anhaos ] ) reads @xmath156^{-1 } x ) /2 - s_{\rm int}\left(\ { \bar x \ } , \ { x \ }
\right ) \right\ } } \ ; , \end{aligned}\ ] ] with @xmath157 bosonic matsubara frequencies @xmath158 , and complex fields @xmath159 . as outlined in the last section and
using a frequency cutoff eq.([matsubaracutoff ] ) flow equations for the @xmath66 can be obtained . here
we focus on the equations in truncation order @xmath111 . for @xmath160 and after introducing @xmath161 we find @xmath162 \ ; , \end{aligned}\ ] ] with the initial condition @xmath163 . at the end of the flow , @xmath164 directly provides the frg approximation @xmath165 for the @xmath118-dependent part of the ground state energy .
the flow equation for the self - energy follows as @xmath166 with the initial condition @xmath167 and the frg approximation for the self - energy @xmath168 . here
@xmath169 denotes the totally symmetric 2-particle vertex which , in contrast to the vertex @xmath93 introduced in the last section , does not contain an energy conserving @xmath125-function and factors of @xmath170 .
it depends on only three frequencies , but the fourth will nevertheless always be included in the following . to derive eqs .
( [ gamma0ao ] ) and ( [ gamma1ao ] ) one has to deal with products of delta functions @xmath171 and terms involving step functions @xmath172 .
these seemingly ambiguous expressions are well defined and unique if the sharp cutoff is implemented as a limit of increasingly sharp broadened cutoff functions @xmath173 , with the broadening parameter @xmath174 tending to zero .
the expressions can then be conveniently evaluated by using the following relation @xcite , valid for arbitrary continuous functions @xmath175 : @xmath176 \to \delta(x-\lambda ) \int_0 ^ 1 f(t ) \ , dt \ ; , \ ] ] where @xmath177 .
for @xmath169 the flow equation reads @xmath178\end{aligned}\ ] ] where @xmath179 stands for two different products of propagators . in the conventional frg scheme it is given by @xmath180 while in the modified scheme @xcite one obtains @xmath181 to explicitly evaluate @xmath182 eq.([morristrick ] ) has to be used , where special care has to be taken for the case @xmath183 . in the conventional scheme
@xmath184 contains a factor @xmath185 and the integral over @xmath186 in eq .
( [ gamma2ao ] ) can be performed analytically . as a function of @xmath118 .
different approximations ( second order perturbation theory [ thin dashed line ] , mean field [ dotted line ] , conventional frg [ thick dashed line ] , modified frg [ dashed - dotted line ] ) are compared to the exact result ( solid line ) .
the inset shows the difference between the exact result and the two frg approximations .
[ fig3],scaledwidth=80.0% ] , but for the energy difference @xmath144 of the first excited state and the ground state .
[ fig4],scaledwidth=80.0% ] , but for the spectral weight @xmath145 of the first peak .
[ fig5],scaledwidth=80.0% ] to numerically solve the set of differential equations ( [ gamma0ao ] ) , ( [ gamma1ao ] ) , and ( [ gamma2ao ] ) we have discretized the frequencies ( which at @xmath119 are continuous ) on a linear mesh @xmath187 with @xmath188 @xcite . by increasing @xmath189 and decreasing @xmath125 convergence
can be achieved up to the required accuracy . for our purposes @xmath190 and @xmath191 turned out to be appropriate .
this leads to a set of roughly @xmath192 coupled equations .
the @xmath43-integration is started at @xmath193 making sure that further increasing @xmath194 does not lead to significant changes in the results .
[ fig3 ] to [ fig5 ] show comparisons of @xmath195 , @xmath144 , and @xmath145 with the different approximations considered here ( second order perturbation theory , mean field , conventional frg , modified frg ) and the exact results . although the approximate frgs correctly reproduce only the first two derivatives with respect to @xmath118 at @xmath196 , they give extremely accurate results even up to @xmath197 , while conventional second order perturbation theory can only be trusted for @xmath198 .
this provides an impressive example of the power of `` renormalization group enhanced perturbation theory '' . comparing the two frg approximations
the modified scheme is roughly a factor of two closer to the exact result and thus indeed a substantial improvement @xcite . to avoid the problem of analytic continuation
( see the next section ) the results for @xmath199 and @xmath200 were obtained by fitting a function @xmath201 with @xmath127 and @xmath202 as fitting parameters to @xmath203^{-1}$ ]
. assuming this fitting form we have used that the spectral function is dominated by the first peak @xcite . for the problem studied also mean field theory leads to fairly accurate results ( but not as good as the frg ) .
this is related to the fact that low - lying eigenstates of the hamiltonian in eq .
( [ anhaos ] ) can be described quite well by the eigenstates of a harmonic oscillator with a shifted frequency determined self - consistently .
in contrast to the anharmonic oscillator studied in the previous section , the single impurity anderson model @xcite @xmath204 consists of two subsystems , namely a `` conduction band '' with continuous energy spectrum described by the first term in eq .
( [ equ : siam ] ) and a localized level ( typically referred to as `` @xmath205''-state ) with energy @xmath206 relative to the chemical potential @xmath207 of the conduction electrons .
two electrons occupying the localized level are in addition subject to a coulomb repulsion @xmath208 .
both subsystems are coupled via the hybridization in the last term .
as the two - body interaction is restricted to a single level the siam falls into the of class of zero - dimensional interacting systems .
while this model looks rather simple , it contains all ingredients that make it a complicated many - body problem .
the bare energy scales of the model are the bandwidth @xmath209 of the band electrons , the local energy @xmath206 , the coulomb repulsion @xmath1 , and the bare level width generated by the hybridization , @xmath210 , where @xmath211 denotes the density of states of the conduction states which is assumed to be slowly varying . in the following
we restrict ourselves to the particle - hole symmetric case @xmath212 .
as in the parameter regime @xmath213 the charge fluctuations on the @xmath205-level from the average value @xmath214 are small , it can effectively be described by a spin antiferromagnetically coupled to the conduction electron spin density @xcite , the so - called kondo model @xcite .
this antiferromagnetic coupling leads to a screening of the local spin by the conduction electrons with a characteristic energy scale @xmath215 , the kondo temperature .
as is apparent from this expression , @xmath216 depends non - analytically on @xmath217 , signalling the occurrence of infrared divergences in perturbation theories in @xmath218 @xcite and a severe problem for computational techniques , namely the task to resolve an exponentially small energy scale . on the other hand , the physics in the regimes @xmath219 and @xmath220 is comparatively simple . for @xmath219
it is governed by charge excitations with energies @xmath206 and @xmath221 with a life - time given by @xmath222 . in the other limit @xmath220
it has been worked out using wilson s nrg @xcite that the system can again be described by a hamiltonian of the form eq .
( [ equ : siam ] ) , but with @xmath223 , @xmath224 and @xmath217 replaced by @xmath225 .
this regime has been coined `` local fermi liquid '' by nozires @xcite .
this effective description yields a narrow resonance at the chemical potential in the spectral function of the @xmath205-green function which in the symmetric case takes the form @xmath226 here @xmath227 and @xmath228 denotes all self - energy contributions of second and higher order in @xmath1 . as a special feature of the symmetric case , already the approximation to only keep the second order self - energy @xmath229 describes the qualitative behavior of the @xmath205-spectral function for different values of @xmath230 correctly @xcite . for @xmath231
there is a lorentzian peak at the chemical potential with a width differing little from @xmath217 . for @xmath232
most of the spectral weight is in the high energy peaks near @xmath233 with a narrow resonance at @xmath234 .
quantitatively the width and shape of this kondo ( or abrikosov - suhl ) resonance is described poorly using @xmath229 , vanishing only @xmath235 . from the results of sect . 3
one expects that the application of the functional renormalization group to the siam leads to improvements over the direct perturbational description of the kondo resonance . for the siam
the frg results can be compared to the outcome of nrg calculations @xcite .
an important aspect of the treatment of the siam with the frg is that , while the numerical effort of the nrg method increases exponentially with the number of impurity degrees of freedom , e.g. in case of additional orbital degrees of freedom or for a system of many magnetic impurities , the increase in computational resources necessary in the frg approach described below is governed at most by a power - law . the starting point of our frg approach are again eqs .
( [ gamma1fl ] ) and ( [ gamma2fl ] ) .
we also use @xmath111 , i.e. we replace the 3-particle - vertex by its initial condition @xmath236 , and consider @xmath119 only .
we again use the frequency cutoff described in eq .
( [ matsubaracutoff ] ) and because we presently concentrate on spectral properties only flow equations for the self - energy and for the 2-particle - vertex ( and not the ground state energy ) will be derived .
note that the calculation of two - particle properties is possible within the same scheme without further difficulties @xcite and will be discussed in a forthcoming publication . for the siam
all green functions @xmath237 different from @xmath238 can be expressed in terms of @xmath238 and green functions of the non - interacting system . for the calculation of @xmath238 the multi - index @xmath239
is replaced by a spin - index @xmath240 and a frequency @xmath186 .
the 2-particle - vertex can then be written as @xmath241 this can be used to perform the sum over the spins in eq .
( [ gamma1fl ] ) and we find for the self - energy @xmath242.\ ] ] in order to derive a flow equation for @xmath243 , with @xmath244 , the spin - sum in eq .
( [ gamma2fl ] ) has to be performed as well .
this leads to the lengthy expression eq .
( [ gamma2:siam ] ) , presented in appendix a. the initial values are given by @xmath245 and @xmath246 .
in the numerical solution of these flow equations we start integrating at finite @xmath247 .
the integration from @xmath248 to @xmath194 can be performed analytically . up to corrections of order
@xmath249 the new initial conditions are @xmath250 and @xmath251 .
@xmath252 and @xmath253 are calculated the same way as for the anharmonic oscillator using eq .
( [ morristrick ] ) , and are given by @xmath254^{-1}-\se{i\nu } } \ ; , \ ] ] with @xmath255^{-1 } = i \nu - \delta(i \nu ) - \epsilon_d$ ] and @xmath256^{-1}-\se{i\nu } } \frac{\h{\nu'}}{\left[\gn{i\nu'}\right]^{-1}-\se{i\nu'}}\ ] ] with @xmath257 .
in our calculations we use the limit of an infinite bandwidth , i.e. @xmath258 . the results of second order perturbation theory for the self - energy can be recovered by replacing the self - energy on the rhs of eqs .
( [ self : siam ] ) and ( [ gamma2:siam ] ) and @xmath259 on the rhs of eq .
( [ gamma2:siam ] ) by their initial values .
it turns out that the frg version using @xmath260 gives good results for @xmath261 , but fails to provide the expected improvement compared to the use of @xmath229 for @xmath262 .
this will be discussed in more detail in a forthcoming publication .
for this reason we follow @xcite and replace @xmath263 by @xmath264 in eq .
( [ gamma2:siam ] ) .
applying eq .
( [ morristrick ] ) yields @xmath265^{-1}-\se{i\nu } } \frac{\h{\nu'}}{\left[\gn{i\nu'}\right]^{-1}-\se{i\nu'}}\\[8 mm ] & -&\ds\ddl\se{i\nu}~\frac{\h{\nu}}{[\left[\gn{i\nu}\right]^{-1}-\se{i\nu}]^2 } \frac{\h{\nu'}}{\left[\gn{i\nu'}\right]^{-1}-\se{i\nu'}}. \end{array}\ ] ] for the integration of the @xmath119-flow equations the continuous frequencies again have to be discretized .
however , in contrast to the anharmonic oscillator it is not sufficient to work with a linear mesh .
because for @xmath232 the kondo resonance is visible only on an exponentially small energy scale around the fermi level , we use a combination of a linear and a logarithmic mesh .
this enables us to recover both , the high energy as well as the low - energy physics keeping the number of frequencies to a manageable size .
the numerical effort grows with the third power of the number of frequencies in the conventional version and with almost the forth power using the modified version , due to the absence of a @xmath125-function in the last term of eq .
( [ kat : siam ] ) . in the frg
one naturally obtains the self - energy for imaginary frequencies . in order to calculate spectral functions ,
an analytic continuation to the real axis is necessary , which is known to be an ill - posed problem . since the results from frg are not subject to statistical errors or noise as for example quantum monte carlo data , we have applied the method of pad approximation @xcite to obtain @xmath266 from @xmath267 .
we find that especially the low - energy part of the spectral function which contains the kondo resonance can be reliably extracted if sufficiently many frequencies close to the fermi level are used ( see below ) . from frg for @xmath268 ( circles ) compared to nrg ( dashed line ) and second order perturbation theory ( dotted line ) .
+ right panel : @xmath269 for frg obtained from pad approximants for the data in left panel .
the inset shows an enlarged view around @xmath270 .
[ fig : siam1],scaledwidth=90.0% ] as an example , we present in fig . [
fig : siam1 ] the calculated @xmath271 ( left panel ) and the corresponding @xmath269 from the pad approximation ( right panel ) for @xmath272 .
the dashed and dotted lines represent results from nrg and second order perturbation theory , respectively .
the inset in the right panel shows an enlarged view of the region around @xmath270 .
apparently , the pad approximation to the frg provides reliable results , in particular around @xmath270 , and recovers the perturbation theory as expected for such small value of @xmath1 .
discrepancies to nrg for large @xmath8 can be traced to broadening effects in the nrg .
the true test for the method however is a comparison of the self - energy to nrg results for larger values of @xmath1 and in particular the behavior on low - energy scales .
such a comparison is presented in fig .
[ fig : siam2 ] for @xmath272 , @xmath273 , and @xmath274 .
for @xmath275 the frg results are still in excellent agreement with the nrg data while second order perturbation theory already deviates . for @xmath276
we observe a significant dependency of the general structures and also of the slope at @xmath277 on details of the discretization mesh .
it seems necessary to have an extremely fine resolution around @xmath277 and a sufficient resolution around @xmath278 . with an exponentially vanishing low - energy scale
these constraints are hard to fulfill with both linear and logarithmic meshes keeping the number of flow equations to a manageable size .
the bottom left pannel contains the best frg data we were able to obtain so far .
a more extended discussion of this issue will be presented in a forthcoming publication . from frg ( circles ) for @xmath272 , @xmath273 and @xmath274
compared to nrg ( dashed line ) and second order perturbation theory ( dotted line ) .
+ bottom right : effective mass @xmath279 from frg ( circles ) , nrg ( squares ) and second order perturbation theory ( triangles).[fig : siam2],scaledwidth=90.0% ] in addition to the self - energy the effective mass @xmath279 obtained from @xmath280 as a function of @xmath1 is depicted , in the bottom right panel of fig .
[ fig : siam2 ] .
this quantity is directly related to the kondo temperature @xcite .
compared to perturbation theory the effective mass is much closer to the very accurate values determined from the nrg . the evolution of the spectral function of the @xmath205-level , @xmath281 for the above values of @xmath1 is collected in fig . [ fig : siam3 ] .
the inset shows a comparison to nrg and perturbation theory for the region around @xmath270 at @xmath282 .
one can see very nicely the development of the sharp resonance in the spectrum and the formation of the hubbard bands with increasing @xmath1 .
-level for the values of @xmath1 in fig .
[ fig : siam2 ] .
the inset shows a comparison to nrg for @xmath282 for the region around @xmath270.[fig : siam3],scaledwidth=90.0% ] due to the insufficient resolution at larger @xmath8 in the calculation for @xmath282 , the hubbard bands come out too broad here .
we believe that an improved discretization of the energy mesh in the solution of the flow equations will remedy that particular problem .
the important region around @xmath270 on the other hand is captured rather well by the frg .
it is in particular noteworthy that the functional form at low frequencies ( see the inset ) apparently does not follow a lorentzian but rather , like the nrg , the more complex scaling form with logarithmic tails predicted by logan _
we have presented an application of the functional renormalization group technique to solve for the dynamics of zero - dimensional interacting quantum problems . as particular examples
, we discussed the anharmonic oscillator and the single impurity anderson model . in both cases the frg proved to be a substantial improvement over conventional low - order perturbation theory and rather close to the very accurate results obtained numerically .
we also investigated the differences between conventional frg and a modification suggested by katanin @xcite , which should improve the accuracy of the method further . that this is indeed true
was shown directly for the anharmonic oscillator .
for the single impurity anderson model it was actually necessary to use this modified version to obtain sensible results for values @xmath283 .
an important aspect of the frg is that an extension of the siam to more complex systems , like e.g. orbital degrees of freedom or systems of coupled impurities is straightforward .
this is in principle also true for wilson s nrg . in the latter ,
however , the exponentially increasing hilbert space renders a practical application quickly impossible . for the frg ,
on the other hand , the major modification will be an increase of the number of equations , which means that the numerical effort will increase at most with a power - law .
this feature makes the frg a possible method to study features of complex impurity systems and in particular a potential `` impurity solver '' for mean - field theories of interacting lattice models like the hubbard model in the framework of the dynamical mean - field theory @xcite or the dynamical cluster approximation @xcite .
furthermore , as it becomes clear from the anharmonic oscillator , the frg is of equal complexity for bosonic and fermionic systems .
this makes opens the possibility to study combinations of such degrees of freedom in impurity models . a question which can be addressed using
the frg is the influence of phonons or magnetic fluctuations on low energy scales . within the dynamical mean - field theory the metal - insulator transition in the presence of phonons ( kondo volume collapse @xcite ) as well as the problem of non - fermi liquid formation in for example
cecu@xmath284au@xmath285 @xcite can be studied .
we thank s. dusuel , w. metzner , m. salmhofer , h. schoeller , and g. uhrig for useful discussions .
this work was supported by the sfb 602 of the deutsche forschungsgemeinschaft ( r.h . and k.s . ) .
is grateful to the bundesministerium fr bildung und forschung for support .
depending on the specific parameterization of the 2-particle - vertex flow equations for the relevant frequency dependent parts of the vertex and for the self - energy can be derived by using the spin conservation on the vertex when the sum over spins in eq .
( [ gamma2fl ] ) is performed . for the form of the the vertex given in eq .
( [ gamma2:siama ] ) the corresponding flow equation for the self - energy is given by eq .
( [ self : siam ] ) , and the equation for the frequency dependent part of the vertex @xmath286 with @xmath244 reads @xmath287{l}\!\!\!\ds(i\nu_1',i\nu_2';i\nu_1,i\nu_2 ) = -\frac{1}{2\pi}\int_{-\infty}^{\infty}d\nu \ds\biggl[\mathcal p^{\lambda}(i\nu , i\nu_1+i\nu_2-i\nu ) \\[5 mm ] \bigl(-\gw(i\nu , i\nu_1+i\nu_2-i\nu;i\nu_1,i\nu_2 ) \gw(i\nu_2',i\nu_1';i\nu_1+i\nu_2-i\nu , i\nu)\\[5 mm ] -\gw(i\nu_1+i\nu_2-i\nu , i\nu;i\nu_1,i\nu_2 ) \gw(i\nu_1',i\nu_2';i\nu_1+i\nu_2-i\nu , i\nu)\bigr)\\[5 mm ] \ds+\biggl\{\mathcal p^\lambda(i\nu ,-
i\nu_1+i\nu_1'+i\nu ) \\[5 mm ] \bigl(2\ , \gw(i\nu_1',i\nu;i\nu_1,-i\nu_1+i\nu_1'+i\nu ) \gw(i\nu_2',-i\nu_1+i\nu_1'+i\nu;i\nu_2,i\nu)\\[5 mm ] - \gw(i\nu_1',i\nu;i\nu_1,-i\nu_1+i\nu_1'+i\nu ) \gw(-i\nu_1+i\nu_1'+i\nu , i\nu_2';i\nu_2,i\nu)\\[5 mm ] - \gw(i\nu , i\nu_1';i\nu_1,-i\nu_1+i\nu_1'+i\nu ) \gw(i\nu_2',-i\nu_1+i\nu_1'+i\nu;i\nu_2,i\nu)\bigr)\\[5 mm ] + ( 1'\leftrightarrow2 ' ; 1\leftrightarrow2)\biggr\}\\[5 mm ] \ds-\biggl\{\mathcal p^\lambda(i\nu ,-
i\nu_1+i\nu_2'+i\nu ) \\[5 mm ] \gw(i\nu , i\nu_2';i\nu_1,-i\nu_1+i\nu_2'+i\nu ) \gw(-i\nu_1+i\nu_2'+i\nu , i\nu_1';i\nu_2,i\nu)\\[5 mm ] + ( 1'\leftrightarrow2 ' ; 1\leftrightarrow2)\biggr\ } \biggr ] \ ; .
\end{array } \end{aligned}\ ] ] again @xmath288 stands either for @xmath289 or @xmath290 .
two other possible parameterizations of the 2-particle vertex are @xmath291{l}\d{\nu_1+\nu_2-\nu_1'-\nu_2 ' } \ds \bigl\{\delta_{\sigma_1,\sigma_2}\delta_{\sigma_1',\sigma_2'}\delta_{\sigma_1,\sigma_1'}\ \gp(i\nu_1',i\nu_2';i\nu_1,i\nu_2)\\[5 mm ] \hspace{-2 cm } + \ds \delta_{\sigma_1,-\sigma_2}\delta_{\sigma_1',-\sigma_2'}\ \bigl(\delta_{\sigma_1,\sigma_1'}\ga(i\nu_1',i\nu_2';i\nu_1,i\nu_2)- \delta_{\sigma_1,-\sigma_1'}\ga(i\nu_2',i\nu_1';i\nu_1,i\nu_2)\bigr ) \bigr\}.\end{array}\end{aligned}\ ] ] + and @xmath292{l}\ds s_{\sigma_1 ' , \sigma_2 ' ; \sigma_1 , \sigma_2}\ \gs(i\nu_1',i\nu_2';i\nu_1,i\nu_2)\\[5 mm ] + \ds t_{\sigma_1 ' , \sigma_2 ' ; \sigma_1 , \sigma_2}\ \gt(i\nu_1',i\nu_2';i\nu_1,i\nu_2)\bigr\}\end{array}\ ] ] with @xmath293 both these parameterizations lead to different set of flow equations for the self - energy and to two sets of flow equations for the frequency dependent functions @xmath294 and @xmath295 for the first and @xmath296 and @xmath297 for the second ( compared to one set for the parameterization eq .
( [ gamma2:siama ] ) ) implying an increased numerical effort .
they are usefull to further investigate the processes occurring during the integration of the flow equations . this way we are able to distinguish the behavior of different channels of the interaction .
wilson , rev .
47 * , 773 ( 1975 ) .
glazek and p.b .
wiegmann , phys .
d * 48 * , 5863 ( 1993 ) ; f. wegner , ann .
physik ( leipzig ) * 3 * , 77 ( 1994 ) .
p. lenz and f. wegner , nucl .
b[fs ] * 482 * , 693 ( 1996 ) ; a. mielke , europhys .
lett . * 40 * , 195 ( 1997 ) ; m. ragwitz and f. wegner , eur .
j. b * 8 * , 9 ( 1999 ) .
uhrig , phys .
b * 57 * , r14004 ( 1998 ) ; c. knetter and g.s .
uhrig , eur .
j. b * 13 * , 209 ( 2000 ) .
s. kehrein and a. mielke , ann .
( new york ) * 252 * , 1 ( 1996 ) ; c. slezak , s. kehrein , th . pruschke and m. jarrell , phys .
b * 67 * , 184408 ( 2003 ) . j. polchinski , nucl .
b * 231 * , 269 ( 1984 ) . c. wetterich , phys .
b * 301 * , 90 ( 1993 ) .
morris , int .
j. mod .
phys . a * 9 * , 2411 ( 1994 ) .
m. salmhofer , _ renormalization _ ( springer , berlin , 1998 ) . as an even simpler `` toy model '' we have studied the classical anharmonic oscillator . besides the one - particle irreducible version for this model we have also investigated the frg schemes using amputated connected green functions @xcite and wick ordered amputated connected green functions @xcite . comparing results within the same order of truncation the irreducible scheme always led to the best results .
d. zanchi and h.j .
schulz , phys .
b * 61 * , 13609 ( 2000 ) .
halboth and w. metzner , phys .
b * 61 * , 7364 ( 2000 ) . c. honerkamp , m. salmhofer , n. furukawa , and t.m . rice , phys .
b * 63 * , 035109 ( 2001 ) .
t. busche , l. bartosch , and p. kopietz , j. phys . :
matter * 14 * , 8513 ( 2002 ) .
v. meden , w. metzner , u. schollwck , and k. schnhammer , phys .
b * 65 * , 045318 ( 2002 ) ; j. low temp .
* 126 * , 1147 ( 2002 ) ; s. andergassen , t. enss , v. meden , w. metzner , u. schollwck , and k. schnhammer , cond - mat/0403517 .
certain aspects of the frequency dependence were taken into account in ref .
@xcite and : c. honerkamp and m. salmhofer , phys . rev .
* b * 67 , 174504 ( 2003 ) .
e.g. see c.m .
bender and t.t .
wu , phys .
rev . * 184 * , 1231 ( 1969 ) ; c.m .
bender and t.t .
wu , phys .
* 27 * , 461 ( 1971 ) ; w. janke and h. kleinert , _ ibid .
_ * 75 * , 2787 ( 1995 ) ; e.j .
weniger , _ ibid . _ * 77 * , 2859 ( 1996 ) ; c.m .
bender and l.m.a .
bettencourt , _ ibid . _ * 77 * , 4114 ( 1996 ) ; t. hatsuda , t. kunihiro , and t. tanaka , _ ibid . _
* 78 * , 3229 ( 1997 ) ; y. meurice , _ ibid .
_ * 88 * , 141601 ( 2002 ) and references therein .
the problem has also been studied using wegner s flow equation technique ; g. uhrig , private communication . j.w .
negele and h. orland , _ quantum many - particle physics _ ( addison - wesley , 1988 ) .
katanin , cond - mat/0402602 .
for symmetry reasons the @xmath142 with even @xmath298 vanish . for the @xmath118 considered here the exact spectral weight
@xmath299 is roughly three orders of magnitude smaller than @xmath145 . thus extracting more than @xmath200 and
@xmath300 is beyond the accuracy of our numerical treatment of the flow equations .
this discretization is not equivalent to considering finite temperatures .
although the frg method is set up for arbitrary @xmath301 ( see sect . [
sec : frg ] ) in most applications calculations were performed at @xmath119 . using a smooth frequency cutoff
frg results for finite @xmath301 for the problem of resonant tunneling in a luttinger liquid are presented in : v. meden , t. enss , s. andergassen , w. metzner , and k. schnhammer , cond - mat/0403655 .
also for the classical anharmonic oscillator compared to the conventional scheme the modified frg @xcite leads to better agreement with the exact result .
anderson , phys .
124 * , 41 ( 1961 ) .
r. schrieffer and p.a .
wolff , phys .
rev . * 149 * , 491 ( 1966 ) .
j. kondo , prog .
32 * , 37 ( 1964 ) .
a.c . hewson in d. edwards and d. melville , eds .
, _ the kondo problem to heavy fermions _
( cambridge university press , 1993 ) . h.r .
krishnamurthy , j.w .
wilkins and k.g .
wilson , phys .
b*21 * , 1003 ( 1980 ) .
press _ et al .
, _ _ numerical recipes in c _ ( cambridge university press , 1993 ) .
d. logan and m. glossop , j. phys . : condens .
matter * 12 * , 985 ( 2000 ) .
pruschke , m. jarrell , and j.k .
freericks , adv . in phys .
* 44 * , 187 ( 1995 ) ; a. georges , g. kotliar , w. krauth , and m.j .
rozenberg , rev .
phys . * 68 * , 13 ( 1996 ) .
maier , m. jarrell , th .
pruschke , and m. hettler , cond - mat/0404055 ( 2004 ) .
j. w. allen and r. m. martin , phys .
lett . * 49 * , 1106 , ( 1982 ) ; j. w. allen and l. z. liu , phys .
b * 46 * , 5047 , ( 1992 ) ; m. lavagna _
_ , phys . lett . *
90a * , 210 ( 1982 ) .
smith and q. si , phys .
b * 61 * , 5184 ( 2000 ) . | we apply the functional renormalization group method to the calculation of dynamical properties of zero - dimensional interacting quantum systems . as case studies we discuss the anharmonic oscillator and the single impurity anderson model .
we truncate the hierarchy of flow equations such that the results are at least correct up to second order perturbation theory in the coupling .
for the anharmonic oscillator energies and spectra obtained within two different functional renormalization group schemes are compared to numerically exact results , perturbation theory , and the mean field approximation . even at large coupling
the results obtained using the functional renormalization group agree quite well with the numerical exact solution .
the better of the two schemes is used to calculate spectra of the single impurity anderson model , which then are compared to the results of perturbation theory and the numerical renormalization group . for small to intermediate couplings
the functional renormalization group gives results which are close to the ones obtained using the very accurate numerical renormalization group method .
in particulare the low - energy scale ( kondo temperature ) extracted from the functional renormalization group results shows the expected behavior . |
the main purpose of this paper is to deepen the study of qualitative theory of differential equations induced by differentiable vector fields of the plane which are not necessarily of class @xmath0 .
for instance , problems addressed in @xcite include global injectivity , uniqueness of solutions and asymptotic stability of the point at infinity of the riemann sphere @xmath1 .
these works , like many others in the literature , extend to the differentiable case previous results concerning @xmath2vector fields .
thus , the present paper pretends to extend to the differentiable case the work in @xcite .
the authors of @xcite guarantee the change of stability at infinity of a one parameter family of @xmath2vector fields as the parameter goes through zero .
this behaviour was presented as hopf bifurcation at infinity .
before @xcite the study of hopf bifurcation at infinity had been approached only from two points of view .
one deals with families which have finitely many parameters and are made up of continuous perturbations of linear systems .
the other point of view deals with polynomial families .
the former case uses the strong domination imposed by the linear part of the vector field on the continuous nonlinear term , which can have sub linear growth @xcite .
the later point of view involves standard poincar compactification @xcite .
both points of view are related to the existence of limit cycles of arbitrarily large amplitudes .
particularly , the case of polynomial systems can be associated to the sixteenth hilbert problem , see for instance @xcite .
in contrast , a different point of view was given in @xcite , where the authors are only interested in the change of stability at infinity .
this point of view has been strongly articulated in other works related to planar systems induced by differentiable vector fields @xcite .
the main motivation of this approach was the definition of hopf bifurcation at infinity given in @xcite ( see also @xcite ) .
concretely , it is said in @xcite that the polynomial family of vector fields @xmath3 presents a hopf bifurcation at infinity when @xmath4 cross @xmath5 if the infinity changes its stability . the paper @xcite deals with a class of one parameter family @xmath6 of planar @xmath0 vector fields defined on the complement of an open ball centered at origin which are also free of singularities .
the authors find the change of stability at infinity when the sign of @xmath7 varies as the parameter crosses zero . in the context of @xcite
, @xmath8 is any @xmath2 extension of @xmath9 with divergence lebesgue almost
integrable ( see section [ sec_prelim ] for details ) .
the families considered in @xcite are neither polynomial nor a perturbation of linear systems , consequently the work in @xcite contains the two point of view described above .
we consider a one parameter family @xmath10 , where @xmath11 has isolated singularities and belongs to a class of differentiable ( not necessarily of class @xmath0 ) planar vector fields defined on the complement of some compact ball centered at the origin .
the main motivation to address the existence of hopf bifurcation at infinity for this family is the work given in @xcite and @xcite . in those papers the authors prove the uniqueness of the positive trajectory starting at every point @xmath12 due some conditions on the eigenvalues of the jacobian matrix of the differentiable vector field .
in addition , results in @xcite allow us to focus on the integral @xmath13 in order to find a change of stability at infinity . in this context , since the qualitative change at infinity of the family @xmath10 may be exhibit from global extensions @xmath14 , the main obstruction to describe the bifurcation is the existence of unbounded sequences of singular points .
in order to avoid this phenomenon , we impose some extra condition on the extension @xmath15 which imply that every vector field of the family induces an injective map .
consequently , every vector field of the family has at most one singular point in the plane .
observe that the one
parameter family considered in this paper is not necessarily a perturbation of linear systems in the sense of @xcite .
moreover , complements the results presented in @xcite for @xmath2vector fields because in our work the existence of isolated singularities of @xmath11 is allowed .
let @xmath16 be the compact ball bounded by @xmath17 with @xmath18 , and suppose that @xmath19 is a differentiable vector field , defined on the complement of @xmath20 on @xmath21 . here , the term differentiable means _
frechet differentiable _ at each point @xmath22 and @xmath23 is the respective derivative .
this derivative is defined as the bounded linear operator induced by the standard jacobian matrix at @xmath24 .
this matrix is denoted by @xmath25 , so @xmath26 and @xmath27 are the typical jacobian determinant and divergence , respectively . on the other hand , if the point @xmath28 is kept fixed , then a _ trajectory _ of @xmath11 starting at @xmath29 is defined as the integral curve @xmath30 , determined by a maximal solution of the initial value problem @xmath31 .
of course , it means that @xmath32 and @xmath33 . in this context , it is useful to have a term that refers to the image of the solution .
hence , we define the _ orbit _ of @xmath34 to be the set @xmath35 , and we identify the trajectory with its orbit . similarly , @xmath36 and @xmath37\cap i_q\}$ ] are the _ semi trajectories _ of @xmath11 , and
they are called _ positive _ and _ negative _ , respectively . in consequence , @xmath38 , and each trajectory has its two limit sets , @xmath39 and @xmath40 .
these limit sets are well defined in the sense that they only depend on the respective solution . and
lastly , such a vector field induces a well defined _ positive semi flow _
( respectively _ negative semi flow _ ) , if the condition @xmath28 implies the existence and uniqueness of @xmath41 ( respectively @xmath42 ) .
the trajectories of a vector field may be unbounded .
one way to obtain some information about the behavior of such solutions is to compactify the plane , so that the vector field is extended to a new manifold that contains the points at infinity . in this context
, the so called alexandroff compactification has been most successful in the study of planar systems induced by @xmath0-vector fields , not necessarily polynomial ( see for example @xcite ) .
to describe the results , @xmath21 is embedded in the riemann sphere @xmath43 .
consequently , @xmath44 is the subspace of @xmath43 with the induced topology , and infinity refers to the point @xmath45 of @xmath43 .
moreover , a vector field @xmath46 ( without singularities ) can be extended to a map @xmath47 ( which takes @xmath45 to @xmath5 ) . in this manner ,
all questions concerning the local theory of isolated singularities of planar vector fields can be formulated and examined in the case of the extended vector field @xmath48 , which coincides with @xmath11 on @xmath49 .
for instance , if @xmath50 is an unbounded semi trajectory of @xmath51 with empty @xmath52limit set , then we declare that @xmath53 _ goes to infinity _ , and we write @xmath54 .
similarly , @xmath55 denotes that @xmath56 _ comes from infinity _ , and it means that @xmath57 is an unbounded semi trajectory whose @xmath58limit set is empty . therefore , it is also possible to talk about the phase portrait of @xmath11 in a neighborhood of @xmath45 , as shown @xcite . throughout this paper , given @xmath59 , a closed ( compact , no boundary ) curve ( @xmath60manifold ) , @xmath61 ( respectively @xmath62 ) denotes the compact disc ( respectively open disc ) bounded by @xmath63 .
thus , the boundaries @xmath64 and @xmath65 are equal to @xmath63 besides homeomorphic to the circle @xmath66 .
the index at infinity was defined firstly for @xmath2vector fields in @xcite and lately in @xcite for differentiable vector fields ( not necessarily of class @xmath0 ) .
the relationship between this index and the stability of the point at infinity of the vector field is also stated in those works in both cases .
consider @xmath51 be a differentiable vector field .
the _ index of @xmath11 at infinity _ , denoted by @xmath67 , is the number of the extended line @xmath68 $ ] given by @xmath69 where @xmath70 is a global differentiable vector field such that : * in some @xmath71 with @xmath72 , both @xmath11 and @xmath73 coincide . * @xmath74 is lebesgue almost
integrable on @xmath75 in the sense of @xcite .
this index is a well defined number in @xmath68 $ ] , and it does not depend on the pair @xmath76 as shown in @xcite ( see also @xcite for the @xmath0 case ) . the definition of hopf bifurcation at infinity for differentiable vector fields presented in this paper is based on the definition of hopf bifurcation at infinity for @xmath0 vector fields given in @xcite and the definition of attractor / repellor given in @xcite for differentiable vector fields . in @xcite ,
the authors relate the index at infinity of a differentiable vector field with the stability of the point at @xmath45 of the riemann sphere .
[ def : atractor repellor ] the point at infinity @xmath45 of the riemann sphere @xmath43 is an _ attractor _ ( respectively , a _ repellor _ ) for @xmath51 if : * there is a sequence of closed curves , transversal to @xmath11 and tending to infinity .
it means that for every @xmath77 there exists a closed curve @xmath78 such that @xmath79 contains @xmath80 and @xmath78 has transversal contact to each small local integral curve of @xmath11 at any @xmath81 .
* for some @xmath82 with @xmath83 all the trajectories @xmath84 starting at a point @xmath85 satisfy @xmath86 , that is @xmath87 goes to infinity ( respectively , @xmath88 , that is @xmath56 comes from infinity ) .
notice that in the @xmath0 case , is equivalent to saying that the vector field @xmath48 induced by @xmath11 on the riemann sphere is locally topologically equivalent in an open neighborhood of the infinity to @xmath89 ( respectively , @xmath90 ) at the origin @xcite .
the authors of @xcite relate the index at infinity of a differentiable vector field with the stability of the point at @xmath45 of the riemann sphere .
[ def : hopf]we will say that the family of differentiable vector fields @xmath91 has at @xmath92 a _ hopf bifurcation at @xmath45 _ if the following two conditions are satisfied : * for @xmath93 ( resp .
@xmath94 ) , the vector field @xmath95 has a repellor at @xmath45 , and for @xmath94 ( resp .
@xmath93 ) , the vector field @xmath95 has an attractor at @xmath45 . *
the vector field @xmath95 has no singularities in @xmath96 , for @xmath97 and some @xmath72 . in order to capture the essential features of
, we remark that the element @xmath98 might be unstable in the sense that the vector field @xmath99 induced by @xmath98 on the riemann sphere might be locally topologically equivalent in an open neighborhood of the infinity to a center at the origin . in this context
, @xmath98 has first degree of instability @xcite in the sense that it is unstable whereas any vector field in @xmath100 , with @xmath4 sufficiently close to @xmath5 , is either stable ( i.e @xmath101 ) or topologically equivalent to @xmath98 ( see figure [ fig : mainf proposition ] ) . in section [ sec : free eigen ] , we also impose some condition on @xmath98 in order to force a specific family @xmath100 to present a hopf bifurcation at @xmath45 as a topological characterization of the bifurcation when @xmath45 reverses its stability .
the authors of @xcite relate the existence of a _
hopf bifurcation at infinity _ for a family of @xmath2vector fields with the _ index at infinity _ of each vector field of the family .
[ teoagg07 ] let @xmath102 be a family of @xmath2vector fields such that : 1 . @xmath103 is lebesgue almost
integrable on @xmath104 , 2 .
@xmath105 is well defined , this index belongs to @xmath68 $ ] , and @xmath106 where @xmath107 , 3 .
@xmath108 implies that the product @xmath109 , 4 .
@xmath110 has no singularities , and the _ poincar index _ at @xmath45 of the vector field @xmath111 ( which extends @xmath110 at @xmath45 ) is less than or equal to @xmath112 .
+ then @xmath113 is a hopf bifurcation at @xmath114 of the family @xmath115 .
is the one we extend to the differentiable case in .
a differentiable planar vector field @xmath116 is called _ dissipative _ on a region @xmath117 if @xmath118 on @xmath117 and the equality only holds on a set with lebesgue measure zero .
see @xcite for more details .
let @xmath119 be the set of differentiable vector fields @xmath120 which are dissipative and have strictly positive jacobian determinant .
this means that all vector field @xmath11 in @xmath119 verifies @xmath121 where @xmath122 is the set of eigenvalues of @xmath25 , for all @xmath123 . in this section
we study the case of dissipative vector fields in @xmath119 .
this is complemented in the next section , where we focus on vector fields that are free of real eigenvalues .
observe that the divergence of every vector field verifying is lebesgue almost
integrable on @xmath124 , in the sense of @xcite .
recall that , a map @xmath125 is a local homeomorphism ( respectively a local diffeomorphism ) if for every @xmath126 there exist open neighborhoods @xmath127 of @xmath24 y @xmath128 of @xmath129 such that the mapping @xmath130 is a continuous ( resp .
differentiable ) bijection whose inverse is also continuous ( resp .
differentiable ) .
consider the family of vector fields @xmath131 where @xmath132 and @xmath11 has some singularity in @xmath124 . in this section
, we also prove the existence of globally injective extensions of @xmath133 to the whole plane whose divergence is lebesgue almost integrable in @xmath21 .
consider @xmath134 . here
, @xmath135 with @xmath136 denotes the continuous foliation given by the level sets @xmath137 and then the leaves of the foliations are differentiable .
the concept of half - reeb component is a natural generalization of the description of the planar foliations given in @xcite .
more precisely , @xmath138 is a _ half reeb component _ of @xmath135 if there is a homeomorphism @xmath139\times[0,2]:0<x+y\leq 2 \big\}\to\mathcal{a},\ ] ] which is a topological equivalence between @xmath140 and @xmath141 with @xmath142 such that : * the segment @xmath143 is sent by @xmath144 onto a transversal section for the foliation @xmath135 in the complement of the point @xmath145 . *
both segments @xmath146 and @xmath147 are sent by @xmath144 onto full half leaves of @xmath135 .
[ lem : existence of the index ] if @xmath148 has some singularity , the following hold : * @xmath149 and a globally injective local homeomorphism @xmath150 such that @xmath151 and @xmath152 and @xmath11 coincide on @xmath153 ; * the index @xmath67 at infinity is a well defined number in @xmath154 . observe that every @xmath155 satisfies @xmath156 and @xmath157 .
thus , the methods of ( * ? ? ?
* proposition 2 ) lead us to obtain that : * any half
reeb component of either @xmath158 or @xmath159 is bounded . under conditions ( a.1 ) and @xmath156 ,
the means developed in the last section of @xcite ( see proposition 5.1 ) can be applied .
consequently , the proof of ( * ? ? ?
* theorem 3 ) gives the existence of a closed curve @xmath63 , surrounding the singularity of @xmath11 and the point at the origin , so that @xmath63 is embedded in the plane , and it admits an exterior collar neighborhood @xmath160 such that : * @xmath161 is a non trivial closed curve surrounding the origin , @xmath162 is an exterior collar neighborhood of @xmath161 and the restriction @xmath163 is a homeomorphism .
* the foliation @xmath158 , restricted to @xmath164 is topologically equivalent to the foliation made up by all the vertical straight lines on @xmath165 . by schoenflies theorem , @xcite , the map @xmath166 can be extended to a homeomorphism @xmath167 with @xmath168 .
we also extend @xmath169 to @xmath170 by defining @xmath171 thus , @xmath172 is a homeomorphism , and @xmath173 ( resp .
@xmath162 ) is a exterior collar neighborhood of @xmath63 ( resp .
@xmath161 ) . furthermore ,
( a.3 ) implies that @xmath152 is a local homeomorphism whose foliation @xmath174 is trivial .
therefore , @xmath175 is injective by ( * ? ? ?
* proposition 1.4 ) . observe that @xmath176 by construction .
hence @xmath177 holds considering @xmath178 with @xmath179 . under these conditions , @xcite gives the existence of some @xmath180 such that the restriction @xmath181 admits a global differentiable extension @xmath73 which divergence is lebesgue almost
integrable on @xmath21 and @xmath182 .
consequently , the index of @xmath11 at infinity @xmath67 is a well defined number of the extended real line @xmath183 .
this concludes the proof .
[ cor:3.2 ] if @xmath148 has some singularity and @xmath184 , then @xmath185 such that : * the restriction @xmath186 satisfies the statement of ; * the vector field @xmath95 has a well defined index at infinity and @xmath187 observe that a direct computation of the eigenvalues in @xmath188 shows that @xmath189 thus , as a consequence of , we can obtain that : @xmath190 therefore , for every @xmath191 , ( a ) follows by . item ( b ) holds by and definition of index at infinity of @xmath95 , provided by @xmath192 [ cor:3.2bis ] let @xmath148 with some singularity .
suppose in addition that @xmath193 then , @xmath194 such that for every @xmath195 $ ] there exists @xmath196 for which the restriction @xmath186 satisfies the statement of lemma [ lem : existence of the index ] .
by hypothesis , @xmath197 for all @xmath198 , then implies that @xmath199.\ ] ] consider @xmath200 .
thus , proposition 2.7 in @xcite implies that , for all @xmath201 $ ] , @xmath202 has the following property : any half
reeb component of either @xmath203 or @xmath204 is bounded .
therefore , the proof of lemma [ lem : existence of the index ] holds verbatim .
[ rem : main ] under conditions of corollary [ cor:3.2bis ] , the set @xmath205\big\}$ ] is a well defined set . in addition , this set is bounded when @xmath206 is continuous .
when the extension given by lemma [ lem : existence of the index ] , corollary [ cor:3.2 ] and corollary [ cor:3.2bis ] is also a local diffeomorphism , we say that the vector field @xmath207 is a _
strong extension _ of the vector field @xmath208 . in this section
we study the stability at infinity of the family of global vector fields @xmath209 where @xmath210 is a vector field such that @xmath211 and the restriction @xmath212 belongs to @xmath119 .
[ prop : without sing ] if there exists @xmath213 such that the map @xmath214 is a local diffeomorphism , for all @xmath215 .
then , @xmath216 such that the restriction @xmath217 has no singularities in @xmath218 , for every @xmath201 $ ] . as a consequence of , the following assumptions are equivalent * @xmath219),}$ ] * @xmath220=\emptyset \quad ( \text{resp .
} \quad \mbox{spc}(y)\cap[0,\varepsilon_0)=\emptyset).}$ ] since the restriction @xmath212 belongs to @xmath119 , condition implies that @xmath221.\ ] ] hence , by ( * ? ? ?
* theorem 2.1 ) , the following holds : @xmath222,\ ] ] ( see also ( * ? ? ? * theorem 3 ) and @xcite ) . consequently , since @xmath223 , implies that there exists @xmath224 such that the restriction @xmath217 has no singularities in @xmath218 , for every @xmath201 $ ] .
thus , the proposition holds .
[ prop : main global]let @xmath225 be as in proposition [ prop : without sing ] .
suppose in addition that @xmath226 such that , for each @xmath227 , * @xmath228 induces a well defined negative semi
flow on @xmath71 , * the index @xmath229 satisfies @xmath230 .
then , for each @xmath201 $ ] , @xmath45 is an attractor ( respectively a repellor ) for the vector field @xmath231 provided by @xmath232 ( respectively @xmath233 ) . observe that , for all @xmath234 @xmath235\ ] ] hence , for all @xmath236 the following hold : * @xmath237 and so @xmath238 is lebesgue almost integrable in @xmath239(@xcite ) ; * @xmath240 is a well - defined number in @xmath154 ( @xcite ) . * for every @xmath241 , there is only one positive semi
trajectory of @xmath228 passing through @xmath24 .
( @xcite ) . therefore , @xmath240 is a well defined number and @xmath242 , for all @xmath243 $ ] .
* we claim that for every @xmath201 $ ] , the condition @xmath244 implies that @xmath228 has a bounded set of periodic trajectories .
since @xmath223 and @xmath228 is injective by , we proceed by contradiction : * we suppose that @xmath245 is an unbounded set of periodic trajectories of @xmath228 such that @xmath246 under these conditions , by using the green theorem in @xmath247 and the arc length element @xmath248 we obtain that @xmath249 where @xmath250 is the unitary outer normal vector to @xmath251 and @xmath252 is the inner product of @xmath253 with @xmath254 .
thus , @xmath255 provided by @xmath256 , for all @xmath254 .
this contradiction with @xmath244 shows that ( * ) never happens , and gives the proof of ( a.4 ) .
furthermore , since @xmath257 $ ] is a compact set , the following is directly obtained applying ( a.4 ) : * @xmath258 such that @xmath228 has no periodic trajectories in @xmath71 as long as @xmath259 and @xmath244 .
hence , @xmath260 generates a positive semi flow on @xmath71 , for all @xmath261 .
in addition , for @xmath262 as in ( a.5 ) we have that for all @xmath259 the following hold : * @xmath263 ; * @xmath264 , @xmath265 , so @xmath260 preserves orientation ; * there exists @xmath266 such that @xmath267 , for any @xmath268 , by the openness of @xmath260 . under these properties
: and ( a.1 ) to ( a.8 ) , the vector field @xmath228 satisfies all the conditions of ( * ? ?
* theorem 26 ) . consequently : * for every @xmath269 there exist a closed curve @xmath78 transversal to @xmath228 contained in the regular set @xmath270 .
in particular , @xmath79 contains @xmath80 and @xmath78 has transversal contact to each small local integral curve of @xmath271 at any @xmath81 .
moreover , for @xmath269 large enough the closed curve @xmath272 is transversal to @xmath228 , near @xmath45 and it is such that @xmath259 , the index @xmath240 and @xmath273 have the same sign . thus , the point at infinity of the riemann sphere @xmath43 is either an attractor or a repellor of @xmath274 .
actually , by ( a.9 ) and ( * ? ? ?
* theorem 28 ) we obtain that : 1 .
if @xmath275 ( respectively @xmath276 ) , then @xmath45 is a repellor ( respectively an attractor ) of the vector field @xmath274 .
it concludes the proof , and the proposition holds .
we are now ready to state our main theorem .
[ main 1]consider the family @xmath277 where @xmath11 is a differentiable vector field in @xmath119 with some singularity and verifying @xmath278 suppose in addition that 1 .
@xmath279 such that the set @xmath280\big\}$ ] induced by strong extensions is bounded ; 2 . for each @xmath94 , @xmath95 induces a well defined negative semi flow and the index @xmath281
$ ] satisfies @xmath282 . then there are @xmath283 and @xmath284 such that the family of vector fields @xmath285 has at @xmath92 a hopf bifurcation at @xmath45 . by the assumption ( 1 ) there
exist a number @xmath286 and a local diffeomorphism @xmath287 with @xmath288 such that @xmath289 in addition , the map @xmath290 such that @xmath291 is also a local diffeomorphism . therefore
, there exist @xmath292 and @xmath293 such that for every @xmath201 $ ] : * the restriction @xmath294 has no singularities in @xmath96 . by proposition [ prop : without sing ] .
* @xmath295 , for all @xmath93 . by corollary [ cor:3.2 ] . * for @xmath93 ,
the restriction @xmath296 has a repellor at @xmath45 , and for @xmath94 , the restriction @xmath296 has an attractor at @xmath45 . by hypotheses ( 2 ) and proposition [ prop : main global ] .
thus , @xmath297 has at @xmath92 a hopf bifurcation at @xmath45 . observe that the vector fields @xmath188 as in satisfy @xmath298 therefore , if @xmath299 is the space of the continuous vector fields of @xmath124 endowed with the topology induced by the uniform convergence on compact sets , the functional @xmath300 is continuous .
[ rem : discontinuity ] if the maps @xmath301 and @xmath302 are differentiable and the family @xmath297 is given by , then the families @xmath303 also have at @xmath92 a hopf bifurcation at @xmath45 . for instance , the map @xmath304 notice that @xmath305 induces a well defined map @xmath306 which is discontinuous at @xmath92 .
in some sense , shows that includes the results where the strong domination imposed by the linear part is used .
furthermore , we do not assume the existence of some open neighborhood of infinity free of singularities , as in theorem [ teoagg07 ] item ( 4 ) .
recall that the results in this work are obtained from the weak hypothesis on the jacobian determinant like , which is natural in order to work with isolated singularities .
in addition , the assumption ( 2 ) in theorem [ main 1 ] is necessary , since the theorem considers the differentiable vector fields , not necessarily of class @xmath0 . in the particular case of @xmath2vector fields , theorem [
main 1 ] directly gives the next illustrative result .
[ cor:1]consider @xmath307 of class @xmath0 .
suppose @xmath308 has a singularity and the maps @xmath309 are orientation preserving local diffeomorphisms , for all @xmath4 in some interval @xmath310 .
if assumption ( 1 ) in theorem [ main 1 ] holds , then the condition @xmath311 implies the existence of @xmath283 and @xmath284 such that @xmath312 has at @xmath92 a hopf bifurcation at @xmath45 . observe that condition extends the case where the spectrum of @xmath95 crosses the imaginary axis transversally when @xmath4 moves from negative to positive values .
this subsection addresses the special case of differentiable vector fields whose jacobian matrix is free of real eigenvalues .
these vector fields induce local diffeomorphisms with zero divergence .
[ cor:3.3 ] let @xmath120 be a differentiable vector field with some singularity .
suppose that the following holds @xmath313 then , @xmath314 @xmath315 such that * the restriction @xmath316 is injective and admits a global differentiable extension @xmath317 with @xmath182 and @xmath318 * the index @xmath67 at infinity is a well defined number in @xmath154 ; * @xmath95 induces a well defined positive ( respectively negative ) semi flow as long as @xmath93 ( respectively @xmath94 ) . moreover , the trajectories of @xmath11 are unique in the sense that only depend of the initial condition . since @xmath188 implies that @xmath319 for all @xmath320 .
proposition 2.7 in @xcite , implies that @xmath202 has the following property : * any half - reeb component of either @xmath203 or @xmath204 is a bounded set .
therefore , the proof of lemma [ lem : existence of the index ] holds verbatim and statements ( a ) and ( b ) remain true .
in addition , implies that : @xmath321 therefore , ( * ? ? ?
* lemma 3.3 ) implies that the vector fields @xmath95 with @xmath93 , and @xmath322 with @xmath94 , induce a well defined positive semi flow .
similarly , ( * ? ? ?
* lemma 2.1 ) gives the uniqueness of the trajectories induced by @xmath11 .
thus , statement ( c ) holds .
[ prop : div 0 ] consider the family of differentiable vector fields @xmath323 , where @xmath11 verifies hypotheses of proposition [ cor:3.3 ] .
if @xmath279 such that the set @xmath280\big\}$ ] induced by strong extensions is bounded , then @xmath226 and @xmath283 such that the family @xmath324 has at @xmath92 a hopf bifurcation at @xmath45 . follows by proposition [ cor:3.3 ] and proposition [ prop : main global ] .
observe that a similar result as described in remark [ rem : discontinuity ] is obtained in the case of families given by .
+ [ rem : center ] consider now the special case of planar @xmath2vector fields @xmath325 such that @xmath211 and @xmath326 the trajectories in @xmath327 induced by @xmath271 are periodic orbits surrounding the origin @xcite .
moreover , a direct application of gives the existence of some @xmath283 , for which the family @xmath328 has at @xmath92 a hopf bifurcation at @xmath45 .
see figure [ fig : mainf proposition ] .
the first author was partially supported by grants micinn-12-mtm2011 - 22956 from spain and cnpq 474406/2013 - 0 from brazil .
the second author was partially supported by pontifcia universidad catlica del per ( dgi:70242.0056 ) , by instituto de cincias matemticas e de computao ( icmc usp : 2013/16226 - 8 ) .
this paper was written while the second author served as an associate fellow at the abdus salam ictp in italy .
he also acknowledges the hospitality of icmc usp in brazil during the preparation of part of this work .
a. a. andronov , e. a. leontovich , i. i. gordon , and a. g. maer , _ theory of bifurcations of dynamic systems on a plane _ , halsted press [ a division of john wiley & sons ] , new york - toronto , 1973 , translated from the russian .
p. diamond , d. rachinskii , and m. yumagulov , _ stability of large cycles in a nonsmooth problem with hopf bifurcation at infinity _ , nonlinear anal .
* 42 * ( 2000 ) , no . 6 , ser . a : theory methods , 10171031 .
a. gasull and j. sotomayor , _ on the basin of attraction of dissipative planar vector fields _ , bifurcations of planar vector fields ( luminy , 1989 ) , lecture notes in math . ,
1455 , springer , berlin , 1990 , pp .
187195 .
v. guez , e. sez , and i. sznt , _ simultaneous hopf bifurcations at the origin and infinity for cubic systems _ , dynamical systems ( santiago , 1990 ) , pitman res .
notes math .
285 , longman sci .
tech . , harlow , 1993 , pp .
4051 .
j. sotomayor and r. paterlini , _ bifurcations of polynomial vector fields in the plane _
, oscillations , bifurcation and chaos ( toronto , ont . , 1986 ) ,
, vol . 8 , amer . math . soc . , providence , ri , 1987 , pp . | this work deals with one parameter families of differentiable ( not necessarily @xmath0 ) planar vector fields for which the infinity reverses its stability as the parameter goes through zero .
these vector fields are defined on the complement of some compact ball centered at the origin and have isolated singularities .
they may be considered as linear perturbations at infinity of a vector field with some spectral property , for instance , dissipativity .
we also address the case concerning linear perturbations of planar systems with a global period annulus .
it is worth noting that the adopted approach is not restricted to consider vector fields which are strongly dominated by the linear part .
moreover , the poincar compactification is not applied in this paper . |
the two hubble deep fields ( hdf ) provide our deepest optical insight into the near and far universe .
each contains approximately 750 galaxies for which magnitudes , colours , photometric redshifts and morphologies have been determined @xcite .
the data have been used by many groups to construct a picture of the assembly and evolution of galaxies .
however every dataset , the hdfs included , suffer from selection biases at some level . here
i hope to briefly highlight the severity of these selection biases and to provide some suggestions as to how to test for , and monitor , these biases .
in particular i advocate the use of the bivariate brightness distribution ( bbd ) as a means of probing galaxy evolution while also tracking the biases which plague our ability to decipher deep data .
finally i put forward my own cautious conclusions as to the evolution of ellipticals , spirals and irregulars since z=1 .
galaxies exist over a vast dynamic range in both luminosity and in surface brightness .
any dataset has an apparent magnitude limit , a surface brightness detection limit and size limits which together create a selection window @xcite .
we view of the galaxy population through this selection window different surveys with differing limitations will sample different regions of the galaxian parameter - space .
unless these biases are monitored discrepancies will manifest between surveys as are seen for example in measures of the local luminosity function @xcite .
these concerns become more paramount when one compares local and deep data , as such biases will cause discrepancies that may erroneously be interpreted as evolution .
additionally as these selection biases are all distance dependent the selection window varies such that comparisons within a single dataset are also susceptible .
the most important of these biases is malmquist bias - i.e. , as we go fainter we select against firstly , low luminosity systems , then normal systems and eventually giant systems . an analogous bias , disney bias @xcite , works in a similar way but in terms of a galaxy s absolute surface brightness such that lower surface brightness systems are preferentially lost first , etc . neglecting either malmquist or disney bias leads to misleading results and as both biases
are ever present it is vital that we , as a community , demonstrate the robustness of our data to these biases prior to interpretation .
[ fig1 ] shows two visualisations of the hdf data highlighting the severity of both the malmquist bias ( left ) and the disney bias ( right ) .
the dashed lines shows the location of the canonical non - evolving zero - redshift @xmath2 point .
v @xmath3 , left ) and ( @xmath4 v @xmath3 , right ) highlight the severe effect of malmquist- and disney- bias which impinge define the selection window through which we view the galaxy population at any redshift .
the two selection limits define the uncertainty due to the range of possible k - corrections .
solid points are spectrscopic redshifts , circles are ellipticals , squares spirals and triangles irregulars , title="fig:",scaledwidth=60.0% ] v @xmath3 , left ) and ( @xmath4 v @xmath3 , right ) highlight the severe effect of malmquist- and disney- bias which impinge define the selection window through which we view the galaxy population at any redshift .
the two selection limits define the uncertainty due to the range of possible k - corrections .
solid points are spectrscopic redshifts , circles are ellipticals , squares spirals and triangles irregulars , title="fig:",scaledwidth=60.0% ] the conclusion we must draw from fig . [ fig1 ] is that the hdf suffers from severe bias beyond z=1 ( for the spectroscopic sample : solid points ) and beyond z=1.25 ( for the photometric sample ) .
any statistical measurement of luminosity functions , star - formation history , evolution etc , will therefore be incomplete and increasingly misleading beyond z=1.25 .
so , does the mean star - formation of the universe really peak at z=1.5 and then turn down ?
[ fig1 ] convinces me that we simply do nt know until we look into the `` unobserved '' regions .
looking at fig .
1 it is also easy to see why one might be drawn to the conclusion that galaxies were both more luminous and of higher surface brightness in the past . again
an alternative and more informed conclusion is that this interpretation _ may _ be purely due to the shifting selection window - we can not say . *
the conclusion is that the hubble deep fields lack the breadth in luminosity and surface brightness beyond z=1.25 to place meaningful constraints on galaxy evolution . *
having concluded that a statistical investigation of galaxy evolution from the hdfs is currently impossible beyond z=1.25 this confines us to investigation galaxy evolution at @xmath5 .
this is not to say that selection biases are not present below @xmath5 but that they have not yet impinged upon the canonical z=0 @xmath2 point .
why is the @xmath2 point so critical ? mainly because it is these galaxies which dominate the luminosity and mass density of the local universe @xcite . to see where and whether they have to evolved we have to be able to `` observe '' any positive- , zero- and negative- evolution about this point .
additional selection lines also exist which can not be shown on fig .
these are the size constraints due to the pixel size ( limiting the highest measurable surface brightness ) and the field - of - view size ( limiting the lowest surface brightness object detectable for any absolute magnitude at some redshift ) . to show these additional selection lines we must simultaneously look at the @xmath6-@xmath4 plane ( our window into the galaxy population ) .
this plane is known as the bivariate brightness distribution and has up until no only be derived for the virgo cluster @xcite - nevertheless it is a powerful tool .
an example bbd , taken from @xcite , is shown as fig .
this shows the hdf window into the galaxy population at z=0.3 - 0.5 .
the absolute magnitude , surface brightness and size selection lines are all shown and the shaded regions indicate areas where the hdf is unable to detect galaxies .
the selection lines are all redshift dependent and control our insight into the galaxy population .
the selection lines are fairly straight - forward to estimate and are laid out in detail in @xcite .
note that those galaxies which appear in the excluded region _ must _ have erroneous photometric redshifts as their detection _ is _ impossible . as an aside the data show a strong luminosity - surface brightness relation and a dearth of luminous low surface brightness galaxies @xcite . * using a bivariate brightness distribution we can extract volume - limited samples for which the selection biases are well defined . *
having advocated a methodology to extract well controlled samples below @xmath7 , we are now in a good position to investigate galaxy evolution since @xmath8 . fig .
[ fig4 ] shows a sequence of bbds for all the hdf ellipticals , spirals and irregulars from z=0 to z=1 for @xmath9 .
the striped areas show the prohibited regions and galaxies that lie within these regions most likely have erroneous photometric redshifts .
the density of galaxies which lie in these excluded regions is of some concern and highlights the desperate need to spectroscopically confirm late - types in particular .
finally then what does fig .
[ fig4 ] tell us about the evolution of the three generic galaxy types .
the dashed line guides the eye and are fitted to the z=0.3 - 0.6 bin for each type .
two conclusions strike home : ( 1 ) the hdf contain insufficient data for a detailed statistical analysis particularly when one folds in the clustering uncertainty , and ( 2 ) , in all cases the 0.3 - 0.6 fit matches the 0.8 - 1.0 fit .
in closing it appears that the current outlook is somewhat depressing , the hdf is insufficiently deep to probe the evolution of the generic population beyond @xmath0 and the volume insufficient to probe evolution below @xmath0 .
we desperately need an order of magnitude more hdfs .
* for the moment the occam s razor solution must surely prevail : within the limited statistics available the hdf is consitent with null / passive evolution since @xmath8 , beyond this even the hdf data are dominated by selection effects . * | the hubble deep fields represent our best opportunity for probing galaxy evolution over a substantive look - back time . however as with any dataset the hdfs are prone to selection biases .
these biases are extremely severe beyond @xmath0 such that a meaningful interpretation of generic galaxy evolution is not possible .
we can however extract well defined volume - limited samples at @xmath1 .
the data are entirely consistent with passive / null - evolution for ellipticals , spirals and irregulars however this concluion is tempered by small number statistics .
alas stringent constraints on galaxy evolution await an order of magnitude increase in the number of hdfs . |
a theorem of shannon @xcite basic to all information theory describes the optimum compression of a discrete memoryless source , showing that the minimum achievable rate is the entropy of the source distribution .
the situation is the following : let @xmath0 be a probability distribution on the finite set @xmath1 .
we call @xmath2 an _ @xmath3code _ for the _ discrete memoryless source @xmath0 _ , if @xmath4 are stochastic maps , with a finite set @xmath5 , such that @xmath6 where @xmath7 denoting the minimal @xmath8 such that an @xmath3 code exists , by @xmath9 , shannon @xcite shows that for @xmath10 @xmath11 with the entropy @xmath12 of the distribution . motivated by the work @xcite , and by a construction in @xcite ( in footnote 4 ) , we study here the following modification of this problem : to each @xmath13 is associated a probability distribution @xmath14 on the finite set @xmath15 ( thus @xmath16 is a stochastic map , or channel , form @xmath1 to @xmath15 ) .
an @xmath3code is now a pair @xmath2 of stochastic maps @xmath17 ( compare with eq .
( [ eq : det : code ] ) ) , and instead of condition ( [ eq : det : condition ] ) we impose @xmath18 where @xmath19 is the @xmath20norm on function on @xmath21 : @xmath22 .
note that for two probability distributions @xmath0 and @xmath23 , @xmath24 equals their _ total variational distance _
@xmath25 of the two .
we define @xmath9 to be the minimal @xmath8 of an @xmath3code .
note that for @xmath26 , and @xmath14 the point
mass @xmath27 in @xmath28 , the new notion of @xmath3code coincides with the previous one . notice further , that we allow probabilistic choices in the encoding and decoding . while it is easy to see that this freedom does not help in shannon s problem , it is crucial for the more general form , that we will study in this paper .
the basic problem of course is to find the optimum rate @xmath29 of compression ( if the limit exists ; otherwise @xmath30 is to be considered ) , and especially the behaviour of this function at @xmath31 . for the case
@xmath32 , i.e. perfect restitution of the distributions @xmath14 , these definitions in principle make sense , but we do nt expect a neat theory to emerge .
instead we define @xmath33 the minimal entropy of the distribution on @xmath5 induced by the encoder @xmath34 ( with the idea that blocks of these @xmath35blocks we may data compress to this rate ) .
obviously @xmath36 , so the limit @xmath37 exists , and is equal to the infimum of the sequence . to evaluate this quantity is another problem we would like to solve .
the structure of this paper is as follows : first we find lower bounds ( section [ sec : lower ] ) , then discuss upper bounds , preferrably by constructing codes : in section [ sec : cr : trick ] we show how the lower bound is approached by using the additional resource of common randomness , in section [ sec : local : fid ] we prove achievability of it under a letterwise fidelity criterion as a consequence of this result , section [ sec : howards ] presents a constructions to upper bound @xmath38 and @xmath39 . in section [ sec : applications ] applications of the results and conjectures are presented : first , we make it plausible that the distillation procedure of @xcite is asymptotically reversible , second we show that shannon s coding theorem allows an `` inverse '' ( at least in situations where unlimited common randomness is around ) , third we give a simple proof that feedback does not increase the rate of a discrete memoryless channel , and fourth demonstrate , how shannon s rate distortion theorem follows as a corollary .
the compression result ( with or without common randomness ) thus reveals a great unifying power in classical information theory .
finally , in section [ sec : quantum ] we discuss extensions of our results to the case of a source of mixed quantum states : the present discussion fits into this models as probability distributions are just commuting mixed state density operators .
let us mention here the previous work on the problem : the major initiating works are @xcite and @xcite .
the latter introduced the distinction between blind and visible coding , and between the block and letterwise fidelity criterion .
in contrast to the pure state case the four possible combinations of these conditions seem to lead to rather different answers . the case of blind coding with either the letter or blockwise fidelity criterion was solved recently by koashi and imoto @xcite . otherwise in this paper
, we will only address the visible case .
an attempt on the letterwise fidelity case with either blind or visible encoding was made in @xcite .
however , an examination of the approach of this work shows that it does not fit into any of the the classes of fidelity criteria proposed by @xcite : for a code @xmath2 one could either apply the _ global _ criterion , which is essentially our eq .
( [ eq : condition ] ) , that is definitely not what is considered in @xcite , there being employed rate distortion theory .
or one could impose that the output @xmath40 is good on the average letterwise ( the _ local _ criterion of @xcite ) : @xmath41 \leq \lambda,\ ] ] where @xmath42 denotes the marginal distribution of @xmath43 on the @xmath44 factor in @xmath21 , and @xmath45 is any distance measure on probability distributions ( that we require only to be convex in the second variable ) . for @xmath46
this is implied by eq .
( [ eq : condition ] ) .
this , too , is not met in @xcite , as there @xmath34 and @xmath47 are constructed as deterministic maps , while to satisfy eq .
( [ eq : l : condition ] ) one needs at least a small amount of randomness . to achieve this
one could base the fidelity condition on looking at individual letter positions of _ source and output simultaneously _ : @xmath48\
! \leq \lambda.\ ] ] condition ( [ eq :
l : condition ] ) being weaker than ( [ eq : condition ] ) , this one is still weaker .
however , this , too , does not coincide with the criterion of @xcite : denoting by @xmath49 the joint distribution of @xmath28 and @xmath50 according to @xmath0 and @xmath16 , i.e. @xmath51 , one considers @xmath52 ( this is implied by eq . ( 1 ) of @xcite for @xmath53 , which in turn is implied by eq .
( [ eq : ll : condition ] ) for @xmath54 ) .
it is not at all clear how to connect this with any of the above : eq .
( [ eq : ks : condition ] ) is about the _ empirical joint distribution _ of letters in @xmath55 and @xmath43 ( assume for simplicity , as indeed the authors of @xcite do , that @xmath34 and @xmath47 are deterministic ) , that is about a distribution created by selecting a position @xmath56 randomly , while eqs .
( [ eq : condition ] ) to ( [ eq : ll : condition ] ) are about distributions created either by the coding process alone or in conjunction with the source .
our view is confirmed in an independent recent analysis of @xcite by soljanin @xcite , to the same effect .
an interesting new twist was added when in @xcite ( and later in a more extended way in @xcite and the recent @xcite ) the use of unlimited common randomness between the sender and receiver was allowed in the visible coding model with blockwise fidelity criterion . as already mentioned , we reproduce this result here in detail , with special attention to the resource of common randomness : we present a protocol for which we prove that it has minimum common randomness consumption in the class of protocols which even simulate full passive feedback of the received signal to the sender .
let the random variable @xmath57 be distributed according to @xmath58
. then we can define @xmath59 by @xmath60 by ( [ eq : condition ] ) we have the markov chain @xmath61 using data processing inequality as follows : @xmath62 with @xmath63 for @xmath31 . to be precise
, one may choose ( for @xmath64 ) @xmath65 employing the following well known result with eq .
( [ eq : condition ] ) .
[ lemma : h : cont ] let @xmath0 and @xmath23 be probability distributions on a set with finite cardinality @xmath66 , such that @xmath67 . then @xmath68 see e.g. @xcite .
thus we arrive at [ satz : lower ] for any @xmath35 and @xmath69 : @xmath70 where @xmath71 is the mutual information of the channel @xmath16 between the input distribution @xmath0 and the output distribution @xmath72 .
@xmath73 by using slightly stronger estimates , we even get [ satz : strong : lower ] for every @xmath10 @xmath74 let @xmath2 be an optimal @xmath3code . from eq .
( [ eq : condition ] ) we find ( by a markov inequality argument ) that @xmath75 denote the intersection of this set with the typical sequences @xmath76 ( see eq .
( [ eq : typical ] ) below ) by @xmath77 , with @xmath78
. then @xmath79 and there exists an @xmath80transmission code @xmath81 for the channel @xmath82 with @xmath83 , see @xcite ( the case of a classical quantum channel @xmath16 was done in @xcite ) . by construction
this is a @xmath84code for the channel @xmath85 .
we want now view @xmath34 as belonging to the message encoder , and @xmath47 as belonging to the message decoder , the resulting code being one for the identical channel on @xmath5 .
let us denote the concatenation of the map @xmath47 with the channel decoder by @xmath86 .
on the other hand , we may replace @xmath34 by a deterministic map @xmath87 , because randomization at the encoder never decreases error probabilities : @xmath88 still is an @xmath84code .
it is now obvious that @xmath89 for every @xmath90 , hence @xmath91 and we are done .
it might be a bit daring to formulate conjectures at this point , so we content ourselves with posing the following questions : [ quest : main ] is it true that for all @xmath10 @xmath92 in fact , we would like to go present a slightly stronger statement : _ question [ quest : main] _ : for every @xmath10 , @xmath93 , @xmath94 , and large enough @xmath35 does there exists a @xmath3code with @xmath95 and with the additional property that @xmath96 here @xmath76 is the set of _ typical sequences _ : @xmath97 where @xmath98 counts the number of occurences of @xmath28 in @xmath55 , and @xmath99 .
observe that by chebyshev s inequality @xmath100 in fact , by employing the chernoff bound we even obtain @xmath101 with these bounds it is easily seen that a positive answer to the latter question implies the same to the former .
but also conversely , it is not difficult to show that a `` yes '' to question [ quest : main ] implies a `` yes '' to question [ quest : main].
the following construction is a generalization and refinement of the one by bennett et al .
@xcite ( footnote 4 ) , found independently by dr , vidal , and cirac @xcite . the idea there is to use common randomness between the sender and the receiver of the encoded messages .
formally this means that @xmath34 and @xmath47 also depend on a common random variable @xmath102 , uniformly distributed and independent of all others .
note that this has a nice expression when viewing @xmath34 and @xmath47 as map valued random variables : here we allow dependence ( via @xmath102 ) between @xmath34 and @xmath47 , while in the initial definition , eq . ( [ eq : code ] ) , @xmath34 and @xmath47 are independent ( as random variables ) .
it seems that the power of allowing the use of common randomness can be understood from this point of view : it is a `` convexification '' of the theory with deterministic or independent encoders and decoders .
it is easy to see that the lower bound of theorem [ satz : lower ] still applies here .
we only have to modify the derivation a little bit : @xmath103 with a slight variant @xmath104 of @xmath105 .
we shall apply an explicit large deviation estimate for sampling probability distributions from @xcite ( extended to density operators in @xcite ) , which we state separately without proof : [ lemma : large : deviation ] let @xmath106 be independent identically distributed ( i.i.d . )
random variables with values in the function algebra on the finite set @xmath107 , which are bounded between @xmath108 and @xmath109 , the constant function with value @xmath110 .
assume that the average @xmath111 .
then for @xmath112 @xmath113\right\ } \leq 2|{{\mathcal{k}}}|\exp\left(-m\frac{\eta^2 s}{2\ln 2}\right)\!,\ ] ] where @xmath114=[(1-\eta)\sigma;(1+\eta)\sigma]$ ] is an interval in the value wise order of functions : @xmath115=\{x:\forall k\ a(k)\leq x(k)\leq b(k)\}$ ] .
@xmath73 before we prove our main theorem , we need three lemmas on exact types and conditional types .
the first is a simple yet crucial observation : [ lemma : types : blurb ] let @xmath16 be a channel from @xmath1 to @xmath15 , @xmath0 a p.d
. on @xmath1 , @xmath116 the induced distribution on @xmath15 and @xmath117 the transpose channel from @xmath15 to @xmath1 .
let @xmath118 , @xmath119 be exact @xmath35types of @xmath1 , @xmath15 , respectively that are marginals of a joint exact @xmath35type @xmath120 of @xmath121 .
consider the uniform distribution @xmath122 on @xmath123 on @xmath123 , which has the property @xmath124 and the channel from @xmath123 to @xmath125 , @xmath126 where @xmath127 is the set of _ conditional exact typical sequences _ of @xmath55 .
then the induced distribution @xmath128 on @xmath125 is the uniform distribution , i.e. @xmath129 and the transpose channel to @xmath130 is indeed @xmath131 , defined by @xmath132 with @xmath133 .
straightforward .
[ lemma : cardinalities ] there is an absolute constant @xmath134 such that for all distributions @xmath0 on @xmath1 , @xmath135 , channels @xmath136 and @xmath94 @xmath137 for @xmath138 , consider a joint @xmath35type @xmath120 on @xmath121 with marginals @xmath118 on @xmath1 and @xmath119 of @xmath15 . then , introducing the channel @xmath139 with @xmath140 : @xmath141 see @xcite .
the third contains the central insight for our construction : [ lemma : covering ] with the hypotheses and notation of lemma [ lemma : types : blurb ] there exist families @xmath142 , @xmath143 , from @xmath144 such that for all @xmath102 @xmath145,\tag{i${}_\nu$}\ ] ] and @xmath146,\tag{ii}\ ] ] for all @xmath147 and @xmath148 that satisfy @xmath149 introduce i.i.d .
random variables , distributed on @xmath125 according to @xmath150 ( i.e. uniformly ) .
then for all @xmath151 : @xmath152 hence lemma [ lemma : large : deviation ] applies and we find @xmath153 by choosing @xmath148 and @xmath147 according to the lemma we enforce that the sum of these probabilities is less than @xmath110 , hence there are actual values of the @xmath154 such that all ( i@xmath155 ) and ( ii ) are satisfied .
with this we are ready to prove : [ satz : sim : feedback : channel ] there exists an @xmath3code @xmath156 with @xmath157 and common randomness consumption @xmath158 in fact , not only the condition ( [ eq : condition ] ) is satisfied but the even stronger @xmath159 suppose @xmath55 is seen at the source , and that its type is @xmath118 .
for each joint @xmath35type @xmath120 of @xmath121 we assume that families @xmath160 as described in lemma [ lemma : covering ] are fixed throughout .
then the protocol the sender follows is : 1 .
choose a joint type @xmath120 on @xmath121 with probability @xmath161 and send it .
note that @xmath120 can be written @xmath140 , with the marginal @xmath118 on @xmath1 and a channel @xmath162 .
2 . if @xmath118 is not typical or @xmath120 is not jointly typical then terminate . 3 . use the common randomness to choose @xmath102 uniformly .
4 . choose @xmath163 according to @xmath164 and send it .
the receiver chooses @xmath165 , using the common randomness sample @xmath102 .
let us first check that this procedure works correctly : for typical @xmath55 we can calulate the distribution of @xmath166 conditional on the event that their joint type is @xmath120 : this is then a distribution on @xmath167 , and we assume @xmath120 to be typical .
@xmath168 with the `` big
notation : @xmath169 signifies any function whose modulus is bounded by @xmath170 . here we have used the definition of the protocol , then lemma [ lemma : types : blurb ] ( for the definition of @xmath171 and the fact that @xmath172 does not depend on @xmath173 ) , then lemma [ lemma : covering ] .
so , the induced distribution is , up to a factor between @xmath174 and @xmath175 , equal to the correct output distribution @xmath176 . now averaging over the typical @xmath120 gives eq .
( [ eq : individual : condition ] ) .
what is the communication cost ?
sending @xmath120 is asymptotically for free , as the number of joint types is bounded by the polynomial @xmath177 .
sending @xmath163 costs @xmath178 bits , with @xmath147 bounded according to lemma [ lemma : covering ] .
that is , @xmath179 on the other hand @xmath180 and we are done .
[ rem : chernoff : better ] in the above statement of theorem [ satz : sim : feedback : channel ] we assumed @xmath181 to be a constant , absorbed into the `` @xmath182 '' in the code length estimate . using the chernoff estimate ( [ eq : typical : prob : exp ] ) on the probabilities of typical sets in the above proof in fact shows the existence of an @xmath3code satisfying ( [ eq : individual : condition ] ) @xmath183 in the line of @xcite , the interpretation of this result is that investing common randomness at rate @xmath184 , one can simultate the noisy channel @xmath16 by a noiseless one of rate @xmath185 , when sending only @xmath0typical words .
considering the construction again , we observe that in fact not only it provides a simulation of the channel @xmath16 , but additionally of the _ noiseless passive feedback_. simply because the sender can read off from his random choices the @xmath166 obtained by the receiver , too .
this observation is the key to show that our above construction is optimal under the hypothesis that the channel _ with noiseless passive feedback _ is simulated : in fact , since both sender and receiver can observe the very output sequence @xmath166 of the channel , which has entropy @xmath186 , they are able to generate common randomness at this rate . since communication was only at rate @xmath185 , the difference must by invested in prepared common randomness : otherwise we would get more of it out of the system than we could have possibly invested .
formally this insight is captured by the following result : [ satz : feedback : channel : opt ] if the decoder of a @xmath3code @xmath2 with common randomness consumption @xmath187 $ ] ( with distribution @xmath188 ) depends deterministically on @xmath102 and @xmath90 ( which is precisely the condition that the encoder can recover the receiver s output ) then @xmath189 for the first inequality introduce the channels @xmath190 , and their induced distributions @xmath191 on @xmath21 and transpose channels @xmath192 with respect to @xmath58 , i.e. @xmath193 then we can rewrite eq .
( [ eq : condition ] ) as @xmath194 this inequality oviously remains valid if we restrict the sum to @xmath195 and replace @xmath196 and @xmath192 by their restrictions to @xmath197 : @xmath198 and @xmath199 , respectively .
on the other hand , choosing @xmath200 , we have @xmath201 which yield @xmath202 hence there exists at least one @xmath102 such that @xmath203 note that , as functions on @xmath204 , @xmath205 so , when we introduce the support @xmath206 of the left hand side , we arrive at @xmath207 from which our claim follows by a standard trick @xcite : let @xmath208 , with @xmath209 .
then @xmath210 and using the fact that @xmath211 this implies @xmath212 now only note that ( since @xmath213 is deterministic ) @xmath214 and by @xmath215 we are done
. now for the second inequality : from the definition we get , by summing over @xmath55 , @xmath216 because the @xmath213 are all deterministic , the distributions @xmath217 are all supported on sets of cardinality @xmath8 .
hence the support @xmath206 of @xmath218 can be estimated @xmath219 .
on the other hand , we deduce @xmath220 which , by the same standard trick @xcite as before , yields our estimate : with @xmath221 , the set @xmath222 satisfies @xmath223 but since for all @xmath224 @xmath225 we can conlude @xmath226 collecting these results we can state [ cor : feedback : channel : opt ] for any simulation of the channel @xmath16 together with its noiseless passive feedback with error @xmath227 , at rate @xmath118 and common randomness consumption rate @xmath228 : @xmath229 conversely , these rates are also achievable
. a simulation of the channel must be in the error bound for _ every _ input @xmath55 , hence eq .
( [ eq : condition ] ) will be satisfied for every distribution @xmath0 .
the lower bounds follow now from theorem [ satz : feedback : channel : opt ] by choosing @xmath0 to maximize @xmath185 and @xmath186 , respectively . to achieve this , the encoder , on seeing @xmath55 reports its type to the receiver ( asymptotically free ) and then they use the protocol of theorem [ satz : sim : feedback : channel ] for @xmath230 , the empirical distribution of @xmath55 .
possibly they have to use the channel at rate @xmath231 to set up additional common randomness beyond the given @xmath232 . at this point
we would like to point out a remarkable parallel of methods and results to the work @xcite : our use of lemma [ lemma : covering ] is the classical case of of the use of its quantum version from @xcite , and the main result of the cited paper is the quantum analog of the present theorem [ satz : sim : feedback : channel ] .
the optimality result there has its classical case formulated in theorems [ satz : lower ] ( and [ satz : strong : lower ] ) and [ satz : feedback : channel : opt ] , and even the construction of the following section has its counterpart there .
the use of common randomness turned out to be remarkably powerful , and it is known in various occasions to make problems more tractable : a major example is the arbitrarily varying channel ( see for example the review @xcite ) . while for discrete memoryless channels it does not lead to improved rates or error bounds , it there allows for a `` reverse '' of shannon s coding theorem @xcite in the sense of simulating efficiently a noisy channel by a noiseless one .
this viewpoint seems to extend to quantum channels as well , assisted by entanglement rather than common randomness : see @xcite .
we shall expand on the power of the `` randomness assisted '' viewpoint in section [ sec : applications ] .
here we show that from the theorem of the previous section a solution to the compression problem under a slightly relaxed distance criterion follows : whereas previously we had to employ common randomness to achieve the lower bound @xmath185 , this will turn out to be unnecessary now . specifically
, our condition will be eq .
( [ eq : l : condition ] ) : [ satz : coding : local ] there exists an @xmath35block code @xmath2 with @xmath233 such that @xmath234 choose an @xmath3code @xmath235 as in theorem [ satz : sim : feedback : channel ] .
obviously this code meets the condition of the theorem , except for the use of common randomness .
we will show that a uniformly random choice among a _ small _ ( subexponential ) number of @xmath102 is sufficient for this to hold .
then the protocol simply is : 1 .
the sender choses @xmath102 uniformly random ( among the chosen few ) , and sends it to the receiver ( at asymptotic rate @xmath108 ) .
2 . she uses @xmath236 to encode , and the receiver uses @xmath213 to decode . by construction
this meets the requirements of the theorem . to prove our claim , note that from theorem [ satz : sim : feedback : channel ] we can infer @xmath237 introduce i.i.d .
random variables @xmath238 , distributed according to @xmath239 . with the notations @xmath240 and @xmath241 we have
@xmath242 denote the minimal nonzero entry of @xmath16 by @xmath243 , and choose @xmath170 so small that for all typical @xmath55 and all @xmath56 @xmath244 by lemma [ lemma : large : deviation ] we obtain @xmath245\text { on } { { \operatorname{supp}\,}}w_{x_k}\right\ } \\ & \hspace{4.7 cm } \leq 2|{{\mathcal{y}}}|\exp\left(-q\frac{\epsilon^2 u}{4\ln 2}\right ) .
\end{split}\ ] ] hence the sum of these probabilities is upper bounded by @xmath246 which is less than @xmath110 for @xmath247 hence there exist actual values @xmath238 such that @xmath248 which is what we wanted to prove : observe that @xmath23 grows only polynomially .
as we remarked already in the introduction , @xcite proposed to prove this result ( and indeed more , being interested in the tradeoff between rate and error ) , but eventually turned to the much softer condition ( [ eq : ks : condition ] ) , which originates from the traditional model of rate distortion theory .
nice though the idea of the previous section is , the lower bound results show that on this road we can not hope to approach the conjectured bound , because without common randomness at hand we have to spend communication at the same rate to establish it ( compare @xcite , appendix , for this rather obvious looking fact ) . in this section
we want to study the _ perfect _ restitution of the probability distributions @xmath14
( i.e. @xmath32 ) : recall that here we want to minimize @xmath249 , and this minimum we call @xmath250 .
obviously @xmath36 , so the limit @xmath37 exists , and is equal to the infimum of the sequence
. then we have [ satz : old : new ] for all @xmath10 @xmath251 it is sufficient to prove the inequality for @xmath252 in place of @xmath253 : fix a @xmath110code @xmath254 with @xmath255 .
then , for @xmath256 choose any @xmath3source code @xmath257 for @xmath258 , which is possible at rate @xmath259 .
then @xmath2 with @xmath260 and @xmath261 is an @xmath3code for the mixed state source with limiting rate @xmath252
. it would be nice if we could prove also an inequality in the other direction , but it seems that a direct reduction like in the previous proof does not exist : for this we would need to take an @xmath3code and convert it to an @xmath262code , increasing the entropy only slightly .
[ fig : network ] . note that we included a sink , edges leading to the sink obviously having probability @xmath110.,title="fig : " ] a nice picture to think about the problem of finding @xmath252 is the following in the spirit of flow networks : from the source we go to one of the nodes @xmath13 , with probability @xmath263 .
then , with a probability of @xmath264 we go to @xmath90 , and from there with a probability of @xmath265 to @xmath266 .
then the condition is that @xmath267 examples of this constructions are discussed in @xcite ( where it was in fact invented ) , and here we want to add some general remarks on optimizing it , as well thoughts on a possible algorithm to do that .
we begin with a general observation on the number of intermediate nodes : [ satz : cd ] an optimal zero error code for @xmath16 requires at most @xmath268 intermediate nodes , with @xmath269 , @xmath270 . for a fixed set @xmath5
the problem is the following : + under the constraints @xmath271 minimize the entropy @xmath272 , where @xmath273 .
observe that for each fixed set of @xmath274 the constraints define a convex admissible region for the @xmath275 , of which a _ concave _ function is to be minimized .
hence , the minimum will be achieved at an extreme point of the region , that we rewrite as follows : @xmath276 an extreme point must be extremal in every of the summand convex bodies @xmath277 .
on the other hand , an extreme point of @xmath277 must meet @xmath278 many of the inequalities ( @xmath279 ) with equality .
since @xmath280 there remain only at most @xmath47 nonzero @xmath275 for every @xmath28 . in particular , only at most @xmath281 many @xmath90 are accessed at all .
in fact , to minimize @xmath272 , at most @xmath268 , otherwise @xmath282 would contain full information about @xmath28 .
[ rem : cd : more ] the last argument can be improved : for @xmath283 we can even assume @xmath284 .
the argument of the proof gives us the idea that maybe by an alternating minimization we can find the optimal code : indeed , conditions ( [ eq : cd : e ] ) and ( [ eq : cd : ed ] ) for fixed @xmath47 are linear in @xmath34 , and the target function is concave ( entropy of a linear function of @xmath34 ) , so we can find it s minimum at an extreme point of the admissible region .
this part is solved by standard convex optimization methods . on the other hand , for fixed @xmath34 , eqs .
( [ eq : cd : d ] ) and ( [ eq : cd : ed ] ) are linear in @xmath47 .
however , variation does not change the aim function .
still we have freedom to choose , and this might be a good rule : let @xmath47 _ maximize _ the conditional entropy @xmath285 .
the rationale is that this entropy signifies the ignorance of the sender about the actual output .
if it does not approach @xmath184 in the limit this means that the protocol simulates partial feedback of the channel @xmath16 , which could be used to extract common randomness .
this amount is a lower bound to what the protocol has to communicate in excess of @xmath185 .
we have , however , no proof that this rule converges to an optimum .
in this section we point out three important connections to other questions , some of which depend on positive answers to the questions [ quest : main ] and [ quest : main]. it is known that if two parties ( say , alice and bob ) have access to many inpendent copies of the pair of random variables @xmath286 ( which are supposed to be correlated ) , then they can , by public discussion ( which is overheard by an eavesdropper ) , create common randomness at rate @xmath287 , almost independent of the eavesdropper s information . for details
see @xcite , where this is proved , and also the optimality of the rate .
one might turn around the question and ask , how much common randomness is required to create the pair @xmath288 approximately .
this question , in the vein of that of the previous subsection , is really about reversibility of transformations between different appearances of correlation .
note that this was confirmed in @xcite for the case of deterministic correlation between @xmath289 and @xmath290 , i.e. @xmath291 , which there was parallelled to entanglement concentration and dilution for pure states .
an affirmative answer to question [ quest : main ] , surprisingly implies that a rate of @xmath287 of common randomness is sufficient , _ with no further public discussion _ to create pairs @xmath292 .
this is done by first creating the distribution @xmath23 of @xmath293 on @xmath5 from the common randomness ( this alice and bob do each on their own ! ) : this may be not altogether obvious as the common randomness is assumed in pure form ( i.e. a uniform distribution on @xmath148 alternatives ) , while the distribution @xmath23 may have no regularity . to overcome this
difficulty fix an @xmath93 and let @xmath294 now we partition the unit interval into the subintervals @xmath295,\end{aligned}\ ] ] and define @xmath296 , @xmath297 .
notice that for @xmath298 the probabilities for @xmath282 s belonging to the same set @xmath299 differ from each other only by a factor between @xmath300 and @xmath301 , and that @xmath302 , because of @xmath303 , by definition of @xmath56 .
hence , defining uniform distributions @xmath304 on @xmath299 for @xmath298 , it is immediate that @xmath305 now the distribution on the @xmath306 in this formula can be approximated to within @xmath307 by a @xmath308type distribution , which in turn can be obtained directly from a uniform distribution on @xmath308 alternatives . in this way we reduced everything to a number of uniform distributions , maybe on differently sized sets ,
all bounded by @xmath8 and a helper uniform distribution on a set of size @xmath308 .
however , it is well known that these can be obtained from a uniform distribution on @xmath309 items within arbitrarily small error . given this distribution on @xmath5 , bob applies @xmath47 , whereas alice applies the transpose channel @xmath310 to @xmath34 .
one readily checks that this produces the joint distribution of @xmath311 , up to arbitrarily small disturbance in the total variational norm .
note that this result would imply a new proof of the optimality of of the rate @xmath312 of common randomness distillation from @xmath311 : because we can simulate the latter pair of random variables with this rate of common randomness , we would obtain a net increase of common randomness after application of the distillation , which clearly can not be .
it was already pointed out that this study has the paper @xcite as one motiviation , with its idea to prove the optimality of shannon s coding theorem by showing that every noisy channel @xmath16 can be simulated by a binary noiseless one operating at rate @xmath313 .
shannon s theorem is understood as saying that the noisy channel can simulate a binary noiseless one of rate @xmath313 .
both simulations are allowed to perform with small error .
note that an affirmation of question [ quest : main] , implies that this can be done , without the common randomness consumption like in section [ sec : cr : trick ] .
as indicated , this provides a proof of the converse to shannon s coding theorem : the idea is that otherwise we could , given a rate of @xmath313 noiseless bits simulate the channel , which in turn could be used to transmit at a rate @xmath314 . the combination of simulation and coding yields a coding method for transmitting @xmath118 bits over a channel providing @xmath313 noisless bits , which is absurd ( in @xcite this reasoning is called `` causality argument '' ) .
theorem [ satz : sim : feedback : channel ] allows us to prove even more : [ satz : feedback : capacity ] for the channel @xmath16 with noisless feedback ( i.e. after each symbol @xmath28 transmitted the sender gets a copy of the symbol @xmath50 read by the receiver , and may react in her encoding ) the capacity is given by @xmath313 .
in fact , for the maximum size @xmath315 of an @xmath3feedback code @xmath316 let an optimal @xmath3feedback code for the channel @xmath82 with noiseless feedback be given .
we will construct an @xmath317code with shared randomness , as follows : choose a simulation of the channel @xmath16 on @xmath35blocks sending @xmath318 bits , and using shared randomness , and with error bounded by @xmath319 ( this is possible by the construction of theorem [ satz : sim : feedback : channel ] see remark [ rem : chernoff : better ] ) .
we shall use @xmath35 independent copies of the feedback code in parallel : in each round @xmath35 inputs symbols are prepared , sent through the channel , yielding @xmath35 respective feedback symbols .
obviously , each round can be simulated with an error in the output distribution bounded by @xmath170 , using our simulation of the channel w ( which , as we remarked earlier , simulates even the feedback ) . in each of the parallel executions of the feedback code
thus accumulates an error of at most @xmath320 , increasing the error probability of the code to @xmath321 .
hence on the block of all the @xmath35 feedback codes we can bound the error probability by @xmath322 .
but this is subexponentially ( in @xmath323 ) close to @xmath110 , so a standard argument applies : first , by considering average error probability we can get rid of the shared randomness : there exists one value of the shared random variable for which the average error probability is bounded by @xmath324 .
then we can argue that there is a subset @xmath325 of the constructed code s message set @xmath326 which has _
maximal _ error probability bounded by @xmath327 and @xmath328 what we achieved so far hence is this : a code of @xmath329 messages with error probability @xmath330 and using @xmath331 noiseless bits .
clearly , we may assume the encoder to be deterministic without losing in error probability .
but then at most @xmath332 messages can be mapped to the same codeword without violating the error condition . collecting everything we conclude
@xmath333^n , \end{split}\ ] ] implying the theorem .
[ rem : feedback ] the _ weak converse _ ( i.e. the statement that the rate for codes with error probability approaching @xmath108 is bounded by @xmath313 ) is much easier to obtain , by simply keeping track of the mutual information between the message and the channel output through the course of operating a feedback code , using some well known information identities , and finally estimating the code rate employing fano s inequality .
let @xmath334 be any _ distortion measure _
, i.e. a non negative real function .
this function is extended to words @xmath335 by letting @xmath336 shannon s rate distortion theorem is about the following problem : construct an @xmath35block code @xmath2 ( which my be chosen to be deterministic ) such that for a given @xmath337 @xmath338 i.e. , the average distortion between source and output word is bounded by @xmath339 . a pair @xmath340 of non negative real numbers is said to be _ achievable _ if there exist @xmath35block codes with code rate tending to @xmath118 and distortion rate asymptotically bounded by @xmath45 .
define the _ rate distortion function _
@xmath341 as the minimum @xmath118 such that @xmath340 is achievable .
[ satz : shannon : r : d ] the rate distortion function is given by the following formula : @xmath342 where @xmath343 is the expected ( single letter ) distortion when using the channel @xmath16 .
the proof of `` @xmath344 '' here is a simple exercise using convexity of mutual information in the channel and standard entropy inequalities .
we can give a simple proof of the `` @xmath345''part of this result , using theorem [ satz : feedback : channel : opt ] : choose some channel @xmath16 satisfying the distortion constraint .
then mapping @xmath55 to @xmath346 obviously satisfies the distortion constraint on the code in the sense that the expected distortion between input and output , over source and channel , is bounded by @xmath339 .
of course , sampling @xmath346 at the encoder and sending some @xmath166 will not meet the bound @xmath185 .
however , we can apply theorem [ satz : feedback : channel : opt ] to approximately simulate the joint distribution of @xmath55 and @xmath166 by using some common randomness @xmath102 and a deterministic code @xmath347 sending @xmath348 bits .
hence , invoking linearity of the definition of @xmath349 , @xmath350 so there must be one @xmath102 such that @xmath351 , which ends our proof . at this point
we would like to advertise our point of view that theorem [ satz : coding : local ] , and even more so theorem [ satz : sim : feedback : channel ] , is what rate
distortion is actually about : the former theorem shows how to simulate a given channel on all individual positions of a transmission , and this is what we need in rate distortion .
in fact , rate distortion theory is unchanged when instead of the one convex condition ( `` distortion bound '' ) on the code we have several , effectively restricting the admissible approximate joint types of input and output to any prescribed convex set in particular a single point . the strength of theorem [ satz : coding : local ] in comparison to such a development of rate
distortion theory lies in the fact that with its help we satisfy the convex conditions in _ every letter _ , not just in the block average . and theorem [ satz : sim : feedback : channel ] gives the analogue of this even with the condition imposed on the whole block , yielding results that are not obtainable by simply applying rate
distortion tools ( see e.g. @xcite ) .
the problem studied in this paper has a natural extension to quantum information theory : now the source emits ( generally mixed ) quantum states @xmath14 on the hilbert space @xmath352 ( @xmath13 ) , with probabilities @xmath263 , and an @xmath3code is a pair @xmath2 of maps @xmath353 where @xmath354 is the set of states on the code hilbert space @xmath355 and @xmath47 is completely positive , trace preserving , and linear .
the condition to satisfy is @xmath356 with the trace norm @xmath19 on density operators .
define , like before , @xmath9 as the minimum @xmath357 of an @xmath3code .
sometimes , the stronger condition @xmath358 will be applied .
notice that this contains our original problem as the special case of a _ quasiclassical _ ensemble , when all the @xmath359 commute ( which means they can be interpreted as probability distributions on a set of common eigenstates ) .
this problem ( with a number of variations , which we explained in the introductory section [ sec : pdsources ] for the classical case ) is studied in @xcite . there ( and previously in @xcite ) it is shown that the lower bound theorem [ satz : lower ] holds in the quantum case , too , with understanding @xmath360 as von neumann entropy : [ satz : mixed : lower ] for all @xmath35 , @xmath181 @xmath361 with a function @xmath63 for @xmath31 .
@xmath73 let us improve this slightly by proving the strong version of this result : [ satz : mixed : strong : lower ] for all @xmath10 @xmath362 by much the same method as the proof of theorem [ satz : strong : lower ] : the changes are that we need the more general code selection result of @xcite , thm .
ii.4 , instead of the classical theorem @xcite , and which we state separately below : if @xmath2 is an optimal @xmath3code , define @xmath363 obviously @xmath364 , so we can apply lemma [ lemma : max : code ] and find an @xmath365transmission code @xmath81 for @xmath82 such that @xmath366 this is an @xmath80code for the channel @xmath85 , with @xmath367 , if we choose @xmath170 small enough . combining @xmath34 with the transmission encoder , and @xmath47 with the transmission decoder , we obtain an @xmath3transmission code for @xmath329 many messages over a noiseless system with hilbert space @xmath355 of dimension @xmath9 . to each message @xmath368 there belongs a decoding operator @xmath369 on the coding space @xmath355 , forming together a povm : @xmath370 . now
to decode correctly with probability @xmath371 , for each @xmath243 we must have @xmath372 on the other hand , by @xmath373 , we conclude @xmath374 and we are done .
[ lemma : max : code ] for @xmath375 there is a constant @xmath376 and @xmath94 such that for every discrete memoryless quantum channel @xmath16 and distributions @xmath0 on @xmath1 the following holds : if @xmath377 is such that @xmath378 then there exists an @xmath3transmission code @xmath2 with the properties @xmath379 @xmath380 see @xcite , thm .
progress on the problem of achievability of this bound is not known to us .
it is remarkable that koashi and imoto @xcite could obtain the exact optimal bound in the case of _ blind _ coding .
it is indirectly defined via a canonical joint decomposition of the source states , but it can be derived from their result that generically the optimum rate is @xmath186 , which is achieved by simply schumacher encoding the ensemble @xmath381 . nevertheless , the results obtained in the classical case are very encouraging , so we state two conjectures : [ conj : qmain : cr ] for @xmath69 there exist @xmath3codes with common randomness , asymptotically achieving transmission rate @xmath185 and common randomness consumption @xmath184 .
if it turns out true , and also question [ quest : main ] has a positive answer , we might even hope that also [ quest : qmain ] for @xmath69 , is @xmath382 [ note that , as in the case of question [ quest : main ] , codes achieving the optimal bound may also be constructed to satisfy eq .
( [ eq : qu : condition : strong ] ) . ]
answers `` yes '' .
the implications of these statements , if they are true , would be of great significance to quantum information theory : not only would we get a new proof of the capacity of a classical quantum channel being bounded by the maximum of the holevo information and for the optimality of common randomness extraction from a class of bipartite quantum sources @xcite , but also the achievability of @xmath185 in the quantum rate distortion problem @xcite with visible coding would follow , that until now has escaped all attempts .
we demonstrated the current state of knowledge in the problem of visible compression of sources of probability distributions and its extension to mixed state sources in quantum information theory .
apart from reviewing the currently known constructions we contributed a better understanding of the resources involved : in particular the use of common randomness in some of them , and providing strong converses . also we showed the numerous applications the result ( and sometimes the conjectures ) have throughout information theory , making the matter an eminent unifying building block within the theory .
we would like to draw the attention of the reader once more to our questions [ quest : main ] and [ quest : qmain ] , and especially the conjecture [ conj : qmain : cr ] offering them as a challenge to continue this work .
research partially supported by sfb 343 `` diskrete strukturen in der mathematik '' of the deutsche forschungsgemeinschaft , by fakultt fr mathematik , universitt bielefeld , by the university of bristol , and by the u.k . engineering and physical sciences research council . c. h. bennett , p. w. shor , j. a. smolin , a. v. thapliyal , `` entanglement assisted classical capacity of noisy quantum channels '' , phys .
letters , vol .
30813084 , 1999 .
by the same authors : `` entanglement assisted capacity of a quantum channel and the reverse shannon theorem '' , e print quant - ph/0106052 , 2001 .
g. kramer , s. a. savari , `` quantum data compression of ensembles of mixed states with commuting density operators '' , e print quant - ph/0101119 . presented at the @xmath383 ams meeting ( hoboken , nj ) , april 2829 , 2001 . | we study the problem of efficient compression of a stochastic source of probability distributions . it can be viewed as a generalization of shannon s source coding problem .
it has relation to the theory of common randomness , as well as to channel coding and rate distortion theory : in the first two subjects `` inverses '' to established coding theorems can be derived , yielding a new approach to proving converse theorems , in the third we find a new proof of shannon s rate distortion theorem . after reviewing the known lower bound for the optimal compression rate
, we present a number of approaches to achieve it by code constructions .
our main results are : a better understanding of the known lower bounds on the compression rate by means of a strong version of this statement , a review of a construction achieving the lower bound by using common randomness which we complement by showing the optimal use of the latter within a class of protocols .
then we review another approach , not dependent on common randomness , to minimizing the compression rate , providing some insight into its combinatorial structure , and suggesting an algorithm to optimize it .
the second part of the paper is concerned with the generalization of the problem to quantum information theory : the compression of mixed quantum states . here , after reviewing the known lower bound we contribute a strong version of it , and discuss the relation of the problem to other issues in quantum information theory . |
since zwickys @xcite early discovery of blue compact galaxies ( bcgs ) a large amount of work has been devoted to investigate their nature .
major questions are what causes the burst , how long it can be sustained and if the heating and mechanical energy it produces eventually will cause the cas to be expelled from the system . recent models ( mac low and ferrara @xcite ) indicate that the galaxy probably can retain its gas if @xmath010@xmath1 . cooling and
reaccretion may later ignite a new burst and in a hubble time a few such bursts may occur .
another possibility is the young galaxy hypothesis .
today we have two hot candidates , izw18 and sbs0335 - 052 , both with extremely low metallicities .
the possible presence of old stars is still debated however ( aloisi et al .
@xcite ; hunter & thronson @xcite ; papaderos et al .
@xcite ) . in both cases signs of merger activity
are evident .
this also seems to be a common property of the most active starburst dwarfs and may be the most likely scenario to explain the burst activity .
different classification systems have been proposed for starburst dwarfs .
telles et al .
@xcite separate two different types , one regular type and the other more luminous , having an irregular structure
. these may be related , modified in their morphology by evolution . in the merger scenario one of the progenitors
has to be a gas rich , metal poor galaxy , i.e. a di galaxy , a low surface brightness galaxy ( lsbg ) or an isolated hi cloud .
we have studied the global properties of a few lsbgs and bcgs to look for possible relationships and find that the global parameters like disk scale length and central surface brightness are quite similar .
it has been claimed ( papaderos et al .
@xcite ) that lsbgs can not be related to bcgs since these two parameters differ significantly .
however , the problem may be related to the lack of very deep images .
the scalelengths are derived from regions too close to the central starburst .
1 from a deep b image of eso 338-ig04 shows that the slope of the profile changes at @xmath227@xmath3 arcsec@xmath4 .
an exponential fit to the outer profile results in a central surface brightness of @xmath524.5 , typical of a faint lsbg .
the difference between a power law fit and an exponential fit is however minor and we need to look at the colours of the halo to determine what type of galaxy may be involved .
2 shows results from surface photometry in the optical / near - ir of the bcg halos and the global colours of our sample of blue lsbgs compared to predictions from our spectral evolutionary models . for comparison the colours of the halo of izw18 ( stlin et al .
, in prep . ) is also indicated .
all lsbgs are starforming so the short burst model does not apply .
thus the ages are @xmath6 2 - 3 gyr .
although we can not go as far out into the halo in the near - ir as in the optical , we find that the halo colours of the bcgs range from those of lsbgs to those of very old red stellar populations .
one of the galaxies , eso 338-ig04 , has an exponential lsb disk and at the same time contains a system of globular clusters ( stlin et al .
@xcite ) with a full range of ages and indications of a few bursts in the past .
the gcs indicate the presence of a relatively massive , very old system .
a probable scenario is therefore a merger between an lsbg and a de / dsph .
other strong supports of merger activities come from a study of the h@xmath7 velocity field in this galaxy and a handful of other luminous bcgs ( stlin et al .
if mergers are the cause of luminous starbursts , why would they be fed by lsbgs and not dis ?
bergvall et al .
@xcite argued that lsbgs are more attractive due to the constraints set by the metallicities and the hi contents of dwarf galaxies in general . under the assumption that the samples available in the literature are representative
, the luminosity - metallicity diagram indicates a luminosity difference between a bcg and a di _ at the same metallicity _ of @xmath8 2 magnitudes . a bursting di would obtain a @xmath9/l@xmath10 drastically lower than what is observed for bcgs
lsbgs have no such problem . fig .
3 illustrates the situation in a slightly different way .
there seems to be a clear distinction between di / is and lsbgs in the o / h - @xmath9 diagram .
it shows very nicely that they are the same family , the lsbgs being the less evolved and more gas rich ones . if we ignite a starburst , the position of a galaxy would shift towards a lower gas mass and a higher metallicity .
this indicates that the best guess progenitors for the burst component of the bcgs are the lsbgs . in the most luminous cases , e.g. eso 338 - 04 and eso 400 - 43 , as for sbs0335
, lsbgs seems to be the only option .
the two most metal poor starburst dwarfs , izw 18 and sbs0335 - 052 , are located in the same region as the lsbgs .
are these two galaxies in an early phase of the merging process ?
the disturbed hi morphology of izw 18 ( van zee et al .
@xcite ) may point to this .
the morphology of the hii / hi cloud of sbs0335 - 052 ( see also pustilnik et al 1999 , in preparation ) also opens the long standing question of the low metallicity .
are we measuring the metallicity of the nearly pristine hi gas cloud illuminated by the starburst or has it something to do with the underlying galaxy ?
a possible scenario that would explain the different properties of dis and bcgs is that the star formation efficiency is ruled by the dark matter potential field .
the dm is more centrally concentrated in des than in dis and this may make the difference .
since the dm dominates the mass distribution in dwarf galaxies , the burst we see in luminous bcgs today therefore mimics the first burst in the history of the de
. a merger between an lsbg and a di would probably be less dramatic .
aloisi , a. , tosi , m. , greggio , l. , 1999 , aj 118 , 302 bergvall n. , stlin g. , pharasyn a. , rnnback j. , masegosa j. : 1998 , in highlights of astronomy , vol .
11 hunter , d.a . ,
thronson , h.a .
, 1996 , apj 461 , 202 mac low , m .- m . , ferrara , a. , 1999 , apj 513 , 42 stlin , g , bergvall , n. , rnnback . ,
j. , 1998 , a&a 335 , 85 stlin , g , amram , p. , masegosa , j. , bergvall , n. , boulesteix . , j. , 1999 , a&as 137 , 419 papaderos , p. , loose , h .- h .
, fricke , k.j . ,
thuan , t.x . , 1998 , a&a 338 , 43 papaderos , p. , loose , h .- h . ,
thuan , t.x . ,
fricke , k.j .
, 1996 , a&as 120 , 207 telles , e. , melnick , j. , terlevich , r. , 1997 , mnras 288 , 78 van zee , l. , westpfahl , d. , haynes , m.p . ,
salzer , j. , 1998 , aj , 115 , 1000 zwicky , f. , 1964 , apj 140 , 1467 | the effects of mergers in low mass galaxies are poorly understood . in this paper
we analyze different observational evidences which support the view that mergers can trigger starbursts in dwarf galaxies .
we discuss the relationships between blue compact galaxies ( bcgs ) , dis , des and low surface brightness galaxies ( lsbgs ) and present some evidences which favour that strong starbursts are caused by mergers between des and lsbgs . |
super - kamiokande ( sk ) @xcite is a cylindrical water cherenkov detector with 50-kton of purified water and about 11000 of 20-inch pmt s .
it is located 1000 m underground ( 2700 m water equivalence ) in kamioka observatory , icrr , university of tokyo in mozumi mine in gifu prefecture in japan .
figure [ sk - photo ] shows the inside of the sk detector .
the fiducial volume of the sk detector is 22.5 kton .
the main observation targets of sk detector are solar neutrinos , atmospheric neutrinos , supernova neutrinos , and nucleon decays . sk is the largest detector in the world for these physics targets . sk has started its observation in 1996 .
table [ run - hist ] shows a summary of the sk experiments . [ cols="<,<,>,>,^ " , ] the systematic uncertainty on the total flux is estimated to be 2.1 @xmath0 .
this is about two thirds of the corresponding sk - i value .
the main contributions to the improvement are the vertex shift , angular resolution , and event selection uncertainties .
the total live time for solar neutrino analysis in sk - iii is 548 days for @xmath1 mev , and 289 days for @xmath2 mev .
the recoil electron s total energy range used for this analysis is from 5.0 to 20 mev . the observed @xmath3b solar neutrino flux via the es reaction is @xmath4 ( @xmath5 ) , which is consistent with the previous sk - i and ii results .
the obtained day - night asymmetry is @xmath6 .
the total @xmath3b and hep flux values are referred from the bp2004 standard solar model ( ssm ) @xcite .
we used the expected spectrum of @xmath3b and hep neutrinos calculated by winter @xcite and bahcall @xcite , respectively .
figure [ fig : sk3energy ] shows the observed recoil electron s energy spectrum divided by the expected signal rate from bp2004 ssm without oscillation .
the line in figure [ fig : sk3energy ] represents the total sk - iii average ( flat data / ssm prediction without msw effect ) .
this result indicates no significant spectral distortion .
a preliminary oscillation analysis has also been done including the sk - iii solar neutrino data .
the two flavor analysis for the determination of the solar oscillation parameters , @xmath7 and @xmath8 was performed with the combined sk - i , ii and iii data . in this analysis ,
the total @xmath9b flux is constrained by the averaged nc flux from sno ncd @xcite and leta @xcite ( @xmath10 ) .
the result is shown in the left side in figure [ fig:2flavor ] .
as shown in the left figure , the low solution is excluded and the lma solution is only allowed to explain the solar neutrino oscillation . and
tan@xmath11 from solar analysis ( preliminary ) .
the left figure shows the results from the sk - i , ii and iii combined analysis with the @xmath9b flux constrained by sno nc measurements .
the right figure presents the result of global analysis including sk and other solar experiments ( sno , borexino , homestake , gallex - gno and sage ) .
the star mark indicates the best - fit parameter set.,width=302 ] we have also done a preliminary global oscillation analysis with other solar experiments . in this analysis ,
the following observation results are used ; sno total cc rates observed in the 306-day pure d@xmath12o phase ( sno - i ) @xcite , 391-day salt phases ( sno - ii ) @xcite , and 385-day ncd phase ( sno - iii ) @xcite , the combined nc rates of leta @xcite and sno - iii , the predicted day - night asymmetry for sno - i and ii , @xmath13 solar neutrino flux from borexino @xcite , homestake @xcite , gallex - gno @xcite , and sage @xcite .
the result is shown in the right side in figure [ fig:2flavor ] .
the best - fit parameter set is @xmath14 and @xmath15 ev@xmath16 .
in addition , combining the above and kamland data @xcite , the best - fit parameter set becomes @xmath17 and @xmath18 ev@xmath16 .
we also performed a preliminary 3-flavor oscillation analysis . in the 3-flavor analysis ,
the calculation of oscillation probability is based on reference @xcite .
the probability can be calculated with three parameters : @xmath7 , @xmath19 , and @xmath20 , assuming @xmath21 .
we fixed @xmath22 ev@xmath16 and the normal hierarchy is assumed .
for the solar neutrino oscillation , the other mixing parameters are irrelevant , but we set @xmath23 and @xmath24 in our calculation . as done for the two - flavor analysis , the oscillation probabilities depending on different zenith angles are calculated .
other experiment data are have been published by homestake , sage , gallex - gno , borexino , sno , and kamland .
the oscillation parameters are scanned in the following regions : @xmath25 ev@xmath26 ev@xmath16 , @xmath27 , and @xmath28 .
figure [ fig:3flavor ] shows the allowed region for sin@xmath29 and tan@xmath30 extracted from this analysis . and
tan@xmath30 from the preliminary 3-flavor analysis .
the thick lines and the star mark show the allowed regions and the best - fit point of the solar global analysis .
the thin lines and the square mark show the allowed regions and the best - fit point of our kamland analysis .
the filled areas and the filled circle mark show the allowed regions and the best - fit point of the combined analysis .
for all regions , the innermost area ( red ) , the middle area ( green ) and the outermost area ( blue ) show 68.3 , 95 , 99.7 % c.l . respectively . , width=226 ] the allowed region from the combined analysis of solar global and kamland analyses is also indicated .
the global solar analysis finds that the upper limit of sin@xmath29 is 0.060 at 95 % c.l . after combining with the kamland result
, the best fit value of sin@xmath29 is found to be @xmath31 and the upper limit is obtained as sin@xmath32 at 95 % c.l .
super - kamiokande - iv is running with the lowest energy threshold in sk .
currently , 100 % efficiency is at @xmath33 mev .
preliminary results of the full 3-flavor atmospheric neutrino oscillation analysis from sk - i+ii+iii data are obtained .
there is no preference for either mass hierarchy in the data , and no significant constraint applied on cp phase at 90 % c.l .
preliminary results of the solar neutrino measurement in sk - iii are obtained . the measured solar @xmath3b flux via es reaction is @xmath34 . in a preliminary 3-flavor global solar neutrino oscillation analysis with kamland
, the best - fit value of @xmath35 is found to be @xmath31 and an upper bound is obtained as @xmath36 at 95 % c.l .
the author gratefully acknowledges the cooperation of the kamioka mining and smelting company .
super - kamiokande has been built and operated from funds provided by the japanese ministry of education , culture , sports , science and technology , the u.s .
department of energy , and the u.s . national science foundation .
this work was partially supported by the research foundation of korea ( bk21 and knrc ) , the korean ministry of science and technology , the national science foundation of china , and the spanish ministry of science and innovation ( grants fpa2009 - 13697-c04 - 02 and consolider - ingenio-2010/cpan ) . the super - kamiokande collaboration , nucl .
instr . and meth .
* a501 * ( 2003 ) 418 .
s. yamada _
et al . _ , ieee trans .
* 57 * ( 2010 ) 428 .
j.n . bahcall and m.h .
pinsonneault , phys .
lett * 92 * ( 2004 ) 121301 .
et al . _ ,
c * 73 * ( 2006 ) 025503 .
et al . _ ,
b ( proc . suppl . )
* 77 * ( 1999 ) 64 .
b. aharmim _
et al . _ ,
* 101 * ( 2008 ) 011301 .
b. aharmim _
et al . _ ,
c * 81 * ( 2010 ) 055504 .
b. aharmim _
et al . _ ,
c , * 75 * ( 2007 ) 045502 .
s.n . ahmed _
et al . _ ,
* 72 * ( 2005 ) 055502 . c. arpesella _ et al .
* 101 * ( 2008 ) 091302 .
_ , astrophys
. j. * 496 * ( 1998 ) 505 .
m. altmann _
et al . _ ,
b * 490 * ( 2000 ) 16 ; also presented at neutrino 2004 by c.cattadori .
j.n . abdurashitov _
c * 60 * ( 1999 ) 055801 .
et al . _ ,
* 100 * ( 2008 ) 221803 . v. barger _
et al . _ ,
d. * 22 * ( 1980 ) 2718 . | the recent results from super - kamiokande ( sk ) are reported . on atmospheric neutrino analysis ,
we have performed a full 3-flavor oscillation analysis with sk - i+ii+iii data . a cpt violation study on atmospheric neutrino
is also done with sk - i+ii+iii data . on solar neutrino analysis , a 3-flavor oscillation analysis with sk - iii data
is performed .
solar neutrino , atmospheric neutrino |
v344 lyrae ( wood et al . 2011 ) is a short - period su uma - type of the dwarf novae ( dn ) subclass to the photometrically active cataclysmic variables ( cvs ) .
su uma s consist of a late - type main - sequence star that fills its roche lobe and transfers its gaseous material to an accretion disk around a white dwarf .
in addition to outbursts ( e.g. , lasota 2001 ) , which are periodic increases in brightness due to a thermal instability in the disk , su uma systems exhibit superoutbursts that last a few times longer and are up to a magnitude brighter . the source to superoutbursts is a tidal instability : the disk expands outward until enough mass has expanded beyond the @xmath4=3 eccentric inner lindblad resonance radius , at which time the outer disk annuli are tidally forced into cyclically changing shape from circular to highly eccentric .
a characteristic of superoutbursts is large - amplitude positive superhumps that have a disk origin ( horne 1984 ) .
in addition to these positive superhumps , negative superhumps are also sometimes found in the same system .
like positive superhumps , negative superhumps also have a disk origin ( montgomery 2012a ) .
systems that show positive or negative superhumps in light curves also sometimes show prograde or retrograde precession , respectively , suggesting superhumps and precession are intrinsically linked .
numerical simulations show disks cyclically changing shape from circular to highly eccentric , producing a saturated positive @xmath5 superhump that has a slightly longer period than the orbital period .
tidal torques on the asymmetric density disk result in disk prograde precession ; the secondary needs to advance slightly more than one orbital period to meet up with the next shape - change cycle , and several cycles are needed for each prograde precession . two asymmetric spiral - like density waves in the outer - to - mid annuli advance @xmath1180@xmath6 in the prograde direction in one positive superhump period as shown in simpson & wood ( 1998 ) , for example , and may be induced when higher order harmonics are present ( montgomery 2012b ) .
other numerical simulations show disks tilting , producing a negative superhump that has a slightly shorter period than the orbital period ( montgomery 2012a ) .
tidal torques on the tilted disk result in disk retrograde precession ( montgomery 2009a ) ; the secondary meets up with the second node slightly before one orbital period , and several cycles are needed for each retrograde precession ( montgomery 2012a ) . for cv dn su uma - type systems and their higher - mass transfer rate analogs , nova - likes ( nls ) , both precessions are seen either in series or simultaneously .
as noted in skillman et al .
( 1998 ) for tt ari , understanding how a differentially rotating fluid disk can manage @xmath7 well - defined precession frequency is difficult in itself , let alone two . to date , both precessions have not yet been shown to occur naturally in a numerical simulation , although some artificially induce ( wood et al .
2011 ) . in this work ,
we show how a differentially rotating fluid disk can simultaneously precess in both directions .
we use the same 25,000 particle sph code ( simpson & wood , 1998 ) that has successfully produced either positive superhumps and prograde precession ( e.g. , simpson & wood , 1998 ) or negative superhumps and retrograde precession ( montgomery 2012a ) .
we use v344 lyrae , an su uma system in the @xmath0 field , as our model .
v344 lyrae experiences outbursts and superoutbursts ; has @xmath10.25 mag positive and @xmath10.8 mag negative superhumps with periods 2.20 and 2.06 hours , respectively ; and has an orbital period 2.11 hours ( wood et al . 2011 ) . in 2
, we briefly discuss the sph code , and in 3 , we present the simulation results . in 4 , we discuss the results , and in 5 , we provide conclusions and future work .
the 3d sph codes used and described in wood et al . ( 2000 ) , montgomery ( 2004 , 2009a ) , wood et al .
( 2009 ) , and wood et al .
( 2011 ) have their roots in simpson ( 1995 ) , a code that applies to non - magnetic cv systems .
these later versions of the root code have e.g. , more particles ( i.e. , more than 100,000 ) , modified particle shapes , varying injection rates .
however , none of these changes have resulted in disks that simultaneously precess in the prograde and retrograde directions without inducing artificial means . to induce both superhump modulations and both precessions ,
the run is stopped after prograde precession has begun , the disk is artificially rotated , and the run is restarted , as shown in some of these works .
although negative superhumps and retrograde precession are now present , the act to induce is artificial .
hence we return to the root code to see if accretion disks can tilt naturally and precess in both directions .
results such as frequency values and amplitudes of signals can then be compared to results obtained from the control group , simulations using the same base code but artificially inducing a 5@xmath6 disk tilt ( montgomery 2009a ) .
the root code ( simpson 1995 ) uses the lagrangian method of sph , has a maximum of 25,000 particles , assumes an ideal gas law with a low adiabatic gamma , does not include radiation effects or magnetic fields . the only unknown is the shukura - sunyaev ( 1973 ) @xmath8 viscosity .
we use nearly the same input parameters in montgomery ( 2012a ) , but this time we want to simulate an su uma system : we assume secondary - to - primary mass ratio @xmath9 , a similar mass ratio to that used in wood et al .
( 2011 ) ; primary mass @xmath10 @xmath11 ; viscosity coefficients @xmath12 and @xmath13 ; smoothing length @xmath14 , and separation distance @xmath15 . in all simulations ,
the primary and secondary stars are treated as point masses , and gas particles are injected through the inner lagrange point @xmath16 to build the disk from scratch .
the simulation produces @xmath5 positive superhumps ( @xmath17 @xmath18 , where @xmath18 is the orbital period , @xmath19 is the positive superhump period ) and prograde precession once the superoutburst is initiated near orbit 65 ( i.e. , @xmath15.7 days in v3444 lyrae if the disk is built from scratch ) .
the outer annuli of the disk cyclically changes shape from circular to highly eccentric ; two asymmetric spiral waves build and dissipate like those shown in montgomery ( 2012b ) and whitehurst ( 1984 , 1988 ) but do not appear to advance ; and only the fundamental apsidal superhump frequency @xmath20 is present in the fourier transform like that shown in montgomery ( 2012b ) .
unlike other attempts to model v344 lyrae ( wood et al . 2011 ) , we continue the simulation beyond orbit 420 ( @xmath137 days in v344 lyrae ) as additional time is needed for settling of relatively colder particles to the disk midplane . until orbit @xmath1965 ( i.e. , @xmath1900 orbital periods or @xmath185 days in v344 lyrae after superoutburst initiation ) , only 2@xmath20 and 3@xmath20 higher order harmonics have since initiated , and spiral like structures in outer annuli weakly advance ( not shown ) . near orbit @xmath1965 , the disk begins to tilt and precess in the retrograde direction .
in addition to the already present apsidal mode frequencies , the fundamental nodal superhump @xmath21 and linear combination frequencies are now present ( see figure 1 and table 1 ) .
two independent precessions are now occurring simultaneously in the same disk , prograde and retrograde .
as shown in figures 2 and 3 , the disk is tilted and precessing in the retrograde direction as well as simultaneously cyclically changing shape from circular to highly eccentric and precessing in the prograde direction .
based on the observational data for v344 lyrae ( wood et al .
2011 ) , @xmath22/@xmath23 = 1.8 where ( @xmath22 = @xmath24 -1 ) , ( @xmath23 = 1 - @xmath25 ) , and @xmath26 is the negative superhump period . the ratio @xmath22/@xmath23 is related to the rate of advance of the line of apsides relative to the rate of regression of the line of nodes
. the simulation does not produce the same fractional period excess @xmath22
( 4.4% , observed ; 7.5% , simulated ) or fractional period deficit @xmath23 ( 2.5% , observed ; 3.9% , simulated ) , but does produce nearly the same @xmath22/@xmath23 ratio . to obtain more accurate individual fractional excesses , better estimates of parameter(s ) like stellar masses are needed . to obtain a more precise ratio of the fractional periods , tidal torques on differing disk thickness along the rim during precession
may need to be included in the definition , which is currently based on a uniform thickness disk that is solely precessing either in the prograde or retrograde direction .
figure 2 is in black and white to emphasize the location of the two spiral density waves in the outer - to - middle annuli relative to the location of the bright spot , the spot where the gas stream strikes the rim of the disk . from time @xmath27=126.6 days
to @xmath27=127.0 days in v344 lyrae , the spiral waves advance in the prograde direction .
when the disk is highly eccentric at @xmath27=126.2 days and @xmath27=127.4 days , the density waves compress with the disk rim ( i.e. , the rim on the far side of the disk , furthest from the secondary star , compresses into the spiral wave on the far side of the disk ; meanwhile , the spiral wave on the near side of the disk compresses into the rim of the disk on the near side ; the compressions are sequential , not simultaneous ) .
when the disk is more circular at @xmath27=126.6 days and @xmath27=127.0 days , the density waves appear in the outer - to - middle disk annuli .
the two waves are asymmetric .
when compared with the control group of artificially tilted disks ( montgomery 2009b ) , the nodal superhump fundamental and higher order harmonic frequencies ( see table 1 for this simulations values ) agree within the uncertainties .
however , the amplitude of the signals do not agree , indicating that the disk tilt in this simulation is tilted less than four degrees . to produce the stronger nodal superhump frequency in v344 lyrae ( wood et al .
2011 ) , the mass transfer rate may need to be increased to a rate more indicative of a system in outburst .
figure 3 shows yellow as a warmer emission color than orange .
the black color emphasizes relatively colder particles , which comprise the disk midplane ; the two density waves ; and the gas stream overflow . the cold , dense disk midplane along the rim is warped as shown at @xmath27=126.6 days and @xmath27=127.4 days .
the disk is also slightly titled as shown at @xmath27=126.2 days and @xmath27=127.0 days , causing the gas stream to flow over the disk for @xmath11/2 orbital period and under the disk form the latter @xmath11/2 orbital period . as suggested in the figure ,
the midplane may be warped enough that portions become part of the disk surface material .
to generate a disk tilt , simulations in montgomery ( 2012a ) and this work use a continuous infall of equal size and equal mass particles , the path of which is in the orbital plane .
the infall strikes the disk at the bright spot and unequally overflows and underflows the disk rim , resulting in the disk s angular momentum @xmath2 becoming misaligned to the stellar host s rotational angular momentum and the system s orbital angular momentum @xmath3 yet the gas stream remains normal to @xmath3 .
the differential flow paths over and under the disk may produce the disk tilt by the lift force ( montgomery & martin , 2010 ) .
once the disk is tilted as a unit , the disk precesses in the retrograde direction by tidal torques from the secondary on the tilted disk as shown in montgomery ( 2012a ) .
more of an undulating warp is generated in this simulation because the disk is simultaneously precessing in both the prograde and retrograde directions .
the changing thickness along the disk rim coupled with the continuous gas stream impact on the titled disk rim results in the discontinuous undulating warp .
early positive superhumps are generated in this simulation s artificial light curves , similar to those generated by the @xmath28 simulation in wood et al .
( 2011 ) , and thus we do not show these light curves in this work .
the observed light curves in v344 lyrae show dominating positive superhumps followed by dominating negative superhumps .
the artificial light curves generated in this work do show dominating early positive superhumps but do not generate prominent negative superhumps . instead ,
modulated superhumps are generated , the modulation being caused by the addition of disk tilt and retrograde precession . to generate dominating negative superhumps , the mass transfer rate may need to be increased to that of a system in outburst , a rate not simulated in this work .
this fine - tuning analysis is reserved for a later paper .
also reserved for a later paper is harmonic synthesis of the simulated light curves and/or harmonic analysis of the observed and simulated light curves that produced figure 1 and table 1 . modulated superhumps caused by simultaneous precessions
are not discussed in wood et al .
( 2011 ) .
although the frequencies in table 1 agree , within the uncertainties , with values reported in an artificially induced tilted disk simulation ( montgomery 2009a ) , the frequency amplitudes do not agree .
as the control group was artificially tilted five degrees , resulting in prominent negative superhumps and large amplitude frequencies , the disk tilt in this simulation is less than four degrees .
although a signal may be present in a periodogram that suggests the disk is precessing in the retrograde direction , the light curve may not show a prominent negative superhump as the disk is not tilted high enough . in this work
, we show negative superhump frequencies and retrograde precession are possible in systems below the period gap ( 2 - 3 hours ) such as in su uma systems like v344 lyrae , v503 cygni ( harvey et al . , 1995 ) , and sdss j210014.12 + 004446.0 ( olech et al .
2009 ) , all of which also show positive superhumps . in montgomery ( 2012a ) , we show negative superhump frequencies and retrograde precession are possible in systems above the period gap such as v378 pegasi ( ringwald et al . , 2012 ) .
thus negative superhumps and/or their signals should be ubiquitous above , below , and in the period gap like tt ari .
observers of cv dns should thus be seeking these ubiquitous negative superhumps and/or frequencies to confirm this prediction .
we note this simulation may not be over , and we report any additional cycles in a follow - up paper .
also , we note that although the code is set to the parameters of an su uma system , many features shown in this work should also apply to other system types such as inner annuli of protoplanetary disks that may contain a planet in a gap .
we find the following outcomes : 1 ) the observed light curves of v344 lyrae shown in wood et al .
( 2011 ) indicate cyclic times when positive superhumps dominate followed by times when negative superhumps dominate .
the simulation of v344 lyrae in this work also produces dominating positive superhumps prior to the addition of the negative superhump frequencies , although the simulation does not produce dominating negative superhumps . as such
, the mass transfer rate may need to be modified in the simulation to that more of a system in outburst to induce the dominating negative superhumps , a test not included in this work . as suggested by this work ,
although a signal in a periodogram may indicate that the disk is precessing in the retrograde direction , the negative superhump in the observed light curve may not be present .
2 ) as shown in montgomery ( 2012a ) and this work , neither a superoutburst ( which is associated with positive superhumps ) nor an outburst is required to induce disk tilt and retrograde precession ( which are associated with negative superhumps ) .
3 ) the simulation of v344 lyrae in this work produces simultaneous disk prograde and retrograde precession , which is not discussed in wood et al .
the observations ( wood et al . 2011 ) do not find a significant signal at the retrograde precession period @xmath29 3.6 days .
similarly , no significant simulated signal at the retrograde precessional period ( @xmath12.2 days ) is found in figure 1 .
4 ) although individual observed and simulated fractional periods @xmath26 and @xmath19 do not agree , nearly the same ratio of fractional period excess @xmath22 to fractional period deficit @xmath23 is obtained .
this result implies that stellar masses are not precise but are accurate in the simulation .
5 ) the frequencies in table 1 agree , within the uncertainties , with values reported in an artificially induced tilted disk simulation that is not precessing in the prograde direction ( montgomery 2009a ) .
the frequency amplitudes do not agree , however , suggesting that the disk is not tilted as high .
6 ) we show in this work that negative superhump frequencies and retrograde precession are possible in systems below the period gap ( 2 - 3 hours ) . in montgomery
( 2012a ) , we show negative superhump frequencies and retrograde precession are possible in systems above the period gap .
thus negative superhumps and/or their signals should be ubiquitous above and below the period gap .
this research was supported in part by nasa through the american astronomical society s small research grant program .
we would like to thank nsf ucf yes recipient and undergraduate student , nicholas howell , who helped analyze disk effects when the first higher order harmonic is present in tilted disks that only precess in the retrograde direction .
we also kindly thank the reviewer whose careful analysis and comments improved this work .
harvey , d. , skillman , d.r . , patterson , j. , & ringwald , f.a .
, 1995 , pasp , 107 , 551 horne , k. , 1984 , nature , 312 , 348 lasota , j .-
, 2001 , new astron .
, 45 , 449 montgomery , m.m . , 2009a , mnras , 394 , 1897 montgomery , m.m . ,
2009b , apj , 705 , 603 montgomery , m.m .
, 2012 , apj , 745 , l25 montgomery , m.m . ,
2012b , in the golden age of cataclysmic variables and related objects , f. giovannelli & l. sabau - graziati ( eds . ) , mem . sait .
83 n.2 montgomery , m.m . , 2004 , phd thesis , florida institute of technology montgomery , m.m .
& bisikalo , d.v . , 2010 , mnras , 405 , 1397 montgomery , m.m .
& martin , e.l . , 2010 , apj , 722 , 989 olech , a. , rutkowski , a. , & schwarzenberg - czerny , a. , 2009 , mnras , 399 , 465 pringle , j.e . , 1996 , 281 , 357 ringwald , f.a . , velasco , k. , roveto , j.j . , & meyers , m.e . , 2012 ,
newa , 17 , 433 shakura s .. & sunyaev r.a . , 1973 ,
a&a , 149 , 135 skillman d.r . , 1998 ,
apj , 503 , l67 simpson , j.c .
, 1995 , apj , 448 , 822 whitehurst , r. 1988 , mnras , 232 , 35 whitehurst , r. , 1994 , mnras , 266 , 35 wood , m.a . ,
montgomery , m.m . , & simpson , j.c . , 2000 , apj , 535 , l39 wood , m.a .
, still , m.d . ,
howell , s.b .
, cannizzo , j.k . ,
smale , j.p .
, 2011 , apj , 741 , 105 wood m.a .
, thomas d. , simpson j.c . , 2009 , 398 , 2110 to the disk rim . twenty orbital periods or @xmath11.2 days in v344 lyrae are shown .
as time progresses , the outer annuli of the disk cyclically change shape from highly eccentric to circular to highly eccentric . during this disk shape - change cycle ,
two asymmetric spiral density waves in the outer - to - middle annuli compress with the disk rim and advance @xmath1180@xmath30 in the prograde direction .
two shape - change cycles are needed to advance the spirals arms once . ] | warps are suspected in disks around massive compact objects . however ,
the proposed warping source non - axisymmetric radiation pressure does not apply to white dwarfs . in this letter
we report the first smoothed particle hydrodynamic simulations of accretion disks in su uma - type systems that naturally tilt , warp , and simultaneously precess in the prograde and retrograde directions using white dwarf v344 lyrae in the @xmath0 field as our model .
after @xmath179 days in v344 lyrae , the disk angular momentum @xmath2 becomes misaligned to the orbital angular momentum @xmath3 . as the gas stream remains normal to @xmath3 , hydrodynamics ( e.g. , the lift force )
is a likely source to disk tilt .
in addition to tilt , the outer disk annuli cyclically change shape from circular to highly eccentric due to tidal torques by the secondary star .
the effect of simultaneous prograde and retrograde precession is a warp of the colder , denser midplane as seen along the disk rim .
the simulated rate of apsidal advance to nodal regression per orbit nearly matches the observed ratio in v344 lyrae . |
the device used for this experiment has been described in @xcite ( see fig .
s3 ) . it is based on an undoped si / sige heterostructure with two layers of electrostatic gates .
two accumulation gates are used to induce a two - dimensional electron gas ( 2deg ) in a 12 nm wide si quantum well 37 nm below the surface and a set of depletion gates is used to form a single quantum dot in the 2deg , and a charge sensor next to this dot . the dot is tuned so it is occupied by just one electron .
two micromagnets placed on top of the accumulation gates produce a local magnetic field gradient .
the sample is attached to the mixing chamber of a dilution refrigerator with base temperature @xmath025 mk and an electron temperature estimated from transport measurements of @xmath0150 mk . for the present gate voltage configuration ,
the valley splitting , @xmath1 , is comparable to the thermal energy , @xmath2 .
microwave excitation applied to one of the gates oscillates the electron wave function back and forth in the dot , roughly along the @xmath3 axis ( fig .
because of the local magnetic field gradient @xmath40.3 mt / nm @xcite , where @xmath5 is the component of the micromagnet field gradient perpendicular to the static magnetic field @xmath6 , the electron is then subject to an oscillating magnetic field @xcite and electron spin transitions can be induced when the excitation is resonant with the spin splitting .
the spin - up probability @xmath7 in response to the microwave excitation is measured by repeated single - shot cycles ( see section iii.a of the supplemental material for details ) .
the initialization and read - out procedures require a zeeman splitting exceeding @xmath8 , which here restricts us to working at @xmath9 mt . when varying the applied microwave frequency and external magnetic field , we observe six distinct resonance peaks [ see fig .
[ fig : fig1b](a ) ] .
the two resonances labeled ( 1 ) and ( 2 ) , not resolved on this scale , are two intra - valley spin resonances , one for each of the two lowest - lying valley states that are thermally occupied @xcite .
they exhibit a @xmath10 @xmath11s and rabi frequencies of order mhz .
the two resonances labeled ( 3 ) and ( 4 ) , similarly not resolved , arise from second harmonic driving of the two intra - valley spin flip transitions .
these transitions too can be driven coherently , with rabi frequencies comparable to those for the fundamental harmonic , as we reported in @xcite .
we focus here on the resonances labeled ( 5 ) and ( 6 ) in fig .
1 , which have not been discussed before .
the frequency of resonance ( 5 ) , @xmath12 , is @xmath07 ghz lower than the fundamental intra - valley spin resonance frequencies , @xmath13 and @xmath14 . from the magnetic field dependence measured above 500 mt
, we extract a @xmath15-factor of about 1.971 @xmath16 0.002 , close to but different from the @xmath15-factors for resonances ( 1 ) and ( 2 ) ( 1.99 @xcite ) . the line width ( fig .
[ fig : fig5](a ) inset and fig .
s4 ) is almost ten times larger than that for the intra - valley resonances , giving a correspondingly shorter @xmath17 of around 100 ns .
around 500 mt , resonance ( 5 ) changes in a way reminiscent of level repulsions and transitions into resonance ( 6 ) ( see fig . [
fig : fig1b](b ) ) . without the change in slope , resonance ( 5 ) would have crossed resonances ( 3 ) and ( 4 ) ; however , the latter do not show any sign of level repulsion and continue their linear dependence on magnetic field .
we interpret these puzzling observations starting with fig .
[ fig : fig1b](d ) .
two sets of zeeman split levels are seen , separated by the energy of the first excited valley - orbit state .
the green ( 1 ) [ ( 2 ) ] and two blue ( 3 ) [ ( 4 ) ] arrows show driving of spin transitions via the fundamental and second harmonic respectively , for the valley - orbit ground [ excited ] state .
we identify resonance ( 5 ) with the transition indicated with the red arrow in which both spin and valley(-orbit ) flip .
it has the same field dependence as resonance ( 1 ) , but ( above 500 mt ) it is offset from resonance ( 1 ) by a fixed amount , which as we can see from fig .
[ fig : fig1b](d ) , is a measure of the valley - orbit splitting , @xmath1 . resonance ( 6 ) is a three photon process in which both spin and valley - orbit states flip .
as we discuss below , hybridization between ( 5 ) and ( 6 ) is possible at magnetic fields @xmath18 , for which the photons shown by the corresponding arrows in fig .
[ fig : fig1b](d ) have the same energy .
we now introduce a simple model hamiltonian that can be used to understand the observed spectroscopic response . this model explains the presence of both the first and second harmonic driven spin resonance as well as the observed inter - valley spin resonance .
we show that resonances such as those observed in fig .
[ fig : fig1b](a , b ) are generic features of a strongly driven four - level system composed of two orbital levels and two spin levels in which there is a coupling between the orbital levels , such as a tunnel coupling . for our case , it is natural to associate the orbital levels with two different valley - orbit states ( see section i of the supplemental material for details ) .
our analysis builds upon the theory proposed by rashba @xcite .
when a spin qubit is driven at a frequency @xmath19 , it responds at one or more frequencies @xmath20 , which may be the same as @xmath19 , but may also be different ( see fig .
[ fig : theory2](a ) ) .
spin resonance is observed if ( i ) the spin is flipped , and ( ii ) @xmath21 , where @xmath22 is the zeeman splitting , @xmath15 is the land @xmath15-factor in silicon , @xmath23 is the bohr magneton , and @xmath24 is the total magnetic field . in electric dipole spin resonance ( edsr ) , the spin flip requires a physical mechanism for the electric field to couple to the spin , such as spin - orbit coupling @xcite . in our experiment ,
an effective spin - orbit coupling due to the strong magnetic field gradient from the micromagnet is the mechanism responsible for spin flips @xcite .
hence , we can say that edsr and its associated spin dynamics provide a tool for observing the mapping @xmath25 .
however , as discussed in @xcite , edsr does not determine the mapping ; determining the resonant frequencies @xmath20 requires including the essential non - linearity in the system , which in this case resides in the orbital sector of the qubit hamiltonian .
we therefore focus on the dynamics of the orbital sector of the hamiltonian ; the mechanism for spin flips is included perturbatively after the charge dynamics have been characterized .
the exact orbital hamiltonian is difficult to write down from first principles , since it likely involves both orbital and valley components @xcite , and depends on the atomistic details of the quantum well interface @xcite . nonetheless , the features of the resonances in fig .
[ fig : fig1b ] emerge quite naturally using a model with one low - lying orbital excited state . referring to fig .
[ fig : fig1b](c ) , in this model , the hamiltonian for the orbital sector is described by a simple two - state hamiltonian , which we write as @xmath26 here , @xmath27 is a detuning parameter , @xmath28 is the tunnel coupling between the generic basis states labeled @xmath29 and @xmath30 , and @xmath31 and @xmath32 are pauli matrices .
we consider a classical ac drive , applied to the detuning parameter : @xmath33 if the quantum dot confinement were purely parabolic , then changing the detuning would not affect the energy splitting between the eigenstates .
however , any nonparabolicity in the dot , which is unavoidable in real devices , will cause the energy splitting to depend on the detuning and will yield a nonlinear response to the driving term , eq .
[ eq : epstm ] . in our hamiltonian
, this effect enters via the tunnel coupling @xmath28 , which causes the qubit frequency to depend on @xmath34 .
our goal is to determine the response of the two - level system to this @xmath34 .
we assume that the basis states are coupled by the applied electric field because they have different spatial charge distributions , and study the time evolution of the dipole moment of the ground state @xmath35 , defined as @xmath36 .
\label{eq : p0m}\ ] ] here , @xmath37 is the distance between the charge in states @xmath29 and @xmath30 is the lateral separation between the two sides of the double quantum dot . for an orbital qubit
, @xmath37 is the lateral separation of the center of mass of the two orbital states . for a pure valley qubit
, @xmath37 is the vertical separation of the even and odd states ( @xmath00.16 nm ) . for a complicated system like a valley - orbit qubit with interface disorder
, @xmath37 will have lateral and vertical components , with the lateral component being usually much larger than the vertical one . in this last case
the exact length @xmath37 will depend on the specifics of the interface disorder ; a reasonable guess would be @xmath380.5 - 5 nm ( see , e.g. ref .
[ 28 ] ) . ] .
rashba has studied hamiltonian ( [ eq : hm ] ) perturbatively in the regime of weak driving and high excitation frequency @xcite . in sections
i and ii of the supplemental material , we present a detailed exposition of our extensions of these investigations into the strong driving regime relevant to resonances ( 5 ) and ( 6 ) . we find that driving this transition involves non - adiabatic processes @xcite whereby the orbital state gets excited , in contrast to the subharmonics reported in @xcite , which as we show here involve only adiabatic processes in the charge sector . here
, we present the results of numerical simulations in this regime and show that the results are consistent with the main features observed experimentally .
the dynamical simulations are performed by setting @xmath39 and solving the schrdinger equation @xmath40 for eqs .
( [ eq : hm ] ) and ( [ eq : epstm ] ) and computing @xmath41 as defined in eq .
( [ eq : p0 m ] ) for a fixed driving frequency @xmath19 ( b ) , we use time steps in the range 0.061 - 0.073 . ] .
the resulting @xmath41 is fourier transformed , yielding a @xmath42 whose peaks reflect the resonant response . finally , we smooth @xmath42 by convolving it with a gaussian of width 0.025 ( in order to take into account noise , which is averaged in the experiment ) . because of the spin - orbit coupling , peaks in @xmath42 correspond to frequencies at which an ac magnetic field is generated , so spin flips will occur at a zeeman frequency given by @xmath43 .
the experiment measures the probability of a spin flip as a function of magnetic field ; via the edsr mechanism , resonances in this probability therefore occur when the peak locations in @xmath42 satisfy @xmath44
. figs .
[ fig : theory2](b ) and ( c ) show the results of our simulations for @xmath42 as a function of both driving frequency @xmath19 and response frequency @xmath20 over a range of parameters analogous to those shown in fig .
[ fig : fig1b ] . from top
to bottom , the dashed blue lines correspond to the fundamental resonance ( 1,2 ) , and the first two subharmonics ( 3,4,9 ) , which are all visible in the simulations .
we also see a resonance , labeled ( 5 ) , running parallel to the fundamental resonance , and a resonance labelled ( 6 ) running parallel to ( 9 ) ( the other resonances are discussed in section i.c of the supplemental material ) .
the upper inset of fig .
[ fig : theory2](b ) highlights the particular resonances in the main figure that should be compared to the experimental data shown in fig .
[ fig : fig1b](a , b ) . an interesting feature of the resonances , observed both in the experiments and theory ( figs . [
fig : theory2](b - d ) ) , is the apparent ` level repulsion ' between resonance lines ( 5 ) and ( 6 ) that takes place near @xmath45 mt .
this magnetic field value is much higher than @xmath46 mt , where the anti - crossing between the states @xmath47 and @xmath48 is expected to occur @xcite ( see the red dashed line in fig .
[ fig : fig1b](d ) ) , but which is outside our measurement window , see above .
instead the observed ` level repulsion ' has a purely dynamical origin , as demonstrated by the fact that the anti - crossing is suppressed in fig .
[ fig : theory2](c ) , where the simulation parameters are identical to fig .
[ fig : theory2](b ) , except for a smaller driving amplitude @xmath49 . in sec .
ii of the supplemental material , we develop a dressed - state theory to describe these strong - driving effects . in this formalism , the quasiclassical driving field of eq .
[ eq : epstm ] is replaced by a fully quantum description of the photon field and its coupling to the ( valley)-orbital hamiltonian of the quantum dot .
the resulting dressed eigenstates describe the hybridized photon - orbital levels , and more generally , the hybridization of orbital , photon , and spin states . in this way , resonances ( 1 ) and ( 2 ) in fig . [ fig : theory2](c ) correspond to single - photon spin flips , while resonances ( 3 ) and ( 4 ) correspond to two - photon spin flips . resonance
( 5 ) involves both a spin flip and a valley - orbit excitation and is parallel to resonance ( 1,2 ) , indicating that it is a single - photon process .
resonance ( 6 ) is parallel to ( 9 ) , indicating that it is a three - photon process .
the physical mechanisms of the resonances are also indicated in fig .
[ fig : fig1b](d ) . in principle
, coupling occurs between all of the dressed states due to the effective spin - orbit coupling in our edsr experiment . in practice however , the orbital - rabi frequency is two orders of magnitude larger than the spin - rabi frequency , so mode hybridization is only observed between resonance ( 5 ) and ( 6 ) , resulting in the level repulsion .
the magnitude of this repulsion provides a convenient way to determine the orbital - rabi frequency , which can not be measured directly due to the fast dephasing of the excited valley - orbit state ( see section iii ) . in sec .
ii of the supplemental material , we estimate this rabi frequency to be about 0.2 ghz .
we now examine the possible origin of the ten times larger line width of resonance ( 5 ) compared to that of the pure spin - flip resonances ( 1 ) and ( 2 ) . given the partial valley nature of transition ( 5 ) and the strong valley - orbit coupling that is typical of si / sige quantum dots @xcite , a plausible candidate decoherence mechanism for this transition is electric field noise . in order to study the sensitivity to electric fields of the respective transitions ,
we show in fig . [
fig : fig5](a ) the dependence of the frequency of resonances ( 1 ) and ( 5 ) on the voltage applied to one of the quantum dot gates , @xmath50 . clearly , resonance ( 5 ) exhibits a much greater sensitivity to gate voltage than resonance ( 1 ) : @xmath0 18.5 mhz / mv for @xmath12 , versus @xmath0 0.5 mhz / mv for @xmath13 .
we also notice that the two resonance shifts as a function of @xmath50 have opposite sign , which indicates that different mechanisms are responsible . for resonance ( 1 ) , we believe that the dominant effect of the electric field is the displacement of the electron wave function in the magnetic field gradient from the micromagnets @xcite ( see section iii.b of the supplemental material ) .
this effect also contributes to the frequency shift of resonance ( 5 ) , but presumably it is masked by the change in valley - orbit splitting ( @xmath1 ) resulting from the displacement of the electron wave function in the presence of interface disorder @xcite .
for instance , moving the electron towards or away from a simple atomic step at the si / sige interface leads to a change of the valley - orbit energy splitting , as shown by the results of numerical simulations reported in fig .
[ fig : fig5](b , c ) @xcite .
as expected , the simulations predict a minimum in the valley - orbit splitting when the wave function is centered around the atomic step , but interestingly it does not vanish , i.e. the opposite signs for the valley - orbit splitting left and right of the atomic step do not lead to complete cancellation ( see section iv of the supplemental material for a more detailed description ) .
the @xmath035 times greater sensitivity of the spin - valley transition frequency to electric fields may contribute to its ten times larger line width compared to the intra - valley spin transition .
the line width of the intra - valley spin transition is believed to be dominated by the 4.7@xmath51 @xmath52si nuclear spins in the host material @xcite .
the nuclear field also affects the spin - valley transition , but obviously only accounts for a small part of the line width here .
we propose that the dominant contribution to the line width of resonance ( 5 ) is low - frequency charge noise .
although not definitive , some evidence for this interpretation is found in fig . [ fig : fig5](d ) , which shows a scatter plot of @xmath12 and one of the dot - reservoir tunnel rates , simultaneously recorded over many hours ( see section iii.c of supplemental material for a more detailed description of the measurement scheme ) .
the dot - reservoir tunnel rate serves as a sensitive probe of local electric fields , including those produced by charges that randomly hop around in the vicinity of the quantum dot ( see fig .
s6 ) @xcite .
the plot shows a modest correlation between the measured tunnel rate and @xmath12 , suggesting that the shifts in time of both quantities may have a common origin , presumably low - frequency charge noise . in this case
, we can also place an upper bound on @xmath53 permitted by charge noise for the intra - valley spin transitions ( 1,2 ) in the present sample ( given the specific magnetic field gradient reported in @xcite ) .
indeed , due to the micro magnet induced gradient in the local magnetic field parallel to @xmath6 , the pure spin transitions are also sensitive to charge noise .
given that @xmath54 110 ns for transition ( 5 ) and the ratio of @xmath0 35 in sensitivity to electric fields , charge noise in combination with this magnetic field gradient would limit @xmath53 to @xmath55 @xmath11s for transitions ( 1,2 ) .
it is important to note that this is not an intrinsic limitation , as the stray field of the micro magnet at the dot location can be engineered to have zero gradient of the longitudinal component , so that to first order charge noise does not affect the frequency and @xmath53 of transitions ( 1,2 ) . at the same time , a strong gradient of the transverse component can be maintained , as is necessary for driving spin transitions . besides its strong sensitivity to static electric fields , we report a surprising dependence of the frequency of resonance ( 5 ) on microwave driving power [ fig . [
fig : fig7](b , c ) ] .
increasing the driving power , the resonance not only broadens but also shifts in frequency , as in an a.c .
stark shift @xcite .
this dynamical evolution is very different from the case of the intra - valley spin resonance , which is power broadened but stays at fixed frequency @xcite .
this frequency shift is , at least for a limited microwave power range , in line with the dynamical level repulsion captured by eq .
s42 , where @xmath56 expresses the energy splitting between resonance ( 5 ) and its asymptote , @xmath57 .
this relation is verified in fig .
[ fig : fig7](c ) for microwave powers of @xmath58 dbm . finally , we attempt to drive coherent oscillations using resonance ( 5 ) at high applied microwave power , recording the spin excited state probability as a function of the microwave burst time .
oscillations are not visible , indicating that the highest rabi frequency we can obtain for resonance ( 5 ) is well below the corresponding @xmath59 of 110 ns .
this is consistent with our estimate that the rabi frequency is of the order of 10 khz , based on the magnitude of the dynamical level repulsion seen in fig .
[ fig : fig1b](a , b ) and the derivation in the section ii.b of supplemental material .
despite its simplicity , the electrical driving of a single electron confined in a single quantum dot can produce a complex spin resonance energy spectrum .
this particularly applies for quantum dots realized in silicon , where the presence of the excited valley - orbit state , close in energy and strongly coupled to the ground state , introduces a substantial non - linearity in the system response to microwave electric fields .
this allows us to observe a transition whereby both the spin and the valley state are flipped at the same time .
we demonstrate how both static external electric fields and electrical noise influence the frequency of this inter - valley spin transition , dominating its coherence properties . much of the dynamics of the spin and valley transitions can be captured in a semi - classical picture , including driving using higher harmonics exploiting non - linearities .
however , under intermediate or strong driving , new phenomena emerge that can not be easily explained except in terms of dressed states that fundamentally involve a quantum mechanical coupling between photons and orbital or spin states . here , we have provided experimental and numerical evidence for the existence of such dressed states of photons and valley - orbit states at strong driving .
we have further estimated the strength of this valley - orbit to photon coupling by comparing our analytical theory to the experiments .
this work provides important experimental and theoretical insight in the role of inter - valley transitions for controlling spin dynamics in silicon based quantum dots .
it also highlights the limitations of valley - based qubits in the presence of strong valley - orbit coupling , due to their sensitivity to electrical noise .
we acknowledge r. schouten and m. j. tiggelman for technical support and the members of the delft spin qubit team for useful discussions .
research was supported by the army research office ( w911nf-12 - 0607 ) , a european research council synergy grant , and the dutch foundation for fundamental research on matter .
e.k . was supported by a fellowship from the nakajima foundation . the development and maintenance of the growth facilities used for fabricating samples
is supported by doe ( de - fg02 - 03er46028 ) , and this research utilized nsf - supported shared facilities at the university of wisconsin - madison .
this first section of the supplemental material presents a theoretical discussion of the model used to understand the electric dipole spin resonance experiments presented in the main text .
we show that experimental features are generic features of a driven four - level system , comprised of two valley - orbit and two spin degrees of freedom , with a tunnel coupling between the orbital states .
the model exhibits conventional resonances , including both fundamental and higher harmonics , as well as novel resonances involving photonically dressed orbital states .
we develop an analytical model that describes the key features of the hybridized dynamical states , based on a simple , 4d dressed - state hamiltonian , and we use this model to determine the rabi frequency for orbital excitations by fitting to the experimental data . the main text reports edsr measurements on an ac - driven spin qubit in a silicon quantum dot .
the system is driven in a microwave regime that is intermediate between weak and strong driving , allowing us to observe semiclassical phenomena , such as conventional spin resonances and spin - flip photon assisted tunneling ( pat ) @xcite , which have been observed previously in quantum dots .
however , the experiments also probe new and intriguing phenomena that are purely quantum mechanical , involving the hybridization of photons , orbitals and spin .
such effects are most conveniently described as dressed states . by developing a formalism that is appropriate for our system
, we can describe the photonically dressed states as pseudospins , and explain the hybridization effects by means of simple 2d and 4d hamiltonians .
moveover , although coherent oscillations of the photon - orbital - spin pseudospins are not observed in our experiments ( in contrast with photon - spin pseudospins , whose coherent oscillations were reported in ref .
@xcite ) , the dressed - state theory still allows us to extract the pat rabi frequency .
our analysis is broken into two parts . in sec .
[ sec : model ] , we describe a simple model for edsr , and provide an intuitive discussion and overview of the resonances observed in the experiments . in sec .
[ sec : dressed ] , we develop a more technical , dressed - state formalism to describe the photonic dressing of the orbital and spin states in our system , which we use to explain and fit the experimental data . in sec .
[ sec : additional measurements ] we provide additional information on the experimental setup and on the measurements . in sec .
[ sec : simulations ] we provide a more exhaustive description of the tight binding simulation realized in order to capture the evolution of the valley - orbit energy splitting as a function of the electron position in the sige / si / sige quantum well ( reported in fig .
when a spin qubit is driven at a frequency @xmath19 , it responds at one or more frequencies @xmath20 , which may be the same as @xmath19 , but may also be different .
spin resonance is observed if ( i ) the spin is flipped , and ( ii ) @xmath21 , where @xmath60 is the zeeman splitting , @xmath61 is the land @xmath15-factor in silicon , @xmath23 is the bohr magneton , and @xmath62 is the applied magnetic field .
the spin flip requires a physical mechanism , such as spin - orbit coupling in edsr @xcite .
indeed , the strong magnetic field gradient of the micromagnet introduces such a coupling @xcite .
rashba has suggested that the mapping @xmath25 occurs entirely within the orbital sector , and that edsr simply provides a tool for observing this mapping @xcite . in this supplemental section ,
we adopt rashba s orbital - based model as our starting point .
several theoretical explanations have been put forth to explain strong - driving phenomena in edsr , such as the generation of higher harmonics @xcite , and other , more exotic effects such as an even / odd harmonic structure @xcite . for silicon dots ,
the exact orbital hamiltonian is difficult to write down from first principles , since it involves both orbital and valley components @xcite , and depends on the atomistic details of the quantum well interface @xcite .
nonetheless , the dynamics can be understood using a minimal model for investigating the physics of the orbital sector , given by the two - level system @xmath63 where @xmath31 and @xmath32 are pauli matrices .
the basis states in this model may correspond to excited orbitals in a single quantum dot @xcite , valley states in a single dot @xcite , localized orbital states in a double dot @xcite , or a hybridized combination of these systems @xcite . in each case , the two basis states have distinct charge distributions that can support edsr through the mechanism described by tokura et al .
@xcite equation ( [ eq : h ] ) is , in fact , the most general theoretical model of a two - level system .
mapping the theoretical model onto a real , experimental system requires specifying the dependence of the effective detuning parameter @xmath27 and the effective tunnel coupling @xmath28 on the voltages applied to the top gates for a given device . to take a simple example that is relevant for silicon quantum dots , we consider such a mapping for the basis comprised of valley states . in this case
, @xmath27 represents a valley splitting , which can be tuned with a vertical electric field , while @xmath28 represents a valley coupling , which can depend on the interfacial roughness of the quantum well and other materials parameters @xcite .
the centers of mass of the charge distributions of the basis states in this example differ by about a lattice spacing in the vertical direction .
it is likely that the silicon dots in the experiments reported here experience an additional valley - orbit coupling , which induces a lateral variation of the charge distributions . in this case
, the gate voltage dependence of @xmath27 and @xmath28 is quite nontrivial ; however eq .
( [ eq : h ] ) should still provide a useful theoretical description .
for convenience in the following analysis , we adopt the picture of a charge qubit , for which the basis states correspond to left " ( @xmath29 ) and right " ( @xmath30 ) .
the eigenstates of @xmath64 are then given by @xmath65 and the energy splitting is given by @xmath66 .
to begin , we consider a classical ac driving field , applied to the detuning parameter : @xmath67 our goal is to determine the response of the two - level system to this @xmath34 .
in contrast with rashba , we will not limit our analysis to the perturbative regime in which @xmath68 , since the experiments reported here are not in that regime . to proceed
, we note that the basis states couple differently to the applied electric field because they have different spatial charge distributions . we specifically consider the time evolution of the dipole moment of the ground state @xmath35 , defined as @xmath69 , \label{eq : p0}\ ] ] where @xmath37 is the distance between the center of mass of the charge in states @xmath29 and @xmath30 . throughout the discussion that follows , when connecting the output of the model to the experimental data , we will identify @xmath70 with the valley - orbit splitting , @xmath1 .
two fundamentally different types of response to the driving are observed in @xmath41 : adiabatic vs. nonadiabatic .
we first consider adiabatic processes .
the adiabatic eigenstates of the hamiltonian , given in eqs .
( [ eq:0 ] ) and ( [ eq:1 ] ) yield a corresponding dipole moment , computed from eq .
( [ eq : p0 ] ) , of @xmath71 = -\frac{el\epsilon}{2e_{01 } } .
\label{eq : p0adiabatic}\ ] ] substituting eq .
( [ eq : epst ] ) into eq .
( [ eq : p0adiabatic ] ) , we see that the adiabatic response of @xmath41 is a nonlinear function of @xmath72 . for example , in the weak driving limit , we can expand eq .
( [ eq : p0adiabatic ] ) to second order in the small parameter @xmath73 , giving @xmath74 here , @xmath75 denotes the energy splitting when @xmath76 .
we see that the nonlinear dependence of the dipole moment on @xmath49 translates into a series of response terms at frequencies @xmath77 . since spin resonances only occur when @xmath21 , these should be observed as subharmonics of the fundamental zeeman frequency : @xmath78 . for the two - level system defined by eq .
( [ eq : h ] ) , we note that all resonances depend on the presence of a coupling between the basis states , @xmath28 . in fig .
1(d ) of the main text , the fundamental resonance condition and its second harmonic are sketched as the green and blue vertical arrows respectively .
we note that adiabatic resonances can be observed in both the strong and weak driving regimes , although the higher harmonics may be suppressed for weak driving , as consistent with eq .
( [ eq : p0expand ] ) .
we next consider nonadiabatic , or landau - zener ( lz ) processes @xcite , which always involve an excitation ( or deexcitation ) of an orbital state .
for the two - state hamiltonian of eq .
( [ eq : h ] ) , lz excitations can arise in two ways : ( i ) a sudden pulse of a control parameter ( e.g. , @xmath27 ) , or ( ii ) strong periodic driving of the control parameter [ e.g. , eq .
( [ eq : epst ] ) ] . in the first case , the qubit
is suddenly projected onto a new adiabatic eigenbasis , which induces a larmor response at the frequency @xmath79 . in the second case , fast periodic excursions of @xmath34 , which does not need to occur at the resonant frequency , can also produce a similar response .
such ` conventional ' lz excitation processes are indicated by the black vertical arrow labeled ( 7 ) in fig .
1(d ) of the main text .- field .
the process indicated in the figure implies an orbital excitation combined with a spin flip , which we discuss below . ] in the experiments reported here , simple lz orbital excitations are undetectable , because the excitation energy @xmath70 is smaller than the fermi level broadening .
however , the edsr spin - flip mechanism allows us to generate and detect more complex processes at higher energies .
the resonance conditions indicated by the red and yellow vertical arrows in fig .
1(d ) are particularly important for these experiments ; when combined with a spin flip , the lz orbital excitation corresponds to resonances ( 5 ) and ( 6 ) shown in figs .
1(a ) and ( b ) .
resonances of this type occur when @xmath80 , corresponding to the driving frequencies @xmath81 .
since these resonance lines are parallel to the conventional spin resonances , rashba has called them satellites " @xcite , although he studied a different driving regime for which no satellites were observed in the experiments .
there are two main theoretical approaches for describing such nonadiabatic phenomena .
since the orbital excitation involves an lz process , it must be described quantum mechanically .
however , the ac driving field can be described with a semiclassical , adiabatic theory like the one described in the previous section . in this approach
, the qubit gains a stckelberg phase " @xcite each time it passes through an energy level anticrossing , resulting in interference effects that can be described using standard lzs theory @xcite . ) ; however , the interference effects observed in our simulations occur even without passing through the anticrossing . ]
the alternative theoretical approach , described below , uses a dressed state formalism that encompasses all of the same interference effects , and plays a key role in understanding the detailed behavior of resonances ( 5 ) and ( 6 ) . in this case , the process can be described fully quantum mechanically as an orbital excitation combined with a photon absorption . in either description ,
the orbital excitation is followed by a spin - flip caused by the orbital dynamics .
results of numerical simulations of eq .
( [ eq : h ] ) are presented in fig .
2(b - d ) of the main text .
here , we discuss the physical interpretation of these results in more detail . figure 2 in the main text shows the results of our simulations for @xmath42 as a function of both driving frequency @xmath19 and response frequency @xmath20 over a range of parameters analogous to those in fig .
figures 2(b)-2(d ) correspond to three different driving amplitudes , with panel ( d ) chosen to match the experimental results in fig .
the dashed blue lines correspond to the conventional esr signals ( the fundamental resonance , and the first two harmonics , top to bottom ) , which are all visible in the simulations . in fig .
2(c ) ( reproduced in fig .
[ fig : theorylabels ] ) , these are labeled ( 1,2 ) , ( 3,4 ) , and ( 9 ) .
the inset on the left - hand side of fig .
2(b ) highlights the particular resonances that should be compared to the experimental data in fig .
in addition to conventional esr resonances , several lz resonances can be seen in the simulations .
three lz resonances , labeled ( 5 ) , ( 6 ) and ( 8) in fig .
2(c ) , are observed between the fundamental resonance and the second harmonic . at high frequencies ,
we note that resonance ( 8) has the same slope as the second harmonic ( 3,4 ) , while resonance ( 5 ) has the same slope as the fundamental ( 1,2 ) . more explicitly , these lines correspond to the resonance conditions given by @xmath82 and @xmath83 , respectively .
the latter resonance is clearly visible in the experiment , and corresponds to the resonance labeled ( 5 ) at high frequencies in fig . 1 .
resonance ( 6 ) has the same slope as the third harmonic ( 9 ) and is visible in the experiment at the lowest frequencies .
its frequency is given by @xmath84 .
an interesting feature of the lz resonances , which is also observed in the experiments ( see fig
. 1(b ) ) , is the apparent level repulsion between the resonance peaks , for instance between ( 5 ) and ( 6 ) .
several of these anticrossings are seen in fig . 2 .
the effect is purely dynamical , as demonstrated by the fact that the anticrossing is suppressed in going from panels ( b ) to ( d ) , as consistent with the smaller driving amplitudes . in the dressed - state formalism
, these dynamical modes correspond to resonances of the hybridized photon - orbital states , where the state composition changes suddenly .
several excitations are observed in the simulations of fig . 2
that are not also observed in the experiments .
for example , the vertical line labeled ( 7 ) in fig .
2(c ) corresponds to the nonresonant ( larmor ) lz excitation discussed in sec .
[ sec : nonadiabatic ] .
this excitation occurs at an energy @xmath85 that is too small to be detected experimentally .
resonance ( 8) is suppressed in the simulations , compared to resonance ( 5 ) , and it is not observed at all in the experiments . a level anticrossing is observed between resonances ( 7 ) and ( 8) in the simulations , which also occurs outside the experimental measurement window .
several other features can be observed in the lower - right corner of fig .
2(b ) , which are very faint , and can also be classified according to the resonance scheme discussed above .
previous approaches to strong driving have assumed that the driving amplitude @xmath49 and the driving frequency @xmath86 are both much larger than the minimum energy gap @xmath28 @xcite
. however , the simulation results shown in fig . 2 , which provide a good descrption of the experiments in fig . 1 ,
suggest that our system falls into an intermediate regime , where the dimensionless parameters all have similar magnitudes , @xmath87 , and @xmath88 .
below , we develop a new dressed state formalism that is appropriate for such an intermediate parameter regime , while treating @xmath49 as weakly perturbative .
this formalism successfully describes all the resonance features observed in figs . 1 and 2 .
in keeping with our emphasis on orbital physics , we will first consider only the charge and photon sectors , and introduce the spin afterwards .
the time - dependent hamiltonian given by eqs .
( [ eq : h ] ) and ( [ eq : epst ] ) is semiclassical ( i.e. , the electric field is treated classically ) .
the same hamiltonian was investigated in ref .
@xcite , in the context of strong driving .
a fully quantum version of the same problem was considered in refs .
@xcite and @xcite , also in the context of strong driving .
the solutions to the semiclassical and quantum hamiltonians are identical , suggesting that the two approaches are interchangeable .
the quantum version is more elegant however , and it is more directly compatible with the dressed state formalism . we therefore adopt the quantum hamiltonian as our starting point , rather than eq .
( [ eq : h ] ) .
the full quantum hamiltonian is given by @xmath89 here , @xmath90 and @xmath91 are the creation and annihilation operators for microwave photons of frequency @xmath19 , and @xmath15 is the electron - photon coupling constant ( we ignore the vacuum fluctuations of the electric field ) .
the semiclassical and quantum hamiltonians are related through the correspondence @xcite @xmath92 where @xmath93 is the photon number operator .
we expect that @xmath94 for gate - driven microwave fields .
the ac detuning amplitude can be related to the ac voltage applied to a top gate , @xmath95 , through the relation @xmath96 , where @xmath97 is the capacitance between the dot and the top gate and @xmath98 is the total capacitance of the dot @xcite .
it is important to note the form of the coupling term in eq .
( [ eq : quantumh ] ) , which arises because the ac signal is applied to the detuning parameter .
the resulting hamiltonian can not be solved exactly , and several approaches have been used to simplify the problem . in the weak driving limit , it is common to combine a unitary transformation with a rotating wave approximation to reduce the coupling in eq .
( [ eq : quantumh ] ) to a transverse form : @xmath99 @xcite .
the latter has been used to describe the interactions between a dot and a microwave stripline @xcite .
alternatively , in the strong driving limit , it is common to retain the longitudinal coupling form of eq .
( [ eq : quantumh ] ) while treating @xmath28 as a perturbation .
this approach has been used to describe the dynamics of a dot in a gate - driven microwave field @xcite . in the intermediate driving regime , which is most relevant for the experiments reported here
, neither of these approaches is appropriate . as discussed in supplemental sec .
[ sec : nonadiabatic ] , the lz satellite peaks in fig .
2 are displaced vertically from the conventional spin resonances ( the center peaks ) by @xmath100 .. ] for example , the satellite peak labeled ( 5 ) in fig .
2(c ) is displaced by @xmath101 from the fundamental resonance ( 1,2 ) , while the satellite labeled ( 8) is displaced by @xmath102 from the conventional second harmonic ( 3,4 ) .
this suggests that we should first diagonalize the uncoupled ( @xmath103 ) hamiltonian , giving @xmath104 , where @xmath105 is the pauli matrix along the orbital quantization axis .
the eigenstates of @xmath106 are labeled @xmath107 , where @xmath108 and @xmath109 , and the energy eigenvalues are given by @xmath110 .
the full hamiltonian becomes @xmath111 where we have defined @xmath112 . to dress the orbital states , we consider the hamiltonian @xmath113 , whose coupling is purely longitudinal .
this hamiltonian has exact solutions , given by @xcite , @xmath114{\ensuremath{|{c}\rangle}}{\ensuremath{|{n}\rangle } } , \label{eq : dressed}\ ] ] with the corresponding energies @xmath115 here , the subscript @xmath116 indicates a longitudinal dressed state .
evaluating the full hamiltonian @xmath117 in the dressed state basis , we obtain diagonal terms given by eq .
( [ eq : dressede ] ) , and off - diagonal terms given by @xmath118 where @xmath119 \\ \times
\frac{(-1)^{m+p-1}(2\alpha)^{m+2p-1}\sqrt{n!(n+m)!}}{p!(m+p-1)!(n - p ) ! } , \label{eq : nu}\end{gathered}\ ] ] and @xmath120 we note that eq .
( [ eq : me1 ] ) follows from the facts that ( i ) the orbital character of @xmath121 is the same as @xmath107 , and ( ii ) @xmath122 .
equations ( [ eq : me2 ] ) and ( [ eq : me3 ] ) are obtained using standard techniques @xcite . we can obtain an approximate form for the rabi frequency @xmath123 for orbital oscillations . since our simulations indicate that @xmath124 , and we have assumed that @xmath94 , eqs .
( [ eq : gcouple ] ) and ( [ eq : alpha ] ) imply that @xmath125 . since the distribution of photon number states comprising @xmath121 is sharply peaked around @xmath126 , we can approximate @xmath127 , yielding @xmath128 \frac{(-2\alpha)^{m-1}}{(m-1)!}\langle n\rangle^{m/2 } .
\label{eq : nulargen}\ ] ] finally , in the limit @xmath129 , we find @xmath130 thus , ignoring geometrical factors , we obtain the general scaling relation @xmath131 .
the dressed state formalism is convenient for identifying and analyzing resonant phenomena .
for example , the resonance condition for an @xmath132-photon orbital excitation is simply given by @xmath133 , or @xmath134 .
when this condition is satisfied , the full hamiltonian approximately decouples into a set of 2d manifolds given by @xmath135 where we have subtracted a constant diagonal term . for a system
initially in the orbital state @xmath136 , eq . ( [ eq : h2d ] ) describes oscillations between @xmath136 and @xmath137 , with the rabi frequency @xmath123 .
these resonances are referred to as photon assisted tunneling ( pat ) , and are well known in quantum dots @xcite . in the present context , tunneling " refers to the generalized tunneling parameter @xmath28 in eq .
( [ eq : h ] ) , which represents the coupling between the different orbital states .
following rashba @xcite , we have argued that the experimental edsr spectrum can be explained entirely within the orbital sector .
the argument is summarized as follows : ( 1 ) the orbital ( two - level ) system is driven at a frequency @xmath19 ; ( 2 ) nonlinearities in the orbital system generate responses at multiples of @xmath19 , corresponding to multiphoton processes ; ( 3 ) strong driving modifies the response by generating additional satellite peaks at frequencies @xmath138 ; ( 4 ) the response frequencies in the orbital sector are transferred directly to the spin sector , where they become the driving frequencies for spin rotations via the edsr mechanism . by associating the response frequency @xmath20 in the orbital system with the characteristic frequency @xmath139 in the spin system , we were able to obtain excellent agreement with the experiments , without invoking any other assumptions about the spin physics . in an alternative explanation of the spin harmonics
, we could consider unrelated nonlinearities in the spin sector , independent of the orbital sector .
two arguments against such an approach are as follows .
first , since nonlinearities occur in the orbital system anyway , occam s razor suggests using the simplest , weak - driving model for the spin dynamics .
second , while the spin rabi frequencies obtained in our previous edsr experiments are @xmath140 mhz @xcite , the rabi frequencies for orbital oscillations are @xmath141 mhz ( see supplemental sec . [
sec : hybrid ] ) ; it is therefore very reasonable to consider strong , photon - mediated driving for the orbital system , but weak driving for the spin system , mediated by the orbital dynamics .
we can formalize such a simple - spin " model by adopting a standard spin hamiltonian , @xmath142)\cdot \bm \tau .
\label{eq : hspin}\ ] ] here , @xmath143 is a superoperator representing the inhomogeneous magnetic field , @xmath144 is the electron position operator , which is related linearly to the dipole moment operator in eq .
( [ eq : p0 ] ) , and @xmath145 is the pauli operator for a ( spin-1/2 ) electron .
the explicit dependence of @xmath144 on @xmath32 indicates that @xmath144 acts only on the orbital variables .
we emphasize that there is no direct coupling between spins and photons in this model .
the quantization axis for the spins is defined as the average of the magnetic field operator : @xmath146 with the corresponding zeeman energy , @xmath147 .
the deviation of @xmath148 from its average value ( e.g. , due to driving ) is given by @xmath149 clearly , @xmath150 contains information about the magnetic field gradient , which is the source of spin - orbit coupling in edsr . transforming @xmath151 to the quantization frame yields @xmath152 , where @xmath153 and @xmath154 .
note that we have identified @xmath155 as the low - energy spin state , in analogy with the orbital system . in the simple - spin model , the photons dress the orbital state but not the spin state .
we may therefore simply extend the longitudinal dressed states to include the bare spin : @xmath156 . in this basis , the diagonal elements of the full hamiltonian are given by @xmath157 note that we have dropped the constant energy shift @xmath158 from eq .
( [ eq : dressede ] ) since it is present for all the basis states and plays no role in a following analysis . in general , @xmath159 is neither parallel nor perpendicular to @xmath160 .
hence , our spin - orbit coupling terms are of the form @xmath161 .
the off - diagonal perturbation in the fully quantum hamiltonian is now given by @xmath162 where the first term describes the orbital - photon coupling , and the second term describes the spin - orbit coupling .
the spin - orbit coupling constants @xmath163 depend , in part , on geometrical factors such as the relative alignment of the magnetic field gradient with respect to the dipole moment of the orbital states .
the term @xmath164 is the conventional edsr matrix element @xcite , describing the flip - flop transition between the orbital and spin states .
the constant @xmath165 is of particular interest for the resonance ( 5 ) .
it describes a hybridization of spin states ( @xmath166 ) , which depends on the orbital occupation , but does not hybridize the orbital states ( @xmath167 ) .
these coupling constants arise through the valley - orbit coupling mechanism , which depends on the local disorder realization at the quantum well interface . since we do not know the disorder potential for our device , we do not attempt to calculate @xmath163 here .
however , we provide estimates for several of these parameters below , based on our experimental measurements . the other contributions to the spin - orbit matrix elements can be evaluated using eqs .
( [ eq : me1])-([eq : me3 ] ) , and the additional results @xmath168 which are also obtained using standard techniques @xcite . here
, @xmath169 is an integer bessel function . in the simple - spin model ,
no photons are absorbed via spin - orbit coupling , so only the @xmath170 terms will be considered in eqs .
( [ eq : me4])-([eq : me6 ] ) .
there are therefore three types of direct ( i.e , first - order ) transitions allowed by eq .
( [ eq : v ] ) : ( i ) pat transitions , with no spin flip , like those described in eq .
( [ eq : me3 ] ) , ( ii ) spontaneous flip - flops between the spin and orbital states ( @xmath171 processes ) , or ( iii ) phase exchange ( @xmath172 ) processes , without photon absorption or emission .
note that direct transitions of the @xmath173 or @xmath165 type are also allowed ; however , they do not conserve energy , and are therefore highly suppressed in resonant transitions .
the only relevant first - order matrix elements are therefore @xmath174 note that eq .
( [ eq : me30 ] ) corresponds to pat , while eqs .
( [ eq : me50 ] ) and ( [ eq : me60 ] ) describe pure spin - orbit coupling , which hybridizes the spin and orbital states , even in the absence of a driving term .
we now consider second - order processes in the perturbation , @xmath175 .
all photon - driven processes with spin flips must be second - order , involving two distinct steps . in the first ,
an orbital state is excited ( deexcited ) while absorbing ( emitting ) one or more photons ; this is a nonresonant lz process . in the second ,
a spin flip occurs , accompanied by a second orbital flip ( an @xmath171 process ) , or a phase change of the orbital state ( an @xmath165 process ) .
processes of the @xmath173 and @xmath172 type are also allowed ; however they do not cause spin flips , and will not be considered in our edsr analysis . we first consider a pure - spin resonance generated by the sequence @xmath176 . here , the difference between the initial and final states involves a spin excitation and @xmath132 absorbed photons .
the intermediate state @xmath177 is off - resonant , and may be eliminated to obtain an effective , second - order matrix element , given by @xmath178 here , we have evaluated the energy denominator at the resonant condition @xmath179 .
similarly , we can consider resonances ( 5 ) , ( 6 ) , and ( 8) type process generated by the sequence @xmath180 . in this case , the difference between the initial and final states involves a spin excitation , a charge excitation , and @xmath132 absorbed photons .
the effective , second - order matrix element for this process is given by @xmath181 corresponding to the resonance condition @xmath182 .
in the limit @xmath183 , which seems appropriate for our experiments ( see below ) , we therefore obtain @xmath184 here , @xmath185 represents the rabi frequency for a pure - spin resonance , and @xmath186 represents the rabi frequency for a combined spin - orbit resonance ( like resonance ( 5 ) ) . keeping all other parameters fixed
, we can estimate the power scaling of the rabi frequencies ( [ eq : rabif ] ) and ( [ eq : rabifp ] ) by using eqs .
( [ eq : gcouple ] ) , ( [ eq : alpha ] ) , and ( [ eq : nulargen ] ) . assuming the limit @xmath187
, we find that @xmath188 , @xmath189 , etc . ; these theoretical results are consistent with experimental results reported in ref .
@xcite . based on eq .
( [ eq : rabif ] ) , we also predict a simple form for the ratio @xmath190 using the results for @xmath191 and @xmath192 in ref .
@xcite , we estimate this ratio to be in the range @xmath193 - 1/2 for typical experiments , which confirms our previous assumptions . for the experiments reported here , we use lower microwave power than in ref .
@xcite , which yields a smaller estimate for this parameter , but makes it impossible to measure @xmath191 and @xmath192 . we can estimate the ratio of the rabi frequencies between different orbital states ( i.e. , different values of @xmath194 ) , but the same edsr mode ( the same value of @xmath132 ) . taking @xmath109 in eq .
( [ eq : rabif ] ) , we obtain the ratio @xmath195 here , we have adopted the notation of ref .
@xcite where @xmath196 ( @xmath197 ) corresponds to @xmath136 ( @xmath198 ) for the case @xmath199 . in ref .
@xcite , it was found that @xmath200 when @xmath201 mt . using these values , eq .
( [ eq : rabiratio ] ) predicts that @xmath202 @xmath11ev or 4 ghz .
we note that , although ref .
@xcite used the same device as we do here , the gates were tuned differently , which could potentially yield to different values for the valley splitting . the spectroscopy shown in fig .
1 of the main text gives a more direct estimate of @xmath203 ghz .
the agreement is reasonable ; however , eq . ( [ eq : rabiratio ] ) also predicts that @xmath204 should not depend on @xmath132 , while ref .
@xcite estimates a ratio of @xmath205 for the first subharmonic resonance ( @xmath206 in their notation ) .
this experimental result is clearly inconsistent with eq .
( [ eq : rabiratio ] ) , which should be greater than 1 .
we point out , however , that the error bars reported in ref .
@xcite were large enough to permit @xmath207 . moreover ,
we note again that strong driving was employed in ref .
@xcite , so that the arguments leading to eq .
( s37 ) could begin to break down . the scaling behavior @xmath208 suggested by eq .
( [ eq : rabif ] ) is consistent with tokura et al .
we can now obtain a rough estimate for the edsr coupling constant @xmath171 by combining the current experimental results with those reported in ref .
@xcite . in ref .
@xcite , a typical value of @xmath209 mhz was obtained ( note that @xmath196 should be smaller in the current experiment , although we do not know its value ) . in the current experiments , we find that @xmath210 ghz in the range of interest ( see fig . 1 ) , and @xmath211 ghz ( see sec . [
sec : hybrid ] , below ) . from eq .
( [ eq : rabif ] ) , we then obtain the estimate @xmath212 mhz .
we can also estimate the parameter @xmath165 . as discussed above ,
this coupling constant describes the orbital ( i.e. , valley)-induced hybridization of the spin states , and determines the strength of the resonances ( 5 ) and ( 6 ) .
specifically , it arises from contributions to the @xmath15-factor tensor that are transverse to the applied field , and which differ slightly between the two different valley states .
it is not possible to compute these valley - induced perturbations without knowledge of the disorder potential . in general
however , the disorder will not be aligned with the magnetic field , so the transverse component of the perturbation in the @xmath15-factor tensor should differ from the total perturbation by a simple geometrical factor of order @xmath213 $ ] . in ref .
@xcite , the total difference in the @xmath15-factors for the two valley states was found to be @xmath214% .
hence , we can estimate the valley - dependent hybridization of the spin states to be @xmath215 . at the magnetic field corresponding to the kink in hybridized resonance ( 5,6 ) in fig . 1 of the main text
, we have @xmath216 mt , yielding @xmath217 mhz , or @xmath218% .
finally , we can estimate a typical rabi frequency for resonance ( 5 ) from eq .
( [ eq : rabifp ] ) , giving @xmath219 khz .
this small value explains why it is not possible to observe coherent oscillations of the resonance ( 5 ) in our experiments . to close this section ,
we discuss the basic features of the simulation results shown in fig .
[ fig : theorylabels ] , in terms of the first and second - order matrix elements obtained in eqs .
( [ eq : me30])-([eq : me60 ] ) and ( [ eq : rabif])-([eq : rabifp ] ) .
first , it is helpful to recall how the spin enters this discussion , since the orbital hamiltonian used in the simulations , eq .
( [ eq : h ] ) , does not explicitly include spin .
it is important to note that fig .
[ fig : theorylabels ] does not describe spin resonance phenomena ; it is simply the fourier transform of the response of the orbital dipole moment to an ac drive . in the simple - spin model
, this response is converted into an ac magnetic signal , with a strength determined by the spin - orbit coupling parameters @xmath163 in eqs .
( [ eq : me50 ] ) and ( [ eq : me60 ] ) , yielding an edsr response in the spin sector .
since there is no spin - orbit coupling in the simulations , we can learn nothing about the spin rabi frequencies .
however , the response spectrum in fig .
[ fig : theorylabels ] should still provide an accurate mapping of the edsr spectrum . to take some examples , the features label ( 1,2 ) and ( 3,4 ) in fig .
[ fig : theorylabels ] correspond to conventional edsr , with spin excitations occurring at the resonance conditions @xmath220 ( or @xmath179 ) when @xmath199 and 2 , respectively .
these features correspond to the experimenal resonances labeled the same way in fig .
1(a ) of the main text .
the corresponding rabi frequencies are given by @xmath185 in eq .
( [ eq : rabif ] ) , and were measured to be in the range of a few mhz in ref .
the feature labeled ( 9 ) in fig .
[ fig : theorylabels ] corresponds to a 3-photon spin resonance , with @xmath221 ; this feature is not observed in the experiments .
the features labeled ( 5 ) , ( 6 ) , and ( 8) occur at the resonance conditions @xmath222 [ or @xmath182 ] , and involve simultaneous spin and orbital excitations . in previous work ,
such processes were referred to as spin - flip pat " @xcite .
the corresponding rabi frequencies are given by @xmath186 in eq .
( [ eq : rabifp ] ) , and have not been measured yet experimentally . in principle , the pat transitions , which occur at the resonance condition @xmath224 ( or @xmath225 ) in eq .
( [ eq : me30 ] ) , could also be observed in the simulations .
however , these transitions correspond to discrete points along the line labeled ( 7 ) in fig . [
fig : theorylabels ] , which makes them difficult to detect .
moreover , the response frequency for pat ( @xmath226 ) is smaller than fermi level broadening , so it can not be detected in our readout scheme .
it is therefore interesting that pat plays a key role in our understanding of the hybridization between dressed states , as discussed in supplementary sec .
[ sec : hybrid ] , and that the repulsion between the dynamical modes allows us to estimate the magnitude of the pat rabi frequency , @xmath123 .
one of the most interesting features in fig .
[ fig : theorylabels ] is labeled ( 7 ) , and corresponds to a pure - lz transition , with an orbital state excitation but no photon absorption . since @xmath170 , according to eq .
( [ eq : me3 ] ) , this can not be described by a direct , first - order process .
instead , it involves a combination of photon - mediated processes , with a _ net _ photon absorption of @xmath170 .
a minimal process of this type must be third - order , since second - order processes would involve an orbital excitation , followed by a deexcitation , leaving the orbital in its initial state .
since the electrons in experiments have spins ( in contrast with our simulations ) , there is also another mechanism that could cause feature ( 7 ) in the experiments : a spin - orbit flip - flop occurring at the resonance condition @xmath227 ( or @xmath228 ) .
this is the direct , first - order process , described by eq .
( [ eq : me60 ] ) , in which the orbital state spontaneously deexcites while causing a spin excitation .
feature ( 7 ) should therefore appear in both the simulations and the experiments , but with different physical origins and different rabi frequencies .
this is a moot point for experiments however , since low - energy processes can not be resolved by our measurement scheme .
dressed states allow us to analyze one of the most intriguing features of the data : the kink in the hybridized resonance ( 5,6 ) , where the resonance line appears to switch between two well - defined slopes , as shown in fig .
1(b ) of the main text .
the two asymptotic resonance lines correspond to the conditions @xmath229 ( or @xmath230 ) far to the left of the kink , and @xmath231 ( or @xmath232 ) far to the right .
the intersection of these two resonance lines occurs when @xmath233 , or @xmath234 , and defines a quadruple resonance . at this special point ,
the arrows in the processes labeled ( 3)-(7 ) in fig .
1(d ) all have the same length .
the hamiltonian for this 4d manifold is given by @xmath235 we can gain intuition about the experiments and simulations by considering a simplified version of @xmath236 .
first , we note that @xmath237 when @xmath94 .
to investigate the dominant behavior , since @xmath238 , we consider the limit @xmath239 .
the resulting hamiltonian @xmath240 does not permit any dynamical hybridization of spin states , and can therefore be used to directly identify the spin resonance conditions .
we find that eq .
( [ eq : h4dsimple ] ) reproduces all the main features of the experimental and simulation results near the kink in hybridized resonance ( 5,6 ) , as indicated in fig .
[ fig : quad ] .
the features observed in figs .
[ fig : quad](b ) and ( c ) reflect sudden changes in the compositions of the eigenstates . to obtain these plots ,
we first diagonalize @xmath236 to obtain its four eigenstates as a function of @xmath241 .
we then compute probability @xmath242 for the @xmath243th state to have spin @xmath244 . here , the states are labeled according to their energy ordering , as they would be for eigenstates of the full 4d hamiltonian in eq .
( [ eq : h4d ] ) . near a spin resonance , the spin - up or spin - down character of the ordered eigenstates changes abruptly . to see these sudden changes in the simulations
, we sum the numerical derivatives , @xmath245 .
the resulting spectrum displays sharp peaks at the resonance locations , as illustrated in fig .
[ fig : quad](b ) .
comparison with the simulation results , reproduced in fig .
[ fig : quad](a ) , shows that only one parameter ( @xmath246 ) is needed to capture the main features of the hybridized resonance ( 5,6 ) . including the other off - diagonal terms in eq .
( [ eq : h4d ] ) leads to small feature changes , such as broadening of the resonance peaks , but does not change the underlying spectrum .
the dynamically induced shifts of the dressed state energies ( i.e. , ac stark shifts ) are dominated by the orbital coupling terms @xmath123 , because @xmath247 .
hence , they are easily computed from eq .
( [ eq : h4dsimple ] ) by diagonalizing the two 2d subsystems . diagonalizing the hamiltonian in this way
yields the four energies @xmath248 in particular , the energy splitting between hybridized resonance ( 5,6 ) and its large-@xmath241 asymptote , @xmath57 , is given by @xmath249 and the energy splitting at the anticrossing between resonances ( 3,4 ) and ( 5,6 ) is given by @xmath250 .
using the asymptotic formula for the rabi frequencies , eq .
( [ eq : nulimit ] ) , and in particular , @xmath251 , we find that @xmath252 in the regime of experimental interest .
we can also obtain analytical expressions for the resonance conditions by equating the energy expressions in eqs .
( [ eq : acstark1 ] ) and ( [ eq : acstark2 ] ) . for the resonant feature labeled ( 3,4 ) in fig .
[ fig : quad ] , we obtain the condition @xmath253 . for ( 5,6 ) , we obtain the conditions @xmath254 note that this equation describes the hybridized resonances on either side of ( 3,4 ) . it is interesting to note that , while the parameters @xmath185 control the spin dynamics and the edsr rabi frequencies , the dressed state energies @xmath255 determine the gross features of the resonance spectrum , and the pat rabi frequency @xmath256 determines the fine features of hybridized resonance ( 5,6 ) . from eqs .
( [ eq : h4d ] ) and ( [ eq : h4dsimple ] ) , we see that resonances ( 5 ) and ( 6 ) correspond to dressed states with the same spin but different orbital states .
the kink in the hybridized resonance is therefore caused by orbital physics .
the correspondence between figs .
[ fig : quad](a ) and [ fig : quad](b ) indicates that all of the resonances in this manifold involve a spin flip .
we have also computed the analogous orbital resonances for the same 4d manifold , corresponding to sudden changes in the charge state , as obtained from the expression @xmath257 .
the results are shown in fig .
[ fig : quad](c ) . in this case
, the central resonance ( 3,4 ) disappears , indicating that it is a pure - spin resonance .
although coherent edsr oscillations were observed in ref .
@xcite , with rabi frequencies in the range @xmath258 - 4 mhz , it was not possible to observe coherent pat oscillations because of their relatively low energy scale .
however , we can now estimate the pat rabi frequency by fitting the analytical theory of eq .
( [ eq : h4dsimple ] ) to the experimental resonance spectrum in fig .
1(b ) of the main text , obtaining the result @xmath259 ghz .
this confirms our previous assumption that @xmath260 . we can also estimate the remaining parameters in the theoretical model of eqs .
( [ eq : h ] ) and ( [ eq : epst ] ) . from eqs .
( [ eq : gcouple ] ) and ( [ eq : nulargen ] ) , we obtain @xmath261 ghz @xmath262 @xmath11ev . using the experimental result @xmath263 ghz @xmath264 @xmath11ev and the rough estimate @xmath265 ( deduced from our simulations )
, we also obtain @xmath266 ghz @xmath267 @xmath11ev . the parameters @xmath268 and @xmath49 that describe the detuning of the orbital basis states due to the driving field are difficult to determine from first principles in silicon devices because they are determined by the degree of valley - orbit mixing , which depends sensitively on the details of the interfacial disorder @xcite .
in contrast , the energy splitting between the orbital states @xmath70 and the pat frequency @xmath256 can be extracted directly from the experiments .
schematic of the device , showing a pattern of split gates , labeled 1 - 10 . for this experiment
we create a single quantum dot ( estimated location indicated by a blue circle ) and a sensing dot ( gates 4 and 5 ) .
the current through the sensing dot is recorded in real time for a fixed voltage bias of 500 @xmath11ev . the voltage pulses and microwave excitation
are applied to gate 2 and 6 respectively .
green semitransparent rectangles show the position of two 200-nm - thick co micromagnets .
the yellow - shaded areas show the location of two accumulation gates , one for the reservoirs and the other for the double quantum dot region.,width=170 ] all the measurements shown here make use of 4-stage voltage pulses ( see inset of fig .
4(a ) ) applied to gate 2 @xcite ( see fig .
[ fig : figs1 ] ) : ( 1 ) initialization to spin - down ( 4 ms ) , ( 2 ) spin manipulation through microwave excitation of gate 8 ( 1 ms ) , ( 3 ) single - shot spin read - out ( 4 ms ) , and ( 4 ) a compensation / empty stage ( 1 ms ) . by repeating this cycle , typically 150 - 1000 times ,
we collect statistics of how often an electron leaves the dot during the detection stage , giving the spin - excited state probability , @xmath7 , at the end of the manipulation stage .
microwave excitation , generated by an agilent e8267d vector source , is applied to gate 6 through the following attenuation chain : dc block / ult-05 keycom coax/20 db attenuator ( 1.7 k)/ nbti coax /10 db attenuator ( 50 mk)/flexible coax / sma connector on the printed circuit board ( pcb ) .
the high frequency signal is combined on the pcb with a dc voltage line by a homemade resistive bias tee ( @xmath269 m@xmath20 , c= 47 nf ) . in fig .
[ fig : fig3 ] we report repeated measurements of resonance ( 5 ) recorded over time , in order to investigate the low frequency noise fluctuations . in the main central panel we report , in blue , the spin resonance peak obtained by averaging directly , over time , the data of panel ( a ) , inside the blue square frame . from its full width at half maximum ( fwhm )
we can get a lower bound for @xmath17 , in this case of 100 ns .
alternatively we can also perform a gaussian fit for each of the time traces in panel ( a ) and then shift the center of each dataset in order to align the centers of each resonance peak on top of each other , as reported in panel ( b ) .
if we now perform the averaging we get the red gaussian , reported in the main central panel(a ) , from which we can extract a @xmath17 of 150 ns .
this last averaging strategy is equivalent to filtering out the very low frequency component of the noise ( @xmath0mhz ) , clearly visible in panel ( a ) .
+ in what follows we report 3 different measurements realized in order to investigate the sensitivity of resonance ( 5 ) to the electrostatic environment and to electrical noise . controlling the electrostatic potential by the voltage applied to the gates defining the qd confinement potential
, we can systematically shift in frequency resonance ( 5 ) ( see fig . 3 ) .
the inset of fig .
[ fig : fig4](a ) shows the peak of resonance ( 5 ) for five different voltage configurations of the gate 2 ( @xmath50 , represented by different colors ) .
the main panel reports a measurement of the g - factor in each of the gate voltage configurations reported in the lower inset of fig .
[ fig : fig4](a ) .
the extracted g - factors are compatible with each other , considering their associated errors , as reported in fig .
[ fig : fig4](b ) .
this implies that , by changing the gate voltage we can modify the larmor frequency of the 2-level system involved in the resonance process , without appreciably modifying the zeeman energy .
this suggests that , in this case , we are directly modifying the valley splitting ( modelled by moving the qd towards a step defect in the qw , as schematically represented in fig .
we can estimate the effective wave function displacement in space ( mainly in the qw plane ) generated by modifying the voltage applied to the gate 2 by 1 mv , by using the shift of the larmor frequency of the intra - valley spin resonance ( red trace in fig .
3(a ) ) due to the magnetic field gradient .
from the simulation of the magnetic field gradient we can estimate an upper bound for the magnetic field gradient of @xmath00.2 mt / nm ( in the y - direction @xmath0gate 2 ) , equivalent to @xmath056 mhz / nm in silicon .
this energy gradient gives a position lever - arm of @xmath00.01 nm / mv , considering the slope of @xmath00.55 mhz / mv of the red curve in fig .
furthermore , by making use of the measured value 18.5 mhz / mv for @xmath12 ( see blue trace in fig . 3(a ) ) and of the estimation above @xmath270 nm / mv , we can conclude that @xmath271 ghz / nm ( where @xmath272 is the vector representing the motion of the electron generated in the si qw by the changing of the voltage on gate 2 ) mhz ) is determined by the valley energy splitting fluctuation caused by charge noise , by using the measured @xmath273 mhz / mv we can translate it into an equivalent standard deviation of gate voltage noise ( for the specific gate 2 used to perform the measurement ) @xmath274 @xmath11ev . ] . an equivalent way to explore this effect
is reported in figures [ fig : fig4](c - d ) where keeping a constant static d.c .
voltage configuration , we change the amplitude of the voltage pulse applied on gate 2 to bring the system in coulomb blockade during the manipulation stage . as summarized in fig .
[ fig : fig4](d ) , by increasing the pulse amplitude @xmath275 , the resonance peak systematically moves toward lower frequencies .
the lever - arm of this process ( @xmath276 ) is around 9 mhz / mv .
this measurement demonstrates the dynamic stark shifting of resonance ( 5 ) and creates opportunities for site - selective addressing ( voltage pulse induced addressability ) . the difference in @xmath277 between the d.c . and
pulsed case can be ascribed to the fact that for the measurement reported in fig .
3(a ) we have to compensate the change in @xmath50 by changing the voltage on another gate , in order to keep a good initialization - readout fidelity .
instead , for the pulsed case we just modify the voltage pulse amplitude on the gate 2 in the manipulation stage ( and the empty stage , to keep the symmetry of the pulse ) , without any further compensation .
the measurements reported above reveal that the electrostatic configuration considerably affects the inter - valley resonance frequency .
next we consider the effect of the electrical noise on the coherence of the inter - valley spin resonance .
we already have clear evidence of the importance of the electric noise from the upper bound of @xmath17 of resonance ( 5 ) , estimated from the fwhm measurement reported in in fig .
[ fig : fig3 ] , which is 10 times shorter than what it should be if just hyperfine fluctuations dominated @xcite .
this results from the higher sensitivity of resonance ( 5 ) to the electrical environment , so also to the electrical noise . in the hypothesis that the hyperfine ( hf ) and the electric noise ( el ) are independent , we can write the total noise affecting this resonance as @xmath278 , from which + @xmath279_{el}\sim 1/\sqrt{([t_2^\ast]_{tot})^{-2}-([t_2^\ast]_{hf})^{-2}}\sim 100 $ ] ns . therefore , we can deduce that the coherence properties of the inter - valley spin resonance are almost completely dominated by the electric noise ( here for @xmath279_{hf}$ ] we use the value 950 ns extracted for the inter - valley spin resonance in @xcite ) . now , we present a measurement ( reported in fig . 3(d ) ) realized to study the correlation between the low frequency fluctuations of resonance ( 5 ) and low frequency electric noise ( low with respect to the electron tunnel rate during the readout stage ) . in this respect
we made use of the following measurement technique ( see fig .
[ fig : figs4 ] ) : + i ) the tunnel rate for an electron jumping inside ( or outside ) a qd from ( into ) an electron reservoir is a sensitive function of the relative energy alignment between the fermi level of the electron reservoir and the electrochemical potential of the single electron inside the qd @xcite . fig .
[ fig : figs4](a ) reports a schematic representation of @xmath280 ( the single electron tunnel rate in ( y - axis ) ) , as a function of the voltage on gate 2 ( @xmath50 ) that controls the relative alignment of the two levels ( measured data represented in fig .
[ fig : figs4](c ) ) .
making @xmath50 more positive the tunnel rate will increase , due to the participation of the excited state in the tunneling process .
each tunnel rate is extracted by recording and processing on the fly ( using an fpga ) the tunnel events during the initialization stage ( @xmath05 ms ) and fitting an exponential to the histogram of the number of events versus time @xmath281 ( fig .
[ fig : figs4](a ) ) . here
@xmath281 is the time between the start of the initialization stage and the actual tunnel - in event .
for the experiment we used @xmath50 around -196 mv ( red points in fig .
[ fig : figs4](c ) ) , which has been optimized for a good readout and initialization process . as schematically represented inside the orange frame of fig .
[ fig : figs4](b ) , if we model the electric noise from the environment as a relative fluctuation of @xmath50 ( @xmath282 ) , we can write @xmath283 . in this way we use @xmath284 , the tunnel rate fluctuations during the initialization stage ( of the 4-stages pulse sequence ) , as a probe of electric noise .
+ ii ) the same 4-stage voltage pulse scheme allows us to perform a rotation of the electron spin and to read it out in the last stage . due to
the coupling with the environmental noise , the spin larmor frequency , @xmath12 , fluctuates ( see blue frame of fig . [
fig : figs4](b ) ) over time , as reported in the cw measurement over time in fig .
[ fig : figs4](d ) .
we fit each resonance trace to a gaussian in order to extract the center of it ( larmor frequency ) .
+ iii ) we repeat this measurement over time ( each trace takes tens of seconds ) , in order to get enough statistics , and plot , as reported in fig .
3(d ) , the extracted @xmath280 ( x - axis ) and larmor frequency fluctuation @xmath285 ( y - axis ) traces , for each measurement cycle .
a more rigorous estimate of the correlation in time between @xmath286 and @xmath287 , evaluated according to the pearson product - moment correlation coefficient gives a value @xmath2880.5 indicating a modest correlation in time .
+ the sensitivity of this procedure is related to the ability to distinguish small charge fluctuations and it is a function of the static gate voltage configuration ( @xmath50 ) chosen for the measurement . in fact , from fig .
[ fig : figs4](c ) we can clearly notice that we are not working yet in the most sensitive gate voltage configuration ( maximal @xmath289 ) .
instead , the specific electrostatic configuration has been chosen in order to optimize the spin initialization - readout .
the simulation reported in fig .
3(c ) of the main text represents the result of a 2d tight binding calculation of valley splitting in the presence of interfacial disorder , in the form of a single atomic step . the fact that the simulation is 2d rather than 3d affects just the magnitude of @xmath70 , but not its qualitative dependence on the step offset .
the quantum well barrier has been chosen to be 160 mev , corresponding to 30@xmath51 ge .
for simplicity , we assumed a parabolic confinement potential for the dot of size @xmath290 nm , corresponding to an orbital energy splitting of @xmath291 mev @xcite .
the electric field and the quantum well width have been chosen to be @xmath292 v / m and 13 nm , respectively .
the experimental quantum well is nominally 12 nm and the experimental electric field is not well known .
there are 3 competing effects that determine the energy splitting between the two lowest eigenstates , which we usually refer as ` valley splitting ' , for simplicity .
( i ) the first is the actual valley splitting , which is maximized when the electron sees no step .
for the electric field assumed here , the valley splitting can vary by about 75 @xmath11ev , as shown in fig .
3(c ) of the main text , depending on the position of the step relative to the dot . in fig .
s7 , we see that the ground and excited states are shifted to the left or right from their unperturbed positions , due to the step . moreover , they are shifted differently , as a signature of valley - orbit coupling .
( ii ) the second effect is related to the electrostatic energy .
when the electron sits to the left of the step , it gains a little energy because the energy is given by @xmath293 , where z is the vertical position . for a single atom step with @xmath294 v / m ,
this energy scale is 200 @xmath11ev , making it comparable to the valley splitting .
this competition between the electrostatic and valley splitting perturbations caused by a step is always present , for any value of the electric field , because the electrostatic energy and the valley splitting both have a linear dependence on @xmath295 .
( iii ) the third competing effect is the confinement energy of the parabolic dot : when the electron center of mass moves to the left or right due to the step , this increases the average energy of the electron in the parabola . as shown in fig .
s7 , the maximum displacement of the dot is 5 - 6 nm , when the step is positioned 10 - 20 nm from the center of the parabola .
this corresponds to an energy shift of about 10 @xmath11ev , which is slightly lower than the maximum valley splitting and electrostatic energy scales .
however , for a step offset of 20 nm , the valley splitting and electrostatic energy scales are also somewhat smaller , resulting in a competition between all three energies .
this competition causes the left - right symmetry of figs .
3(c ) and s7 to be broken .
the competition also makes it difficult to gain much intuition about the final results . without the competing effects
, we would expect the valley splitting to go to zero as the ground state changes from ` even ' to ` odd ' in the valley parameter .
but the electron always wants to maximize the energy splitting between the two eigenstates , and the competing effects ( in particular , the shifting of the electrons to the left and right ) allow it to keep a nonzero energy splitting at all times .
it is interesting that this minimum valley splitting ( @xmath025 @xmath11ev ) is very close to the one we observed in our experiments .
37ifxundefined [ 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 10.1126/science.1148092 [ * * , ( ) ] link:\doibase 10.1088/0268 - 1242/24/6/064004 [ * * , ( ) ] link:\doibase 10.1038/nnano.2012.160 [ * * , ( ) ] link:\doibase 10.1038/nnano.2013.140 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.227601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.166801 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.195417 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.106802 [ * * , ( ) ] link:\doibase 10.1103/physrevb.64.125316 [ * * , ( ) ] link:\doibase 10.1063/1.2387975 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.115324 [ * * , ( ) ] link:\doibase 10.1038/nphys475 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.85.961 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.205433 [ * * , ( ) ] link:\doibase 10.1038/ncomms3069 [ * * , ( ) ] link:\doibase 10.1038/ncomms4860 [ * * , ( ) ] link:\doibase 10.1038/nnano.2014.153 [ * * , ( ) ] link:\doibase 10.1038/nnano.2014.216 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.201401 [ * * , ( ) ] link:\doibase 10.1038/nmat3585 [ * * , ( ) ] link:\doibase 10.1038/nature00976 [ * * , ( ) ] link:\doibase 10.1038/nature03815 [ * * , ( ) ] link:\doibase 10.1038/ncomms1561 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.075417 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.126804 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.047202 [ * * , ( ) ] link:\doibase 10.1038/nphys1053 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.241305 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1098/rspa.1932.0165 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.88.035310 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.140503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.94.035438 [ * * , ( ) ] link:\doibase 10.1063/1.1815041 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.72.3339 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.235315 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.115409 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.247002 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2010.03.002 [ * * , ( ) ] link:\doibase 10.1126/science.1119678 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.87.246601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.257003 [ * * , ( ) ] link:\doibase 10.1103/physreva.69.042302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.046807 [ * * , ( ) ] , , and , eds .
, link:\doibase 10.1002/9783527617197 [ _ _ ] ( , ) link:\doibase 10.1103/physrevlett.73.3443 [ * * , ( ) ] link:\doibase 10.1063/1.2591432 [ * * , ( ) ] | the valley degree of freedom is intrinsic to spin qubits in si / sige quantum dots .
it has been viewed alternately as a hazard , especially when the lowest valley - orbit splitting is small compared to the thermal energy , or as an asset , most prominently in proposals to use the valley degree of freedom itself as a qubit . here
we present experiments in which microwave electric field driving induces transitions between both valley - orbit and spin states .
we show that this system is highly nonlinear and can be understood through the use of dressed photon - orbital states , enabling a unified understanding of the six microwave resonance lines we observe .
some of these resonances are inter - valley spin transitions that arise from a non - adiabatic process in which both the valley and the spin degree of freedom are excited simultaneously . for these transitions , involving a change in valley - orbit state
, we find a tenfold increase in sensitivity to electric fields and electrical noise compared to pure spin transitions , strongly reducing the phase coherence when changes in valley - orbit index are incurred .
in contrast to this non - adiabatic transition , the pure spin transitions , whether arising from harmonic or subharmonic generation , are shown to be adiabatic in the orbital sector .
the non - linearity of the system is most strikingly manifest in the observation of a dynamical anti - crossing between a spin - flip , inter - valley transition and a three - photon transition enabled by the strong nonlinearity we find in this seemly simple system .
1.5truecm a spin-1/2 particle is the canonical two - level quantum system .
its energy level structure is extremely simple , consisting of just the spin - up and spin - down levels .
therefore , when performing spectroscopy on an elementary spin-1/2 particle such as an electron spin , only a single resonance is expected corresponding to the energy separation between the two levels .
recent measurements have shown that the spectroscopic response of a single electron spin in a quantum dot can be much more complex than this simple picture suggests .
this is particularly true when using electric - dipole spin resonance , where an oscillating electric field couples to the spin via spin - orbit coupling @xcite .
first , due to non - linearities in the response to oscillating driving fields , subharmonics can be observed @xcite , and the non - linear response can even be exploited for driving coherent spin rotations @xcite .
second , due to spin - orbit coupling , the exact electron spin resonance frequency in a given magnetic field depends on the orbital the electron occupies @xcite . in silicon or germanium quantum dots ,
the conduction band valley is an additional degree of freedom @xcite , and the electron spin resonance frequency should depend on the valley state as well @xcite . as a result , when valley or orbital energy splittings are comparable to or smaller than the thermal energy , thermal occupation of the respective levels leads to the observation of multiple closely spaced spin resonance frequencies @xcite .
the picture becomes even richer when considering transitions in which not only the spin state but also the orbital quantum number changes .
such phenomena are common in optically active dots @xcite , but have been observed also in electrostatically defined ( double ) quantum dots in the form of relaxation from spin triplet to spin singlet states @xcite and spin - flip photon - assisted tunneling @xcite .
however , that work is all in semiconductor quantum dots with no valley degree of freedom , and the degree to which valleys -often treated as weakly coupled to each other and orbital states - couple to each other to enable microwave - driven transitions that change spin has not been explored .
furthermore , the investigation of resonant transitions involving the valley degree of freedom is very important in the context of the new qubit architecture recently proposed for si quantum dots @xcite , based on the valley degree of freedom to encode and process quantum information . here
, we report transitions where both the spin and valley - orbit state flip in a si / sige quantum dot .
we demonstrate that we can stark shift the transitions , and we compare the sensitivity to electric fields to the case of pure spin transitions , including the impact on phase coherence .
we find that the valley - orbit coupling strongly affects the coherence properties of the inter - valley spin resonances .
we show that a theory incorporating a driven four - level system comprised of two valley - orbit and two spin states subject to strong ac driving provides a consistent description of these transitions , as well as all the previously reported transitions for this system .
this theory also explains the observation of a dynamical level repulsion , which can be understood effectively and compactly using a dressed - state formalism . |
when dealing with complicated multilevel systems , such as atoms or molecules , it is necessary to look for allowed restrictions of the hilbert space which can lead to simplifications of the computational work .
thus , one usually forgets states which are not populated initially and not coupled , either _ directly _ or _ indirectly _ , to initially occupied states .
it is sometimes possible to go beyond this first step and isolate some subset of initially occupied states if they are only _ weakly _ and _ non - resonantly _ coupled to the others .
the effective dynamics of such a subset can then be approximately described by a hamiltonian of smaller dimensions than the original one , in which the effect of couplings outside the relevant subspace is accounted for by additive energy shifts and couplings .
the procedure which allows one to get rid of the irrelevant states and derive this effective hamiltonian is called _
adiabatic elimination_. one of the simplest examples of such a situation is the case of a three - level lambda system , the low levels of which are initially populated and non - resonantly coupled to the initially empty upper level via detuned harmonic perturbations : through adiabatic elimination , it is possible to reduce the problem to an oscillating two - level system , which has been widely studied , for instance in atomic physics @xcite .
the aim of the present paper is to understand how the adiabatic elimination procedure works on this simple example and how it should be performed .
after briefly presenting the model ( sec . [ model ] ) , we show that an ansatz commonly used in the literature to adiabatically eliminate the excited state may lead to ambiguous results ( sec .
[ roughadel ] ) . in order to better understand this scheme and explicit its conditions of applicability
, we then review two rigorous approximation methods to treat the problem : the first one relies on the solution of the schrdinger equation and consists in neglecting the fast oscillating terms in the exact expression of the amplitude of the excited state , which is then injected back into the dynamical equations for the lower states ( sec .
[ adelres ] ) ; the second one resorts to the green s function formalism and makes use of the pole approximation which is discussed in detail ( sec .
[ green ] ) .
in this paper , we shall consider an atomic lambda system consisting of two lower states @xmath0 coupled to an excited level @xmath1 via two off - resonance lasers : the detunings are denoted @xmath2 @xmath3 where @xmath4 and @xmath5 are the frequencies of the lasers ( cf figure [ fig1]a ) .
[ ptb ] fig1.eps turning to the rotating frame defined by the transformation @xmath6 where @xmath7{ccc}% \frac{\delta}{2 } & 0 & 0\\ 0 & \omega_{a}-\omega_{b}-\frac{\delta}{2 } & 0\\ 0 & 0 & \omega_{a}-\frac{\delta_{a}+\delta_{b}}{2}% \end{array } \right]$ ] and @xmath8 , and performing the rotating wave approximation , one gets ( cf figure [ fig1]b ) @xmath9{ccc}% -\frac{\delta}{2 } & 0 & \frac{\omega_{a}^{\ast}}{2}\\ 0 & \frac{\delta}{2 } & \frac{\omega_{b}^{\ast}}{2}\\ \frac{\omega_{a}}{2 } & \frac{\omega_{b}}{2 } & \delta \end{array } \right ] , \label{ham}%\ ] ] where @xmath10 , and @xmath11 denote the rabi frequencies of the lasers coupling @xmath12 to @xmath13 and @xmath14 to @xmath13 , respectively .
note that we have implicitly chosen the origin of the energies between the two states @xmath12 and @xmath15 ( see figure [ fig1]b ) . from now on
, we shall assume @xmath16 .
if the system is initially prepared in a superposition @xmath17 , the excited state will then essentially remain unpopulated , while second - order transitions will take place between the two lower states :
this constitutes the so - called raman transitions , which play an important role in atomic and molecular spectroscopy , and have recently become important processes in laser cooling and trapping @xcite and in quantum computing proposals with ions @xcite , atoms @xcite , and solid - state systems @xcite . in these conditions
, it is natural to restrict the hilbert space to the relevant states @xmath12 and @xmath15 and to describe their dynamics by a @xmath18 effective hamiltonian @xmath19 . in the following sections , we describe and discuss different ways to derive @xmath19 .
the usual way to eliminate the excited state from the schrdinger equation , written for the state @xmath20{l}% \alpha\\ \beta\\
\gamma \end{array } \right ] $ ] , @xmath21{l}% i\dot{\alpha}\left ( t\right ) = -\frac{\delta}{2}\alpha + \frac{\omega_{a}^{\ast}}{2}\gamma\\ i\dot{\beta}\left ( t\right ) = \frac{\delta}{2}\beta+\frac { \omega_{b}^{\ast}}{2}\gamma\\ i\dot{\gamma}\left ( t\right ) = \frac{\omega_{a}}{2}\alpha + \frac{\omega_{b}}{2}\beta+\delta\gamma \end{array } \right . \label{esnat}\ ] ] consists in claiming @xmath22 , which implies , by solving the last equation @xmath23 and injecting ( [ gamel ] ) back into the dynamical equations for @xmath24 and @xmath25 , which yields @xmath26{l}% \alpha\\ \beta% \end{array } \right ] = h_{eff}\left [ \begin{array } [ c]{l}% \alpha\\ \beta% \end{array } \right ] , % \label{sysdif}\ ] ] where the effective two - level hamiltonian writes@xmath27{ll}% \frac{\delta}{2}+\frac{\left\vert \omega_{a}\right\vert ^{2}}{4\delta } & \frac{\omega_{r}^{\ast}}{2}\\ \frac{\omega_{r}}{2 } & -\frac{\delta}{2}+\frac{\left\vert \omega _ { b}\right\vert ^{2}}{4\delta}% \end{array } \right ] , \quad \omega_{r}\equiv \frac{\omega_{a}\omega_{b}^{\ast}}{2\delta } = \left\vert \omega_{r}\right\vert e^{i\phi}. \label{hameff}\ ] ] if we had performed the same calculation in a shifted picture defined by the transformation @xmath28{l}%
\widetilde{\alpha}_{\eta}\\ \widetilde{\beta}_{\eta}\\ \widetilde{\gamma}_{\eta}% \end{array } \right ] $ ] , we would have obtained @xmath21{l}% i\dot{\widetilde{\alpha}}_{\eta}\left ( t\right ) = \left ( \eta\delta-\frac{\delta}{2}\right ) \widetilde{\alpha}_{\eta}+\frac{\omega _ { a}^{\ast}}{2}\widetilde{\gamma}_{\eta}\\ i\dot{\widetilde{\beta}}_{\eta}\left ( t\right ) = \left ( \eta\delta+\frac{\delta}{2}\right ) \widetilde{\beta}_{\eta}+\frac{\omega _ { b}^{\ast}}{2}\widetilde{\gamma}_{\eta}\\ i\dot{\widetilde{\gamma}}_{\eta}\left ( t\right ) = \frac{\omega _ { a}}{2}\widetilde{\alpha}_{\eta}+\frac{\omega_{b}}{2}\widetilde{\beta}_{\eta } + \left ( 1+\eta\right ) \delta\widetilde{\gamma}_{\eta}. \end{array } \right .
% \label{esshift}\ ] ] and , applying the same ansatz @xmath29 as before ( we assume @xmath30 ) , we would have derived @xmath31 which yields the effective hamiltonian @xmath32 in the shifted picture . finally , subtracting @xmath33 from @xmath32 we would have obtained the effective hamiltonian @xmath34 @xmath35{cc}% \frac{\delta}{2}+\frac{\left\vert \omega_{a}\right\vert ^{2}}{4\delta\left ( 1+\eta\right ) } & \frac{\omega_{r,\eta}^{\ast}}{2}\\ \frac{\omega_{r,\eta}}{2 } & -\frac{\delta}{2}+\frac{\left\vert \omega _ { b}\right\vert ^{2}}{4\delta\left ( 1+\eta\right ) } % \end{array } \right ] , \quad \omega_{r,\eta}\equiv\frac{\omega_{a}\omega_{b}^{\ast } } { 2\delta\left ( 1+\eta\right ) } = \left\vert \omega_{r,\eta}\right\vert e^{i\phi_{\eta}}\ ] ] we are thus led to the obviously unphysical conclusion that the effective hamiltonian , @xmath36 the effective dynamics of the system depends on the picture where the elimination is performed .
this raises questions about the ansatz we used : in which picture , if any , does it apply , and what is its physical meaning ? to understand better when and how to employ this scheme , we investigate two rigorous elimination methods in the next sections , which yield both an unambiguous expression of the effective hamiltonian and the level of accuracy of the approximation achieved .
in this section , we propose a straightforward elimination scheme based on the analysis of the exact expression of the amplitude @xmath37 of the excited state : resorting to a simple mathematical argument , we identify its relevant part @xmath38 which mainly contributes to the dynamics of @xmath39 ; injecting it back into the dynamical equations , we then derive an unambiguous expression for @xmath19 which yields an approximation of the dynamics of the system , the accuracy of which can be quantified .
this method moreover allows us to specify the applicability conditions of the previous scheme .
the derivation of the exact solutions of the schrdinger equation can be straightforwardly performed by finding the eigenenergies of the system and using the boundary condition @xmath40 to determine the coefficients of the fourier decompositions of the different amplitudes .
we do not reproduce these calculations but only summarize the results which are useful for our purpose .
introducing the reduced variables @xmath41 such that @xmath42 , @xmath43 , with @xmath44 , @xmath45 , and @xmath46 are not uniquely defined by these relations : the point is only to identify a common infinitesimal parameter @xmath47 for systematic expansion .
] , one readily shows @xmath21{l}% \alpha\left ( t\right ) = \sum_{k=1}^{3}a_{k}e^{-i\delta x_{k}t}\\ \beta\left ( t\right ) = \sum_{k=1}^{3}b_{k}e^{-i\delta x_{k}t}\\ \gamma\left ( t\right ) = \sum_{k=1}^{3}c_{k}e^{-i\delta x_{k}t}% \end{array } \right .
\label{exactsol}%\ ] ] where @xmath48 are the solutions of the equation @xmath49 and the coefficients @xmath50 are determined by the boundary conditions .
table [ tabres ] displays the expansions in @xmath47 of these different parameters .
.[tabres]expansions of the different parameters of the exact solutions . [ cols="^,^,^",options="header " , ] the rigorous solutions are represented in figure [ fig4 ] , in the regime @xmath51 : the amplitudes @xmath24 and @xmath25 show an oscillating behavior , which , as expected , is much alike a two - level rabi oscillation ; @xmath37 oscillates much faster than @xmath24 and @xmath25 .
[ ptb ] fig4.eps the comparison between the exact solutions and the approximations obtained in the previous subsection ( see figure [ fig5 ] ) shows that the previous elimination method provides a valuable approximation only in what we called `` the natural picture '' . in the following ,
we clarify why this is so through a straightforward and rigorous elimination scheme .
[ ptb ] fig5.eps consider the differential equation@xmath52 the solution of which writes @xmath53 if @xmath54 , @xmath55 and @xmath56 , with @xmath57 , @xmath58 and @xmath59 , one gets@xmath60 if one is interested in a solution valid to first order in @xmath47 , the @xmath61 term can be neglected , or , equivalently , one can forget the corresponding term directly in the differential equation which then reduces to @xmath62 in the same way , one straightforwardly shows that the solutions of the equation @xmath63 where @xmath64 is an arbitrary real number and all the other parameters obey the same relations as before , coincide ( to order @xmath65 ) with those of the simpler equation @xmath66 let us now return to our initial problem . injecting the exact expression ( [ exactsol ] ) of @xmath37 into ( [ esnat ] )
, we get the equations @xmath67 which are of the form ( [ eqnat ] ) , with @xmath68 neglecting the last term of each of these equations , or equivalently , replacing @xmath37 by its `` relevant '' component can easily be shown to be equal , up to second order terms in @xmath47 , to the average of @xmath37 over the period of its highest harmonic component @xmath69 . ]
@xmath70 then leads to approximate forms for @xmath39 , valid to order @xmath65 terms .
let us now relate @xmath38 to @xmath71 : we have @xmath72 + x_{3}\delta \gamma\left ( t\right ) \\
& = & i\dot{\gamma}_{rel}\left ( t\right ) -x_{3}\delta\gamma _ { rel}+x_{3}\delta\gamma\left ( t\right ) \\ & = & \frac{\omega_{a}}{2}\alpha+\frac{\omega_{b}}{2}\beta+\delta\gamma\end{aligned}\ ] ] whence @xmath73 and , as @xmath74 we deduce from ( [ gamrel ] ) that @xmath75 this expression coincides with ( [ gamel ] ) : it means that , in the natural picture , the rough adiabatic elimination procedure indeed leads to a valid approximation of the actual dynamics of the system , to order @xmath65 , which agrees with the remark we made about figure [ fig5 ] and legitimates the expression ( [ hameff ] ) for the effective hamiltonian . if we turn to the shifted picture defined by the transformation @xmath76 , the exact expression of the amplitude of the excited state is now @xmath77 and leads to the dynamical equations @xmath78 which are of the form ( [ eqshift ] ) . by similar calculations as above
, one shows that a good approximation of the dynamics of the system is then obtained through replacing @xmath79 by its relevant part @xmath80 this expression differs from the rough adiabatic elimination result ( [ gamshift ] ) . as can be easily checked , however , when @xmath81 , ( [ gamshift ] ) and ( [ gamrelshift ] ) only differ by @xmath61 terms which are not significant at the level of accuracy we consider .
the rough elimination procedure thus appears as a practical ( though physically not motivated ) trick which works as the long as the origin of the energies lies in the neighbourhood of the midpoint energy @xmath82 of the two lower states , or , to be more explicit , as long as @xmath83 .
the rigourous method we have employed here allowed us to clarify the applicability conditions of the rough procedure but it requires the exact solution of the schrdinger equation . in the next section , we present a more systematic and elegant approach based on the green s function formalism which can be used to generalise the results presented here to more complicated level schemes @xcite .
this overview summarizes the basic features of the green s function and projection operator formalism first introduced in @xcite and extensively presented in @xcite .
given a system of hamiltonian @xmath84 , comprising a leading part @xmath85 and a perturbation @xmath86 , one defines the green s function @xmath87 as follows @xmath88 the evolution operator of the system can then be derived through the formula@xmath89 where @xmath90 and @xmath91 denote two parallel lines just above and below the real axis , oriented from the right to the left and from the left to the right , respectively .
let @xmath92 be the subspace spanned by some relevant eigenstates of @xmath85 , and let @xmath93 and @xmath94 be the orthogonal projectors on @xmath92 and @xmath95 , respectively .
then , one can show @xmath96 where the displacement operator @xmath97 is defined by@xmath98 in our case , @xmath99{ccc}% -\frac{\delta}{2 } & 0 & 0\\ 0 & \frac{\delta}{2 } & 0\\ 0 & 0 & \delta \end{array } \right ] , \quad \widehat{v}=\frac{\hbar}{2}\left [ \begin{array } [ c]{ccc}% 0 & 0 & \omega_{a}^{\ast}\\ 0 & 0 & \omega_{b}^{\ast}\\ \omega_{a } & \omega_{b } & 0 \end{array } \right]\ ] ] whence @xmath100{cc}% \left\vert \lambda_{a}\right\vert ^{2 } & \lambda_{a}^{\ast}\lambda_{b}\\ \lambda_{a}\lambda_{b}^{\ast } & \left\vert \lambda_{b}\right\vert ^{2}% \end{array } \right ] .\ ] ] to compute @xmath101 one has to calculate the residues of @xmath102 at its poles which are the solutions of @xmath103 = 0 $ ] . setting @xmath104 , one gets@xmath105 which coincides with ( [ eqfreq ] ) and thus leads to the same results @xmath106 summarized in table [ tabres ] .
the associated residues @xmath107 are readily found to be @xmath108{ll}% 1 & \frac{\lambda_{a}\lambda_{b}^{\ast}% } { 2\lambda}\epsilon \\
\frac{\lambda_{a}^{\ast}\lambda_{b}}{2\lambda}\epsilon & 0 \end{array } \right ] + o\left ( \epsilon^{2}\right ) , \\
\mathcal{r}_{2 } & = & e^{-i\delta x_{2}t}\left [ \begin{array } [ c]{ll}% 0 & -\frac{\lambda_{a}\lambda_{b}^{\ast}}% { 2\lambda}\epsilon \\ -\frac{\lambda_{a}^{\ast}\lambda_{b}}{2\lambda}\epsilon & 1 \end{array } \right ] + o\left ( \epsilon^{2}\right ) , \\ \mathcal{r}_{3 } & = & o\left ( \epsilon^{2}\right)\end{aligned}\ ] ] when @xmath109 , while when @xmath110 @xmath111{ll}% \left\vert \lambda_{a}\right\vert ^{2 } & \lambda_{a}\lambda_{b}^{\ast } \\
\lambda_{a}^{\ast}\lambda_{b } & \left\vert \lambda_{b}\right\vert ^{2 } \end{array } \right ] + o\left ( \epsilon^{2}\right ) , \\
\mathcal{r}_{2 } & = & \frac{e^{-i\delta x_{2}t}}{\left\vert \lambda _ { a}\right\vert ^{2}+\left\vert \lambda_{b}\right\vert ^{2}}\left [ \begin{array } [ c]{ll}% \left\vert \lambda_{b}\right\vert ^{2 } & -\lambda_{a}\lambda_{b}^{\ast } \\
-\lambda_{a}^{\ast}\lambda_{b } & \left\vert \lambda _ { a}\right\vert ^{2 } \end{array } \right ] + o\left ( \epsilon^{2}\right ) , \\
\mathcal{r}_{3 } & = & o\left ( \epsilon^{2}\right ) .
\end{aligned}\ ] ] in both cases , the third pole only contributes to the second order in @xmath47 to the evolution operator ( [ projevol ] ) , whereas the first two are @xmath112 .
to the first order , the last pole can thus be omitted : this constitutes the so - called _ pole approximation_.
a simple and straightforward way to implement the pole approximation is to replace @xmath113 by @xmath114{cc}% \left\vert \lambda_{a}\right\vert ^{2 } & \lambda_{a}^{\ast}\lambda_{b}\\ \lambda_{a}\lambda_{b}^{\ast } & \left\vert \lambda_{b}\right\vert ^{2}% \end{array } \right]\ ] ] in ( [ projevol ] ) which boils down to replacing @xmath115 by @xmath116 in the expression of @xmath117 .
one readily shows that , as desired , the replacement of @xmath118 by @xmath119 in @xmath120 discards the irrelevant pole and its associated residue , while leaving the others unchanged ( up to @xmath61 terms ) .
moreover @xmath117 can now be put under the form @xmath121 where @xmath122 is a @xmath18 hermitian matrix , independent of @xmath123 which can be interpreted as the effective hamiltonian governing the dynamics of the reduced two - dimensional system @xmath124 : a simple calculation again leads to the expression ( [ hameff ] ) for @xmath19 .
let us finish this section by some remarks .
the calculations above have been performed assuming that the energy @xmath82 at the midpoint of the two ground states spanning @xmath92 is zero : if one shifts the energies so that @xmath125 , all the previous expressions and calculations hold , up to the replacement of @xmath123 by its translated @xmath126 , as can be easily checked .
the expression of the effective hamiltonian thus becomes @xmath127 which is of course consistent with the particular case @xmath128 considered above .
it is interesting to note that @xmath19 always takes the same form , wherever one chooses the origin of the energies , _
i.e. _ the sum of the projected unperturbed hamiltonian and the operator @xmath129 evaluated at the middle energy of the subspace @xmath92 . finally , let us also note that if @xmath129 is evaluated at an energy @xmath130 , the previous approximation remains valid , as the residues will only be affected by @xmath61 terms , which are not significant at the level of accuracy considered .
the goal of this article was to clarify the scheme usually employed to derive the effective two - dimensional hamiltonian of a lambda system excited by off - resonant lasers : in particular we have reviewed two methods which enabled us to rigorously derive the effective dynamics of the system , up to terms to well - established order of magnitude in a small parameter ; this study also allowed us to specify the applicability conditions of the rough elimination procedure .
the second of these schemes relies on the green s function formalism and the use of projectors on the different relevant subspaces of the state space of the system : this fairly general and elegant tool naturally leads to generalisations to more complicated multilevel systems , as shall be considered in a forthcoming paper . | this paper deals with different ways to extract the effective two - dimensional lower level dynamics of a lambda system excited by off - resonant laser beams .
we present a commonly used procedure for elimination of the upper level , and we show that it may lead to ambiguous results . to overcome this problem and better
understand the applicability conditions of this scheme , we review two rigorous methods which allow us both to derive an unambiguous effective two - level hamiltonian of the system and to quantify the accuracy of the approximation achieved : the first one relies on the exact solution of the schrdinger equation , while the second one resorts to the green s function formalism and the feshbach projection operator technique . |
since the advent of string theory , theories of spacetime with extra dimensions become prevalent in physics .
the idea of extra dimensional spacetime can be dated back to the work of kaluza and klein @xcite , who tried to propose a theory that can unify classical electrodynamics and gravity .
recently , intensive investigations on the casimir effect in spacetime with extra dimensions are undergoing . in the context of string theory , casimir effect was studied in @xcite .
the possible roles played by casimir energy as dark energy or cosmological constant was discussed in @xcite . the use of casimir effect in stabilizing extra dimensions were considered in @xcite . in the braneworld scenario ,
casimir effect was also considered in @xcite .
the influence of the extra dimensions on the casimir force acting on a pair of parallel plates in macroscopic ( 3 + 1)-dimensional spacetime was studied in @xcite . in the pioneering work of casimir @xcite
, it was shown that the casimir force acting on a pair of parallel perfectly conducting plates in ( 3 + 1)-dimensional spacetime is attractive .
it was confirmed later in the work of mehra @xcite and brown and maclay @xcite that the thermal correction would not change the sign of the casimir force .
the recent works @xcite explored the possible influence of the extra dimensions to the magnitude and sign of the casimir force . in @xcite ,
the casimir effect on a pair of parallel plates in spacetime with one extra dimension compactified to a circle was considered .
generalizations to extra dimensional space with more dimensions and more complicated geometries were considered in @xcite .
further generalizations to finite temperature casimir effect were studied in @xcite . in the works
@xcite , the spacetimes considered are the generalized kaluza - klein ( kk ) models of the form @xmath0 with metric @xmath1 where @xmath2 is the usual ( 3 + 1)-@xmath3 metric on the minkowski spacetime @xmath4 and @xmath5 is a riemannian metric on the @xmath6-dimensional compact internal space @xmath7 . in this model
, the metric is factorizable . hence the geometrical structures of the macroscopic manifold @xmath4 and the internal manifold @xmath7 are independent .
it has been concluded that the casimir force acting on a pair of parallel plates due to a scalar field with homogeneous boundary conditions where dirichlet conditions are imposed on both plates ( @xmath8 conditions ) or neumann conditions are imposed on both plates ( @xmath9 conditions ) is always attractive , at either zero or any finite temperature . on the other hand , for mixed boundary conditions where one of the plates assumes
dirichlet boundary condition and the other one assumes neumann boundary condition ( @xmath10 conditions ) , the casimir force is always repulsive . in @xcite ,
the spacetime considered is the randall - sundrum ( rs ) brane model .
this model was proposed in @xcite to solve the hierarchy problem between the planck and the electroweak scale . in this model ,
the underlying spacetime is a five - dimensional anti - de sitter space ( ads@xmath11 ) with background metric @xmath12this metric is non - factorizable .
the extra dimension with coordinate @xmath13 is compactified on the orbifold @xmath14 .
the metric of the underlying minkowski spacetime depends on the extra dimension through the warp factor @xmath15 , where @xmath16 determines the degree of curvature of the ads@xmath11 space .
there are two types of rs brane models , denoted by rsi and rsii respectively . in rsi , there are two 3-branes with equal and opposite tensions , one invisible and one visible , localized at @xmath17 and @xmath18 respectively , where @xmath19 is the compactification radius of the extra dimension .
the @xmath20-symmetry is realized by @xmath21 , @xmath22 .
the standard model fields are localized on the visible brane .
rsii can be considered as a limiting case of rsi where @xmath23 , i.e. , one brane is located at infinity . in relation to casimir effect on parallel plates
, rs model was generalized to @xmath24-branes with @xmath6-compact dimensions compactified to an @xmath6-torus embedded in a @xmath25-dimensional spacetime with background metric @xmath26 in @xcite and @xcite , it was concluded that the zero temperature casimir force acting on a pair of parallel plates in either the ( 4 + 1)-d rs model or its extension due to a massless scalar field with @xmath8 boundary conditions is always attractive .
the methods used in @xcite and @xcite involve approximations to the tower of masses induced by the extra dimension @xmath14 , and the attractive nature of the casimir force is not obvious from its analytical expressions .
it is also not clear whether the approximations used in deriving the casimir force would affect the conclusion about the sign of the casimir force .
therefore , it is desirable to obtain an exact expression for the casimir force . as mentioned in @xcite ,
the rs scenario is the simplest case of warped geometries .
the higher dimensional warped geometries deserve more attention especially in connection with string theory , which asserts that our spacetime should has eleven dimension . in this article
, we consider generalized rs model as in @xcite whose background metric is @xmath27compared to the model , the internal space now is an arbitrary @xmath6-dimensional compact manifold with riemannian metric @xmath5 .
it can be considered as a kk model embedded in a rs model .
therefore we call this model randall - sundrum - kaluza - klein ( rskk ) model .
our concern here is the casimir force acting on a pair of parallel plates rather than the casimir force acting on the branes which was considered in @xcite . as in most of the works about casimir effect on parallel plates in higher dimensional spacetime
, we regard the parallel plates as co - dimension one hyperplanes in the spacetime , and the field is assumed to propagate in the bulk .
we derive the exact formulas for the finite temperature casimir force and show that warped extra dimensions can not change the attractive or repulsive nature of the casimir force .
more precisely , it will be concluded that for @xmath8 or @xmath9 boundary conditions , the casimir force is always attractive ; whereas for @xmath10 boundary conditions , the casimir force is always repulsive .
the units used are such that @xmath28 .
in this section , we derive the casimir force acting on a pair of parallel plates in the rskk model with background metric due to a scalar field @xmath29 of mass @xmath30 with equation of motion @xmath31 using separation of variables , @xmath32the equation of motion for @xmath33 is equivalent to the following two equations : @xmath34 eq . is the equation of motion for a scalar field with effective mass @xmath35 in the kk spacetime . for the first equation in , the general solutions can be expressed in terms of the bessel functions of the first and second kinds @xmath36 and @xmath37 : @xmath38where @xmath39 in the massless case , there is an additional solution given by @xmath40constant and @xmath41 .
we first consider the rskki model , where two @xmath24-branes are localized at @xmath17 and @xmath18 respectively . in this case ,
boundary conditions imposed on the branes at @xmath17 and @xmath42 give rise to discrete spectrums of @xmath43 .
there are various possibilities @xcite of boundary conditions .
we choose the one considered in @xcite , where neumann boundary conditions are imposed on the branes , i.e. , @xmath44 . in the massless case ,
the constant functions satisfy this boundary conditions .
the general solution satisfies the boundary conditions at @xmath17 and @xmath18 if and only if @xmath45this gives a nontrivial solution to @xmath46 if and only if @xmath47 for simplicity , we only consider the massless case , i.e. , the @xmath48 case from now on . in this case , @xmath49 and becomes @xmath50multiplying eq . with @xmath51 and integrating over @xmath13 from @xmath17 to @xmath18 , we find that the effective masses @xmath35 should satisfy @xmath52 .
therefore we only consider the real solutions of .
one can verify that if @xmath35 is a solution of , so is @xmath53 .
however , @xmath35 and @xmath53 give rise to linearly dependent solutions of @xmath46 .
therefore we only need to consider the positive solutions of .
let @xmath54 be the effective mass corresponding to the solution @xmath55 = constant and let @xmath56 be all the positive solutions of .
denote by @xmath57 a nontrivial solution to with @xmath58 which satisfies the neumann boundary conditions at @xmath17 and @xmath18 . to find the casimir force acting on a pair of parallel plates
, we use the piston approach @xcite ( see fig . [ f1 ] ) .
= .45 the casimir force acting on the piston located at @xmath59 is given by @xmath60where @xmath61 is the casimir energy in a cavity of the form @xmath62\times[0,l_2]\times[0 , l_3]\times n^n \times s^1/\mathbb{z}_2 $ ] defined by @xmath63the summation runs through all @xmath64 which are the eigenfrequencies of the field @xmath29 satisfying the equation of motion , with appropriate boundary conditions on the boundary of the cavity . by letting @xmath65 , we obtain the casimir force acting on a pair of parallel plates embedded orthogonally inside an infinitely long rectangular cylinder @xcite : @xmath66 in this article , we impose dirichlet boundary conditions on the walls @xmath67 of the rectangular cylinder which are transversal to the @xmath68 direction . on the @xmath68 direction , we consider different combinations of boundary conditions .
a complete set of independent solutions to the equation with @xmath58 is given by @xmath69where for @xmath70 , @xmath71respectively for @xmath72 and @xmath10 boundary conditions .
for @xmath73 , @xmath74 is an eigenfunction of the laplace operator @xmath75 on @xmath7 with eigenvalue @xmath76 . by convention
, @xmath77 corresponds to the constant functions on @xmath7 .
the eigenfrequency of @xmath78 is @xmath79where @xmath80 for @xmath72 and @xmath10 boundary conditions respectively .
the finite temperature casimir force can be calculated in the same way as in @xcite , which gives @xmath81the sum of the terms with @xmath82 is the finite temperature casimir force acting on a pair of parallel plates in the kk model due to a massless scalar field @xcite . notice that each summand in the summation of is positive .
therefore , the sign of the casimir force is governed by the factor @xmath83 in front of the summation , which is negative for @xmath84 and positive for @xmath85 . as a result
, we find that the casimir force is always attractive at any temperature for @xmath8 and @xmath9 boundary conditions , but always repulsive for @xmath10 boundary conditions .
this shows that the warped extra dimension can not change the sign of the casimir force , but it increases the strength of the casimir force . for the influence of the internal extra dimension , notice that the sum of the terms with @xmath86 in corresponds to the casimir force in the absence of the internal space @xmath7 .
using again the fact that each summand of the summation in is positive , one finds that the extra dimensions enhance the magnitude of the casimir force .
moreover , the casimir force becomes stronger in the presence of more extra dimensions .
when the size of the internal manifold shrinks to zero , only the terms with @xmath86 in give a nonzero limit : @xmath87which is the finite temperature casimir force acting on a pair of parallel plates in the rsi model . as was proved in @xcite , when @xmath88 , the leading term of the casimir force is of order @xmath89 .
divide by @xmath89 and take the limit where @xmath90 , one finds that the casimir force density acting on a pair of infinite parallel plates in rskki model is @xmath91 here @xmath92 is the modified bessel function of the second kind .
the sum of the @xmath93 term is understood as @xmath94 taking the zero temperature limit of , we find that the zero temperature casimir force acting on a pair of parallel plates in the rskki model is @xmath95 in the limit where @xmath89 is large , we obtain the zero temperature casimir force density acting on a pair of infinite parallel plates : @xmath96the sum of the terms in with @xmath97 should be understood as @xmath98 which is the the zero temperature casimir force density @xmath99 acting on a pair of infinite parallel plates in the @xmath100-dimensional minskowski spacetime due to a massless scalar field . in the limit of vanishing internal space @xmath7 , we obtain the zero temperature casimir force density acting on a pair of infinite parallel plates in the rsi model : @xmath101 by using the approximation @xmath102 in as in @xcite , we find that in the case of @xmath8 boundary conditions , i.e. , @xmath103 , we obtain the same result as derived in @xcite ( formula ( 2.18 ) ) . in the general case where there are @xmath6 extra dimensions compactified to a torus @xmath104 on the branes , with radius @xmath105 respectively
, the authors in @xcite used the approximation @xmath106which is only good if the internal manifold has dimension @xmath107 . using this approximation in with @xmath103
, it seems that there are still some discrepancies between the result of @xcite ( see formula ( 32 ) in @xcite ) and our result in the terms corresponding to @xmath82 .
however , if one applies the chowla - selberg formula for epstein zeta functions to the three terms on the first two lines of ( 32 ) in @xcite , one would recover the sum of the terms with @xmath82 .
compared to the works of @xcite , our results and do not use any approximations to the effective masses .
they are exact results .
moreover , our formulas and show manifestly that the casimir force is attractive for dirichlet - dirichlet boundary conditions .
since rskkii model is a limit of the rskki model when the compactifying radius @xmath108 of the extra dimension with coordinate @xmath13 becomes infinite , one would expect that the same conclusions about the sign of the casimir force still hold for the rskkii model . from
, one can show that as @xmath109 , @xmath110 is approximately equal to @xmath111 for some @xmath112 .
this implies that @xmath113 can be considered as the eigenvalues coming from an extra dimension compactified to a ( twisted ) torus of radius @xmath114 .
therefore shows that for @xmath108 large enough , the magnitude of the casimir force is increased if we increase the compactifying radius @xmath108 . in the special cases considered in @xcite ,
this behavior was confirmed by the figures in @xcite .
in fact , the same argument as in @xcite shows that the casimir force will be proportional to the radius @xmath115 of the extra dimension .
therefore when all the parameters except @xmath108 are kept fixed and @xmath116 , the casimir force is proportional to @xmath117 , and if @xmath108 goes to infinity in such a way that @xmath118 is kept fixed , then the casimir force is proportional to @xmath108 . as a result
, we can not take the limit @xmath116 on the casimir force directly . in @xcite ,
finite results have been claimed for the casimir force density for the rsii model . in these articles ,
the authors retained the @xmath82 terms in or , and change the summation over @xmath119 to an integral over @xmath30 .
the latter is tantamount to dividing by the factor @xmath117 and taking the limit @xmath116 on the summation of @xmath119 terms .
this combination of retaining the @xmath82 term and changing the summation to integral is not equivalent to the direct @xmath116 limit .
as we have discussed , one should expect that in the limit @xmath116 , the casimir force density acting on infinite parallel plates increases beyond all bounds .
nevertheless , it is easy to see that the procedure used in @xcite gives casimir force that is attractive for homogeneous boundary conditions and repulsive for mixed boundary conditions .
as is discussed earlier , the presence of the extra dimensions enhances the casimir force
. it will be interesting to study whether the extra dimensions give a significant increase to the casimir force .
it is sufficient to consider the case with @xmath8 boundary conditions .
we discuss first the case of the rsi model where there is no internal space . at any temperature , numerical computations of and show that for plate separation @xmath120 in the range @xmath121 nm @xmath122 @xmath123 nm , the correction to the casimir force in @xmath100-d spacetime is less than @xmath124 for @xmath1253ev . for @xmath126 nm ,
the correction to the casimir force becomes @xmath127 when @xmath128 1.40ev ; and for @xmath129 nm , the correction to the casimir force becomes @xmath127 when @xmath128 0.13ev .
as is well known , the resolution of the hierarchy problem requires @xmath130 . in this case , the correction to the casimir force is significant if @xmath131 gev . if @xmath16 is of planck scale @xmath132gev , then @xmath133400gev @xmath134 3ev , and the correction to the casimir force would be too small to be observed in the current casimir experiments .
this situation can be changed if there exists an internal space that has size comparable to the plate separation .
assume that the internal space is @xmath135 with radius @xmath136 . in the absence of the extra dimension @xmath14 or when @xmath137 3ev
, the correction to the casimir force is less than @xmath124 if @xmath138 .
if @xmath139 , then the correction will grow to @xmath127 . when @xmath140 , the correction is @xmath141 .
however , it is believed that the internal manifold should be compactified to a size much smaller than @xmath142 nm , therefore the existence of internal space would not be able to be detected by the present casimir experiments which measures casimir force for separations larger than @xmath142 nm .
= .49 = .49 in fig .
[ f2 ] and fig .
[ f3 ] , we show the casimir pressures and for @xmath8 boundary conditions in @xmath100-dimensional minskowski spacetime , in rsi spacetime with @xmath143gev and @xmath144 , in kk spacetime where the internal manifold is a circle with radius @xmath145 nm , and in rskki spacetime , at zero temperature and at temperature @xmath146mev ( @xmath147 k ) respectively .
the parameters are chosen so that there are significant differences between the casimir pressures in various spacetimes .
compare fig .
[ f2 ] and fig .
[ f3 ] , we see that high temperature has profound effect on the strength of the casimir force , as dictated by the linear dependence of the casimir force in temperature in the high temperature regime .
= .49 = .49
we have given a brief discussion about the casimir force acting on a pair of parallel plates in higher dimensional warped spacetime model which are generalizations of @xmath148-d randall - sundrum spacetime model .
the contributions of this article are the followings .
we have derived exact formulas for the casimir force acting on a pair of parallel plates due to a massless scalar field without using any approximations . from the exact formulas
, we showed that the casimir force is always attractive for dirichlet - dirichlet or neumann - neumann boundary conditions , and repulsive for dirichlet - neumann conditions .
although the discussions in this article are restricted to the extension of randall - sundrum model , which we call randall - sundrum - kaluza - klein model , and specific boundary conditions have been imposed on the branes , the discussions in this article can be applied in a more general context . notice that the change in the boundary conditions on the branes only alters the spectrum of the effective masses , and the change in the geometry of the extra dimensions only alter the spectrum of the internal manifold . therefore the results of this article can be applied to more general spacetime model with extra dimensions .
one can deduce as in this article that in the absence of tachyonic modes , the sign of the casimir force due to a scalar field acting on a pair of parallel plates in the macroscopic @xmath100-dimensional minskowski spacetime only depends on the boundary conditions imposed on the plates , and is not changed by the presence of extra dimensions .
the extra dimensions only enhance the magnitude of the casimir force .
it should be mentioned that there is a recent work @xcite that proposed a different perspective on the casimir force acting on parallel plates in spacetime with extra dimensions .
the approach considered in this article permits the field to propagate in the bulk , and we are in fact considering the casimir force acting on a codimension one hyperplane . in @xcite , it was shown that if the plates are localized on the visible brane , the correction to the casimir force should be much smaller .
finally , we would also like to mention that this article is a revised version of the preprint @xcite . a recent interesting work @xcite generalized our work @xcite and discussed the casimir effect on parallel plates in the usual @xmath148-d rs model , but with general robin boundary conditions on the plates . in case of @xmath72 or @xmath10 boundary conditions ,
the authors of @xcite confirmed that they obtained the same result as our eq . .
the author would like to thank a. flachi for explaining his work to us .
we would also like to thank the anonymous referee for the helpful comments .
this project is funded by ministry of science , technology and innovation , malaysia under e - science fund 06 - 02 - 01-sf0080 .
i. brevik and a. a. bytsenko , in mathematical methods in physics : proceedings of the 1999 londrina winter school , world scientific ( 2000 ) .
l. hadasz , g. liambiase and v. v. nesterenko , phys .
d * 62 * ( 2000 ) , 025011 .
r. obousy and . g. cleaver , arxiv:0810.1096 .
r. obousy and a. a. saharian , arxiv:0906.4601 .
a.a . saharian and m.r .
setare , phys .
b * 552 * ( 2003 ) , 119 .
a. knapman and d.j .
toms , phys .
d * 69 * ( 2004 ) , 044023 .
saharian and m.r .
setare , phys .
b * 584 * ( 2004 ) , 306 .
saharian and m.r .
setare , nucl .
b * 2005 * , 406 . a. a. saharian , _ casimir densities in brane models with compact internal spaces _ , in _ the problems of modern cosmology : a volume in honour of prof
. s. d. odintsov on the occasion of his 50th birthday , tomsk _ , pp . 284 | in this article , we derive the finite temperature casimir force acting on a pair of parallel plates due to a massless scalar field propagating in the bulk of a higher dimensional brane model .
in contrast to previous works which used approximations for the effective masses in deriving the casimir force , the formulas of the casimir force we derive are exact formulas .
our results disprove the speculations that existence of the warped extra dimension can change the sign of the casimir force , be it at zero or any finite temperature . |
the study of information inequalities is a subfield of information theory that describes linear constraints on the entropies of finite collections of jointly distributed discrete random variables .
historically , the known information inequalities were orignally all special cases of shannon s conditional mutual information inequality @xmath0 , but later were generalized to other types of inequalities , called non - shannon inequalities .
information inequalities have been shown to be useful for computing upper bounds on the network coding capacities of certain networks .
analagously , the study of linear rank inequalities is a topic of linear algebra , which describes linear constraints on the dimensions of collections of subspaces of finite dimensional vector spaces .
in fact , the set of all information inequalities can be viewed as subclass of the set of all linear rank inequalities .
information inequalities hold over all collections of a certain number of random variables . in constrast
, linear rank inequalities may hold over only certain vector spaces , such as those whose scalars have particular field characteristics . in this paper , we present two new linear rank inequalities over finite fields , which are not information inequalities , and with the peculiar property that they only hold for certain fields , depending on the associated vector space .
the first inequality is shown to hold over all vector spaces when the field characteristic is anything but three ( theorem [ thm : t8 ] ) , but does not always hold when the field characteristic is three ( theorem [ thm : t8-char3 ] ) .
in contrast , the second inequality is shown to hold over all vector spaces when the field characteristic is three ( theorem [ thm : nont8 ] ) , but does not always hold when the field characteristic is not three ( theorem [ thm : non - t8-non - char3 ] ) .
we also show how these inequalities can be used to obtain bounds on the capacities of certain networks ( corollaries [ cor : t8-capacity ] and [ cor : nont8-capacity ] ) .
it will be assumed that the reader has familiarity with linear algebra , finite fields , information theory , and network coding .
nevertheless , we will give some brief tutorial descriptions of these topics for completeness .
in 2000 , ahlswede , cai , li , and yeung introduced the field of network coding @xcite and showed that coding can outperform routing in directed acyclic networks .
there are presently no known algorithms to determine the capacity or the linear capacity of a given network .
in fact , it is not even known if such algorithms exist .
information inequalities are linear inequalities that hold for all jointly distributed random variables , and shannon inequalities are information inequalities of a certain form @xcite .
both are defined in section [ section : lri ] .
it is known @xcite that all information inequalities containing three or fewer variables are shannon inequalities .
the first `` non - shannon '' information inequality was of four variables and was published in 1998 by zhang and yeung @xcite . since 1998
, various other non - shannon inequalities have been found , for example , by lnnika @xcite , makarychev , makarychev , romashchenko , and vereshchagin @xcite , zhang @xcite , zhang and yeung @xcite , dougherty , freiling , and zeger @xcite , and mat @xcite .
additionally , in 2007 , mat demonstrated an infinite collection of independent non - shannon information inequalities @xcite and there were necessarily an infinite number of such inequalities . in 2008 ,
xu , wang , and sun @xcite also gave an infinite list of inequalities but did not establish their necessity .
there is a close connection between information inequalities and network coding @xcite .
capacities of some networks have been computed by finding matching lower and upper bounds @xcite .
lower bounds have been found by deriving coding solutions .
upper bounds have been found by using information inequalities and treating the sources as independent random variables that are uniformly distributed over the alphabet .
one `` holy grail '' problem of network coding is to develop an algorithm to compute the coding capacity of an arbitrary network .
if such an algorithm exists , information inequalities may potentially play a role in the solution .
it has been shown that linear codes are insufficient for network coding in general @xcite . however , linear codes may be desirable to use in practice due to ease of analysis and implementation .
it has been shown that the coding capacity is independent of the alphabet size @xcite .
however , the linear coding capacity is dependent on alphabet size , or more specifically the field characteristic .
in other words , one can potentially achieve a higher rate of linear communication by choosing one characteristic over another . to provide upper bounds for
the linear coding capacity for a particular field one can look at linear rank inequalities @xcite .
linear rank inequalities are linear inequalities that are always satisfied by ranks of subspaces of a vector space .
all information inequalities are linear rank inequalities but not all linear rank inequalities are information inequalities .
the first example of a linear rank inequality that is not an information inequality was found by ingleton @xcite .
information inequalities can provide an upper bound for the capacity of a network , but this upper bound would hold for all alphabets . therefore , to determine the linear coding capacity over a certain characteristic one would have to consider linear rank inequalities . all linear rank inequalities up to and including five variables are known and none of these depend on the vector spaces field characteristics @xcite .
the set of all linear rank inequalities for six variables has not yet been determined .
characteristic - dependent linear rank inequalities are given , for example , in @xcite and @xcite .
an inequality is given in @xcite which is valid for characteristic two and another inequality is given which is valid for every characteristic except for two .
these inequalities are then used to provide upper bounds for the linear coding capacity of two networks . in the present paper , we give two characteristic - dependent linear rank inequalities on eight variables .
one is valid for characteristic three and the other is valid for every characteristic except for three .
these inequalities are then used to provide upper bounds for the linear coding capacity of two networks .
it is our intention that the techniques presented here may prove useful or otherwise motivate further progress in determining network capacities . in this section a very brief review of matroids
is given which will enable discussion in subsequent sections of a matroid - based method for constructing a particular network that helps in the derivation of the linear rank inequalities presented in this paper .
a matroid is an abstract structure that captures a notion of independence " that is found in finite dimensional vector spaces , graphs , and various other mathematical topics .
we will follow the notation and results of @xcite . a _ matroid _
, @xmath1 , is a pair @xmath2 , where @xmath3 is a finite set and @xmath4 is a set of subsets of @xmath3 that satisfies the following properties : 1 .
@xmath6 , if @xmath7 , then @xmath8 .
3 . @xmath6 , if @xmath9 and @xmath10 , then @xmath11 such that @xmath12 .
the sets in @xmath4 are called _
independent sets_. if a subset of @xmath3 is not in @xmath4 , then it is called _
dependent_. an example of a matroid is obtained from linear algebra .
let @xmath13 be a finite field and let @xmath14 be the vector space of all @xmath15-dimensional vectors whose components are elements of @xmath13 .
suppose @xmath16 is an @xmath17 matrix over @xmath13 .
let @xmath18 and @xmath4 be the set of all @xmath19 such that the multiset of columns of @xmath16 indexed by the elements of @xmath20 is linearly independent in the vector space @xmath14 . then @xmath21 is a matroid called the _ vector matroid _ of @xmath16 .
a matroid is said to be _
representable _ over the field @xmath13 if it is isomorphic to some vector matroid over @xmath14 .
for example , if @xmath13 is the binary field and @xmath22 where @xmath23 denote the columns of @xmath16 from left to right , then @xmath21 is a vector matroid of @xmath16 , where @xmath24 and @xmath25 a _ base _ is a maximal independent set .
let @xmath26 denote the set of all bases of a matroid @xmath1 . in our example
, @xmath27 it is well known that all the bases of a matroid are of the same cardinality .
if we let @xmath19 and @xmath28 , then it is easy to see that @xmath29 is a matroid .
the _ rank _ of @xmath20 , denoted by @xmath30 , is defined to be the cardinality of a base in @xmath31 . in our example , @xmath32 .
a _ circuit _ is a minimal dependent set .
the circuits in our example are @xmath33 . in this section
we will use the information theoretic concepts of entropy and mutual information to define and use the linear algebraic concept of linear rank inequalities .
connections between information inequalities and linear rank inequalities is also discussed .
let @xmath34 be collections of discrete random variables over a finite alphabet @xmath35 , and let @xmath36 be the probability mass function of @xmath16 .
the _ entropy _ of @xmath16 is defined by @xmath37 the _ conditional entropy _ of @xmath16 given @xmath38 is @xmath39 the _ mutual information _ between @xmath16 and @xmath38 is @xmath40 and the _ conditional mutual information _ between @xmath16 and @xmath38 given @xmath41 is @xmath42 we will make use of the following basic information - theoretic facts @xcite : @xmath43 the equations ( [ h5])-([h9 ] ) were originally given by shannon in 1948 @xcite , and can all be obtained from the single inequality @xmath44 .
let @xmath45 be a positive integer , and let @xmath46 be subsets of @xmath47 .
let @xmath48 for @xmath49 .
a linear inequality of the form @xmath50 is called an _ information inequality _ if it holds for all jointly distributed random variables @xmath51 . as an example , taking @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , and using ( [ h8 ] ) shows that @xmath59 is an information inequality .
a _ shannon information inequality _ is any information inequality that can be expressed as a finite sum of the form @xmath60 where each @xmath61 is a nonnegative real number .
any information inequality that can not be expressed in the form above will be called a _
non - shannon information inequality_. linear rank inequalities are closely related to information inequalities .
in fact , in order to describe linear rank inequalities we will borrow notation from information theory to use in the context of linear algebra in the following manner .
suppose @xmath16 and @xmath38 are subspaces of a given vector space @xmath62 , and let @xmath63 denote the span of @xmath64 .
we will let @xmath65 denote the rank of @xmath16 , and let @xmath66 denote the rank of @xmath63 .
the meanings of some other information theoretic notation in the context of linear algebra then follows from - .
specifically , note that the conditional entropy notation @xmath67 denotes the excess rank of subspace @xmath16 over that of subspace @xmath68 , or equivalently , the codimension of @xmath68 in @xmath16 ; and the mutual information notation @xmath69 denotes the rank of @xmath68 .
a _ linear rank inequality _ over a vector space @xmath62 is a linear inequality of the form in , that is satisfied by every assignment of subspaces of @xmath62 to the variables @xmath70 .
all information inequalities are linear rank inequalities over all finite vector spaces , but not all linear rank inequalities are information inequalities . for background material on these concepts ,
the reader is referred to hammer , romashchenko , shen , and vereshchagin @xcite .
the first known example of a linear rank inequality over all finite vector spaces that is not an information inequality is the _ ingleton inequality _
@xcite : @xmath71 to see that the ingleton inequality is not an information inequality , let @xmath72 be binary random variables , and let @xmath73 with probabilities : @xmath74 then the ingleton inequality fails since : @xmath75 in this section , we will briefly review some concepts of network coding .
this will enable the discussion later in this paper of our construction of linear rank inequalities using networks constructed from two particular matroids ( t8 and non - t8 ) .
for more details on network coding , see @xcite .
a _ network _ is a finite , directed , acyclic multigraph with messages and demands .
messages _ are arbitrary vectors of @xmath76 symbols over a finite alphabet @xmath77 . each network edge carries a vector of @xmath78 symbols from @xmath77 .
each message originates at a particular node called the _ source node _ for that message and is required by one or more _
demand nodes_. when we draw a network , a message variable appearing above a node indicates the message is generated by such node and [ fig : nont8 ] , for convenience , we label source messages above nodes lying in both the top and bottom layers in each diagram .
this is meant to indicate that there is , in fact , a separate ( but hidden ) distinct node for each such source message , whose out - edges go directly to the nodes labeled by the source message in the top and bottem layers . ] , and a message variable appearing below a node indicates the message is demanded by such node , for a given network , the values of @xmath76 and @xmath78 can be chosen in order to implement certain codes and to obtain certain throughput @xmath79 .
the inputs to a network node are the vectors carried on its in - edges as well as the messages , if any , generated at the node .
the outputs of a network node are the packets carried on its out - edges as well as any demanded messages at the node .
each output of a node must be a function only of its inputs .
a _ coding solution _ for the network is an assignment of such functions to the network edges .
when the values of @xmath76 and @xmath78 need to be emphasized , the coding solution will be called a @xmath80-coding solution .
capacity _ of a network is defined as : @xmath81 a solution is called a _ linear solution _ , if the alphabet @xmath77 is a finite field and the edge functions are linear ( i.e. linear combinations of their input vectors where the coefficients are matrices over the field ) . the _ linear capacity _ is defined the same as the capacity but restricting solutions to be linear .
it is also easily verified that if @xmath82 is a message , then @xmath83 , and if @xmath82 is a vector carried by an edge , then @xmath84 . the butterfly network with source messages @xmath82 and @xmath85 , generated by source nodes @xmath86 and @xmath87 , respectively .
demand nodes @xmath88 and @xmath89 demand messages @xmath85 and @xmath82 , respectively . ]
let us illustrate a method for finding capacity bounds by examining the well - known butterfly network , depicted in figure [ fig : butterfly ] .
we assume the network messages @xmath82 and @xmath85 are independent , @xmath76-dimensional , random vectors with uniformly distributed components .
then in any solution it must be the case that @xmath90 since @xmath85 is a function of @xmath82 and @xmath91 , and also that @xmath92}\nonumber\\ & \leq h(x , y , z ) & \text { [ from \eqref{h9}]}\nonumber\\ & = h(x , z ) + h(y|x , z ) & \text{[from \eqref{h1}]}\nonumber\\ & = h(x , z ) & \text{[from ( \ref{butterfly:1})]}\nonumber\\ & \le h(x ) + h(z ) & \text{[from \eqref{h8}]}\nonumber\\ & \le k + n. \label{rv - inequality}\end{aligned}\ ] ] this implies @xmath93 , or equivalently @xmath94 .
since this bound holds for all choices of @xmath76 and @xmath78 , the coding capacity must be at most @xmath95 .
on the other hand , a solution with @xmath96 is obtained by taking @xmath97 over any finite field alphabet , so the coding capacity is at least @xmath95 .
thus the coding capacity for the butterfly network is the same as the linear coding capacity which is exactly equal to @xmath95 .
the inequalities in were based on random variables @xmath98 .
later , in the proofs of corollaries [ cor : t8-capacity ] and [ cor : nont8-capacity ] , we will obtain bounds on the capacities of networks by using linear rank inequalities , instead of information inequalities . in those cases , certain vector subspaces
will be used instead of random variables , but the procedure will appear similar .
in this section , we given some technical lemmas which will be useful for proving the main results of the paper . if @xmath16 is a subspace of vector space @xmath62 , and @xmath99 is a subspace of @xmath16 , then we will use the notation @xmath100 to represent the codimension of @xmath99 in @xmath16 .
we will omit the subscript when it is obvious from the context which space the codimension is with respect to .
[ lemma1 ] @xcite let @xmath62 be a finite dimensional vector space with subspaces @xmath16 and @xmath38 .
then the subspace @xmath101 has codimension at most @xmath102 in @xmath62 .
we know @xmath103 .
then adding @xmath104 to both sides of the inequality gives @xmath105 .
thus , @xmath106 .
[ lemma2 ] @xcite let @xmath16 and @xmath38 be vector spaces over the same finite scalar field and with subspaces @xmath99 and @xmath107 , respectively .
let @xmath108 be a linear function such that @xmath109 .
then the codimension of @xmath99 in @xmath16 is at most the codimension of @xmath107 in @xmath38 .
suppose a base for @xmath16 consists of a base for @xmath99 together with the vectors @xmath110 .
let @xmath111 be field elements which are not all zero .
then @xmath112 , so @xmath113 . thus , the vectors @xmath114 are linearly independent over the subspace @xmath107 , and therefore @xmath115 .
[ lemma3 ] @xcite let @xmath16 and @xmath38 be vector spaces over the same finite scalar field , let @xmath107 be a subspace of @xmath38 , and let @xmath108 be a linear function .
then @xmath116 on a subspace of @xmath16 of codimension at most the codimension of @xmath107 .
let @xmath117 . then @xmath109 and the result follows from lemma [ lemma2 ] .
[ lemma4 ] @xcite let @xmath62 be a finite dimensional vector space and let @xmath118 be subspaces of @xmath62 .
then for @xmath119 , there exist linear functions @xmath120 such that @xmath121 on a subspace of @xmath38 of codimension @xmath122 .
let @xmath123 be a subspace of @xmath38 defined by @xmath124 .
the subspace on which this lemma holds is @xmath123 .
if @xmath125 , then the lemma would be trivially true .
so , assume that @xmath126 , and let @xmath127 be a basis for @xmath123 . for each @xmath128
, choose @xmath129 for @xmath130 such that @xmath131 . for each @xmath132 , define a linear mapping @xmath133 so that @xmath134 for all @xmath135 and @xmath136 . then extend @xmath137 arbitrarily to @xmath138 .
now we have linear functions @xmath139 such that @xmath121 on @xmath123 .
the dimension of @xmath123 is @xmath140 , so the codimension of @xmath123 is @xmath141 .
[ lemma5 ] @xcite let @xmath62 be a finite - dimensional vector space and let @xmath142 , and @xmath41 be subspaces of @xmath62 .
let @xmath143 and @xmath144 be linear functions such that @xmath145 on @xmath16 .
then @xmath146 on a subspace of @xmath16 of codimension at most @xmath147 .
let @xmath148 be the kernel of @xmath149 . clearly , @xmath149 maps @xmath16 into @xmath150 and since @xmath149 is linear the rank of its domain is at most the sum of the ranks of its kernel and range , so @xmath151 [ lemma6 ] @xcite let @xmath62 be a finite dimensional vector space and let @xmath152 be subspaces of @xmath62 . for
each @xmath132 let @xmath153 be a linear function such that @xmath154 on @xmath16 .
then @xmath155 on a subspace of @xmath16 of codimension at most @xmath156 .
first we apply lemma [ lemma5 ] to @xmath157 and @xmath158 to get @xmath159 on a subspace @xmath160 of @xmath16 of codimension at most @xmath161 . then apply lemma [ lemma5 ] to @xmath162 and @xmath163 to get @xmath164 on a subspace @xmath165 of @xmath160 of codimension at most @xmath166 .
continue on until we apply lemma [ lemma5 ] to @xmath167 and @xmath168 to get @xmath169 on a subspace @xmath170 of @xmath171 of codimension at most @xmath172 .
now @xmath170 is a subspace of @xmath16 of codimension at most @xmath173 , on which @xmath174 .
[ lem inj ] let @xmath175 be subspaces of a vector space @xmath62 and let @xmath176 , and @xmath177 be functions such that @xmath178 and @xmath179 .
if @xmath180 on @xmath16 and @xmath181 is injective on @xmath38 , then @xmath182 is injective on @xmath183 .
let @xmath184 .
we know @xmath185 on @xmath186 because @xmath187 for all @xmath188 . since @xmath189 , we know @xmath190 , which implies @xmath191 for some @xmath192 .
similarly , we know @xmath193 for some @xmath194 .
so , we have @xmath195 and @xmath196 .
if we assume @xmath197 , then we have @xmath198 . since @xmath181 is injective on @xmath38 , we know @xmath199 .
thus @xmath200 , which implies @xmath201 .
since @xmath180 on @xmath16 , we know @xmath202
. thus @xmath182 is injective on @xmath183 .
in this section , we use the known t8 matroid to construct a `` t8 network '' , and then in turn we use the t8 network to guide a construction of a `` t8 linear rank inequality '' that is shown to hold for all vector spaces having finite scalar fields of characteristic not equal to @xmath203 .
then we show that the t8 inequality does not necessarily hold when such scalar fields have characteristic @xmath203 . finally , we determine the exact coding capacity of the t8 network and its linear coding capacity over finite field alphabets of characteristic @xmath203 , as well as a linear capacity upper bound for finite field alphabets whose characteristic is not @xmath203 .
the t8 matroid is representable over a field if and only if the field is of characteristic 3 .
figure [ fig : t8 ] is a network whose dependencies and independencies are consistent with the t8 matroid .
it was designed by the construction process described in @xcite , and we will refer to it as the t8 network .
theorem [ thm : t8 ] uses the t8 network as a guide to derive a linear rank inequality valid for every characteristic except for 3 .
we refer to the inequality in the following theorem as the _
t8 linear rank inequality_. and @xmath205 generated at hidden source nodes with certain hidden out - edges pointing to corresponding displayed nodes @xmath86 , @xmath206 , @xmath88 , and @xmath207@xmath208 ( which are labeled by incoming messages above such nodes ) .
the nodes @xmath207@xmath208 each demand one message , as labeled below such nodes . ,
width=510 ] [ thm : t8 ] let @xmath209 , and @xmath210 be subspaces of a vector space @xmath62 whose scalar field is finite and of characteristic other than @xmath203 .
then the following is a linear rank inequality over @xmath62 : @xmath211 such that @xmath212 now combining some functions we obtained from lemma [ lemma4 ] gives four new functions : @xmath213 using ( [ p3:eq:1])([p3:eq:4 ] ) , ( [ p3:eq:11 ] ) , lemma [ lemma1 ] , and lemma [ lemma3 ] we know the sum of these four functions is equal to @xmath4 on a subspace of @xmath16 of codimension at most @xmath214 . now applying lemma [ lemma6 ] and lemma [ lemma1 ] to the functions @xmath215 , @xmath216 , @xmath217 , and @xmath218 , we get a subspace @xmath219 of @xmath16 of codimension at most @xmath220 on which @xmath221 similarly , we get a subspace @xmath99 of @xmath16 of codimension at most @xmath222 on which @xmath223 we get a subspace @xmath107 of @xmath38 of codimension at most @xmath224 on which @xmath225 we get a subspace @xmath226 of @xmath38 of codimension at most @xmath227 on which @xmath228 we get a subspace @xmath229 of @xmath41 of codimension at most @xmath230 on which @xmath231 we get a subspace @xmath232 of @xmath41 of codimension at most @xmath233 on which @xmath234 we get a subspace @xmath235 of @xmath205 of codimension at most @xmath236 on which @xmath237 let @xmath238 . considering ( [ p3:eq : bhat2 ] ) and ( [ p3:eq : chat3 ] ) , we can apply lemma [ lem inj ] to show that @xmath239 is injective on @xmath240 . by ( [ p3:eq : bhat3 ] )
, we know @xmath241 let @xmath242 . considering again ( [ p3:eq : bhat2 ] ) and ( [ p3:eq : chat3 ] )
, we can apply lemma [ lem inj ] to show that @xmath243 is injective on @xmath244 . by ( [ p3:eq : chat2 ] )
, we know @xmath245 let @xmath246 . considering ( [ p3:eq : abar1 ] ) and ( [ p3:f5f25 ] ) , we can apply lemma [ lem inj ] to show that @xmath247 is injective on @xmath248 . by ( [ p3:eq : abar3 ] ) , we know @xmath249 is injective on @xmath248 which implies @xmath250 let @xmath251 . considering ( [ p3:eq : cbar3 ] ) and ( [ p3:eq : dbar4]),we can apply lemma [ lem inj ] to show that @xmath252 is injective on @xmath253 .
then by ( [ p3:eq : cbar4 ] ) , we know @xmath254 let @xmath255 . considering ( [ p3:eq : bbar2 ] ) and , we can apply lemma [ lem inj ] to show that @xmath256 by ( [ p3:eq : bbar4 ] ) , we know @xmath257 which implies @xmath258 let us define the functions @xmath259 where @xmath260 and @xmath261 are the restrictions of the functions @xmath262 and @xmath263 to the sets @xmath248 and @xmath264 , respectively . now , considering ( [ p3:eq : bbar2 ] ) , ( [ p3:eq : bhat2 ] ) , ( [ p3:eq : cbar3 ] ) , and ( [ p3:eq : chat3 ] ) we have @xmath265 } \label{p3:f1}\\ f_2 & = -f_5f_{15}g_{14 } \mbox { on $ f_{14}\overline{a}^\ast$ } & { & [ \mbox{from ( \ref{p3:eq : abar3 } ) } ] } \label{p3:f2}\\ f_3 & = -f_6f_{15}g_{14 } \mbox { on $ f_{14}\overline{a}^\ast$ and } f_3 = -f_9f_{18}f_2 \mbox { on $ f_{17}\overline{c}$ } & { & [ \mbox{from ( \ref{p3:eq : abar4 } ) , ( \ref{p3:eq : cbar4 } ) } ] } \label{p3:f3}\\ f_4 & = -f_7f_{21}g_{20 } \mbox { on $ f_{20}\overline{b}^\ast$ } & { & [ \mbox{from ( \ref{p3:eq : bbar1 } ) } ] } \label{p3:f4}\\ f_6 & = -f_9f_{21}g_{20 } \mbox { on $ f_{20}\overline{b}^\ast$ } & { & [ \mbox{from ( \ref{p3:eq : bbar4 } ) } ] } \label{p3:f6}\\ f_7 & = -f_4f_{20}f_8 \mbox { on $ f_{21}\overline{b}$ } & { & [ \mbox{from ( \ref{p3:eq : bbar1 } ) } ] } \label{p3:f7}\\ f_9 & = -f_6f_{20}f_8 \mbox { on $ f_{21}\overline{b}$ } & { & [ \mbox{from ( \ref{p3:eq : bbar4 } ) } ] } \label{p3:f9}\\ f_{10 } & = -f_4f_{25}f_{11 } \mbox { on $ f_{27}\widehat{b}$ and } f_{10 } = -f_7f_{28}f_{12 } \mbox { on $ f_{29}\widehat{c}$ } & { & [ \mbox{from ( \ref{p3:eq : bhat1 } ) , ( \ref{p3:eq : chat1 } ) } ] } \label{p3:f10}\\ f_{11 } & = -f_8f_{28}f_{12 } \mbox { on $ f_{29}\widehat{c}$ } & { & [ \mbox{from ( \ref{p3:eq : chat2 } ) } ] } \label{p3:f11}\\ f_{12 } & = -f_5f_{25}f_{11 } \mbox { on $ f_{27}\widehat{b}$}. & { & [ \mbox{from ( \ref{p3:eq : bhat3 } ) } ] } \label{p3:f12}\end{aligned}\ ] ] next , we provide upper bounds for the codimensions of @xmath248 , @xmath240 , @xmath264 , @xmath244 , and @xmath253 . from ( [ p3:eq : bhat2 ] ) , we know @xmath266 is injective on @xmath267 and @xmath268 is injective on @xmath226 .
these facts will be used to arrive on lines ( [ p3:bhatstar:1 ] ) and ( [ p3:bhatstar:2 ] ) . from ( [ p3:eq : chat3 ] ) , we know @xmath269 is injective on @xmath232 , which will also be used to arrive on line ( [ p3:bhatstar:2 ] ) .
lemma [ lemma1 ] will be used to arrive on ( [ p3:bhatstar:3 ] ) .
@xmath270 from ( [ p3:eq : abar1 ] ) , we know @xmath271 is injective on @xmath272 and @xmath273 is injective on @xmath99 .
these facts will be used on lines ( [ p3:abarstar:1 ] ) and ( [ p3:abarstar:2 ] ) . from ( [ p3:f5f25 ] ) , we know @xmath274 is injective on @xmath240 , which will also be used to arrive on line ( [ p3:abarstar:2 ] ) .
lemma [ lemma1 ] will be used to arrive on ( [ p3:abarstar:3 ] ) .
@xmath275 from ( [ p3:eq : cbar3 ] ) , we know @xmath162 is injective on @xmath276 and @xmath277 is injective on @xmath229 .
these facts will be used to arrive on lines ( [ p3:cbarstar:1 ] ) and ( [ p3:cbarstar:2 ] ) . from ( [ p3:eq : dbar4 ] ) , we know @xmath278 is injective on @xmath235 , which will also be used on line ( [ p3:cbarstar:2 ] ) .
lemma [ lemma1 ] will be used to arrive on ( [ p3:cbarstar:3 ] ) .
@xmath279 from ( [ p3:eq : bbar2 ] ) , we know @xmath280 is injective on @xmath281 and @xmath282 is injective on @xmath107 .
these facts will be used to arrive on lines ( [ p3:bbarstar:1 ] ) and ( [ p3:bbarstar:2 ] ) . from ( [ p3:f9f18 ] ) , we know @xmath283 is injective on @xmath253 , which will also be used on line ( [ p3:bbarstar:2 ] ) .
lemma [ lemma1 ] will be used to arrive on ( [ p3:bbarstar:3 ] ) .
@xmath284 from ( [ p3:eq : chat3 ] ) , we know @xmath285 is injective on @xmath286 and @xmath269 is injective on @xmath232 .
these facts will be used to arrive on lines ( [ p3:chatstar:1 ] ) and ( [ p3:chatstar:2 ] ) . from ( [ p3:eq : bhat2 ] ) , we know @xmath268 is injective on @xmath226 , which will also be used on line ( [ p3:chatstar:2 ] ) .
lemma [ lemma1 ] will be used to arrive on ( [ p3:chatstar:3 ] ) .
@xmath287 let @xmath288 .
now , we will assume @xmath289 satisfies conditions ( [ p3:ass1])([p3:ass7 ] ) .
the justifications can be found below .
@xmath290 now , we need to make two assumptions on @xmath289 simultaneously . @xmath291 to justify ( [ p3:ass3 ] ) , first we know @xmath263 is injective on @xmath264 by ( [ p3:f6f20 ] ) .
then by lemma [ lemma3 ] , we know @xmath292 on a subspace of @xmath16 of codimension at most @xmath293 . by ( [ p3:f5f25 ] ) , we also know @xmath274 is injective on @xmath240 . then by lemma [ lemma3 ] , we know @xmath294 on a subspace of @xmath16 of codimension at most @xmath295 . then using lemma [ lemma1 ] , we have @xmath296 on a subspace of @xmath16 of codimension at most @xmath297 . conditions ( [ p3:ass4])([p3:ass6 ] ) can be justified similarly . to justify ( [ p3:ass7 ] ) ,
first we know @xmath277 is injective on @xmath229 by ( [ p3:eq : cbar3 ] ) .
then by lemma [ lemma3 ] , we know @xmath298 on a subspace of @xmath16 of codimension at most @xmath299 . by ( [ p3:f14 ] ) , we also know @xmath262 is injective on @xmath248 . then by lemma [ lemma3 ] ,
we know @xmath300 on a subspace of @xmath16 of codimension at most @xmath301 . then using lemma [ lemma1 ] , we have @xmath302 on a subspace , @xmath303 , of @xmath16 of codimension at most @xmath304 . since @xmath262 is injective on @xmath248 , the function @xmath305 is defined on @xmath303 . using the same technique as before we can show that @xmath306 on a subspace , @xmath307 , of codimension with respect to @xmath303 at most @xmath308 .
thus both conditions are true on @xmath307 , which has codimension with respect to @xmath16 at most @xmath309 .
our final goal is to show that @xmath310 for some @xmath82 so that we may conclude that @xmath311 if the characteristic is 3 .
we will accomplish this by using ( [ p3:eq : ahat1 ] ) and by proving that @xmath312 .
first we must show that @xmath314 . by ( [ p3:eq : ahat2 ] )
, we know @xmath315 then by using ( [ p3:f11 ] ) and condition ( [ p3:ass5 ] ) , we have @xmath316 now , by using ( [ p3:f1 ] ) and condition ( [ p3:ass7 ] ) , we have @xmath317 by ( [ p3:f8f28 ] ) , we know @xmath280 is injective on @xmath318 . by condition ( [ p3:ass4 ] ) , we know @xmath319 . by condition ( [ p3:ass6 ] ) , we know @xmath320 . by condition ( [ p3:ass5 ] ) , we know @xmath321 . using ( [ p3:eq : chat3 ] ) , we know @xmath322 .
thus , we have @xmath323 by ( [ p3:eq : ahat4 ] ) , we have @xmath324 then by using ( [ p3:f6 ] ) and condition ( [ p3:ass3 ] ) , we have @xmath325 now , by using ( [ p3:f3 ] ) and condition ( [ p3:ass7 ] ) , we have @xmath326 by ( [ p3:f9f21 ] ) , we know @xmath327 is injective on @xmath328 . by condition ( [ p3:ass4 ] ) , we know @xmath329 . by condition ( [ p3:ass3 ] ) , we know @xmath292 so @xmath330 . by condition ( [ p3:ass6 ] ) , we know @xmath331 .
thus , we have @xmath332 now , setting ( [ p3:claim1:1 ] ) and ( [ p3:claim1:2 ] ) equal to each other , we have @xmath333 by ( [ p3:f4 ] ) and condition ( [ p3:ass3 ] ) , we know @xmath334 using ( [ p3:claim1:3 ] ) , we have @xmath335 then using ( [ p3:f10 ] ) and condition ( [ p3:ass5 ] ) , we know @xmath336 first we must show that @xmath338 . by ( [ p3:eq : ahat3 ] ) ,
we know @xmath339 then by using ( [ p3:f12 ] ) and condition ( [ p3:ass5 ] ) , we have @xmath340 now , by using ( [ p3:f2 ] ) and condition ( [ p3:ass7 ] ) , we have @xmath341 by ( [ p3:f5f25 ] ) , we know @xmath342 is injective on @xmath343 . by condition ( [ p3:ass3 ] ) , we know @xmath294 . by condition ( [ p3:ass5 ] ) , we know @xmath344 . now , using ( [ p3:eq : bhat2 ] ) , we know @xmath345 . by condition ( [ p3:ass7 ] ) , we know @xmath346 .
thus , we have @xmath347 by ( [ p3:eq : ahat4 ] ) , we have @xmath348 then using ( [ p3:f9 ] ) and condition ( [ p3:ass4 ] ) , we have @xmath349 now , by using ( [ p3:f3 ] ) and condition ( [ p3:ass7 ] ) , we have @xmath350 by ( [ p3:f6f20 ] ) , we know that @xmath351 is injective on @xmath352 . by condition ( [ p3:ass3 ] ) , we know @xmath292 . by condition ( [ p3:ass4 ] ) , we know @xmath329 .
now , using ( [ p3:eq : bbar2 ] ) , we know @xmath353 . by condition ( [ p3:ass7 ] ) , we know @xmath354 .
thus , we have @xmath355 now , setting ( [ p3:claim2:1 ] ) and ( [ p3:claim2:2 ] ) equal to each other , we have @xmath356 by ( [ p3:f7 ] ) and condition ( [ p3:ass4 ] ) , we know @xmath357 using ( [ p3:claim2:3 ] ) , we have @xmath358 then using ( [ p3:f10 ] ) and condition ( [ p3:ass5 ] ) , we know @xmath359 now , by and the two claims , we have @xmath360 thus if the field has characteristic 3 , then @xmath361 no nonzero @xmath289 can satisfy all of the conditions ( [ p3:ass1])([p3:ass7 ] ) , so we must have @xmath362 [ thm : non - t8-non - char3 ] for each prime number @xmath363 there exists a vector space @xmath62 with a finite scalar field of characteristic @xmath36 such that the non - t8 inequality in theorem [ thm : nont8 ] is not a linear rank inequality over @xmath62 .
let @xmath62 be the vector space of @xmath364-dimensional vectors whose components are from @xmath365 , and define the following subspaces of @xmath62 : @xmath366 we have : @xmath367}\nonumber\\ & = h(x|a , c , d ) & { & [ \mbox{from $ ( 1,0,1,1)=(1,0,0,0)+(0,0,1,0)+(0,0,0,1)$}]}\nonumber\\ & = h(y|a , b , d ) & { & [ \mbox{from $ ( 1,1,0,1)=(1,0,0,0)+(0,1,0,0)+(0,0,0,1)$}]}\nonumber\\ & = h(z|a , b , c ) & { & [ \mbox{from $ ( 1,1,1,0)=(1,0,0,0)+(0,1,0,0)+(0,0,1,0)$}]}\nonumber\\ & = h(a|b , w , x ) & { & [ \mbox{from $ ( 1,0,0,0)=(1,0,1,1)+(0,1,0,0)-(0,1,1,1)$}]}\nonumber\\ & = h(c|a , w , y ) & { & [ \mbox{from $ ( 0,0,1,0)=(0,1,1,1)+(1,0,0,0)-(1,1,0,1)$}]}\nonumber\\ & = h(b|c , x , y ) & { & [ \mbox{from $ ( 0,1,0,0)=(1,1,0,1)+(0,0,1,0)-(1,0,1,1)$}]}\nonumber\\ & = h(d|a , w , z ) & { & [ \mbox{from $ ( 0,0,0,1)=(0,1,1,1)+(1,0,0,0)-(1,1,1,0)$}]}\nonumber\\ & = h(b|d , x , z ) & { & [ \mbox{from $ ( 0,1,0,0)=(1,1,1,0)+(0,0,0,1)-(1,0,1,1)$}]}\nonumber\\ & = h(c|d , y , z ) & { & [ \mbox{from $ ( 0,0,1,0)=(1,1,1,0)+(0,0,0,1)-(1,1,0,1)$}]}\nonumber\\ & = h(a|w , x , y , z ) & { & [ \mbox{from $ ( 1,0,0,0)=3^{-1 } ( ( 1,0,1,1){+}(1,1,0,1){+}(1,1,1,0 ) { - } 2(0,1,1,1))$}]}. \label{eq : zero - entropies - nont8}\end{aligned}\ ] ] we know @xmath368 , also , we have @xmath369 so , if the inequality in theorem [ thm : nont8 ] were to hold over @xmath62 , then we would have @xmath370 which is impossible . for the non - t8 network ,
the linear coding capacity is at most @xmath371 over any finite field alphabet of characteristic equal to @xmath203 .
the linear coding capacity over finite field alphabets of characteristic not @xmath203 and the coding capacity are all equal to @xmath95 .
[ cor : nont8-capacity ] let @xmath13 be a finite field alphabet . consider a @xmath80 linear solution of the non - t8 network over @xmath13 , such that the characteristic of @xmath13 is @xmath203
let @xmath16 , @xmath38 , @xmath41 , @xmath205 be message random variables in the t8 network , that are uniformly distributed over vectors in @xmath372 .
let @xmath123 , @xmath20 , @xmath373 , @xmath210 be the resulting random variables associated with the corresponding labeled edges of t8 in figure [ fig : nont8 ] . equations now hold with random variables @xmath374 are taken as random variables ( i.e. not as subspaces as in theorem [ thm : non - t8-non - char3 ] ) by lemma [ lem : conditional - dimensions ] : @xmath375}\nonumber\\ & = h(x|a , c , d ) & { & [ \mbox{from $ ( n_3,n_4)$}]}\nonumber\\ & = h(y|a , b , d ) & { & [ \mbox{from $ ( n_5,n_6)$}]}\nonumber\\ & = h(z|a , b , c ) & { & [ \mbox{from $ ( n_7,n_8)$}]}\nonumber\\ & = h(a|b , w , x ) & { & [ \mbox{from $ n _ { 9}$}]}\nonumber\\ & = h(c|a , w , y ) & { & [ \mbox{from $ n_{10}$}]}\nonumber\\ & = h(b|c , x , y ) & { & [ \mbox{from $ n_{11}$}]}\nonumber\\ & = h(d|a , w , z ) & { & [ \mbox{from $ n_{12}$}]}\nonumber\\ & = h(b|d , x , z ) & { & [ \mbox{from $ n_{13}$}]}\nonumber\\ & = h(c|d , y , z ) & { & [ \mbox{from $ n_{14}$}]}\nonumber\\ & = h(a|w , x , y , z ) & { & [ \mbox{from $ n_{15}$}]}\end{aligned}\ ] ] and since the source message @xmath72 are independent random variables , we have @xmath376 so the non - t8 inequality in theorem [ thm : nont8 ] reduces to @xmath377 now , since @xmath378 and @xmath379 , we have @xmath380
so , the linear coding capacity over characteristic 3 is at most @xmath381 the non - t8 network has a scalar linear solution over every characteristic except for 3 by using the following edge functions ( here we are using the notations @xmath374 to denote edge variables rather than vector spaces ) : @xmath382 and decoding functions : @xmath383 we know the coding capacity is at most 1 because there is a unique path from source @xmath16 to node @xmath207 ( through node @xmath384 ) . since the coding capacity is at least as large as the linear coding capacity for characteristics other than 3
, we conclude that the coding capacity is exactly equal to 1 .
r. dougherty , c. freiling , and k. zeger , `` linear rank inequalities on five or more variables , '' _ arxiv 0910.0284 _ , 2012 .
r. dougherty , c. freiling , and k. zeger , `` linear network codes and systems of polynomial equations , '' _ ieee transactions on information theory _ ,
54 , no . 5 , pp . 23032316 , 2008 .
k. makarychev , y. makarychev , a. romashchenko , and n. vereshchagin , `` a new class of non - shannon - type inequalities for entropies , '' _ communications in information and systems _ , vol . 2 , no . 2 ,
pp . 147166 , 2002 . | two characteristic - dependent linear rank inequalities are given for eight variables .
specifically , the first inequality holds for all finite fields whose characteristic is not three and does not in general hold over characteristic three .
the second inequality holds for all finite fields whose characteristic is three and does not in general hold over characteristics other than three .
applications of these inequalities to the computation of capacity upper bounds in network coding are demonstrated .
[ theorem ] corollary [ theorem ] proposition [ theorem ] remark [ theorem ] algorithm [ theorem ] conjecture [ theorem ] example [ theorem ] definition |
two liquid state theories are presently available which provide an accurate description of the liquid - vapour transition of simple systems including the critical region .
a first one is the so - called self - consistent ornstein - zernike approximation ( scoza ) devised initially by hye and stell @xcite and later adapted for numerical calculations @xcite .
it is based on the assumption that the direct correlation function behaves , to a good approximation , at long range ( attractive region of the potential ) as the potential with a density and temperature dependent prefactor which is determined by imposing consistency between the energy and compressibility routes to the thermodynamics .
the second approach put forward by parola and reatto @xcite , incorporates renormalisation group ideas by gradually taking into account fluctuations of longer and longer wavelengths .
it is based on the hierarchical reference theory ( hrt ) of fluids truncated at lowest order by means of an orpa - like ( optimised random phase approximation ) ansatz ( see ref .
@xcite for review ) . both predict non - classical critical exponents @xcite .
thermodynamic properties and phase diagrams based on the scoza and hrt approaches have been reported for a variety of potentials ( limiting ourselves to continuum systems ) including the hard core yukawa ( hcy ) potential @xcite , the square well potential @xcite , the lennard - jones ( lj ) potential @xcite , the gaussian potential @xcite , ultrasoft potentials @xcite , asakura - oosawa pair potentials @xcite , girifalco potentials @xcite , binary hard core yukawa mixtures @xcite , etc .. most of the comparisons with simulation data are so far restricted to short or medium range potentials .
the aim of this article is to extend such comparisons to long range interactions . a convenient choice to do this is the hcy potential as the range can be varied from short range appropriate for modelling colloidal suspensions or protein solutions to medium range , where it mimics the familiar lj potential @xcite , to long range . for long range interactions
we have considered a third approach , an optimised mean field theory ( omf ) @xcite , exact in the kac limit @xcite , which can be worked out analytically for the hcy potential ( see section 6 ) .
a further advantage of the hcy potential is that the repulsive part of the potential which in scoza , hrt and omf is used as a reference potential is a hard sphere potential whose properties are accurately known @xcite .
last , an approximate analytical solution ( mean spherical approximation ) is available for the hcy potential @xcite which drastically reduces the computational efforts of the solution of the scoza equations . for short inverse screening lengths @xmath0 of the yukawa potential , @xmath1 ( @xmath2 hard sphere diameter ) , a comparison between scoza , hrt and simulation results for the liquid - vapour coexistence curve is available in @xcite . in this work
we extend the domain of @xmath0 from @xmath3 ( where the hcy potential approximates the lj potential ) to @xmath4 , the infinite long range potential . with respect to simulations , in this domain
the range of the yukawa potential generally exceeds the dimensions of the simulation box ( for typical system sizes of a thousand particles ) and some care is necessary to properly treat the long range of the potential .
this has been done both by using periodic boundary conditions ( b.c . ) and performing an ewald ( ew ) sum @xcite and by using hyperspherical b.c . @xcite .
the remainder of the paper is organised as follows : in section [ mod ] we define the hcy interaction .
section [ sim ] provides a description of the two monte carlo ( mc ) methods used to treat the long range of the yukawa interaction , ewald sum and hyperspherical method .
after summarising the three considered theories , scoza , hrt and omf in section [ th ] we compare , in section [ results ] , the numerical results obtained with the different theories for the liquid - vapour coexistence curve and the thermodynamic properties along the coexistence boundary with the simulation results .
the conclusions are presented in section [ conclusion ] .
an appendix describes some properties of yukawa charge distributions relevant for elaborating the omf theory .
in our system particles interact by the hard - core yukawa potential @xmath5 in the limit @xmath6 the potential gets infinitely long range and infinitely weak ( kac limit ) .
we define an inverse reduced temperature @xmath7 , where @xmath8 is the boltzmann constant , @xmath9 the temperature and @xmath10 has dimension of an electric charge .
the reduced density is @xmath11 .
numerical simulations of fluids or plasmas involving yukawa interactions @xmath12 can be performed either in a cube of side @xmath13 with periodic b.c .
( @xmath14 ) or on the surface of a four dimensional ( 4d ) sphere of centre @xmath15 and radius @xmath16 ( equation @xmath17 ) , the hypersphere @xmath18 for short
. in both geometries the helmholtz equation @xmath19 can be solved analytically . in the case of @xmath14
the solution of the helmholtz equation can be reexpressed in terms of ewald sums with good convergence properties@xcite .
the use of an ewald potential is of course only necessary when the range of the yukawa potential exceeds the linear dimensions of the simulation box .
molecular dynamics simulations of simple models of yukawa plasmas were performed recently with this method@xcite .
for the potential model eq .
( [ potential ] ) the ewald sum for the energy is given by @xcite @xmath20 \big /2
|{\bf r}_{ij}+ l{\bf n}| \nonumber \\ & & - \frac{2\pi}{v } q^2 \alpha^{*2 } \sum_{{\bf k } = 0}^\infty \displaystyle \frac{e^{- ( { \bf k}^2 + \alpha^2 ) / 4 \eta^2}}{{\bf k}^2 + \alpha^2 } f({\bf k})f^*({\bf k } ) + u_{self } \end{aligned}\ ] ] where @xmath21 \big /2 l
|{\bf n}| \nonumber \\ & & - \frac{2\pi}{v } q^2 \alpha^{*2 } \sum_{{\bf k } = 0}^\infty \displaystyle \frac{e^{- ( { \bf k}^2 + \alpha^2 ) / 4 \eta^2}}{{\bf k}^2 + \alpha^2 } \nonumber \\ & & + q^2 \alpha^{*2 } \frac{\eta}{\sqrt \pi } { e^{- \alpha^2 / 4 \eta^2 } } - q^2 \alpha^{*2 } \ ; \frac{\alpha}{2 } { \rm erfc } \big ( \displaystyle \frac{\alpha}{2 \eta } \big ) .
\end{aligned}\ ] ] and @xmath22,\end{aligned}\ ] ] in eq .
( [ i2 ] ) @xmath23 , @xmath13 is the box length , @xmath24 the volume and @xmath25 denotes the complementary error function .
the prime in the sum over @xmath26 , with @xmath27,@xmath28 integers , restricts it to @xmath29 for @xmath30 . the parameter @xmath31 governs the convergence of the real - space and reciprocal contributions to the energy . with @xmath32 ,
adopted in our calculations , only terms with @xmath30 need to be retained in eq .
( [ i2 ] ) .
the sum in reciprocal space extends over all lattice vectors @xmath33 with @xmath34 . in the @xmath18 geometry the green s function of the helmholtz equation for the yukawa fluid
is obtained analytically @xcite : @xmath35 where @xmath36 .
the pair potential @xmath37 in @xmath18 is isotropic and depends only on the geodesic length @xmath38 .
the geodesic distance between two points @xmath39 and @xmath40 of @xmath18 is given by @xmath41 the configurational energy of a system of @xmath42 yukawa hard spheres in @xmath18 is therefore given by @xmath43 in equation ( [ confi ] ) @xmath44 denotes the hard sphere potential in @xmath18 , i.e. @xmath45 the presence of the constant @xmath46 in the r.h.s .
of equation([confi ] ) accounts for the fact that the energy is defined up to an additive constant .
the choice of @xmath46 is a recurrent problem of simulations in @xmath18 ( see e.g. the discussion in reference@xcite ) .
here we define @xmath46 by requiring that the self energies in @xmath18 and the usual euclidean space @xmath47 coincide , i.e. , @xmath48 which gives @xmath49 recent monte carlo simulations of yukawa plasmas have been performed with this method@xcite . here
we apply it to a liquid of attractive yukawa spheres .
the self - consistent ornstein - zernike approximation in its most general formulation is a liquid - state theory that introduces in the relation between the direct correlation function @xmath50 and the pair potential @xmath51 one or more state - dependent parameters that are determined by imposing consistency between the different routes to thermodynamics leading to a partial differential equation ( pde ) for the unknown parameter(s ) .
the scoza originally grew out of the semi - analytic solution of the mean spherical approximation ( msa ) for hcy potentials @xcite and it is this scoza formulation that is considered in the present work .
it is based on the ornstein - zernike equation supplemented with an msa - type closure relation @xmath52 here @xmath53 is the pair distribution function , @xmath54 is the direct correlation function of the hard - core reference system given e.g. in the waisman parameterisation @xcite and @xmath55 is the state - dependent parameter , that is not given in advance but is determined so that thermodynamic consistency is obtained between the energy and compressibility routes ( @xmath56 ) .
the consistency requirement leads to a pde of the form @xmath57 where @xmath58 is the reduced ( with respect to the ideal gas ) dimensionless isothermal compressibility given by the compressibility route and @xmath59 is the excess ( over ideal gas ) internal energy per volume provided by the energy route , i.e. @xmath60 . once @xmath61 is expressed in terms of the excess internal energy the pde eq.([consist ] ) can be transformed into a pde for @xmath59 @xmath62 for the general case of an arbitrary attractive tail of the pair interaction the determination of @xmath63 must be done fully numerically @xcite and requires an enormous amount of computational cost .
however , this procedure is not necessary for the fluid considered here : for certain pair potentials , like the hard - core yukawa potential or hard - core potentials with linear combinations of yukawa and exponential tails - so called sogami - ise fluids , the coefficient @xmath63 can be obtained semi - analytically due to the availability of the semi - analytic msa solution for these potentials .
details of the determination of @xmath63 can be found in ref .
@xcite , @xcite and @xcite . while the former formulation of scoza in ref .
@xcite is based on the laplace transformation route for solving the underlying msa for hard - core yukawa fluids , we have used in the present work the reformulation described in refs .
@xcite and @xcite which is based on the wertheim - baxter factorisation method for the solution of the msa .
while both solution methods are equivalent for the potential considered here , the wertheim - baxter factorisation method is more flexible and allows a broader applicability of the scoza @xcite .
the pde is a quasilinear diffusion equation that has been solved numerically via an implicit finite - difference algorithm in the region @xmath64\times[0,1]$ ] with suitable initial and boundary conditions ( see ref .
@xcite for more details ) .
we have chosen a density and temperature grid spacing of @xmath65 and @xmath66 .
up to now scoza has been applied to a number of discrete @xcite and continuum systems @xcite and the results showed - when compared with computer simulations - that the theory yields very accurate predictions for the thermodynamic and structural properties throughout the temperature - density plane , even near the phase coexistence and in the critical region .
indeed , scoza is one of very few liquid state theories that do not show serious problems in the critical region and even exhibit some form of scaling with non - classical , partly ising - like critical exponents @xcite . in hrt
the pair potential @xmath51 is divided into a repulsive reference potential part @xmath67 and a longer - ranged attractive part @xmath68 . in the present case
the reference system is the hard - sphere fluid , whose thermodynamics and correlations are accurately described by the carnahan - starling equation of state @xcite and the verlet - weis parameterisation of the two - body radial distribution function @xcite .
the hrt differs from the conventional liquid state approaches in the way the attractive perturbation is dealt with : in order to take into account the effect of fluctuations , @xmath68 is introduced gradually via a sequence of @xmath69 systems whose interaction potential @xmath70 results from the sum of @xmath71 and a perturbation @xmath72 .
the latter term is defined by its fourier transform @xmath73 which coincides with @xmath74 for @xmath75 , and is identically vanishing for @xmath76 .
as @xmath69 evolves from @xmath77 to @xmath78 , the interaction @xmath79 goes from the reference part @xmath67 to the full potential @xmath51 . in other words the parameter @xmath69 plays the role of an infrared cutoff whose effect consists in depressing fluctuations on length scales larger than @xmath80 : the critical fluctuations are recovered only in the limit @xmath81 .
the corresponding evolution of thermodynamics and correlations of increasing order can be determined via perturbation theory and is described by an exact hierarchy of integro - differential equations . as mentioned above close to a critical point and at large length scales , this hierarchy becomes equivalent to a formulation of the momentum - space renormalisation group @xcite .
unlike current liquid - state theories hrt preserves the correct convexity of the free energy also below the critical temperature , producing flat isotherms in the coexistence region which correspond to infinite compressibility and constant chemical potential .
the first equation of the hierarchy gives the evolution of the helmholtz free energy @xmath82 of the partially interacting system in terms of its two - body direct correlation function in momentum space @xmath83 and the full perturbation @xmath74 : @xmath84 where @xmath85 .
the quantities @xmath86 , @xmath87 are linked to @xmath82 and @xmath83 by the relations @xmath88 -\frac{1}{2 } \ , \rho\!\int\!\!\frac{d^{3}{\bf k}}{(2\pi)^{3 } } \left[\phi(k)-\phi_{q}(k)\right ] \label{amod } \\ & & { \cal c}_{q}(k ) = c_{q}(k)+\phi(k)-\phi_{q}(k ) \
, , \label{cmod}\end{aligned}\ ] ] where @xmath89 is the volume and @xmath90 the number density .
these modified free energy and direct correlation function have been introduced in order to remove the discontinuities which appear in @xmath82 and @xmath83 at @xmath78 and @xmath91 respectively as a consequence of the discontinuity of @xmath92 at @xmath91 .
indeed , they represent the free energy and direct correlation function of the fully interacting fluid as given by a treatment such that the fourier components of the interaction with wavelengths larger than @xmath80 are not really disregarded , but approximately taken into account by mean - field theory . in the @xmath81
limit the modified quantities coincide with the physical ones , once the fluctuations have been fully included .
for @xmath77 the definitions in eqs .
( [ amod ] ) and ( [ cmod ] ) reduce , instead , to the well - known mean - field expressions of the free energy and direct correlation function .
a simple approximation scheme consists in truncating the hierarchy at the first equation , supplementing it with a suitable closure relation based on some ansatz for @xmath87 . as in previous applications
, our form of @xmath87 has been inspired by liquid - state theories , @xmath93 where @xmath94 is the fourier transform of the direct correlation function of the hard - sphere fluid , and @xmath95 , @xmath96 are _ a priori _ unknown functions of the thermodynamic state and of q. the function @xmath96 is determined by the core condition , i.e. , the requirement that the radial distribution function @xmath97 is vanishing for every @xmath69 whenever the interparticle separation is less than the hard - sphere diameter @xmath2 . via eq .
( [ closure ] ) we can write the core condition in terms of an integral equation for @xmath98 . @xmath95 is adjusted so that @xmath87 satisfies the compressibility rule . for each q - system
this constraint gives the reduced compressibility of the fluid as the structure factor evaluated at zero wavevector , and can be expressed in terms of the modified quantities @xmath86 , @xmath87 as @xmath99 this equation can be regarded as a consistency condition between the compressibility route and a route in which the thermodynamics is determined from the helmholtz free energy as obtained from eq.([hrteq ] ) . here
the compressibility rule ( [ sum ] ) plays a fundamental role .
in fact , when @xmath95 in eq.([closure ] ) is determined via eq.([sum ] ) and the resulting expression for @xmath87 is used in eq.([hrteq ] ) , one obtains a partial differential equation for @xmath86 which reads @xmath100 \ , , \label{hrtclos}\ ] ] where we have set @xmath101 , @xmath102 and @xmath103 \varphi(k ) \ , .
\label{psi}\ ] ] eq.([hrtclos ] ) is integrated numerically from @xmath77 down to @xmath78 . at each integration step
, @xmath96 is determined by the core condition @xmath104 , @xmath105 .
this condition acts as an auxiliary equation for determining @xmath106 via eq.([psi ] ) . for this specific work we considered the density interval @xmath107 $ ] to numerically solve the differential equation for the free energy .
the spacing of the density grid is @xmath108 , consequently the error in the determination of the coexistence densities is of the same order of magnitude .
concerning the boundary conditions : for low density the system behaves as an ideal gas while at @xmath109 we assumed the validity of the standard orpa approximation .
two remarks are worth mentioning : first , eq .
( [ closure ] ) assumes that the fluid direct correlation function depends linearly on the perturbation @xmath110 .
this ansatz is especially appropriate when the perturbation range is much longer than that of the reference part of the interaction , i.e. longer than the particle size .
the second remark relates to the implementation of the core condition : the inverse fourier transform of the function @xmath96 has been expanded in a series of legendre polynomials for @xmath111 and the series has been truncated after a finite number of terms ( typically five ) .
the evolution equations for the expansion coefficients were then determined by differentiating with respect to @xmath69 the integral equation for @xmath98 , i.e. the core condition , and subsequently projecting it on the polynomials used in the expansion . however ,
the resulting equations are coupled to the evolution equation [ hrteq ] for the free energy , and this makes them difficult to handle . therefore , as described in detail in [ 21 , 22 ] , in the derivative of @xmath98 with respect to q the long - wavelength contributions containing the isothermal compressibility of the q - system were disregarded .
physically this amounts to a decoupling of the short- and the long - range evolutions of the correlations .
this approximation , as well as the previous one , are fully justified for long - range perturbations , where the short- and the long - range parts of the correlations are mainly affected by the reference and the perturbation terms respectively , but become more problematic for short - range interactions where the interplay between excluded - volume and cohesion effects is much stronger . however , this effect does not concern the present study
. we consider , quite generally , a fluid of identical hard spheres of diameter @xmath2 with additional isotropic pair interaction @xmath112 .
since @xmath51 is an arbitrary function of @xmath113 for @xmath114 , one can assume that @xmath51 has been regularised in the core in such a way that its fourier transform @xmath115 is a well behaved function of @xmath116 and that @xmath117 is a finite quantity . at equilibrium at inverse temperature @xmath118 and chemical potential @xmath119 ( grand canonical ( gc ) ensemble )
the pressure @xmath120 of the fluid is a convex function of @xmath121 ( at fixed @xmath118 ) @xcite even before the thermodynamic limit has been taken @xcite . as a consequence @xmath122 is continuous ,
its derivatives with respect to @xmath123 exist and it admits a legendre transform @xmath124 with respect to @xmath123 defined as @xmath125 the gc free energy @xmath124 is then a convex function of the density and its legendre transform with respect to @xmath90 coincides with @xmath122 .
it was shown in ref.@xcite that in the case of an attractive interaction ( i.e. @xmath126 for all @xmath10 ) we have the inequality @xmath127 here the subscript mean field ( mf ) emphasises that @xmath128 is the van der waals free energy@xcite .
note that @xmath129 $ ] is _ a priori _ non convex , notably in the two - phase region .
> from here on we specialise to the yukawa interaction @xmath130 with @xmath131 . since @xmath132 , equation ( [ boundf ] ) does not appear to be very useful .
however , it follows from appendix a that we can replace @xmath133 by the regular function @xmath134 where @xmath135 is the interaction energy of two spherical distributions @xmath136 of yukawa charges of effective charge @xmath137 .
the interaction is regularised in the core but remains unchanged for @xmath138 . since @xmath139 , eq .
( [ boundf ] ) is still valid and will give a non trivial upper bound for the free energy .
similar to what has been done in reference@xcite for the coulomb interaction we seek a best upper bound for @xmath140 by minimising the quantity @xmath141 to this end we consider variations of @xmath142 with respect to the charge distribution @xmath143 defined by eq .
( [ tau ] ) and work out the stationary condition ( @xmath144 ) @xmath145 first , it follows from equation ( [ q ] ) that @xmath146 second , as a consequence of the convolution relations @xmath147 and @xmath148 one has @xmath149 and thus @xmath150 finally , since @xmath151 , we have @xmath152 collecting results ( [ eq1 ] ) , ( [ eq2 ] ) and ( [ eq3 ] ) one gets @xmath153 note that we can impose @xmath154 since , if @xmath143 is solution of the stationary condition ( [ statio ] ) , then @xmath155 , where @xmath156 is an arbitrary constant , is also a solution ( @xmath137 is then multiplied by @xmath157 but @xmath158 is left unchanged )
. therefore eq.([statio ] ) can be recast in the form @xmath159 of course , eq .
( [ resuv ] ) is valid only for @xmath160 .
the first contribution in the r.h.s of eq .
( [ resuv ] ) is clearly identified with the potential created by a uniform density of yukawa charges of density @xmath90 while the second contribution stems from a charge discontinuity at @xmath161 .
the solution is therefore of the form @xmath162 , where @xmath163 is a constant determined by the condition @xmath154 , yielding @xmath164 ( @xmath165 packing fraction ) .
furthermore , from eq .
( [ q ] ) it follows that @xmath166 so that the potential @xmath167 is seen to be the superposition of two potentials created by spherical surface and volume distributions , the expressions of which are given in appendix a by eqs .
( [ surf ] ) and ( [ vol ] ) , respectively .
one finds @xmath168 one can check that @xmath167 indeed verifies eq.([resuv ] ) . finally , the optimised upper bound for the free energy
can then be shown to be equal to @xmath169 an optimised mean field ( omf ) theory for the fluid can now be devised by considering the landau function@xcite @xmath170 at given @xmath118 and @xmath123 , the density @xmath90 is the one which minimises the landau function @xmath171 . in the limit @xmath172 , @xmath173 and one
thus revovers the free energy of the kac model for @xmath174 .
a comparison between the three theoretical approaches and simulation data for the liquid - vapour coexistence curve is made in figs .
[ figure1 ] -[figure4 ] for @xmath175 and 0.1 .
values for the densities , internal energies , chemical potentials and helmholtz free energies along the coexistence curve are presented in tables [ table1]- [ table4 ] .
these data may be valuable for reference to future approaches .
the simulation results have been obtained with the gibbs ensemble monte carlo ( gemc ) method @xcite using mostly a total of 1000 particles ( ew b.c . ) and 2000 particles ( @xmath18 b.c . ) .
the number of cycles generated after equilibration varied from 7 - 40 @xmath176 according to temperature , each cycle comprising with equal probabilities , displacement of the @xmath42 particles , volume change , or exchange of 500 ( ew b.c . ) or 1000 particles ( @xmath18 b.c . ) .
error bars ( two standard deviations ) were calculated by averaging densities , energies or histograms over blocks of 1 million cycles .
expectedly , uncertainties are largest near the critical temperature where fluctuations of density of the coexisting phases are important and hence convergence slow .
finite size effects additionally affect the critical behaviour , but also influenced to some extent the determination of the coexistence density on the liquid side at least for the smallest value @xmath177 .
thus , for the @xmath18 b.c . , an increase of the number of particles from 1000 to 2000 typically lowered @xmath178 by about @xmath179 .
the opposite trend was observed with the ew b.c . where @xmath178 increased by @xmath180 by an increase of @xmath42 from 1000 to 1872 ( cf .
table [ table4 ] ) .
it is likely that , for @xmath177 , small finite size effects are still present in both sets of results .
the agreement between hrt , scoza and mc simulations turns out to be extremely good over the range of @xmath0 values considered .
small differences occur only at the highest and lowest values of @xmath0 . at @xmath3
the critical temperature of hrt is slightly below that predicted by scoza .
previous work @xcite showed , however , that by further increasing the value of @xmath0 the hrt critical temperature eventually gets higher than that of scoza . at @xmath177 the coexistence curves , including the critical regions , of scoza and omf are very close ( cf .
[ figure4 ] ) ; as the omf theory is expected to be nearly exact at this small value of @xmath0 it is very likely that scoza becomes exact also in the kac limit , though no formal proof seems to be available .
this is further confirmed by the close agreement between critical parameters of scoza and omf at @xmath181 ( see table [ table5 ] ) .
the coexistence curve of hrt is found to be slightly narrower than that of scoza though in the near critical region the agreement between the two theories is quite good . as pointed out in ref .
@xcite the hrt flow can not be solved conveniently in the kac limit .
inspection of fig .
[ figure3 ] shows that the omf theory rapidly deteriorates with increase of @xmath0 .
figure [ figure5 ] compares hrt , scoza , omf and mc simulation results for the internal energy along the coexistence curve at @xmath182 .
the agreement is similar to that for the densities . the variation with @xmath0 of the critical temperature and density for scoza and hrt , including data of ref .
@xcite for @xmath183 , is given in table [ table5 ] .
no attempt has been made to determine critical data in the simulations by fitting data away from the critical point by some power law expression .
because of the finite size effects in the critical region uncertainties of the results would be large and of no use to validate the theoretical predictions . in the gemc method
the order parameter @xmath184 is believed to have mf behaviour ( critical exponent @xmath185 ) in the critical region once the correlation length exceeds the linear dimension @xmath13 of the simulation box .
in fact we find that for @xmath186 , @xmath187 varies linearly over the whole temperature region considered yielding an effective exponent @xmath188 ) .
such an analysis was not conclusive for @xmath3 .
we have compared the predictions for the liquid - vapour coexistence curve of a long - range yukawa fluid obtained from advanced theoretical approaches with gibbs ensemble monte carlo simulations .
concerning the simulations care has to be taken when the range of the potential exceeds the length of the simulation box .
this was done by performing an ewald sum in the case of a cubic simulation box and by using hyperspherical boundary conditions .
the theoretical approaches that we applied comprised the self - consistent ornstein - zernike approximation ( scoza ) , the hierarchical reference theory ( hrt ) , as well as an optimised mean field theory ( omf ) . while the omf yields the exact result in the limit of infinite range of the potential , it deteriorates with decreasing interaction range . on the other hand ,
hrt and scoza turn out to be in perfect agreement with simulation results over the whole interaction range considered .
this study complements a recent comparison @xcite for the case of intermediate and short range yukawa fluid .
this work was supported by a grant through the programme dactions intgres amadeus under project nos . 06648pb and fr 09/2007
, by the hochschuljubilumsstiftung der stadt wien under project number 1757/2007 and by the marie curie european network mrtn - ct-2003 - 504712 .
federica lo verso thanks davide pini , alberto parola and luciano reatto for helpful discussions .
99 j. s. hye , g. stell .
_ j. chem .
phys . , _ * 67 * , 439 ( 1977 ) .
j. s. hye , g. stell .
phys . , _ * 52 * , 1071 ( 1984 ) .
r. dickman , g. stell .
lett . , _ * 77 * , 996 ( 1996 ) .
d. pini , g. stell , r. dickman .
e , _ * 57 * , 2862 ( 1998 ) .
d. pini , g. stell , j.s .
j. thermophys . ,
_ * 19 * , 1029 ( 1998 ) .
a. parola , l. reatto , _ phys .
rev . a , _ * 31 * , 2417 ( 1985 ) .
a. parola , l. reatto .
phys . , _ * 44 * , 211 ( 1995 ) .
j. s. hye , d. pini , g. stell _ physica a , _ * 279 * , 213 ( 2000 ) .
d. pini , g. stell , n. b. wilding .
phys . , _ * 95 * , 483 ( 1998 ) . c. caccamo , g. pellicane , d. costa , d. pini , g. stell _ phys
e , _ * 60 * , 5533 ( 1999 ) .
d. pini , ge jialin , a. parola , l. reatto .
lett . , _ * 327 * , 209 ( 2000 ) .
d. pini , a. parola , l. reatto .
_ j. phys . :
matter , _ * 18 * , s2305 ( 2006 )
. a. reiner , g. kahl .
_ j. chem .
_ * 117 * , 4925 ( 2002 ) .
e. schll - paschinger , a. l. benavides , r. castaeda - priego .
_ j. chem .
_ * 123 * , 234513 ( 2005 ) .
a. meroni , a. parola , l. reatto .
rev . a , _ * 42 * , 6104 ( 1990 ) .
m. tau , a. parola , d. pini , l. reatto .
e , _ * 52 * , 2644 ( 1995 ) .
e. schll - paschinger , a. reiner .
_ j. chem .
_ * 125 * , 164503 ( 2006 ) .
j. s. hye , a. reiner .
_ j. chem .
, _ * 125 * , 104503 ( 2006 ) .
b. m. mladek , g. kahl , m. neumann .
_ j. chem .
_ * 124 * , 064503 ( 2006 ) .
f. lo verso , m. tau , l. reatto .
_ j. phys .
: condens.matter_ * 15 * , 1505 ( 2003 ) .
f. lo verso , d. pini , l. reatto . _ j. phys . : condens . matter _
* 17 * , 771 ( 2005 ) .
f. lo verso , r. l. c. vink , d. pini , l. reatto .
e _ * 73 * , 61407 ( 2006 ) .
e. schll - paschinger , g. kahl .
_ europhys .
letters , _ * 63 * , 538 ( 2003 ) .
d. costa , g. pellicane , c. caccamo , e. schll - paschinger , g. kahl _ phys .
* 68 * , 021104 ( 2003 ) .
e. schll - paschinger , g. kahl .
_ j. chem .
phys . , _ * 118 * , 7414 ( 2003 )
. e. schll - paschinger , d. levesque , j .- j .
weis , g. kahl .
_ j. chem .
phys . , _ * 122 * , 024507 ( 2005 ) .
e. schll - paschinger , g. kahl .
_ j. chem .
_ * 123 * , 134508 ( 2005 ) .
d. pini , m. tau , a. parola , l. reatto .
e , _ * 67 * , 046116 ( 2003 ) .
_ * 101 * , 1617 ( 2003 ) .
m. kac , g.e .
uhlenbeck , p.c .
hemmer , _ j. math , phys . ,
_ * 4 * , 216 ( 1963 ) .
n. f. carnahan , k.e .
starling _ _ j .
phys.,__*51 * , 635 ( 1969 ) .
l. verlet , j. j. weis .
rev . a,__*5 * , 939 ( 1972 ) .
d. henderson e. w. grundke .
_ j. chem .
, _ * 63 * , 601 ( 1975 ) . e. waisman . _ mol .
, _ * 25 * , 45 ( 1973 ) .
l. blum , j.s .
_ j. stat .
_ * 19 * , 317 ( 1978 ) .
g. salin , j .-
m . caillol .
_ j. chem .
phys . , _ * 113 * , 10459 ( 2000 ) .
caillol , d. gilles .
_ j. stat .
, _ * 100 * , 905 ( 2000 ) .
caillol , d. gilles .
_ j. stat .
, _ * 100 * , 933 ( 2000 ) .
g. salin , j .-
m . caillol .
, _ * 88 * , 65002 - 1 - 4 ( 2002 ) . g. salin , j .-
m . caillol .
_ phys . of plasmas ,
_ * 10 * , 1220 ( 2003 ) .
caillol , d. gilles .
_ j. phys .
a : math . gen .
, _ * 36 * , 6343 ( 2003 ) .
g. kahl , e. schll - paschinger , g. stell .
_ j. phys . condens .
matter , _ * 14 * , 9153 ( 2002 ) .
e. schll - paschinger .
_ j. chem .
_ * 120 * , 11698(2004 ) .
hansen , i.r .
_ theory of simple liquids _ academic press , london ( 1986 ) .
see , for instance , k. g. wilson , j. b. kogut _ phys .
c _ 12 , 75 ( 1974 ) .
see , for instance , m. e. fisher _ critical phenomena _ , lecture notes in physics , vol .
* 186 * , f. j. w. hahne ( ed . ) , springer , berlin , ( 1982 ) .
f. nicoll , t. s. chang .
lett . , _ * 62a * , 287 ( 1977 ) . n. goldenfeld . _
lectures on phase transitions and the renormalization group _ addison - wesley ( 1992 ) .
_ j. phys .
a : math . gen . ,
_ * 35 * , 4189 ( 2002 ) .
_ j. stat .
_ * 115 * , 1461 ( 2004 ) .
a.z . panagiotopoulos . _
phys . , _ * 61 * , 813 ( 1987 ) .
d. frenkel , b. smit .
_ understanding molecular simulation _ academic press , london ( 1996 ) .
_ * 104 * , 1931 ( 2006 ) .
y. rosenfeld .
1993 , _ j. chem .
phys . , _ * 98 * , 8126 ( 1993 )
. y. rosenfeld .
e , _ * 47 * , 2676 ( 1993 ) .
y. rosenfeld .
phys . , _ * 88 * , 1357 ( 1996 ) .
y. rosenfeld , g. chabrier .
_ j. stat .
_ * 98 * , 283 ( 1997 ) .
coexistence densities , internal energy @xmath189 , excess chemical potential @xmath119 and helmholtz free energy @xmath190 at coexistence for @xmath191 .
@xmath192 is the carnahan - starling hs free energy @xcite .
l l l l l l l l @xmath193 & & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 + & scoza&0.174 & 0.464 & -1.055 & -2.650 & -2.737 & & + 0.635 & hrt & 0.181 & 0.458 & -1.106 & -2.607 & & 0.963 & 2.521 + & mc - ew&0.175(10)&0.430(15 ) & -1.06(4 ) & -2.43(3 ) & -2.74(1 ) & & + + & scoza&0.161 & 0.485 & -0.982 & -2.770 & -2.765 & & + 0.63 & hrt & 0.166 & 0.477 & -1.030 & -2.735 & & 0.890 & 2.657 + & mc - ew&0.171(8)&0.470(15 ) & -1.05(2 ) & -2.67(2 ) & -2.76(1 ) & & + + & scoza&0.138 & 0.515 & -0.862 & -2.996 & -2.825 & & + 0.62 & hrt & 0.142 & 0.509 & -0.901 & -2.967 & & 0.774 & 2.900 + & mc - ew&0.149(5)&0.501(4 ) & -0.94(2 ) & -2.92(2 ) & -2.83(1 ) & & + + & scoza&0.105 & 0.565 & -0.686 & -3.407 & -2.953 & & + 0.60 & hrt & 0.108 & 0.560 & -0.724 & -3.386 & & 0.609 & 3.333 + & mc - ew&0.109(2)&0.553(4 ) & -0.73(1 ) & -3.34(2 ) & -2.96(1 ) & & + + & scoza&0.082 & 0.606 & -0.556 & -3.803 & -3.094 & & + 0.58 & hrt & 0.083 & 0.602 & -0.578 & -3.787 & & 0.486 & 3.741 + & mc - ew&0.083(2)&0.594(4 ) & -0.58(1 ) & -3.74(2 ) & -3.10(1 ) & & + + & scoza&0.063 & 0.643 & -0.452 & -4.204 & -3.251 & & + 0.56 & hrt & 0.065 & 0.640 & -0.490 & -4.192 & & 0.396 & 4.155 + & mc - ew&0.063(2)&0.631(4 ) & -0.46(1 ) & -4.13(3 ) & -3.25(3 ) & & + + & scoza&0.049 & 0.676 & -0.367 & -4.614 & -3.424 & & + 0.54 & hrt & 0.050 & 0.674 & -0.390 & -4.606 & & 0.317 & 4.573 + & mc - ew&0.050(2)&0.670(4 ) & -0.39(1 ) & -4.57(2 ) & -3.47(8 ) & & + + & scoza&0.0284 & 0.738 & -0.235 & -5.503 & -3.834 & & + 0.50 & hrt & 0.029 & 0.736 & -0.258 & -5.500 & & 0.201 & 5.469 + & mc - ew&0.029(2)&0.733(4 ) & -0.25(1 ) & -5.47(3 ) & -3.76(10 ) & & + + @xmath201@xmath202 + @xmath203@xmath204 [ table1 ] table 2 .
same as table 1 but for @xmath205 + @c l
l l l l l l l @xmath193 & & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 + & scoza&0.174 & 0.396 & -0.987 & -2.209 & -2.773 & & + 0.905 & hrt & 0.176 & 0.393 & -1.000 & -2.195 & & 0.955 & 2.169 + & mc - ew&0.183(2)&0.395(10 ) & -1.025(50 ) & -2.19(5 ) & -2.77(1 ) & & + + & scoza&0.135 & 0.450 & -0.788 & -2.583 & -2.864 & & + 0.88 & hrt & 0.136 & 0.448 & -0.767 & -2.477 & & 0.758 & 2.557 + & mc - ew&0.135(2)&0.439(5 ) & -0.78(2 ) & -2.52(2 ) & -2.87(1 ) & & + + & scoza&0.112 & 0.484 & -0.676 & -2.852 & -2.944 & & + 0.86 & hrt & 0.114 & 0.482 & -0.687 & -2.842 & & 0.650 & 2.825 + & mc - ew&0.112(2)&0.474(4 ) & -0.67(1 ) & -2.79(2 ) & -2.95(1 ) & & + + & scoza&0.095 & 0.514 & -0.586 & -3.107 & -3.028 & & + 0.84 & hrt & 0.096 & 0.512 & -0.593 & -3.098 & & 0.561 & 3.083 + & mc - ew&0.096(2)&0.508(5 ) & -0.59(2 ) & -3.07(3 ) & -3.04(2 ) & & + + & scoza&0.0806 & 0.541 & -0.511 & -3.360 & -3.119 & & + 0.82 & hrt & 0.081 & 0.540 & -0.513 & -3.353 & & 0.485 & 3.341 + & mc - ew&0.085(3)&0.543(6 ) & -0.54(2 ) & -3.37(3 ) & -3.11(2 ) & & + + & scoza&0.0685 & 0.567 & -0.446 & -3.615 & -3.216 & & + 0.80 & hrt & 0.069 & 0.566 & -0.451 & -3.610 & & 0.424 & 3.600 + & mc - ew&0.070(2)&0.564(4 ) & -0.447(15 ) & -3.60(3 ) & -3.21(1 ) & & + + & scoza&0.058 & 0.591 & -0.389 & -3.873 & -3.321 & & + 0.78 & hrt & 0.059 & 0.590 & -0.401 & -3.868 & & 0.372 & 3.859 + & mc - ew&0.062(2)&0.594(6 ) & -0.420(15 ) & -3.89(3 ) & -3.30(2 ) & & + + & scoza&0.049 & 0.614 & -0.339 & -4.140 & -3.434 & & + 0.76 & hrt & 0.050 & 0.613 & -0.353 & -4.133 & & 0.324 & 4.124 + & mc - ew&0.050(2)&0.612(3 ) & -0.344(5 ) & -4.120(15 ) & -3.43(1 ) & & + + & scoza&0.041 & 0.636 & -0.294 & -4.413 & -3.557 & & + 0.74 & hrt & 0.042 & 0.635 & -0.304 & -4.407 & & 0.280 & 4.398 + & mc - ew&0.042(2)&0.634(4 ) & -0.30(1 ) & -4.440(25 ) & -3.55(1 ) & & + + @xmath201@xmath202 + @xmath203@xmath204 [ table2 ] table 3 .
same as table 1 but for @xmath206 + @c
l l l l l l l l @xmath193 & & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 + & scoza&0.164 & 0.367 & -0.923 & -2.063 & -2.824 & & + 1.045 & hrt & 0.163 & 0.371 & -0.916 & -2.086 & & 0.908 & 2.082 + & mc - ew&0.169(8)&0.355(8 ) & -0.93(3 ) & -1.99(2 ) & -2.82(1 ) & & + & mc - s3&0.182(5)&0.384(3 ) & -1.02(3 ) & -2.15(3 ) & & & + + & scoza&0.157 & 0.377 & -0.885 & -2.131 & -2.839 & & + 1.04 & hrt & 0.156 & 0.381 & -0.881 & -2.153 & & 0.873 & 2.149 + & mc - ew&0.160(8)&0.370(8 ) & -0.89(4 ) & -2.05(4 ) & -2.84(1 ) & & + & mc - s3&0.165(6)&0.384(6 ) & -0.93(3 ) & -2.172(22 ) & & & + + & scoza&0.133 & 0.412 & -0.764 & -2.377 & -2.899 & & + 1.02 & hrt & 0.132 & 0.415 & -0.760 & -2.394 & & 0.752 & 2.390 + & mc - ew&0.130(6)&0.400(6 ) & -0.75(2 ) & -2.31(4 ) & -2.90(1 ) & & + & mc - s3&0.133(2)&0.415(3 ) & -0.77(1 ) & -2.39(2 ) & & & + + & scoza&0.114 & 0.442 & -0.670 & -2.602 & -2.962 & & + 1.00 & hrt & 0.114 & 0.445 & -0.670 & -2.620 & & 0.662 & 2.617 + & mc - ew&0.118(4)&0.439(4 ) & -0.689(15 ) & -2.58(2 ) & -2.96(1 ) & & + & mc - s3&0.111(1)&0.441(2 ) & -0.655(50 ) & -2.60(1 ) & & & + + & scoza&0.099 & 0.469 & -0.594 & -2.817 & -3.030 & & + 0.98 & hrt & 0.099 & 0.470 & -0.594 & -2.826 & & 0.587 & 2.824 + & mc - ew&0.103(4)&0.467(4 ) & -0.62(1 ) & -2.74(2 ) & -3.02(1 ) & & + & mc - s3&0.095(1)&0.466(2 ) & -0.573(54 ) & -2.800(15 ) & & & + + & scoza&0.086 & 0.493 & -0.528 & -3.028 & -3.101 & & + 0.96 & hrt & 0.086 & 0.494 & -0.527 & -3.035 & & 0.520 & 3.032 + & mc - ew&0.085(4)&0.485(4 ) & -0.52(1 ) & -2.99(1 ) & -3.11(1 ) & & + & mc - s3&0.0846(11)&0.494(3 ) & -0.521(7 ) & -3.037(15 ) & & & + + & scoza&0.075 & 0.516 & -0.471 & -3.239 & -3.178 & & + 0.94 & hrt & 0.076 & 0.516 & -0.476 & -3.240 & & 0.469 & 3.238 + & mc - ew&0.076(2)&0.512(2 ) & -0.468(5 ) & -3.21(1 ) & -3.17(2 ) & & + & mc - s3&0.072(1)&0.509(2 ) & -0.442(4 ) & -3.20(1 ) & & & + + & scoza&0.0656 & 0.538 & -0.420 & -3.451 & -3.259 & & + 0.92 & hrt & 0.066 & 0.537 & -0.422 & -3.447 & & 0.417 & 3.446 + & mc - ew&0.068(2)&0.539(3 ) & -0.43(1 ) & -3.46(2 ) & -3.27(3 ) & & + & mc - s3&0.0678(13)&0.547(3 ) & -0.437(8 ) & -3.52(2 ) & & & + + & scoza&0.0572 & 0.559 & -0.374 & -3.667 & -3.346 & & + 0.90 & hrt & 0.058 & 0.557 & -0.380 & -3.658 & & 0.374 & 3.656 + & mc - ew&0.058(2)&0.556(3 ) & -0.375(10 ) & -3.64(2 ) & -3.34(1 ) & & + & mc - s3&0.058(1)&0.566(3 ) & -0.382(7 ) & -3.72(2 ) & & & + + @xmath201@xmath202 + @xmath203@xmath204 [ table3 ] table 4 .
same as table 1 but for @xmath207 + @c
l l l l l l l l @xmath193 & & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 + & scoza&0.158 & 0.357 & -0.900 & -2.030 & -2.863 & & + 1.10 & hrt & 0.161 & 0.353 & -0.916 & -2.010 & & -0.916 & -2.010 + & mc - ew&0.17(1)&0.340(15 ) & -0.96(2 ) & -1.96(5 ) & -2.86(2 ) & & + & mc - s3&0.167(5)&0.361(5 ) & -0.93(3 ) & -2.05(2 ) & & & + + & scoza&0.145 & 0.375 & -0.834 & -2.155 & -2.891 & & + & hrt & 0.148 & 0.371 & -0.850 & -2.132 & & -0.850 & -2.131 + 1.09 & mc - ew&0.145(10)&0.360(15 ) & -0.82(1 ) & -2.05(1 ) & -2.89(1 ) & & + & mc - ew1&0.145(4)&0.362(3 ) & -0.82(2 ) & -2.07(2 ) & -2.895(5 ) & & + & mc - s3&0.160(6)&0.392(3 ) & -0.92(2 ) & -2.25(2 ) & & & + + & scoza&0.134 & 0.391 & -0.778 & -2.270 & -2.918 & & + 1.08 & hrt & 0.138 & 0.387 & -0.800 & -2.244 & & -0.800 & -2.244 + & mc - ew&0.136(5)&0.380(5 ) & -0.79(2 ) & -2.21(2 ) & -2.92(2 ) & & + & mc - s3&0.1384(24)&0.395(3 ) & -0.8018(15 ) & -2.29(2 ) & & & + + & scoza&0.124 & 0.406 & -0.730 & -2.379 & -2.947 & & + & hrt & 0.128 & 0.401 & -0.749 & -2.347 & & -0.749 & -2.347 + 1.07 & mc - ew&0.126(2)&0.393(3 ) & -0.73(1 ) & -2.32(2 ) & -2.95(1 ) & & + & mc - ew1&0.127(2)&0.399(3 ) & -0.74(1 ) & -2.34(2 ) & -2.95(1 ) & & + & mc - s3&0.1260(15)&0.4103(21 ) & -0.737(8 ) & -2.402(13 ) & & & + + & scoza&0.116 & 0.420 & -0.686 & -2.484 & -2.977 & & + 1.06 & hrt & 0.120 & 0.415 & -0.708 & -2.452 & & -0.708 & -2.452 + & mc - ew&0.121(4)&0.416(4 ) & -0.716(15 ) & -2.46(2 ) & -2.97(1 ) & & + & mc - s3&0.118(2)&0.423(3 ) & -0.697(11 ) & -2.544(15 ) & & & + + & scoza&0.1082 & 0.434 & -0.647 & -2.587 & -3.007 & & + & hrt & 0.112 & 0.427 & -0.667 & -2.547 & & -0.667 & -2.547 + 1.05 & mc - ew&0.109(2)&0.425(3 ) & -0.65(1 ) & -2.54(2 ) & -3.000(15 ) & & + & mc - ew1&0.113(2)&0.433(3 ) & -0.68(1 ) & -2.59(2 ) & -3.00(1 ) & & + & mc - s3&0.111(2)&0.442(2 ) & -0.660(7 ) & -2.63(2 ) & & & + + & scoza&0.101 & 0.446 & -0.610 & -2.689 & -3.038 & & + 1.04 & hrt & 0.107 & 0.439 & -0.638 & -2.644 & & -0.638 & -2.644 + & mc - ew&0.102(1)&0.439(2 ) & -0.641(5 ) & -2.64(1 ) & -3.04(1 ) & & + & mc - s3&0.1047(13)&0.456(3 ) & -0.630(9 ) & -2.75(2 ) & & & + + & scoza&0.089 & 0.470 & -0.546 & -2.887 & -3.103 & & + 1.02 & hrt & 0.093 & 0.462 & -0.571 & -2.837 & & -0.570 & -2.837 + & mc - ew&0.091(1)&0.466(2 ) & -0.555(5 ) & -2.86(1 ) & -3.10(1 ) & & + & mc - s3&0.091(1)&0.479(2 ) & -0.558(6 ) & -2.94(1 ) & & & + + & scoza&0.078 & 0.493 & -0.489 & -3.085 & -3.173 & & + & hrt & 0.083 & 0.483 & -0.519 & -3.026 & & -0.519 & -3.026 + 1.00 & mc - ew&0.077(3)&0.483(5 ) & -0.48(1 ) & -3.028(25 ) & -3.17(1 ) & & + & mc - ew1&0.078(2)&0.487(5 ) & -0.484(14 ) & -3.05(3 ) & -3.185(15 ) & & + & mc - s3&0.079(1)&0.499(2 ) & -0.492(6 ) & -3.126(13 ) & & & + + & scoza&0.0572 & 0.559 & -0.374 & -3.667 & -3.346 & & + 0.98 & hrt & 0.074 & 0.503 & -0.472 & -3.215 & & -0.472 & -3.215 + & mc - ew&0.057(1)&0.539(3 ) & -0.373(6 ) & -3.55(2 ) & -3.35(3 ) & & + + @xmath201@xmath202 + @xmath203@xmath204 [ table4 ] table 5 .
critical parameters .
results based on the carnahan - starling hard sphere free energy @xcite for the reference system .
+ @c l l l @xmath208 & & @xmath209 & @xmath210 + 0.0 & omf&1.13194 & 0.24913 + + @xmath211 & scoza&1.1319 & 0.2491 + + 0.1 & scoza&1.1294 & 0.2495 + & hrt & 1.129 & 0.250 + & omf & 1.13108 & 0.24915 + + 0.5 & scoza&1.0718 & 0.2586 + & hrt & 1.073 & 0.260 + + 1.0 & scoza&0.9264 & 0.2793 + & hrt & 0.925 & 0.279 + + 1.8 & scoza&0.6527 & 0.3145 + & hrt & 0.650 & 0.314 + + @xmath212 & scoza&0.173 & 0.3895 + & hrt & 0.175 & 0.394 + + @xmath213 & scoza&0.0187 & 0.4575 + & hrt & 0.019 & 0.424 + + @xmath201 from ref . .
due to a different definition , the temperatures of ref .
must be multiplied by the factor @xmath214 .
[ table5 ] [ figure1 ] [ figure2 ] [ figure3 ] [ figure4 ] [ figure5 ]
the electrostatics of yukawa charge distributions is similar to , but not identical with the electrostatics of usual coulomb charge distributions . in this appendix
we extend previous results obtained for special forms of spherical distributions of yukawa charges @xcite to a general spherical distribution @xmath215 which satisfies : @xmath216 we denote by @xmath133 the yukawa potential created by a unit point charge .
we have @xmath131 and for its fourier transform @xmath217 . the yukawa potential @xmath218 created by this distribution at point @xmath219 is given by the convolution of @xmath220 and @xmath221 @xmath222 where @xmath223 denotes the infinitesimal spherical angle about vector @xmath224 .
the integral @xmath225 reads @xcite @xmath226 the potential @xmath227 for @xmath228 is not particularly useful but outside the sphere ( i.e. for @xmath229 one infers from equations ( [ v ] ) and ( [ i ] ) the remarkable result as in the case of the coulomb potential , the potential outside the spherical distribution of charge @xmath133 is still a yukawa potential with the same screening parameter @xmath0 but with a different charge @xmath137 . since @xmath231 the effective charge @xmath137 is larger than the bare charge @xmath232 of the distribution .
of course in the limit @xmath6 one has @xmath233 and gauss law is recovered .
we now prove the theorem @xmath234 for @xmath235 it is obvious ; for @xmath236 we note that from the expression ( [ q ] ) of @xmath137 and from equation([v ] ) it follows that @xmath237 where @xmath238 .
since @xmath239 is a decreasing function of @xmath113 and @xmath240 we have @xmath241 from which inequality ( [ inequav ] ) follows ( note that @xmath143 must be positive ) . for a sufficiently regular distribution @xmath143 , the potential @xmath227 is a bounded smooth function for @xmath242 , in particular @xmath243 is finite in general ( see examples at the end of the appendix ) .
also in this case gauss theorem generalises easily .
indeed , for the case where the two distributions do not overlap , i.e. @xmath250 , @xmath251 can be replaced by @xmath252 in equation ( [ w ] ) and thus @xmath253 note that the fourier transform @xmath254 is positive definite contrary to @xmath255 .
in addition , we have the theorem @xmath256 the proof is trivial : @xmath257 the function @xmath135 is in general bounded in the core and in particular @xmath258 is finite .
this charge smearing process provides an easy way to regularise the yukawa potential @xmath133 in the core which preserves the positivity of the interaction .
finally , we give explicit expressions for @xmath137 , @xmath167 for @xmath259 , and the value @xmath258 in simple cases relevant for sect .
thus for a surface distribution @xmath260 one finds @xmath261 for a volume distribution @xmath262 ( @xmath263 step function ) one has @xmath264 | two liquid state theories , the self - consistent ornstein - zernike equation ( scoza ) and the hierarchical reference theory ( hrt ) are shown , by comparison with monte carlo simulations , to perform extremely well in predicting the liquid - vapour coexistence of the hard core yukawa ( hcy ) fluid when the interaction is long range .
the long range of the potential is treated in the simulations using both an ewald sum and hyperspherical boundary conditions .
in addition , we present an analytical optimised mean field theory which is exact in the limit of an infinitely long range interaction .
the work extends a previous one by caccamo _
et al _ [ _ phys . rev .
e , _ * 60 * , 5533 ( 1999 ) ] for short range interactions .
* keywords * : yukawa potential , critical phenomena , monte carlo simulations , scoza , hrt . |
the primary focus in kaon physics today is the two golden modes : and .
these modes are interesting as there is essentially no theoretical ambiguity in extracting fundamental ckm parameters from measurements of the branching ratios@xcite .
the intrinsic theoretical uncertainty in b ( ) is @xmath07% and is even smaller in b ( ) , only @xmath02% ; in both cases the hadronic matrix element can be extracted from the ( @xmath1 ) decay rate .
the unitarity of the ckm matrix can be expressed as @xmath2with the three vectors @xmath3 converging to form a very elongated triangle in the complex plane . the length of first vector @xmath4 is precisely determined from @xmath1 decay .
the length of the third side @xmath5 is measured by and the height of the triangle , @xmath6 , can be measured by . branching ratio measurements of the two modes , along with the well known @xmath1 , will completely determine the unitarity triangle .
comparison of ckm parameters as measured from the golden , and modes , provide the best opportunity to over - constrain the unitary triangle and to search for new physics .
in particular , comparisons of * @xmath7 from and from the ratio of the mixing frequencies of @xmath8 and @xmath9 mesons @xcite , * @xmath10 from b()/b ( ) and from the time dependent asymmetry in the decay @xcite offer outstanding opportunities to explore the standard model ( sm ) picture of _ cp_violation .
the sm prediction for the branching ratios are b ( ) = @xmath11 and b ( ) = @xmath12 @xcite .
in addition , an upper limit on b ( ) can be derived with small uncertainty from the current limit on @xmath9 mixing ; this limit is b ( ) @xmath13 @xcite . [ [ section ] ] the alternating gradient synchrotron ( ags ) as a high - energy physics ( hep ) research facility has had a broad and rich program in kaon physics , culminating in the observation of two events by e787 . with the recent successful commissioning of the relativistic heavy ion collider ( rhic ) ,
the primary role of the ags has shifted to become an injector of heavy ions for rhic .
nevertheless , the ags remains the highest intensity proton synchrotron in the world and is designed to be available for @xmath020 hours / day when not filling rhic , and as such retains an important role in the us hep program .
doe and bnl have approved and agreed to fund one new hep experiment to run at the ags between rhic fills : the e949 experiment seeks to make a precise measurement of the branching ratio b ( ) . while doe approved e949 to run for 60 weeks , the fy03 budget of the president of the united states does not include running for e949 . at this point
the e949 experiment will be terminated after only 12 weeks of running , unless the us congress restores funding .
the fate of the e949 experiment will not be known until congress initiates and then completes the appropriations process between june and october of 2002 .
the completion of the main injector ( mi ) at fnal allows for the simultaneous operation of a fixed target program along with the tevatron collider . the next step in the pursuit of will be the ckm experiment , which plans to use a modest fraction of the mi protons ( @xmath14 ) , extracted over a 1-second spill with minimal bunching of the proton beam , and will push the sensitivity to the current limits of theoretical precision .
the first signal was observed in the 1995 data sample of the e787 experiment @xcite .
no new events were seen in the data sample from 199697 @xcite , and with a background of @xmath15 events and a signal of one event a branching ratio of b ( ) = @xmath16 was measured .
that event was in fact in a very clean region of the predefined signal region with a sm - signal to background ratio of 35 . an analysis of the final e787 data sample from the 1998 - 99 run has recently been reported @xcite . with a measured background of @xmath17 , one new event was observed .
the final plot of range vs. energy from the combined e787 199599 data sample for events passing all other cuts is shown in figure [ fig : pnn ] .
the branching ratio , as determined from these two events , is @xmath18 this branching ratio is a factor of two higher than expected in the sm and is higher than allowed by the current limit on @xmath9 mixing .
of course , the uncertainty on the br measurement is large due to limited statistics and new data from the e949 experiment are eagerly awaited .
the new event found in the 1998 data sample is in a relatively clean region of the accepted signal region : the sm signal to background ratio for this event is 3.6 .
an event display for this event , as well as the previous event , is shown in figure [ fig : event ] .
* 1995 event * * 1998 event * limits on and @xmath19 can be obtained ( these are 1-@xmath20 limits except for @xmath6 which is 90% cl ) , @xmath21 even with the large statistical error , this new measurement provides a non - trivial contribution to global fits of the ckm parameters @xcite .
the constraints on @xmath19 from this result are shown in figure [ fig : rho_eta ] .
the constraints from the other golden b modes , and are shown on the same plot .
one can immediately see that @xmath22 is tightly constrained to a narrow crescent by and .
new data from the successor to e787 , e949 , will make a significant contribution to our knowledge of the ckm parameters .
in addition , e787 has searched for the decay in the pion kinematic region below the ( @xmath23 ) peak @xcite .
this region contains more of the phase space , but is complicated by a significant background from decays with the @xmath24 scattering in the scintillating fiber target and shifting its kinematics into the search region .
the data from the 1996 run of e787 has been analyzed ( @xmath020% of the entire e787 data sample ) .
one event was observed in the search region , consistent with the background estimate of @xmath25 .
this implies an upper limit on b ( ) @xmath26 ( 90% c.l . ) , and is consistent with the 2 events observed above the @xmath23 peak and the sm spectrum .
some additional reduction of the background levels in the remaining e787 data may be possible , but the major focus will shift to the new e949 experiment , which has significantly enhanced photon veto capabilities that will further suppress this background .
in addition , the next experiment after e949 , ckm at fnal , will be essentially free of this background since there is no stopping target .
e949 is an upgraded version of the e787 experiment , planning to capitalize on the full ags beam to collect data at 14 times the rate of the e787 run in 1995 .
the new detector has substantially upgraded photon veto capabilities , enhanced tracking , triggering , monitoring , and daq capability , and will run at a higher ags duty factor and a lower kaon momentum ( with an increased fraction of stopped kaons ) .
it has been designed to reach a sensitivity of at least 5 times beyond e787 and observe 510 sm events .
the background level for e949 measurement of b ( ) above the @xmath23 peak is reliably projected from e787 data to be @xmath010% of the standard model signal .
e949 should see up to 10 sm events ( or 20 events at the branching ratio measured by e787 ) within the next couple of years .
this is an exciting opportunity to make a significant contribution to quark mixing and _ cp_violation that should be fully exploited .
a history of the search for is shown in figure [ fig : hist ] .
the next step towards a precision measurement of b ( ) will be the ckm experiment at fnal .
ckm has been given scientific ( stage1 ) approval by fnal and could be running by 2007 .
ckm plans to use a novel technique for : a decay in flight experiment , with redundant kinematic constraints from a conventional momentum spectrometer and a novel velocity spectrometer based on rich counters .
ckm expects to observe 100 sm signal events in a two year run , using the main injector simultaneously with the tevatron .
the background is expected to be @xmath010% of the sm signal , predominantly from .
a vacuum of @xmath27 torr is required to minimize backgrounds from kaon interactions in the decay volume .
ckm will require less than 20% of the flux from the main injector , but will require a slow extracted spill of @xmath01 second duration .
ckm will run with a 50 mhz 22 mev / c beam with rf separators and about 70% k@xmath28 purity .
figure [ fig : hist ] shows a projected measurement of b ( ) from ckm , assuming the current central value of the branching ratio .
[ [ section-1 ] ] the decay is even cleaner theoretically and is _ cp_-violating .
however , it is even more challenging experimentally as all of the particles involved are neutral .
presently , the best limit on is derived in a model - independent way @xcite from the e787 measurement of : @xmath29 of course , it is desirable to observe this mode directly in order to extract a second constraint on the ckm matrix parameters .
the current best direct limit is derived from a ktev search for high transverse momentum @xmath30 s decaying via @xmath31 . from the full 1997 data set , ktev observed no events with an expected background of @xmath32 and set a 90%-cl limit @xcite of @xmath33 due to the small @xmath31 branching ratio all future experiments plan to use the more copious @xmath34 mode .
ktev also made a search in this mode in a special one day test , with a highly collimated ` pencil ' beam and observed one background event , most likely from a neutron interaction , and set a 90%-cl limit @xcite of b ( ) @xmath35 .
the next generation of experiments will start with e391a at the high energy accelerator research organization ( kek ) which hopes to reach a sensitivity of @xmath36 .
this experiment will use a technique similar to ktev , with a pencil beam , high quality calorimetry and very efficient photon vetos .
this is the first experiment dedicated to searching for and aims to close the window for non - sm contributions to the decay .
it will also serve as a test bed for the experimental techniques necessary to observe in future experiments .
beam tests were started in 2001 and the first data - taking run is scheduled for 2003 .
this experiment plans to eventually move to the japanese hadron facility ( jhf ) in @xmath02007 , and attempt to push to a sensitivity of @xmath37(1000 ) events .
the national science board of the national science foundation ( nsf ) has approved the construction of the two new large experiments at the bnl ags : kopio and meco , as components of the rare symmetry violation proposal ( rsvp ) .
rsvp is planned to be one of the next major research equipment construction projects at the nsf .
the kopio experiment is designed to discover the decay and measure its branching ratio to @xmath020% .
kopio will make use of a time - of - flight technique to measure the momentum of the @xmath38 and will operate at a large targeting angle to improve the @xmath39 resolution and soften the neutron spectrum to reduce @xmath30 hadroproduction .
all possible aspects of the decay will be measured : the photon directions will be measured in a pre - radiator , the times and energy will be precisely measured in a shashlyk calorimeter , the time of the kaon s creation is determined by the proton bunch width from the ags ( @xmath0250 psec ) .
kopio will have a substantial photon veto system and make the same sort of background measurements as e787 , directly from the data , with two independent tools for attacking the major background , , through both kinematics and photon veto .
kopio expects to observe 50 sm events , with a background of 50% .
this will allow a determination of @xmath6 to 10% .
kopio is expected to start data collection in @xmath02006 .
the next decade will be an exciting time for improved understanding of _ cp_-violation and quark mixing .
it is quite likely that precise measurements of all four golden modes will be made : at the b - factories ; , most likely at the tevatron ; and at bnl , kek and fnal .
these measurements will allow a precise determination of ckm parameters and provide a critical test of the sm picture of _ cp_-violation .
i would like to thank takao inagaki , bob tschirhart , taku yamanaka , takeshi komatsubara , shojiro sugimoto , david jaffe and laurie littenberg for discussions and/or comments on this paper .
this work was supported in part under us department of energy contract # de - ac02 - 98ch10886 .
99 a. buras , hep - ph/0101336 ( 2001 ) ; a. buras and r. fleischer , phys . rev . *
d64 * , 115010 ( 2001 ) .
g. bucahalla and a. buras , nucl .
b548 * , 309 ( 1999 ) .
g. bucahalla and a. buras , phys . lett .
* b333 * , 221 ( 1994 ) ; g. bucahalla and a. buras , phys . rev .
* d54 * , 6782 ( 1996 ) ; y. nir and m.p .
worah , phys . lett . *
b423 * , 319 ( 1998 ) ; s. bergmann and g. perez , jhep * 8 * , 034 ( 2000 ) .
y. grossman and y. nir , phys .
b398 * , 163 ( 1997 ) .
g. dambrosio and g. isidori , hep - ph/0112135 s. adler , _ et al .
lett . * 79 * , 2204 ( 1997 ) .
s. adler , _ et al .
lett . * 84 * , 3768 ( 2000 ) .
s. adler , _ et al .
rev . lett . * 88 * , 041803 ( 2002 )
. s. adler , _ et al .
b537 * , 211 ( 2002 ) .
y. grossman and y. nir , phys .
b398 * , 163 ( 1997 ) .
a. alavi - harati _ et al .
_ , phys . rev . *
d61 * , 072006 ( 2000 )
. j. adams , _ et al .
b447 * , 240 ( 1999 ) . | the experimental program for the study of the rare kaon decays , , is summarized .
a review of recent results is provided along with a discussion of prospects for the future of this program .
the primary focus of the world - wide kaon program is the two golden modes : and .
the first step in an ambitious program to precisely measure both branching ratios has been successfully completed with the observation of two events by e787 .
the e949 experiment is poised to reach an order of magnitude further in sensitivity and to observe @xmath010 standard model events , and the ckm experiment should observe @xmath0100 sm events by the end of this decade .
limits on the neutral analog have been set by ktev and within the next couple of years will be pushed by e391a .
measurements of the branching ratio should be made within the next 10 years by kopio , with a goal of @xmath050 events , and at the jhf , with a goal of up to 1000 events .
= 11.6pt |
the birkhoff polytope , which we will denote as @xmath1 , has been extensively studied and generalized .
it is defined as the convex hull of the @xmath2 permutation matrices as vectors in @xmath3 .
many analogous polytopes have been studied which are subsets of @xmath1 ( see e.g. @xcite ) . in contrast
, we study a polytope containing @xmath1 .
we begin with the following definitions .
[ def : asm ] alternating sign matrices ( asms ) are square matrices with the following properties : * entries @xmath4 * the entries in each row and column sum to 1 * nonzero entries in each row and column alternate in sign @xmath5 @xmath6 [ fig : n3asm ] the total number of @xmath2 alternating sign matrices is given by the expression @xmath7 mills , robbins , and rumsey conjectured this formula @xcite , and then over a decade later doron zeilberger proved it @xcite .
shortly thereafter , kuperberg found a bijection between asms and the statistical physics model of square ice with domain wall boundary conditions ( which is very similar to the _ simple flow grids _ defined later in this paper ) , and gave a shorter proof using insights from physics @xcite .
for a detailed exposition of the conjecture and proof of the enumeration of asms , see @xcite .
see figure [ fig : n3asm ] for the seven @xmath8 asms .
the @xmath9th alternating sign matrix polytope , which we will denote as @xmath10 , is the convex hull in @xmath3 of the @xmath0 alternating sign matrices . from definition [ def : asm ]
we see that permutation matrices are the alternating sign matrices whose entries are nonnegative .
thus @xmath1 is contained in @xmath10 .
the connection between permutation matrices and asms is much deeper than simply containment .
there exists a partial ordering on alternating sign matrices that is a distributive lattice .
this lattice contains as a subposet the bruhat order on the symmetric group , and in fact , it is the smallest lattice that does so ( i.e. it is the macneille completion of the bruhat order ) @xcite .
given this close relationship between permutations and asms it is natural to hope for theorems for @xmath10 analogous to those known for @xmath1 . in this paper
we find analogues for @xmath10 of the following theorems about the birkhoff polytope ( see the discussion in @xcite and @xcite ) .
* @xmath1 consists of the @xmath0 nonnegative doubly stochastic matrices ( square matrices with nonnegative real entries whose rows and columns sum to 1 ) . *
the dimension of @xmath1 is @xmath11 .
* @xmath1 has @xmath12 vertices .
* @xmath1 has @xmath13 facets ( for @xmath14 ) where each facet is made up of all nonnegative doubly stochastic matrices with a 0 in a specified entry .
* @xmath1 projects onto the permutohedron .
* there exists a nice characterization of its face lattice in terms of elementary bipartite graphs @xcite .
as we shall see in theorem [ thm : asmchasing ] , the row and column sums of every matrix in @xmath10 must equal 1 .
thus the dimension of @xmath10 is @xmath11 because , just as for the birkhoff polytope , the last entry in each row and column is determined to be precisely what is needed to make that row or column sum equal 1 . in section [ sec : properties ] we prove that @xmath10 has @xmath15 $ ] facets and its vertices are the alternating sign matrices .
we also prove analogous theorems about the inequality description of @xmath10 ( section [ sec : asm ] ) , the face lattice ( section [ sec : facelattice ] ) , and the projection to the permutohedron ( section [ sec : properties ] ) .
see @xcite for background and terminology on polytopes .
the alternating sign matrix polytope was independently defined in @xcite in which the authors also study the integer points in the @xmath16th dilate of @xmath10 calling them _ higher spin alternating sign matrices_.
the main theorem about the birkhoff polytope is the theorem of birkhoff @xcite and von neumann @xcite which says that the birkhoff polytope can be described not only as the convex hull of the permutation matrices but equivalently as the set of all nonnegative doubly stochastic matrices ( real square matrices with row and column sums equaling 1 whose entries are nonnegative ) .
the inequality description of the alternating sign matrix polytope is similar to that of the birkhoff polytope .
it consists of the subset of doubly stochastic matrices ( now allowed to have negative entries ) whose partial sums in each row and column are between 0 and 1 .
the proof uses the idea of von neumann s proof of the inequality description of the birkhoff polytope @xcite .
note that in @xcite behrend and knight approach the equivalence of the convex hull definition and the inequality description of @xmath10 in the opposite manner , defining the alternating sign matrix polytope in terms of inequalities and then proving that the vertices are the alternating sign matrices .
[ thm : asmchasing ] the convex hull of @xmath0 alternating sign matrices consists of all @xmath0 real matrices @xmath17 such that : @xmath18 call the subset of @xmath3 given by the above inequalities @xmath19 .
it is easy to check that the convex hull of the alternating sign matrices is contained in the set @xmath19 .
it remains to show that any @xmath20 can be written as a convex combination of alternating sign matrices .
let @xmath20 .
let @xmath21 and @xmath22 .
thus the @xmath23 are the row partial sums and the @xmath24 are the column partial sums .
it follows from ( [ eq : partialcolumnsum ] ) and ( [ eq : partialrowsum ] ) that @xmath25 for all @xmath26 .
also , from ( [ eq : sumagain ] ) and ( [ eq : sumagain2 ] ) we see that @xmath27 for all @xmath26 .
if we set @xmath28 we see that every entry @xmath29 satisfies @xmath30 .
thus @xmath31 using von neumann s terminology , we call a real number @xmath32 _ inner _ if @xmath33 .
we construct a circuit in @xmath34 such that the partial sum between adjacent matrix entries in the circuit be an inner .
so we rewrite the matrix @xmath34 with the partial sums between entries as shown below .
@xmath35 begin at the vertex to the left or above any inner partial sum ; if no such partial sum exists , then @xmath34 is an alternating sign matrix .
then there exists an adjacent inner partial sum by ( [ eq : rcrel ] ) . by repeated application of ( [ eq : rcrel ] ) to each new inner partial sum , a path can then be formed by moving from entry to entry of @xmath34 along inner partial sums . since @xmath34 is of finite size and all the boundary partial sums are 0 or 1 ( i.e. non inner )
, the path eventually reaches an entry in the same row or column as a previous entry yielding a circuit in @xmath34 whose partial sums are all inner . using this circuit
we can write @xmath34 as a convex combination of two matrices in @xmath19 , each with at least one more non inner partial sum , in the following way .
@xmath36 label the corner matrix entries in the circuit alternately ( @xmath37 ) and ( @xmath38 ) . define @xmath39 where @xmath23 , @xmath40 , @xmath41 , and @xmath42 are taken over respectively the row partial sums to the right of a ( @xmath38 ) corner along the circuit ,
the row partial sums to the right of a ( @xmath37 ) corner , the column partial sums below a ( @xmath38 ) corner , and the column partial sums below a ( @xmath37 ) corner .
subtract @xmath43 from the entries labeled ( @xmath38 ) and add @xmath43 to the entries labeled ( @xmath37 ) .
subtracting and adding @xmath43 in this way preserves the row and column sums and keeps all the partial sums weakly between 0 and 1 ( satisfying ( [ eq : partialcolumnsum])([eq : sumagain2 ] ) ) , so the result is another matrix @xmath44 in @xmath19 with at least one more non inner partial sum than @xmath34 .
now give opposite labels to the corners in the circuit in @xmath34 and subtract and add another constant @xmath45 in a similar way to obtain another matrix @xmath46 in @xmath19 with at least one more non inner partial sum than @xmath34 .
then @xmath34 is a convex combination of @xmath44 and @xmath46 , namely @xmath47 .
therefore , by repeatedly applying this procedure , @xmath34 can be written as a convex combination of alternating sign matrices ( i.e. matrices of @xmath19 with no inner partial sums ) .
now that we can describe the alternating sign matrix polytope in terms of inequalities , let us use this inequality description to examine some of the properties of @xmath10 , namely , its facets , its vertices , and its projection to the permutohedron . to make the proofs of the next two theorems more transparent
, we introduce modified square ice configurations called _ simple flow grids _ which will be used more extensively in section [ sec : facelattice ] . consider a directed graph with @xmath48 vertices : @xmath13 ` internal ' vertices @xmath49 and @xmath50 ` boundary ' vertices @xmath51 , @xmath52 , @xmath53 , and @xmath54 where @xmath55 .
these vertices are naturally depicted in a grid in which vertex @xmath49 appears in row @xmath56 and column @xmath57 .
define the _ complete flow grid _
@xmath58 to be the directed graph on these vertices with edge set @xmath59 for @xmath55 .
so @xmath58 has directed edges pointing in both direction between neighboring internal vertices in the grid , and also directed edges from internal vertices to neighboring border vertices .
[ definition : simpleflowgrid ] a _ simple flow grid _ of order @xmath9 is a subgraph of @xmath58 consisting of all the vertices of @xmath58 for which four edges are incident to each internal vertex : either four edges directed inward , four edges directed outward , or two horizontal edges pointing in the same direction and two vertical edges pointing in the same direction .
[ prop : sfgasmbij ] there exists an explicit bijection between simple flow grids of order @xmath9 and @xmath0 alternating sign matrices .
given an asm @xmath60 , we will define a corresponding directed graph @xmath61 on the @xmath13 internal vertices and @xmath50 boundary vertices arranged on a grid as described above .
let each entry @xmath62 of @xmath60 correspond to the internal vertex @xmath49 of @xmath61 . for neighboring vertices @xmath63 and @xmath64 in @xmath61 let there be a directed edge from @xmath63 to @xmath64 if the partial sum from the border of the matrix to the entry corresponding to @xmath63 in the direction pointing toward @xmath64 equals 1 . by the definition of alternating sign matrices
, there will be exactly one directed edge between each pair of neighboring internal vertices and also a directed edge from an internal vertex to each neighboring border vertex .
vertices of @xmath61 corresponding to 1 s are sources and vertices corresponding to @xmath65 s are sinks .
the directions of the rest of the edges in @xmath61 are determined by the placement of the 1 s and @xmath65 s , in that there is a series of directed edges emanating from the 1 s and continuing until they reach a sink or a border vertex .
thus @xmath61 is a simple flow grid .
also , given a simple flow grid we can easily find the corresponding asm by replacing all the sources with 1 s and all the sinks with @xmath65 s . thus simple flow grids are in one - to - one correspondence with asms ( see figure [ fig:5by5grid ] ) .
@xmath66{5by5griddwbc.eps } \hspace{.7 in } \includegraphics[scale=0.25]{matrix.eps}\ ] ] simple flow grids are , in fact , almost the same as configurations of the six - vertex model of square ice with domain wall boundary conditions ( see the discussion in @xcite ) , the only difference being that the horizontal arrows point in the opposite direction .
recall that for @xmath67 the birkhoff polytope has @xmath13 facets ( faces of dimension one less than the polytope itself ) .
( @xmath68 is simply a line segment , so the number of facets equals the number of vertices which is 2 . )
each facet of the birkhoff polytope consists of all nonnegative doubly stochastic matrices with a zero in a fixed entry , that is , where one of the defining inequalities is made into an equality .
the analogous theorem for @xmath10 is the following .
[ theorem : facets ] @xmath10 has @xmath15 $ ] facets , for @xmath67 . note that the @xmath69 defining inequalities for @xmath70 given in ( [ eq : partialcolumnsum ] ) and ( [ eq : partialrowsum ] ) can be restated as @xmath71 @xmath72 for @xmath55 .
we have rewritten the statement that the row and column partial sums from the left or top must be less than or equal to @xmath73 as the row and column partial sums from the right and bottom must be greater than or equal to @xmath74 . by counting these defining inequalities
, one sees that there could be at most @xmath75 facets , each determined by making one of the above inequalities an equality .
it is left to determine how many of these equalities determine a face of dimension less than @xmath76 . by symmetry
we can determine the number of facets coming from the inequalities @xmath77 for @xmath78 and then multiply by 4 . since the full row and column sums always equal @xmath73 , the equalities such as @xmath79 yield the empty face ( @xmath80 ) .
also , @xmath81 is implied from the fact that @xmath82 ( @xmath83 ) .
the inequalities @xmath84 for all @xmath85 , i.e. the entries in the first column are nonnegative , imply that @xmath86 for @xmath87 ( @xmath88 ) , thus each of these sets is a face of dimension less than @xmath76 , and similarly for @xmath89 for @xmath87 ( @xmath90 ) the partial sums of the last column .
so we are left with the @xmath91 inequalities @xmath92 for @xmath93 and @xmath94 along with the inequality @xmath95 . for our symmetry argument to work
, we do not include @xmath96 in our count since @xmath96 is also an inequality of the form @xmath97 .
thus @xmath10 has at most @xmath15 $ ] facets , given explicitly by the @xmath98 sets of all @xmath70 which satisfy one of the following : @xmath99 @xmath100 they are facets ( not just faces ) since each equality determines exactly one more entry of the matrix , decreasing the dimension by one .
recall that a directed edge in a simple flow grid @xmath61 represents a location in the corresponding asm @xmath60 where the partial sum equals 1 , thus a directed edge missing from @xmath61 represents a location in @xmath60 where the partial sum equals 0 .
thus we can represent each of the @xmath101 facets of ( [ eq : faceteq ] ) as subgraphs of the complete flow grid @xmath58 from which a single directed edge has been removed : @xmath102 or @xmath103 with @xmath104 .
we can represent the facets of ( [ eq : corners ] ) as subgraphs of @xmath58 from which two directed edges have been removed : @xmath105 and @xmath106 , @xmath107 and @xmath108 , @xmath109 and @xmath110 , or @xmath111 and @xmath112 .
now given any two facets @xmath113 and @xmath114 , it is easy to exhibit a pair of asms @xmath115 such that @xmath116 lies on @xmath113 and not on @xmath114 . include the directed edge(s ) corresponding to @xmath114 but not the directed edge(s ) corresponding to @xmath113 in @xmath117 , then do the opposite for @xmath118 .
thus each of the @xmath15 $ ] equalities gives rise to a unique facet .
for @xmath67 , the number of facets of @xmath10 on which an asm @xmath60 lies is given by @xmath119 each @xmath74 around the border of @xmath60 represents one facet .
thus the number of facets corresponding to border zeros of @xmath60 equals @xmath120 .
then there are @xmath121 facets represented by directed edges pointing in the opposite directions to the directed edges in the @xmath122 interior array of @xmath61 .
the sum of these numbers gives the above count .
even though @xmath10 is defined as the convex hull of the asms , it requires some proof that each asm is actually an extreme point of @xmath10 .
the vertices of @xmath10 are the @xmath0 alternating sign matrices .
fix an @xmath0 asm @xmath60 . in order to show that @xmath60 is a vertex of @xmath10 , we need to find a hyperplane with @xmath60 on one side and all the other asms on the other side .
then since @xmath10 is the convex hull of @xmath0 asms , @xmath60 would necessarily be a vertex .
consider the simple flow grid corresponding to @xmath60 . in any simple flow grid
there are , by definition , @xmath123 directed edges , where for each entry of the corresponding asm there is a directed edge whenever the partial sum in that direction up to that point equals 1 . since the total number of directed edges in a simple flow grid is fixed , @xmath60 is the only asm with all of those partial sums equaling 1 .
thus the hyperplane where the sum of those partial sums equals @xmath124 will have @xmath60 on one side and all the other asms on the other .
thus the @xmath0 asms are the vertices of @xmath10 .
another interesting property of the asm polytope is its relationship to the permutohedron . for a vector @xmath125 with distinct entries , define the permutohedron @xmath126 as the convex hull of all vectors obtained by permuting the entries of @xmath127
that is , @xmath128 also , for such a vector @xmath127 , let @xmath129 be the mapping from the set of @xmath0 real matrices to @xmath130 defined by @xmath131 it is well known , and follows immediately from the definitions , that @xmath126 is the image of the birkhoff polytope under the projection @xmath129 .
[ prop : birkpermut ] let @xmath1 be the birkhoff polytope and @xmath127 be a vector in @xmath130 with distinct entries .
then @xmath132 this result is one of many classical results about the birkhoff polytope dating back to hardy , littlewood , and plya @xcite @xcite .
see @xcite for a nice summary of relevant results .
the next theorem states that when the same projection map is applied to @xmath10 , the image is the same permutohedron whenever @xmath127 is a decreasing vector . for the proof of this theorem we will need the concept of majorization @xcite .
[ def : maj ] let @xmath133 and @xmath63 be vectors of length @xmath9
. then @xmath134 ( that is @xmath133 is _ majorized _ by @xmath63 ) if @xmath135 } \le \sum_{i=1}^k v_{[i]},&\mbox{for } 1\le k\le n-1\\ \sum_{i=1}^n u_i = \sum_{i=1}^n v_i&\end{cases}\ ] ] where the vector @xmath136},u_{[2]},\ldots , u_{[n]})$ ] is obtained from @xmath133 by rearranging its components so that they are in decreasing order , and similarly for @xmath63 .
let @xmath127 be a decreasing vector in @xmath130 with distinct entries .
then @xmath137 it follows from proposition [ prop : birkpermut ] and @xmath138 that @xmath139 .
thus it only remains to be shown that @xmath140 .
let @xmath127 be a decreasing @xmath9vector ( so that @xmath141}$ ] ) and @xmath142 an @xmath0 asm .
then there is a proposition of rado which states that for vectors @xmath133 and @xmath63 of length @xmath9 , @xmath134 if and only if @xmath133 lies in the convex hull of the @xmath12 permutations of the entries of @xmath63 @xcite .
therefore the proof will be completed by showing @xmath143 . by definition [ def : maj ]
we need to show @xmath144}&\le \sum_{j=1}^k z_{j } , \hspace{.4 in } 1\le k\le n-1\\ \label{eq : proj2 } \sum_{j=1}^n ( z x)_j&= \sum_{j=1}^n z_j\end{aligned}\ ] ] where the @xmath57th component @xmath145 of @xmath146 is given by @xmath147 . to verify ( [ eq : proj2 ] ) note that since @xmath148 , @xmath149 to prove ( [ eq : proj1 ] ) we will show that @xmath150 given any @xmath151 , so that in particular @xmath152}\le \sum_{j=1}^{|j| } z_j$ ] .
we will need to verify the following : @xmath153 to prove ( [ eq : vim ] ) note that @xmath154 since @xmath155 .
but also , since @xmath156 and @xmath157 we have that @xmath158 to prove ( [ eq : vik ] ) observe , @xmath159 since the columns of @xmath34 sum to 1 . therefore using ( [ eq : vim ] ) and ( [ eq : vik ] ) we see that @xmath160 thus @xmath161 and so @xmath146 is contained in the convex hull of the permutations of @xmath127 .
therefore @xmath162 .
another nice result about the birkhoff polytope is the structure of its face lattice @xcite .
associate to each permutation matrix @xmath34 a bipartite graph with vertices @xmath163 and @xmath164 where there is an edge connecting @xmath165 and @xmath166 if and only if there is a 1 in the @xmath49 position of @xmath34 .
such a graph will be a perfect matching on the complete bipartite graph @xmath167 .
a graph @xmath168 is called elementary if every edge is a member of some perfect matching of @xmath168 .
[ prop : elementary ] the face lattice of the birkhoff polytope is isomorphic to the lattice of elementary subgraphs of @xmath167 ordered by inclusion . this lattice structure was first identified by billera and sarangarajan in @xcite and @xcite , but the set of faces itself was first characterized and studied extensively by brualdi and gibson in @xcite and @xcite using certain 0 - 1 matrices which correspond trivially to elementary subgraphs of @xmath167 .
other relevant results were also obtained by balinski and russakoff in @xcite and @xcite .
a similar statement can be made about the face lattice of the asm polytope using simple flow grids ( see definition [ definition : simpleflowgrid ] ) in place of perfect matchings , the complete flow grid @xmath58 instead of the complete bipartite graph @xmath167 , and _ elementary flow grids _ in place of elementary graphs .
an _ elementary flow grid _
@xmath168 is a subgraph of the complete flow grid @xmath58 such that the edge set of @xmath168 is the union of the edge sets of simple flow grids .
now for any face @xmath169 of @xmath10 define the grid corresponding to the face , @xmath170 , to be the union over all the vertices of @xmath169 of the simple flow grids corresponding to the vertices .
that is , @xmath171 thus @xmath170 is an elementary flow grid since its edge set is the union of the edge sets of simple flow grids .
now we wish to define the converse , that is , given an elementary flow grid @xmath168 we would like to know the corresponding face @xmath172 of @xmath10 .
define @xmath172 to be the convex hull of the vertices of @xmath10 whose corresponding simple flow grids are contained in the elementary flow grid @xmath168 .
so let @xmath173 recall that we can represent each of the facets of @xmath10 either as subgraphs of the complete flow grid @xmath58 from which one of the directed edges in the set @xmath174 , @xmath103 @xmath175 @xmath176 has been removed or from which one of the pairs of directed edges in the set @xmath177 , @xmath178 , @xmath179 , @xmath180 has been removed .
thus each of the directed edges in @xmath181 and the first of each pair of directed edges in @xmath182 that are not in @xmath168 represent facets that contain @xmath172 .
let the collection of these directed edges be called @xmath183 and their corresponding facets @xmath184 .
let @xmath185 be the intersection of these facets .
thus @xmath186 is a face of @xmath10 and @xmath187 .
we wish to show that @xmath172 equals @xmath186 .
so suppose @xmath188 .
then since @xmath186 is a face of @xmath10 and @xmath172 is defined as the convex hull of vertices of @xmath10 there exists an additional vertex @xmath189 of @xmath10 such that @xmath190 .
but @xmath191 must be missing the directed edges @xmath192 since @xmath189 , thus all the directed edges of @xmath191 must be in @xmath168 .
therefore @xmath193 so that @xmath194 which is a contradiction .
so @xmath195 .
thus @xmath172 is a face of @xmath10 since it is the intersection of faces of @xmath10 .
it can easily be seen that @xmath196 and @xmath197 . also if @xmath113 and @xmath114 are faces of @xmath10 then @xmath198 if and only if @xmath199 . thus elementary flow grids are in bijection with the faces of @xmath10 ( if we also regard the empty grid as an elementary flow grid ) .
elementary flow grids can be made into a lattice by inclusion , where the join is the union of the edge sets and the meet is the largest elementary flow grid made up of the directed edges from the intersection of the edges sets .
this discussion yields the following theorem : [ thm : facelattice ] the face lattice of @xmath10 is isomorphic to the lattice of all @xmath0 elementary flow grids ( or equivalently all @xmath0 square ice configurations with domain wall boundary conditions ) ordered by inclusion .
the dimension of any face of @xmath10 can be determined by looking at @xmath170 as in the following theorem .
the characterization of edges of @xmath10 is analogous to the result for the birkhoff polytope which states that the graphs representing edges of @xmath1 are the elementary subgraphs of @xmath167 which have exactly one cycle @xcite @xcite @xcite . given an elementary flow grid @xmath168 , define a _ doubly directed region _ as a collection of cells in @xmath168 completely bounded by double directed edges but containing no double directed edges in the interior ( see figure [ fig : redbluecircuits ] ) .
let @xmath200 denote the number of doubly directed regions in @xmath168 .
[ thm : dim ] the dimension of a face @xmath169 of @xmath10 is the number of doubly directed regions in the corresponding elementary flow grid @xmath170 . in particular ,
the edges of @xmath10 are represented by elementary flow grids containing exactly one cycle of double directed edges . to a 3-dimensional face of @xmath201
we proceed by induction on the dimension of the face of @xmath10 .
the simple flow grid corresponding to any asm @xmath60 has no double directed edges , thus @xmath202 .
now suppose for every @xmath203dimensional face of @xmath10 , the number of doubly directed regions of the elementary flow grid corresponding to the face equals @xmath203 .
let @xmath169 be an @xmath204dimensional face of @xmath10 and @xmath205 an @xmath203dimensional subface of @xmath169 .
we assume @xmath206 and wish to show that @xmath207 .
now @xmath170 is the elementary flow grid whose edge set is the union of the edge sets of @xmath208 and @xmath61 over all asms @xmath60 in @xmath209 .
every vertex in a simple flow grid must have even indegree and even outdegree .
therefore , if we wish to obtain @xmath61 from @xmath210 , where @xmath211 is an asm in @xmath205 , by reversing some directed edges , the number of directed edges reversed at each vertex must be even . thus taking the union of the directed edges of @xmath210 with the directed edges of @xmath61 forms one or more circuits of double directed edges , where at least one of the double directed edges is not in @xmath208 .
therefore @xmath170 has at least one more doubly directed region than @xmath208 , so @xmath212 .
then since @xmath213 equals the complete flow grid @xmath58 , we have that @xmath214 . therefore moving up the face lattice
one rank increases the number of doubly directed regions by exactly one , so @xmath207 .
see figure [ fig : redcircuit ] for the elementary flow grid representing the edge in @xmath215 between @xmath216 ]
this work is based on research which is a part of the author s doctoral thesis at the university of minnesota under the direction of dennis stanton .
the author would like to thank professor stanton for the many helpful discussions and encouragement .
m. l. balinski and a. russakoff , some properties of the assignment polytope , 1972 .
programming , * 3 * , 257258 . m. l. balinski and a. russakoff , on the assignment polytope , 1974 .
, * 16 * , 516525 . r. e. behrend and v. a. knight , higher spin alternating sign matrices , 2007 . electron
. j. combin . , * 14 * , no .
l. j. billera and a. sarangarajan , all @xmath74-@xmath73 polytopes are traveling salesman polytopes , 1996 .
combinatorica , * 16 * , no .
2 , 175188 . l. j. billera and a. sarangarajan , the combinatorics of permutation polytopes , 1996 .
dimacs ser .
discrete mat .
24 , amer . math .
soc . , 123
. g. birkhoff , three observations on linear algebra , 1946 .
revista a. , * 5 * , 147151 .
d. m. bressoud , proofs and confirmations ; the story of the alternating sign matrix conjecture , maa spectrum , mathematical association of america , washington , dc , 1999 .
r. a. brualdi , combinatorial matrix classes , cambridge university press , 2006 .
r. a. brualdi and p. m. gibson , the assignment polytope , 1976 .
programming , * 11 * , 97101 .
r. a. brualdi and p. m. gibson , convex polyhedra of doubly stochastic matrices .
i. applications of the permanent function , 1977 .
j. combin .
theory ser .
a , * 22 * , 194230 . g. h. hardy , j. e. littlewood , and g. plya , some simple inequalities satisfied by convex functions , 1929 . messenger math . * 58 * , 145152 . g. h. hardy , j. e. littlewood , and g. plya , inequalities ( second edition ) , cambridge university press , 1952 .
g. kuperberg , another proof of the asm conjecture , 1996 .
math . not .
, * 3 * , 139 - 150 .
a. lascoux and m. schtzenberger , treillis et bases des groupes de coxeter , 1996 .
j. combin . , * 3 * , no .
2 . a. w. marshall and i. olkin , inequalities : theory of majorization and its applications , mathematics in science and engineering , * 143 * , academic press inc .
[ harcourt brace jovanovich publishers ] , new york , 1979 .
w. h. mills , d. p. robbins , and h. rumsey , jr .
, alternating sign matrices and descending plane partitions , 1983 .
j. combin .
theory ser .
a , * 34 * , no .
3 , 340359 . l. mirsky and z. wahrsch , results and problems in the theory of doubly stochastic matrices , 1963 . verw .
gebiete , * 1 * , 319334 .
r. rado , an inequality , 1952 .
j. london math .
soc . , * 27 * , 16
. j. von neumann , a certain zero sum two person game equivalent to the optimal assignment problem , contributions to the theory of games vol . 2 , 1953 .
annals of mathematics studies no .
28 , princeton university press , 512 .
v. a. yemelichev , m. m. kovalev , and m. k. kravtsov , polytopes , graphs , and optimisation , cambridge university press , 1984 .
d. zeilberger , proof of the alternating sign matrix conjecture , 1996 .
j. combin .
* 3 * , no .
2 . g. m. ziegler , lectures on polytopes , springer - verlag , new york , 1995 . | we define the alternating sign matrix polytope as the convex hull of @xmath0 alternating sign matrices and prove its equivalent description in terms of inequalities .
this is analogous to the well known result of birkhoff and von neumann that the convex hull of the permutation matrices equals the set of all nonnegative doubly stochastic matrices .
we count the facets and vertices of the alternating sign matrix polytope and describe its projection to the permutohedron as well as give a complete characterization of its face lattice in terms of modified square ice configurations .
furthermore we prove that the dimension of any face can be easily determined from this characterization . |
* preparation of a spin - imbalanced gas in an optical lattice * we realize the fermi - hubbard model using a degenerate mixture of two zeeman states ( @xmath77 and @xmath78 , numbered up from lowest energy ) in the ground state hyperfine manifold of @xmath23li placed in an optical lattice .
the global spin imbalance @xmath79 can be varied continuously from 0 to @xmath25 while the temperature @xmath80 remains relatively unaffected . to create the sample we load a magneto - optical trap ( mot ) from a zeeman slower , then use a compressed mot stage before loading into a @xmath81 deep optical trap and evaporating near the feshbach resonance .
we stop the evaporation before feshbach molecules form and transfer the atoms to a highly anisotropic ` light sheet ' trap with aspect ratio @xmath82 where it undergoes further evaporation near where @xmath83 .
next we transfer to the final trap geometry where a beam provides radial confinement and a accordion lattice with trapping frequency @xmath84 provides axial confinement ( for further details see @xcite ) .
the spin populations are imbalanced by evaporating the mixture in a magnetic gradient of up to along the same direction as the magnetic bias field , which we set in the range 75- depending on the targeted imbalance .
we then tune the bias field to set the scattering length and load into a 2d square lattice with a long ramp to a depth of @xmath85 .
* details of the 2d lattice * we use a non - separable 2d lattice with spacing formed by four interfering passes of a single vertically polarized beam .
the lattice geometry is shown in fig .
[ fig : lattice ] and described in @xcite . compared to the commonly used lattice configuration created by two non - interfering orthogonal
retro - reflected beams , this lattice has a spacing a factor of @xmath86 larger and a much larger depth because of the 4-fold interference , features that facilitate quantum gas microscopy .
here we have to take into account that the two incoming lattice beams are not exactly at a angle . to compensate the resulting tunneling asymmetry in the lattice and retain a symmetric lattice we attenuate the retro - reflected laser power .
the potential at the atoms is given by where we align the coordinate axes with the principle axes of the lattice .
the lattice beams travel in the @xmath88 and @xmath89 directions . here , @xmath90 is the full lattice depth , @xmath91 the angle between the first two incoming lattice beams , @xmath92 and @xmath93 the electric field amplitude attenuation factor of the retro - reflected beam . for @xmath94
the lattice constant is @xmath95 . the lattice beam waist is .
* calibration of hubbard parameters * we extract our lattice parameters from atom resolved images , finding that the lattice axes are at an angle of @xmath96 and the lattice constants are @xmath97 and @xmath98 from which we extract the intersection angle of the lattice beams , @xmath99 .
we experimentally compensate for the tunneling asymmetry by finding the retro - reflection power that gives symmetric tunneling by making the singles - density correlators equal along both lattice axes .
we calibrate our lattice depth by measuring the frequencies of the three @xmath100 bands in a deep lattice using lattice amplitude modulation , and compare these with the 2d band structure calculation , including the measured lattice angle and the retro - reflection attenuation factor @xmath93 as a fit parameter . at the best fit @xmath101 , the inferred depth of the lattice at which our measurements are performed is @xmath102 , where @xmath103 . from
that we obtain tight - binding tunneling values @xmath104 hz , @xmath105 hz ( @xmath106 ) , confirming our experimental procedure for making the tunneling symmetric . the reduction of the lattice depth across the cloud due to the gaussian profile of the lattice beams leads to an increase in the tunneling by at the edge of the cloud compared to the central value .
we also have a non - zero but negligible diagonal tunneling @xmath107 , due to the non - separability of the lattice we measure the interaction energy @xmath0 using radio frequency spectroscopy .
we transfer atoms from state @xmath108 to @xmath109 and resolve the frequency shift between singly and doubly occupied sites .
we determine @xmath110 taking into account a small correction due to weak final state interactions .
the experimentally measured value agrees with the value determined from band structure calculations including higher band corrections @xcite to within 10% .
* measurement of spin correlations * after the imbalanced gas is loaded adiabatically into the optical lattice , we freeze tunneling dynamics by increasing the lattice depth within to 55 @xmath102 .
after that , we convert the @xmath111 mixture to a @xmath112 mixture by driving a radio - frequency landau - zener transition between states @xmath113 and @xmath109 in with efficiency of @xmath114 .
this step is necessary because an rf transition between state @xmath108 and @xmath113 is forbidden , so we can not perform pseudospin rotations with the @xmath115 mixture .
next we hide all doublons in our sample by converting them to molecules , employing a ramp across a narrow feshbach resonance near . then we either drive a @xmath45 spin rotation to measure correlations for the spin component perpendicular to the effective magnetic field , or omit this to measure correlations for the spin component parallel to the field .
we image the final atom distribution by increasing the lattice depth to 2500 @xmath102 within , ramping up the light sheet to provide axial confinement , and then collecting fluorescence photons during raman cooling .
raman imaging is not spin sensitive , so to measure spin correlations we remove one spin state from the trap before imaging .
we take three sets of images , one with @xmath116 atoms removed , one with @xmath117 atoms removed , and one with no removal pulse . *
spin removal * for removing atoms in a particular spin state , we illuminate the cloud with a resonant pulse .
the length of this pulse was determined by observing atom loss for different pulse lengths .
the atom loss curve exhibits two well - separated time scales , which we interpret as resonant heating and off - resonant heating of the two spin states , and fit with the sum of two decaying exponentials @xcite .
we find time constants and , from which we estimate that @xmath118 of resonantly blown atoms remain , and up to @xmath119 of the off - resonant state atoms are ejected .
the resonant pulse also pumps some of state @xmath120 ( @xmath121 ) atoms into state @xmath122 ( @xmath123 ) .
we take advantage of the cycling state @xmath124 to calibrate these values . to measure the probability of pumping @xmath120 into @xmath122
we prepare a @xmath125 mott insulator and compare the results blowing both states with blowing only state @xmath126 .
we find that we pump of the atoms from @xmath120 to @xmath122 .
we separately measured the probability of pumping @xmath121 into state @xmath123 to be by preparing a @xmath127 mott insulator and comparing blowing both states with blowing @xmath121 only .
* handling doublons * we observe that resonantly blowing one of the spin states does not lead to loss of a doublon , but ejects only a single atom .
the other atom , which has a different spin state , remains behind .
this is in contrast to the behavior observed during raman imaging , where doublons are lost through light - assisted collisions . as the spin correlator
can not be easily extracted with the extra atoms from the doublon blowing , we hide doublons by ramping the magnetic field over the narrow @xmath128 feshbach resonance near thereby converting doublons to feshbach molecules before resonantly blowing . to estimate the fidelity of doublon hiding , we first prepare a band insulator .
we take images with neither doublon hiding nor blowing to determine the number of singles in the band insulator .
we then compare this to the number of singles we observe after performing doublon hiding and removing one of the spin states with a blowing pulse , which leaves behind single atoms on sites where the doublon hiding failed .
we find a doublon hiding fidelity of @xmath129 . *
raman imaging * we use a raman cooling scheme for lithium similar to the ones demonstrated before @xcite but with slight differences .
we use a lin - perp - lin configuration for the raman beams in a retro - reflected configuration .
the raman axis is along the diagonal of the square lattice and tilted about @xmath130 out of plane .
the optical pumping beam is circularly polarized .
we use a magnetic offset field of about which is in the plane of the lattice and perpendicular to the raman axis .
the optical pumping beam is tuned to the red of the @xmath131 line .
we typically operate with focused to a waist of about for the pump beam and focused to a waist of for each raman beam .
we estimate our rabi frequency to be @xmath132 by measuring power broadened raman spectra for different powers .
we perform raman imaging for and collect approximately @xmath133 photons per atom using a special optics custom objective with a working distance of @xmath134 mm . an achromatic doublet with @xmath135
gives a magnification of @xmath136 , which we verified using kapitza - dirac scattering of a molecular bec .
our objective is designed to be diffraction limited for and corrected for our thick fused silica vacuum window .
we image the atoms with a zyla 4.2 cmos camera ( andor technology ) with quantum efficiency of @xmath137% near .
* image analysis * we reconstruct the atom distribution in the lattice following the image reconstruction procedure described in @xcite ( fig .
[ fig : reconstr ] ) .
the reconstruction allows us to reduce the data from the fluorescence pictures to binary matrices that are the basis for all further data processing . our measured point spread function ( psf )
has a full - width half - max of , slightly larger than expected for our numerical aperture of 0.5 .
we estimate fidelity errors due to raman imaging imperfections by taking 40 consecutive images with our standard imaging time of of the same atom cloud and determine the shot - to - shot differences .
this leads to a hopping rate during one picture of and a loss rate of .
in addition , while holding the atoms in a deep lattice for spin manipulations and doublon hiding , we lose of the atoms , leading to a net detection efficiency of @xmath138 .
* extracting density and spin correlations * we are able to measure several different kinds of correlators with the techniques described previously . recall that we define the singles density of spin @xmath4 on lattice site @xmath139 by @xmath140 .
we also define @xmath141 . the first correlator we measure is the singles density correlator .
this is the natural quantity we measure with the microscope without any blowing pulses or doublon hiding . in this case , the doublons are removed by light - assisted collisions during the imaging .
the correlator is defined by , @xmath142 where @xmath143 . if we insert a resonant blowing pulse to eject one spin state and additionally perform doublon hiding , we measure singles density correlations of one spin state , @xmath144 this correlator is closely related to the spin correlator , but with additional contributions due to the presence of holes and doublons .
we define @xmath38 and @xmath148 analogously . in the experiment we can not distinguish between @xmath14 and @xmath149 spin correlations and
effectively measure @xmath150 . the fermi - hubbard model with an applied field along the @xmath145-direction retains in - plane rotational symmetry , so @xmath151 .
* systematic errors in measuring correlations * atoms that are lost from the trap after tunneling dynamics are frozen can reduce our measured correlations .
we estimate this effect by introducing the probability of an atom being lost before imaging , @xmath152 , and the probability @xmath153 that we image a single when it is not present on a site .
@xmath153 is a result of both hopping and recapture of atoms during imaging .
we find @xmath154 to first order in @xmath155 .
using our previous estimates for imaging loss rate , hopping rate , and loss during hold times , leads to an expected reduction of @xmath156 in all measured correlators .
imperfections in the spin imaging process can also effect correlators .
our blowing pulse off - resonantly ejects an atom in the spin state of interest with probability @xmath157 and fails to eject an atom in the other spin state with probability @xmath158 , leading to @xmath159 the errors considered here are due to the resonant blowing pulses and do not effect the value of @xmath160 .
however , they do effect the @xmath41 through the @xmath161 .
the reduction in these correlators is @xmath162 \\
c^z(\mathbf{d } ) - \tilde{c}^z(\mathbf{d } ) & = & \left(\epsilon_{f,\uparrow } + \epsilon_{f,\downarrow}\right ) c^z(\mathbf{d } ) + \sum_\sigma \left(\epsilon_{f\uparrow } + \epsilon_{f\downarrow } - 2 \epsilon_\sigma \right ) b_\sigma(\mathbf{d}),\end{aligned}\ ] ] where we have used the identity @xmath163 . for the spin correlator , the magnitude of the first term is at most @xmath164 , while the second two are at most @xmath165 .
imperfect doublon hiding also effects our correlators .
if we fail to hide a doublon with probability @xmath166 , we find @xmath167 nlce predicts the correlator on the right hand side of this equation is at most @xmath168 ( see supplement to @xcite ) .
this effect has magnitude at most @xmath169 . * the numerical linked cluster expansion * the numerical linked - cluster expansions ( nlces ) @xcite combine high - temperature series expansions ( htses ) with exact diagonalization techniques to calculate exact finite - temperature properties of quantum lattice models in an extended temperature range in the thermodynamic limit .
the formalism can be summarized in the following equation @xmath170 where @xmath171 , an extensive property per lattice site in the thermodynamic limit , is expressed in terms of contributions [ or weights , @xmath172 of all the clusters @xmath173 that can be embedded in the lattice and are not related by any translational symmetry .
the smallest cluster is a single site , the next smallest is a nearest - neighbor bond , and so on .
if the hamiltonian is also invariant under point group symmetries of the underlying lattice , the number of terms in eq .
[ eq : nlce ] can be reduced by considering only clusters that are topologically distinct and multiplying their weights by a multiplicity factor @xmath174 , which represents the number of ways per site cluster @xmath173 can be embedded on the lattice .
the weights @xmath175 are in turn calculated using the inclusion - exclusion principle ; writing eq .
[ eq : nlce ] for the finite cluster @xmath173 , we can express @xmath175 as @xmath176 where @xmath177 is the extensive property of interest on the finite cluster @xmath173 , and @xmath32 are all the clusters that can be embedded in @xmath173 ( sub - clusters of @xmath173 ) .
unlike in the htses , where @xmath177 are expressed as perturbative expansions in terms of powers of the inverse temperature , in the nlces , we obtain @xmath177 exactly ( to all powers in the inverse temperature ) using full diagonalization , leading to a better convergence .
we use the site expansion nlce in which the results in the @xmath178th order contain contributions from all clusters with @xmath178 sites or less .
we carry out the expansion to the 9th order . only in the last order , and for each of the two values of @xmath13 ( 8.0 , and 14.0 ) , we diagonalize the hubbard hamiltonian in the presence of a magnetic field for 112 topologically distinct 9-site clusters that have no particular symmetry . we choose a fine grid for the temperature and work in the grand canonical ensemble ; for @xmath179 at half filling , we fix the chemical potential and choose a fine grid for the magnetic field . similarly for @xmath180 away from half filling ,
we fix the magnetic field and choose a fine grid for the chemical potential .
these allow us to have a high resolutions for the polarization and the density , respectively @xcite .
the converged results are valid in the thermodynamic limit and contain no systematic or statistical errors .
however , due to the finite number of terms in our series the convergence is lost below a temperature that is dependent on @xmath13 , the polarization and the density . to extend the convergence to lower temperatures , we take advantage of numerical re - summation techniques . to minimize the possibility of introducing systematic errors
, we use two vastly different algorithms , namely , wynn resummation with four cycles of improvement and the euler resummation for the last six terms in the series @xcite , and present results when they agree within a few percent . in the nlce
, one can in principle compute correlation functions at distances as far as the largest clusters in the series extend .
however , as this distance increases , a smaller number of clusters contribute to the series , and so , one should expect the lowest convergence temperature to increase with distance . in fig .
[ fig : spincorr ] , we show spin correlations from the nlce at @xmath179 for distances as large as @xmath181 . we find a very good qualitative agreement with the experimental results shown in fig .
3 of the main text .
the nlce results are at a fixed @xmath182 , so we do not consider the temperature variations that may exist from one polarization to the next in the experimental data .
* determinantal quantum monte carlo * the determinantal or auxilliary field quantum monte carlo ( qmc ) for interacting fermions is an unbiased algorithm that provides statistically exact " answers for the energy and correlation functions as a function of temperature .
we start with the hubbard model in terms of fermion operators @xmath183 where the single particle part of the hamiltonian is here , @xmath1 is the hopping amplitude between sites , @xmath187 is the chemical potential that controls the density of fermions and @xmath188 is the zeeman field that controls the imbalance between majority up and minority down spin fermions .
the experiments are performed in the canonical ensemble for a fixed total number of fermions @xmath189 and a fixed magnetization @xmath190 whereas the simulations are done in the grand canonical ensemble for a fixed @xmath187 and @xmath188 . for a bipartite lattice , the repulsive hubbard model maps onto an attractive hubbard model through a particle - hole transformation on the down spins and an additional @xmath191 phase shift on the b sublattice given by @xmath192 , @xmath193 . with this mapping
the chemical potential and zeeman field get interchanged @xmath194 and @xmath195 in the repulsive and attractive hubbard models .
we will return to this mapping when we discuss the nature of the sign problem toward the end of the section . with the hamiltonian defined above , we next divide up the imaginary time interval @xmath196 $ ] into @xmath197 segments of width @xmath198 .
a trotter break up of the two non - commuting operators @xmath80 and @xmath199 can now be done on each time - slice with errors on the order of @xmath200 , which gives @xmath201^m\approx [ e^{-\delta \tau t } e^{-\delta \tau v}]^m $ ] . by introducing auxiliary fields at each space - time point @xmath202 in a path integral representation of the partition function in imaginary time ,
typically we use ising auxiliary fields @xmath203 at each space - time point , the quartic interaction terms can be factored into quadratic forms , given by @xmath204 with @xmath205 determined by @xmath206 and @xmath207 . the path integral of the interacting problem is thus converted to an effectively non - interacting action that describes the motion of a fermion propagating through fluctuating auxiliary field configurations , given by : @xmath208 & = & -t\sum_{\langle i , j\rangle\sigma } [ c_{i\sigma}^\dagger c_{j \sigma } + h.c . ] \\
\nonumber & + & \sum_{i \sigma } [ -(\mu + u/2+\sigma h ) + \sigma \lambda s(i,\tau ) ] n_{i \sigma}.\end{aligned}\ ] ] the fermions can now be integrated out yielding a partition function @xmath209 as a sum of determinants over auxiliary field configurations that are sampled using monte carlo methods , where @xmath210 $ ] .
the negative sign or complex phase of this determinant is the source of the so - called sign - problem " in qmc simulations that can limit the ability to go to low temperatures away from half filling .
( teal and black ) , @xmath211 ( purple ) , @xmath212 ( blue ) , and @xmath213 ( brown ) , open symbols for balanced populations and closed symbols for @xmath214 , triangles for @xmath215 and squares for @xmath16 . ] in fig .
[ fig : sign ] we show the average sign as a function of density at various temperatures for zero field and a finite field . the important points to note are : ( i ) the magnetic field does not generate a sign problem
since it essentially maps to @xmath187 in the attractive hubbard model for which the up and down determinants have the same sign , hence the product is always positive , for any density .
( ii ) at half filling the average sign is unity , hence no sign problem , for any temperature and field .
( iii ) for @xmath16 ( fig .
[ fig : sign ] , black squares ) where most experimental data were taken , the sign problem is not severe and does not affect the quality of the data at the lowest temperatures considered .
( iii ) away from half filling the sign is most severe around @xmath216 filling for @xmath215 ; the average sign nose dives with decreasing temperature . within this density region , for @xmath217 increased statistics is used to improve the quality of the data and we were able compare our qmc with the experimental data for the density and the polarization .
for even smaller temperatures ( fig .
[ fig : sign ] , down triangles ) we have not used qmc data to compare with experiments .
the error bars reported in the figures are statistical .
the qmc simulations were performed for @xmath218 lattices with @xmath219 for @xmath16 and @xmath220 for @xmath215 .
each data point was obtained from up to 60 independent runs , each with 4000 sweeps through the lattice . * sign change of spin @xmath145-correlation for large polarizations * to understand why @xmath57 becomes negative near @xmath221 at half filling , it is instructive to consider what happens in the heisenberg model with a nearly fully polarized gas .
we can regard the fully polarized gas as a vacuum state and the minority spins as a dilute gas of magnons @xcite .
these magnons are bosons and can form a bose einstein condensate ( bec ) at low temperatures .
we will find that the bec off - diagonal order is associated with the spin correlations perpendicular to the magnetization , and the density correlation of the magnons are associated with the spin correlations parallel to the magnetization . to make this argument more concrete , we rewrite the heisenberg hamiltonian using operators defined by @xmath222 , @xmath223 and @xmath224 , @xmath225 the @xmath226 and @xmath227 satisfy bosonic commutation relations for @xmath228 . to avoid difficulty at @xmath229 we introduce an infinite on - site repulsion , or hard - core constraint , and
can then regard @xmath230 as the creation operator for a boson on site @xmath231 @xcite . from the middle term in in eq .
[ eq : magnonbec ] , we see the field @xmath188 is the chemical potential of the magnons ( up to a constant offset ) .
the last term describes nearest neighbor repulsion of magnons .
the first term describes the bosons hopping . transforming to momentum space ,
this term becomes @xmath232 with @xmath233 where @xmath234 is the lattice constant .
in contrast to the typical case , the hopping term here is positive and the condensate forms at the band minimum , @xmath235 .
condensation is signalled by non - zero expectation value of off - diagonal density matrix elements .
for system of volume @xmath199 and total particle number @xmath236 these elements are @xmath237 using the identity @xmath238 , we see the bec off - diagonal order is associated with spin correlations perpendicular to the magnetization , and the anti - ferromagnetic checkerboard is associated with the condensate quasimomentum , @xmath239 .
the spin correlations parallel to the magnetization are associated with the density correlations of the magnons , which can be written @xmath240 .
these are the density correlations of a liquid of bosons on the lattice , with hard core exclusion and nearest - neighbor repulsion .
these are expected to be typical liquid correlations , starting negative at short distances and oscillating and damping with distance , first becoming positive at a distance of order the inter - particle spacing .
as we go to high polarization and thus low density of these bosons , this inter - particle distance grows and moves beyond the diagonal neighbor distance , resulting in the sign change in @xmath57 . 14 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevlett.117.093601 [ * * , ( ) ] link:\doibase 10.1103/physreva.73.033605 [ * * , ( ) ] link:\doibase 10.1103/physreva.71.050701 [ * * , ( ) ] link:\doibase 10.1126/science.aag1430 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.263001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.213002 [ * * , ( ) ] link:\doibase 10.1038/nature09378 [ * * , ( ) ] link:\doibase 10.1126/science.aag3349 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.187202 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.cpc.2012.10.008 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.053611 [ * * , ( ) ] link:\doibase 10.1103/physreve.75.061119 [ * * , ( ) ] http://www.jetp.ac.ru/cgi-bin/dn/e_060_04_0781.pdf [ * * , ( ) ] link:\doibase 10.1103/revmodphys.86.563 [ * * , ( ) ] | understanding the magnetic response of the normal state of the cuprates is considered a key piece in solving the puzzle of their high - temperature superconductivity @xcite .
the essential physics of these materials is believed to be captured by the fermi - hubbard model @xcite , a minimal model that has been realized with cold atoms in optical lattices @xcite . here
we report on site - resolved measurements of the fermi - hubbard model in a spin - imbalanced atomic gas , allowing us to explore the response of the system to large effective magnetic fields .
we observe short - range canted antiferromagnetism at half - filling with stronger spin correlations in the direction orthogonal to the magnetization , in contrast with the spin - balanced case where identical correlations are measured for any projection of the pseudospin .
the rotational anisotropy of the spin correlators is found to increase with polarization and with distance between the spins .
away from half - filling , the polarization of the gas exhibits non - monotonic behavior with doping for strong interactions , resembling the behavior of the magnetic susceptibility in the cuprates @xcite .
we compare our measurements to predictions from determinantal quantum monte carlo ( dqmc ) @xcite and numerical linked cluster expansion ( nlce ) @xcite algorithms and find good agreement .
calculations on the doped system are near the limits of these techniques , illustrating the value of cold atom quantum simulations for studying strongly - correlated materials .
ultracold quantum gases have emerged as a powerful tool to study strongly correlated many - body physics . a two - component fermi gas in an optical lattice
can realize the repulsive hubbard model , which describes fermions in a periodic potential with onsite interaction @xmath0 and tunneling matrix element @xmath1 between neighboring sites @xcite .
the recent introduction of quantum gas microscopes for fermionic atoms @xcite has led to rapid development in the experimental study of the 2d hubbard model .
the number - squeezed nature of the mott insulating phase previously inferred from bulk measurements @xcite has been explicitly revealed .
furthermore , site - resolved measurements probe antiferromagnetic correlations beyond the nearest neighbor @xcite , which was not possible in previous studies @xcite . in this work ,
we investigate the fermi - hubbard model with imbalanced spin populations described by the hamiltonian @xmath2 here @xmath3 is the creation operator for a fermion with spin @xmath4 on site @xmath5 and @xmath6 . theoretical studies of spin - imbalance in the hubbard model have predicted an interesting magnetic structure in trapped gases arising from the interplay of spin - imbalance and antiferromagnetic and stoner instabilities @xcite .
experimentally , the polarization of our two - component atomic fermi gas is a controllable quantity that is conserved due to the absence of spin - relaxation mechanisms .
thermodynamically , a non - zero polarization corresponds to the introduction of an effective zeeman field @xmath7 , where @xmath8 are the chemical potentials of the two components . a starting point for understanding the low temperature phase diagram at half - filling and strong interactions
is the heisenberg hamiltonian @xmath9 which is a good approximation for the hubbard model in this regime . here
, @xmath10 is the vector spin operator , @xmath11 is its projection along the direction of the magnetization and @xmath12 is the superexchange coupling . at zero field ,
the su(2 ) symmetry of the hamiltonian leads to isotropic antiferromagnetic correlations that decay exponentially in 2d with a correlation length that diverges at zero temperature . for a non - zero field
, the heisenberg model has a finite temperature kosterlitz - thouless transition to a canted antiferromagnet @xcite ( fig .
[ fig:1]a ) , which accommodates magnetization along the field while still benefiting from the superexchange interactions by building up long - range antiferromagnetic correlations of the spin components perpendicular to the magnetization .
while the heisenberg model provides useful insight in this regime , with finite @xmath13 or with doping it is necessary to consider the full hubbard model .
( yellow ) and @xmath14 ( blue ) spin projections versus local polarization @xmath15 at half - filling .
we show nlce ( dashed lines ) and dqmc ( small circles ) results at @xmath16 corrected for our detection efficiency of 0.96 ( see methods ) and uncorrected nlce results ( solid lines ) .
we display a temperature band from @xmath17 to @xmath18 . *
b * , anisotropy @xmath19 of nearest - neighbor ( large green circles ) and diagonal neighbor ( red squares ) spin correlations with nlce ( solid lines ) and dqmc ( small circles ) results .
@xmath19 is insensitive to detection efficiency .
error bars are s.e.m .
experimental data averaged over @xmath20 images and azimuthally . ]
we realize the two - dimensional fermi hubbard model using a degenerate mixture of two hyperfine ground states @xmath21 and @xmath22 of @xmath23li in an optical lattice .
the global spin imbalance @xmath24 can be varied continuously from 0 to @xmath25 by evaporating the gas in the presence of a magnetic gradient leading to preferential loss of one of the spin states .
we work at a scattering length of 448(9 ) @xmath26 , where @xmath26 is the bohr radius , obtained by adjusting a bias magnetic field in the vicinity of a broad feshbach resonance centered at 690 g. the imbalanced mixture is prepared in a single layer ( for details see @xcite ) and subsequently loaded adiabatically into a 2d square lattice of depth @xmath27 , where @xmath28 hz . here , @xmath29
khz is the recoil energy . for these parameters ,
we obtain @xmath30 , where strong antiferromagnetic correlations are expected at half - filling in the balanced gas .
fluorescence images obtained with quantum gas microscopy techniques enable us to identify singly occupied sites in the lattice , regardless of the spin state ( see methods ) .
doubly occupied sites undergo light assisted collisions and appear empty .
we can also identify singly occupied sites where the atoms are projected onto a chosen eigenstate of @xmath31 by illuminating the cloud with a short pulse of resonant light that ejects atoms in the other eigenstate ( fig .
[ fig:1]b , c ) . by first converting atoms on doubly occupied sites to deeply bound molecules
, we ensure that they are not affected by this light pulse and they are subsequently lost in light - assisted collisions during imaging . from these images
we extract the azimuthally - averaged density of atoms on singly - occupied ( @xmath32 ) sites in a particular spin ( @xmath4 ) state @xmath33 and the total density @xmath34 , shown in fig .
[ fig:1]d for an imbalanced gas .
we observe a plateau in @xmath35 over an extended region near the center of the trap , indicating the formation of a spin - imbalanced mott - insulating state .
the deviation of @xmath35 from unity within the mott insulator is primarily due to doublon - hole quantum fluctuations , which are non - negligible at our interaction strength .
we characterize the local polarization in terms of the measured quantities @xmath36 .
this definition coincides with the true polarization in the absence of doubly occupied sites . at the accessible temperatures , the local polarization is constant throughout the mott insulator .
we measure the spin correlators in the mott insulator both for the spin components along the field , @xmath37 , and orthogonal to it , @xmath38 , by extending techniques previously introduced in experiments on balanced gases @xcite .
the spin correlator at a displacement @xmath39 between two sites is given by @xmath40 , where the brackets denote an ensemble and azimuthal average .
@xmath41 is obtained from the singles density correlators @xmath42 and @xmath43 taking into account the effect of doublons and holes ( see methods ) .
to extract @xmath44 , we insert an additional radiofrequency pulse to rotate the spins by @xmath45 before initiating the measurement protocol . the measured nearest and next - nearest neighbor correlators @xmath46 are shown versus @xmath15 in fig .
[ fig:2]a . as the polarization is increased , we observe an overall decrease in correlations , with @xmath37 decreasing faster than @xmath38 .
we define the correlator anisotropy as @xmath47 . for an almost unpolarized gas @xmath48 we measure @xmath49 for the nearest neighbor and @xmath50 for the next - nearest neighbor .
the consistency of these values with zero verifies the su(2 ) symmetry of the hubbard hamiltonian at @xmath51 .
the anisotropies are shown versus @xmath15 in fig .
[ fig:2]b .
the system s preference to build correlations in the plane orthogonal to the field can already be understood at the level of a classical heisenberg model , since spins oriented with the staggered magnetization in @xmath52 the plane can lower their energy by uniformly canting in the direction of the field . in the quantum system , strong quantum fluctuations in 2d reduce the magnitude of the nearest - neighbor correlator in the balanced gas from 1 to 0.36 in the ground state @xcite , and thermal fluctuations and imaging fidelity further reduce it to the experimentally measured value of @xmath53 . at non - zero polarization
, we observe that the correlator anisotropy is stronger when the sites are further apart .
for example , at @xmath54 , @xmath55 for the nearest neighbor , while @xmath56 for the next - nearest neighbor .
the increase of the correlation anisotropy with distance can be partly understood by considering what happens at lower temperatures as we approach the kosterlitz - thouless transition . there
the @xmath38 correlations become long range while the @xmath37 do not , so at long distance @xmath19 approaches one .
insight into the range of the antiferromagnetic order can be gained by examining 2d plots of the spin correlators as a function of the displacement vector between the sites , shown in fig .
[ fig:3 ] .
the checkerboard pattern is visible for displacements of up to four sites along the diagonal in the almost unpolarized gas , and the overall decrease of all correlations with polarization , as well as the relative suppression of @xmath37 relative to @xmath38 is evident .
the @xmath38 correlations remain antiferromagnetic at all polarizations , but the @xmath37 correlations can be viewed as the density correlations of the gas of minority spins @xcite whose modulation becomes longer wavelength as the density of this gas decreases .
this leads to a change in the sign of @xmath57 near @xmath58 which can be seen in figs .
[ fig:2]a and [ fig:3 ] .
the observation of this percent - level negative correlation is only possible because of the superb sensitivity of quantum gas microscopy .
we compare our measurements to results from dqmc and nlce simulations of the fermi - hubbard model @xcite at half - filling in the presence of a chemical potential imbalance with the temperature left as a free parameter . for the balanced gas ,
the measured nearest neighbor correlators give a temperature of @xmath59 .
the temperature shows a mild increase with polarization , rising to @xmath60 at @xmath61 .
the calculated anisotropy is almost independent of temperature over this range , and shows excellent agreement with the experiment ( fig .
[ fig:2]b ) . .
we show nlce ( dashed lines ) and dqmc ( small circles ) results corrected for our detection efficiency , and uncorrected nlce results ( solid lines ) .
we show a temperature band from @xmath17 to @xmath18 , which is determined from spin correlators .
* inset * , singles density correlations versus density for @xmath62 ( yellow ) , @xmath63 ( purple ) , and @xmath64 ( red ) .
lines are guides to the eye .
error bars are s.e.m .
experimental data averaged over @xmath20 images and azimuthally . ] in addition to spin correlations , our data gives access to density correlations between singly occupied sites @xmath65 ( fig .
[ fig:4 ] ) .
previous analysis of this correlator in balanced gases revealed a dominant positive contribution at half - filling from doublon - hole virtual excitations and a smaller negative contribution from hole - hole correlations due to pauli repulsion @xcite . for increasing polarization , singles density correlations decrease as pauli blocking suppresses double occupancy in the gas .
the results of nlce and dqmc calculations at @xmath16 show good agreement with the data using the same temperature range extracted from the spin correlations . in the inset of fig .
[ fig:4 ] , we show the singles density correlation versus @xmath66 for different polarizations . as was observed in the balanced case , the correlator changes sign as the filling is reduced , an effect that has been attributed to pauli repulsion in the metallic regime @xcite .
this repulsion becomes more pronounced as the polarization is increased , leading to negative correlations over a wider range of fillings .
the polarization profile of the imbalanced gas in the trap gives insight into the spin susceptibility of the hubbard model in the doped regime . for strong interactions
we observe that the in - trap polarization profile can exhibit non - monotonic behavior as shown in fig .
[ fig:5]a for @xmath67 obtained by increasing the scattering length to 793(12 ) @xmath26 . for these experiments ,
we extract the true polarization @xmath68 , rather than @xmath15 ( see methods ) .
the local polarization shows a shallow rise near the edge of the mott insulator , then drops in the metallic region , before rising rapidly at the edge of the cloud .
these effects can be understood qualitatively in terms of the magnetic susceptibilities of the gas at different fillings . at half - filling ( @xmath69 )
, the susceptibility is expected to be that of an antiferromagnet @xmath70 , while for small doping at our temperatures there is a nondegenerate gas of holes in the lower hubbard band and as a result a weak maximum in the susceptibility . at intermediate hole doping
, the susceptibility crosses over to that of a metal , @xmath71 , which is smaller than @xmath72 for large @xmath13 .
similar behavior has been observed in the cuprates in the normal phase @xcite and studied theoretically @xcite . at even lower filling , @xmath73 and
there is no filled fermi sea to hinder spins from aligning with the effective field , leading to an enhanced magnetic susceptibility .
we show the polarization versus density in fig .
[ fig:5]b . and
compare to nlce and dqmc calculations in the local density approximation .
the data is at a constant effective field since the trapped gas is in chemical equilibrium .
the strength of the field , @xmath74 , is extracted from the experimental polarization at half - filling , which exhibits only a weak dependence on temperatures for @xmath75 .
these calculations reproduce the non - monotonic behavior of the polarization .
we have presented the first study of a 2d spin - imbalanced fermi - hubbard system in a regime near the edge of what state - of - the - art numerical techniques can simulate .
our quantum simulations qualitatively reproduce the non - monotonic behavior of the magnetic susceptibility with doping in the cuprates .
on the other hand , the high effective fields reached in our experiments allow us to explore canted antiferromagnetism , physics that is inaccessible in the cuprates at laboratory magnetic fields .
interesting future directions for both experimental and theoretical work include investigation of spin - imbalance in the attractive 2d hubbard model where fulde - ferrell - larkin - ovchinnikov superfluid correlations should be detectable at the entropies achieved in repulsive experiments @xcite and study of the kosterlitz - thouless transition in the imbalanced repulsive gas @xcite which would require lower temperatures .
finally , the achievement of cold spin - imbalanced clouds in an optical lattice suggests a new route for local entropy reduction using adiabatic demagnetization cooling , a technique previously demonstrated in bosonic lattice experiments @xcite . [ [ acknowledgements ] ] acknowledgements + + + + + + + + + + + + + + + + this work was supported by the nsf ( grant no .
dmr-1607277 ) , the david and lucile packard foundation ( grant no .
2016 - 65128 ) , and the afosr young investigator research program ( grant no . fa9550 - 16 - 1 - 0269 ) .
w.s.b . was supported by an alfred p. sloan foundation fellowship .
p.t.b . was supported by the dod through the ndseg fellowship program .
e.k . was supported by the nsf ( grant no .
dmr-1609560 ) .
t.p . was supported by cnpq , faperj , and inct on quantum information . n.t .
acknowledges funding from the nsf ( grant no .
dmr-1309461 ) .
we thank m. rigol for providing exact diagonalization results for the imbalanced system at @xmath76 to benchmark the nlce . [ [ author - contributions ] ] author contributions + + + + + + + + + + + + + + + + + + + + p.t.b .
, d.m . ,
e.g .- s .
, p.s . , and s.k .
performed the experiment and data analysis .
e.k . performed the nlce calculations .
t.p .
and n.t . performed the dqmc calculations .
w.s.b . and d.a.h .
provided scientific guidance in experimental and theoretical questions .
all authors contributed to the interpretation of the data and the writing of the manuscript .
35ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/revmodphys.78.17 [ * * , ( ) ] link:\doibase 10.1126/science.235.4793.1196 [ * * , ( ) ] link:\doibase 10.1038/nature07244 [ * * , ( ) ] link:\doibase 10.1126/science.1165449 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.66.763 [ * * , ( ) ] link:\doibase 10.1103/physrevd.24.2278 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.187202 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-070909 - 104059 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.193001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.213002 [ * * , ( ) ] link:\doibase 10.1038/nphys3403 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.063406 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.263001 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/18/2/023016 [ * * , ( ) ] link:\doibase 10.1126/science.aad9041 [ * * , ( ) ] link:\doibase 10.1126/science.aag1430 [ * * , ( ) ] link:\doibase 10.1126/science.aag1635 [ * * , ( ) ] link:\doibase 10.1126/science.aag3349 [ * * , ( ) ] link:\doibase 10.1126/science.1236362 [ * * , ( ) ] link:\doibase 10.1038/nature14223 [ * * , ( ) ] http://arxiv.org/abs/1607.00392 [ ( ) ] link:\doibase 10.1103/physreva.81.023628 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.013616 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.054419 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.117.093601 [ * * , ( ) ] http://www.jetp.ac.ru/cgi-bin/dn/e_060_04_0781.pdf [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.066406 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.053611 [ * * , ( ) ] link:\doibase 10.1103/physrevb.40.8872 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.62.957 [ * * , ( ) ] link:\doibase 10.1103/physrevb.40.2254 [ * * , ( ) ] link:\doibase 10.1063/1.352636 [ * * , ( ) ] link:\doibase 10.1103/physrevb.48.3380 [ * * , ( ) ] link:\doibase 10.1103/physrevb.94.075157 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.195301 [ * * , ( ) ] |
in @xcite , we discussed the measurement of luminosity distances of events using the advanced and virgo ground - based interferometric gravitational - wave ( ) detectors . in the main letter ,
an algorithm was introduced for rapidly extracting directionally dependent distance estimates from observations and illustrated the typical shape of volume reconstructions during early advanced .
finally , we argued that the structure and distance information can be leveraged to guide searches of likely nearby host galaxies for x - ray , optical , and infrared counterparts of mergers .
this supplement provides the following supporting material .
first , in [ sec : format ] , we document a file format for the rapid transmission of volume reconstructions in alerts . it is based on and is backward - compatible with the localization formation that we introduced in @xcite and that was employed in alerts that were sent in (;
second , in
[ sec : data ] , we describe the online data release , which provides a browsable collection of simulated sky maps .
third , in [ sec : tutorial ] , we provide a python primer for performing basic operations on sky maps , all the way through selecting a list of the most likely host galaxies . in [ sec : algorithm ] , we provide additional details of the position reconstruction algorithm . finally , in [ sec : faithfulness ] , we show that the algorithm produces faithful representations of the full probability distributions .
the reader who is interested in leveraging distance information for planning follow - up observations or performing archival research needs only consult [ sec : format ] and [ sec : tutorial ] .
the localization for a single candidate is stored as a (;
@xcite ) file .
the file contains a single binary table @xcite that represents a (;
@xcite ) all - sky image .
the table has four floating - point columns , listed in table [ table : fitscolumns ] , which represent four channels of the image .
the first column , ` prob ` , is simply the probability that the source is contained within the pixel @xmath0 that is centered on the direction @xmath1 , the same as in the localization format . the second and third columns , ` distmu ` and `
diststd ` , are the ansatz location and scale parameters , respectively . the fourth column , `
distnorm ` , is the ansatz normalization coefficient , included for convenience . in pixels on the sky that contain very little probability ,
sometimes the conditional distance distribution can not be represented using the ansatz .
this is signaled by ` distmu`=@xmath2 , ` distsigma`=@xmath3 , and ` distnorm`=0 .
> llll prob & @xmath4 & pixel@xmath5 & probability that the source is contained in pixel @xmath0 , centered on the direction @xmath1 + distmu & @xmath6 & mpc & ansatz location parameter of conditional distance distribution in direction @xmath1 , or @xmath2 if invalid + distsigma & @xmath7 & mpc & ansatz scale parameter of conditional distance distribution in direction @xmath1 , or @xmath3 if invalid + distnorm & @xmath8 & mpc@xmath9 & ansatz normalization coefficient , or @xmath10 if invalid the header , an example of which is shown in table [ table : fitsheader ] , provides metadata including the utc time of the trigger and the list of instruments that contributed to the localization .
the header also provides values for ` distmean ` and ` diststd ` , respectively , being the posterior mean and standard deviation of distance marginalized over the whole sky .
> l > l > l simple & t & conforms to fits standard + bitpix & 8 & array data type + naxis & 0 & number of array dimensions + extend & t + xtension & bintable & binary table extension + bitpix & 8 & array data type + naxis & 2 & number of array dimensions + naxis1 & 16384 & length of dimension 1 + naxis2 & 3072 & length of dimension 2 + pcount & 0 & number of group parameters + gcount & 1 & number of groups + tfields & 4 & number of table fields + ttype1 & prob + tform1 & 1024e + tunit1 & pix-1 + ttype2 & distmu + tform2 & 1024e + tunit2 & mpc + ttype3 & distsigma + tform3 & 1024e
+ tunit3 & mpc + ttype4 & distnorm + tform4 & 1024e + tunit4 & mpc-2
+ pixtype & healpix & healpix pixelisation + ordering & nested & pixel ordering scheme , either ring or nested + coordsys & c & ecliptic , galactic or celestial ( equatorial ) + extname & xtension & name of this binary table extension + nside & 512 & resolution parameter of healpix + firstpix & 0 & first pixel # ( 0 based ) + lastpix & 3145727 & last pixel # ( 0 based ) + indxschm & implicit & indexing : implicit or explicit + object & coinc_event : coinc_event_id:18951 & unique identifier for this event + instrume & h1,l1 & instruments that triggered this event + date - obs & 2010 - 09 - 03t06:12:26.60324 & utc date of the observation + mjd - obs & 55442.2586412414 & modified julian date of the observation + date & 2015 - 04 - 13t10:17:11 & utc date of file creation + creator & bayestar_localize_coincs.py & program that created this file + distmean & 68.54061620909769 & posterior mean distance in mpc + diststd & 17.14006463067744 & posterior standard deviation of distance in mpc +
an online data release provides a browsable catalog of simulated localizations .
one may select events from or .
events may be sorted by detector network ( a one- or two - letter combination consisting of ` h ' for , ` l ' for , ` v ' for virgo ) , 90% credible volume in mpc@xmath11 , 90% credible area in deg@xmath12 , or . for each event
, a or lalinference file may be downloaded .
a screen shot of the data release is shown in fig . [
fig : screenshot ] .
in this section , we provide some python sample code to perform some simple manipulations of sky maps .
the triple greater - than signs ( ` > > > ` ) and triple - dots ( ` ... ` ) are the python interactive prompt ; the reader should type everything on the line _
after _ these .
these examples will work in python 2.7 and later on linux or unix systems .
if the reader does not already have a python environment of preference , we suggest the anaconda python distribution for desktop use or the lightweight miniconda variant for computing clusters . only the astropy , healpy , and numpy packages are essential for working with the localizations , but the examples below will also use matplotlib , scipy , and astroquery .
all of these packages can be installed with pip : .... $ pip install astropy astroquery healpy matplotlib scipy .... for all of the samples below , start by importing the healpy for working with files , the numpy for vector operations , matplotlib for plotting , and scipy for probability functions : .... $ python > > > import healpy as hp > > > import numpy as np > > > from matplotlib import pyplot as plt >
> > from scipy.stats import norm .... next , select download an example sky map from the data release . in this example , we use the simulated event that is shown in figs . 1 and 2 of @xcite
. a convenient way to download it is using astropy s ` download_file ` utility , which will retrieve the file and cache it locally : .... > > > from astropy.utils.data import download_file > > > url = ( ' https://dcc.ligo.org/p1500071/public ' ... + ' /18951_bayestar.fits.gz ' ) > > > filename = download_file(url , cache = true ) .... the new localization format is backward - compatible with the format introduced in @xcite . by default , when we read the healpix file with healpy ( or any other common - place healpix library or tool ) , we get just the first layer , the probability sky map : .... > > > prob = hp.read_map(filename ) .... to read both the probability layer and the three additional distance layers , we need to pass the optional ` field= ` parameter to healpy : .... > > > prob , distmu , distsigma , distnorm = hp.read_map ( ... filename , field=[0 , 1 , 2 , 3 ] ) .... or slightly more concisely : .... > > > prob , distmu , distsigma , distnorm = hp.read_map ( ... filename , field = range(4 ) ) .... last , it will be useful for subsequent healpy calls to have the resolution on hand : .... > > > npix = len(prob ) > > > npix 3145728 > >
> nside = hp.npix2nside(npix ) > > > nside 512 .... in this example , we compute the probability per steradian or per deg@xmath12 that the source is in a given direction .
let s take as an example the following equatorial coordinates : .... > > >
ra , dec = 137.8 , -39.9 .... which , coincidentally , happen to be the true simulated position to the source .
healpy uses
physicist s " spherical coordinates ( @xmath13 ) , with @xmath14 $ ] being the colatitude from the north celestial pole in radians , and @xmath15 being the right ascension in radians .
we convert ....
> > > theta = 0.5 * np.pi - np.deg2rad(dec ) > >
> phi = np.deg2rad(ra ) ....
next , we use healpy to look up the index of the pixel that contains that direction : .... > > > ipix = hp.ang2pix(nside , theta , phi ) > > > ipix 2582288 .... healpy will tell us the area per pixel in steradians at the current resolution : .... > > > pixarea = hp.nside2pixarea(nside ) > > > pixarea 3.994741635118857e-06 .... or in deg@xmath12 : .... > > >
pixarea_deg2 = hp.nside2pixarea(nside , degrees = true ) > > > pixarea_deg2 0.013113963206424481 .... all that is left to do is look up the probability contained within pixel ` ipix ` and ( if desired ) divide by the area per pixel to obtain the probability per steradian : .... > > > dp_da = prob[ipix ] / pixarea > > > dp_da 7.4387317043042076 .... or the probability per deg@xmath12 : .... > > >
dp_da_deg2 = prob[ipix ] / pixarea_deg2 > > > dp_da_deg2 0.0022659672582507331 .... next , we calculate the conditional distance distribution along a given line of sight , which is the probability per unit distance under the assumption that the source is in a given direction .
we will use the same sky position as in the example above .
we lay out a grid in distance along that line of sight : .... > > > r = np.linspace(0 , 150 ) .... then , we plug everything into the ansatz distribution : .... > > >
dp_dr = r**2 * distnorm[ipix ] * norm ( ... distmu[ipix ] , distsigma[ipix]).pdf(r ) .... finally , we plot the result : .... > > > plt.plot(r , dp_dr ) > > > plt.xlabel('distance ( mpc ) ' ) > > > plt.ylabel('prob mpc$^{-1}$ ' ) > > > plt.show ( ) .... now , we calculate the probability density per mpc@xmath11 at a point .
we will use the same right ascension and declination as above and a distance of 74.8mpc : .... > > > r = 74.8 .... finally , .... > > >
dp_dv = prob[ipix ] * distnorm[ipix ] * norm ( ... distmu[ipix ] , distsigma[ipix]).pdf(r ) / pixarea > > > dp_dv 3.1173200109121657e-05 .... as our next example , we compute the marginal distance distribution , the probability density per unit distance integrated over the entire sky : .... > > > r = np.linspace(0 , 150 ) > > > dp_dr = [ np.sum(prob * rr**2 * distnorm ... * norm(distmu , distsigma).pdf(rr ) ) for rr in r ] .... finally , we plot the result : .... > > > plt.plot(r , dp_dr ) >
> > plt.xlabel('distance ( mpc ) ' ) >
> > plt.ylabel('prob mpc$^{-1}$ ' ) > > > plt.show ( ) .... as our final example , we will generate a ranked list of galaxies . for the purpose of this demonstration
, we will use the (; @xcite ) because it is a flux - limited all - sky spectroscopic redshift catalog .
this greatly simplifies the issues of completeness , sky coverage , and accuracy of redshift estimates .
first , download the entire catalog from vizier @xcite using astroquery : .... > > > from astroquery.vizier import vizier > > > vizier.row_limit = -1 > > > cat , = vizier.get_catalogs('j/apjs/199/26/table3 ' ) .... according to @xcite , the luminosity function is well fit by a schechter function with a cutoff absolute magnitude of @xmath16 and a power - law index of @xmath17 .
we find the maximum absolute magnitude @xmath18 for a completeness fraction of 0.5 : .... > >
> from scipy.special import gammaincinv > > > completeness = 0.5 > > > alpha = -1.0 >
> > mk_star = -23.55 > > > mk_max = mk_star + 2.5 * np.log10 ( ... gammaincinv(alpha + 2 , completeness ) ) > >
-23.947936347387156 .... we select only galaxies with positive redshifts and absolute magnitudes greater than @xmath18 : .... > >
> from astropy.cosmology import wmap9 as cosmo > > > from astropy.table import column >
> > import astropy.units as u > > > import astropy.constants as c > > > z = ( u.quantity(cat['cz ' ] ) / c.c).to ( ...
u.dimensionless_unscaled ) > > > mk = cat['ktmag ' ] - cosmo.distmod(z ) > > > keep = ( z > 0 ) & ( mk < mk_max ) > > > cat = cat[keep ] > > > z = z[keep ] .... then , we calculate the luminosity distance and index of each galaxy : .... > > > r = cosmo.luminosity_distance(z).to('mpc').value > >
> theta = 0.5 * np.pi - cat['_dej2000'].to('rad').value > > >
phi = cat['_raj2000'].to('rad').value > > > ipix = hp.ang2pix(nside , theta , phi ) .... we find the probability density per unit volume at the position of each galaxy : .... > > >
dp_dv = prob[ipix ] * distnorm[ipix ] * norm ( ... distmu[ipix ] , distsigma[ipix]).pdf(r ) / pixarea .... finally , we sort the galaxies by descending probability density and take the top 50 : .... > > >
top50 = cat[np.flipud(np.argsort(dp_dv))][:50 ] > > > top50['_raj2000 ' , ' _ dej2000 ' , ' ktmag ' ] < table masked = true length=50 > _
raj2000 _
dej2000 ktmag deg deg mag float64 float64 float32 --------- --------- ------- 344.01190 36.36136 8.772 343.81122 36.67177 9.958 137.19089 -38.60788 9.566 334.86545 29.39581 9.835 359.81589 46.88923 9.307
0.00695 47.27456 9.499 ... ... ... 123.16494 -16.05073 9.727 341.26642 33.99616 9.799 339.33075 34.44790 9.204 137.27219 -35.90259 10.822 188.13953 -68.53886 9.609 339.01483 33.97575 10.032 ....
the volume reconstruction algorithm consists of a computationally trivial postprocessing stage that is added to the two established ligo / virgo methods for localization of events , the rapid triangulation code @xcite , and the lalinference parameter estimation pipeline @xcite .
the conditional mean and standard deviation of distance are extracted from as described in
[ sec : bayestar ] and from lalinference as explained in
[ sec : lalinference ] below .
then , the mean and standard deviation are converted to the ansatz parameters as described in [ sec : method - of - moments ] .
@xcite is a rapid position reconstruction algorithm for mergers .
its inputs are a trio of numbers for each detector : the matched - filter estimates of the arrival time , phase , and amplitude at each site .
it marginalizes over polarization angle , inclination angle , coalescence time , and distance by performing low - order gaussian quadrature integration in a series of nested loops .
the output is a all - sky map of posterior probability , consisting of @xmath19 equal - area pixels .
the distance prior is a power law of @xmath20 , with @xmath21 being supplied by the user and normally set to @xmath22 for a spatially homogeneous source population . to evaluate the distances , we run two more times , with @xmath23 and @xmath24 .
the resulting three sky maps are denoted @xmath25 , @xmath26 , and @xmath27 .
the healpix - sampled marginal sky posterior @xmath28 , conditional mean distance @xmath29 , and conditional standard deviation of distance @xmath30 are then given by @xmath31 finally , the moments @xmath29 and @xmath30 are converted to the ansatz parameters @xmath32 , @xmath33 , and @xmath34 using the procedure described in [ sec : method - of - moments ] below
. takes about a minute to run @xcite ; the conventional one - dimensional sky map is ready with a response time of a few minutes .
since we will now run three times , the total number of computations will increase by about a factor of 3 .
fortunately , since is able to make effective use of many cpu cores , we can offset the modest increase in computational cost by moving the analysis to a machine with more cores , resulting in a negligible overall change in running time .
lalinference @xcite is the advanced bayesian parameter estimation library .
it includes several algorithms that perform full modeling of the signal and stochastic sampling of the parameter space .
the inputs to lalinference are the time series from all of the detectors .
the output is a cloud of sample points drawn from the posterior .
the samples are converted to a smooth multidimensional probability distribution by clustering them into @xmath35 disjoint sets , each consisting of @xmath36 spatially neighboring points , and building a for each cluster . in cartesian
coordinates , the smoothed distribution is given by a double sum over the clusters and the samples within each cluster : @xmath37.\end{gathered}\ ] ] here , @xmath38 is a weight associated with cluster @xmath0 , and each cluster is described by its covariance @xmath39 and @xmath36 samples @xmath40 .
the @xmath41 denotes the matrix determinant .
the stochastic sampling takes hours to weeks depending on the sophistication of the waveform models that are used and on the treatment of uncertainty in detector calibration .
it takes up to tens of minutes to build the .
we can exactly calculate the conditional mean and standard deviation of the distance for the posterior .
first , we evaluate eq .
( [ eq : kde ] ) at the position @xmath42 : @xmath43 , \intertext{with } c_i & = ( \boldsymbol{n } { \ensuremath{^{^\mathsf{t}}}}{\boldsymbol c_i}^{-1 } \boldsymbol{n})^{-1 } , \\ x_{ij } & = ( \boldsymbol{n } { \ensuremath{^{^\mathsf{t}}}}{\boldsymbol c_i}^{-1 } \boldsymbol x_{ij } ) c_i , \text { and } \\ w_{ij } & = \frac{1}{2\pi } \sqrt{\frac{c_i}{\left| \boldsymbol c_i\right| } } \exp\left[\frac{1}{2}\left(\frac{{x_{ij}}^2}{c_i } - \boldsymbol x_{ij}{\ensuremath{^{^\mathsf{t}}}}\boldsymbol{c_i}^{-1}\boldsymbol x_{ij}\right)\right ] w_i.\end{aligned}\ ] ] we compute the integrals of 1 , @xmath44 , and @xmath45 , weighted by @xmath46 : @xmath47\right ) , \text { and } \\ b & = \sqrt{\frac{c_i}{2\pi } } \exp\left[\frac{-{x_{ij}}^2}{2c_i}\right].\end{aligned}\ ] ] then @xmath25 , @xmath26 , and @xmath27 are converted to @xmath48 , @xmath29 , and @xmath30 using equations ( [ eq : bayestar - rho])([eq : bayestar - s ] ) .
finally , @xmath49 and @xmath50 are converted to @xmath32 , @xmath33 , and @xmath51 using the procedure described in [ sec : method - of - moments ] below . for both and lalinference
, the parameters of the ansatz distribution are extracted using the method of moments .
the ansatz is that the conditional distribution of distance is described by the function @xmath52 r^2 \\
\text{for } r & \geq 0 .
\nonumber\end{aligned}\ ] ] the @xmath53th moment of the distance ansatz is @xmath54 r^{2+n } \ , { \ensuremath{\mathrm{d}}}r.\ ] ] the conditional mean and standard deviation of the ansatz distribution are @xmath55 our task is , given the conditional mean @xmath49 and standard deviation @xmath50 as measured from the actual posterior probability distribution , to numerically solve the following system of equations for @xmath32 and @xmath33 : @xmath56 we can reduce this to a single equation by defining @xmath57 and @xmath58 .
with this substitution , the moments can be written as @xmath59 the function @xmath60 is the upper tail of the normal distribution , @xmath61 is the normal distribution function , and @xmath62 is the hazard function .
then equations ( [ eq : hat - m ] ) and ( [ eq : hat - s ] ) become @xmath63 the derivative of the left - hand side , @xmath64 , is given by @xmath65 with @xmath66 we solve equations ( [ eq : mom - f ] ) and ( [ eq : mom - fprime ] ) for @xmath67 using steffensen s method , an accelerated newton solver . starting from an initial value of @xmath68
, the solution converges to machine precision in 10 iterations .
finally , we calculate @xmath32 , @xmath33 , and @xmath35 as follows : @xmath69 in the rare event that the solution does not converge , or yields an invalid value such that @xmath70 , we set @xmath71
the ansatz guarantees that the first two moments of distance are exactly reproduced along all .
however , we must ask how accurately the ansatz represents the 3d posterior as a whole . the @xmath72@xmath72 plot graphical test , popularized in the parameter estimation literature by @xcite , compares two populations by plotting their cumulative distributions against each other .
if the two distributions match , then the result should be a diagonal line . in our case
, we compare the to the lalinference posterior samples by projecting both the and the posterior samples along the distribution s three principal axes , yielding three @xmath72@xmath72 tests . as shown in fig .
[ fig : kde - ansatz - comparison]a , the plot is nearly diagonal , indicating that the is a faithful representation of the posterior samples .
we then compare the with the ansatz by drawing samples from the ansatz distribution ( fig .
[ fig : kde - ansatz - comparison]b ) .
some deviation is perceptible ; in the most extreme cases we find a maximum difference in credible levels of about 5% .
@xmath72@xmath72 tests of the conditional distance distribution itself along individual generally also agree within 5% or better , except in directions of low probability ( small @xmath4 ) .
we test the statistical self - consistency of the entire ensemble of simulated events in fig .
[ fig : pp ] . here
, we show a cumulative histogram of the number of simulated events whose true and coordinates are found within a given credible level .
we find that both the sky maps and the ansatz are self - consistent within a binomial 95% tolerance band due to the finite sample size of 250 events .
our interpretation is that the ansatz is a reasonable approximation of the full 3d posterior , in the sense that a stated 50% credible volume has a 50%@xmath73% chance of containing the source .
the most obvious alternative to the ansatz is a densely sampled 3d grid , or a stack of 2d sky maps for a series of distance shells .
either would be just as conceptually simple , but computationally cumbersome due to size .
the is an accurate representation of the posterior , but is expensive to evaluate because it is a sum of @xmath74 gaussians . as a rapidly available data product and as a tool for real - time observation planning , the ansatz distribution is a reasonable compromise . [ 2d]two - dimensional [ 2 + 1d]2 + 1dimensional [ 2mrs]2mass redshift survey [ 3d]three - dimensional [ 2mass]two micron all sky survey [ advirgo]advanced virgo [ ami]arcminute microkelvin imager [ agn]active galactic nucleus [ agn]active galactic nuclei [ aligo]advanced [ askap]australian pathfinder [ atca]australia telescope compact array [ atlas]asteroid terrestrial - impact last alert system [ bat]burst alert telescope [ batse]burst and transient source experiment [ bayestar]bayesian triangulation and rapid localization [ bbh]binary black hole [ bhbh] [ bh]black hole [ bns]binary neutron star [ carma]combined array for research in millimeter - wave astronomy [ casa]common astronomy software applications [ cfh12k]canada france
hawaii @xmath75 pixel ccd mosaic [ crts]catalina real - time transient survey [ ctio]cerro tololo inter - american observatory [ cbc]compact binary coalescence [ ccd]charge coupled device [ cdf]cumulative distribution function [ cgro]compton gamma ray observatory [ cmb]cosmic microwave background [ crlb]cramr rao lower bound [ wb]coherent waveburst [ daswg]data analysis software working group [ dbsp]double spectrograph [ dct]discovery channel telescope [ decam]dark energy camera [ des]dark energy survey [ dft]discrete fourier transform [ em]electromagnetic [ er8]eighth engineering run [ fd]frequency domain [ far]false alarm rate [ fft]fast fourier transform [ fir]finite impulse response [ fits]flexible image transport system [ flops]floating point operations per second [ fov]field of view [ fov]fields of view [ ftn]faulkes telescope north [ fwhm]full width at half - maximum [ gbm]gamma - ray burst monitor [ gcn]gamma - ray coordinates network [ gmos]gemini multi - object spectrograph [ grb]gamma - ray burst [ gsc]gas slit camera [ gsl]gnu scientific library [ gtc]gran telescopio canarias [ gw]gravitational wave [ hawc]high - altitude water erenkov gamma - ray observatory [ hct]himalayan chandra telescope [ healp]hierarchical equal area isolatitude pixelization [ heasarc]high energy astrophysics science archive research center [ hete]high energy transient explorer [ hfosc]himalaya faint object spectrograph and camera [ hmxb]high - mass x - ray binary [ hmxb]high - mass x - ray binaries [ hsc]hyper suprime - cam [ iact]imaging atmospheric erenkov telescope [ iir]infinite impulse response [ imacs]inamori - magellan areal camera & spectrograph [ imr]inspiral - merger - ringdown [ ipac]infrared processing and analysis center [ ipn]interplanetary network [ ptf]intermediate [ ism]interstellar medium [ iss]international space station [ kagra]kamioka gravitational - wave observatory [ kde]kernel density estimator [ kn]kilonova [ kne]kilonovae [ lat]large area telescope [ lcogt]las cumbres observatory global telescope [ lho ] hanford observatory [ lib]lalinference burst [ ligo]laser interferometer observatory [ grb]low - luminosity [ lloid]low latency online inspiral detection [ llo ] livingston observatory [ lmi]large monolithic imager [ lofar]low frequency array [ los]line of sight [ loss]lines of sight [ lmc]large magellanic cloud [ lsb]long , soft burst [ lsc ] scientific collaboration [ lso]last stable orbit [ lsst]large synoptic survey telescope [ lt]liverpool telescope [ lti]linear time invariant [ map]maximum a posteriori [ mbta]multi - band template analysis [ mcmc]markov chain monte carlo [ mle ] estimator [ ml]maximum likelihood [ mou]memorandum of understanding [ mous]memoranda of understanding [ mwa]murchison widefield array [ ned]nasa / ipac extragalactic database [ nsbh]neutron star black hole [ nsbh] [ nsf]national science foundation [ nsns] [ ns]neutron star [ o1 ] s first observing run [ o2 ] s second observing run [ lib]omicron+ [ ot]optical transient [ p48]palomar 48 inch oschin telescope [ p60]robotic palomar 60 inch telescope [ p200]palomar 200 inch hale telescope [ pc]photon counting [ pessto]public eso spectroscopic survey of transient objects [ psd]power spectral density [ psf]point - spread function [ ps1]pan - starrs 1 [ ptf]palomar transient factory [ quest]quasar equatorial survey team [ raptor]rapid telescopes for optical response [ reu]research experiences for undergraduates [ rms]root mean square [ rotse]robotic optical transient search [ s5 ] s fifth science run [ s6 ] s sixth science run [ saa]south atlantic anomaly [ shb]short , hard burst [ shgrb]short , hard [ ska]square kilometer array [ smt]slewing mirror telescope [ s / n]signal - to - noise ratio [ ssc]synchrotron self - compton [ sdss]sloan digital sky survey [ sed]spectral energy distribution [ sgrb]short [ sn]supernova [ sn]supernova [ i]type ia [ i]type ic [ i - bl]broad - line type ic [ i - bl]broad - line type ic [ svd]singular value decomposition [ tarot]tlescopes action rapide pour les objets transitoires [ tdoa]time delay on arrival [ tdoa]time delays on arrival [ td]time domain [ toa]time of arrival [ toa]times of arrival [ too]target - of - opportunity [ too]targets of opportunity [ uffo]ultra fast flash observatory [ uhe]ultra high energy [ uvot]uv / optical telescope [ vhe]very high energy [ vista@eso]visible and infrared survey telescope [ vla]karl g. jansky very large array [ vlt]very large telescope [ vst@eso ] survey telescope [ wam]wide - band all - sky monitor [ wcs]world coordinate system [ w.s.s.]wide - sense stationary [ xrf]x - ray flash [ xrf]x - ray flashes [ xrt]x - ray telescope [ ztf]zwicky transient facility we thank the aspen center for physics and nsf grant # 1066293 for hospitality during the conception , writing , and editing of this paper . we thank p shawhan and f tombesi for detailed feedback on the manuscript .
the online data release is available at https://dcc.ligo.org/p1500071/public/html .
this is document p1500071-v7 . | this is an online supplement to the letter of @xcite , in which we demonstrated a rapid algorithm for obtaining joint estimates of sky location and luminosity distance from observations of mergers with advanced and virgo . we argued that combining the reconstructed volumes with positions and redshifts of possible host galaxies can provide large - aperture but small instruments with a manageable list of targets to search for optical or infrared emission . in this supplement
, we document the new -based file format for localizations of gravitational - wave transients .
we include python sample code to show the reader how to perform simple manipulations of the sky maps and extract ranked lists of likely host galaxies .
finally , we include mathematical details of the rapid volume reconstruction algorithm .
= -1 |
the supernova remnant g296.5 + 10.0 was observed using the australia telescope compact array at 1.4 ghz with an angular resolution of 30 arcseconds .
signals from the dual linear polarized feeds were combined to produce images of the remnant in linear polarisation .
the rotation measure synthesis method ( brentjens & de bruyn , 2005 ) was employed to obtain an rm spectrum at each pixel in the image ( figure 1 ) .
the 104 mhz bandwidth was split into 13 @xmath0 8 mhz spectral channels , giving sufficient coverage in @xmath1@xmath2 to remove any n@xmath3 ambiguities in the rm values .
x - ray data in the energy range 0.12.4 kev were taken from the rosat archive , via nasa s skyview .
these complimentary data allow us to estimate the electron density in the remnant .
in order to disentangle the electron density ( n@xmath4 ) and line - of - sight magnetic field ( @xmath5@xmath6 ) we measured the x - ray intensity at every point in the rosat image , after regridding and smoothing them to match the atca image . rm is defined as the integral of n@xmath4 and @xmath5@xmath6 along the line of sight .
the thermal x - ray surface brightness , s@xmath7 , is proportional to n@xmath4@xmath2 .
therefore if rm and s@xmath7@xmath8 are correlated , the faraday rotation is occurring within the thermal plasma ( matsui et al .
a plot of rm vs. s@xmath7@xmath8 ( figure 2 ) shows that there is no clear correlation between rm and s@xmath7@xmath8 in the remnant , which means that the rms originate from the ism / halo along the line - of - sight .
therefore the rm variations across the remnant may reflect the structure of the magnetised interstellar medium ( haverkorn et al .
2004 ) and the smooth component may tell us about the overall magnetic field morphology within the remnant ( ransom et al .
this will form the basis of our further analysis of these data .
the following analytical steps will be taken : ( 1 ) modelling of the rm structure across the remnant to determine the large - scale structure and morphology of the magnetic field ( i.e. linear , spiral , toroidal ) in the remnant ( 2 ) analysis of smaller - scale rm variations across the remnant , to probe the turbulent scale - sizes in the ism and galactic halo . | radio continuum emission from the supernova remnant g296.5 + 10.0 was observed using the australia telescope compact array . using a 104 mhz bandwidth split into 13 @xmath0 8 mhz spectral channels , it was possible to produce a pixel - by - pixel image of rotation measure ( rm ) across the entire remnant .
a lack of correlation between rm and x - ray surface brightness reveals that the rms originate from outside the remnant . using this information
, we will characterise the smooth component of the magnetic field within the supernova remnant and attempt to probe the magneto - ionic structure and turbulent scale sizes in the ism and galactic halo along the line - of - sight . |
`` we live in a global world '' has become a clich @xcite .
historically , the exchange of goods , money , and information was naturally limited to nearby locations , since globalization was effectively blocked by spatial , territorial , and cultural barriers @xcite . today
, new technology is overcoming these barriers and exchange can take place in an increasingly international arena @xcite .
nevertheless , geographical proximity still seems to be important for the trade of goods @xcite as well as for mobile phone communication @xcite and scientific collaboration @xcite .
however , since the internet allows information to travel more easily and rapidly than goods , it remains unclear what are the effective barriers of global information exchange .
as information exchange requires shared interests , we therefore need to better understand global connections in interest , and the factors that form these connections .
although globalization of information has been discussed extensively in the research literature @xcite , currently there is no method to quantitatively map bilateral information interests from large - scale data . without such a method , it becomes difficult to justify qualitative statements about , for example , the complex interplay between shared values and conflict on a global scale .
we use data mining and statistical analysis to device a measure of bilateral information interests , and use this measure to construct a world map of information interests . to study interests on a global scale , we use the free online encyclopedia wikipedia , which has evolved into one of the largest collaborative repositories of information in the history of mankind @xcite .
the free online encyclopedia consists of almost 300 language editions , with english being the largest one .
this multi - lingual encyclopedia captures a wide spectrum of information in millions of articles .
these articles undergo a peer - reviewed editing process without a central editing authority .
instead , articles are written , reviewed , and edited by the public .
each article edit is recorded , along with a time - stamp , and , if the editor is unregistered , the computer s ip address .
the ip address makes it possible to connect each edit to a specific location .
therefore we can use wikipedia editors as sensors for mapping information interest to specific countries . in this paper , we use co - editing of the same wikipedia article as a proxy for shared information interests . to find global connections , we look at how often editors from different countries co - edit the same articles . to infer connections of shared interest between countries , we develop a statistical model and represent significant correlations between countries as links in a global network .
structural analysis of the network suggests that interests are polarized by factors related to geographical proximity , language , religion and historical background .
we quantify the effects of these factors using regression analysis and find that information exchange indeed is constrained by the impact of social and economic factors connected to shared interests .
as one of the largest and most linguistically diverse repositories of human knowledge , wikipedia has become the world s main platform for archiving factual information @xcite
. one important feature of wikipedia is that every edit made to an article is recorded .
thanks to this detailed data , wikipedia provides a unique platform for studying different aspects of information processes , for example , semantic relatedness of topics @xcite , collaboration @xcite , social roles of editors @xcite , and the geographical locations of wikipedia editors @xcite . in this work , we used data from wikipedia dumps to select a random sample from the english wikipedia edition , which is the largest and most widespread language edition . in total , the english edition has around 10 million articles , including redirects and duplicates .
since retrieving the editing histories of all articles is computationally demanding , we randomly sampled more than six million articles from this set . for each english article
, we retrieved the complete editing history of the same article in all language editions that the english wikipedia page links to .
finally we merged all language editions together to create a global editing history for each article . for each edit ,
the editing history includes the text of the edit , its time - stamp , and , for unregistered editors , the ip address of the editor s computer . from the ip address associated with the edit
, we retrieved the geolocation of the corresponding editor using an ip database . for the purpose of spatial analysis
, we limited the analysis to edits from unregistered editors , because data on the location for most of the registered wikipedia editors are unavailable .
the resulting dataset contains more than six million ( 6,285,753 ) wikipedia articles and about 140 million edits in total .
we use these edits to create interest profiles for countries .
we identify the interest profile of a country by aggregating the edits of all wikipedia editors whose ips are recorded in the country .
if an article is co - edited by editors located in different countries , we say that the countries share a common interest in the information of the article .
in other words , we connect countries if their editors co - edit the same articles . indirectly , we let individuals who edit wikipedia represent the population of their country . while wikipedia editors in a country certainly do not represent a statistically unbiased sample , there is a higher tendency that they edit contents that are related to the country in which they live @xcite .
therefore , we approximate the interest profile of a country with collective editing behavior of editors in that country .
inferring the location of all editors on the country level is non - trivial .
although we have data on all edits , we do not know the location of registered editors because their ips are not recorded .
one proposed approach to tackle this problem makes use of circadian rhythms of editing activity to infer the location of the editors @xcite .
this method approximates the longitude of a location but provides little information about its latitude .
therefore , we must limit the analysis to the activity of unregistered editors with recorded ip addresses .
this will arguably affect the results .
not only do registered editors contribute to 70% of all 140 million edits , they also have somewhat different behavior . for example , many of the most active registered users take on administrative functions , develop career paths , or specialize in covering selected topics @xcite . on the other hand , some unregistered editors are involved in vandalism , but often their activity nevertheless indicates their interest .
while we can only speculate about how including registered editors would affect the results , unregistered editors can nevertheless provide useful information about shared interests between countries . from the co - editing data , we create a network that represents countries as nodes and shared interests as links .
the naive approach is to use the raw counts of co - edits between countries as weighted links .
the problem with this approach is that it is biased toward the number of editors in each country .
some countries may be strongly connected , not because of evident shared interests but merely as a result of a large community of active wikipedia editors . to address this problem
, we propose a statistical validation method that filters out connections that could exist only due to size effects or noise .
the filtering method assumes a multinomial distribution and determines the expected number of co - occurring edits from the empirical data . in other words , we infer significant links in a bipartite system in which countries are in one set and articles are in the other set .
there are other existing methods to evaluate the significant correlation between entities in bipartite systems .
for example , @xcite proposed a systematic approach to one - mode projections of bipartite graphs for different motifs . in another work
, @xcite used the hypergeometric distribution and measured the @xmath0-value for each subset of the bipartite network .
moreover , @xcite proposed a community detection method to classify topics to articles more efficiently , and @xcite used a disparity filtering method to infer significant weights in networks .
finally , @xcite adopted a statistical approach to determine significant links between languages in various written documents .
however , the model that we use has the advantages that it makes it easy to account for size variation inside an article and to compute the @xmath1-scores for analyzing the country - based editor activity .
in this section we outline the formalisation of the model .
we link countries based on their co - occurring edits over all wikipedia articles . for a specific article @xmath2
, we calculate the link weight between all pairs of countries that edited the article , as follows : if editors in country @xmath3 have edited an article @xmath4 times , and editors in country @xmath5 have edited the same article @xmath6 times , then the countries empirical link weight , @xmath7 , is calculated as : @xmath8 since the total number of articles is over six million , most country pairs have co - edited at least one article . therefore , the aggregation of all articles results in numerous links between countries , and the countries with relatively large editing activities become highly central . accordingly , we can not know if the link exists by chance , or because countries actually tend to edit the same articles more frequently than expected .
to determine which links are statistically significant , we compare the empirically observed link weights with the weight given by a null model . in the null model , we assume that each edit comes from a country randomly picked proportionally to its total number of edits .
more specifically , the random assignments are performed by drawing the countries from a multinomial distribution .
that is , for each edit , country @xmath3 is selected proportional to its cumulative editing activity , @xmath9 , where @xmath10 is the total number of edits for all articles .
note that each edit is sampled independently from all other edits , and that the cumulative edit activity of a country in the null model on average will be the same as the observed one .
this null model preserves the average level of activity of the countries , but randomizes the temporal order and the articles that countries edit .
figure [ fig : model ] shows an example of this reshuffling scheme with four articles . from the null model
, we can analytically compute the expected probability , @xmath11 , that two countries @xmath3 and @xmath5 edit the same article @xmath2 ( detailed derivation in the methods ) : @xmath12 where @xmath13 is the total number of edits in article @xmath2 . to compare the empirical and expected link values , we compute standardized values , so called @xmath1-scores . for countries
@xmath3 and @xmath5 and article @xmath2 , the @xmath1-score @xmath14 is defined as @xmath15 where the standard deviation @xmath16 , is computed in the method section .
the @xmath1-scores are useful for comparisons of weights , since they account for the large variations that exist in the articles edit histories .
we then sum over all articles to find the cumulative @xmath1-score for countries @xmath3 and @xmath5
@xmath17 using the bonferroni correction , we consider a link to be significant if the probability of observing the total @xmath1-score is less than @xmath18 , where @xmath19 is the number of countries . since the total @xmath1-score is a sum over many independent variables , we can approximate the expected total @xmath1-score distribution with a normal distribution .
the normal distribution has average value @xmath20 and standard deviation @xmath21 , where @xmath22 is the number of wikipedia articles .
thus , the threshold for the significant link weight is @xmath23 , where @xmath24 is derived from the condition that @xmath25 , where @xmath26 countries and @xmath27 is the standard gaussian distribution ( with zero average and unit variance ) .
if the total @xmath1-score is larger than the threshold , we create a link between countries @xmath3 and @xmath5 with weight @xmath28 according to @xmath29 in summary , the interest model maintains the average level of activity of the countries and randomizes the articles that they edit . by comparing results from the interests model and empirical values , we can identify significant links between countries .
to investigate the effective barriers of global information exchange , we first identify large - scale structures among the thousands of links between countries . in this way
, we can highlight the groups of countries that share interest in the same information . to reveal such groups among the pairwise connections , we use a network community detection method based on random walks as a proxy for interest flows . while the community - detection method we use is good at breaking chains of links
, we may connect some countries primarily based on strong connections with common countries and not between themselves . nevertheless , simplifying and highlighting important structures provide a valuable map to investigate the large - scale structure of global information exchange . in our clustering approach ,
we first build a network of countries connected with the significant links we find in our filtering . to identify groups of countries , we envision an editor game in which editors from different countries are active in sequence . in this relay race , a country exchanges information to another country proportional to the weight of the link between the countries .
accordingly , the sequence of countries forms a random walk and certain sets of countries with strong internal connections will be visited for a relatively long time .
this process is analog to the community - detection method known as the map equation @xcite .
here we use the map equation s associated search algorithm infomap @xcite to identify the groups of countries we are looking for and to reveal the large - scale structure .
we will discuss the results at four levels of detail , from the big picture to the detailed dynamics , and highlight different potential mechanisms for barriers of information exchange .
first , we will show a global map of countries with shared information interests , and continue with the interconnections between the clusters .
then we will consider each cluster separately and examine the interconnections between countries within the clusters .
finally , we will apply multiple regression analysis to examine explanatory variables to the highways and barriers of information exchange .
the world map of information interests suggests that cultural and geopolitical features can explain the division of countries . between the 234 countries , we identified 2,847 significant links that together form a network of article co - edits . by clustering the network
, we identified 18 clusters of strongly connected countries ( see supplementary table for a detailed list of countries in each cluster ) .
the resulting network is illustrated as a map in fig .
[ fig : full_worldmap_m1 ] , where countries of the same cluster share the same color .
the map suggests that the division of countries can be explained by a combination of cultural and geopolitical features .
for example , the united states and canada share a long geographical border and extensive mutual trade , and are clustered together despite the fact that other english - speaking countries are not . moreover , religion is a plausible driver for the formation of the cluster of countries in the middle east and north africa , as well as the cluster of russia and the orthodox eastern - european countries @xcite .
another factor in the formation of shared information interests is language .
for example , countries in central and south america are divided into two clusters with portuguese and spanish as common languages in each cluster , respectively .
colonial history can also shape similarity in interests , as in the cluster of portugal , angola and brazil , as well as the cluster of former soviet union countries @xcite .
overall , there is strong empirical evidence that geographical proximity , common religion , shared language , and colonial history can explain the division of countries . to examine the connections between clusters
, we looked at the network structure at the cluster level .
the network in fig.[fig : worldfrance ] shows the connections between the clusters of countries illustrated in fig .
[ fig : full_worldmap_m1 ] with the same color coding .
connections tend to be stronger between clusters of geographically proximate countries also at this level .
interestingly , the middle east cluster in turquoise has the largest link strengths to other clusters , forming a hub that connects east and west , north and south . interpreting the strong connections as potential highways for information exchange , the middle east is not only a melting pot of ideas , but also plays an important role in the spread of information . to get better insights into how the clusters are shaped
, we zoomed into the country networks within clusters . in the upper left corner of fig .
[ fig : worldfrance ] , we show the strongest connections within the central european cluster .
it suggests that countries are linked based on the overlap of the official languages @xcite .
for example , belgium has three official languages , dutch , french and german .
indeed , belgium is connected closely with the netherlands , france , and luxembourg .
we observed the same pattern in other clusters , and the triad of switzerland , germany , and austria is another example of strongly linked countries with a shared language .
just to illustrate what interests can form the bilateral connections , we looked at a number of concrete examples .
first , we ranked the articles according to their significant @xmath1-scores for each pair of countries .
then we looked at the top - ranked articles and here report the results for two european country pairs : germany austria in the european cluster and sweden norway in the scandinavian cluster .
the articles with the most significant co - edits relate to local and regional interests , including sports , media , music , and places ( see supplementary table ) . for example
, the top - ranked articles in the germany austria list include an austrian singer who is also popular in germany , and an austrian football player who is playing in the german league .
the top - ranked articles in the sweden norway list shows a similar pattern of locally related topics , for example , a host of a popular tv show simultaneously aired in sweden and norway , a swedish football manager who has been successful both in sweden and norway , and a music genre that is nearly exclusive to scandinavian countries .
altogether , the top articles suggest that an important factor for co - editing is related interests , which in turn may be an effect of shared language , religion , or colonial history , as well as geographical proximity or large volume of trade between countries .
to quantify the impact of social and economic factors behind the shared interests , we performed a multiple regression quadratic assignment procedure ( mrqap ) analysis .
this method is specifically suited when there are collinearity and autocorrelation in the data @xcite .
we performed the mrqap using the ` netlm ` function in the ` sna ` r package @xcite .
the dependent variables in the regression model were the significant @xmath1-scores that we obtained from the data .
our independent variables were geographical proximity , trade @xcite , colonial ties , shared language , and shared religion , as suggested by the analysis of the map of information interests ( see the supplementary information s2 for a more detailed description of the data ) .
lrrrrr & @xmath30 & @xmath31 & @xmath32 & @xmath33 & @xmath34 + intercept & 0.41 & 0.3 & 2.33 & 2.33 & 2.28 + shared language & 0.91@xmath35 ( 69 ) & 0.82@xmath35 ( 64 ) & 0.77@xmath35 ( 60 ) & 0.75@xmath35 ( 58 ) & 0.74@xmath35 ( 57 ) + shared religion & & 2.76@xmath35 ( 46 ) & 2.6@xmath35 ( 44 ) & 2.6@xmath35 ( 43 ) & 2.44@xmath35 ( 40 ) + log distance & & & -0.23@xmath35 ( -23 ) & -0.23@xmath35 ( -23 ) & -0.23@xmath35 ( -23 ) + colonial tie & & & & 4.5@xmath35 ( 22 ) & 4.35@xmath35 ( 21 ) + log trade & & & & & 0.03@xmath35 ( 10 ) + adjusted r - squared & 0.13 & 0.19 & 0.20 & 0.21 & 0.22 + f - statistic & 7,774 & 3,590 & 2,610 & 2,110 & 1,716 + df & 30,874 & 30,873 & 30,872 & 30,871 & 30,870 + all independent variables show significant correlation with the data ( see table [ table : regression ] ) . to observe the variation between different independent matrices , we examined them in different models . in model @xmath30 , we examined the influence of shared language , which explains 13% of our observation . in model @xmath31 ,
we added shared religion , which increases the power of the model to 19% . in model @xmath32
, we included the geographical proximity .
it slightly increases the r - squared and has negative relation to the observed @xmath1-scores , since short distance corresponds to high proximity . in models @xmath33 and @xmath34 , respectively , we added colonial ties and trade . including all these explanatory variables into the regression model enabled us to increase the explanatory power of the model to 22% .
the correlation of each variable with the observed @xmath1-scores can be inferred from the @xmath36-statistic .
shared language shows the strongest association , followed by shared religion , geographical proximity , colonial ties , and volume of trade ( see table [ table : regression ] ) .
the influence of language on shared interests is not surprising .
it is well known that interests are formed by cultural expression and public opinion , and language is an important platform for these expressions @xcite .
that the relation between language and interests is important has also been demonstrated by the surprisingly small overlap between languages in wikipedia @xcite and the variation in the editing of controversial topics @xcite .
moreover , the influence of religion is in line with the huntington s thesis that the source of division between people in the post - cold war period is primarily rooted in cultural differences and religion @xcite .
similar results were found in other studies that analyzed twitter and email communication worldwide @xcite .
overall , the analysis reveals that information exchange is constrained by the impact of social and economic factors connected to shared interests . in other words , globalization of the technology
does not bring globalization of the information and interests .
language , religion , geographical proximity , historic background , and trade are potential driving factors to polarize the information interests .
these results coincide with earlier works that highlight the impact of the colonization , immigration , economics , and politics on the cultural similarities and diversities @xcite .
by simplifying and highlighting the important structure in the myriad edits of wikipedia , we provide a world map of shared information interests .
we find that information interests despite globalization are diverse , and that the highways and barriers of information exchange are formed by social and economic factors connected to shared interests . in descending order
, we find that language , religion , geographical proximity , historic background , and trade explain the diversity of interests .
while technological advances in principle have made it possible to communicate with anyone in the world , these social and economic factors limit us from doing so and information interests remain diverse .
questions remain how different social and economic factors affect different regions , how they relate to conflicts on a global scale , and how the impact of these factors changes over time
. it would therefore be interesting to extend the methodology to track changes over time .
to find connections in interest , we measure the co - occurring edits of two countries in the same articles .
we quantify the connection with an empirical weight that is computed as the product of the countries edit activities in the article . for a wikipedia article @xmath2 ,
if the total edit activity of country @xmath3 is denoted by @xmath4 , and for country @xmath5 is @xmath6 , then we calculate the empirical weight , @xmath7 , according to @xmath37 as the total edit activity of countries differs , the probability that countries appear together in a certain article varies .
if the total number of edits for all articles is @xmath10 , then the expected proportion of edits for country @xmath3 is @xmath38 this is the probability of country @xmath3 making a random edit overall , and this is the null model we use to filter noisy connections in the interest model .
assume that there is a total of @xmath39 countries , and a total of @xmath40 edits for article @xmath0 .
let @xmath41 denote the number of edits from country @xmath42 .
then the probability of any particular combination of edits for the various countries follows a multinomial distribution @xmath43 with the above distribution , we can compute the expected probability of the co - occurrence of two countries @xmath3 and @xmath5 in an article @xmath44 following the multinomial theorem , we can now calculate the mean , variance and covariance matrices for the occurrence of a country pair .
the mean value of the co - occurrence of two countries becomes @xmath45 using the multinomial theorem multiple times , one can also compute the variance : @xmath46 thus , the standard deviation of the pair , @xmath16 , is the square root of equation .
this equation enters into the definition of the @xmath1-score in equation .
the sum of the @xmath1-scores is then approximated with a gaussian distribution .
this approximation is well justified by the large number of articles . in practice ,
already @xmath47 articles give a good approximation , as shown by the numerical simulations in the supplementary information .
arazy , o. , ortega , f. , nov , o. , yeo , l. , and balila , a. ( 2015 ) . functional roles and career paths in wikipedia . in _ proceedings of the 18th acm conference on computer
supported cooperative work & social computing _ , pages 10921105 .
auer , s. and lehmann , j. ( 2007 ) . what have innsbruck and leipzig in common ? extracting semantics from wiki content . in _ the semantic web : research and applications _ , pages 503517 .
springer .
bleich , e. ( 2005 ) .
the legacies of history ?
colonization and immigrant integration in britain and france .
, 34(2):171195 .
butts , c. t. ( 2008 ) .
social network analysis with sna .
, 24(6):151 .
cairncross , f. ( 2001 ) . .
harvard business press .
callahan , e. s. and herring , s. c. ( 2011 ) .
cultural bias in wikipedia content on famous persons .
, 62(10):18991915 .
castells , m. ( 2011 ) . , volume 2 .
john wiley & sons .
dekker , d. , krackhardt , d. , and snijders , t. a. ( 2007 ) .
sensitivity of mrqap tests to collinearity and autocorrelation conditions .
, 72(4):563581 .
edler , d. and rosvall , m. ( 2015 ) .
the infomap software package .
http://www.mapequation.org .
fagiolo , g. , reyes , j. , and schiavo , s. ( 2010 ) .
the evolution of the world trade web : a weighted - network analysis .
, 20(4):479514 .
feldman - bianco , b. ( 2001 ) .
razilians in portugal , portuguese in brazil : constructions of sameness and difference 1 . , 8(4):607650 .
friedman , t. l. ( 2000 ) . .
macmillan .
gelfand , m. j. , raver , j. l. , nishii , l. , leslie , l. m. , lun , j. , lim , b. c. , duan , l. , almaliach , a. , ang , s. , arnadottir , j. , et al .
differences between tight and loose cultures : a 33-nation study . , 332(6033):11001104 .
hale , s. a. ( 2014 ) .
multilinguals and wikipedia editing . in _ proc . conf . on web science _ , pages 99108 .
hecht , b. and gergle , d. ( 2010a ) .
the tower of babel meets web 2.0 : user - generated content and its applications in a multilingual context . in _ proc .
sigchi conf . on human factors in computing systems _ ,
pages 291300 .
hecht , b. j. and gergle , d. ( 2010b ) . on the `` localness '' of user - generated content . in _ proceedings of the 2010 acm conference on computer supported cooperative work _
, cscw 10 , pages 229232 , new york , ny , usa . acm .
hennemann , s. , rybski , d. , and liefner , i. ( 2012 ) .
the myth of global science collaboration collaboration patterns in epistemic communities .
, 6(2):217225 .
huntington , s. p. et al .
the clash of civilizations ?
kaluza , p. , klzsch , a. , gastner , m. t. , and blasius , b. ( 2010 ) .
the complex network of global cargo ship movements .
, 7(48):10931103 .
keegan , b. , gergle , d. , and contractor , n. ( 2012 ) .
do editors or articles drive collaboration ? : multilevel statistical network analysis of wikipedia coauthorship . in _ proc .
acm conf . on computer supported cooperative work _ ,
pages 427436 .
kimmons , r. m. ( 2011 ) . understanding collaboration in wikipedia .
, 16(12 ) .
kose , m. a. and ozturk , e. o. ( 2014 ) .
a world of change .
, page 7 .
krackhardt , d. ( 1988 ) .
predicting with networks : nonparametric multiple regression analysis of dyadic data .
, 10(4):359381 .
lambiotte , r. , blondel , v. d. , de kerchove , c. , huens , e. , prieur , c. , smoreda , z. , and van dooren , p. ( 2008 ) .
geographical dispersal of mobile communication networks .
, 387(21):53175325 .
lancichinetti , a. , sirer , m. i. , wang , j. x. , acuna , d. , krding , k. , and amaral , l. a. n. ( 2015 ) .
high - reproducibility and high - accuracy method for automated topic classification . , 5:011007 .
lieberman , m. d. and lin , j. ( 2009 ) .
you are where you edit : locating wikipedia contributors through edit histories . in _
icwsm_. mesgari , m. , okoli , c. , mehdi , m. , nielsen , f. . , and lanamki , a. ( 2014 ) .
`` the sum of all human knowledge '' : a systematic review of scholarly research on the content of wikipedia .
overman , h. g. , redding , s. , and venables , a. ( 2003 ) . .
blackwell publishing .
pan , r. k. , kaski , k. , and fortunato , s. ( 2012 ) .
world citation and collaboration networks : uncovering the role of geography in science .
radinsky , k. , agichtein , e. , gabrilovich , e. , and markovitch , s. ( 2011 ) . a word at a time : computing word relatedness using temporal semantic analysis . in _ proc .
on world wide web _ , pages 337346 .
risse , t. ( 2001 ) . a european identity ?
europeanization and the evolution of nation - state identities . in cowles , m. g. , caporaso , j. a. , and risse - kappen , t. , editors ,
_ transforming europe : europeanization and domestic change_. cornell university press ithaca , ny . ronen , s. , gonalves , b. , hu , k. z. , vespignani , a. , pinker , s. , and hidalgo , c. a. ( 2014 ) .
links that speak : the global language network and its association with global fame . .
rosvall , m. , axelsson , d. , and bergstrom , c. t. ( 2009 ) .
the map equation .
, 178(1):1323 .
rosvall , m. and bergstrom , c. t. ( 2008 ) .
maps of random walks on complex networks reveal community structure . , 105(4):11181123 .
serrano , m. . ,
bogu , m. , and vespignani , a. ( 2007 ) .
patterns of dominant flows in the world trade web . , 2(2):111124 .
serrano , m. . ,
bogu , m. , and vespignani , a. ( 2009 ) .
extracting the multiscale backbone of complex weighted networks .
, 106(16):64836488 . state , b. , park , p. , weber , i. , and macy , m. ( 2015 ) .
the mesh of civilizations in the global network of digital communication .
, 10(5):e0122543 .
subramanian , a. and wei , s .- j . ( 2007 ) .
the wto promotes trade , strongly but unevenly .
, 72(1):151175 .
tgil , s. ( 1995 ) . .
siu press .
trk , j. , iiguez , g. , yasseri , t. , san miguel , m. , kaski , k. , and kertsz , j. ( 2013 ) .
opinions , conflicts , and consensus : modeling social dynamics in a collaborative environment .
, 110(8):088701 .
tumminello , m. , miccich , s. , lillo , f. , piilo , j. , and mantegna , r. n. ( 2011 ) . statistically validated networks in bipartite complex systems .
, 6(3):e17994 .
usunier , j .- c . and lee , j. ( 2005 ) . .
pearson education .
welser , h. t. , cosley , d. , kossinets , g. , lin , a. , dokshin , f. , gay , g. , and smith , m. ( 2011 ) .
finding social roles in wikipedia . in _ proc .
iconference _ , pages 122129 .
yasseri , t. , spoerri , a. , graham , m. , and kertsz , j. ( 2014 ) .
the most controversial topics in wikipedia : a multilingual and geographical analysis . in fichman , p. & hara , n. , editor , _ global wikipedia : international and cross - cultural issues in online collaboration .
_ rowman & littlefield .
yasseri , t. , sumi , r. , and kertsz , j. ( 2012 ) .
circadian patterns of wikipedia editorial activity : a demographic analysis . , 7(1):e30091 .
zweig , k. a. and kaufmann , m. ( 2011 ) . a systematic approach to the one - mode projection of bipartite graphs .
, 1(3):187218 .
we thank alcides v. esquivel , daniel edler , claudia wagner , markus strohmaier and micheal macy for valuable discussions .
we also thank the wikimedia foundation for providing free access to the data .
, a.l . and m.r .
were supported by the swedish research council grant 2012 - 3729 .
the authors declare that they have no competing financial interests .
correspondence and requests for materials should be addressed to f.k . .
the datasets generated during and/or analysed during the current study are available in the `` google sites '' repository , `` https://sites.google.com/site/mappingbilateralwiki/ '' . | * abstract * we live in a global village where electronic communication has eliminated the geographical barriers of information exchange .
the road is now open to worldwide convergence of information interests , shared values , and understanding .
nevertheless , interests still vary between countries around the world .
this raises important questions about what today s world map of information interests actually looks like and what factors cause the barriers of information exchange between countries . to quantitatively construct a world map of information interests ,
we devise a scalable statistical model that identifies countries with similar information interests and measures the countries bilateral similarities . from the similarities we connect countries in a global network and find that countries can be mapped into 18 clusters with similar information interests . through regression we find that language and religion
best explain the strength of the bilateral ties and formation of clusters .
our findings provide a quantitative basis for further studies to better understand the complex interplay between shared interests and conflict on a global scale .
the methodology can also be extended to track changes over time and capture important trends in global information exchange . |
while the standard picture of the hierarchical build - up of large - scale structure in the universe has been supported by ever increasing observational evidence , it is still an ongoing task to understand thoroughly the detailed baryonic physics associated with galaxy formation , such as gas accretion , star formation and feedback .
direct observation of these gaseous processes is challenging . on the other hand ,
given the large galaxy samples currently available to us , such as sdss dr7 @xcite , statistical analysis of galaxy properties has proved to be invaluable to probe the underlying physics affecting galaxy formation and evolution .
the morphology - density correlation for galaxies , uncovered already in the 1930s @xcite , shows a clear segregation , with early - type galaxies preferentially existing in high density regions @xcite .
further investigation has indicated that it is the star - formation related properties , such as color and emission - line flux , that are more directly correlated with environmental density @xcite . at fixed color ,
the residual dependence of morphology on the environmental density is rather weak @xcite .
it is also emphasized by , e.g. , @xcite , @xcite and @xcite , that the environment dependence is in fact quite local .
while galaxies with abundant close neighbors have clearly different properties from those of isolated galaxies , on scales larger than @xmath10 , environmental effects probably have little influence on galaxy properties @xcite .
theoretically , it has long been known from extensions of press - schecter theory that massive halos preferentially reside in dense environments @xcite .
galaxies therein have properties that are different from those in small halos . on the other hand ,
the excursion set theory with sharp - k filter and with dynamics described by the spherical collapse model predicts that the formation history of halos of fixed mass should depend only on smaller - scale structure not on their large - scale environment ( e.g. , * ? ? ?
incorporating the ellipsoidal collapse model into excursion set theory gives halo mass functions and bias in better agreement with numerical simulations @xcite .
in its simple version , however , the theory takes into account the ellipsoidal collapse by considering the average tidal field , and thus also predicts no correlations between the formation history of halos and their specific large - scale environment @xcite .
recent numerical simulations show that halo properties can be strongly correlated with their environment .
halo assembly bias reveals that old halos tend to cluster more strongly than their younger counterparts , and this bias is particularly significant for low mass halos @xcite . using filters other than the sharp - k filter in the excursion set theory leads to non - markovian random walks , which in turn generate correlations between halo formation history and large - scale environment ( e.g. , * ? ? ?
@xcite shows that under the spherical collapse model , the correlations arising purely from this non - markovianity are expected to be stronger for more massive halos , in disagreement with the trend seen in simulations .
taking into account ellipsoidal collapse and the effects of large - scale environment on halo formation statistically @xcite and dynamically @xcite , the age dependent assembly bias can naturally arise and be more apparent for less massive halos .
it is also noted by @xcite that environmental density plays the determining role in the virialization redshift of halos , and that the morphology of large - scale structure contributes mostly to its scatter .
however , given a fixed halo mass , the analytical model of @xcite predicts that the median formation redshift of halos decreases with increasing environmental density , defined as the linear density fluctuation smoothed over @xmath11 .
this is inconsistent with the results from simulations .
@xcite emphasize that for galactic - scale halos of fixed mass , an early formation epoch is closely related to a reduced mass - growth rate at late times that is mainly due to the tidal effects of neighboring massive halos .
as galactic halos are enhanced in filaments near massive halos , the dependence of their formation epoch on environmental density is expected ( see also , e.g. , * ? ? ?
* ; * ? ? ?
* ) . for cluster - scaled halos , however ,
@xcite show that tidal effects from large - scale structure can not play an important role in the environment dependence of formation history .
gas physics allows more complicated interactions between galaxies and their environment .
galaxies are more subject to ram pressure ( e.g. , * ? ? ?
* ) and shock heating ( e.g. , * ? ? ?
* ) in denser regions .
both the cooling efficiency and the gas compression rate are directly related to environmental density .
it has been observed in hydrodynamical simulations that most cold flows are highly anisotropic and follow dark matter filaments ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . in this paper
, we use sdss dr7 data to examine the dependence of galaxy properties on their large - scale environment .
we focus on the influence of the morphology of environment over and above the well - known dependence on environment density .
this analysis may provide important constraints on the theories and hypotheses discussed above .
it can also help clarify if large - scale environment should be included in the halo occupation distribution model ( e.g. , * ? ? ?
* ; * ? ? ?
the rest of the paper is organized as follows . in section 2 , we describe our measurements of large - scale tidal field with sdss dr7 . in section 3 , we present our results .
comparisons with results from simulations with semi - analytical modeling of galaxy formation are shown in section 4 .
section 5 contains discussions and our conclusions .
the base data set adopted in this paper is sdss dr7 @xcite .
sdss used a dedicated wide - field 2.5 m telescope @xcite located at apache point observatory to obtain images in the @xmath12 bands over an area of @xmath13 and spectra of @xmath14 million selected objects among which galaxies amount to @xmath15 million .
the photometric data was calibrated through ubercalibration @xcite , which used the overlap between adjacent imaging runs to tie together the photometry of all imaging observations .
fibers arranged on plates @xmath16 in radius were assigned with an efficient tiling algorithm to spectroscopic targets selected with certain photometric criteria . due to the fiber collision effect
, only one spectroscopic measurement can be done for a group of targets with separations less than @xmath17 arsec .
overlaps of tiles can help to reduce the number of missed targets .
the overall incompleteness for galaxies is @xmath18 .
our test shows that this incompleteness together with @xmath19 incompleteness caused by mismeasurement of redshift in spectra and bright star blocking is negligible for our measurements of the large - scale tidal field .
our analysis is based on the large - scale structure sample vagc dr72 of new york university value added catalogue ( nyu - vagc;@xcite ) .
it is constructed from sdss , first , 2mass , 2dfgrs and pscz .
the sdss part is updated for dr7 .
there are @xmath20 million galaxies in vagc dr72 . from vagc dr72
, we build up three volume - limited samples in which galaxies have absolute r - band magnitude brighter than @xmath21 , @xmath22 and @xmath23 respectively , ( @xmath2418 , @xmath2419 and @xmath2420 hereafter ) .
we consider all spectroscopic galaxies with listed apparent magnitude brighter than r = 18 .
all magnitudes mentioned in this paper are petrosian magnitudes , and are k - corrected to redshift @xmath25 . for @xmath2418 , @xmath2419 and @xmath2420 ,
the respective number of galaxies is @xmath26 , @xmath27 and @xmath28 .
we mainly focus on the @xmath2420 sample because its volume is large enough for accurate measurements of the large - scale structure geometry and it does not suffer galaxy bias severely since @xmath29 in the luminosity function is a little brighter than @xmath30 . the vagc provides us information on the size and the stellar mass for each galaxy , so that we can directly draw four parameters , concentration @xmath31 , @xmath32 color , stellar mass @xmath29 , and the surface stellar mass density @xmath33 , where @xmath34 and @xmath35 are defined as the radii containing @xmath36 and @xmath37 of the petrosian flux , respectively . in vagc , calculations for the stellar mass @xmath38 are based on five - band photometric data and redshift . @xcite compare this stellar mass measurement with spectroscopically based stellar mass measurements by @xcite , and show that the two results agree very well .
we also add the amplitude of the 4000 @xmath39 break d@xmath40(4000 ) into our analysis as a clear indicator of stellar population age , from the mpa / jhu sdss data ( http://www.mpa-garching.mpg.de/sdss/ ) .
there are @xmath41 galaxies in @xmath2420 that have counterparts in mpa / jhu data set .
nearest neighbor .
right panel : scatter plot of the environmental density @xmath42 and ellipticity @xmath0 .
the solid line is the average @xmath0 at different @xmath42 and the dashed lines show @xmath43 @xmath0 range around the average @xmath0 . , title="fig:",width=158 ] nearest neighbor .
right panel : scatter plot of the environmental density @xmath42 and ellipticity @xmath0 .
the solid line is the average @xmath0 at different @xmath42 and the dashed lines show @xmath43 @xmath0 range around the average @xmath0 . ,
title="fig:",width=158 ] $ ] for a slice of sdss dr7 . the environmental density and ellipticity
are encoded with two - dimensional color scheme shown in the insertion .
the smoothing scale is @xmath44.,width=317 ] from a theoretical point of view , the tidal environment of a galaxy can be characterized by the second derivatives of the gravitational potential , @xmath45 .
the potential can not , however , be measured directly , and needs to be calculated through the distribution of galaxies in redshift space .
such a measurement can be affected by , e.g. , galaxy bias , redshift - space distortion and survey geometry . here
we describe our procedures to calculate the potential field , leaving detailed discussion to section 4 . for calculations of the potential field
, we put the volume - limited sample into a large cubic box with side @xmath46 .
galaxies are assigned onto a @xmath47 uniform grid with cic interpolation to give the number density of galaxies in each cell .
we remove galaxies from those cells that happen to cross the survey boundary , and then assign the average number density of galaxies to all the empty cells . filling the regions outside the survey volume with the average number density
allows large voids with size comparable to the survey volume to be identified properly .
this galaxy number density field is then smoothed with a gaussian filter with a smoothing scale @xmath7 , and the potential field and further the tidal tensor field @xmath45 at each grid point are calculated from the smoothed density fluctuation field .
the tidal environment of each galaxy is obtained by linearly interpolating this discrete tidal tensor field from the mesh to its position . without considering the specific orientation of the tidal tensor ,
its properties are fully characterized by its three eigenvalues @xmath48 .
one way to classify the tidal environment of a galaxy uses the signs of @xmath49 ( e.g. , * ? ? ?
* ; * ? ? ?
specifically , one can define a point as ( i ) `` cluster '' , if all three eigenvalues are positive ; ( ii)``filament '' , if one eigenvalue is negative ; ( ii)``sheet '' , if two eigenvalues are negative ; and ( iv)``void '' , if all the three eigenvalues are negative .
these definitions can show clearly the filament - node nature of large - scale structure at least in a two - dimensional cut .
however , this classification is highly degenerate with a classification by density . as shown in the left panel of figure 1 , the overdensity distributions of points in these different categories are very different , and thus the dependence of galaxy properties on such `` tidal environments '' may mainly reflect a density dependence . to address the dependence of galaxy properties on tidal properties other than overdensity , we here describe the tidal environment of galaxies with parameters describing ellipticity @xmath0 and prolateness @xmath50 . specifically , we define @xmath0 and @xmath50 as follows @xmath51 where @xmath52 .
we find that these definitions , with the number @xmath53 added to the denominators , can minimize the correlations between @xmath0 ( @xmath50 ) and @xmath42 , as seen in the right panel of figure 1 .
figure 2 shows a 2-dimensional color rendering of @xmath54 $ ] in a 2-dimension cut through the @xmath2420 sample .
one can see that the cluster - filament - cluster structure is very well represented .
high density and high ellipticity bridges link high density and low ellipticity clusters . in the following analysis
, we mainly focus on the @xmath0-dependence of galaxy properties , since the @xmath42-dependence has been well explored in previous work .
as can be seen from figure 2 , giant filaments link clusters together .
these filaments have similar densities to the outskirts of clusters .
by introducing the additional parameter @xmath0 , however , these two regions can be separated clearly .
the main purpose of our study is to investigate if galaxy properties depend on the morphology of their large - scale environment in addition to the known density dependence . to do that
, we consider galaxy properties in the @xmath55 $ ] plane , and analyze the correlations of these properties with @xmath0 and @xmath56 .
specifically , we divide the @xmath55 $ ] plane into @xmath57 cells , where @xmath58 $ ] and the typical range of @xmath56 is @xmath59 with its precise range varying somewhat from one smoothing scale to another .
we then extend the plane to a periodic @xmath60 grid by reflecting the original cells with respect to @xmath61 and a line of the lowest value of @xmath56 .
at each cell , we calculate the total number of galaxies @xmath62 $ ] and the sum of the considered galaxy property @xmath63 $ ] .
smoothing is then applied to both @xmath64 and @xmath65 with a gaussian smoothing function @xmath66 , where subscript @xmath67 denotes that the smoothing scale is @xmath67 cells , and the average @xmath68 is obtained at each grid point by @xmath69 , where the subscript @xmath70 represents the smoothed quantities . to quantify the statistical significance of the @xmath0-dependence of galaxy properties , we need to estimate the statistical errors for the smoothed quantities properly .
our sample contains @xmath71 million galaxies .
when they are distributed onto @xmath57 cells in the @xmath55 $ ] plane , we expect that poisson fluctuations due to the limited number of galaxies associated with each cell will dominate the statistical errors .
we thus estimate the error in each cell as @xmath72 , where @xmath73 is the standard deviation of the considered galaxy property over the whole sample , and @xmath74 and @xmath75 are the smoothed cell count with the smoothing functions @xmath66 and @xmath76 , respectively . in table 1 , we list the values of @xmath73 for different galaxy properties for @xmath2420 .
.average value and standard deviation of different galaxy properties for @xmath2420 [ cols="<,^,^,^,^,^,^",options="header " , ] figure 3 shows the results for the color @xmath32 , where @xmath77 as derived from the sample @xmath2420 and shown in table 1 .
the three different panels correspond to three different smoothing scales to represent the large - scale environment , @xmath78 , @xmath79 and @xmath44 from left to right .
the @xmath32 values are shown in color in each panel with the scale shown to the right of figure 3 .
for the superposed white contour lines , the interval between the adjacent lines is @xmath80 .
the inside - out black lines show the error contours of @xmath80 , @xmath81 , @xmath82 and @xmath83 , respectively .
the dependence of @xmath32 on the density @xmath42 is clearly seen in figure 3 .
the slope of the @xmath32 contours reflects whether the ellipticity @xmath0 also affects the galaxy color in addition to @xmath42 .
we can see a weak correlation between @xmath32 and @xmath0 , and this correlation is scale - dependent . for @xmath78 ,
an anti - correlation is seen .
this reverses with increasing @xmath7 , and is positive for @xmath84 and larger . for @xmath85 ,
the correlation is nearly null . from figure 3
, we see that galaxy colors depend strongly on environmental density @xmath42 and weakly on environmental ellipticity @xmath0 .
this behavior are scale - dependent .
the question arises then what smoothing scale one should adopt in order to reveal the most fundamental dependence of galaxy properties on their environment .
previous studies have shown that galaxy properties are influenced by their environmental densities mainly locally @xcite .
on the other hand , if we simply smooth the environments on a scale larger than expected for any physical impact , we may still see an apparent environment - density dependence which arises merely due to the correlation between the environmental density on the considered smoothing scale and that on the scale where environmental effects are the strongest .
this can also apply to the apparent @xmath0-dependence of galaxy properties shown in figure 3 .
the scale - dependent switch from anti - correlation to positive - correlation with @xmath0 for galaxy colors seen in figure 3 indicates to us that the best smoothing scale should be around the transition scale where no @xmath0-correlations are observed .
furthermore , this transition scale should correspond to the scale where the density dependence of galaxy properties is the strongest . in appendix
a , we discuss further this scale - dependence of the environmental - density influence on galaxy properties .
we find that the scale where the environmental - density dependence of galaxy properties is the strongest is somewhat different for high and low density regions with the values of @xmath86 and @xmath87 , respectively .
this difference suggests that an adaptive smoothing scale may be more appropriate .
a natural choice is the characteristic distance between the considered galaxy and its neighbors .
our analysis show that for @xmath2420 sample , the distance to a galaxy s @xmath88-nearest - neighbor is an optimal adaptive smoothing scale , which has a distribution peaked at @xmath89 and is close to the peak scale where galaxy properties and environment density have the strongest correlation in both high and low density regions . for @xmath2419 and @xmath2418 samples
, this optimal scale should correspond to the distance to a galaxy s @xmath90-nearest- neighbor and @xmath91-nearest - neighbor respectively , due to the fact that the galaxy density is higher . in figure 4
, we show @xmath32 contours as in figure 3 .
the right and left panels are for fixed smoothing scales of @xmath92 and @xmath93 , respectively .
the middle panel shows results using the @xmath88-nearest - neighbor distance as an adaptive smoothing scale . in comparison with the middle panel of figure 3 ,
the null - dependence on @xmath0 is more cleanly seen for the adaptive smoothing . to investigate
if the dependence on @xmath0 seen in the left and right panels of figure 4 ( the same as those of figure 3 ) is physical or due to the correlations between @xmath94 at the considered scale and @xmath42 at the ` optimal adaptive smoothing scale ' , we calculate the latter correlations .
specifically , we define @xmath94 at the considered smoothing scale and at the optimal adaptive smoothing scale as @xmath95 and @xmath96 , respectively .
we compute the @xmath97-contours in the @xmath95 plane , which are shown as black lines in the left and right panels of figure 4 .
it is seen that the @xmath97-contours align with the @xmath32-contours ( white lines ) very well .
this demonstrates that the observed @xmath0-dependence is mainly attributable to the correlation between @xmath95 and @xmath97 .
some deviation between the @xmath32-contours and the @xmath97-contours are seen in high density regions .
these are , however , insignificant compared to the statistical errors .
together with the null - correlation with @xmath0 for the optimal adaptive smoothing , we conclude that no significant physical dependence of galaxy colors on @xmath0 is detected by our analysis .
we have studied the dependence of galaxy color on @xmath94 carefully . here
we examine the tidal dependence of other galaxy properties .
galaxy properties can be divided into three classes : ( 1 ) star - formation - related properties such as galaxy color and d@xmath40(4000 ) ; ( 2 ) morphology - related properties such as concentration ; and ( 3 ) extensive quantities such as the total stellar mass @xmath29 and galaxy size .
previous studies have shown that star - formation - related properties depend directly on the environmental density @xcite , but morphology - related ones are not independently correlated with environmental density @xcite . here
we analyze the @xmath94-dependence of different galaxy properties .
figure 5 presents the environmental dependence of @xmath98 .
the left panel shows a scatter plot of @xmath98 .
the red lines are contours of constrained mean value of @xmath99 . from the scatter plot , we see that @xmath100 , which reflects stellar population ages and therefore star formation histories , correlates with total stellar mass @xmath38 .
the directions of the red contours indicate that both depend on environmental density as measured by our optimal adaptive smoothing . in the right panel of figure 5
, we show the @xmath94 dependence for @xmath100 given the stellar mass @xmath38 . specifically , the contours of the deviation @xmath101 are shown , where @xmath102 is the average value of @xmath100 in stellar mass bins .
the environmental - density dependence of the contours are clearly seen , which , in agreement with the trend seen in the left panel , demonstrates the extra density dependence of @xmath100 in addition its dependence of @xmath38 .
on the other hand , no dependence on the shape parameter @xmath0 is detected . in figure 6
, we show a similar analysis for the concentration parameter @xmath31 .
the scatter plot of @xmath103 is shown in the left panel with the @xmath99 contours superposed on it .
two clumps are seen in the scatter plot with one at low value of @xmath100 and small value of @xmath64 , and the other at high @xmath100 and large @xmath64 .
the lower - left clump corresponds to galaxies with recent star formation , and the upper - right clump is associated with galaxies with quenched star formation .
the @xmath99 contour lines are nearly vertical , showing no detectabe independent correlation between @xmath64 and the environmental density . in the right panel of figure 6 , we show the @xmath104 contours in the @xmath105 $ ] plane , where @xmath106 is the average value of @xmath64 given @xmath100 .
the chaotic pattern seen here reflects the lack of any environmental dependence of the concentration @xmath64 other than its @xmath97 dependence through the correlations with @xmath100 .
, d@xmath40(4000 ) ) .
the @xmath99 contours are shown in red .
right panel : the contours of d@xmath40(4000)-@xmath107d@xmath40(4000)@xmath108 , where @xmath107d@xmath40(4000)@xmath108 is the average value of d@xmath40(4000 ) given stellar mass.,title="fig:",width=158 ] , d@xmath40(4000 ) ) .
the @xmath99 contours are shown in red .
right panel : the contours of d@xmath40(4000)-@xmath107d@xmath40(4000)@xmath108 , where @xmath107d@xmath40(4000)@xmath108 is the average value of d@xmath40(4000 ) given stellar mass.,title="fig:",width=158 ] ( 4000 ) , @xmath64 ) .
the @xmath99 contours are shown in red .
right panel : contours of @xmath104 , where @xmath106 is the average value of @xmath64 given d@xmath40(4000).,title="fig:",width=158](4000 ) , @xmath64 ) .
the @xmath99 contours are shown in red .
right panel : contours of @xmath104 , where @xmath106 is the average value of @xmath64 given d@xmath40(4000).,title="fig:",width=158 ] in unit of for sdss .
the boundary is the 0.2@xmath73 uncertainty contour , where @xmath73 is the standard deviation of galaxy size for the whole sample .
the optimal adaptive smoothing is applied.,width=317 ] we also analyze the environmental dependence of galaxy size defined as @xmath34*d in unit of h@xmath109kpc , where d denotes the distance to a galaxy .
the results are shown in figure 7 , again using our optimal adaptive smoothing .
it is seen that in high density regions , galaxy size has no detectable correlation with either environmental density or ellipticity . in low density regions , on the other hand , there is a notable correlation between galaxy size and @xmath110 in addition to its correlation with @xmath97 .
we will show in section 4.3 that in low density regions , the survey boundary effects can bias the @xmath97 and @xmath110 measurements to lower values .
this can induce an artificial correlation with @xmath110 , which is considerable for small - volume samples such as @xmath2418 . for @xmath2420 , however , the survey volume is large and the boundary effects are insignificant .
thus the correlations seen in the lower left region of figure 7 may indicate that both the density and the ellipticity of the environment , can affect the galaxy size .
galaxies are systematically smaller in lower @xmath111 ) regions . due to the still limited statistics of sdss , the environmental @xmath110 dependence of galaxy size needs to be investigated further with future large surveys that can provide much improved statistics .
our analysis so far has measured ellipticity and density on the same scale .
the optimal adaptive smoothing is chosen to be the distance from a galaxy to its 3rd - nearest neighbor .
this scale has a distribution peaked between 1.5h@xmath109mpc and 2.5h@xmath109mpc , depending on environmental density . in order to show the influence of larger scales , especially the influence of the large filaments linking cluster as seen in figure 2
, we also analyze the environmental dependence of galaxies properties on ellipticity and density measured on two different scales .
we fixs the smoothing for density to the optimal adaptive smoothing and we change the smoothing scale for ellipticity . for no smoothing scale larger than the 2h@xmath109mpc
do we see a correlation of galaxy properties with environment ellipticity .
we thus conclude that large scale structure beyond galaxy groups has no additional influence on individual galaxies .
we have so far focused our analysis on the ellipticity @xmath0 of large - scale environment .
however , we have also looked for the dependence of galaxy properties on the environmental prolateness @xmath50 .
we find that smoothing according to the distance to the @xmath88-nearest - neighbor of a galaxy , all of the considered galaxy properties are independent of @xmath50 .
in semi - analytic models of galaxy formation ( sam ) , our best understanding on dark matter halo formation is combined with the complex baryonic physics associated with galaxy formation as represented by physically / observationally motivated prescriptions involving a set of parameters .
while earlier sams were based on using the extended press - schechter theory to construct dark matter halo merger trees , modern sams take full advantage of high resolution n - body simulations to follow the formation and evolution of dark matter halos .
galaxies are then formed within these halos by modeling physical processes such as cooling , condensation , star formation , supernova and agn feedback . here
we analyze the sam galaxy catalog from @xcite , for which the dark matter distribution was taken from the millennium simulation @xcite .
we compare the results with those from the sdss . to test observational effects
, we also construct mock sdss dr7 catalogs from the sam data @xcite .
specifically , we address the following issues 1 . a comparison of the environmental - density dependence of galaxy properties in the sam and the sdss ; 2 . a comparison of the behavior of the tidal dependence of galaxy properties ; 3 . the influence of various observational effects , such as redshift - space distortion and survey geometry on the results of our tidal analysis . to analyze the scale dependence of the color - density relation , we calculate the correlation coefficient between galaxy color @xmath32 and environmental density @xmath42 at different smoothing scales
the correlation coefficient is defined as @xmath112 ^ 2\rangle}\sqrt{\langle(\delta \delta_o)^2\rangle}},\ ] ] where @xmath113 and @xmath114 .
the results are shown in figure 8 .
we divide galaxies into two groups of high ( solid lines ) and low densities ( dashed lines ) as described in appendix a ( the green lines in figure a1 ) .
the red , blue and green lines are , respectively , for the sdss , the full volume of the sam , and for the mock sdss .
first , we see similar trends for the scale dependence of the color - density correlation in the sam and the sdss data .
the sam ` optimal smoothing scale ' for the strongest correlation is also consistent with that of sdss , which is @xmath115 for high density regions and @xmath116 for low density regions . on the other hand ,
the amplitude of the correlation coefficient is significantly larger for the sam than for the sdss .
the sam predicts an environment - density dependence of galaxy colors stronger than is observed , in accordance with other studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we note that while the tidal interactions of dark matter halos are naturally accounted for in the sam , tidal disruption of galaxies is not . for massive clusters ,
however , it is expected that their strong tidal forces may destroy galaxies ( e.g. , * ? ? ?
as those satellite galaxies are mostly red , such disruptions can decrease the level of clustering of red galaxies , and therefore reduce the discrepancies in color - density relation between the sam and the observations in high density regions .
further , this can also alleviate the overly - fast - growth problem for central galaxies in the sam @xcite .
it has also been argued that hot gas in satellite galaxies may not be immediately stripped away by ram pressure when they merge into a larger system as often assumed in sams .
therefore a reservoir of gas can be replenished to maintain the star formation for a relatively long period of time ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in low density regions ,
the sam also predicts stronger color - density correlations than seen in the sdss , which reflects the smaller color spread in the sam compared to the sdss .
the green lines in figure 8 are the results from the sdss dr7 mock catalogue constructed from sam data .
the differences between the green and blue lines reflect observatioal effects which will be discussed in 4.3 .
to study the tidal dependence of galaxy properties , we perform the same analysis for the sam data as shown in 3.2 for the sdss .
the results are shown in figure 9 .
the black lines here represent results from the mock catalogue , and will be discussed in 4.3 .
the three panels correspond to the smoothing with @xmath78 , with the @xmath88-nearest - neighbor distance , and with the @xmath117 .
it is seen that the contour behaviors are qualitatively similar to those of figure 4 for the sdss . in the sam
, the distance to the @xmath88-nearest neighbor also corresponds to the peak scale of the color - density dependence , at which , the tidal @xmath0-dependence of galaxy properties is nearly null .
the density @xmath97 at the optimal adaptive smoothing scale plays the dominant role for the environmental effects .
the weak negative / positive @xmath32-@xmath0 correlations at smaller / larger smoothing scales again reflect correlations between @xmath94 at the smoothing scale and @xmath97 .
a notable difference between figure 9 and figure 4 is the density dependence of color , which is significantly stronger in figure 9 than that shown in figure 4 .
this is consistent with the result shown in figure 8 , which also show that the sam predicts a considerable stronger color - density correlation than seen in observations . to make a realistic comparison between the results for the sam and the sdss ,
different observational effects must be taken into account .
we thus build mock catalogs from the sam data following @xcite .
we first test the influence of fiber collisions which reduce the number of redshift measurements for galaxies in close pairs by about @xmath118 .
because our smoothing scale is nearly always larger than @xmath119 , fiber collisions do not affect our analysis significantly .
the most notable influences from obervations are the redshift distortion effect and the effect from the survey geometry .
the redshift - space distortion affects mainly the high density regions and plays a role of `` mixing '' . in redshift space ,
galaxies in massive clusters present a finger - of - god configuration along the line of sight . on the other hand ,
galaxies in filaments surrounding a massive cluster have a tendency to be flowing into the cluster , which squashes the filaments along the line of sight .
in other words , galaxies in filaments and galaxies in massive clusters are somewhat mixed in redshift space .
thus the color scatter in a given density bin becomes larger in redshift space , resulting in a lower color - density correlation coefficient in high density regions , as shown by the green solid line in figure 8 .
furthermore , the redshift - space distortion makes the measured environmental density systematically lower than that measured in real space .
the ellipticity calculated in the redshift space is systematically lower / higher for regions with high / low @xmath0 in real space .
that is , filaments become rounder and clusters tend to be more anisotropic in redshift space . such effects cause color mixing in the @xmath120 $ ] plane , and thus lead to flatter color contours with respect to the ellipticity @xmath0 and weaker correlations with @xmath42 ( see the black contour lines in figure 9 ) . in low density regions , measurements of environmental density and ellipticity
are mainly affected by survey geometry .
both are biased to lower values . in the left panel of figure 10 , we show the value of @xmath0 with respect to the distance to the survey boundary for sdss sample @xmath2420 .
different lines represent different environmental density ranges with longer dashes corresponding to higher densities .
physically , we do not expect any correlations between @xmath0 value and the distance to the survey boundary . for high - density lines , they are indeed nearly flat . for low - density lines , however , the measured @xmath0 is systematically lower for galaxies closer to the survey boundary .
the effect is significant for galaxies that are within @xmath121 of the boundary .
this is clearly an observational effect due to padding the regions outside the survey boundary with the average number density .
such a measurement bias on @xmath0 can lead to artificial correlations between galaxy properties and environmental ellipticity .
the significance of the effect depends on the the fraction of galaxies that are within @xmath121 of the boundary .
volume - limited samples with smaller volume are affected more . in the right panel of figure 10
, we show the ratio of color in high @xmath0 regions to the color in low @xmath0 regions vs. minus the absolute magnitude of galaxies -@xmath24 . here
high and low @xmath0 are defined as above the @xmath122 level and below the @xmath123 level with respect to the average @xmath0 ( see the right panel of figure 1 ) .
we present results for different sdss volume - limited samples , @xmath2418 , @xmath2419 and @xmath2420 from smaller to larger volume .
for @xmath2418 , we see the decrease of the color ratio with respect to -@xmath24 of galaxies .
that is , for relatively faint galaxies , the ones in low @xmath0 regions tend to be bluer than those in high @xmath0 regions . for bright galaxies ,
the ratio is close to @xmath124 . for @xmath2419 and @xmath2420
, there is no detectable dependence of the color ratio on -@xmath24 , and the ratio is always close to @xmath124 . to test
if the @xmath24-dependence of the color ratio seen in @xmath2418 is physical or due to the boundary effect , we construct subsamples of galaxies with distances less than @xmath125 from @xmath2419 and @xmath2420 , respectively . this cut corresponds roughly to the distance limit of @xmath2418 .
the results for the @xmath2419 and @xmath2420 subsamples are shown by unfilled green triangles and unfilled red squares , respectively . the apparent correlation with @xmath24 of galaxies
is then seen .
this demonstrates that the correlation seen for @xmath2418 sample is largely due to the boundary effect . as shown in the left panel of figure 10 , in low density regions , the @xmath0 value is biased to a lower value for galaxies closer to the survey boundary . because galaxies in low density regions tend to be bluer than those in high density regions , this bias leads to an increase in the number of relatively blue galaxies in low @xmath0 regions , and thus increases the color ratio between high @xmath0 and low @xmath0 regions .
the increase of the color ratio is more significant for fainter galaxies because the observed ones reside in a smaller volume around the observer and thus suffer more from the boundary effect . in figure 11
, we show color contours in the @xmath126 $ ] plane for @xmath2418 . for @xmath2418 ,
the optimal adaptive smoothing scale for a galaxy corresponds to the distance to its @xmath127th nearest neighbor . comparing to the middle panel of figure 4 , we see a notable correlation between color and ellipticity in low - density regions .
this however , is largely due to the survey boundary effect that biases both the environmental density and ellipticity toward lower values in low density regions .
ranging from @xmath128 to @xmath129 .
longer - dashed lines represent the results for higher density bins .
right panel : the ratio of the mean @xmath32 in high ellipticity regions to that in low ellipticity regions for @xmath24 20 ( red filled circle ) , @xmath2419 ( green filled circle ) and @xmath2418 samples ( blue filled circle ) , respectively .
the high and low ellipticity regions are defined as the regions above the @xmath130 dashed line and below the @xmath131 dashed line shown in the right pane of figure 1 .
the results for subsamples of @xmath2420 and of @xmath2419 cut at @xmath125 are shown by red open squares and green open triangles , respectively .
, title="fig:",width=158 ] ranging from @xmath128 to @xmath129 .
longer - dashed lines represent the results for higher density bins .
right panel : the ratio of the mean @xmath32 in high ellipticity regions to that in low ellipticity regions for @xmath24 20 ( red filled circle ) , @xmath2419 ( green filled circle ) and @xmath2418 samples ( blue filled circle ) , respectively . the high and low ellipticity regions are defined as the regions above the @xmath130 dashed line and below the @xmath131 dashed line shown in the right pane of figure 1 . the results for subsamples of @xmath2420 and of @xmath2419 cut at @xmath125 are shown by red open squares and green open triangles , respectively . ,
title="fig:",width=158 ] 18 . the environmental smoothing scale for a galaxy
is taken to be the distance to its @xmath91-nearest neighbor .
the boundary is the @xmath83 uncertainty contour , where @xmath73 = 0.168 is the standard deviation of galaxy color of the whole @xmath2418 sample .
, width=317 ]
in this paper we examine the tidal dependence of galaxy properties in the nyu - vagc sample and compare it with the predictions of the sam . to separate the environment - morphology dependence from the environmental - density dependence , we construct the density fluctuation @xmath42 , the ellipticity @xmath0 and the prolateness @xmath50 from the three eigenvalues of the tidal tensor of the potential field calculated from the spatial distribution of galaxies , where @xmath0 and @xmath50 are nearly independent of @xmath42 . it should be noted that this potential field can be biased with respect to the potential field from the underlying dark matter distribution .
the well - known galaxy bias refers to the ratio of the density fluctuation amplitude of galaxy distribution and that of the dark matter distribution . for @xmath2420 , the volume - limited sample with absolute magnitude brighter than @xmath30 ,
our results show that while the environmental - density dependence of galaxy properties is indisputable , the dependence on @xmath0 and @xmath50 is rather weak , except perhaps for galaxy size .
it is further shown that such weak correlations depend on the smoothing scale considered for the environment .
there exists a particular scale , @xmath132 , at which , the environmental-@xmath0 ( @xmath50 ) dependence nearly vanishes .
this corresponds to the optimal adaptive smoothing scale where the environmental - density dependence of galaxy properties is the strongest .
it varies from high density regions to low density regions , and corresponds well to the distance to each galaxy s @xmath88-nearest - neighbor for @xmath133 sample . at smoothing scales larger / smaller than this `
optimal smoothing scale ' , the weak correlations between galaxy properties and @xmath0 are negative / positive .
we demonstrate , however , that these correlations result mainly from correlations between @xmath94 at the considered smoothing scale and the environmental density @xmath97 at the optimal adaptive smoothing scale . in other words ,
we find no physical influence of environment morphology on galaxy properties .
furthermore , it is @xmath97 that plays the dominant role for the environmental effects , and the apparent density dependence on other smoothing scales is largely due to the correlations between @xmath42 and @xmath97 .
this indicates that , galaxy properties in @xmath2420 are affected mainly by their nearby environments .
our analysis shows that galaxy size is independent of environment in high - density regions . in low - density regions ,
some correlation with @xmath110 in addition to the dependence on @xmath97 is detected .
galaxies in lower @xmath134 regions tend to be smaller .
such correlations need to be further explored with future observations . with nyu - vagc @xmath2419 ,
@xcite analyze galaxies in groups , and find that galaxy properties depend mainly on the properties of their host groups , and are not affected independently by the environment on larger scales(@xmath135 ) . as they only consider galaxies in groups
, it is appropriate to compare their results with ours in high density regions . from figure 8 , we can see that the ` peak scale ' in high density regions is @xmath115 , consistent with the typical group scale in their studies .
therefore our results are in agreement with theirs concerning the dominant importance of the nearby environment .
this in turn validates the current halo occupation distribution model , which assumes that galaxy properties depend only on those of their parent halos with no additional influence from larger scale .
@xcite also examines the tidal dependence of galaxy properties .
they found that there is a tendency for elliptical galaxies to be preferentially situated in low @xmath0 regions with density in the range @xmath136 . for @xmath137 ,
spiral galaxies are likely in regions with large @xmath0 .
we note that their considered scale for environmental effects is @xmath138 .
thus their results can loosely be compared with our analysis for @xmath139 ( but note the different @xmath42 range ) . with the correspondence between color and morphology of galaxies , the positive correlation of galaxy color with @xmath0 seen in right panel of figure 4
is qualitatively consistent with the trend found by @xcite although their definition of @xmath0 is different from ours . on the other hand
, our analysis indicates that this @xmath0-dependence is largely due to the correlations between @xmath0 and @xmath97 and therefore has little physical significance . comparing the sdss results with those from our sam , we see qualitative agreement between the two ,
although the sam predicts a stronger dependence of galaxy properties on environmental density .
our results are also in accordance with the theoretical study of @xcite , which showed that environmental effects on dark matter halo formation mainly reflect environmental density , and are influenced little by @xmath0 and @xmath50 .
@xcite found that the anti - correlation between the formation epoch of galactic halos and their environment density is , in fact , mainly attributable to tidal suppression by neighboring massive halos . because of the enhanced environment density near massive halos , this suppression shows up as an assembly bias , and provides a possible explanation for galaxy variation with environment density . on the other hand , @xcite chose the smallest eigenvalue @xmath140 of the tidal tensor as an indicator of the strength of tidal field , and they found a stronger dependence of halo formation epoch on @xmath140 than on environment density .
our results manifest a strong degeneracy between @xmath140 and environment density . just as for ellipticity , @xmath140 can be geometrically more strongly correlated with environment density on the optimal smoothing scale than on a larger scale .
further work is needed to clarify whether the dependence of halo formation history on environment morphology can be attributed to this geometry effect , or whether baryon physics actually make difference between halos and galaxies . by constructing mock catalogs from sam data , we have examined the influence of observational effects on our analysis .
redshift - space distortion is significant in high density regions .
it mixes galaxies in filaments and galaxies in clusters causing an increase of color scatter at given density , and therefore reduces the color - density correlation coefficient considerably . the survey boundary affects environment measurements mainly in low density regions , biasing both the environmental density and the ellipticity to lower values .
the effect is notable for galaxies within @xmath121 of the survey boundary .
while it is negligible for @xmath2420 , the effect is considerable for @xmath2418 which occupies a relatively small volume and therefore has a large fraction of its galaxies close to the boundary .
it induces an artificial correlation between galaxy properties and the environmental ellipticity for @xmath2418 . in summary
, our analysis of sdss data shows that in addition to environment density , there is no significant further dependence of galaxy properties on the morphology of large scale structure .
geometrically , both ellipticity and density on one smoothing scale correlate strongly with each other , and with ellipticity / density on other smoothing scales .
if the smoothing scale is not chosen properly , an appeared dependence of galaxy properties on both environment ellipticity and environment density arises which is merely due to geometry .
we find that for the optimal adaptive smoothing scale , the dependence on density is maximized and the dependence on ellipticity and prolateness is null .
we thank cristiano porciani and guinevere kauffmann for detailed discussions in the early stages of this project , and cheng li for his help on dealing with the sdss data .
this research is supported in part by the nsfc of china under grants 10773001 and 11033005 , 11173001 and the 973 program no .
2007cb815401 , as well as by the erc under advanced grant 246797 `` galformod '' . 99 abazajian k. , adelman - mccarthy j.k . ,
agueros , m. , 2009 , apjs , 182 , 543 ball n.m . , loveday j. , brunner r.j . , 2008 , mnras , 383 , 907 bond j.r . , myers s.t . , 1996 , apjs , 103 , 1 bamford s.p . , rojas , a.l . , nichol r.c . , 2008 , mnras , 391 , 607 blanton m.r . , schlegel d.j . ,
strauss m.a . , 2005 ,
apj , 128 , 2562 blanton m.r . ,
eisenstein d. , hogg d.w . , 2005 , apj , 629 , 143 blanton m.r . , berlind a.a . , 2007 ,
apj , 664 , 791 bond j.r .
, efstathiou g. , kaiser n. , 1991 , apj , 379 , 440 christlein d. , zabludoff a. , 2005 , apj , 621 , 201 coil a. , newman j.a . &
croton d. , 2008 , apj , 672 , 153 cowan n.b .
, ivezic . , 2008 , apj , 674 ,
13 dalal n. , white m. , bond j. r. , shirokov a. , 2008 , apj , 687,12 dekel a. birnboim y. , 2006 , mnras , 368,2 de lucia g. , blaizot j. , 2007 , mnras , 375 , 2 desjacques v. , 2008 , mnras , 388 , 638 diemand j. , 2007 , apj , 667 , 859 dressler , a. , 1978 , apj , 223,765 croton d.j .
, gao l. , white s.d.m . , 2007 , mnras , 374 , 1303 gao l. , springel v. , white s.d.m . , mnras , 363 , 66 gao l. , white s.d.m . , 2007 , mnras , 377 , 5 gunn j.e .
, siegmund w.a . , mannery e.j . , 2006 ,
aj , 131 , 2332 guo , q. , white , s.d.m . , boylan - kolchin , m. , & de luica , g. , 2011 , mnras , 413 , 101 frenk c.s . , white s.d.m
, bode p. , apj , 525 , 554 font a.s .
, bower r.g .
, mccarthy i.g . , mnras , 389 , 1619 hahn o. , corollo m. , porciani c. , 2007 , mnras , 381,41 hahn o. , porciani c. , dekel a. , 2009 , mnras , 398 , 1742 henriques b. m. , bertone , s. , thomas p.a . , 2008 , mnras , 383 , 1649 hubble e.p .
, 1936 , the realm of the nebulae .
new haven , ct : yale univ .
press harker g. , cole s. , helly j. , 2006 , mnras , 367 , 1039 jing y.p .
, suto y. , mo h.j . , 2007 , apj , 657 , 665 kauffmann g. , heckman t.m .
, white s.d.m .
kauffmann g. , white s.d.m . ,
heckman t. m. , 2004 , mnras , 353 , 713 keres d. , katz n. , weinberg d.h . , 2005 , mnras , 363 , 2 keselman j.a . ,
nusser a. , 2007 , mnras , 382 , 1853 kim h.s . ,
baugh c.m . , cole s. , 2009 , mnras , 400 , 1527 lacey c. , cole s. , 1993 , mnras , 262 , 627 lee j. , lee b. , 2008 , apj , 688 , 78 li c. , kauffmann g. , jing y.p , 2006 , mnras , 368 , 21 li c. , jing y.p , kauffmann g. , 2007 , mnras , 376 , 984 li c. , white s.d.m . , 2009 , mnras , 398 , 2177 ma c. p. , maggiore m. , riotto a. , zhang j. , 2010 , mnras in press ( arxiv:1007.4201 ) mo , h. j. & white , s.d.m . , 1996 , mnras , 282 , 347 mori m. , burkert a. , 2000 , apj , 538 , 559 park c. , choi y.y . , vogeley m.s . , 2007 , apj , 658 , 898 sandvik h.b . , mller sheth r.k . , mo h.j . , tormen g. , 2001 , mnras , 323 somerville r.s . , lemson g. , sigad y. , 2001 , mnras , 320 , 289 springel v. , white s.d.m .
, jenkins a. , 2005 , nature , 435 , 629 oemler a. , 1974 , apj , 194 , 1 padmanabhan n. , schlegel , d.j . , finkbeiner d. p. et al . , 2008 , apj , 674 , 1217 postman m. , geller m.j . , 1984 , apj , 281 , 53 tinker j.l . , conroy c. , norberg p. , 2008 ,
apj , 686 , 53 york d.g . , adelman j. , anderson j.e . , 2000 ,
aj , 120 , 1579 wang h.y . , mo h.j . , jing y.p
, 2007 , mnras , 375 , 633 wang l. , jing y.p . , 2010 , mnras , 402 , 1796 wechsler r.h . , zentner a.r . , bullock j.s .
white s. d. m. , 1996 , in schaeffer r. , silk j. , spiro , m. , zinn - justin j. , eds , les houches session lx , cosmology and large sclae structure .
elsevier , amsterdam , p.77 weinmann s.m .
, van den bosch f.c .
, yang x.h . , mo h.j . , 2006 , mnras , 366 2 zhang y.c .
, yang x.h . , faltenbacher a. , 2009 , apj , 706 , 747 [ lastpage ]
previous studies have shown that age - related properties , such as galaxy color , are correlated with environmental density and that this correlation is scale dependent @xcite . here
we examine the scale dependence of the color - density relation quantitatively and in more detail . for each galaxy , we calculate its environmental density at two different smoothing scales , denoted as @xmath141 and @xmath142 .
we then analyze the dependence of galaxy color on @xmath143 .
the results are shown in figure a1 , where @xmath144 is taken to be the reference scale . in each panel ,
the black dots show the values of @xmath145 $ ] for all the galaxies , and the red lines are @xmath32 contours .
the green line denotes our dividing line for high and low density regions , which is the @xmath146 line passing through the point of @xmath147 , where @xmath148 is the corresponding value of @xmath149 given @xmath150 .
it is seen that for @xmath151 , the color - density dependence , @xmath152 , is mainly on @xmath153 , and the additional dependence on @xmath142 is very weak . for @xmath154 , a certain level of @xmath142-dependence
is seen in low density regions . for @xmath155
, on the other hand , the @xmath142-dependence appears in high density regions .
these results clearly demonstrate the scale dependence of the color - density relation .
there exists a special scale where the environmental density affects galaxy colors the most .
the value of this ` optimal adaptive smoothing scale ' is @xmath116 for @xmath2420 , and varies somewhat from high to low density regions .
our detailed analysis show that the variation is from @xmath156 in high density regions to @xmath157 in low density regions , consistent with that shown in figure 8 for the scale dependence of the color - density correlation coefficient .
our further studies find that a better way to quantify the environmental effects on a galaxy is to use adaptive smoothing , for example distance to its n@xmath158-nearest neighbor , which takes into account the variations of the environmental density naturally .
we have tested the distributions for the n@xmath158-nearest neighbor distances , where @xmath159 are considered .
we notice that the distance to the @xmath88-nearest neighbor for @xmath133 sample is peaked at @xmath89 with a range mainly between @xmath10 and @xmath44 , therefore it represents the ` peak scale ' suitably and adaptively , and we propose it as the optimal adaptive smoothing scale for the @xmath133 sample .
the corresponding optimal adaptive smoothing scales for the @xmath160 and @xmath161 samples are respectively 6th - nearest and 8th - nearest neighbors . | using volume - limited samples drawn from the sloan digital sky survey data release 7 ( sdss dr7 ) , we measure the tidal environment of galaxies , which we characterize by the ellipticity @xmath0 of the potential field calculated from the smoothed spatial number density @xmath1 of galaxies .
we analyze if galaxy properties , including color , d@xmath24000 , concentration and size correlate with @xmath0 , in addition to depending on @xmath1 .
we find that there exists a transition smoothing scale at which correlations / anti - correlations with @xmath0 reverse .
this transition scale is well represented by the distance to the 3rd - nearest - neighbor of a galaxy in a volume limited sample with @xmath3 which has a distribution peaked at @xmath4 @xmath5mpc .
we further demonstrate that this scale corresponds to that where the correlation between the color of galaxies and environmental density @xmath1 is the strongest . for this optimal smoothing @xmath6
no additional correlations with @xmath0 are observed . the apparent dependence on tidal ellipticity @xmath0 at other smoothing scales @xmath7 can be viewed as a geometric effect , arising from the cross correlation between @xmath8 and @xmath9 .
we perform the same analysis on numerical simulations with semi - analytical modeling ( sam ) of galaxy formation .
the @xmath0 dependence of the galaxy properties shows similar behavior to that in the sdss , although the color - density correlation is significantly stronger in the sam .
the ` optimal adaptive smoothing scale ' in the sam is also closely related to the distance to the 3rd - nearest - neighbor of a galaxy , and its characteristic value is consistent with , albeit slightly smaller than for sdss .
[ firstpage ] galaxies : evolution - galaxies : formation - galaxies : stellar content |
during last one decade , theoretical research on co - evolution of species in eco - systems and the statistics of extinctions have been strongly infleunced by the pioneering interdisciplinary works of per bak and his collaborators @xcite . in the same spirit , we address some fundamental questions of evolutionary ecology from the perspective of statistical physics . how did the higher species emerge in eco - systems inhabited initially only by primitive forms of life , like bacteria and plankton @xcite ?
the available record of the history of life , written on stone in the form of fossils , is incomplete and ambiguous .
an alternative enterprise seeks to recreate the evolution on a computer by simulating theoretical models @xcite . in this paper
we propose a theoretical model that not only addresses the question raised above but also provides a versatile conceptual tool for studying evolutionary ecology .
in particular , it describes both the `` macro''-evolutionary processes ( e.g. , origin , evolution and extinction of species ) as well as `` micro''-evolution , ( e.g. , age - distribution in the population of a species , mortality rates , etc . ) .
if watched over a short period of time , the dynamics of the eco - system appears to be dominated by birth and death of the individual organisms as well as by the prey - predator interactions .
however , over longer period of time , one would see not only extinction of some species but also the appearance of new ones . besides , in many situations , macro - evolutionary changes occur at rates that are comparable to those of the ecological processes @xcite .
the artificial separation of this process into `` ecological '' time scales and `` geological '' time scales @xcite has been made in many earlier theoretical works only for the convenience of modelling .
the `` ecological '' models , that describe population dynamics in detail using , for example , the lotka - volterra equations @xcite usually ignore the slow macro - evolutionary changes in the eco - system ; hardly any effects of these would be observable before the computer simulations would run out of computer time @xcite . on the other hand , in order to simulate the billion - year old history of life on earth with a computer , the elementary time steps in `` evolutionary '' models have to correspond to thousands of years , if not millions ; consequently , the finer details of the ecological processes over shorter periods of time can not be accounted for by these models in any explicit manner @xcite
. limitations of these approaches are well known @xcite .
moreover , most of the recent computer models of ageing @xcite focus attention on only one isolated species and , therefore , can not capture macro - evolutionary phenomena like , for example , extinctions which depend crucially on the prey - predator interactions .
we wish to develop one single theoretical model which would be able to describe the entire dynamics of an eco - system since the first appearance of life in it up till now and in as much detail as possible .
this dream has now come closer to reality , mainly because of the availability of fast computers @xcite .
it has become feasible now to carry out computer simulations ( in - silico experiments ) of eco - system models where , each time step would correspond to typical times for `` micro''-evolution while each of the simulations is run long enough to capture `` macro''-evolution .
the prey - predator relations in any eco - system are usually described graphically in terms of food webs @xcite .
more precisely , a food web is a directed graph where each node is labelled by a species name and each directed link indicates the direction of flow of nutrient ( i.e. , _ from _ a prey _ to _ one of its predators ) .
we incorporate in our model the hierarchical organization of the species at different trophic levels of the food web . in real eco - systems ,
the food web is a slowly evolving dynamic network .
for example , species are known to change their food habits @xcite .
these changes in diets may be caused by scarcity of the normal food and abundance of alternative food resources .
moreover , higher organisms appear through speciation in an eco - system that initially began with only simple forms of life .
these not only occupy new trophic levels but also introduce new prey - predator interactions with the existing species .
therefore , it is also desirable that these _ self - organizing _ features of natural eco - systems should be reproduced , at least qualitatively , by the theoretical models . the aim of this paper is to propose a model that would capture the desirable features of eco - systems outlined above .
higgs , mckane and collaborators @xcite have developed a model , called the webworld model , which was aimed at linking the ecological modeling of food web architecture with the evolutionary modeling of speciation and extinction .
the spirit of our model is very similar although the details of the mathematical formulation of the two models are quite different .
we model the eco - sytem as a dynamic _ hierarchical _ network .
the `` micro''-evolution , i.e. , the birth , growth ( ageing ) and natural death of the individual organisms , in our model is captured by the intra - node dynamics .
the macro-evolution , e.g. , adaptive co - evolution of the species , is incorporated in the same model through a slower evolution of the network itself over longer time scales .
moreover , as the model eco - system evolves with time , extinction of species is indicated by vanishing of the corresponding population ; thus , the number of species and the trophic levels in the model eco - system can fluctuate with time .
furthermore , the natural process of speciation is implemented by allowing re - occupation of vacant nodes by mutated versions of non - extinct species .
each node of the network represents a niche that can be occupied by at most one species at a time .
the number of nodes in the trophic level @xmath0 is @xmath1 where @xmath2 is a positive integer .
we assume only one single species at the highest level @xmath3 .
the allowed range of @xmath0 is @xmath4 , where @xmath5 is a time - dependent number in our new model . in other words , in contrast to all cited earlier models , the numerical value of @xmath6 in our new model is not put in by hand , but is an _
emergent property _ of the eco - system .
the prey - predator interaction between two species that occupy the nodes @xmath7 and @xmath8 at two adjacent trophic levels is represented by @xmath9 ; the three possible values of @xmath9 are @xmath10 and @xmath11 .
the sign of @xmath9 indicates the direction of trophic flow , i.e. _ from the lower to the higher _ level .
@xmath9 is @xmath12 if @xmath7 eats @xmath8 and it is @xmath13 if @xmath8 eats @xmath7 . if there is no prey - predator relation between the two species @xmath7 and @xmath8 , we must have @xmath14 .
although there is no direct interaction between species at the same trophic level in our model , they can compete , albeit indirectly , with each other for the same food resources available in the form of prey at the next lower trophic level .
we now argue that the elements of the matrix @xmath15 account not only for the _ inter_-species interactions but also for the _ intra_-species interactions arising from the competition of individual organisms for the same food resources .
let @xmath16 be the number of all prey individuals for species @xmath7 on the lower trophic level , and @xmath17 be @xmath2 times the number of all predator individuals on the higher trophic level .
since we assume that a predator eats @xmath2 prey per time interval , @xmath16 gives the amount of total food available for species @xmath7 , and @xmath17 is the total contribution of species @xmath7 to the pool of food required for all the predators on the higher level .
if the available food @xmath16 is less than the requirement , then some organisms of the species @xmath7 will die of _ starvation _ , even if none of them is killed by any predator . the _ intra_-species competition among the organisms of the same species for limited availability of resources , other than food , imposes an upper limit @xmath18 of the allowed population of each species ;
@xmath18 is time - independent parameter in the model .
thus , the total number of organisms @xmath19 at time @xmath20 is given by @xmath21 .
if @xmath22 is larger than @xmath17 then food shortage will be the dominant cause of premature death of a fraction of the existing population of the species @xmath7 .
on the other hand , if @xmath23 , then a fraction of the existing population will be wiped out primarily by the predators . in order to capture the _ starvation deaths and killing by the predators _ , in addition to the natural death due to ageing , a reduction of the population by @xmath24 is implemented at every time step , where @xmath25 is the population of the species @xmath7 that survives after the natural death .
@xmath26 is a constant of proportionality .
if this leads to @xmath27 , species @xmath7 becomes extinct .
we assume that the simplest species occupying the lowest trophic level always get enough resources that neither natural death nor predators can affect their population .
an arbitrary species @xmath7 is _ collectively _ characterized by @xcite :
+ ( i ) the _ minimum reproduction age _ @xmath28 , + ( ii ) the _ birth rate _ @xmath29 , + ( iii ) the _ maximum possible age _ @xmath30 that depends only on the trophic level occupied by the species .
+ an individual of the species @xmath7 can reproduce only after attaining the age @xmath28 . whenever an organism of this species gives birth to offsprings , @xmath29 of these
are born simultaneously .
none of the individuals of this species can live longer than @xmath30 , even if an individual manages to escape its predators or starvation . with probability @xmath31 per unit time ,
each of the species randomly increases or decreases , with equal probability , their @xmath32 and @xmath33 by unity .
( @xmath32 is restricted to remain in the interval from @xmath34 to @xmath35 , and @xmath36 . ) moreover , with the same probability @xmath31 per unit time , they also re - adjust one of the links @xmath15 from prey and one of the links @xmath15 to predators @xcite .
if the link @xmath37 to the species @xmath7 from a _ higher _ level species @xmath38 is non - zero , it is assigned a new value of @xmath39 .
on the other hand , if the link @xmath9 to a species @xmath7 from a _ lower _ species @xmath8 is zero , the new values assigned are @xmath40 .
these re - adjustments of the incoming and outgoing ( in the sense of nutrient flow ) interactions are intended to capture the facts that each species tries to minimize predators but look for new food resources .
the niches ( nodes ) left empty because of extinction are re - filled by new species , with probability @xmath41 per unit time @xcite .
all the simultaneously re - filled nodes in a trophic level of the network originate from _ one common ancestor _ which is picked up randomly from among the non - extinct species at the same trophic level .
all the interactions @xmath15 of the new species are identical to those of their common ancestor .
the characteristic parameters @xmath32 , @xmath33 of each of the new species differ randomly by @xmath10 from the corresponding parameters for their ancestor .
however , occasionally , all the niches at a level may lie vacant . under such circumstances ,
all these vacant nodes are to be filled by a mutant of the non - extinct species occupying the closest _ lower _ populated level .
as stated above , the lowest level , that is populated by the simplest species , never goes extinct ; the possible ageing of the species at the lowest level @xcite is not relevant here .
all the individual organisms of the new species are assumed to be newborn babies that begin ageing with time just like the other species . since space
does not enter explicitly in our model , it does not distinguish between sympatric and allopatric speciation @xcite . in order to understand why the total number of trophic levels in food webs usually lie between @xmath42 and @xmath43
, we allowed adding a new trophic level to the food web , with a small probability @xmath44 per unit time , provided the total bio - mass distributed over all the levels ( including the new one ) does not exceed the total bio - mass available in the eco - system .
this step is motivated by the fact that real ecosystems can exhibit growing bio - diversity over sufficiently long period of time .
increase of the number trophic level means the diversification at the erstwhile topmost level as well as all the lower levels and the emergence of yet another dominating species that occupies the new highest level .
the total number of levels , which determines the lengths of the food chains , depends on several factors , including the available bio - mass @xcite . at each time step , each individual organism @xmath45 of the species @xmath7 gives birth _ asexually _ to @xmath29 offsprings with a probability @xmath46 .
we also assume the _ time - dependent _ probability @xmath46 is a product of two factors .
one of these two factors decreases linearly with age , from unity , attainable at the minimum reproduction age , to zero at the maximum lifespan .
the other factor is a standard verhulst factor @xmath47 which takes into account the fact that the eco - system can support only a maximum of @xmath18 individual organisms of each species .
thus , @xmath46 is equal to the verhulst factor at @xmath48 .
each individual organism , irrespective of its age , can meet its natural death .
however , the probability @xmath49 of this natural death depends on the age of the individual . in order to mimic age - independent constant mortality rate in childhood
, we assume the probability @xmath49 of `` natural '' death ( due to ageing ) to be a constant @xmath50 $ ] , ( with a small fraction @xmath51 ) , so long as @xmath52 .
however , for @xmath53 , the probability of natural death is assumed to increase following the gompertz law @xmath54 $ ] .
note that , for a given @xmath35 and @xmath32 , the larger is the @xmath33 the higher is the @xmath49 for any age @xmath55 .
therefore , in order maximize reproductive success , each species has a tendency to increase @xmath33 for giving birth to larger number of offsprings whereas the higher mortality for higher @xmath33 opposes this tendency @xcite .
however , even with a constant @xmath56 we found qualitatively similar results .
the state of the system is updated in discrete time steps where each step consists of a sequence of six stages : + _ i- birth _ _ ii- natural death _ _ iii- mutation _ _
iv- starvation death and killing by prey _
_ v- speciation
_ _ vi- emergence of new trophic level _ in all our simulations we began with random initial condition , except for @xmath57 for all species , mostly with only _ three _ levels in the food web , and
let the eco - system evolve for @xmath58 time steps before we started collecting ecological and evolutionary data from it ; these data were collected for the subsequent @xmath59 time steps where the longest runs were for @xmath60 .
we have not observed any qualitative differences in the data for @xmath61 and @xmath62 , keeping all the other parameters same .
most of our simulations were carried out with @xmath63 , as we did not observe qualitative differences between the data for @xmath63 and @xmath64 in test runs .
the maximum lifespans in the levels were assumed to be @xmath65 starting from the highest level .
several theories , based on extremely simple models , claim that the distribution of the lifetimes of the species should follow a power law with a slope of @xmath66 on the log - log plot .
the distributions of the lifetimes of the species in our model are shown in fig.[fig-1 ] for a few different sets of parameter values .
although our data do not rule out an approximate power law over limited regime of lifetimes , one single power law over the entire range of lifetimes seems impossible . in fig.[fig-2 ] we plot the distributions of the minimum reproductive age @xmath32 of the species for several different sets of values of the model parameters .
although over relatively short time scales of observation this distribution appears quite broad it narrows down with evolution and the non - zero values of this distribution correspond to reasonable values of age . due to the randomness in the evolutionary process , occassionally
, all of the niches in a level ( except the lowest one ) may lie vacant .
we have monitored @xmath67 and also @xmath68 , the number of those levels at time @xmath20 in which at least one niche is occupied by a non - extinct species . in fig.[fig-3 ] we plot @xmath69 as a function of time for one single run .
this clearly shows how , over geological time scales , @xmath6 reaches @xmath43 . in this run , the sixth level ( the highest one )
emerges after @xmath70 time steps .
it also demonstrates that at all stages of evolution , the number @xmath68 keeps fluctuating . during the very late stages
, @xmath69 keeps fluctuating between @xmath71 and @xmath43 , although @xmath69 is more often @xmath43 than @xmath71 for all times beyond @xmath72 .
the ratio of occurrences of _ six _ levels and _ five _ levels in the eco - system stabilized only after @xmath73 time steps .
we have also computed the distributions ( histograms ) of @xmath69 by averaging the data over large number of runs . as shown in fig.[fig-4 ] , the distribution becomes narrower for longer runs and the trend indicates than in the extreme long time limit would be sharply peaked around one single value of @xmath69 , as indicated by the fig.[fig-3 ] .
in this paper we have introduced a theoretical model of eco - systems with a generic hierarchical trophic level structure . because of data collection at suffciently short intervals ,
we have been able to monitor the ecological phenomena like , for example , birth , ageing and death of individual organisms and , hence , the population dynamics of the species .
we have also been able to run our simulations upto sufficiently long times ( @xmath74 , with stationarity achieved at around @xmath73 ) so that the model also accounted for macro - evolutionary phenomena like extinctions of species as well as speciation that leads not only to emergence of species at the existing levels of complexity but also to higher species that occupy an altogether new trophic level in the food web . from the infinite _ possible _ life forms , we start with one or a few , and then let our ecosystem grow in diversity and complexity until the limitations of biomass restrict it to hundreds of individuals in dozens of species , organized into about five trophic levels . although our model food web is hierarchical , it is not a tree - like structure .
the hierarchical architecture helps us in capturing a well known fact that in the normal ecosystems the higher is the trophic level the fewer are the number of species .
it is well known that the body size and abundance of a species are strongly correlated to their positions as well as to their interactions with other species in the food web @xcite . if we neglect parasites and herbivorous insects on trees , then , in general , predators are fewer in number and bigger in size as compared to their prey species @xcite .
this is very naturally incorporated in the hierarchical food web structure of our model .
let us assume that in the model the body size of individual organisms on each level @xmath75 is about @xmath2 times smaller than that on its predator level @xmath76 . on the other hand ,
the maximum possible populations of organisms , including all the nodes , in a level @xmath0 is @xmath2 times that at the level @xmath77 .
consequently , the maximum amount of biomass on each level is , approximately , the same .
since each individual organism appears explicitly in our model , one could , at least in principle , assign a genome to each individual and describe darwinian selection which takes place at the level of organisms @xcite .
unfortunately , additional ad - hoc assumptions would be required to relate the genome with the reproductive success @xcite . instead of introducing an ad - hoc mathematical formula to relate genotype with phenotype , we have worked directly with phenotype , particularly , quantities that decide the reproductive success of the organisms ; these quantities are @xmath32 , @xmath35 and @xmath33 . from the perspective of self - organization ,
the new model surpasses all cited previous models as not only the characteristic collective properties of the species but even the nature of inter - species interactions as well as the total number of trophic levels in the food web are determined by self - organization of the eco - system .
p. bak and k. sneppen , _ phys .
lett . _ * 71 * , 4083 ( 1993 ) ; m. paczuski , s. maslov and p. bak , phys .
e * 53 * , 414 ( 1996 ) .
sol , s.c.manrubia , m.benton , and p.bak , _ nature _ * 388 * , 764 ( 1997 ) .
+ p. bak and s. boettcher , physica d * 107 * , 143 ( 1997 ) .
r. v. sol , s. c. manrubia , m. benton , s. kauffman and p. bak , _ trends in ecology and evolution _ , * 14 * , 156 ( 1999 ) .
+ j. maynard - smith , _ the theory of evolution _ , ( cambridge university press , 1993 ) ; j. maynard - smith , _ shaping life : genes , embryos and evolution _ ( weidenfeld and nicolson , orion publishing group , london , 1998 ) .
+ e. mayr , _ what evolution is _ , ( basic books , 2001 ) .
+ s. b. carroll , _ nature _ * 409 * , 1102 ( 2001 ) .
+ b. drossel , _ adv .
phys . _ * 50 * , 209 ( 2001 ) .
+ m. e. j. newman , and r. g. palmer , _ modeling extinction _ , ( oxford university press , 2002 ) .
+ j. n. thompson , _ trends in ecol .
_ * 13 * , 329 ( 1998 ) .
+ g. f. fussmann , s. p. ellner , and n. g. hairston jr , _ proc .
b _ * 270 * , 1015 ( 2003 ) .
+ a. i. khibnik and a. s. kondrashov , _ proc .
soc . _ * 264 * , 1049 ( 1997 ) .
+ n. s. goel , s. c. maitra and e. w. montroll , _ rev .
_ * 43 * , 231 ( 1971 ) .
j. hofbauer , and k. sigmund , k. _ the theory of evolution and dynamical systems _ ( cambridge university press , 1988 ) .
+ j. d. murray , _ mathematical biology _
( springer , 1989 ) .
+ s. kauffman , _ the origins of order : self - organization and selection in evolution _
( oxford university press , 1993 ) .
+ g. caldarelli , p.g .
higgs and a.j .
mckane , j. theor .
* 193 * , 345 ( 1998 ) .
+ b. drossel , p.g .
higgs and a.j .
mckane , j. theor .
biol . * 208 * , 91 ( 2001 ) .
+ c. quince , p.g .
higgs and a. j. mckane , in : _ biological evolution and statistical physics _
, eds . lssig .
m. and valleriani , a. ( eds ) ( springer , 2002 ) , p.281 - 298 .
+ d. stauffer , in : _ biological evolution and statistical physics
lssig , m. and valleriani , a. ( springer , 2002 ) , p.255 - 267
. + m. hall , k. christensen , s. a. di collobiano and h. j. jensen , _ phys .
e _ * 66 * , 011904 ( 2002 ) .
+ s. a. di collobiano , k. christensen and h. j. jensen , _ j. phys . a _ * 36 * , 883 ( 2003 )
. + d. chowdhury , d. stauffer and a. kunwar , _ phys .
. lett . _ * 90 * , 068101 ( 2003 ) .
+ d. chowdhury and d. stauffer , phys .
e * 68 * , 041901 ( 2003 ) .
+ p. a. rikvold and r. k. p. zia , 2003 punctuated equilibria and 1/f noise in a biological coevolution model with individual - based dynamics , nlin.ao/0306023 .
+ a. aszkiewicz , sz .
szymczak and s. cebrat , _ int .
c _ * 14 * , ( 2003 ) .
+ s. l. pimm , _ food webs _ ( chapman and hall , london , 1982 ) .
+ g. a. polis and k. o. winemiller , ( eds . ) _ food webs : integration of patterns and dynamics _ ( chapman and hall , new york , 1996 ) .
+ b. drossel and a. j. mckane , 2003 modelling food webs , p. 218
in : _ handbook of graphs and networks - from the genome to the internet _ , eds . s. bornholdt , and h. g. schuster , ( wiley - vch , weinheim , 2003 ) + j. e. cohen , t. luczak , c. m. newman and z. -m .
zhou , _ proc .
b _ * 240 * , 607 ( 1990 ) .
+ j. n. thompson , _ science _ * 284 * , 2116 ( 1999 ) . + r. v. sol and s. c. manrubia , phys .
e * 54 * , r42 ( 1996 ) ; _ phys .
e _ * 55 * , 4500 ( 1997 ) ; see also r. v. sol , in : _ statistical mechanics of biocomplexity _ , lec .
notes in phys .
( springer , 1999 ) . + m. r. orr and t. b. smith , _ trends ecol .
* 13 * , 502 ( 1998 ) .
+ m. ackermann , s. c. stearns and u. jenal , _ science _ * 300 * , 1920 ( 2003 ) .
+ u. dieckmann and m. doebeli , _ nature _ * 400 * , 354 ( 1999 ) ; m. doebeli and u. dieckmann , _ am .
* 156 * , s77-s101 ( 2000 ) .
see also the other articles on speciation in the same issue .
+ d. m. post , 2002 _ trends ecol .
* 17 * , 269 ( 2002 ) .
+ c. k. ghalambor and t. e. martin , _ science _ * 292 * , 494 ( 2001 ) .
+ j. e. cohen , s. l. pimm , p. yodzis and j. saldana , _ j. animal ecology _ , * 62 * , 67 ( 1993 ) .
+ j. e. cohen , t. jonsson and s. r. carpenter , 2003 _ proc .
usa _ * 100 * , 1781 ( 2003 ) .
+ g. c. williams , _ natural selection : domains , levels and challenges _
( oxford university press , 1992 ) .
+ e. mayr , _ proc .
usa _ * 94 * , 2091 ( 1997 ) . | we propose a generic model of eco - systems , with a _ hierarchical _ food web structure . in our computer simulations
we let the eco - system evolve continuously for so long that that we can monitor extinctions as well as speciations over geological time scales . _
speciation _ leads not only to horizontal diversification of species at any given trophic level but also to vertical bio - diversity that accounts for the emergence of complex species from simpler forms of life .
we find that five or six trophic levels appear as the eco - system evolves for sufficiently long time , starting initially from just one single level .
moreover , the time intervals between the successive collections of ecological data is so short that we could also study `` micro''-evolution of the eco - system , i.e. , the birth , ageing and death of individual organisms . |
the modeling of complex materials based upon molecular dynamics simulation has been one of the remarkable advances in theoretical condensed matter physics . whether the potentials chosen are empirical or
_ ab initio _
, remarkable insights have accrued for diverse problems in materials physics and beyond . there is , however , an unsatisfying point to the logic of md simulation : it does not make use of all the information available about a material under study notably the information implied by experiments . simulations often can not achieve agreement with experiment because of short simulation times , small system sizes or inaccuracies in the interactions . successful prediction of _ new _ properties is more likely for models in agreement with existing data .
imposition of experimental information may be important in phase - separated or other complex materials for which obtaining a suitable starting structure may be difficult , and for which short md time scales preclude the emergence of such structures in the model .
a different approach to model construction implemented by mcgreevy@xcite and colleagues is the so - called reverse monte carlo " ( rmc ) method .
here , one explicitly sets out to make an atomistic model which agrees with experiments .
rmc has been widely used to model a variety of complex disordered materials .
this is accomplished by making monte carlo moves which drive a structural model toward exact agreement with one or more experiments . in practice
, rmc is the ideal method to explore the _ range _ of configurations which are consistent with experiment(s ) . without adequate limitation to a physical " subspace of configuration space ,
it is unlikely to produce a satisfactory model .
that is , only a subset of rmc models [ which match the experiment(s ) ] is physically realistic ( consistent with accurate interatomic interactions ) .
the imposition of topological / chemical bonding constraints in rmc can ameliorate this problem , but not remove it entirely@xcite .
the mathematical structure of constrained rmc is a constrained optimization traveling salesman " problem . in our previous implementation of constrained rmc we formed a positive definite ( quadratic ) cost or
penalty " function @xmath1 , which was then minimized ( ideally , but not practically ) to zero for a structural model which exactly satisfies all constraints imposed : [ eq-1 ] = _ j=1^k _ i=1^m_j _ i^j \{f^j_e(q_i)- f^j_c(q_i)}^2 + _ l=1^l _ l p_l where @xmath2 is related to the uncertainty associated with the experimental data points , @xmath3 is the number of experimental data sets employed , @xmath4 is the number of data for the @xmath5 set and @xmath6 is the number of additional ( non - experimental ) constraints included .
the quantity @xmath7 is the appropriate generalized variable associated with experimental data @xmath8 and @xmath9 is the penalty function associated with each additional constraint and @xmath10 .
additional " constraints can be of many different forms ( for example , one may impose chemical or topological ordering , or phase separated units within a continuous random network ) the coordinates of atoms are changed according to monte carlo moves , which is akin to a simulated annealing minimization of our cost function @xmath1 .
the method is easy to implement , though care must be taken to include the minimum number of independent constraints possible to reduce the likelihood of getting stuck " in spurious minima .
we have shown that inclusion of suitable constraints leads to models of a - si much improved compared to rmc models using only the structure factor ( first term of eq .
[ eq-1 ] ) as constraint@xcite .
as the creation of models of complex materials is a difficult task , it is of obvious advantage to incorporate _ all _ possible information in fabricating the model .
we assert that an ideal model of a complex material should ( 1 ) be a minimum ( metastable or global ) of a suitable energy functional faithfully reproducing the structural energetics , ( 2 ) should contradict no experiments . when stated in these terms , our criterion seems quite obvious , yet current simulation schemes do not simultaneously accommodate both criteria , but focus only on one or the other . in this paper
, we merge the power of _ ab initio _ molecular simulation with the _ a priori _ information of experiments to create models consistent with experiments and the chemistry implied by accurate interatomic interactions . to obtain joint agreement , we unite md with the reverse monte carlo ( rmc ) method .
we name the scheme `` experimentally constrained molecular relaxation '' ( ecmr ) .
one can understand our scheme as a way to
tune " a structural model using md within the space of _ experimentally realistic _ models as defined by rmc .
we choose a troublesome and complex material with a long experimental and modeling history : g - gese@xmath0 . from an algorithmic perspective , our scheme has important advantages .
for example , to model a glass like gese@xmath0 or sio@xmath0 using first principles methods , the method of choice is to form an equilibrated liquid , use some dissipative dynamics to simulate an ( unphysically ) rapid quench of the liquid into an arrested phase and finally to relax this to a local energy minimum , presumably at astronomical fictive temperature ( high potential energy ) .
usually some repeated annealing " cycles are also used .
if one is interested in a glassy phase all the work of forming and equilibrating the liquid is redundant , and it is a pious hope that the arrested liquid will resemble a real glassy phase . evidently the likelihood for success is strongly affected by topological and chemical similarity of the melt to the physical amorphous phase . if complex ordering self - organization " or phase separation occurs in the physical amorphous phase , the short simulations of conventional _ ab initio _ schemes will surely miss these important structural features . in this vein , we have used ecmr to construct models of a - si with intermediate range order on a nanometer length scale@xcite by inclusion of fluctuation electron microscopy@xcite .
we note that successful techniques do exist to tackle the time - scale problem@xcite , though these do not enable the inclusion of experimental information .
our method is efficient enough to enable the creation of a 647 atom model of g - gese@xmath11 using only a workstation .
the method is at least a factor of five faster than a comparable quench from the melt simulation , with its inherent limitations .
the obvious means to incorporate interatomic interactions into an rmc simulation is to add a constraint to minimize the magnitude of the force on all the atoms according to some energy functional or to minimize the total energy . for an _ ab initio _
hamiltonian this is expensive , since monte carlo minimization of eq .
[ eq-1 ] requires a large number of energy / force calls .
thus , we have instead employed a simple self consistent " iteration scheme ( indicated in fig .
[ fig1 ] ) : ( 1 ) starting with an initial configuration c@xmath12 , minimize @xmath1 to get c@xmath0 , 2 ) steepest - descent quench c@xmath0 with an _ ab initio _ method to get c@xmath13 , ( 3 ) subject the resulting configuration to another rmc run ( minimize @xmath1 again ) , repeat steps ( 2 ) and ( 3 ) until both the md relaxed model and rmc models no longer change with further iteration . in this paper , we limit ourselves to the first term in eq .
1 ( the experimental static structure factor ) , though additional constraints certainly could be employed . for the rmc component of the iteration
, we make the conventional choice of using monte carlo for the minimization .
this is simple and does not require gradients ( and thus allows the use of non - analytic terms in eq .
[ eq-1 ] @xcite , if desired ) .
we emphasize that our method is _ flexible_. it s logic suggests that one should include whatever experimental information is available . in this paper
we limit ourselves to the pair - correlation functions . in principle
, other experiments could be included as well .
these might be costly to include ( for example to compel agreement with the vibrational density of states , the dynamical matrix would be required at each iteration ) .
the method is equally suited to fast empirical potentials , which would allow studies of very large models .
it is also possible to force a close fit to some restricted range of data , and a less precise fit elsewhere if desired .
our scheme also provides insight into the topological signatures of different constraints ( experimental or otherwise ) .
chemical and or topological constraints could also be maintained as part of the rmc iteration .
our method can be understood as a way to minimize an effective potential energy function @xmath14 , where @xmath15 is the potential energy of the configuration ( denoted by @xmath16 ) , @xmath17 , and @xmath18 is a non - negative cost function enforcing experimental ( or other ) constraints as in eq .
empirically , we find that it is possible to find configurations that simultaneously approximately minimize both terms ( which implies that the choice of @xmath19 is not very important ) .
it is also clear that our method is really _ statistical _ : in general one should generate an ensemble of conformations using ecmr .
for adequately large models , self averaging can be expected ; in this study of large ( 647 atom ) models of g - gese@xmath0 we find similar results for two runs ; for small systems a proper ensemble average is required .
the method is new and as such needs to be studied and developed in a number of ways .
nevertheless , we show in this paper that it is relatively easy to model a particularly challenging material with significant advantages in both experimental plausibility of the model and computational efficiency of the algorithm . ,
we apply ecmr to glassy gese@xmath0 , a classic glass forming material with challenging physical and technical issues : ( 1 ) it displays nanoscale order : a first sharp diffraction peak " ( fsdp ) is observed in neutron diffraction measurement , ( 2 ) the packing of gese tetrahedra involves both edge- and corner- sharing topologies ; ( 3 ) the material has interesting photoresponse ( understanding of which requires the electronic structure ) , ( 4 ) the material is difficult to simulate with _ ab initio _ techniques@xcite .
the model used in our calculation consists of 647 atoms in a cubic box of size 27.525 . in the nomenclature of fig .
[ fig1 ] , c@xmath12 is obtained by constraining the coordination number ( 2 for se , 4 for ge ) and the bond - angle distribution of se - ge - se to an approximate gaussian with an average bond angle 109.5@xmath20 .
the initial network was generic " and included none of the detailed local chemistry of ge and se aside from the coordination and chemical ordering ( bond _ angles _ were not constrained in rmc loops ) .
equal weighting was used for all experimental points in this paper . using the method of isotopic substitution ,
salmon and petri @xcite were able to separately measure the three ( ge - se , ge - ge and se - se ) partial structure factors of g - gese@xmath0 .
we jointly enforced all three partials ( in real space ) in the rmc component of the loop in fig .
the md relaxation was done with fireball@xcite .
it was found that after the fourth iteration , s(q ) hardly changed . in table
i , we show the average force per atom at the beginning of each call to md relaxation ; good convergence is observed .
subsequent discussion in this paper is for the last step of the md , with forces less than @xmath21 .
it was not obvious to us in the beginning that rmc and first principles interatomic interactions could be made self - consistent " , but for this system at least , reasonable convergence is possible .
it is likely that some initial conformations c@xmath12 will get stuck " and require a new start , but we have not encountered difficulty with this yet . in fig .
[ fig2 ] , we compare the rmc , ecmr and experimental structure factors . here , the rmc model is that obtained by starting with the generic c@xmath22 configuration , and forcing agreement on the experimental @xmath23 ( without any other constraints ) . while the agreement is very good , it is not perfect
this is to be expected for three reasons : ( 1 ) consistency between data and hamiltonian is never exact ; ( 2 ) our cell contains 647 atoms , which is compared to the thermodynamic limit and ( 3 ) we chose to constrain our model using real space data , which involves fourier transforms and windowing ( this introduces only small errors in this data set ) . in fig .
2 we highlight the differences between experiment@xcite , a quench from the melt model@xcite and the new ecmr model . in the inset of figure 2 , we also illustrate the level of agreement using a pure rmc approach , which is similar to the ecmr result and notably better than quench from the melt . for reference
, we have reproduced the full partial structure factors elsewhere@xcite .
note in fig .
[ fig2 ] that the first sharp diffraction peak ( fsdp ) is well reproduced , ( very close in width and centering , and much improved from all previous models in height ) . moreover , as for our decorate and relax " ( dr ) method@xcite , the large @xmath7 structure closely tracks experiment ( unlike the experience for quench from the melt models which are too liquid - like and therefore decay too rapidly for large q ) .
these desirable features are of course built in " ; we show here that the ecmr method does preserve every important feature of the structure of the glass manifested in s(q )
. neutron - weighted static structure factor , comparing ecmr model , experiment@xcite and a quench from the melt made with the same hamiltonian used with ecmr@xcite .
inset : blowup of small - q region showing initial rmc model ( eg , enforcing experimental structure factor , but without ecmr iterations ) , experiment@xcite and quench from the melt model due to cobb _ et al._@xcite .
the first sharp diffraction peak is closely reproduced by ecmr and rmc , and is present but weak in the quenched model .
, width=4,height=3 ] an important indicator of network topology and medium range order of gese@xmath0 glass is the presence of edge - sharing and corner - sharing tetrahedra .
raman spectroscopy @xcite and neutron diffraction @xcite studies have indicated that 33% to 40% of ge atoms are involved in edge sharing tetrahedra .
the fraction in our model is found to be 38% .
this was not built in " to our modeling , and is a pleasing prediction arising from the procedure .
we also have observed that 81% of ge atoms in our model are four - fold coordinated of which approximately 75% form predominant ge - centered structural motifs @xmath24 while 6% are ethane - like @xmath25 units .
the remaining ge atoms are three - fold coordinated and are mostly found to be bonded as ge se@xmath13 units . on the other hand ,
the percentage of two- , three- and one - fold coordinated se atoms are 72% , 18% and 10% respectively .
mssbauer experiments , where sn was used as a ge probe @xcite , estimated that the fraction of ge involved in dimers is 16% which is again in favor of our model .
.[tab : table1 ] the convergence of ecmr described in the text . [ cols="<,^,^,^,^,^,^,^",options="header " , ]
in summary , we have proposed a new method which enables the inclusion of _ a priori _ information ( experimental or otherwise ) into molecular simulation .
we have shown that the method is effective for a challenging material , g - gese@xmath0 .
we thank dr.p.s . salmon for providing us with experimental data , and professor
himanshu jain for helpful discussions .
we acknowledge the support of national science foundation for support under grants nos .
dmr-0074624 , dmr-0205858 and dmr-0310933 . | an ideal atomistic model of a disordered material should contradict no experiments , and should also be consistent with accurate force fields ( either _ ab initio _ or empirical ) .
we make significant progress toward jointly satisfying _ both _ of these criteria using a hybrid reverse monte carlo approach in conjunction with approximate first principles molecular dynamics .
we illustrate the method by studying the complex binary glassy material g - gese@xmath0 . by constraining the model to agree with partial structure factors and _ ab initio _ simulation ,
we obtain a 647-atom model in close agreement with experiment , including the first sharp diffraction peak in the static structure factor .
we compute the electronic state densities and compare to photoelectron spectroscopies .
the approach is general and flexible . |
a guiding principle in physics is that physical laws should be expressed in a form that is independent of any particular coordinate system .
_ fluid mechanics _ is a field theory of newtonian mechanics that has galilean symmetry : that is to say , it is covariant under transformations of the galilei group .
two symmetries ( _ i.e. _ transformation invariances ) are known as subgroups of the galilei group : translation ( space and time ) and space - rotation . in the present formulation of flows of an ideal fluid , we seek a scenario which has a formal equivalence with the gauge theory in physics .
the gauge theory provides a basis for reflection on the similarity between fluid mechanics and other physical fields .
the gauge principle requires a physical system under investigation to have a _ symmetry _ , _
i.e. _ a gauge invariance ( an invariance with respect to a certain group of transformations ) . following this principle , an attempt has been made in kambe ( 2007 ) to study a gauge symmetry of flow fields with respect to translational transformations .
the formulation started from a galilei - invariant lagrangian of a system of _ point masses _ which is known to have _ global _ gauge symmetries with respect to both translation and rotation . and
appendix for the definition of global gauge symmetry ( _ i.e. _ global gauge invariance ) and local gauge symmetry .
] it was then extended to flows of a fluid , a _ continuous _
material characterized with mass density and entropy density . in addition to the global symmetry , _ local _ gauge invariance of a lagrangian is required for such a continuous field .
thus , the convective derivative of fluid mechanics ( _ i.e. _ the lagrange derivative ) is identified as a _
covariant derivative _ , which is a building block in the framework of gauge theory . based on this
, appropriate lagrangian functionals are defined for motion of an ideal fluid .
euler s equation of motion is derived from the action principle .
it is found that a general solution of this formulation is equivalent to the classical clebsch solution ( clebsch 1859 ; lamb 1932 ) , in which the vorticity has a particular form and the helicity vanishes . in most traditional formulations , the continuity equation and isentropic condition
are taken into account as constraints for the variations of the action principle by using lagrange multipliers , where the _ isentropy _ means that each fluid particle of an ideal fluid keeps its entropy value along its trajectory ( but the fluid is not necessarily _
homentropic_. ) in the present formulation however , those equations are derived from the action principle .
the total lagrangian density consists of kinetic energy and internal energy ( with negative sign ) , supplemented with additional two terms associated with translation invariances of mass and entropy ( see ( 27 ) of 3.1 ) .
as is interpreted in detail in 4.1 , it is remarkable that the lagrangians associated with the latter two terms are given such forms as can be eliminated from the action represented with the particle coordinate @xmath1 ( _ i.e. _ lagrangian coordinate ) , so that the euler - lagrange equations represented by the independent varibales @xmath3 are uninfluenced by the additional terms .
such a form of lagrangian has close analogy to the chern - simons term in the gauge theory .
earlier papers ( kambe , 2003a , b ) investigated the rotation symmetry of the velocity field @xmath4 in a preliminary way , and found that the vorticity @xmath5 is in fact a gauge field associated with the rotation symmetry .
however in the previous studies including kambe ( 2007 ) , the vorticity and helicity were not taken into account in satisfactory manner .
present formulation tries to account for the vorticity and helicity more properly and comprehensively .
thus , the main theme of the present paper is the rotation symmetry . a new structure of the rotation symmetry
is given in 6 ( after a review of the rotational transformations in 5 ) by the following lagrangian @xmath0 : @xmath6 \,\rangle \,{\mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$}},\ ] ] where @xmath7 is a vector potential and @xmath8 \equiv { \partial}_t { \mbox{\boldmath $ \omega$}}+({\mbox{\boldmath $ v$}}\cdot\nabla){\mbox{\boldmath $ \omega$}}-({\mbox{\boldmath $ \omega$}}\cdot\nabla){\mbox{\boldmath $ v$}}+ ( \nabla\cdot { \mbox{\boldmath $ v$ } } ) { \mbox{\boldmath $ \omega$}}$ ] .
this is derived from a lagrangian form characteristic of a _ topological _ term known in the gauge theory .
this new term yields non - vanishing rotational component of the velocity field , and provides a source term of non - vanishing helicity .
this is also the chern - simons term , describing non - trivial topology of vorticity field , _ i.e. _ mutual linking of vorticity lines .
the vorticity equation is derived as an equation for the gauge field .
these are described in [ s6 ] ( lagrangian @xmath9 associated with rotation symmetry ) , [ s7 ] ( corrections to the variation ) , and [ s8 ] ( significance of @xmath0 ) . with regard to the variational formulation of fluid flows , the works of eckart ( 1938 , 1960 ) and herivel ( 1955 ) are among the earliest to have influenced current formulations .
their variations are carried out in two ways : _
i.e. _ a lagrangian approach and an eulerian approach . in both approaches , the equation of continuity and the condition of is entropy are added as constraint conditions on the variations by means of lagrange multipliers .
the lagrangian approach is also taken by soper ( 1976 ) . in this relativistic formulation those equation are derived from the equations of current conservation .
several action principles to describe relativistic fluid dynamics have appeared in the past ( see soper ( 1976 ,
4.2 ) for some list of them ) , not only that of taub ( 1954 ) .
firstly , in the lagrangian approach , the euler - lagrange equation results in an equation equivalent to euler s equation of motion in which the acceleration term is represented as the second time derivative of position coordinates of the lagrangian representation . in this formulation
, however , there is a certain degree of freedom in the relation between the lagrangian particle coordinates @xmath10 and eulerian space coordinates @xmath11 .
namely , the relation between them is determined only up to an unknown rotation .
this point has not been considered seriously in the past , although it is fundamental for the formualtion of flows .
this will be discussed more precisely in the sections 4.2 and 8.2 .
a common feature of herivel ( 1955 ) and eckart ( 1960 ) is that both arrive at the clebsch solution in the eulerian representaion ( see sec.[a1-gi ] and [ 3b ] ) .
however , it was remarked by bretherton ( 1970 ) that the clebsch representation has only local validity in the neighbourhood of a chosen point if vortex lines are knotted or linked . in the second approach , referred to as the _ eulerian _ description , the action principle of an ideal fluid results only in potential flows if the fluid is homentropic , as shown in kambe ( 2003a , 2007 ) .
however it is well - known that , even in such a homentropic fluid , it should be possible to have rotational flows .
this is a long - standing problem ( serrin 1959 ; lin 1963 ; seliger & whitham 1968 ; bretherton 1970 ; salmon 1988 ) .
lin ( 1963 ) tried to resolve this difficulty by introducing a _ constraint _ as a side condition , imposing invariance of lagrangian particle labels @xmath10 along particle trajectories .
the constraints for the variation are defined by using lagrange multipliers ( functions of positions ) , which are called potentials. in addition , the continuity equation and isentropic condition are also taken into account by using lagrange multipliers . however , the physical significance of those potentials introduced as the lagrange multipliers is not clear .
mysteriously , the lagrange multiplier for the continuity equation becomes the velocity potential for the irrotational part of the velocity .
gauge theory for fluid flows provides a crucial key to resolve these issues , which is the main target of the present investigation .
symmetries imply conservation laws . according to the relativistic formulation ( _ e.g._soper 1976 ; landau & lifshitz 1987 ( ch.xv ) ) , as long as the flow fields obey the euler - lagrange equation , an energy - momentum tensor @xmath12 must obey the conservation equation of the form @xmath13 ( considered in sec . [ 4b ] below ) . the energy - momentum tensor @xmath12 in the present case is given in sec.[4-flv ] .
a similar field - theoretic approach is taken in jackiw ( 2002 ) by applying the ideas of particle physics to fluid mechanics in terms of hamiltonian and poisson brackets , both relativistically and nonrelativistically , and extension to supersymmetry is also considered .
in soper ( 1976 ) and jackiw ( 2002 ) , the nonrelativistic part follows the traditional approach and no explicit gauge - theoretic consideration is given to fluid mechanics in the sense of the present formulation .
the appendix describes the background of the present theory : ( a ) known gauge invariances in two systems of electromagnetism and clebsch form of fluid flows , ( b ) related aspects in quantum mechanics and relativistic theory , and ( c ) brief scenario of gauge principle .
the formulation of kambe ( 2007 ) started from a galilei - invariant lagrangian of a system of _ point masses _ , which is known to have _ global _ gauge symmetries with respect to both translation and rotation .
concept of _ local _ transformation is an extention of the _ global _ transformation to describe a continuous field . in order to represent a continuous distribution of mass ,
the discrete positions of point masses are replaced by continuous parameters @xmath14 of particles in a bounded sub - space @xmath15 ( of consideration ) in the three - dimensional euclidean space @xmath16 .
the spatial position of each mass particle labelled by @xmath1 ( lagrange parameter ) is denoted by @xmath17 , a function of @xmath1 as well as the time @xmath18 .
an eulerian space point is denoted by @xmath19 .
the fluid particle occupying the point @xmath2 at time @xmath18 is denoted by @xmath20 .
are defined so as to represent the _ mass coordinate _ such that @xmath21 , where @xmath22 and @xmath23 .
a variable such as @xmath24 denotes the one with respect to the particle @xmath1 . ] for a continuous distribution of mass ( _ i.e. _ for a _ fluid _ ) , lagrangian functional is defined as @xmath25 where @xmath26 and @xmath27 denote @xmath28 and @xmath29 ( @xmath30 ) respectively , and @xmath31 is a lagrangian density .
suppose that an infinitesimal transformation is expressed by @xmath32 and @xmath33 , where @xmath34 and @xmath35 , a component of @xmath36 is denoted by @xmath37 , an arbitrary differentiable variation field .
the resulting variation of the lagrangian density @xmath38 is @xmath39 this does not vanish in general .
in fact , assuming the euler - lagrange equation , @xmath40 , we obtain @xmath41 global invariance , _
@xmath42 for arbitrary constants @xmath43 ( @xmath30 ) , requires that the total momentum defined by @xmath44 must be invariant . in the local transformation
, however , the variation field @xmath43 depends on the time @xmath18 and space point @xmath26 , and the variation @xmath45 does not vanish in general .
according to the gauge principle of appendix [ a3-gp ] , non - vanishing of @xmath46 is understood as meaning that a _ new field _
@xmath47 must be taken into account in order to achieve _ local gauge invariance _ of the lagrangian @xmath48 . to that end
, we try to replace the partial time derivative @xmath49 by a _
covariant derivative _ @xmath50 , where the derivative @xmath50 is defined by @xmath51 . in dynamical systems like the present one , the _ time derivative _ is the primary object to be considered in the analysis of local gauge transformation , as implied by the expression ( [ dl3 ] ) .
thus , the time derivatives @xmath52 and @xmath53 are replaced by @xmath54 and @xmath55 , where @xmath56 is understood to denote @xmath2 . a free - field lagrangian @xmath57 is defined by @xmath58 where @xmath59 denotes the mass element in a volume element @xmath60 of the @xmath26-space with @xmath61 the mass - density , and @xmath62 .
symmetries to be investigated are the _ translation symmetry _ and _ rotation symmetry _ , which exist in the discrete system .
the lagrangian @xmath57 defined by ( [ lfvv ] ) for a continuous field has such global symmetries , inherited from the discrete system .
it is a primary concern here to investigate whether the system of fluid flows satisfies local invariance .
we consider the translation symmetry first , and in the second half of the present analysis we investigate the rotation symmetry .
a translational transformation from one frame @xmath63 to another @xmath64 moving with a relative velocity @xmath65 is a galilean transformation in newtonian mechanics .
the transformation is defined by @xmath66 this is a sequence of global translations with a parameter @xmath18
. corresponding transformation of velocity is as follows : @xmath67 . from ( [ gtx ] ) ,
differential operators @xmath68 and @xmath69 ( @xmath70 ) are transformed according to @xmath71 and @xmath72 , where @xmath73 and @xmath74 .
there is a transformation invariance of the convective derivative @xmath75 : namely @xmath76 .
the velocity field @xmath77 is _ defined _ by the particle velocity , _
i.e. _ @xmath78 consider the following infinitesimal transformation : @xmath79 this may be regarded as a local gauge transformation between non - inertial frames .
in fact , the transformations ( [ lgtx ] ) is understood to mean that the coordinate @xmath2 of a fluid particle at @xmath80 in the frame @xmath63 is transformed to the new coordinate @xmath81 of @xmath64 , given by @xmath82 .
therefore , its velocity @xmath83 is transformed by @xmath84 this is interpreted as follows : the origin of a local frame is displaced by @xmath85 where its coordinate axes are moving ( without rotation ) with the velocity @xmath86 in accelerating motion ( non - inertial frame ) .
this implies that the velocity @xmath4 is transformed according to ( [ lgtu ] ) , since the local frame is moving with the velocity @xmath86 .
note that , in this transformation , the points @xmath2 and @xmath81 are the same with respect to the space @xmath63 . in view of the transformation ( [ lgtx ] ) , the time derivative and spatial derivatives are transformed as @xmath87 in this local transformation too
, there is an invariance of the convective derivative @xmath88 .
in fact , from ( [ lgdx ] ) and ( [ lgtu ] ) , we obtain @xmath89 where @xmath90 . the last term is @xmath91 by ( [ lgxi ] ) .
thus , we find the invariance of @xmath92 by ( [ lgdt ] ) .
this invariance ( _ i.e. _ the transformation symmetry ) implies that the covariant derivative @xmath50 is in fact given by the convective derivative @xmath92 : @xmath93 ( see kambe ( 2007 ) for its direct derivation ) . for a scalar function @xmath94 , we have @xmath95 ( lagrange derivative of @xmath94 ) .
it is important to recognize that this derivative @xmath96 has an intrinsic meaning .
that is , we have the invariance : @xmath97 , with respect to the local galilean transformation ( [ lgtx ] ) , since a scalar means @xmath98 . in general , vectors or tensors
need not have this property . by using the covariant derivative @xmath50
, we can define the velocity @xmath99 by @xmath100 .
in fact , each component @xmath101 of the particle label @xmath1 is a scalar , and satisfies @xmath102 , since the particle with the label @xmath101 moves with the velocity @xmath103 by the definition ( [ def - v ] ) . setting @xmath104 for the particle position
, we have @xmath105 where @xmath106 . in addition , applying @xmath50 to ( [ lgtx ] ) and using @xmath107 , we have @xmath108 ( consistent with ( [ lgtu ] ) ) .
thus , the particle velocity @xmath99 can be defined by @xmath109 .
the gauge transformation of sec.[2c - lgt ] defines the transformations : @xmath110 and @xmath111 .
namely , the variations of @xmath112 and @xmath113 are given by @xmath114 where @xmath115 .
next , let us consider acceleration of a fluid particle
. differentiating ( [ lgtu ] ) with respect to @xmath18 again , we obtain @xmath116 this describes a relation between accelerations in the two frames @xmath63 and @xmath64 .
let us define the covariant derivative of the velocity @xmath117 by @xmath118 .
variation of @xmath119 is given by @xmath120 where @xmath121 from ( [ lgdt ] ) ( to the first order of @xmath36 ) .
then , using @xmath122 and @xmath123 from ( [ lgdx ] ) , the transformation of @xmath124 is given by @xmath125 , namely , @xmath126 thus , we obtain @xmath127 it is found that the covariant derivative @xmath124 transforms just like the acceleration @xmath128 according to ( [ lgta ] ) .
this implies that @xmath124 represents the particle acceleration @xmath129 : @xmath130 according to the formulation of kambe ( 2007 ) based on the gauge principle , the total lagrangian is defined by @xmath131 as far as the translation symmetry is concerned , where @xmath132 the variables @xmath99 , @xmath61 , @xmath133 and @xmath134 are the velocity vector , density , internal energy ( per unit mass ) and entropy ( per unit mass ) of the fluid , @xmath135 and @xmath136 are gauge potentials associated with the mass density @xmath61 and entropy density @xmath137 respectively .
the velocity is defined by @xmath138 , where @xmath139 . from [ 2e ] , it is almost obvious that the two lagrangians @xmath140 and @xmath141 of ( [ l - g0 ] ) are invariant with respect to the local galilean transformation ( _ i.e. _ local gauge transformation ) .
the action @xmath142 and the action principle @xmath143 are defined by @xmath144 , \label{action } \\ { \delta}\,i & = & { \delta}\,\int_{t_0}^{t_1}\int_m { \mbox{\textrm{d}}}t\,{\mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$}}\ l_{\rm t}=0,\label{actpri1 } \\ & & l_{\rm t } \equiv \half\,\rho\,\langle\,{\mbox{\boldmath $ v$ } } , { \mbox{\boldmath $ v$}}\,\rangle - \rho\,{\epsilon}(\rho , s ) - \rho\ { \mbox{\textrm{d}}}_t\phi - \rho s\ { \mbox{\textrm{d}}}_t\psi\ , . \label{l - total}\end{aligned}\ ] ] where @xmath145 is the lagrangian density .
there is a certain thermodynamic property that must be taken into consideration . that is the definition of an ideal fluid .
namely , there is no dissipation of kinetic energy into heat , _
i.e. _ there is no heat production within the ideal fluid . by thermodynamics ,
if there is no heat production , we have @xmath146 ( @xmath147 : temperature ) .
then , @xmath148 where @xmath149 is the fluid pressure , @xmath150 is the enthalpy defined by @xmath151 , and @xmath152 denotes @xmath134 being fixed .
however , the entropy @xmath134 may not be uniform and depend on @xmath2 through the initial condition
. one of the aims of the present analysis is to show that the previous formulation ( kambe 2007 ) based on the symmetry of parallel translation alone is not sufficient to represent rotational flows with non - vanishing helicity .
the rotaional symmetry will be considered from 5 below .
writing the lagrangian density ( [ l - total ] ) as @xmath153 $ ] , we take variations of the variables @xmath99 , @xmath61 , @xmath134 and potentials @xmath154 and @xmath155 , where variations are assumed independent . substituting the varied variables @xmath156 , @xmath157 @xmath158 , @xmath159 and @xmath160 into @xmath161 $ ] and writing its variation as @xmath162
, we obtain @xmath163 where @xmath164 is used .
the variation fields are assumed to vanish on the boundary surface enclosing the domain @xmath15 , as well as at both ends @xmath165 and @xmath166 of time integration of the action @xmath142 ( where @xmath15 is chosen arbitrarily ) , as is usually the case .
thus , the action principle @xmath167 for independent variations @xmath168 , @xmath169 and @xmath170 results in @xmath171 using ( [ dvel ] ) and ( [ del - s ] ) , we have @xmath172 under this relation and ( [ del - s ] ) , the equation ( [ drho1 ] ) can be rewritten as @xmath173 this is regarded as an integral of motion . from the variations of @xmath174 and @xmath175 , @xmath176 { \delta}\psi & : & \hskip5 mm { \partial}_t(\rho s)+ \nabla\cdot(\rho s { \mbox{\boldmath $ v$ } } ) = 0\ , .
\label{dpsi1}\end{aligned}\ ] ] using ( [ dphi ] ) , the second equation reduces to the _ adiabatic _ equation : @xmath177 thus , we have obtained the continuity equation ( [ dphi ] ) and entropy equation ( [ dpsi2 ] ) from the action principle .
it is shown in kambe ( 2007 ) that the present solution is equivalent to the classical clebsch solution ( clebsch 1859 ) . with the velocity ( [ dvel ] )
, the vorticity @xmath178 is given by @xmath179 this implies that the vorticity is connected with non - uniformity of entropy . using ( [ drho2 ] )
, we find that the following euler s equation of motion is satisfied : @xmath180 in this case , the helicity vanishes ( bretherton , 1970 ) . for a general velocity field , the helicity @xmath181 defined by @xmath182 \equiv \int_v \
{ \mbox{\boldmath $ \omega$}}\cdot { \mbox{\boldmath $ v$}}\ { \mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$}}\label{hel - v}\ ] ] is a measure of linkage and knottedness of vortex lines , and does not vanish in general ( see sec.[8a ] ) . for a homentropic fluid
in which the entropy @xmath134 is a uniform constant @xmath183 at all points , we have @xmath184 , @xmath185 and @xmath186 from ( [ del - e ] ) since @xmath187 . in addition , the motion is _
irrotational_. in fact , from ( [ dvel ] ) , we have @xmath188 where @xmath189 , and @xmath190 from ( [ om - sp ] ) .
the integral ( [ drho2 ] ) becomes @xmath191 .
the euler s equation ( [ euler - eq1 ] ) reduces to @xmath192 note that the left hand side is the convective time derivative for the irrotational velocity @xmath193 : @xmath194 thus , as far as the action principle is concerned for a homentropic fluid , euler s equation of motion reduces to that for _ potential _ flows of a perfect fluid .
traditionally , this property is considered to be a defect of the formulation of the eulerian variation described in the previous section , because the action principle should yield the equations for rotational flows as well . in order to remove this ( apparent ) flaw , lin ( 1963 ) introduced the condition for the conservation of the identity of particles denoted by @xmath195 , which is represented by an additional subsidiary lagrangian of the form @xmath196 .
this introduces three potentials @xmath197 as a set of lagrange multipliers of conditional variation .
problem is that physical significance of @xmath198 is not clear ( bretherton 1970 ; seliger & whitham 1968 ; salmon 1988 ) .
let us recall that we have been considering the lagrangian @xmath199 satisfying the symmetry of parallel translation .
however , the flow field has another symmetry of rotational invariance ( kambe 2003a , b ) .
the equation ( [ dpsi2 ] ) implies that the entropy @xmath134 plays the role to identify each fluid particle .
furthermore , local rotation is captured by the expression ( [ om - sp ] ) . however , in a homentropic fluid , there is no such machinery to identify each fluid particle .
gauge invariance with respect to local rotation could be a candidate instead of @xmath134 .
insofar as the flow field is characterized by the translation symmetry alone , we have arrived at the above result , _
i.e. _ the flow field should be _ irrotational _ if the fluid is homentropic . the fluid motion is driven by the velocity potential @xmath200 , where @xmath201 with @xmath154 and @xmath155 the gauge potentials .
let us consider another variational formulation with the lagrangian represented by the particle coordinates @xmath202 .
independent variables are denoted with @xmath203 ( @xmath204 ) with @xmath205 being the time variable written as @xmath206 ( @xmath207 ) .
the letter @xmath206 is used instead of @xmath18 in combination with the particle coordinates @xmath208 ( @xmath30 ) .
the physical - space position of a particle is expressed by @xmath209 , or @xmath210 .
its velocity @xmath211 is given by @xmath212 .
the derivative @xmath213 is equivalent to the covariant derivative @xmath50 : @xmath214 mass density @xmath61 is defined by the relation , @xmath215 , where @xmath216 denotes the volume element @xmath217 . using the jacobian determinant @xmath218 of the transformation from @xmath1-space to @xmath219-space
, we obtain @xmath220 the total lagrangian ( [ lam - total ] ) with ( [ gldk ] ) and ( [ l - g0 ] ) is rewritten as @xmath221 the action @xmath142 is defined by integration of @xmath199 with respect to the time @xmath206 : @xmath222 in this definition , the third integral @xmath223 can be integrated with respect to @xmath206 , and expressed as @xmath224 { \mbox{\textrm{d}}}^3{\mbox{\boldmath $ a$}}$ ] , where @xmath225= \phi|_{\tau_2 } - \phi|_{\tau_1}$ ] is the difference of @xmath154 at the ends @xmath226 and @xmath227 and is independent of @xmath228 $ ] .
likewise , the fourth integral can be expressed as @xmath229 s\ , { \mbox{\textrm{d}}}^3{\mbox{\boldmath $ a$}}$ ] , because the entropy @xmath134 is independent of @xmath206 according to ( [ dpsi2 ] ) .
this means that the gauge potentials @xmath154 and @xmath155 do not appear in the variations of the action @xmath142 for @xmath228 $ ] .
in other words , the gauge potentials @xmath154 and @xmath155 make their appearance only in the action represented with the physical space coordinates @xmath2 . the fact that the last two lagrangians in ( [ lt-2 ] ) do not contribute to the variations ( for @xmath228 $ ] ) implies that local galilean symmetry is in fact an internal symmetry .
thus , by omitting the integrated terms , we have @xmath230 where @xmath231 , @xmath232 ( @xmath233 ) , and @xmath234 . in the @xmath1-space ,
the point @xmath1 is motion - less , while in the @xmath2-space , there is a flow of particles driven by the potentials @xmath154 and @xmath155 according to ( [ dvel ] ) .
the euler - lagrange equation can be writtten as @xmath235 ( @xmath236 ) .
the energy - momentum tensor @xmath237 is defined by @xmath238 where @xmath30 .
as long as ( [ el - x ] ) is valid together with @xmath239 , we have a conservation equation @xmath240 .
in fact , we have @xmath241 = 0\end{aligned}\ ] ] this is the noether theorem ( noether ( 1918 ) , weinberg ( 1995 ) ) . by using the lagrangian @xmath242 of ( [ i - ldt ] )
, we obtain @xmath243 .
the internal energy @xmath133 depends on the density @xmath61 which in turn depends on @xmath231 by ( [ jacob ] ) , where the entropy @xmath134 depends on @xmath244 only .
hence , we have @xmath245 by using the thermodynamic relation given in ( [ del - e ] ) . in view of ( [ jacob ] ) , we obtain @xmath246 thus , the energy - momentum tensor @xmath237 are found as follows ( eckart 1960 ) : @xmath247 introducing the symbols @xmath248 to denote @xmath249 cyclically , another twelve components are @xmath250 the noether conservation law @xmath240 reduces to the momentum equations ( see below ) for @xmath251 : @xmath252 where @xmath253 , and other two equations are obtained with cyclic permutation of @xmath249 . integrating this with respect to @xmath206 between the limits 0 and @xmath18
, we find the weber s transformation ( lamb 1932 , art.15 ) : @xmath254 where @xmath255 .
for @xmath256 , the conservation law @xmath240 describes the energy equation : @xmath257 + { \partial}_b \left [ p \frac{{\partial}(x , y , z)}{{\partial}(a , \tau , c ) } \right ] + { \partial}_c \left [ p \frac{{\partial}(x , y , z)}{{\partial}(a , b , \tau ) } \right ] = 0 .
\label{ceq-0}\ ] ] in the eulerian description with the independent variables @xmath258 instead of @xmath259 , this reduces to @xmath260 this can be transformed to the following form of conservation of energy : @xmath261=0 . \label{ceq-0-x}\ ] ] using ( [ ta0 ] ) and ( [ taa ] ) , the equation ( [ eqm - a-1 ] ) reduces to the equation for the acceleration @xmath262 : @xmath263 which is known as the lagrangian form of the equation of motion ( lamb 1932 , art.13 ) .
this is obtained by noting that @xmath264 and @xmath265 in view of ( [ dtu - a ] ) , the equation ( [ eqm - a-2 ] ) is equivalent to euler s equation of motion in the eulerian description of the barotropic relation @xmath266 : @xmath267 velocities @xmath268 in the @xmath1-space may be determined by ( [ web - tr ] ) with given initial conditions of @xmath269 and @xmath270 at @xmath271 . however , in order to transform it to the @xmath2-space , there is some freedom .
the equation ( [ web - tr ] ) is invariant by rotational transformations of a displacement vector @xmath272 since the left hand side is of the form of a scalar product with respect to @xmath273 , so that the vector @xmath273 is not uniquely determined ( the density @xmath61 and enthalpy @xmath150 are not changed by this transformation since it is volume - preserving ) .
the same freedom applies to the acceleration @xmath262 of ( [ eqm - a-2 ] ) as well .
rotational transformations are considered in the next section .
this formulation helps to eliminate the arbitrariness .
next , in order to equip the action @xmath142 with an additional mathematical structure of rotational symmetry ( to be considered in [ s6 ] ) , we summarize the rotational transformations here , and consider the rotational gauge symmetry .
the related gauge group is the rotation group @xmath274 .
an infinitesimal rotation is described by the lie algebra @xmath275 of three dimensions .
the basis vectors of @xmath275 are denoted by @xmath276 , which satisfy the commutation relations : @xmath277 = { \varepsilon}_{jkl}\ , e_l \ , , \label{rcomm}\ ] ] where @xmath278 is the completely skew - symmetric third - order tensor , and @xmath279,\hskip5 mm e_2 = \left [ \begin{array}{ccc } 0 & 0 & 1 \\ 0 & 0 & 0 \\ - 1 & 0 & 0 \end{array } \right],\hskip5 mm e_3 = \left [ \begin{array}{ccc } 0 & - 1 & 0 \\ 1 & 0 & 0 \\ 0 & 0 & 0 \end{array } \right ] .
\label{rt2}\ ] ] a rotation operator is defined by @xmath280 where @xmath281 ( @xmath30 ) are infinitesimal parameters : @xmath282\,.\ ] ] then , an infinitesimal rotation of the displacement vector @xmath283 is expressed by @xmath284 .
this is also written as @xmath285 , where @xmath286 is an infinitesimal angle vector .
preliminary studies of the rotational symmetry of fluid flows are given in kambe ( 2003a , b ) . however , more details are elaborated in the following subsections [ 5a ] @xmath287 [ 5c ] , and the section [ s6 ] proposes a new lagrangian @xmath9 to account for the rotation symmetry .
we consider local rotation of a fluid element about an arbitrary reference point @xmath288 within a fluid .
the non - abelian property of the rotational transformation requires infinitesimal analysis . at a neighboring point @xmath289 for a small @xmath290
, we consider its local transformation , @xmath291 .
rotational gauge transformation of a point @xmath290 in a frame @xmath63 ( an inertial frame , say ) to @xmath292 in a non - inertial frame @xmath64 is expressed as @xmath293 instead of ( [ lgtx ] ) for the translation .
the variation @xmath294 is defined by an infinitesimal rotation @xmath295 ( skew - symmetric ) applied to the displacement vector @xmath290 : @xmath296 .
the corresponding transformation of velocity ( @xmath297 ) is obtained from ( [ lrts ] ) as follows : @xmath298 instead of ( [ lgtu ] ) and ( [ lgxi ] ) where @xmath299 .
as before ( sec .
[ 2c - lgt ] ) , in this transformation the points @xmath290 and @xmath292 are the same point with respect to the frame @xmath63 .
this means that the coordinate frame @xmath64 is rotated by an angle @xmath300 with respect to @xmath63 with a fixed origin @xmath288 . at the origin @xmath288 ( @xmath301 ) , we have the following : @xmath302 from ( [ lrtv ] ) and ( [ lrxi ] ) , where @xmath303 .
let us consider the time - derivative @xmath304 in the euler equation ( [ euler1 ] ) for a homentropic fluid .
it is proposed that this term be replaced with @xmath305 as @xmath306 where @xmath307 is an operator defined by @xmath308 for scalar fields @xmath309 , and @xmath310 is an axial vector defined by @xmath311 @xmath312 .
is used here for brevity instead of bold - face three - vector @xmath99 . ]
the tensor @xmath313 is skew - symmetric ( @xmath314 ) .
we consider variations at the origin ( @xmath301 ) .
then , we have @xmath315 from ( [ dv0 ] ) .
variation of @xmath309 is defined by . ]
@xmath316 using @xmath317 from ( [ lgdt ] ) , we obtain the variation of @xmath305 : @xmath318 @xmath319 where @xmath320 can be eliminated by using the following equality obtained from ( [ rcomm ] ) : @xmath321 @xmath322 . setting @xmath323 , we obtain @xmath324 + 2{\partial}_t \theta^k\,e_k\,v = \hat{\theta } \times { \mbox{\textrm{d}}}_t v + 2\,{\partial}_t\hat{\theta } \times v. \label{dntv2}\ ] ] the second term of the last expression denotes the coriolis term in the rotating system .
thus , it is found that the covariant derivative @xmath305 behaves like the acceleration in a rotating frame , as far as @xmath307 transforms according to ( [ dom ] ) . in the previous analysis for the translation symmetry
, it was shown that the left hand side of ( [ euler1 ] ) , @xmath325 , is in fact equal to the particle acceleration @xmath326 of ( [ mdpot ] ) . in the present section taking into account the rotation symmetry ,
the partial derivative @xmath49 is replaced with the covariant derivative @xmath327 .
thus , the particle acceleration @xmath328 should be defined by @xmath329 where @xmath310 is the gauge field with respect to the rotation symmetry .
it can be verified that the gauge field @xmath330 in fact coincides with the vorticity @xmath331 by the requirement of galilean invariance of the covariant derivative @xmath332 of ( [ cdrot ] ) , as follows . under a global galilean transformation from one frame @xmath333 to another @xmath334 which is moving with a uniform velocity @xmath65 relative to @xmath63 , the position vector @xmath335 and velocity vector @xmath336 in the frame @xmath337
are given by @xmath338 and @xmath339 .
transformation laws of derivatives are @xmath340 and @xmath341 . applying these and replacing @xmath99 by @xmath342 , the covariant derivative of ( [ cdrot ] )
is transformed to @xmath343 since @xmath65 is a constant vector and @xmath344 .
we require the covariance , _
i.e. _ @xmath345 .
namely , the right hand side should be equal to @xmath346 .
therefore , @xmath347 on using the following vector identity : @xmath348 , for the constant vector @xmath65 .
equation ( [ gfgi ] ) is satisfied identically , if @xmath349 the second relation holds by the galilean transformation since @xmath350 and @xmath351 .
thus , the galilean invariance of @xmath332 results in @xmath352 namely , the gauge field @xmath330 _ coincides _ with the vorticity @xmath178 .
consequently , the covariant derivative @xmath332 ( particle acceleration ) is given by @xmath353 using the vector identity @xmath354 the last expression ( [ cdrot * ] ) is transformed to the convective derivative of @xmath99 ( _ i.e. _ the particle acceleration ) : @xmath355 thus , we have @xmath356 . so far , @xmath289 was a local position vector around a fixed point @xmath288 . in general , using @xmath2 in place of @xmath289 , we arrive at the same definition : @xmath357 .
associated with the rotation symmetry , we try to introduce an additional lagrangian according to the gauge principle .
the property of @xmath358 and @xmath359 considered below the equation ( [ lt-2 ] ) suggests how to find such a lagrangian .
it was seen there that the lagrangians represented with the particle coordinates @xmath202 can be integrated with respect to the time @xmath206 and eliminated in the action to deduce the euler - lagrange equation , while they are non - trivial in the physical - space coordinates @xmath360 because of the non - trivial jacobian matrix @xmath361 of the transformation . as a result ,
the action principle yields the continuity equation and entropy equation .
thus the gauge potentials are regarded as _
agents with mechanical functions that drive fluid motions in the physical @xmath2-space , whereas they disappear in motionless @xmath1-space .
it is important to observe that , in the lagrangian ( [ lt-2 ] ) , the integrands of the last two integarals are of the form @xmath362 , because the entropy @xmath134 and mass element @xmath363 are independent of @xmath206 .
the action is defined by @xmath364\ , { \mbox{\textrm{d}}}\tau { \mbox{\textrm{d}}}^3{\mbox{\boldmath $ a$}}$ ] , where @xmath365 denotes the part of the first two terms of ( [ lt-2 ] ) .
this property is the simplest case of representation of topology in the gauge theory . in the context of rotational flows
, it is known that the helicity ( or hopf invariant , arnold & khesin ( 1998 ) ) defined by ( [ hel - v ] ) describes non - trivial topology of vorticity field , _ i.e. _ mutual linking of vorticity lines .
this is closely related with the chern - simons term in the gauge theory .
the chern - simons term lives in one dimension lower than the original four - space - time ( @xmath366 ) of the action @xmath142 because a topological term in the action is expressed in a form of total divergence ( @xmath367 ) and characterizes topologically non - trivial structures of the gauge field ( chern 1979 ; jackiw 1985 ; desser _ et al .
_ 1982 ) . here , we follow the formulation of sec.[4a ] and look for a @xmath206-independent field directly .
the @xmath206-independent field can be found immediately from eq .
( [ eqm - a-1 ] ) . taking the curl of this equation with respect to the coordinates @xmath249
, we obtain @xmath368 since @xmath369 , where @xmath370 , and @xmath371 is defined by ( [ ta0 ] ) .
it is found that @xmath372 is independent of @xmath206 .
, satisfying the mass invariance @xmath373 . ]
hence , one may write as @xmath374 , where the right hand side is a vector depending on @xmath1 only .
it is useful to formulate this property in the framework of exterior differential forms and lie derivative .
vector @xmath375 is understood as a transformed form of the velocity @xmath376 into the @xmath1-space .
this is seen on the basis of a 1-form @xmath377 defined by @xmath378 where @xmath379 , @xmath380 , @xmath381 , _ etc._and @xmath382 .
hence @xmath383 is equivalent to ( [ ta0 ] ) .
its differential @xmath384 gives a two - form @xmath385 : @xmath386 where @xmath387 , and @xmath388 is the vorticity .
thus , it is seen that @xmath389 is the vorticity transformed to the @xmath1-space , and the 2-forms @xmath390 and @xmath391 are surface forms defined in the footnote , and @xmath392 . ] .
the equation ( [ dtom - a ] ) is transformed to the @xmath206-derivative of the 2-form @xmath393 , which is understood as the lie derivative : @xmath394 .
its representation in the @xmath395 space is given as follows .
it is useful to define two _ tangent vectors _ @xmath396 and @xmath397 ( tangent to the flow generated by @xmath99 ) : @xmath398 where @xmath397 is regarded as a tangent vector dual to the 1-form @xmath377 of ( [ v1xx ] ) .
the lie derivative @xmath399 can be defined by @xmath400 according to the symbols of differential calculus and tangent vector , where @xmath401 is also a differential operator along a streamline generated by the flow @xmath402 . by applying the lie derivative @xmath399 on the 2-form @xmath393 ( frankel 1997 , ch.4.2 ) , we obtain @xmath403 \cdot { \mbox{\boldmath $ s$}^2 } + ( \nabla\cdot { \mbox{\boldmath $ v$ } } ) { \mbox{\boldmath $ \omega$}}\cdot { \mbox{\boldmath $ s$}^2 } \nonumber \\ = \
{ \mbox{$\mathcal{l}$}}_w { \omega}^2 \hskip11 mm & = & e_w[{\mbox{\boldmath $ \omega$ } } ] \cdot { \mbox{\boldmath $ s$}^2 } , \label{ldom}\end{aligned}\ ] ] where @xmath404 $ ] is the lie bracket defined by @xmath405 , and @xmath406 \equiv { \partial}_t { \mbox{\boldmath $ \omega$}}+ [ { \mbox{\boldmath $ v$}},{\mbox{\boldmath $ \omega$ } } ] + ( \nabla\cdot { \mbox{\boldmath $ v$ } } ) { \mbox{\boldmath $ \omega$}}= { \partial}_t { \mbox{\boldmath $ \omega$}}+ \nabla \times ( { \mbox{\boldmath $ \omega$}}\times { \mbox{\boldmath $ v$ } } ) , \label{veq}\ ] ] where @xmath407 is used . thus , we obtain the vorticity equation from ( [ ldom ] ) : @xmath408 next , let us introduce a gauge vector - potential @xmath409 in the @xmath1-space , and define its 1-form @xmath410 by @xmath411 the exterior product of @xmath410 and @xmath393 results in a volume form @xmath412 multiplied by a scalar @xmath413 : @xmath414 where @xmath415 is a scalar product of @xmath416 and @xmath389 . in the @xmath417 space , the same exterior product
is given by @xmath418 it is obvious that the scalar product @xmath419 is invariant under local rotational transformations in the @xmath2 space .
thus , it is proposed that a _ possible _ type of lagrangian is of the form , @xmath420,\ ] ] where the time derivative @xmath421 was taken out of the inner - product since @xmath389 is independent of @xmath206 .
it follows that @xmath0 can be dropped in the action @xmath142 expressed with the coordinate @xmath1 except the integrated terms depending on the initial and final times of @xmath206-integration .
however , this lagrangian becomes non - trivial when transformed to the physical - space coordinate @xmath2 because of the non - trivial jacobian of transformation .
it is shown later that @xmath416 may be understood as a kind of vector potential associated with the gauge field @xmath389 . from the equality of ( [ dvphi3 ] ) and ( [ dvphi4 ] )
, we have @xmath422 . using the definition ( [ lda - x ] ) , @xmath423 = { \mbox{$\mathcal{l}$}}_{w}\,\left [ a^1 \wedge { \omega}^2 \right]= \left[{\mbox{$\mathcal{l}$}}_{w } \,a^1 \right ] \wedge { \omega}^2,\ ] ] where ( [ ldom ] ) was used . with the same calculus as in ( [ dvphi4 ] ) ,
the right hand side is given by @xmath424 \wedge { \omega}^2 = \langle { \mbox{$\mathcal{l}$}}_w { \mbox{\boldmath $ a$}},\ , { \mbox{\boldmath $ \omega$}}\rangle \,{\mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$}}$ ] , where the lie derivative of a 1-form @xmath410 is defined by @xmath425 ( _ e.g. _ frankel ( 1997 ) ,
4.2c ) is a scalar function , @xmath426 . ] .
thus , the lagrangian @xmath0 in the @xmath2 variables is @xmath427 integrating this by parts and using the relation @xmath428 , we have @xmath429 where int@xmath430 denotes the integral over the surface @xmath431 bounding @xmath15 .
this is the proposed lagrangian @xmath0 in the physical space @xmath2 , where @xmath432 the lagrangian @xmath0 of ( [ la - x1 ] ) is an integral of @xmath433 $ ] .
given a vector field @xmath434 , the value @xmath435 $ ] taken on the vector @xmath434 measures the derivative of @xmath410 ( as one moves along the trajectory of @xmath396 , _
i.e. _ the particle path ) evaluated on a vector field @xmath434 frozen to the flow generated by @xmath396 . using ( [ bi ] ) in ( [ la - x1 ] ) and carrying out integration by parts , we obtain @xmath6 \,\rangle \,{\mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$ } } , \label{la - x4}\ ] ] with @xmath8
$ ] defined in ( [ veq ] ) , where the integrated terms are dropped . here
, we redefine @xmath0 by this equation instead of ( [ la - x1 ] ) .
hence , the expressions ( [ la - x1 ] ) and ( [ la - x2 ] ) must be modified by additional terms of surface integral over @xmath431 , which are to be expressed symbolically with int@xmath430 .
this lagrangian ( [ la - x4 ] ) provides a source term of non - vanishing helicity .
an example of such a flow will be given in the last section [ 8c ] .
the lagrangian @xmath0 of ( [ la - x2 ] ) should be included in the total lagrangian to describe rotational flows of an ideal fluid .
accordingly , the lagrangian density of the total lagrangian ( excluding surface terms ) is defined by @xmath436 & \equiv & \half\,\rho\,\langle\,{\mbox{\boldmath $ v$ } } , { \mbox{\boldmath $ v$}}\,\rangle - \rho\,{\epsilon}(\rho , s ) -\rho\ , ( { \partial}_t+{\mbox{\boldmath $ v$}}\cdot\nabla)\phi \nonumber \\[1 mm ] & & \hspace*{-5 mm } -\rho s\ , ( { \partial}_t+{\mbox{\boldmath $ v$}}\cdot\nabla)\psi -\rho\,\langle \,{\mbox{\boldmath $ b$}},\ , { \mbox{\boldmath $ v$}}\,\rangle . \label{ld - ursps}\end{aligned}\ ] ] as before ( sec . [ 3b ] ) , we take variations of the field variables @xmath99 , @xmath61 , @xmath134 and potentials @xmath154 and @xmath155 . in addition , we must now include the variation of the new variable @xmath7 .
independent variations are taken for those variables . substituting the varied variables @xmath156 , @xmath157 , @xmath158 , @xmath159 , @xmath160 and @xmath437 into @xmath438 $ ] and writing its variation as @xmath162
, we obtain @xmath439 where remaining terms denoted by @xmath440 are omitted since they are the same as ( [ d - ld ] ) . the new term @xmath441 is defined by @xmath442 thus , the variational principle , @xmath167 for independent variations of @xmath168 , @xmath169 , @xmath443 , _ etc . _ results in @xmath444 where the omitted terms @xmath440 are the same as before .
thus , the continuity equation ( [ dphi ] ) and entropy equation ( [ dpsi1 ] ) are obtained as before .
the equation ( [ dvel - r ] ) gives an expression for the velocity @xmath99 : @xmath445 the vorticity @xmath428 is represented by @xmath446 the second term @xmath447 is a new term leading to non - vanishing vorticity in a homentropic fluid of uniform @xmath134 .
the equation ( [ da - om ] ) is the same as ( [ veq0 ] ) by ( [ veq ] ) .
_ potential part and rotational part _ : the potential part of the present problem is described by ( [ drho - b ] ) , and the rotational part by ( [ veq0 ] ) . in order to show that this is consisitent with the traditional formulation , suppose that the vorticity equation ( [ da - om ] ) is solved and a velocity field @xmath77 is found . in view of the relation @xmath448 ,
the velocity @xmath99 should satisfy the following : @xmath449 for a scalar function @xmath450 .
obviously , taking the curl of this equation reduces to ( [ veq0 ] ) . on the other hand , according to sec.[4b ] , we have the euler s equation of motion ( [ eeq ] ) , because the new lagrangian @xmath0 has no effect in the total lagrangian of @xmath1 space owing to its construction , and hence generates no new term . using the identity ( [ vi - s63 ] )
, this reduces to the following : @xmath451 defining @xmath452 by @xmath453 , the equation ( [ drho - b ] ) is @xmath454 by using ( [ vel - a ] ) . in fact , it can be shown that both expressions ( [ ee-9a ] ) and ( [ ee-9b ] ) are equivalent with this @xmath452 , as follows .
first we note the following identities : @xmath455 the second is an identity with @xmath99 given by ( [ vel - a ] ) . using these
, one can verify the following : @xmath456 = { \partial}_t { \mbox{\boldmath $ v$}}+ ( { \mbox{\boldmath $ v$}}\cdot \nabla ) { \mbox{\boldmath $ v$}}+\nabla h.\ ] ] the right hand side vanishes by eq.([eeq ] ) .
therefore , we obtain @xmath457 . \label{eq - w}\ ] ] it is not difficult to show that the equation ( [ ee-9a ] ) reduces to this equation for the above @xmath452 and the total velocity @xmath99 of ( [ vel - a ] ) , where the total vorticity @xmath178 is given by ( [ vor - a ] ) . the equation ( [ eq - w ] ) is a new equation to be satisfied by the rotational component @xmath441 . if @xmath458 ( uniform ) , then @xmath459 ( @xmath460 ) , and @xmath461 . subtracting ( [ eq - w ] ) from ( [ ee-9b ] ) , we obtain @xmath462=0 $ ] , which is equivalent to ( [ drho8 ] ) .
thus there is no contradiction between the traditional formulation and the present one .
new aspects of the present formulation are described in the next section , in addition to the expression ( [ eq - w ] ) for the rotational component @xmath441 .
we consider here the significance of the lagrangian @xmath0 associated with rotation symmetry and its outcome derived from the variational principle .
let us recall that @xmath463 is defined by ( [ la - x3 ] ) . using ( [ bi ] ) ,
the expression of @xmath464 is rewritten as @xmath465 therefore , defining a vector @xmath466 with @xmath467 , we have @xmath468 . \label{rbb}\ ] ] variation of @xmath0 of ( [ la - x2 ] ) with respect to a variation @xmath469 is given by @xmath470 \,{\mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$}},\ ] ] where @xmath471 from ( [ rbb ] ) for the variation @xmath469 on keeping @xmath7 ( hence @xmath466 ) fixed .
therefore , we have @xmath472 this leads to the following expression for @xmath441 defined by ( [ dla - v ] ) : @xmath473 + { \mbox{\boldmath $ \omega$}}\times { \mbox{\boldmath $ b$ } } , \label{bw}\ ] ] where @xmath474 \equiv { \partial}_t{\mbox{\boldmath $ b$}}+ \nabla \times ( { \mbox{\boldmath $ b$}}\times { \mbox{\boldmath $ v$}})$ ] from ( [ veq ] ) . for flows of a homentropic fluid with constant entropy @xmath183 , the velocity is given by ( [ vel - a ] ) : @xmath475 ( @xmath476 ) .
therefore , the helicity is @xmath477 \ , { \mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$ } } , \label{hlcty}\ ] ] since @xmath478 .
thus it is found that the term @xmath474 $ ] generates the helicity .
transformation from the lagrangian @xmath1 space to eulerian @xmath479 space is determined locally by nine components of the matrix @xmath480 . however , in the previous solution considered in sec.[4b ] , we had three relations ( [ web - tr ] ) between the @xmath2-velocity @xmath481 and the @xmath1-velocity @xmath482 , and another three relations ( [ eqm - a-2 ] ) between the @xmath2-acceleration @xmath483 and the @xmath1-acceleration @xmath484 .
these six relations are not sufficient to determine the nine matrix elements @xmath480 .
but now , we are equipped with a new structure of rotational symmetry which is expected to resolve this indefiniteness .
in fact , the remaining three conditions are found from the equation ( [ om2ss ] ) connecting the @xmath2-vorticity @xmath485 and the @xmath1-vorticity @xmath486 . from ( [ om2ss ] ) , these are @xmath487 with other two equations for @xmath488 and @xmath489 determined cyclically .
there are three vectors ( velocity , acceleration and vorticity ) determined by evolution equations subject to initial conditions in each space of @xmath2 and @xmath1 coordinates .
transformation relations of the three vectors suffice to determine the nine matrix elemets @xmath480 locally .
thus , the transformation between the lagrangian @xmath1 space and eulerian @xmath479 space is determined uniquely . in this sense ,
the equation of the vorticity is essential for the uniqueness of the transformation between lagrangian and eulerian coordinates . as an example of flows of non - zero helicity
, we consider a particular flow field induced by tangling of two vortices , _
i.e. _ a rectilinear line - vortex l linked with a vortex ring r of infinitesimal vortex - cores , whose cross - sections are both circular and strengths are @xmath490 and @xmath491 respectively .
the helicity @xmath181 is immediately found as @xmath492 ( see below ) , hence it is non - zero . therefore , this flow field can not be given a representation of a clebsch - type form .
we consider a flow of an incompressible fluid of density @xmath493 in a cylindrical coordinate frame @xmath494 .
suppose that we have a rectilinear line - vortex l of strength @xmath495 coinciding with the @xmath26-axis ( axis of cylindrical symmetry ) and a vortex ring r of strength @xmath496 which coincides temporarily with a circle of radius @xmath497 in the plane @xmath498 centered at the origin . by symmetry consideration
, it is readily seen for this initial configuration that the vortex ring @xmath497 simply translates parallel to the @xmath26 axis with a constant velocity @xmath499 ( say ) without change of its form and its radius ( apart from rotaional motion around the @xmath26 axis without changing its form ) , while the rectilinear vortex @xmath242 stays at the same position without change . with respect to this coordinate frame , the velocity field ( at the initial instant ) of the rectilinear vortex
l is represented by a vector potential @xmath500 , while that of the vortex ring r is given by @xmath501 .
explicit expressions of velocity are @xmath502 and @xmath503 respectively , and @xmath504 their vorticities are @xmath505 where @xmath506 is the delta function , and @xmath507 denotes the cartesian coordinates in the plane perpendicular to the @xmath26-axis . hence , the helicity is given by @xmath508\ , { \mbox{\textrm{d}}}^3{\mbox{\boldmath $ x$}}.\ ] ] by elementary calculus , we find that the integral of each term gives the same value @xmath509 .
thus we obtain non - zero helicity @xmath510 ( moffatt 1969 ) .
the potential @xmath467 can be determined in the following way . in reference to the coordinate frame moving with the vortex ring r , the velocity field is steady ( and the vorticity too ) . in this frame
, a uniform velocity @xmath511 is added , and total velocity is given by @xmath512 where @xmath513 , independent of @xmath18 . in the equation ( [ bw ] ) : @xmath514 + { \mbox{\boldmath $ \omega$}}\times { \mbox{\boldmath $ b$}}$ ] , the vorticity @xmath178 is nonzero only on the rectilinear line l and the circular curve r , and vanishes ( @xmath190 ) at all other points . at points
neither on l nor on r , we have the equation : @xmath515 , \quad e_w[{\mbox{\boldmath $ b$}}]= { \partial}_t{\mbox{\boldmath $ b$}}+ \nabla \times({\mbox{\boldmath $ b$}}\times { \mbox{\boldmath $ v$ } } ) .
\label{b - lr}\ ] ] let us seek a general solution to this equation for a given velocity field @xmath4 .
suppose that we have a particle - path solution @xmath516 satisfying @xmath517 , where @xmath518 for @xmath519 and @xmath520 .
next , we introduce a particular field @xmath521 defined by @xmath522 which satisfies the equation : @xmath523= { \partial}_t{\mbox{\boldmath $ b$ } } _ * + \nabla \times({\mbox{\boldmath $ b$ } } _ * \times { \mbox{\boldmath $ v$}})=0 $ ] ( cauchy s solution ) for an arbitrary initial field @xmath524 .
a general solution to the equation ( [ b - lr ] ) is given by @xmath525 where @xmath526 is a time - independent field and a particular solution to ( [ b - lr ] ) .
in fact , substituting ( [ gb - lr ] ) into ( [ b - lr ] ) , we have @xmath527 , in view of ( [ v - chi ] ) and @xmath523= 0 $ ] .
thus , we obtain @xmath528 where @xmath155 is a scalar function satisfying @xmath529 . then , the solution @xmath530 is given by @xmath531
following the scenario of the gauge principle of field theory , it is found that the variational principle of fluid motions can be reformulated successfully in terms of a covariant derivative and lagrangians defined appropriately .
the lagrangians are determined such that a gauge invariance is satisfied under local galilean transformations including the rotational symmetry .
the fluid material is characterized thermodynamically in terms of mass density and entropy density .
in addition , local rotation ( _ i.e. _ vorticity ) of the material is also taken into account by a newly - introduced lagrangian @xmath0 .
a gauge - covariant derivative @xmath532 was defined in the present formulation including the rotational symmetry ( it was @xmath50 up to 5 for the translation symmetry alone ) .
galilean invariance requires that the covariant derivative should be the _ convective _ time derivative following the motion of the background material , _
i.e. _ the so - called lagrange derivative . the covariant derivative is an essential building block of the gauge theory . applying @xmath532 to the position
coordinate @xmath533 of a fluid particle @xmath1 , we obtain the particle velocity @xmath99 .
it is found that the vorticity @xmath178 is in fact a gauge field associated with the rotational symmetry .
lastly , an example of flow with non - vanishing helicity is presented together with a general solution of the vector potential for the rotational part . the variational principle ( _ i.e. _ the action principle ) in the space of particle coordinate @xmath1 results in the euler - lagrange equation
. this yields the noether thorem , _
i.e. _ conservation equations of energy - momentum .
the equations thus obtained are the lagrangian form of the equation of motion and energy equation , which are transformed to euler s equation of motion and the conservation equation of total energy @xmath534 in eulerian space . the variational principle with respect to eulerian space coordinate @xmath2 results in ( i ) the expression of velocity @xmath99 in terms of gauge potentials ,
( ii ) an integral describing a potential part of the velocity , ( iii ) the equation of continuity , ( iv ) the entropy equation . in addition to these results obtained in the previous case of translation symmetry alone ,
new results in the present formulation are as follows .
( v ) the velocity @xmath99 includes a new rotational term @xmath441 , ( vi ) the vorticity equation ( [ da - om ] ) is deduced from the variational principle and ( vii ) the equation ( [ eq - w ] ) is derived for the rotational component @xmath441 .
it is noteworthy that the new lagrangian @xmath0 of ( [ la - x4 ] ) yields a source term for the helicity defined by ( [ hlcty ] ) .
transformation from the lagrangian @xmath1 space to eulerian @xmath2 is determined locally by nine components of the matrix @xmath480 .
however , in the solution obtained previously ( kambe 2007 ) without taking account of rotational symmetry , we had only six relations for the velocity ( [ web - tr ] ) and acceleration ( [ eqm - a-2 ] ) including the matrix element @xmath480 .
both of the equations are invariant under rotational transformations of a displacement vector @xmath272 , so that the vector @xmath273 is not uniquely determined . in the improved formulation taking account of the rotational symmetry with three additional equations from ( [ om2ss ] ) , there are three vectors ( velocity , acceleration and vorticity ) determined by evolution equations subject to initial conditions in each space of @xmath2 and @xmath1
. transformation relations of the three vectors suffice to determine the nine matrix elements @xmath480 locally .
thus , the transformation between the lagrangian @xmath1 space and eulerian @xmath479 space is determined uniquely . in this sense ,
the equation ( [ omx - oma ] ) for the vorticity is essential to the uniqueness of the transformation between lagrangian and eulerian coordinates .
accordingly , the present variational formulation is self - consistent and comprehensively describes flows of an ideal fluid .
the author is grateful to professor antti niemi for discussions on the chern - simons term of the gauge theory , and also deeply appreciates the careful comments by professor alex craik on the original manuscript , which have led to significant improvement of the expressions .
he is also much obliged to the referees and the editor for the comments to improve the manuscript .
* appendix : background of the theory [ a - back ] * this appendix reviews the background of the theory and describes the scenario of the gauge principle in a physical system .
in the theory of electromagnetism , it is well - known that there is an invariance under a _ gauge transformation _ of electromagnetic potentials consisting of a scalar potential @xmath154 and a vector potential @xmath7 .
the electric field @xmath535 and magnetic field @xmath466 are represented as @xmath536 , and @xmath537 , where @xmath538 is the light velocity .
the idea is that the fields @xmath535 and @xmath466 are unchanged by the following transformation : @xmath539 , where @xmath540 , @xmath541 with @xmath542 being an arbitrary differentiable scalar function of position @xmath2 and time @xmath18 .
gauge invariance is also known in certain rotational flows of the clebsch representation ( eckart 1960 ) .
its velocity field is represented as ( sec.[3b ] ) : @xmath543 where @xmath544 and @xmath134 are scalar functions , and @xmath155 satisfies the equation @xmath545 .
let @xmath546 and @xmath547 be a second set giving the same velocity field @xmath99 , which implies the following : @xmath548 .
this relation will hold true by the relation @xmath549 , if @xmath550 and @xmath551 .
therefore , @xmath134 and @xmath155 are determined only up to such a _ contact _ transformation , and @xmath154 transforms by the addition of the generating function @xmath552 .
it can be shown that the two sets of triplet give the identical flow field .
thus , there exist some freedom in the expressions of physical fields in terms of potentials .
in _ quantum mechanics _ , schrdinger s equation for a charged particle of mass @xmath553 and electromagnetic fields are invariant with respect to a gauge transformation .
this is as follows . in the _ absence _ of electromagnetic fields , schrdinger s equation for a wave function @xmath155 of a free particle @xmath553
is written as @xmath554 \equiv i \hbar\ , { \partial}_t \psi - \frac{1}{2m}\ p_k^2 \
\psi= 0 $ ] , where @xmath555 is the momentum operator ( @xmath556 for @xmath30 ) . in the _ presence _ of electromagnetic fields , schrdinger s equation for a particle with an electric charge @xmath557
is obtained by the transformation : @xmath558 where @xmath559 is a four - vector potential , actually a covector denoted with lower indices . a point in space - time
is denoted by a four vector with upper indices , @xmath560 with @xmath561 .
thus , we obtain the equation with electromagnetic fields : @xmath562 \ \equiv \ i \hbar\ , { \partial}_t \psi - e\phi\,\psi - \frac{1}{2m}\ , \big [ ( -i \hbar ) ( { \partial}_k + \frac{e}{i\hbar c } a_k ) \big]^2 \
\psi = 0.\ ] ] from the four potential @xmath563 , we obtain @xmath535 and @xmath466 .
suppose that a wave function @xmath564 satisfies the equation @xmath565=0 $ ] , and consider the following set of transformations of @xmath564 and @xmath566 : @xmath567\ \psi(x^\mu ) , \label{s82-psi } \\ a_k & \to & a_k'=a_k+{\partial}_k \beta , \hskip8 mm \phi \to \phi'=\phi - c^{-1 } { \partial}_t \beta , \end{aligned}\ ] ] where @xmath568 .
then , it is shown readily that the transformed function @xmath569 satisfies the schrdinger equation @xmath570=0 $ ] .
this is the _ gauge invariance _ of the system of an electric charge in electromagnetic fields .
the system is said to have a gauge symmetry .
a _ relativistic _ lagrangian ( _ i.e. _ lorentz - invariant ) is defined for fluid flows by soper ( 1976 ) , and the energy - momentum tensor @xmath12 is derived from it , where @xmath571 with 0 denoting the time component . in this relativistic formulation , the continuity equation and an energy equation of fluid flows
are derived from the conservation of particle numbers and an equation of energy conservation @xmath572 .
expanding the energy density @xmath573 and energy flux @xmath574 with respect to an infinitesimal ratio @xmath575 ( @xmath30 , where @xmath538 is the light velocity and @xmath27 a representative fluid velocity ) , the continuity equation is obtained from the lowest order terms ( equivalent to the conservation of particle number ) , while nonrelativistic energy conservation is obtained from the next order terms proportional to @xmath576 .
the entropy equation is another form of the energy conservation .
in the gauge theory of particle physics ( weinberg 1995 ; frankel 1997 ) , a free - particle lagrangian @xmath577 $ ] is defined first for the wave function @xmath564 of a particle with an electric charge .
let us consider the following gauge transformation : @xmath578 .
if @xmath579 is invariant under the transformation when @xmath580 is a constant , it is said that @xmath579 has a _
global _ gauge invariance . in spite of this
, it often happens that @xmath579 is _ not _ invariant for a function @xmath581 , _ i.e. _ @xmath579 is not gauge - invariant _ locally_. in this circumstance , it is instrumental to introduce a new field in order to acquire local gauge invariance .
if the new field ( a gauge field ) is chosen appropriately , local gauge invariance can be recovered . in the section
b , the _ local _ gauge - invariance was acquired by replacing @xmath582 with @xmath583 ( see ( [ s12 - 0 ] ) ) , where @xmath584 , and @xmath585 is the electromagnetic potential ( and termed a _ connection form _ in mathematics ) . the operator @xmath586 is called the _ covariant derivative_. thus , when the original lagrangian is not locally gauge invariant , the _ principle of local gauge invariance _ requires a new gauge field to be introduced in order to acquire local gauge invariance , and the lagrangian is to be altered by replacing the partial derivative with the _ covariant _ derivative including the gauge field .
this is the _
weyl s gauge principle_. in mathematical terms , suppose that we have a group @xmath587 of transformations and an element @xmath588 ( for @xmath589 with @xmath15 a space where @xmath155 is defined ) and that the wave function @xmath590 is transformed as @xmath591 . in the previous example , @xmath592 and the group is @xmath593 . introducing the gauge field @xmath594 allows us to define a covariant derivative @xmath595 as a generalization of the partial derivative @xmath596 that transforms as @xmath597 .
if we operates the right hand side on @xmath155 , we obtain @xmath598 where @xmath599 .
thus , we obtain @xmath600 , showing that @xmath601 transforms in the same way as @xmath155 itself .
frankel t 1997 .
_ the geometry of physics _ _ an introduction_. cup .
herivel jw 1955 .
the derivation of the equations of motion of an ideal fluid by hamilton s principle .
cambridge phil .
soc . _ * 51 * , 344 - 349 .
lin cc 1963 .
hydrodynamics of helium ii , in _
liquid helium_. proc .
`` enrico fermi '' xxi ( ny : academic press ) , 93 - 146 .
moffatt hk 1969 .
_ j. fluid mech . _ * 35 * , 117 .
noether e 1918 .
invariante variationsproblem .
_ klg - ges .
gttingen , math . physik _
* kl , 2 * , 235 . | on the basis of the gauge principle of field theory , a new variational formulation is presented for flows of an ideal fluid .
the fluid is defined thermodynamically by mass density and entropy density , and its flow fields are characterized by symmetries of translation and rotation .
the rotational transformations are regarded as gauge transformations as well as the translational ones .
in addition to the lagrangians representing the translation symmetry , a structure of rotation symmetry is equipped with a lagrangian @xmath0 including the vorticity and a vector potential bilinearly .
euler s equation of motion is derived from variations according to the action principle .
in addition , the equations of continuity and entropy are derived from the variations .
equations of conserved currents are deduced as the noether theorem in the space of lagrangian coordinate @xmath1 . without @xmath0 , the action principle results in the clebsch solution with vanishing helicity .
the lagrangian @xmath0 yields non - vanishing vorticity and provides a source term of non - vanishing helicity . the vorticity equation is derived as an equation of the gauge field , and the @xmath0 characterizes topology of the field .
the present formulation is comprehensive and provides a consistent basis for a unique transformation between the lagrangian @xmath1 space and the eulerian @xmath2 space .
in contrast , with translation symmetry alone , there is an arbitrariness in the transformation between these spaces .
( 2008 ) , accepted on 1 december 2007 . received 6 february 2007 ; revised 23 november 2007 ; accepted 1 december 2007 + communicated by y. fukumoto .
gauge priciple , variational formulation , ideal fluid , rotation symmetry , chern - simons term |
the electron spin transport is playing growing role in the fastly developing new directions of the modern high - tech electronics , the so - called * spintronics * @xcite .
in most solids the conduction electron system is coupled by hyperfine interaction @xcite to the underlying system of nuclear spins , which may profoundly modify the electron spin transport at low temperatures , thus leading to * nuclear spintronics * effects , which will be described in this review . on the other hand ,
the application of high magnetic fields is a very powerful tool for studying the electronic properties of a large variety of metals , semiconductors and superconductors . due to the landau quantization of electron motion in sufficiently strong magnetic field ,
most of the transport properties , such as magnetization , conductivity etc
. experience magnetic quantum oscillations ( qo ) @xcite .
the landau quantization is most spectacularly manifest in the electronic magneto - transport in low dimensional conductors .
striking examples are the celebrated quantum hall effects ( qhe ) @xcite .
apart from the anomalous enhancement of the well known qo in two - dimensional electron systems ( 2des ) one expects also strong qo in physical properties which are not sensitive to the magnetic field in isotropic three dimensional metals .
it was suggested in @xcite that in quasi - two - dimensional metals under strong magnetic fields the nuclear spin lattice relaxation rate @xmath0 should exhibit strong magnetic oscillations .
this should be compared with the korringa relaxation law @xcite usually observed in three - dimensional normal metals , which results in a magnetic field independent nuclear spin - relaxation rate .
this line of research seems to be useful in dense quasi - two - dimensional electronic systems , as is the case of synthetic metals ( gic s etc . ) and low - dimensional organic compounds . a completely new line of research , the hyperfine interaction between nuclear and electron spins in low dimensional and correlated electron systems , has been developed during the two last decades both theoretically [ 8 - 39 ] , and experimentally [ 40 - 59 ] .
the growing attention is attracted by the hyperfine interactions in a ) qhe : theory [ 8 - 19 ] and experiment [ 40 - 53 ] ; b ) mesoscopics and nano - systems : theory [ 20 - 26 ] and experiment [ 54 - 55 ] ; c ) normal and superconducting metals : theory [ 27 - 31 ] and experiment [ 58 - 59 ] .
very recently nuclear spin memory devices @xcite and the indirect hyperfine interaction via electrons between nuclear spin qubits in semiconductor based quantum computer proposals [ 33 - 39 ] attracts sharply growing attention . here
i will outline the main theoretical concepts and some experimental achievements in the new and quickly developing field of * nuclear spintronics*.
in metals and doped semiconductors , usually , the leading contribution to the spin - lattice relaxation process is due to the _ hyperfine fermi contact _ interaction between the nuclear spins and the conduction electron spins @xcite .
this interaction is represented by the hamiltonian : @xmath1 where @xmath2 is the nuclear gyromagnetic ratio , @xmath3 is the nuclear spin and @xmath4 is the magnetic field on the nuclear site , produced by electron orbital and spin magnetic moments : @xmath5 here @xmath6 is the electron radius - vector , @xmath7 is the electron spin operator , @xmath8 is the bohr magneton and @xmath9 is the electronic @xmath9-factor . the nuclear spin - lattice relaxation rate @xmath10 , caused by the hyperfine fermi contact interaction between the nuclear spins and the conduction electron spins , is related to the local spin - spin correlation function through the equation : @xmath11 where @xmath12 , @xmath13 are the transverse components of the electron spin density operator at the nuclear position * r * , and @xmath14 is the nuclear magnetic resonance frequency .
the rate of the nuclear spin - relaxation in metals is , usually , proportional to the temperature and to the square of the electronic density of states at the fermi energy ( the korringa law @xcite ) .
this follows from the following expression : @xmath15\delta ( e_{f}-e_{i}+\gamma _ { n}h_{0}).\ ] ] at low temperatures : @xmath16\propto k_{b}t$ ] @xmath17 , where @xmath18 is the temperature in the energy units , we arrive at the linear in temperature dependence of @xmath19 @xmath20 which is the well known korringa law @xcite . here
@xmath21 is the electron density of states at the fermi level . in high magnetic fields and in systems with reduced dimensionality this simple argumentation does not hold , since the electron spectrum acquires field induced @xcite or size quantized @xcite energy gaps .
it was conjectured in @xcite and confirmed experimentally , see e.g. @xcite , that in qhe systems the nuclear spin relaxation rate should have an activation behavior @xmath22 where @xmath23 is either @xmath24 , the electron zeeman gap ( odd filling factors ) or @xmath25 , the landau levels gap ( even filling factors ) , instead of the usual korringa law .
the discreteness of the electron spectrum manifests , at finite temperatures , in an activation type of the magnetic field dependence of the nuclear spin relaxation rate , @xmath26as it seen from eq .
( [ t1exp ] ) .
this behaviour is similar to that of the magnetoresistance @xmath27 in the qhe , @xcite , see fig . 2 .
.,width=480 ] this unusual magnetic field dependence of the nuclear spin relaxation reflects the fact that the energy gaps in the spectrum of two - dimensional electrons in strong magnetic fields ( either zeeman splitting or the landau levels gap ) are three orders of magnitude larger than the nuclear zeeman energy .
indeed , the energy needed to reverse the spin of an electron in the external magnetic field @xmath28 is @xmath29 , which is much larger ( by a factor of @xmath30 , @xmath31 and @xmath32 being the nuclear and free electron masses ) than the energy @xmath33 provided by reversing the nuclear spin .
therefore the delta function in eq .
( 2 ) can not be realized and the simultaneous spin flip of the nuclear and the electron spins ( flip - flop ) in landau levels , fig .
3 , is severely restricted by the energy conservation . on the other hand , in isotropic @xmath34 electron systems , in a strong magnetic field : @xmath35 , the kinetic energy of the electron motion parallel to the field should change in order to ensure the energy conservation of the process shown in fig . 4 .
thus , in the isotropic
@xmath34 model , the electron spin - flip will be accompanied by a simultaneous change of the landau level and of the kinetic energy parallel to the field , e@xmath36 according to @xmath37 while this is impossible for an ideal @xmath38 system in a strong magnetic field , it may take place in quasi - two - dimensional conductors , as is the case in superlattices , for certain regions of parameters .
electron systems.,width=437 ] because of the existence of energy gaps in the electron spectrum of a 2des under strong magnetic fields ( the qhe systems ) , finite nuclear spin relaxation times @xmath39 could be expected only if 2des is subjected to different kinds of external potentials .
broadened landau levels may overlap , thus providing a nonzero density of states of both spin projections at the fermi energy .
this will result in finite relaxation time for excited nuclear spins interacting with the conduction electrons on the landau levels .
description of these processes , performed in the framework of the finite temperature matzubara diagram techniques , is presented in @xcite .
an important question in the quantum hall effect theory is the role of the edge states : electronic orbits magnetically confined to the sample boundary . while the number of edge states may be small compared to the bulk states , their contribution to @xmath39 can be very important , since they possess a homogeneous energy spectrum @xcite . in sufficiently clean heterojunctions , however , where the fractional qhe ( fqhe ) and wigner crystallization could be observed , the mechanisms mentioned above are extremely inefficient . at finite temperatures the energy conservation in the flip - flop process
can be fulfilled by absorbing a phonon [ 13 - 15 ] .
the mechanism , considered in @xcite , consists of two processes in which , first , electron and nuclear spins are simultaneously reversed by the hyperfine interaction , and , second , only electron spins are reversed by their coupling to the lattice strain . in these processes , a phonon is either absorbed or emitted to satisfy the energy conservation law .
both the zeeman split - level energy ( @xmath40 ) and chemical potential ( @xmath41 ) depend on a magnetic field .
the dominant contribution to nuclear relaxation comes from the transition between two spin - split levels that are near the chemical potential .
the contributions from other landau levels that are far away from the chemical potential are proportional to @xmath42 $ ] and are negligible under the qhe condition .
the value at a minimum depend strongly on the value of exchange interaction . at low temperatures ( @xmath43 ) ,
the nuclear spin relaxation rate varies exponentially with inverse temperature ( i.e. , @xmath44 $ ] ) because the number of thermally excited phonons in the reservoir becomes exponentially small as the temperature is decreased . at higher temperatures ( @xmath45 ) , on the other hand ,
the relaxation rate varies linearly with temperature ( i.e. , @xmath46 ) .
nucleus - mediated spin - flip transitions in gaas quantum dots have been studied in @xcite . in these papers the electron spin relaxation times were calculated taking into account the mechanism , where the phonons provide the necessary energy and nuclear spins take care for the spin conservation during the electron transition between zeeman split states in quantum dots .
the nuclear spin relaxation in 2deg , associated with the dipole - dipole interaction between the nuclear and the electronic spins was studied by ovchinnikov et al .
the principle physical difference between these two mechanisms is that while in the contact interaction the total spin should be conserved and an electron has to flip its spin in order to relax the nuclear spin , the conserved quantity in the dipole - dipole interaction is the total ( spin plus orbital ) momentum .
therefore an electron can relax a nuclear spin without changing its own spin state , just by shifting its center of orbit .
this mechanism should be prevailing in clean samples , i.e. under the fqhe conditions .
much of the recent attention paid to hyperfine interactions under conditions of the quantum hall effect is connected with correlation effects in 2des .
this is based on the notion of a spin - exciton : the elementary excitation over the zeeman gap dressed by the coulomb interaction @xcite.this results in a strong enhancement ( up to a factor of 100 , as is the case in gaas ) of the effective @xmath47 -factor .
spin exciton and skyrmion dispersion : the energy gap of large momentum excitons is two times larger than the skyrmion s one . in the insets the @xmath38 electron spin arrangements for spin excitons ( left ) and skyrmions ( right ) is shown.,width=443 ] due to the coulomb interaction the spin - excitons are bound states of electron - hole pairs which , unlike the individual electrons or holes , can propagate freely under the influence of a magnetic field due to their zero electric charge .
these elementary excitations are , therefore , chargeless particles with a nearly parabolic dispersion in the low - k limit . at @xmath48
the gap is equal to the bare
zeeman splitting . the energy spectrum of spin - excitons on the ground landau level : n=0 is @xcite : @xmath49{\frac{e^{2}}{\kappa a_{h } } } .
\label{espex0}\ ] ] in the parabolic approximation ( small exciton momenta ) , the dispersion relation reads : @xmath50 where @xmath51 is the definition of the spin - exciton mass .
the invariance of the energy gap with respect to the electron- electron interaction is associated with the fact that in creating a quasielectron - quasihole pair excitation at the same point in space ( i.e. with center of mass momentum @xmath48 ) the energy decrease due to the coulomb attraction is exactly cancelled by the increase in the exchange energy .
thus the energy gap for the creation of a widely separated ( i.e. with @xmath52 ) quasielectron - quasihole pair ( large spin - exciton ) is equal to the exchange energy associated with the hole .
external potentials , like long range potential fluctuations , periodically modulated heterostructures etc .
, may reduce substantiality the spin - exciton gap . in @xcite
the spin - exciton in a periodically modulated two - dimensional electron gas under strong magnetic fields was investigated .
it was shown there that a periodic external potential , with a period in the submicron range , can remove the energy gap in the spin - exciton dispersion at certain values of the spin - exciton momenta .
this may result in several interesting phenomena .
for example , the rate of nuclear spin relaxation and depolarization processes in heterojunctions will be strongly enhanced .
it is easy to see why the zeeman energy gap can be removed by the modulation potential in an ideal model of noninteracting electrons : consider a noninteracting electron - hole pair with opposite spins , which are subject to an external magnetic field and a periodic potential with amplitude @xmath53 and period @xmath54 by appropriately selecting the spatial separation between the electron and the hole along the modulation direction the energy difference between them can be reduced to zero .
the minimal amplitude @xmath53 required to satisfy this condition is @xmath55 , the spatial separation between the electron and the hole involved is half of the modulation period @xmath56,and the corresponding wave number ( along the y - direction ) is @xmath57 . the picture is significantly more complicated , however , in the presence of the strong coulomb interaction , where bound spin - excitons exist .
maniv et al .
@xcite , have considered the effect of vacuum quantum fluctuations in the qh ferromagnetic state on the decoherence of nuclear spins .
it was shown there that the virtual excitations of spin excitons , which have a large energy gap ( on the scale of the nuclear zeeman energy ) above the ferromagnetic ground state energy , lead to fast incomplete decoherence in the nuclear spin system .
it is found that a system of many nuclear spins , coupled to the electronic spins in the @xmath38 electron gas through the fermi contact hyperfine interaction , partially loses its phase coherence during the short ( electronic ) time @xmath58 , even under the ideal conditions of the qhe , where both @xmath39 , and @xmath59 are infinitely long .
the effect arises as a result of vacuum quantum fluctuations associated with virtual excitations of spin waves ( or spin excitons ) by the nuclear spins .
the incompleteness of the resulting decoherence is due to the large energy gap of these excitations whereas the extreme weakness of the hyperfine interaction with the @xmath38 electron gas guarantees that the loss of coherence of a single nuclear spin is extremely small .
the manipulation of the nuclear spins is carried out through spin flip - flop processes , associated with the transverse part of the interaction hamiltonian @xmath60 , i.e. @xmath61 $ ] , where @xmath62 are the transverse components of the nuclear spin operators , and @xmath63 are the corresponding components of the electron spin density operators .
here @xmath64 , @xmath65 are the electron field operators with spin projections @xmath66 .
the longitudinal part of @xmath60 , @xmath67 , which commutes with the hamiltonian @xmath68 , and so leaves the nuclear spin projections along @xmath69 unchanged , can still erode quantum coherence in the nuclear spin system @xcite . to simplify the analysis it is assumed that the nuclei under study have spin @xmath70 , and
so the corresponding spin operators are expressed in terms of fermionic creation and annihilation operators , @xmath71 , as @xmath72 in this case the transverse components @xmath73 , @xmath74 , are up to a proportionality constant , just the off diagonal elements ( or coherences ) of the density matrix of a single nuclear spin ( qubit ) @xcite . the decay of these elements with time , which determines the rate of decoherence of a single qubit , can be thus found from the equations of motion for the operators @xmath71 in the heisenberg representation @xmath75 considering a single nuclear spin and evaluating its rate of decoherence due to the coupling with a bath of spin excitons , assuming that initially , at time @xmath76 , the electronic system is in its ground ( qh ferromagnetic ) state @xmath77 that the time dependence of the coherence @xmath78 is given by @xmath79 , where @xmath80.$ ] this result is similar to the expression found by palma et al .
@xcite in an artificial model of pure decoherence , i.e. when energy transfer between the qubit and its environment is not allowed .
the remarkable feature of this expression is due to the presence of the energy gap @xmath81 in the spin exciton spectrum , which is typically much larger than the nuclear zeeman energy @xmath82 . during a short time scale , of the order of @xmath58 ,
the coherence @xmath83 of a single nuclear spin diminishes and then saturates for a very long time ( i.e. of the order of the relaxation time @xmath59 ) at @xmath84 , where @xmath85 ^{2}}% \sim 1.\ ] ] for @xmath86 heterostructure the coupling constant @xmath87 is typically of the order of @xmath88 @xcite .
iordanskii et al .
@xcite have studied nuclear spin relaxation taking into account the creation of spin - excitons @xcite in the flip - flop process .
the energy for the creation of a spin - exciton can be provided by the long range impurity potential in a process , where the electron turns its spin while its center of orbit is displaced to a region with lower potential energy . as shown in fig .
6 , the overlap of the initial and final location of the electron wave functions , centered at @xmath90 and @xmath91 respectively , is : @xmath92 $ ] . here
@xmath93 is the nuclear position .
nuclear spin relaxation by the conduction electron spin in the vicinity of a potential fluctuation is effective when the nuclear spin is positioned in the region of the overlapping initial and final states of the electron wave function .
the energy conservation in the spin - exciton creation process can be written in the form : @xmath94 this expression defines the gradient of electric potential caused by the impurity , sufficient to create a spin - exciton during a flip - flop process .
the probability of finding such a fluctuation is exponentially small : @xmath95.$ ] the momentum * p * of the spin - exciton is small and therefore the expansion in p can be performed everywhere in the final expression for electronic density of states ( dos ) : @xmath96 here @xmath97 depends on the following combination of physical parameters : @xmath98 @xmath99 reaches the maximum , usually , at @xmath100 .
a complete expression for the nuclear depolarization rate from the golden rule formula at t=0 @xcite is given by @xmath101 where @xmath102 are filling factors for the electrons with spin up and spin down , respectively , n is the highest occupied landau level and @xmath103 are the laguerre polynomials . here
the system of units is used , where @xmath104 and @xmath105 .
the dynamics of the nuclear spins is governed by their interactions with each other and with their environment . in the regime of interest ,
these interactions are mediated by 2des .
various time scales are associated with this dynamics .
the relaxation time @xmath39 is related to energy exchange and thermalization of spins .
quantum mechanical decoherence / dephasing will occur on the time scale @xmath59 .
theoretical calculation of the nuclear - spin dephasing / decoherence time scale @xmath59 for in qhe systems with impurities was presented in @xcite .
the hyperfine interaction between the electron and the nuclear spins , @xmath107 , can be split into two parts @xmath108 where @xmath109 corresponds to the coupling of the electrons to the diagonal part of nuclear spin operator @xmath110 , and @xmath111 to its off - diagonal part .
it follows @xcite , that @xmath112 can be incorporated into the nuclear - spin energy splitting , redefining the hamiltonian of the nuclear spin as @xmath113 , where @xmath114 .
the relevant terms in the full hamiltonian can be expressed solely in terms of the nuclear - spin operators and spin - excitation operators @xcite .
it is assumed there that initially , at time @xmath76 , the nuclear spin is polarized , while the excitons are in the ground state , @xmath115 where @xmath116 is the ground state of spin - excitons .
since the hamiltonian conserves the total number of elementary excitations in the system , the most general wave function can be written as @xmath117 corresponding to the nuclear spin in the ground state and @xmath118 describing the single - exciton state with the wave vector @xmath119 .
equations of motion for the coefficients @xmath120 and @xmath121derived from the schr - dinger equation are : @xmath122 the relaxation rate and the added phase shift of the nuclear - spin excited - state probability amplitude @xmath123 are given by the real and imaginary parts of the pole , respectively : @xmath124 and @xmath125 so that @xmath126 $ ] .
it is obvious that due to the large gap in the spin - exciton spectrum , @xmath127 , the energy conservation in the flip - flop process can not be satisfied , and so in the absence of interaction with impurities , @xmath128 .
interactions with impurities will modify these solution , and , as a consequence , the energy conservation condition .
in particular , if the impurity potential is strong enough , it can provide additional energy to spin - excitons , so that their energy can fluctuate on the scale of order @xmath129 thus making nuclear - spin relaxation possible .
this mechanism was identified in @xcite , and it corresponds to large fluctuations of the impurity potential @xmath130 , which usually occur with a rather small probability , so @xmath39 is very large for such systems .
the perturbative solution does not describe the energy relaxation ( @xmath39 ) , but it does yield the phase shift due to the impurity potential .
this phase shift , when averaged over configurations of the impurity potential , produces a finite dephasing time , @xmath59,which can be calculated considering the reduced density matrix of the nuclear spin , given by @xmath131_{u}. \label{eq26moz}\ ] ] here the trace is partial , taken over the states of the spin - excitons , while the outer brackets denote averaging over the impurity potential .
the trace over the spin - excitons can be carried out straightforwardly because within the leading - order perturbative approximation used here they remain in the ground state ; all excitations are virtual and contribute only to the phase shift .
the diagonal elements of @xmath132 are not influenced by virtual excitations and remain constant .
the off - diagonal elements of @xmath133 contain the factors @xmath134 $ ] .
it is the averaging of these quantities over the white - noise impurity potential @xmath130 that yields dephasing of the nuclear spin .
this averaging can be done by utilizing the relation @xmath135 _ { u}=\exp [ -{\frac{1}{2}}\left ( [ \phi ] _ { u}\right ) ^{2 } ] $ ] . apart from the direct nuclear spin relaxation , important information about the electron system can be obtained from nuclear spin diffusion processes .
this is the case when the nuclear spins are polarized in a small part of a sample as it was experimentally observed in @xcite . to explain these experimental observations , bychkov et al
@xcite have suggested a new mechanism for indirect nuclear spin coupling via the exchange of spin excitons .
the spin diffusion rate from a given nuclear site @xmath136 within the polarized region is proportional to the rate of transition probability p(@xmath136 ) for the polarization of the nuclear spin @xmath137 , located at @xmath136 , to be transferred to a nuclear spin @xmath138 , positioned at @xmath139 , outside the polarized region , via the exchange of virtual spin excitons , fig .
the virtual character of the spin - excitons , transferring the nuclear spin polarization , removes the problem of the energy conservation , typical for a single flip - flop process .
furthermore , the virtual spin - excitons are neutral entities , which can propagate freely in the presence of a magnetic field . in this model the electron interactions play a crucial role : the kinetic energy of a spin exciton is due to the coulomb attraction between the electron and the hole
. thus the proposed mechanism yields the possibility of transferring nuclear spin polarization over a distance much longer than the magnetic length @xmath140 .
the long range nature of this mechanism is of considerable importance when the size of the region of excited nuclear spins , @xmath141 , is much larger than the magnetic length @xmath142 .
as it is shown in @xcite , the potential of the nuclear spin - spin interaction , mediated by the exchange of spin - excitons , is a monotonic function of the distance between the two nuclei with the asymptotics : @xmath143 where @xmath144 this is typical for the interaction , mediated by the exchange of quasiparticles with an energy gap at @xmath145 , as is the case for the spin - exciton dispersion .
the range @xmath146 of this potential is determined by the critical wave number @xmath147 as it follows from the uncertainty principle : @xmath148 .
the negative sign of this interaction corresponds to attraction between the nuclear spins and may cause , at sufficiently low temperatures , a ferromagnetically ordered nuclear state in qhe systems .
spin - excitons constitute the building blocks , from which more complex spin textures can be formed .
for example , at finite densities the spin excitons condense @xcite into skyrmion - antiskyrmion pairs .
these unusual topological excitations @xcite in the spin distribution of real 2des were observed , in nmr experiments , near filling factor @xmath149 @xcite .
further evidences for skyrmions in 2des were found in transport and optical experiments .
skyrmions , in qhe systems , are the topologically nontrivial spin excitations around filling factor @xmath149 @xcite which arise as a condensate of interacting spin excitons @xcite .
the coulomb interaction acts to enlarge the skyrmion size while the zeeman splitting tends to collapse skyrmions .
the interplay between these factors determines the final distribution of spins within a skyrmion , and its characteristic length scales .
the resulting radius @xmath150 corresponds to the region where both these energies are of the same value , and grows weakly to infinity as the g - factor goes to zero @xcite , thus reflecting the importance of the long range coulomb repulsion associated with the skyrmion charge in the zero g - factor limit .
a canonical @xmath151 transformation from the fully polarized ground state , where all spins are oriented along a single axis , to a state , consisting of a macroscopic number of differently oriented spins , each of which is slightly rotated with respect to its nearest neighbors in space was is developed in @xcite .
it is found also that contrary to the lowest landau level , these energies are positive for both skyrmions and antiskyrmions . it should be noted that the spin - rotation transformation employed in @xcite is unitary and does not change the total number of electrons .
thus by going to the new state @xmath152 from the fully polarized ground state @xmath153 , is exactly equal to half the total energy required to create a well separated electron - hole pair ( large spin exciton ) .
if the zeeman splitting is not neglected , however , the increase in magnetic energy associated with the rotated spins determines a length scale for the spatial size of the defect . to take into account zeeman spin splitting the hartree - fock ( hf ) energy functional
should be corrected by the additive term @xmath154 @xcite , where @xmath155 is the zeeman splitting energy .
the corresponding equation for the vectorial field @xmath156 can be obtained by variation of the corrected energy functional with respect to @xmath157 under the constraint @xmath158 .
the result of such a calculation is : @xmath159 , \label{eqforn}\ ] ] where @xmath160 and @xmath161 is the effective bohr radius ( note that @xmath162 is the free electron mass ) . in nmr experiments on skyrmions [ 43,46 - 48,51 ]
the nuclear spins are strongly polarized .
the sample inhomogeneity may result in a strong inhomogeneity of the hyperfine field , fig . 8 , and therefore spatially varying electron zeeman splitting .
this may result in a strong localization of skyrmions @xcite , resulting in the combined topologically nontrivial electron - nuclear spin excitation , the skyr - nuon , fig .
as we have seen in previous sections , the nuclear spin polarization , once created , remains finite for macroscopically long times @xcite .
intensive experimental studies @xcite of this phenomenon have provided a more detailed knowledge on the hyperfine interaction between the nuclear and electron spins in heterojunctions and quantum wells .
it was observed that the nuclear spin relaxation time is rather long ( up to 10@xmath163 ) and the hyperfine field acting on the charge carriers spins is extremely high , up to 10@xmath164 @xcite .
the nontrivial physics in this subject is based on a fact that the discrete nature of the electron spectrum in these systems will result in exponentially long dependence of the nuclear spin relaxation times @xmath39 on the system parameters @xcite , i.e. t@xmath165 @xmath166 ( here @xmath167 is the electron energy level spacing and @xmath168 is the temperature ) .
we assume here that similar law should take place also in the nanostructures with well defined size quantization of the electron spectrum .
note that in this case @xmath39 is very sensitive to the potential fluctuations , caused by the inhomogeneous distribution of impurities in a heterojunction @xcite .
a family of new physical effects in nanostructures with strong spin - orbital coupling appears when the electron spin degeneracy is lifted by a hyperfine field of polarized nuclei .
indeed , the combined action of a strong nuclear polarization and the spin - orbit interaction , breaks the time reversal symmetry in a mesoscopic system .
a good illustration for such * meso - nucleo - spinic * effects is the hyperfine aharonov - bohm effect ( habe ) in mesoscopic rings @xcite , caused by hyperfine interaction , as is explained in what follows .
persistent currents ( pc ) in mesoscopic rings reflect the broken clock wise - anticlock wise symmetry of charge carriers momenta caused , usually , by the external vector potential .
experimentally pcs are observed when an adiabatically slow time dependent external magnetic field is applied along the ring axis .
the magnetic field variation results in the oscillatory , with the magnetic flux quantum @xmath169 ( or its harmonics ) period behavior of the diamagnetic moment ( the pc ) , which is the manifestation of the usual aharonov - bohm effect ( abe ) .
it was suggested in @xcite that in a quantum ring with a nonequilibrium nuclear spin population the persistent current will exist , even in the absence of external magnetic field .
it is shown there , that the abe like oscillations of pc with time will appear , during the time interval of the order of nuclear spin relaxation time @xmath39 .
the physics of this phenomenon can be understood along the following lines .
the hyperfine field , caused by the nonequilibrium nuclear spin population breaks the spin symmetry of charged carriers . combined with a strong spin - orbital ( so ) coupling , in systems without center of inversion @xcite , it results in the breaking of the rotational symmetry of diamagnetic currents in a ring . under the topologically nontrivial spatial nuclear spin distribution ,
the hyperfine field produces an adiabatically slow time variation of the berry phase of the electron wave function .
the time variation of this topological phase results in observable oscillations of a diamagnetic moment ( the pc ) .
it is one of a series of meso - nucleo - spinic effects , which may take place in mesoscopic systems with broken symmetry , due to the combined action of the hyperfine field and spin - orbital interaction . due to the contact hyperfine interaction eq .
( [ hint1 ] ) it follows that once the nuclear spins are polarized , i.e. if @xmath170 , the charge carriers spins feel the effective , hyperfine field @xmath171 which lifts the spin degeneracy even in the absence of external magnetic field . in gaas
/ algaas one may achieve the spin splitting due to hyperfine field of the order of the one tenth of the fermi energy @xcite .
let us suppose therefore that the charge carriers spin orientation is partially polarized during the time interval of the order of @xmath39 .
it is quite obvious that the topologically nontrivial spin texture combined with the spin - orbit interaction will result in a pc .
taking the bychkov - rashba term @xcite @xmath172 \mathbf{\nu } , \label{hsorashba1}\ ] ] where @xmath173 for holes with @xmath174 ( @xmath175 is the free electron mass ) , and @xmath176 for electrons , @xmath177,@xmath178 are the charge carrier spin and momentum and @xmath179 is the normal to the surface .
it can be rewritten in the form @xmath180 where @xmath181 @xmath182 stands for a nonequilibrium carriers spin population . under the conditions of a topologically nontrivial orientation of @xmath183 the wave function of a charge carrier encircling the ring gains the phase shift similar to the one in an external magnetic field like in the ordinary abe .
this phase shift can be estimated as follows @xmath184 where @xmath185 is the ring perimeter . to observe the oscillatory persistent current connected with the adiabatically slow time - dependent @xmath186 @xmath185
is supposed to be less than the phase breaking length . taking the realistic values for @xmath187 and @xmath188 we estimate @xmath189 which shows the experimental feasibility of this effect .
the standard definition of the spontaneous diamagnetic current is @xmath190 where f is the electron free energy and @xmath191 is the external ( probe ) magnetic flux .
the oscillations of pc arise due to the exponential time dependence of the phase @xmath192 in eq .
( [ phi1 ] ) , with the time constant @xmath193from the analogy with the standard abe in a mesoscopic ring , we expect the following form for a persistent current in an one dimensional quantum ring at low enough ( t @xmath194 ) temperature @xmath195 here @xmath196 is the initial phase value . the marking difference between the periodical time dependence of standard aharonov - bohm oscillations , which are observed usually under the condition of linear time variation of the applied magnetic field and the hyperfine driven oscillations which die off due to the exponential time dependence of the nuclear polarization . in @xcite
it was proposed to use the inhomogeneous hyperfine field to create so - called nuclear spin polarization induced ( nspi ) magnetic structures , such as magnetic quantum dots , wires , rings , superlattices etc .
the spin splitting ( @xmath197 ) due to a hyperfine magnetic field can be comparable to the fermi energy of 2deg , so that the electrons in the region , where nuclear spins are polarized , will occupy the energetically more favorable states with the spins opposite to @xmath198 .
the inhomogeneous nuclear polarization acts on the electrons as the effective confining potential @xmath199 .
this effective confining potential can be used to create different nanostructures with polarized electrons in them .
the method of local nuclear spin optical polarization allows to create different nspi quantum structures ( quantum dots , rings , wires , etc . ) using the same sample and different illumination masks
. the time evolution of the hyperfine magnetic field in nspi quantum structures , due to the nuclear spin relaxation and the nuclear spin diffusion , leads to variation of the number of transverse modes and corresponding electron energies at a constant gate potential that can be directly measured by transport experiments .
the dependence of the conductance at zero temperature on the number of transverse modes in the conductor is given by the landauer formula @xmath200 , where @xmath168 is the average electron transmission probability , @xmath201 is the number of the transverse modes and the factor @xmath202 stands for the spin degeneracy .
it is assumed that the transition probability @xmath168 is independent of the energy in a small interval between the chemical potentials of the reservoirs .
usually , the number of the transverse modes , defined by the effective width of the conductor , is controlled by the gate voltage and the conductance is changed in discrete steps @xmath203 .
it is underlined in @xcite that due to the spin selective effective potential the height of the conductance steps is just @xmath204 which is a half of the conductance quantum @xmath205 .
the experimental feasibility is based on the method of optical nuclear spin polarization @xcite .
the resolution of the optical illumination of the sample can be high enough .
usual optic technique allows to create the light beams of the width of the order of the wave length ( @xmath206 ) , by using near fields optics the beam width can be sufficiently reduced ( @xmath207 ) .
hence a nspi qw of the width of @xmath208 can be created by the modern experimental technique . in semiconductor heterostructures having supreme quality
, the electron mean free path can be as large as @xmath209 and nspi qw will operate in the quantum regime .
in bulk semiconductors , the study of the conduction electron magnetotransport under the influence of the nonequilibrium nuclear spin polarization was started a long time ago by measurement of the magnetoresistance in bulk insb @xcite . very recently a refined detailed study of the influence of a weak hyperfine field on the hall effect in bulk insb under a strong magnetic field ( one occupied landau level ) was performed @xcite .
apart of the usual hall effect , caused by the drift of electrons in the crossed electric and magnetic fields , there exist also another contribution to the nondiagonal components of the conductivity tensor , the anomalous hall effect @xcite ( ahe ) , which is caused , in nonmagnetic semiconductors , by the spin - orbit interaction combined with the carrier magnetization .
it was proposed in @xcite that in the strong nuclear spin polarization limit and at low magnetic field this effect may give a leading contribution to the anomalous hall effect . due to this interaction ,
electrons with their spin polarization parallel to the magnetization axis will be deflected at right angles to the directions of the electric current and of the magnetization while electrons with antiparallel spin polarization will be deflected in the opposite direction .
thus , if the two spin populations are not equal there appears a net current in the transverse direction .
until now , studies of the anomalous hall effect have been limited to the case where the carrier magnetization is induced by the magnetic field .
the magnetic field , however , produces a much larger normal hall effect which makes experimental studies quite difficult . recording the hall voltage in a sample with a dynamically polarized nuclear spins , during a time scale of the order of the nuclear spin relaxation time , which can be very long on the scale of typical electron equilibration processes , gives a direct measure of the spin dependent component of the hall voltage .
if the experiment is designed in such a way that the hyperfine field is much higher then the external magnetic field the hyperfine field driven anomalous hall effect can be observed .
insb is well suited for observation of the hahe effect because of the strong spin - orbit coupling and the strong , due to the high atomic numbers of the atoms involved , hyperfine interaction ( @xmath210 ev , @xmath211 ev , @xmath212 , @xmath213 ) . the maximal nuclear - spin polarization for all isotopes of insb yields @xmath214 @xcite .
it is outlined in @xcite that the hahe _ _ _ _ gives the possibility to observe genuine spin dependent anomalous hall effect which relies on the natural relaxation of nuclear spins and facilitates the experiment since the anomalous hall voltage is not hidden by the much larger ordinary hall effect .
the possibility that the hyperfine interaction between the conduction electron spins and nuclear spins may result in hyperfine residual resistivity ( hrr ) _ _ _ _ in clean conductors at very low temperatures was studied theoretically by dyugaev et al . @xcite .
apart from the fundamental nature of this problem , the natural limitations on the mean free path are decisive in semiconductor based high speed electronic devices , like heterojunctions and quantum wells .
the space periodicity of nuclei plays no specific role , as long as the nuclear spins are disordered and act as magnetic impurities @xcite with the concentration @xmath216 .
this scattering is not operative at extremely low temperatures , in the @xmath217 region when the nuclear spins are ferro- or antiferromagnetically ordered .
the hrr , arising due to the fermi ( contact ) hyperfine interaction between the nuclear and the conduction electron spins can be written in atomic units as : @xmath218 here @xmath219 @xmath220 are the electron and the nucleon masses , respectively ; @xmath221 ev and @xmath222 is the fine structure constant . in metals the effective electron - nuclear interaction
constant is @xmath223 where the fermi energy @xmath224 varies in a wide interval @xmath225 .
@xmath226 is @xmath227 for @xmath228 and @xmath229 for the rare earth metals .
the total residual resistivity is therefore a sum of the impurity @xmath230 and the nuclear spin @xmath231 contributions : @xmath232 here @xmath233 in atomic units : @xmath234 sec .
the nuclear contribution to resistivity starts to be operative when the impurity concentration is @xmath235 . in the limit of an ideally pure ( @xmath236 ) metal the universal residual resistivity @xmath237is , therefore @xmath238 and the mean free path
is limited by @xmath239 cm .
this yields @xmath240 cm in @xmath228 and @xmath241 cm for the rear earth metals .
it is interesting to note that in materials with even - even nuclei ( zero spin ) , like @xmath242 , @xmath243 , @xmath244 , @xmath245 and isotopically clean graphite @xmath246 , where the electron - nuclear scattering is absent , the _ hrr _ would not be observed .
the temperature and magnetic field dependence of the _ hrr _ contains reach information on the hyperfine interaction between the conduction electrons and the nuclear spins .
usually , the temperature and the magnetic field dependence of residual resistivity due to nonmagnetic impurities is due mostly to the mesoscopic effects , and is vanishing in the limit @xmath247 . in a magnetic field
such that @xmath248 the magnetic impurities freeze out and the kondo effect is quenched . in order to freeze out the nuclear spins however one should apply much higher magnetic fields , @xmath249 .
therefore in the temperature interval @xmath250 @xmath251 the nuclear spin contribution may prevail even in metals with magnetic impurities . in metals like @xmath252 the nuclear magnetic moments
@xmath253 and even without external magnetic field their @xmath254 degeneracy is lifted partially by the quadrupole effects ( in the case of cubic crystal symmetry the quadrupole splitting of the nuclear levels may happen due to the defects @xcite and dislocations @xcite ) . while the normal metals have a quite similar electronic structure , the experimentally observed temperature dependence of the dephasing time @xmath255 is quite different .
this was shown in @xcite , where the value of @xmath256 was defined by the magnetoresistance measurements of long metallic wires @xmath257 in a wide temperature interval @xmath258k . in @xmath259 and @xmath260
wires @xcite@xmath261 saturates at low temperatures which contradicts the standard theory @xcite . strangely enough
the ag wires do not show saturation @xcite at the lowest temperatures , in accordance with @xcite .
it is conjectured in @xcite that the influence of the quadrupole nuclear spin splitting on the phase coherence time @xmath256 can be a clue to this puzzle .
indeed , the nuclear spins of both @xmath259 and @xmath260 have a strong quadrupole moment ( @xmath262 ) and may act as inelastic two - level scatterers once their degeneracy is lifted by the static impurities @xcite , dislocations @xcite and other imperfections .
it is known that ( see @xcite and references therein ) the nondegenerate two level scatterers may introduce inelastic phase breaking scattering of conduction electrons .
this may be not the case for @xmath263 nuclei since their spin is @xmath264 . in this case
the quadrupole splitting of nuclei spins by imperfections is negligible . in the absence of magnetic zeeman splitting
therefore the nuclear spins in @xmath263 samples will act just as a set of elastic scatterers , and the temperature dependence of @xmath256 should obey the standard theory @xcite .
the problem of coexistence of the superconducting and magnetic ordering , in spite of its long history @xcite , is still among the enigmas of modern condensed matter physics .
most of the theoretical and experimental efforts were devoted to studies of the coexistence of electron ferromagnetism and superconductivity .
the possibility of a reduction of @xmath265 by the nuclear ferromagnetism was outlined by dyugaev et al . in @xcite and theoretically studied in more details in @xcite .
it was experimentally observed by several groups @xcite .
it was recently conjectured @xcite that the hyperfine part of the nuclear - spin - electron interaction may result in the appearance of a nonuniform superconducting order parameter , the so called fulde - ferrel - larkin - ovchinni - kov state ( fflo ) @xcite .
the fflo state was thought originally to take place in superconductors with magnetically ordered magnetic impurities @xcite .
the main difficulty , however , in the observation of the fflo caused by magnetic impurity ordering is in the simultaneous action of the electromagnetic and exchange parts of the magnetic impurities on the superconducting order . in most of the
known superconductors the
part destroys the superconducting order before the exchange
part modifies the bcs condensate to a nonuniform fflo state .
the situation may change drastically in the case of nuclear spin ferromagnetic ordering .
indeed , the nuclear magnetic moment @xmath266 is at least three orders of magnitude smaller than the electron bohr magneton @xmath267 so that the electromagnetic part of the nuclear spin fields is quite low , compared to that of the magnetic impurities . on the other hand
the exchange part is strongly dependent on the nuclear charge @xmath268 .
a growing number of models for electron and nuclear spin based memory cells @xcite and quantum information processing ( or quantum computing - qc ) has been recently proposed @xcite-@xcite .
we will concentrate in what follows on the models based on the manipulation of nuclear spins in bulk semiconductor @xcite , heterostructures @xcite and quantum wires and rings @xcite . in @xcite
we have proposed a quantum computer realization based on hyperfine interactions between the conduction electrons and nuclear spins embedded in a two - dimensional electron system in the quantum - hall effect . for modifications and improvements of this model see a recent review @xcite .
maniv et al .
@xcite have suggested a following physical process of preparing a coherent state in qh ferromagnets .
let us assume that at time @xmath269 the filling factor was tuned to a fixed value @xmath270 and then kept constant until @xmath76 . if @xmath271 then at @xmath76 the nuclear spin system is in the ground state corresponding to the @xmath38 electron system at @xmath272 .
suppose that at time @xmath76 the filling factor is quickly switched ( i.e. on a time scale much shorter than @xmath273 ) back to @xmath149 so that the nuclear spin system is suddenly trapped in its instantaneous configuration corresponding to @xmath274 .
thus the nuclear spins for a long time @xmath275 ( @xmath276 @xmath273 ) will find themselves almost frozen in the ground state corresponding to the 2des at @xmath272 , since @xmath277 .
the main idea behind this scenario stems from the experimental observation of a dramatic enhancement of the nuclear spin lattice relaxation rate @xmath278 , and of a sharp decrease of the knight shift in optically pumped nmr measurements , as the filling factor is shifted slightly away from @xmath279 .
the prevailing interpretation of these closely related effects , associates them with the creation of skyrmions ( or antiskyrmions ) in the electron spin distribution as the @xmath38 electron system moves away from the quantum hall ferromagnetic state at @xmath149 .
the off - diagonal long range magnetic order @xcite existing in this state , corresponds to a complex order parameter which is coupled locally through the hyperfine interaction to the nuclear spins @xcite .
it is thus expected that the proposed manipulation of the nuclear spin system can be performed in a phase coherent fashion over a spatial region with size of the order of the skyrmion radius .
the latter effect is still sufficiently weak to enable the survival of a coherent state of a large number of qubits in the computer memory at @xmath149 . to find an upper bound for
the size of such a memory let us consider @xmath280 independent nuclear spins located at various sites @xmath281 in the quantum well .
a number @xmath282 , stored in the computer memory , corresponds to the direct product of @xmath280 pure nuclear spin states @xmath283 @xmath284 @xmath285 , where @xmath286 , and @xmath287 is a nuclear state with spin projection @xmath288 located at the site @xmath289 . to be able to start a significant quantum computing process , however , a coherent superposition of such products , i.e. @xmath290 ( see e.g. @xcite ) , should be prepared at time @xmath291
. this superposition may be represented more transparently for our purposes by the direct product of mixed spin up and spin down states , @xmath292 , with the normalization @xmath293 .
the mixing is expected to take place most efficiently via the flip - flop processes with the electron spins during the manipulation period when @xmath294 .
let us further assume that the initial state @xmath295 is a completely coherent state , so that each of the numbers of length @xmath280 has an equal probability .
this is a typical state required in carrying out efficient quantum calculations @xcite . in this state @xmath296 for any @xmath297 and @xmath298 it is evident that due to the complete coherence of the initial state @xmath299 the survival probability @xmath300 @xmath301 ^{n}\approx \exp \left\ { -\frac{1}{2}n\left [ 1-\func{re}j\left ( t\right ) \right ] \right\ } \approx e^{-\frac{1}{2}n\gamma \left ( t\right ) } \ ] ] depends only on the decoherence factor @xmath302 .
the decay of @xmath303 therefore follows @xmath304 $ ] , saturating at @xmath305 $ ] for @xmath275 @xmath276 @xmath58 . despite the much larger drop in the level of coherence , the time scale over which the coherence diminishes
is the same as in the case of a single qubit .
the most popular and intensively developed is the kane model @xcite , where the qubits are the nuclear spin of the @xmath306p donor placed near to the surface of _ si_. the nuclear spins are manipulated by electric gates above the _ si _ surface via the atomic electrons .
the two - qubit interaction is provided by the overlapping outer electrons of two neighboring @xmath306p donors .
the kane model has stimulated detailed theoretical and experimental activity .
the strength of the hyperfine interaction is proportional to the probability density of the electron wavefunction at the nucleus . in semiconductors
, the electron wavefunction extends over large distances through the crystal lattice .
two nuclear spins can consequently interact with the same electron , leading to electron - mediated or indirect nuclear spin coupling .
voltages applied to metallic gates in a semiconductor device may be used tocontrol the hyperfine interaction and , correspondingly , the electron - mediated interaction between the nuclear spins . a quantum mechanical calculation proceeds by the precise control of three external parameters : \i ) gates above the donors control the strength of the hyperfine interactions and hence the resonance frequency of the nuclear spins beneath them ; \ii ) gates between the donors turn on and off electron - mediated coupling between the nuclear spins ; \iii ) a globally applied a.c .
magnetic field b@xmath307 flips nuclear spins at resonance .
custom adjustment of the coupling of each spin to its neighbors and to b@xmath308 enables different operations to be performed on each of the spins simultaneously .
finally , measurements are performed by transferring nuclear spin polarization to the electrons and determining the electron spin state by its effect on the orbital wavefunction of the electrons , which can be probed using capacitance measurements between adjacent gates .
an important requirement for a quantum computer is to isolate the qubits from any degrees of freedom that may lead to decoherence .
if the qubits are spins on a donor in a semiconductor , nuclear spins in the host are a large reservoir with which the donor spins can interact .
consequently , the host should contain only nuclei with spin _
i = 0_. this simple requirement unfortunately eliminates all _ iii - v _ semiconductors as host candidates , because none of their constituent elements possesses stable _
i=0 _ isotopes .
iv _ semiconductors are composed primarily of _ i = 0 _ isotopes and can in principle be purified to contain only _
i = 0 _ isotopes . the only _
i = 1/2 _ shallow ( group _ v _ ) donor in si is @xmath306p . at sufficiently low @xmath306p concentrations at temperature _
t = 1.5 k _
, the electron spin relaxation time is thousands of seconds and the @xmath309p nuclear spin relaxation time exceeds 10 hours .
it is likely that at millikelvin temperatures the phonon limited @xmath306p relaxation time is of the order of 10@xmath310 seconds @xcite , making this system ideal for quantum computation .
recent calculations of relaxation and decoherence times for this system are presented in @xcite . in @xcite a new system
is proposed , which consists of nuclear spins ( qubits ) embedded into a mesoscopic ring , fig .
11 , or a finite length quantum wire , fig . 12 .
the hyperfine interaction of the electrons in the system with nuclear spins leads to an effective indirect nuclear spin interaction @xmath311 where @xmath312 is magnetic moment of a nucleus , @xmath313 and @xmath314 are functions of the system parameters presented in fig
it is obtained in @xcite that the effective nuclear spins interaction exhibits sharp maxima , fig .
13 , as function of the magnetic field and nuclear spin positions which opens the way to manipulate qubits with almost atomic precision .
the selective nuclear spin interaction can be obtained by changing external parameters of the system , as is shown in fig .
14 . in @xcite the expression for indirect nuclear spin interaction between nuclear spins in a mesoscopic system , consisting of a finite length quantum wire or ring with nuclear spins placed in the zero nuclear spin matrix , is obtained .
interaction between any two qubits , which is necessary for two - qubit operations , is performed by the electrons in the wire .
it is very sensitive to the system parameters : nuclear spin location , number of electrons , magnetic field and geometry of the system .
its dependence on the system parameters is completely different from indirect nuclear spin interaction in @xmath38 and @xmath34 metals : by varying the external parameters ( magnetic field and number of electrons ) one can control with almost atomic precision the nuclear spin interaction strength by creating maxima of the amplitude of electron wave function on some qubits and zero on the other .
the connections to the measuring system , preparation of the initial state and performing of the one - qubit operations using the nmr could be similar to the existing experimental suggestions @xcite .
the decoherence time of the nuclear spins in mesoscopic systems is expected to be long enough to perform the quantum computation , since the discrete electron spectrum in mesoscopic systems imposes restriction on the flip - flop processes and the nuclear spin relaxation time at law temperatures is expected to have an activation behavior . ) , as function of the magnetic flux through the ring.,width=390 ]
the measurement of the nuclear spin - relaxation in heterojunctions is a challenging experimental problem , since the direct detection of the nmr signals in solids requires usually @xmath315 nuclei .
the number of nuclear spins interacting with the two - dimensional electrons is however , much smaller : @xmath316 .
the first successful measurements of the magnetic field dependence of @xmath26 under qhe conditions were performed in a series of elegant experiments by the k. von klitzing group , @xcite . combining electron - spin resonance ( esr ) and resistivity measurement techniques they have observed the shifting of the esr resonance frequency by the hyperfine field of nonequilibrium nuclear spin population , which is the well known overhauser shift @xcite . in this experiment
the @xmath38 electron zeeman splitting is tuned to the pumping frequency . the angular momentum gained by a 2deg electron , excited to the upper zeeman branch ,
is then transferred to the nuclear spins , thus creating a nonequilibrium nuclear spin population .
these measurements show a close similarity between the magnetic field dependence of the nuclear spin - relaxation rate and the magnetoresistance in quantum hall effect , as it was suggested theoretically in @xcite , thus demonstrating clearly the importance of the coupling of nuclear spins to the conduction electron spins in the nuclear relaxation processes in these systems .
various experimental techniques were used since and in what follows we will describe shortly the main developments and achievements in experimental studies of the hyperfine coupling between the nuclear spins and the electrons in qhe , mesoscopic and superconducting systems .
another way of measuring the nuclear spin relaxation and diffusion in a heterojunction under strong magnetic field by transport techniques ( spin - diode ) was demonstrated by kane et al .
they have reported measurements performed on spin diodes : junctions between two coplanar 2deg s in which @xmath317l on one side and @xmath318 1 on the other .
the fermi level e@xmath319 crosses between spin levels at the junction .
in such a device the 2deg is highly conducting except in a narrow region ( with a width of the order of several hundred angstroms ) where @xmath320 wald et al .
@xcite presented the experimental evidence for the effects of nuclear spin diffusion and the electron - nuclear zeeman interaction on interedge state scattering .
polarization of nuclear spins by dc current has proven to be a rich source for new , not yet understood completely , phenomena .
this is the case , for example of anomalous spikes in resistivity around certain fractional filling factors , observed by kronmuller et al .
@xcite and in resistively detected nmr in qhe regime reported by desrat et al .
influence of nonequilibrium nuclear spin polarization on hall conductivity and magnetoresistance was observed and studied in detail by gauss et al .
@xcite . in 1994
barrett et al .
@xcite observed , for the first time , a sharp nmr signal in multi - quantum wells , using the lampel @xcite technique of polarizing the nuclear spins by optical pumping of interband transitions with near - infrared laser light _
( opnmr)_. polarization of nuclei results in a significant enhancement of the nmr sensitivity , since the resonance in a two - level system results in equalizing their population .
the difference in the population is , obviously , maximal when the spins are completely polarized .
detailed studies of the knight shift data suggested @xcite that the usually accepted picture of electron spins , aligned parallel to the external field , should be modified to include the possibility of topologically nontrivial nuclear spin orientations , the skyrmions .
optical polarization of nuclear spins was used also as a tool for reducing the zeeman splitting of @xmath38 electrons by kukushkin et al .
this resulted in a noticeable enhancement of the skyrmionic excitations .
similar results are reported by vitkalov et al .
@xcite .
very recently , a modern ultra - sensitive nmr spin - echo technique was employed to study the physics of qhe @xcite in gaas / algaas multi - quantum well heterostructures .
the spin polarization of 2des in the quantum limit was investigated and the experimental data support the noninteracting composite fermion model in the vicinity of the filling factor @xmath321 . using the same technique the polarization of 2des near @xmath322
was investigated @xcite .
it was shown there that a quantum phase transition from a partially polarized to a fully polarized state can be driven by increasing the ratio between the zeeman and coulomb energies .
an amazing phenomenon , following from the hyperfine coupling between the electron and the nuclear spins , is the giant enhancement of the low temperature heat capacity of gaas quantum wells near the filling factor @xmath323 , discovered in 1996 by bayot et al .
@xcite . as other thermodynamic properties , it experiences quantum oscillations , following from the oscillatory density of states @xmath324 at the fermi level . at about @xmath325 , and in clean samples , bayot et al
@xcite have observed anomalous deviations from the free electron model , in the specific heat value ( up to four orders of magnitude ) for the filling factor in the range @xmath326 @xmath327 .
their explanation is that in this interval of parameters , the electron system couples strongly to nuclear spin system with a concentration of several orders of magnitude larger than the electron one .
this raises a question about the origin of the strong coupling between the electron and the nuclear spins in the interval @xmath326 @xmath327 .
the guess is the skyrmions , since they are predicted to appear just in the same interval of the filling factor .
additional support for this mechanism is in the results of @xcite , where the disappearance of the nuclear spin contribution to the heat capacity was reported , as the ratio between the zeeman and coulomb energies exceeds a certain critical value .
the zeeman splitting of electrons was modified in these experiments by tilting the magnetic field .
a new very promising technique for measuring spatially varying nuclear spin polarization within a gaas sample is reported in @xcite . in the force detected nmr ( fdnmr ) the sample
is mounted on a microcantilever in an applied magnetic field .
a nearby magnetic particle creates a gradient of magnetic field which exerts a force on the magnetized sample and triggers the cantilever oscillations .
fdnmr is capable to perform the magnetic resonance imaging of the sample with a very high accuracy .
this method can be useful in defining the spatial distribution of nuclear spin polarization in non homogeneous samples .
this information may be crucial for understanding different peculiarities of data obtained by previously described methods .
a new world of the low - temperature physics of the hyperfine interactions in superconducting metals opens in the @xmath328 region , where the nuclear spins start ordering , thus reducing the critical magnetic field of superconductors as has been recently discovered by the pobell group @xcite .
they have studied the magnetic critical field @xmath329 of a metallic compound @xmath330 where the superconductivity sets up at @xmath331 they have observed , in @xmath332 , the nuclear spin ferromagnetic ordering at @xmath333 it was observed in these experiments that the magnetic critical field @xmath334 g is lowered by almost a factor of two at @xmath335
in writing this review the author benefited enormously from the colleagues and co - workers : p.
averbuch , m. azbel , s.e .
barret , c. berthier , yu . bychkov , a. dyugaev , e. ehrenfreund , v.i .
falko , v. fleurov , l.p .
gorkov , v.l .
gurevich , m. horvati , s.v .
iordanskii , a.g.m .
jansen , k. von klitzing and his group , k. kikoin , ju.h .
kim , b.i .
lembrikov , d. loss , t. maniv , s. meshkov , d. mozyrsky , yu .
ovchinnikov , yu . pershin , v. privman , f. pobell , m. potemski , late a. rozhavsky , a. shik , i. shlimak , b. spivak , p.c.e .
stamp , p. wyder , a. zyuzin , whose knowledge , intuition and professional skills were crucial in going through the jungle of hyperfine interactions in such exotic systems , as superconducting , quantum hall , mesoscopic and nano - sytems .
a. abragam , _ the principles of nuclear magnetism _
( clarendon , oxford , 1961 ) ; c.p .
slichter , _ principles of magnetic resonance _ ( springer verlag , berlin , 3rd ed , 1991 ) ; j. winter , _ magnetic resonance in metals _ ( clarendon press , oxford , 1971 ) . k. von klitzing , g. dorda , and m. pepper , phys .
* 45 * , 494 ( 1980 ) ; d.c .
tsui , h.l .
strmer , and a.c .
gossard , phys .
. lett . * 48 * , 1559 ( 1982 ) ; r.b .
laughlin , phys .
lett . * 50 * , 1395 ( 1983 ) ; _ perspectives in quantum hall effect _
, eds : s. das sarma and a.p .
pinczuk ( wiley , new york , 1997 ) .
mani , w.b .
johnson , v. narayanamurti , v. privman , and y.h .
zhang , physica e * 12 * , 152 ( 2002 ) ; j.m .
tailor , c.m .
markus , and m.d .
lukin , phys .
* 90 * , 206803 ( 2003 ) ; a. imamoglu , e. knill , l. tian , and p. zoller , phys .
lett . * 91 * , 017402 ( 2003 ) .
p. khandelwal , n.n .
kuzma , s.e .
barrett , l.n .
pfeiffer , and k.w .
west , phys .
lett . * 81 * , 673 ( 1998 ) ; a.e .
dementyev , p. khandelwal , n.n .
kuzma , s.e .
barrett , l.n .
pfeiffer , and k.w .
west , phys .
83 * , 5074 ( 1999 ) ; n.n .
kuzma , p. khandelwal , s.e .
barrett , l.n .
pfeiffer , and k.w .
west , science * 281 * , 5377 , 686 ( 1999 ) .
s. melinte , e. grivei , v. bayot , and m. shayegan , phys .
lett . * 82 * , 2764 ( 1999 ) ; s. melinte , n. freytag , m. horvatic , c. berthier , l.p .
levy , v. bayot , and m. shayegan , phys .
lett . * 84 * , 354 ( 2000 ) .
a. bychkov , t. maniv , i.d .
vagner , and p. wyder , phys . rev
. lett . * 73 * , 2911 ( 1994 ) ; t. maniv , yu .
a. bychkov , a. kaplunovsky , and i.d .
vagner , physica b * 204 * , 134 ( 1995 ) ; yu.a .
bychkov , t. maniv .
vagner , and p. wyder , europhys . lett .
* 40 * , 557 ( 1997 ) .
lee and c.l .
kane , phys .
* 64 * , 1313 ( 1990 ) ; s.l .
sondhi , a. karlhede , s.a .
kivelson , and e.h .
rezayi , phys .
b * 47 * , 16419 ( 1993 ) ; l. brey , h.a .
fertig , r. cote , and a.h .
macdonald , phys .
lett . * 75 * , 2562 ( 1995 ) . for
a review see s.m .
girvin , in : _ topological aspects of low dimensional systems _
, ed . : a. comtet , t. jolicoeur , s. ouvry , and f. david ( springer - verlag , berlin and les editions de physique , les ulis , 2000 ) ; cond - mat/9907002 ( 1999 ) . b. altshuler , a. aronov , and d. khmelnitskii , j. phys .
c * 15 * , 7367 ( 1982 ) ; b. altshuler and a.g .
aronov , in : _ electron - electron interaction in disordered systems _ , eds : a.l .
efros and m. pollak ( north - holland , amsterdam , 1985 ) . | the electron spin transport in condensed matter , * spintronics * , is a subject of rapidly growing interest both scientifically and from the point of view of applications to modern and future electronics . in many cases
the electron spin transport can not be described adequately without accounting for the hyperfine interaction between electron and nuclear spins . under extreme conditions of high magnetic fileds , ultra - low temperatures , ultra high isotopical cleanness etc . , the nuclear spins in these sytems are very promising candidates for the qubits : the basic elements of emerging quantum memory , logics and hopefully quantum computers . here
we review the progress in the * nuclear spintronics * i.e. in physics and applications of hyperfine interactions in such exotic systems , as superconducting , quantum hall , mesoscopic and nano - systems.*pacs * : 73.20.dx ; 31.30.gs ; 71.70.ej ; 74.80.fp ; 03.67.lx ; 76.60.-k |
the study of mechanisms that allow the formation of structures such as sand dunes and ripples at the bottom of a fluid flow plays a crucial role in the understanding of coastal dynamics .
the modeling of these phenomena is particularly complex since we must not only solve the navier - stokes or saint - venant equations with equation for sediment transport , but also take into account the evolution of the bottom . instead of solving the whole system fluid flow , free surface and free bottom ,
nonlocal models of fluid flow interacting with the bottom were introduced in @xcite . among these models , we are interested in the following nonlocal conservation law @xcite : @xmath0 ( x ) - \partial_{xx}^2 u(t , x ) = 0 & t \in ( 0,t ) , x \in \mathbb{r } ,
\\ u(0,x)=u_0(x ) & x \in \mathbb{r } , \end{cases } \label{fowlereqn2}\ ] ] where @xmath1 is any given positive time , @xmath2 represents the dune height ( see fig . [ dune ] ) and @xmath3 is a nonlocal operator defined as follows : for any schwartz function @xmath4 and any @xmath5 , @xmath6 ( x ) : = \int_{0}^{+\infty } |\xi|^{-\frac{1}{3 } } \varphi''(x-\xi ) d\xi . \label{nonlocalterm}\ ] ] is the depth water , @xmath7 the free surface and @xmath8 the seabottom . ]
equation is valid for a river flow over an erodible bottom @xmath9 with slow variation and describes both accretion and erosion phenomena @xcite .
see @xcite for numerical results on this equation .
+ the nonlocal term @xmath3 can be seen as a fractional power of order @xmath10 of the laplacian with the bad sign .
indeed , it has been proved @xcite @xmath11-\varphi''\right ) ( \xi ) = \psi_{\mathcal{i}}(\xi ) { \mathcal{f}}\varphi ( \xi ) \label{fourier2}\ ] ] where @xmath12 with @xmath13 positive constants , @xmath14 denotes the fourier transform defined in and @xmath15 denotes the euler function .
one simple way to establish this fact is the derivation of a new formula for the operator @xmath3 , see proposition [ intfor2 ] . for causal functions ( i.e. @xmath16 for @xmath17 ) ,
this operator is up to a multiplicative constant , the riemann - liouville fractional derivative operator which is defined as follows @xcite @xmath18 [ liouville0 ] therefore , the fowler model has two antagonistic terms : a usual diffusion and a nonlocal fractional anti - diffusive term of lower order .
this remarkable feature enabled to apply this model for signal processing .
indeed , the diffusion is used to reduce the noise whereas the nonlocal anti - diffusion is used to enhance the contrast @xcite .
+ recently , some results regarding this equation have been obtained , namely , existence of travelling - waves @xmath19 where @xmath20 and @xmath21 represents wave velocity , the global well - posedness for @xmath22-initial data , the failure of the maximum principle and the local - in - time well - posedness in a subspace of @xmath23 @xcite .
notice that the travelling - waves are not necessarily of solitary type ( see @xcite ) and therefore may not belong to @xmath24 , the space where a global well - posedness result is available . in @xcite ,
the authors prove local well - posedness in a subspace of @xmath25 but fail to obtain global existence .
+ an interesting topic is to know if the shape of this travelling - wave is maintained when it is perturbed .
this raises the question of the stability of travelling - waves .
but before interesting ourselves to this problem , we have to show first the global existence of perturbations around these travelling - waves .
hence in this paper , we prove the global well - posedness in an @xmath22-neighbourhood of a regular travelling - wave , namely @xmath26 . to prove this result
, we consider the following cauchy problem : @xmath27(x ) - \partial^2_{xx } v(t , x ) = 0 & t \in ( 0,t ) , x \in { \mathbb{r } } , \\
v(0,x)=v_{0}(x ) & x \in { \mathbb{r } } , \end{cases } \label{fowlermodif2}\ ] ] where @xmath28 is an initial perturbation and @xmath1 is any given positive time .
+ to prove the existence and uniqueness results , we begin by introducing the notion of _ mild _ solution ( see definition [ mild2 ] ) based on duhamel s formula , in which the kernel @xmath29 of @xmath30 appears .
the use of this formula allows to prove the local - in - time existence with the help of a contracting fixed point theorem .
the global existence is obtained thanks to an energy estimate .
this approach is classical : we refer for instance to @xcite .
+ the plan of this paper is organised as follows .
in the next section , we define the notion of mild solution to and we give some properties on the kernel @xmath29 of @xmath31 that will be needed in the sequel .
sections [ uniqueness ] and [ existence ] are , respectively , devoted to the proof of the uniqueness and the existence of a mild solution for .
section [ regularity ] contains the proof of the regularity of the solution . +
* notations .
* + - the norm of a measurable function @xmath32 is written @xmath33 for @xmath34 .
+ - we denote by @xmath14 the fourier transform of @xmath35 which is defined by : for all @xmath36 @xmath37 and @xmath38 denotes the inverse of fourier transform .
+ - the schwartz space of rapidly decreasing functions on @xmath39 is denoted by @xmath40 + - we write @xmath41 .
+ - we denote by @xmath42 the space of all bounded continuous real - valued functions on @xmath39 with the norm @xmath43 .
+ - we write for any @xmath44 , @xmath45 \times { \mathbb{r}}\right):= \lbrace u \in c\left ( ( 0,t ] \times { \mathbb{r}}\right ) ; \ , \partial_t u , \partial_x u , \partial_{xx}^2 u \in c\left ( ( 0,t ] \times { \mathbb{r } } ) \right\rbrace .\ ] ] - we denote by @xmath46 the space of test functions on @xmath47 and @xmath48 denotes the distribution space .
here is our main result .
+ let @xmath44 and @xmath28 .
there exists a unique mild solution @xmath49 of ( see definition [ mild2 ] ) which satisfies @xmath50 ; l^{2}({\mathbb{r } } ) \right ) \mbox { and } v(0 , \cdot ) = v_0 \mbox { almost everywhere.}\ ] ] moreover , if @xmath51 then @xmath52 \times { \mathbb{r}}\right)$ ] and satisfies @xmath53 - \partial_{xx}^2 v = 0,\ ] ] on @xmath54 \times { \mathbb{r}}$ ] , in the classical sense or equivalently , @xmath55 is a classical solution of equation . + [ theoreme1 ]
[ mild2 ] let @xmath56 and @xmath57 .
we say that @xmath58 is a * mild solution * to if for any @xmath59 : @xmath60 where @xmath61 is the kernel of the operator @xmath62 and @xmath63 is defined in .
the expression is the duhamel formula and is obtained using the spatial fourier transform .
[ kernel20 ] the kernel @xmath29 satisfies : 1 .
@xmath64 , @xmath65 and @xmath66 , 2 .
@xmath67 , + @xmath68 , 3 .
@xmath69 @xmath70 , \hspace{0.3 cm } \vert \vert \partial_{x}k\left(t , \cdot \right)\vert \vert_{l^2\left({\mathbb{r}}\right)}\leq k_{0 } t^{-3/4}$ ] , 4 .
@xmath71 @xmath72 , \hspace{0.3 cm } \vert \vert \partial_{x}k\left(t,\cdot \right)\vert \vert_{l^1\left({\mathbb{r}}\right)}\leq k_{1 } t^{-1/2}$ ] .
( red ) and @xmath73 s ( blue ) , width=453,height=257 ] an interesting property for the kernel @xmath29 is the non - positivity ( see figure [ kernelfig20 ] ) and the main consequence of this feature is the failure of maximum principle @xcite .
we use again this property to show that the constant solutions of the fowler equation are unstable @xcite .
[ remarkestimation ] using plancherel formula , we have for any @xmath28 and any @xmath74 $ ] @xmath75 where @xmath76 . for all @xmath4 and all @xmath77 , @xmath6(x ) = \frac{4}{9 } \int_{-\infty}^{0 }
\frac{\varphi(x+z ) - \varphi(x ) - \varphi'(x)z}{|z|^{7/3 } } \ , dz . \label{integralformula}\ ] ] the proof is based on simple integrating by parts . the regularity and
the rapidly decreasing of @xmath78 ensure the validity of the computations that follow .
we have @xmath79 there is no boundary term at infinity ( resp . at zero ) because @xmath78 is a rapidly decreasing function on @xmath39 ( resp .
@xmath78 is smooth ) .
using integral formula , @xcite proved that @xmath11 \right)(\xi ) = 4 \pi^2 \gamma(\frac{2}{3})|\xi| ^{4/3 } \left ( -\frac{1}{2}+ i \frac{\sqrt{3}}{2 } \ , sgn ( \xi ) \right ) { \mathcal{f}}\varphi(\xi).\ ] ] notice that @xmath80 \right)(\xi ) = 4 \pi^2 \gamma(\frac{2}{3 } ) ( i \xi ) ^{4/3 } $ ] which is consistent with remark [ liouville0 ] : up to a multiplicative constant @xmath81 $ ] is @xmath82 .
[ fourier0 ] let @xmath83 and @xmath84 .
then @xmath81 \in h^{s-4/3}({\mathbb{r}})$ ] and we have @xmath85||_{h^{s-4/3}({\mathbb{r } } ) } \leq 4 \pi^2 \gamma(\frac{2}{3 } ) ||\varphi||_{h^s({\mathbb{r}})}.\ ] ] [ estimnonlocal ] for all @xmath83 and all @xmath84 , we have , using remark [ fourier0 ] @xmath86^{1/2 } , \\ & = & 4 \pi^2 \gamma(\frac{2}{3 } ) ||\varphi||_{h^s({\mathbb{r}})}.\end{aligned}\ ] ] from the previous proposition , we deduce that for all @xmath83 and all @xmath84 , @xmath81 \in h^{s-4/3}({\mathbb{r}})$ ] . in particular
, using the sobolev embedding @xmath87 , we deduce that @xmath88 is a bounded linear operator .
[ remarkfourier ] [ welldefined ] let @xmath56 , @xmath57 and @xmath89 .
then , the function @xmath90 \rightarrow k(t,\cdot ) \ast v_0 - \int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast w(s,\cdot)ds\ ] ] is well - defined and belongs to @xmath91;l^2({\mathbb{r}}))$ ] ( being extended at @xmath92 by the value @xmath93 ) . from proposition
[ kernel20 ] , it easy to see that @xmath94 is well - defined and that for any @xmath74 $ ] , @xmath95 .
@xmath96 so by young inequalities @xmath97 exists and using the estimates on the gradient ( item 3 and 4 of proposition [ kernel20 ] ) we deduce that @xmath94 is well - defined and @xmath98 .
+ let us prove the continuity of @xmath94 .
by the second item of proposition [ kernel20 ] , we have that the function @xmath74\rightarrow k(t,\cdot ) \ast v_{0}$ ] is continuous and it is extended continuously up to @xmath99 by the value @xmath100 .
we define the function @xmath101\rightarrow \int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast w(s,\cdot)ds.\ ] ] now , we are going to prove that @xmath102 is uniformly continuous . for any @xmath103 , young inequalities imply @xmath104 where @xmath105 are such that @xmath106 . since @xmath107 , the dominated convergence theorem implies that @xmath108 moreover , using the estimates on the gradient ( item 3 and 4 of proposition [ kernel20 ] ) , we have the following inequality @xmath109 where @xmath110 is a positive constant and @xmath111 .
from , we obtain that @xmath112 .
hence , the function @xmath102 is continuous and this completes the proof of the continuity of @xmath94 . using the semi - group property of the kernel @xmath29
, we have for all @xmath113 and all @xmath114 $ ] , @xcite @xmath115 [ semigroupe ]
let us first establish the following lemma .
let @xmath56 and @xmath116 .
for @xmath117 , let @xmath118 and define @xmath119 as in proposition [ welldefined ] by : @xmath120 then , @xmath121 ; l^{2}({\mathbb{r } } ) ) } \leq \left\ { \begin{array}{ll } 4 k_0 t^{1/4 } ||w_1 - w_2 ||_{l^\infty((0,t);l^{1}({\mathbb{r } } ) ) } & \mbox{if } w_i \in l^\infty((0,t);l^{1}({\mathbb{r } } ) ) , \\ 2 k_1 \sqrt{t } \
, ||w_1 - w_2 ||_{l^\infty((0,t);l^{2}({\mathbb{r } } ) ) } & \mbox{if } w_i \in l^\infty((0,t);l^{2}({\mathbb{r } } ) ) .
\end{array } \right.\ ] ] for all @xmath122 $ ] , we have @xmath123 hence with the help of young inequalities , we get @xmath124 it then follows that @xmath125 using again the estimates of the gradient of @xmath29 ( see proposition [ kernel20 ] ) , we conclude the proof of this lemma .
let @xmath56 and @xmath126 .
there exists at most one @xmath127 which is a mild solution to .
let @xmath128 be two mild solutions to and @xmath122 $ ] . using the previous lemma , we get @xmath129 ; l^{2}({\mathbb{r } } ) ) } \leq 2 k_0 t^{1/4 } ||v_1 ^ 2 - v_2 ^ 2 ||_{l^\infty((0,t);l^{1}({\mathbb{r } } ) ) } + 2 k_1 \sqrt{t } ||u_\phi v_1 -u_\phi v_2 ||_{l^\infty((0,t);l^{2}({\mathbb{r}}))}.\ ] ] since , @xmath130 ; l^{2}({\mathbb{r } } ) ) } \label{mmax}\ ] ] with @xmath131 ; l^{2}({\mathbb{r } } ) ) } + ||v_2||_{c([0,t ] ; l^{2}({\mathbb{r } } ) ) } , $ ] + then @xmath129 ; l^{2}({\mathbb{r } } ) ) } \leq ( 2 m k_0 t^{1/4 } + 2 k_1 t^{1/2}||u_\phi ||_{c^1_b({\mathbb{r}})})||v_1 - v_2 ||_{c([0,t ] ; l^{2}({\mathbb{r } } ) ) } .\ ] ] therefore , @xmath132 on @xmath133 $ ] for any @xmath74 $ ] satisfying @xmath134 .
since @xmath135 and @xmath136 are continuous with values in @xmath24 , we have that @xmath137 on @xmath138 $ ] where @xmath139 is the positive solution of the following equation + @xmath140 i.e. @xmath141 + to prove that @xmath142 on @xmath143 $ ] , let us define @xmath144 \mbox { s.t } v_1 = v_2 \hspace{0.1 cm } [ 0 , t ] \rbrace\ ] ] and we assume that @xmath145 . by continuity of @xmath135 and @xmath136
, we have that @xmath146 . using the semi - group property ,
see remark [ semigroupe ] , we deduce that @xmath147 are mild solutions to with the same initial data @xmath148 which implies , from the first step of this proof , that @xmath149 for @xmath150 $ ] .
finally , we get a contradiction with the definition of @xmath151 and we infer that @xmath152 .
this completes the proof of this proposition .
[ existlocal ] let @xmath116
. there exists @xmath153 that only depends on @xmath154 and @xmath155 such that admits a unique mild solution @xmath156 ; l^{2}({\mathbb{r } } ) ) \cap c((0,t_\ast ] ; h^{1}({\mathbb{r } } ) ) .$ ] moreover , @xmath94 satisfies @xmath157 } t^{1/2 } ||\partial_x v(t , \cdot)||_{l^2({\mathbb{r } } ) } < + \infty.\ ] ] for @xmath158 ; l^{2}({\mathbb{r } } ) ) \cap c((0,t ] ; h^{1}({\mathbb{r}}))$ ] , we consider the following norm @xmath159;l^2({\mathbb{r}}))}+\sup_{t \in ( 0,t ] } t^\frac{1}{2 } ||\partial_{x } v(t,\cdot)||_{l^2({\mathbb{r } } ) } \label{norme}\ ] ] and we define the affine space @xmath160;l^2({\mathbb{r } } ) ) \cap c((0,t];h^1({\mathbb{r } } ) ) \mbox { s.t .
$ v(0,\cdot)=v_0 $ and $ |||v|||
< + \infty$ } \right\}.\ ] ] it is readily seen that @xmath161 endowed with the distance induced by the norm @xmath162 is a complete metric space . for @xmath163 ,
we define the function @xmath164 \rightarrow k(t,\cdot ) \ast v_0 - \frac{1}{2 } \int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast v^2(s,\cdot ) \ , ds -\int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast u_\phi v(s,\cdot ) \ , ds.\ ] ] from proposition [ welldefined ] , @xmath165 ; l^2({\mathbb{r}}))$ ] and satisfies @xmath166 .
+ + * first step : @xmath167 . *
since @xmath168 the dominated convergence theorem implies that for any @xmath169 , @xmath170 therefore , the function @xmath171 is continuous and since @xmath14 is an isometry of @xmath22 , we deduce that @xmath172 is continuous .
we have then established that @xmath173 is continuous . moreover , from proposition [ kernel20 ] , we have @xmath174 let @xmath175 denote the function @xmath176 let us now prove that @xmath177 ; h^{1}({\mathbb{r}}))$ ] .
we first have @xmath178 using young inequalities and proposition [ kernel20 ] , we get @xmath179;l^{2}({\mathbb{r } } ) \right ) } \int_{0}^{t } k_0 ( t - s)^{-3/4}s^{-1/2 } ds \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_{x } v(s,.){\vert \vert}_{l^2({\mathbb{r } } ) } \nonumber \\ & \ , & \hspace{2 cm } + \int_{0}^{t}k_1 ( t - s)^{-1/2}s^{-1/2 } ds \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_{x } ( u_{\phi } v)(s,\cdot){\vert \vert}_{l^2({\mathbb{r}})}. \nonumber \\\end{aligned}\ ] ] we then obtain @xmath180;l^{2}({\mathbb{r } } ) ) } t^{-1/4 } \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_{x } v(s,\cdot){\vert \vert}_{l^2({\mathbb{r } } ) } \nonumber \\ & \ , & \hspace{2 cm } + k_1 j \sup_{s
\in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_{x}(u_{\phi } v)(s,\cdot ) { \vert \vert}_{l^{2}({\mathbb{r } } ) } , \end{aligned}\ ] ] where @xmath181 and @xmath182 , @xmath183 being the beta function defined by @xmath184 as @xmath185 then @xmath186 } s^{1/2 } { \vert \vert}\partial_{x } v(s,\cdot){\vert \vert}_{l^2({\mathbb{r } } ) } < \infty \mbox { and } \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_{x}(u_\phi v)(s,\cdot ) { \vert \vert}_{l^{2}({\mathbb{r } } ) } < \infty.\ ] ] we then deduce that @xmath187 is in @xmath22 and so @xmath188 for all @xmath74 $ ] .
+ let us now prove that @xmath189 is continuous on @xmath54 $ ] with values in @xmath22 . + for @xmath190 and
@xmath74 $ ] , we define @xmath191 since @xmath192 ; l^{1}({\mathbb{r}}))$ ] and @xmath193 ; l^{2}({\mathbb{r}}))$ ] then proposition [ welldefined ] implies that @xmath194\rightarrow l^{2}({\mathbb{r}})$ ] is continuous .
moreover , we have for any @xmath74 $ ] and @xmath195 , @xmath196 } s^{1/2 } { \vert \vert}\partial_{x } ( u_{\phi } v)(s,\cdot ) { \vert \vert}_{l^{2}}.\end{aligned}\ ] ] it then follows that @xmath197 } ||\partial_x w(t,\cdot)-(\partial_x w)_\delta(t,\cdot)||_{l^2({\mathbb{r}})}\rightarrow 0 \mbox { as } \delta \rightarrow 0.\ ] ] we next infer that @xmath198;l^2({\mathbb{r}}))$ ] because it is a local uniform limit of continuous functions .
hence , we have established that @xmath165 ; l^{2}({\mathbb{r } } ) ) \cap c((0,t ] ; h^{1}({\mathbb{r } } ) ) $ ] . to prove that @xmath167 , it remains to show that @xmath199 .
using and , we have @xmath200 } t^{1/2 } { \vert \vert}\partial_x \theta v(t , \cdot ) { \vert \vert}_{l^{2 } } & \leq & k_1 { \vert \vert}v_0 { \vert \vert}_{l^{2}}+ k_0 i t^{1/4 } \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_x v(s , \cdot ) { \vert \vert}_{l^{2 } } { \vert \vert}v { \vert \vert}_{c \left ( [ 0,t];l^{2 } \right ) } \nonumber \\ & + & k_1 j t^{1/2 } \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_x ( u_{\phi } v)(s , \cdot ) { \vert \vert}_{l^{2}}.\end{aligned}\ ] ] finally , we have @xmath201 .
+ * second step : * we begin by considering a ball of @xmath161 of radius @xmath202 centered at the origin @xmath203 where @xmath204 .
take @xmath205 and let us now prove that @xmath206 maps @xmath207 into itself .
we have @xmath208 by remark [ remarkestimation ] , we get @xmath209 where @xmath210 .
moreover , since @xmath211 and with the help of proposition [ kernel20 ] , we get @xmath212 therefore , for @xmath56 sufficiently small such that @xmath213 we get that @xmath214 .
+ to finish with , we are going to prove that @xmath206 is a contraction . + for @xmath215 , we have for any @xmath216 @xmath217;l^1({\mathbb{r } } ) ) } \\ & \ , & \hspace{1.5 cm } + \ , 2 k_1 t^{1/2 } { \vert \vert}u_\phi { \vert \vert}_{c^{1}_{b}({\mathbb{r}})}{\vert \vert}v - w{\vert \vert}_{c([0,t];l^2({\mathbb{r } } ) ) } , \end{aligned}\ ] ] and since , @xmath218;l^1({\mathbb{r } } ) ) } & \leq & ( { \vert \vert}v{\vert \vert}_{c([0,t];l^2({\mathbb{r}}))}+{\vert \vert}w{\vert \vert}_{c([0,t];l^2({\mathbb{r}}))}){\vert \vert}v - w{\vert \vert}_{c([0,t];l^2({\mathbb{r } } ) ) } , \\ & \leq & 2 r { \vert \vert}v - w{\vert \vert}_{c([0,t];l^2({\mathbb{r}}))},\end{aligned}\ ] ] we get @xmath219;l^2({\mathbb{r}}))}.\ ] ] moreover @xmath220 } s^{1/2 } { \vert \vert}(v \partial_x v - w\partial_x w)(s , \cdot ) { \vert \vert}_{l^{1}({\mathbb{r } } ) } \\ & \ , & \hspace{1 cm } + \ , k_1 j \sup_{s \in ( 0,t ] } s^{1/2 } { \vert \vert}\partial_x \left ( u_{\phi}(v -w ) \right ) ( s,\cdot ) { \vert \vert}_{l^{2}({\mathbb{r}})}.\end{aligned}\ ] ] and since @xmath221 then @xmath222 therefore , we obtain @xmath223 finally , using and , we get @xmath224|||v - w||| . \\\end{aligned}\ ] ] * last step : conclusion . * for any @xmath153 sufficiently small such that holds true and @xmath225 @xmath206 is a contraction from @xmath226 into itself .
the banach fixed point theorem then implies that @xmath206 admits a unique fixed point @xmath156 ; l^{2}({\mathbb{r } } ) ) \cap c((0,t_\ast ] ; h^{1}({\mathbb{r}}))$ ] which is a mild solution to .
let @xmath116 and @xmath51 .
there exists @xmath227 that only depends on @xmath154 and @xmath228 such that admits a unique mild solution @xmath229 ; l^{2}({\mathbb{r } } ) ) \cap c((0,t_\ast ' ] ; h^{2}({\mathbb{r } } ) ) .$ ] moreover , @xmath94 satisfies @xmath230 } t^{1/2 } ||\partial_{x } v(t , \cdot ) ||_{l^2({\mathbb{r } } ) } < + \infty \hspace{0.3 cm } \mbox { and } \hspace{0.3 cm } \sup_{t \in ( 0,t ' _ * ] } t ||\partial^2_{xx } v(t , \cdot ) ||_{l^2({\mathbb{r } } ) } < + \infty.\ ] ] [ regularityh2 ] to prove this result , we use again a contracting fixed point theorem . but this time , it is the gradient of the solution @xmath94 which is searched as a fixed point .
+ from proposition [ existlocal ] , there exists @xmath153 which depends on @xmath231 and @xmath232 such that @xmath156 ; l^{2}({\mathbb{r } } ) ) \cap c((0,t_\ast ] ; h^{1}({\mathbb{r } } ) ) $ ] is a mild solution to .
since @xmath233 ; h^{1}({\mathbb{r}}))$ ] , we can consider the gradient of @xmath234 for any @xmath235 $ ] .
let then @xmath236 and @xmath237 $ ] .
we consider the same complete metric space @xmath161 defined in the proof of proposition [ existlocal ] and we take the norm @xmath238 defined in : @xmath239;l^2({\mathbb{r } } ) ) \cap c((0,t^{'}_*];h^1({\mathbb{r } } ) ) \mbox { s.t .
$ w(0,\cdot)=w_0 $ and $ |||w|||
< + \infty$ } \right\},\ ] ] with the initial data @xmath240 .
+ we now wish to apply the fixed point theorem at the following function @xmath241 & \rightarrow & k(t,\cdot ) \ast w_0 - \int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast \left ( \bar{v } w \right ) ( s,\cdot)ds \\ & \ , & \hspace{1 cm } - \int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast \left ( \partial_x ( u_{\phi } ) \bar{v } \right)(s , \cdot ) \ , ds \\ & \ , & \hspace{1 cm } - \int_{0}^{t } \partial_x k ( t - s,\cdot ) \ast \left ( u_{\phi } w \right ) ( s , \cdot ) \ , ds , \end{aligned}\ ] ] where @xmath242 .
first , we leave to the reader to verify that @xmath243 maps @xmath161 into itself .
the proof is similar to the one given in proposition [ existlocal ] .
+ for any @xmath244 , we have from young inequalities and remark [ remarkestimation ] @xmath245 ; h^{1}({\mathbb{r } } ) \right ) } \int_{0}^{t } ||\partial_x k(t - s , \cdot)||_{l^{1}({\mathbb{r } } ) } ds \\ & \ , & \hspace{1 cm } + \ , ||u _ { \phi } ||_{c^1_b({\mathbb{r } } ) } ||| w ||| \int_{0}^{t } ||\partial_x k(t - s , \cdot)||_{l^{1}({\mathbb{r } } ) } ds , \end{aligned}\ ] ] and from proposition [ kernel20 ] , we get @xmath246 ; h^{1 } \right ) } + 2 k_1 t'^{1/2 } _ * ||u _ { \phi } ||_{c^1_b } ||| w ||| .
\label{estim1h2}\end{aligned}\ ] ] differentiating @xmath247 w.r.t the space variable , we obtain @xmath248 and developing , we get @xmath249(s , \cdot ) \ , ds \\ & \ , & - \int_0^t \partial_x k(t - s , \cdot ) \ast \left [ \partial_x^2(u_\phi ) \ , \bar{v } + \partial_x(u_\phi ) \ , \partial_x \bar{v}\right ] ( s , \cdot ) \ , ds \\ & \ , & - \int_0^t \partial_x k(t - s , \cdot ) \ast \left [ \partial_x(u_\phi ) \ , w + u_\phi \ , \partial_x w \right](s , \cdot ) \ , ds . \end{aligned}\ ] ] now , from young inequalities , we have @xmath250 \ , ds \\ & \ , & + \int_0^t ||\partial_x k(t - s , \cdot)||_{l^1 } \left [ || \partial_x(u_\phi ) \ , w ( s , \cdot)||_{l^2 } + ||u_\phi \ , \partial_x w(s , \cdot ) ||_{l^2 } \right ] \ , ds.\end{aligned}\ ] ] finally , from proposition [ kernel20 ] , we obtain @xmath251 ; h^{1 } ) } \sup_{s \in ( 0 , t ' _ * ] } s^{1/2 } ||\partial_x w ( s , \cdot)||_{l^2 } \\ & \ , & + 4 \ , k_1 \ , t^{1/2 } \ , ||u_\phi ||_{c^2_b } || \bar{v}||_{c([t_0 ; t^ { ' } _ * ] ; h^{1 } ) } + 2 \ , k_1 \ , t^{1/2 } \ , ||u_\phi ||_{c^1_b } ||| w ||| \\ & \ , & + \int_0^t k_1 \ , ( t - s)^{-1/2 } \ , s^{-1/2 } \ , ds
\ , ||u_\phi||_{c^2_b } \sup_{s\in ( 0 , t ' _ * ] } s^{1/2 } ||\partial_x w ( s , \cdot)||_{l^2}. \end{aligned}\ ] ] in other words , we have for all @xmath252 $ ] @xmath253 ; h^{1 } ) } |||w ||| \nonumber \\ & \ , & + k_0 \ , i \ , t_{*}^{'1/4 } \ , || \bar{v}||_{c([t_0 ; t^ { ' } _ * ] ; h^{1 } ) } \ , |||w ||| + 4 \ , k_1 \ , t_{*}^{'}\ , ||u_\phi ||_{c^2_b } || \bar{v}||_{c([t_0 ; t^ { ' } _ * ] ; h^{1 } ) } \nonumber \\ & \ , & + 2 \ , k_1 \ , \pi \ , t_{*}^ { ' } \ , ||u_\phi ||_{c^1_b } ||| w ||| + k_1 \ , t_{*}^{'1/2 } \ , ||u_\phi||_{c^2_b } \ , |||w ||| , \label{estim2h2}\end{aligned}\ ] ] where @xmath254 . hence , using and , we get @xmath255 for some positive constant @xmath256 which depends on @xmath257 ; h^{1}({\mathbb{r}}))}$ ] and @xmath258 . +
we next leave to reader to verify that : for any @xmath259 , @xmath260 is a positive constant which depends on @xmath257 ; h^{1}({\mathbb{r}}))}$ ] and @xmath258 . + then
, if @xmath261 satisfies @xmath262 ; h^{1}({\mathbb{r}}))}(2 t ' _ * + t^{'1/2 } _ * ) \\ & \ , & \hspace{6.5 cm } + \ , c \ , r \ , ( t^{'1/4 } _ * + t^{'1/2}_*+ t^{'3/4 } _ * + t^ { ' } _ * ) \leq r,\end{aligned}\ ] ] and @xmath263 @xmath264 is a contraction , where @xmath265 is ball of @xmath161 of radius @xmath202 centered at the origin . using a contracting point fixed theorem , it exists a unique fixed point , which we denote by @xmath175 .
but it is easy to see that @xmath266 taking into account the space derivated of the duhamel formulation .
thanks to a uniqueness argument , we deduce that @xmath267 and thus that @xmath268 ; h^{2}({\mathbb{r}}))$ ] , which completes the proof of this lemma .
let us now prove the global - in - time existence of mild solution @xmath94 .
let @xmath116 , @xmath51 and @xmath56 .
then , there exists a ( unique ) mild solution @xmath158 ; l^{2}({\mathbb{r}}))\cap c((0,t ] ; h^{2}({\mathbb{r}}))$ ] to .
moreover , @xmath94 satisfies the pde in the distribution sense . *
first step : @xmath94 is a distribution solution .
* taking the fourier transform w.r.t the space variable in , we get for all @xmath122 $ ] and all @xmath36 , @xmath269 define @xmath270 classical results on ode imply that @xmath271 is differentiable w.r.t the time with + @xmath272 let us now prove that all terms in are continuous with values in @xmath22 . since , @xmath273 ; h^{1}({\mathbb{r}}))$ ] then @xmath274 ; l^{2}({\mathbb{r}}))$ ] .
we thus deduce that @xmath275 and @xmath276 are continuous with values in @xmath24 .
moreover , equation implies that @xmath277 and so @xmath278 is continuous with values in @xmath22 .
indeed , + @xmath279 } | \psi_{\mathcal{i}}(\xi ) |^2 ||g(t,\cdot ) ||^2_{l^2({\mathbb{r } } ) } + c \int_{{\mathbb{r}}\setminus ( -1,1 ) } |\xi^2 g(t,\xi ) |^2 \ , d\xi , \\ & \leq & \sup_{\xi \in [ -1,1 ] } | \psi_{\mathcal{i}}(\xi ) |^2 ||g(t,\cdot ) ||^2_{l^2({\mathbb{r } } ) } \\ & \ , & \hspace{2.2 cm } + \ , c \ , \int_{{\mathbb{r}}\setminus ( -1,1 ) } | { \mathcal{f } } ( \partial^2_{xx }
v(t,\cdot ) ) - \xi ^2 e^{-t \psi_{\mathcal{i}}(\xi ) } { \mathcal{f}}v_0 |^2 \ , d\xi , \\ & \leq & \sup_{\xi \in [ -1,1 ] } | \psi_{\mathcal{i}}(\xi ) |^2 ||g(t,\cdot ) ||^2_{l^2({\mathbb{r } } ) } + \tilde{c } || v(t,\cdot ) ||^2_{h^{2}({\mathbb{r } } ) } \\ & \ , & \hspace{2.5 cm } + \tilde{c}|| v_0||_{l^{2}({\mathbb{r}})}^2 + c ||v(t , \cdot)||_{h^2}||v_0||_{l^2 } , \\ &
< & \infty,\end{aligned}\ ] ] because @xmath63 behaves at infinity as @xmath280 .
@xmath281 are two positive constants .
hence , we have that the function @xmath282 is continuous .
finally , we have proved that all the terms in are continuous with values in @xmath22 .
therefore , from , we get that @xmath283;l^2({\mathbb{r}},\mathbb{c } ) ) $ ] and then @xmath284 moreover , @xmath122\rightarrow e^{-t \psi_{{\mathcal{i } } } } { \mathcal{f}}v_0 \in l^{2}({\mathbb{r}},\mathbb{c})$ ] is @xmath285 with @xmath286 from equation , we infer that @xmath287 is @xmath285 on @xmath143 $ ] with values in @xmath22 with @xmath288 since @xmath14 is an isometry of @xmath22 , we deduce that @xmath289 ; l^{2}({\mathbb{r}}))$ ] and by , we get @xmath290 + \partial_{xx}^2 v(t,\cdot)-\partial_x(\frac{v^2}{2})(t,\cdot)-\partial_x(u_\phi v)(t,\cdot ) . \label{egalitedistribution}\end{aligned}\ ] ] we are now going to prove that @xmath94 satisfies the pde in the distribution sense .
let us note @xmath291 + \partial_{xx}^2 v(t,\cdot)-\partial_x(\frac{v^2}{2})(t,\cdot)-\partial_x(u_\phi v)(t,\cdot)\ ] ] and let us show that @xmath292 by definition , we have for any @xmath293 and @xmath294 : @xmath295 therefore , it is enough to prove that @xmath296 i.e. @xmath297 in the sense of @xmath298 .
but by , we have that the function @xmath299 is @xmath300 and @xmath301 in the classical sense , which proves that the mild solution @xmath94 is a distribution solution of . + * second step : a priori estimate . * by the first step
, we have @xmath302-\partial_{xx}^{2 } v=0\ ] ] in the distribution sense .
therefore , multiplying this equality by @xmath94 and integrating w.r.t the space variable , we get : @xmath303-v_{xx}\right)v \ : dx + \int_{{\mathbb{r } } } \left(u_\phi v\right)_{x } v \ : dx=0\ ] ] because the nonlinear term is zero . indeed , integrating by parts
, we have @xmath304 there is no boundary term from the infinity because for all @xmath74 $ ] , @xmath305 . using and the fact that @xmath306-\partial_{xx}^2 v ) v \ , dx$ ] is real , we get @xmath307-\partial_{xx}^2 v ) v \ ,
dx = \int_{\mathbb{r}}{\mathcal{f}}^{-1}(\psi_{\mathcal{i}}{\mathcal{f}}v ) v \ , dx = \int_{\mathbb{r}}\psi_{\mathcal{i}}|{\mathcal{f}}v|^2 d\xi = \int_{\mathbb{r}}\mbox{re}(\psi_{\mathcal{i } } ) |{\mathcal{f}}v|^2 d\xi.\ ] ] moreover , since @xmath308 we have @xmath309 using , and , we obtain @xmath310 where @xmath311 and @xmath312 . finally , we get for all @xmath122 $ ] the following estimate @xmath313 + * last step : global - in - time existence . * up to this point , we know thanks to proposition [ existlocal ] and lemma [ regularityh2 ] that there exists @xmath314 such that @xmath156 ; l^2({\mathbb{r } } ) ) \cap c((0,t_\ast ] ; h^2({\mathbb{r}}))$ ] is a mild solution of on @xmath315 $ ] .
let us define @xmath316 to prove the global - in - time existence of a mild solution , we have to prove that @xmath317 , where @xmath1 is any positive constant .
assume by contradiction that @xmath145 . with again the help of proposition [ existlocal ] , there exists @xmath318 such that for any initial data @xmath319 that satisfy @xmath320 it exists a mild solution @xmath175 on @xmath321 $ ] . using , we have that @xmath322 satisfies .
therefore , using an argument of uniqueness , we deduce that + @xmath323 for all @xmath324 . to finish with
, we define @xmath325 by @xmath326 on @xmath327 and @xmath328 for @xmath329 $ ] .
hence , @xmath325 is a mild solution on @xmath330 $ ] with initial datum @xmath331 , which gives us a contradiction .
this section is devoted to the proof of the existence of classical solutions @xmath94 to .
let @xmath116 , @xmath51 and @xmath44 . the unique mild solution @xmath158 ; l^{2}({\mathbb{r}}))\cap c((0,t ] ; h^{2}({\mathbb{r}}))$ ] of belongs to @xmath332 \times { \mathbb{r}}\right ) $ ] and satisfies @xmath333 - \partial^2_{xx } v = 0,\ ] ] on @xmath54 \times { \mathbb{r}}$ ] in the classical sense .
+ [ classicalsense ] * first step : @xmath334-regularity in space .
* let us take any @xmath335 $ ] as initial time and let @xmath336 $ ] . differentiating the duhamel formulation two times w.r.t the space
, we get for any @xmath337 $ ] , @xmath338 where @xmath339 and @xmath340 . since @xmath273 ;
h^2({\mathbb{r}}))$ ] then @xmath341 ; l^{2}({\mathbb{r } } ) \right)$ ] and from the sobolev embedding @xmath342 , we get that @xmath343 ; l^1({\mathbb{r}})\cap l^{2}({\mathbb{r } } ) ) $ ] .
let us now define the following functions @xmath344 for all @xmath345 , we have thanks to cauchy - schwartz inequality @xmath346 where @xmath347 denotes the translated function @xmath348 .
+ therefore , for all @xmath345 and all @xmath349 $ ] , we deduce that @xmath350 } ||\mathcal{t}_{(x - y)}\left(\bar{u}_{i}(s , \cdot ) \right ) - \bar{u}_i(s , \cdot ) ||_{l^{2}({\mathbb{r } } ) } , \label{continuity}\end{aligned}\ ] ] where @xmath351 .
then , @xmath352 is uniformly continuous with values in @xmath22 as a continuous function on a compact set @xmath353 $ ] .
therefore , for any @xmath354 , there exists a finite sequence @xmath355 such that for any @xmath356 $ ] , there exists @xmath357 such that @xmath358 therefore , using we have @xmath359 } ||\mathcal{t}_{(x - y)}\left(\bar{u}_{i}(s , \cdot ) \right ) -\mathcal{t}_{(x - y)}\left(\bar{u}_{i}(s_j , \cdot ) \right)||_{l^{2 } } \\ & + & 4 k_0 t'^{1/4 } \left\lbrace ||\mathcal{t}_{(x - y)}\left(\bar{u}_{i}(s_j ,
\cdot ) \right ) - \bar{u}_{i}(s_j , \cdot)||_{l^{2 } } + \sup_{s \in [ 0,t ] } ||\bar{u}_{i}(s , \cdot ) - \bar{u}_{i}(s_j , \cdot)||_{l^{2 } } \right\rbrace .\end{aligned}\ ] ] and since @xmath360 , we get @xmath361 and since the translated function is continuous in @xmath24 , we have @xmath362 as @xmath363 . hence ,
@xmath364 taking the infimum w.r.t @xmath365 , we infer that @xmath366 is continuous w.r.t the variable @xmath367 . moreover , arguing as the proof of proposition [ welldefined ] , we get that @xmath368 ; l^{2}({\mathbb{r } } ) \right)$ ] . from classical results
, we then deduce that @xmath366 is continuous w.r.t the couple @xmath369 on @xmath370 \times { \mathbb{r}}$ ] .
+ moreover , since @xmath371 , we can easily check that @xmath372 is continuous on @xmath54 \times { \mathbb{r}}$ ] .
finally , we get that @xmath373 \times { \mathbb{r}}\right)$ ] and since @xmath151 is arbitrary in @xmath54 $ ] , we conclude that @xmath374 \times { \mathbb{r}}\right)$ ] . + * second step : @xmath285-regularity in time . * from proposition [ global ] , we know that the terms @xmath375 and @xmath376 $ ] have the same regularity .
moreover , by the first step of this proposition , we have that @xmath377 \times { \mathbb{r}})$ ] and from sobolev embeddings and remark [ remarkfourier ] , we deduce that @xmath378 and @xmath379 $ ] belong to @xmath380 \times { \mathbb{r } } ) $ ] . finally , we obtain that @xmath381 \times { \mathbb{r}})$ ] and thus @xmath382 \times { \mathbb{r}})$ ] .
the proof of this proposition is now complete . , _ an introduction to fractional derivatives , fractional differential equations , to methods of their solution and some of their applications .
_ , mathematics in science and engineering , * 198 * academic press , san diego , ( 1999 ) . | we investigate a fractional diffusion / anti - diffusion equation proposed by andrew c. fowler to describe the dynamics of sand dunes sheared by a fluid flow . in this paper , we prove the global - in - time well - posedness in the neighbourhood of travelling - waves solutions of the fowler equation .
+ * keywords : * nonlinear and nonlocal conservation law , fractional anti - diffusive operator , duhamel formulation , travelling - wave , global - in - time existence . + * mathematics subject classification : * 35l65 , 45k05 , 35g25 , 35c07 . |
the dissociative recombination ( dr ) of small molecular ions that collide with electrons plays an important role in interstellar diffuse and dense clouds .
it is well known that these clouds constitute building material for new stars .
the importance of dr cross sections as parameters in star formation models , for such astrophysically relevant ions as h@xmath2 and hco@xmath0 , is one of the reasons why dr has been extensively studied in laboratory experiments @xcite . for diatomic molecules
the process is well understood and described theoretically @xcite . until recently @xcite though , theory was unable to model the dr process in triatomic ions , except in cases where a neutral dissociative state crossed the ionic born - oppenheimer surface in the franck - condon region , generating a rapid rate @xcite . in triatomics ,
a key complication is the fact that multiple vibrational and rotational degrees of freedom must generally be taken into account . and
in addition to the greater computational burden of treating more dimensions quantum mechanically , the addition of these new degrees of freedom can also lead to new conceptual issues , related to the degeneracy of vibrational or rovibrational levels in certain triatomic ions ( e.g. , in both h@xmath1 and hco@xmath0 ) .
this can cause an intrinsic instability of the corresponding neutral molecules @xcite that causes them to distort away from the symmetric geometry . consider an incident electron that interacts with a closed - shell triatomic ion having a degenerate vibrational mode .
if the symmetry group @xmath3 of the resulting neutral complex has at least one degenerate irreducible representation , the electronic partial wave components with angular momentum @xmath4 typically contain at least one such representation , whereby the corresponding electronic states of the neutral molecule are degenerate in the clamped - nuclei approximation .
due to the jahn - teller theorem @xcite , first formulated by landau @xcite , this degenerate electronic state strongly interacts with a degenerate vibrational state of the @xmath3 group .
the interaction leads to either a quadratic ( renner - teller ) @xcite or linear conical ( jahn - teller ) intersection , where `` quadratic '' or `` linear '' refers to the dependence of the adiabatic eigenvalues on the symmetry distortion coordinate ; e.g. , the electronically - degenerate linear molecule being considered in the present article displays a quadratic behavior near the symmetry point , as a function of the bending angle @xmath5 . in electron - induced reactions ,
electron capture can be followed by dissociation or autoionization of the recombined system .
( in some rare cases , recombination could also occur radiatively following electron capture , but this is rarely important in molecular systems having an open dissociation channel and it will be ignored throughout this study . )
since the degenerate electronic state causes the instability of the symmetric configuration , the recombined molecule quickly distorts away from the symmetry point to remove the degeneracy , after which the autoionization channel typically becomes energetically closed , and the molecule eventually finds a pathway along which it can dissociate ( provided there is an electronic state of the molecule that is open for dissociation ) . in a system for which no direct crossing of a neutral dissociative state with the ground state of the ion is present , this indirect mechanism can become dominant , and it can still produce large dr cross sections .
this is expected to be particularly true for molecules containing hydrogen .
work by some members of the present collaboration have recently found that this is what happens in h@xmath6 collisions @xcite . for hco@xmath0 , a closed - shell linear ion in its ground state
, the picture appears to be similar .
the lowest doubly - degenerate vibrational @xmath7 states are coupled to the electronic states @xmath8 of the neutral system through the renner - teller interaction , resulting in a large probability for recombination .
although renner - teller coupling is a well - known phenomenon in spectroscopy , the effect of renner - teller coupling on electron - molecule collisions has not been studied until recently .
_ @xcite have made an important contribution to understanding the role of renner - teller physics in collisions between an electron and the co@xmath9 molecule . in a completely @xmath10 @xmath11 treatment
, the authors employed a time - dependent framework to describe the nuclear motion of the negative ion co@xmath12 formed during the collision .
the nuclei move on the two - component @xmath13 and @xmath14 electronic complex potentials corresponding to the doubly - degenerate @xmath15 molecular state at linear configuration .
the two components interact through the renner - teller coupling mechanism .
the aim of the present study is to understand the mechanism of dr in hco@xmath16 and , in particular , to interpret its large measured rate .
this ion was first detected in space and then synthesized in the laboratory @xcite .
experimental measurements of dr in hco@xmath0 were realized in a number of merged - beam , afterglow - plasma , and storage ring experiments @xcite . in ref .
@xcite we have presented the potential energy surfaces for the ground molecular state of hco@xmath0 and the lowest states of hco up to principal quantum number @xmath17 , calculated as functions of all three vibrational coordinates .
the potentials were then used to estimate the dr rate .
since the splitting between the two @xmath8 electronic states for the principal quantum number @xmath18 was found to be small , it was suggested that the renner - teller non - born - oppenheimer coupling should not play a significant role in dr of hco@xmath0 , therefore , it was not included in the model of ref .
however , since even a small coupling between degenerate states can in principle result in a large dr probability , we investigate here the role of renner - teller coupling in hco in more detail , and find that we must revise that initial expectation .
the article is organized in the following way . in section [ sec : ham_rt ] we construct the model hamiltonian from the results of the _ ab initio _ calculation .
the hamiltonian includes the nonadiabatic renner - teller coupling and depends parametrically on the nuclear configurations .
section [ sec : cs ] describes the quantum defect approach to obtain the reaction matrix from the hamiltonian .
it also describes how the dr cross section is obtained from the reaction matrix .
section [ sec : results ] presents and discusses the results of the calculation : the dr cross section and the thermally averaged dr rate .
section [ sec : auoionization ] considers autoionization in detail .
autoionization is a process that competes with dr and it decreases the dr cross section .
section [ sec : stefano_paper ] compares the present results with a previous theoretical study @xcite .
finally , section [ sec : conclusion ] discusses our conclusions .
the starting point of the present theoretical approach is the molecular hamiltonian @xmath19 of hco , which we represent as @xmath20 , where @xmath21 is the ionic hamiltonian , @xmath22 is the kinetic energy of the incident electron , and @xmath23 describes the electron - ion interactions .
we assume that the ion is in its ground electronic state .
@xmath21 and @xmath23 depend on the four internuclear coordinates @xmath24 , where @xmath25 and @xmath26 are the c - h and c - o internuclear distances , @xmath27 is the bending angle , which is zero for linear configurations .
the electronic energies are independent of the angle @xmath28 , which represents the azimuthal orientation of the bending .
we assume that the incident electron is initially captured in one of the neutral rydberg states , which can be approximately characterized by the principal quantum number @xmath29 , the orbital angular momentum @xmath30 , and its projection on the molecular axis @xmath31 .
in our model we have included only the @xmath32,@xmath33 , @xmath34 , @xmath35 and @xmath36 states .
the symbols @xmath37 imply two opposite - sense electronic angular momenta associated with the different signs of @xmath31 .
linear combinations of the @xmath38 states give states @xmath39 and @xmath40 , symmetric and antisymmetric with respect to reflection in the plane containing the molecular axis .
the other three states are symmetric .
these states were established in ref .
@xcite to be the most important for dr , since they exhibit the largest dependence on the vibrational coordinates .
therefore , @xmath41 assumes a block - diagonal form with an infinite number of @xmath42 blocks corresponding to @xmath43 in addition to the continuum . since the only off - diagonal couplings included in our analysis
are those among the three @xmath44 states , for simplicity we specify only these states in the formulas below .
the @xmath44-block of @xmath41 in the basis of the @xmath32 , @xmath34 and @xmath45 states has the form @xcite @xmath46 where @xmath47 and @xmath48 are the electronic energies of the @xmath49 and @xmath50 states at the linear ionic configuration ; @xmath51 and @xmath52 are the real , non - born - oppenheimer coupling elements .
we denote both the ( @xmath53 ) and ( @xmath54 ) couplings as renner - teller ( rt ) couplings , whereas in some previous studies , only the former is denoted by this term . the couplings @xmath55 and @xmath52 depend on @xmath56 , and @xmath57 .
the diagonalization of @xmath58(% \mathcal{q})$ ] is accomplished by the unitary transformation matrix @xmath59 @xcite : @xmath60 with the abbreviations @xmath61 when diagonalized , the hamiltonian becomes @xmath62u=\mathrm{{diag}\{v_{\pi ' ' } , v_{\sigma } , v_{\pi ' } \}\,,}\ ] ] where @xmath63 the adiabatic potential energy surfaces @xmath64 are known from _
ab initio _ calculations ( see ref .
@xcite for a detailed description ) . from eq .
( [ eq:21c ] ) we obtain @xmath65 therefore , the matrices @xmath59 in eq .
( [ eq : u_trans ] ) and @xmath23 in eq .
( [ eq : hint ] ) are obtained from @xmath66 .
as in our previous dr studies @xcite , we employ multichannel quantum defect theory and need to construct the reaction matrix @xmath67 , which is related to the potential as in ref . @xcite .
( we note , however , that the reaction matrix @xmath67 of ref .
@xcite is not the @xmath67-matrix of mqdt , which includes an additional factor of @xmath68 . )
first we introduce the diagonal quantum defect matrix @xmath69 , whose nonvanishing elements @xmath70 are @xmath71 $ ] , where @xmath72 is the electronic state index .
therefore , the diagonal form of the reaction matrix is directly obtained from the _ ab initio _ calculation . for fixed @xmath73 a.u . and @xmath74 a.u .
labeling the states with electronic momentum projection @xmath75 is an approximation and for large values of @xmath76 it is not appropriate .
however , the four @xmath77 , @xmath78 , @xmath79 , and @xmath80 states are always uncoupled from @xmath81 states for nonlinear geometries . in the absence of the non - diagonal matrix elements in eq .
[ eq : hint ] the @xmath82 and @xmath83 states would have the same quantum defects .
the renner - teller parameters are completely determined by the geometry dependence of @xmath84 and @xmath85 defects ( eq . [ eq : rt_param ] ) . ]
the coupling @xmath86 in eq .
[ eq : hint ] is linear with @xmath5 for small @xmath5 .
correspondingly , the splitting between adiabatic electronic energies @xmath87 and @xmath88 is quadratic with respect to @xmath5 , and , therefore , the splitting between quantum defects @xmath89 and @xmath90 is also quadratic .
this can be seen in fig .
[ qd ] , where @xmath89 is almost constant because it has a different symmetry and is unchanged due to the renner - teller effect .
the quantum defect @xmath90 is approximately quadratic due to the renner - teller effect until @xmath91 , beyond which the linear approximation for @xmath92 is not adequate . in a preliminary calculation
we used such a linear dependence of @xmath92 , because we have found that we can use formulas eqs .
[ eq : u_trans ] , [ eq : w ] , [ eq : u ] , [ eq : ve ] , and [ eq : rt_param ] and fit the model hamiltonian accurately to the _ ab initio _ energies without the additional approximation of linearity of @xmath92 .
the result of the preliminary calculation is consistent with the present calculation using the accurate fit to the _ ab initio _ energies .
the next step is to bring the reaction matrix into the representation of the hamiltonian of eq .
[ eq : hint ] . since the transformation matrix @xmath59 is known from eq .
[ eq : u ] , the @xmath67 matrix in that representation reads @xmath93 the matrices @xmath94 , @xmath95 are diagonal for the @xmath96 and @xmath97 states .
we used quantum defects from ref .
@xcite with @xmath17 for @xmath96 and @xmath18 for the other states .
figure [ qd ] gives the quantum defects as functions of @xmath5 , with @xmath98 and @xmath26 fixed at the ionic equilibrium values .
ion as a function of the @xmath25 distance for different projections @xmath99 of the vibrational angular momentum .
the curves of the same color correspond to the vibrational states with the same quantum number @xmath100 but different @xmath101 .
the energies of the lowest vibrational states with @xmath102 and @xmath103 are shown by two horizontal lines . ] once the reaction matrix @xmath104 is obtained , the dr treatment is along the lines of refs . @xcite with the following differences .
we choose here the @xmath25 distance as the dissociative adiabatic coordinate .
this assumption is justified because experimentally the h+co channel is largely dominant at low energies .
it has been long known that the ch+o channel is endothermic by 0.17 ev and could have an energy barrier , while the oh product is not observed ( see ref . @xcite and references therein ) .
recently , branching ratios have been measured @xcite for the dco@xmath105 dissociative recombination , confirming that the d+co channel is by far the dominant , with a ratio of 0.88 .
we keep the @xmath106 distance frozen at its ionic equilibrium value .
this assumption was made to simplify the treatment but it could result in an underestimation of the dr cross section . in order to account for the co vibration one might use the hyper - spherical coordinates similar to ref .
@xcite , or else normal mode coordinates as in mccurdy _
although the co bond is frozen , the renner - teller coupling physics is included in our model ; we show below that it does increase the dr cross section significantly
. we will briefly address below the possible influence of the co vibration .
consequently , for every @xmath25 distance we calculate matrix elements of the `` large '' reaction matrix @xmath107 @xmath108 where @xmath109 are vibrational wavefunctions of hco@xmath0 parametrically dependent on @xmath25 , while @xmath26 is frozen .
they are eigenfunctions of @xmath21 with eigenvalues @xmath110 .
each index @xmath111 in @xmath112 can be represented as @xmath113 , where @xmath114 specifies the projection of the vibrational angular momentum on the molecular axis , while the index @xmath101 distinguishes different vibrational states with the same @xmath114 .
the rotation of the whole molecule is not considered .
[ energies ] shows the vibrational eigenenergies @xmath115 as a function of @xmath25 .
it is informative to compare the vibrational energies obtained in our simplified approach with the exact calculation accounting for all four vibrational coordinates puzzarini96 .
we can not compare the absolute values of the energies because the co stretch is frozen , however , we can compare the energies of a few vibrational levels @xmath116 with respect to the energy @xmath117 of the ground vibrational level @xmath118 . table [ tab : vibr_ene ]
compares the vibrational energies obtained in this study with accurate results from ref .
the overall accuracy of the present calculated vibrational levels is about 12 % or less .
since the vibrational wave function error is of order @xmath119 , a conservative estimate of the error in our calculations of the final cross section would be of order @xmath120 .
.the table demonstrates the accuracy of the vibrational energies obtained in the present study neglecting the co vibration with the exact calculation from ref .
since the co bond is frozen , we do nt provide energies for excited @xmath121 modes .
the overall error is about 12 % , which translates into about 25 % for vibrational wave functions .
the energies are given in @xmath122 . [ cols="<,<,<",options="header " , ] table [ tab:1 ] conveys an idea for the order of magnitude of the parameters in eq .
[ eq : p_adiab ] .
this table shows two examples of avoided crossings that lie well below the energy at which the electron can be captured . at such large values of @xmath123 , the landau - zener probability @xmath124 ,
depends only weakly on the electron energy .
the two other examples are taken from the region around the left turning point where the velocity of motion could be small or large depending on the energy of the electron . in cases like this one , we calculated the probability @xmath125 for a range of values of the kinetic energy @xmath123 .
there are many such avoided crossings that occur near left turning points and comparatively fewer at right turning points . during half ( @xmath126 ) of an oscillation period
, the system goes through @xmath127 avoided crossings .
we estimate that @xmath127 is about 10 for a typical resonance curve .
therefore , the total predissociation probability @xmath128 is somewhere in the range 0.001 - 0.01 . in the above estimation of the predissociation probability
we have assumed that only two states interact at each avoided crossing .
however , in reality many of the avoided crossing can not be considered as strictly two - state ones : three or more states @xmath129 can participate .
the probability of multistate crossing might conceivably give a larger predissociation probability .
the autoionization probability @xmath130 during time @xmath126 is @xmath131 , where @xmath132 is the total autoionization width .
the largest @xmath132 is about @xmath133 a.u .
, @xmath134 , where @xmath135 is the frequency of oscillations in the ground vibrational level , @xmath136 a.u . thus , @xmath137 0.002 .
this is presumably an upper bound on the autoionization probability . showing the interval of @xmath25 corresponding to the left turning point of the vibrational motion . for convenience ,
the energy of the ground ionic curve @xmath138 was subtracted from all the data . in the figure
, curves with positive and small negative slopes correspond to rydberg states converging to the nearest ionic thresholds .
the probability to capture the electron tends to be high for such states , diminishing however at high principal quantum numbers as @xmath139 . ] as one can see , the autoionization probability could be competitive with predissociation in our model if we take the smallest ratio @xmath140 , but it is more likely that the ratio is of the order of 2 - 3 and therefore , more favorable to predissociation . in the previous theoretical study of dr in h@xmath1 @xcite ,
the ratio was significantly larger .
an important difference with the h@xmath141 study is the two degrees of freedom taken into account when adiabatic potentials @xmath142 were calculated . as a result ,
the density of resonance curves was much higher ( factor 20 - 40 ) .
the main contribution to the high density of resonance curves was from the states with small principal quantum number corresponding to highly excited ionic channels .
this increased significantly the probability of predissociation .
in fact , in the case of hco , the vibrating co bond should also increase the density of steep resonance curves in figs .
[ resonances_m1 ] and [ resonances_m1_2 ] .
this is because if the vibration along co is quantized , each of the curves shown in fig .
[ energies ] will produce a series additional ionic curves with different co quanta .
therefore , the number of ionic potential will be larger and , correspondingly , the density of resonance curves and avoided crossings in figs .
[ resonances_m1 ] and [ resonances_m1_2 ] will be increased too .
this should increase the dr cross section ( more states to be captured to ) and the ratio @xmath143 ( more avoided crossings ) .
another approximation is that the @xmath144 rydberg states have dissociative character and are poorly described by this model .
however , we believe that the dominant dr pathways are triggered by an initial capture into higher rydberg states , rather than being directly captured into @xmath144 states .
in the previous theoretical study of dr in hco@xmath105 published by some of us,@xcite the renner - teller coupling was not taken into account because it appeared to have a small effect on the potential surfaces of excited molecular states of the hco molecule .
however , renner - teller coupling involves degenerate @xmath145 states and therefore should play an important role in electron - ion scattering especially when @xmath84 and @xmath85 molecular potentials approach closely to each other .
although renner - teller coupling was not included in the previous study , some nonadiabatic effects were considered there .
the quantum defects @xmath146 for @xmath147 states obtained from _ ab initio _ calculation demonstrate sharp variations at some configurations @xmath148 of the three hco nuclei .
therefore , instead of using the three adiabatic defects @xmath146 , a numerical diabatization procedure was applied to give a @xmath148-dependent 3x3 matrix @xmath149 of quantum defects .
the @xmath148 dependence of the matrix is weaker and the matrix has more information about nonadiabatic effects than the diagonal matrix @xmath146 .
on the other hand the nonadiabatic effects represented by non - diagonal elements of @xmath146 might have a smaller effect on the dynamics of the electron comparing with the nonadiabatic couplings involving degenerate @xmath145 states .
indeed , the dr cross section obtained in the previous study @xcite is much smaller than the cross section of the present treatment . to verify that the increase in the dr cross section is indeed due to the renner - teller coupling
, we artificially set the non - diagonal elements in the hamiltonian of eq .
[ eq : hint ] to zero and calculated again the cross section for @xmath150 .
the resulting dr cross section is about a factor of three smaller than the cross section shown in fig .
[ cs ] and on average it is close to the result of the previous study .
it is worth mentioning that the calculation of ref .
@xcite accounts for all three vibrational degrees of freedom but the very high vibrational levels that generate the steep rydberg @xmath144 curves in fig .
[ resonances_m1 ] ( which have large widths and therefore lead to an increase in the capture probability ) were omitted ; these two effects would tend to balance each other , and might give roughly the same cross section as in the present study when off - diagonal couplings are switched off .
we speculate that if the co bond vibration was included in the present treatment the effect of the renner - teller effect would likely be even more pronounced . to summarize our discussion of this point
, we conclude that the renner - teller effect appears to be the most important nonadiabatic coupling in the dissociative recombination of hco@xmath105 .
the theoretical cross section obtained in this study is inversely - proportional to the incident energy below @xmath151 ev ( see fig .
[ cs ] ) . above 0.1 ev
the cross section falls more rapidly with increasing energy .
this is caused by the fact that at 0.1 ev an additional ionic channel becomes open .
that ionic channel is responsible for the large probability to capture the electron . above that ionization threshold , the density of rydberg states ( in which capture can occur ) in much lower , as can be seen from fig .
[ resonances_m1 ] , and therefore the capture probability drops . moreover , resonance states above that ionization threshold have another competing decay channel for autoionization , which steals flux away from the dr observable . for the ground
vibrational level @xmath152 ( dot - dashed line in fig .
[ cs ] ) , that additional open channel is @xmath153 , for the first excited level there are two nearby levels : @xmath154 and @xmath155 . a similar behavior of the cross section was observed in dr of h@xmath141 @xcite : the cross section drops down once a new ionic channel becomes open .
the cross section from the previous theoretical study of hco@xmath105 ( the solid line in fig .
[ cs ] ) does not show that behavior .
it behaves smoothly as the energy crosses the new ionization threshold .
this is because the renner - teller effect was not accounted for : hence transitions of the type @xmath156 are forbidden by the different symmetries of those vibrational states .
another difference with the previous study is the manner in which the final cross section is calculated . in ref .
@xcite , the scattering matrix @xmath157 for electron - ion collisions was explicitly calculated , but only in the electron - ion subset of channel space with no channel indices explicitly referring to dissociation .
all of the channels @xmath158 are vibronic states . however , some of the channels @xmath158 are open for dissociation , and as a result , the @xmath157 matrix is not completely unitary .
the defect from unitarity was used to calculate the dissociative flux and the corresponding dr cross section .
the resulting cross section from this method has a rich structure with many rydberg series of resonances .
the cross section was then averaged over an appropriate thermal distribution of electron energies .
the averaged cross section still has a number of resonances that are not seen in the comparatively low resolution experiment . in the present paper
we have employed a different averaging procedure , as was discussed above .
this procedure gives a very smooth curve for the cross section .
in summary , we would like to stress the following : * we have developed a theoretical approach to obtain the cross section for dissociative recombination of hco@xmath105 .
the approach is based on a model reaction matrix @xmath67 that includes the non - born - oppenheimer renner - teller coupling between doubly - degenerate vibrational modes of hco@xmath105 and the degenerate continuum states of the incident electron .
the structure of the matrix @xmath67 is derived directly from the hamiltonian that accounts for the renner - teller physics , with elements of the constructed @xmath67-matrix , if diagonalized , that reproduce the quantum defects obtained from _ ab initio _ calculations . *
we have shown that even a modest renner - teller splitting can cause a greatly enhanced dr rate .
the previous theory without the renner - teller coupling gave a lower dr cross section .
the present dr cross sections are still generally about a factor of 2 lower than the experimental cross sections . * the vibration along the co bond was not included in the present model .
we believe that the inclusion of the co bond should increase the density of hco states available to be populated during the electron - ion collision .
this will probably increase the dr cross section , but the quantitative amount remains to be determined by future studies .
another error expected to be of order 70% is introduced by our use of an adiabatic approximation in the @xmath25 coordinate , and this should also be improved in future calculations .
* in the present treatment the dissociation occurs predominantly through the indirect pathway , i.e. through rydberg states of hco .
the estimated autoionization probability is comparable with the predissociation probability , but it seems likely that inclusion of co vibrations may increase the probability of predissociation . * the approach is quite general and can be applied to other closed - shell linear molecular ions , such as hcs@xmath105 .
this work has been supported by the national science foundation under grant no .
phy-0427460 and grant no .
phy-0427376 , by an allocation of nersc supercomputing resources and by an allocation of ncsa supercomputing resources ( project no .
phy040022 ) .
l. acknowledges support from the swedish research council and the gran gustafsson foundation .
a. p. hickman , j.phys .
b : at . mol .
phys . * 20 * , 2091 ( 1987 ) .
a. giusti , j. phys .
b * 13 * , 3867 ( 1980 ) .
a. giusti - suzor , j. n. bardsley , c. derkits , phys .
a * 28 * , 682 ( 1983 ) .
a. giusti - suzor , ch .
jungen , j. chem . phys . * 80 * , 986 ( 1984 ) . k. nakashima , h. takagi , and h. nakamura , j. chem .
phys . * 86 * , 726 ( 1987 ) . c. jungen , s. c. ross , phys .
a * 55 * , r2503 ( 1997 ) .
i. f. schneider , c. stromholm , l. carata , x. urbain , m. larsson , a. suzor - weiner , j. phys . b. : at . mol .
30 * , 2687 ( 1997 ) .
a. e. orel and k. c. kulander , phys .
lett . * 71 * , 4315 ( 1993 ) .
v. kokoouline , c. h. greene , b.d .
esry , nature * 412 * , 891 ( 2001 ) .
a. le padellec , c. sheehan , d. talbi , and j. b. a. mitchell , j. phys .
b : at . mol .
30 * , 319 ( 1997 ) .
v. poterya , j. l. mclain , n. g. adams , and l. m. babcock , j. phys . chem . * 109 * , 7181 ( 2005 ) . | a theoretical description of the dissociative recombination process for the hco@xmath0 ion suggests that the nonadiabatic renner - teller coupling between electronic and vibrational degrees of freedom plays an important role .
this finding is consistent with a recent study of this process for another closed - shell molecule , the h@xmath1 ion , where jahn - teller coupling was shown to generate a relatively high rate .
the cross section obtained here for the dissociative recombination of hco@xmath0 exhibits encouraging agreement with a merged - beam experiment . |
since its discovery in 1995 at the tevatron@xcite , the top quark remains the heaviest known fundamental particle to date . with a mass of 172.6 @xmath1 1.4 gev / c@xmath2@xcite , it is considered to be intimately connected with the mechanism of electroweak symmetry breaking in the standard model of elementary particle physics ( sm ) and also to be sensitive to physics beyond the framework of the sm .
this article reports recent measurements by the cdf and d collaborations that probe the sm expectations for deviations both in the production and decay of the top quark . after a brief outline of the top quark properties within the sm framework in section [ sec : smtop ] ,
the current status of measured top quark pair production rates is given section [ sec : xsec ] , followed by a section on searches for top quark production beyond the sm .
the subsequent three sections describe measurements probing the top quark decay in terms of branching fractions , search for flavour - changing neutral currents and the helicity of the @xmath3 boson in the top quark decay , respectively .
a conclusion is given in the final section [ sec : conclusion ] .
within the framework of the sm , top quark production at the tevatron proceeds mainly in pairs : @xmath4 via the strong interaction ( 85% @xmath5 annihilation and 15% gluon - gluon fusion ) .
the corresponding production cross section has been evaluated at next - to - leading order ( nlo ) qcd using two different approaches : one calculation considers soft gluon corrections up to next - to - next - to - next - to leading logarithmic ( nnnll ) terms and some virtual terms in a truncated resummation , yielding 6.77 @xmath1 0.42 pb for a top quark mass of 175 gev / c@xmath2@xcite , while another calculation using the nlo calculation with ll and nll resummation at all orders of perturbation theory gives 6.70@xmath6 pb for a top quark mass of 175 gev / c@xmath2@xcite . if a pdf uncertainty is combined linearly with the theoretical uncertainty for the first result similar to what is done for the second result both predictions exhibit not only similar central values but also similar relative uncertainties of @xmath712 - 13% . due to its large mass
, the top quark has an extremely short lifetime of approximately , which makes it decay before it can form hadrons a unique feature setting it apart from all other quarks .
since the top quark mass is well above the threshold for @xmath8 decays with @xmath9 being one of the down - type quarks @xmath10 , this two - body decay dominates the top quark decay .
as each quark flavour contributes to the total decay rate proportional to the square of the respective ckm matrix element @xmath11 , top decays into @xmath12 and @xmath13 are strongly suppressed with respect to the dominant decay @xmath14 .
consequently , top quark pair events contain a @xmath15 and a @xmath16 quark from the @xmath17 decay , and depending on the decay modes of the two @xmath3 bosons , the observed top quark pair final states can be divided into three event classes : 1 . in _ dilepton _ events , both @xmath3 bosons decay leptonically , resulting in a final state containing two isolated high-@xmath18 leptons , missing transverse energy corresponding to the two neutrinos and two jets .
this final state constitutes @xmath75% of the @xmath17 events ( not counting @xmath19 leptons ) and gives the cleanest signal but suffers from low statistics .
lepton+jets _ events , one @xmath3 boson decays leptonically , the other one hadronically , resulting in one isolated high-@xmath18 lepton , and four jets .
events in the @xmath20+jets or @xmath21+jets channels yield @xmath729% of the branching fraction ( @xmath734% when including leptonic @xmath19 decays ) and provide the best compromise between sample purity and statistics .
3 . in _ all - hadronic _ events , both @xmath3 bosons decay to @xmath22 pairs , resulting in a six - jet final state . with a branching fraction of @xmath746% ,
this final state represents the biggest fraction of @xmath17 events , but it is also difficult to separate from the large background of multijet production . all of these final states contain two @xmath15-jets from the hadronisation of the ( anti- ) @xmath15 quarks , and additional jets can arise from initial and final state radiation .
top quark pair production cross section measurements provide a unique test of the predictions from perturbative qcd calculations at high transverse momenta . analysing all three event classes
allows both the improvement of statistics of top events and studies of properties and important checks for physics beyond the sm that might result in enhancement / depletion in some particular channel via novel production mechanisms or decay modes .
the following subsections give an overview of the cross section measurements pursued at the tevatron rather than quoting single cross section results , with the exception of the most precise single measurement to date , obtained by d in the lepton+jets channel .
all current measurements are summarised in figure [ fig : xsecsummary ] .
a typical selection of dilepton events requires two isolated high @xmath18 leptons , and at least two central energetic jets in an event .
the most important physics background processes containing both real leptons and are z/@xmath23+jets production with @xmath24 and the production of dibosons ( @xmath25 ) .
instrumental backgrounds are to be considered as well , arising from misreconstructed due to resolution effects in z/@xmath23+jets production with @xmath26 , and also from @xmath3+jets and qcd multijet production where one or more jets fake the isolated lepton signature .
to ensure proper description of the instrumental backgrounds , these are usually modelled using collider data , while for the physics backgrounds typically monte carlo simulation is used .
a further enhancement of the signal fraction in the selected data samples is possible by requiring additional kinematical event properties like the scalar sum of the jet @xmath18s @xmath27 to be above a certain threshold or rejecting events where both selected leptons have like - sign electric charge .
the obtained purities in such selected samples are usually quite good with a signal to background ratio ( s / b ) better than 2 at least , although signal statistics are low .
the acceptance for dilepton final states can be enhanced by loosening the selection to require only one fully reconstructed isolated lepton ( @xmath28 ) in addition to an isolated track ( `` @xmath29+track analysis '' ) . in particular , such a selection allows the inclusion of `` 1 prong '' hadronic @xmath19 decays .
the top quark production cross section was recently measured for the first time also in the lepton+tau final state by d@xcite , using events with hadronically decaying isolated taus and one isolated high @xmath18 electron or muon . to separate real taus from jets ,
a neural network was used , and the sample purity was enhanced by requiring @xmath15-jet identification ( see section [ lpjxsec ] ) in the selected events .
the result is shown together with the other measurements in figure [ fig : xsecsummary ] .
-tagged jets ( right)@xcite.__,title="fig : " ] -tagged jets ( right)@xcite.__,title="fig : " ] a typical lepton+jets selection requires one isolated high @xmath18 lepton ( @xmath20 or @xmath21 which includes @xmath30 , @xmath31 ) , and at least 4 jets , yielding samples with a s / b around 1/2 .
the dominant physics background to be considered here comes from @xmath3+jets production while the main instrumental background arises from qcd multijet production where a jet fakes the isolated lepton signature .
the cross section can be extracted from such a selected sample either purely based on topological and kinematical event properties combined in a multivariate discriminant to separate the @xmath17 signal from background or by adding identification of @xmath15-jets . since topological analyses do not depend on the assumption of 100% branching of @xmath32 , they are less model - dependent than tagging analyses . on the other hand ,
requiring @xmath15-jet identification is a very powerful tool in suppressing the background processes , which typically exhibit little heavy flavour content . with @xmath15-jet identification
, the top signal can also be easily extracted from lower jet multiplicities , where topological analyses need to impose additional selection criteria like cutting on @xmath33 to be able to extract the signal .
in addition , @xmath15-tagged analyses can provide very pure signal samples , easily exceeding a s / b @xmath34 10 in selections requiring at least four jets with two identified b - jets ( see for example figure [ fig : ljetsxsec ] ) .
the identification of @xmath15-jets can be based on the long _ lifetime _ of b hadrons resulting in significantly displaced secondary vertices with respect to the primary event vertex or large significant impact parameters of the corresponding tracks .
a combination of this type of information in a neural network tagging algorithm yields @xmath15-tagging efficiencies of about 54% while only about 1% of light quark jets are misidentified as @xmath15-jets hence the improved s / b in tagged analyses .
another way to identify @xmath15-jets is to reconstruct _ soft leptons _ inside a jet originating from semileptonic
so far only soft-@xmath21 tagging has been deployed in @xmath17 analyses .
the most precise @xmath17 production cross section measurement to date with a relative uncertainty of 11% has been performed by d on 0.9 fb@xmath0 of data in the lepton+jets channel@xcite . for this measurement , two complementary analyses based on a kinematic likelihood discriminant and on @xmath15-tagging ( see figure [ fig : ljetsxsec ] ) were combined and yield @xmath35 for a top quark mass of 175 gev / c@xmath2 . comparing this measurement with the theory prediction
, the top quark mass can be extracted as well , yielding @xmath36 gev / c@xmath37 in good agreement with the world average .
a first @xmath19+jets cross section analysis using events with hadronically decaying isolated taus and lifetime @xmath15-tagging was performed as well by d the result is shown together with other measurements in figure [ fig : xsecsummary ] .
the all - hadronic final state is studied by requiring events with at least six central energetic jets and no isolated high @xmath18 leptons . due to the overwhelming background from qcd multijet production with a cross section orders of magnitude above that of the signal process
, @xmath15-jet identification is mandatory for this final state .
further separation of signal and background is achieved by using multivariate discriminants based on topological and kinematical event properties .
figure [ fig : xsecsummary ] provides an overview of recent cross section measurements performed by cdf and d .
all measurements show good agreement with the sm prediction and with each other .
the single best measurements are approaching a relative precision of @xmath38 = 10% that should be achievable for the datasets of 2 @xmath39 already at hand and provide stringent tests to theory predictions.with increasing datasets , these measurements naturally start to become limited by systematic uncertainties rather than statistical ones , but it will be possible to further constrain the systematic uncertainties as well using additional data .
cross section measurements form the foundation for all further property analyses like the ones described in the subsequent sections of this article by characterising the datasets enriched in top quark pairs and providing the necessary understanding of object identification , background modelling and sample composition .
various beyond the sm theories predict the existence of a massive @xmath40-like boson that could decay into @xmath17 and hence add a resonant production mode to the sm process .
any such additional production would be visible in the @xmath17 invariant mass distribution provided the resonance @xmath41 decaying to @xmath17 is sufficiently heavy and narrow .
invariant mass distribution in lepton+jets data with four or more jets .
right : expected and observed 95% c.l .
upper limits on @xmath42@xcite.__,title="fig : " ] invariant mass distribution in lepton+jets data with four or more jets .
right : expected and observed 95% c.l .
upper limits on @xmath42@xcite.__,title="fig : " ] both cdf and d perform a search for a generic heavy resonance @xmath41 of narrow width ( @xmath43 ) compared to the detector mass resolution in @xmath15-tagged lepton+jets datasets . the @xmath17 invariant mass spectrum is reconstructed using either a kinematic fit to the @xmath17 production hypothesis ( cdf ) or directly from the four - momenta of the up to four leading jets , the lepton and the neutrino momentum ( d ) .
the latter approach was shown to provide better sensitivity for large resonance masses than the previously used constrained kinematic fit and also allows the inclusion of data with fewer than four jets in case that jets merged . as both experiments observe no significant deviation from the sm expectation , 95% c.l .
upper limits on @xmath42 are given for values of @xmath44 between 450 and 900 gev / c@xmath2 ( cdf ) respectively 350 and 1000 gev / c@xmath2 ( d , see figure [ fig : ttresonance ] ) .
both experiments provide 95% c.l .
mass limits for a leptophobic top - colour - assisted technicolour @xmath45 boson as a benchmark model . using 955 pb@xmath0 ,
cdf finds @xmath46 720 gev / c@xmath2 ( expected limit : 710 gev / c@xmath2)@xcite while d finds @xmath46 760 gev / c@xmath2 ( expected limit : 795 gev / c@xmath2)@xcite using 2.1 fb@xmath0 of data . instead of a new colour singlet particle decaying into @xmath17 as described in the previous subsection , there could also be a new massive colour octet particle @xmath47 contributing to @xmath17 production .
such a `` massive gluon '' production mode would interfere with the corresponding sm production process . assuming a sm top decay , cdf has performed a search for a corresponding contribution by comparing the @xmath17 invariant mass distribution in a 1.9 fb@xmath0 @xmath15-tagged lepton+jets dataset with the sm expectation .
as the largest discrepancy with respect to the sm observed is 1.7@xmath48 for the explored mass and width range 400 gev / c@xmath2 @xmath49 800 gev / c@xmath2 , 0.05 @xmath50 0.5 , upper and lower limits are provided on the corresponding coupling strengths of the massive gluon@xcite .
invariant mass distribution in lepton+jets data with at least four jets .
right : corresponding observed @xmath17 differential cross section , compared to the sm expectation@xcite.__,title="fig:",height=170 ] invariant mass distribution in lepton+jets data with at least four jets .
right : corresponding observed @xmath17 differential cross section , compared to the sm expectation@xcite.__,title="fig:",height=170 ] contributions beyond the sm in @xmath17 production could manifest themselves in either resonances , broad enhancements or more general shape distortions of the @xmath17 invariant mass spectrum . a very generic way to search for such effects
is to measure the @xmath17 differential cross section @xmath51 and compare the shape with the sm expectation .
cdf reconstructs the @xmath17 invariant mass spectrum in a 1.9 fb@xmath0 @xmath15-tagged lepton+jets dataset ( see figure [ fig : diffxsec ] ) by combining the four - vectors of the four leading jets , lepton and missing transverse energy . after subtracting the background processes , the distortions in the reconstructed distribution due to detector effects , object resolutions and geometric / kinematic acceptance
are corrected for by the application of a regularised unfolding technique . from the unfolded distribution , the @xmath17 differential cross section @xmath51
is extracted and its shape is compared with the sm expectation . the shape comparison yields good agreement with the sm , yielding an anderson - darling p - value of 0.45@xcite .
@xmath15-tags for @xmath17 events with @xmath53 jets as a function of @xmath54 .
right : 68% and 95% c.l .
statistical uncertainty contours in the r vs. @xmath55 plane@xcite.__,title="fig : " ] @xmath15-tags for @xmath17 events with @xmath53 jets as a function of @xmath54 .
right : 68% and 95% c.l .
statistical uncertainty contours in the r vs. @xmath55 plane@xcite.__,title="fig : " ] assuming the validity of the sm , specifically the existence of three fermion generations , unitarity of the ckm matrix and insignificance of non-@xmath3 boson decays of the top quark ( see section [ sec : fcncdecay ] ) , the ratio of branching fractions @xmath54
= @xmath56 ) simplifies to @xmath57 , and hence is strongly constrained : @xmath58 at 90% c.l.@xcite .
deviations of @xmath54 from unity could for example be caused by the existence of a fourth heavy quark generation .
the most precise measurement of @xmath54 thus far has been performed by d in the lepton+jets channel using data corresponding to an integrated luminosity of 900 pb@xmath0 . by comparing the event yields with 0 , 1 and 2 or more @xmath15-tagged jets and using a topological discriminant to separate the @xmath17 signal from background in events with 0 @xmath15-tags
, @xmath54 can be extracted together with the @xmath17 production cross section @xmath55 simultaneously ( see figure [ fig : d0rmeas ] ) .
this measurement allows the extraction of @xmath55 without assuming @xmath59 , yielding @xmath60 ( stat+syst ) and @xmath61 ( stat+syst ) @xmath1 0.50 ( lumi ) pb for a top quark mass of @xmath62 gev / c@xmath37 in agreement with the sm prediction@xcite .
the occurrence of flavour changing neutral currents ( fcnc ) a decay of type @xmath63 with @xmath64 and @xmath65 is strongly suppressed in the sm and expected to occur at a rate below @xmath66 , well out of reach of being observed at the tevatron .
consequently , any observation of fcnc decays would signal physics beyond the sm .
cdf has performed a search for @xmath67 in a 1.9 fb@xmath0 dataset of @xmath68 jets events with @xmath69 , assuming a sm decay of the second top quark @xmath70 .
since the event signature does not contain any neutrinos , the events can be fully reconstructed .
the best discriminant found to separate signal from background processes is a mass @xmath71 variable that combines the kinematic constraints present in fcnc decays .
the signal fraction in the selected dataset is determined via a template fit in signal samples with 0 or @xmath721 @xmath15-tags and a background - enriched control sample to constrain uncertainties on the background shape and normalisation ( see figure [ fig : cdffcnc ] ) . since the observed distributions are consistent with the sm background processes , a 95% c.l .
upper limit on the branching fraction @xmath73 of 3.7% is derived@xcite .
this is the best limit on @xmath74 to date .
boson helicity fits ( left : d@xcite , right : cdf@xcite ) .
the ellipses show the 68% and 95% c.l .
contours around the measured data points .
the stars show the sm expectation ; the triangle denotes the physically allowed region where @xmath76 and @xmath77 sum to one or less.__,title="fig:",height=188 ] boson helicity fits ( left : d@xcite , right : cdf@xcite ) .
the ellipses show the 68% and 95% c.l .
contours around the measured data points .
the stars show the sm expectation ; the triangle denotes the physically allowed region where @xmath76 and @xmath77 sum to one or less.__,title="fig:",height=188 ] assuming a massless @xmath15-quark , the top quark decay in the v@xmath78a charged current weak interaction proceeds only via a left - handed ( @xmath79 30% ) and a longitudinal ( @xmath80 70% ) fraction of @xmath3 boson helicities .
the helicity of the @xmath3 boson is reflected in the angular distribution @xmath81 of its decay products , with @xmath82 being the angle of the down - type decay products of the @xmath3 boson ( charged lepton respectively @xmath83- or @xmath84-quark ) in the @xmath3 boson rest frame with respect to the top quark direction . any observed right - handed fraction @xmath85 would indicate physics beyond the sm .
cdf has measured the @xmath3 boson helicity fractions in 1.9 fb@xmath0 of @xmath15-tagged lepton+jets data comparing the @xmath81 distribution of leptons in data to templates for longitudinal , right- and left - handed signal plus background templates . when fitting both @xmath86 and @xmath87 simultaneously , the result is @xmath88 and @xmath89@xcite .
d has measured the @xmath3 boson helicity fractions using the @xmath81 distributions in dilepton and lepton+jets events including their hadronic @xmath3 boson decays in 1 fb@xmath0 of data , yielding @xmath90 and @xmath91@xcite .
both measurements agree with the sm at the 1@xmath48 level ( see figure [ fig : whelmeas ] ) .
a wealth of top quark analyses is being pursued at the tevatron , probing the validity of the sm with unprecedented precision .
the measured top quark pair production rates are found to be consistent with the sm expectation across the decay channels , with the most precise measurements surpassing the precision of theory predictions .
there is no evidence thus far for contributions beyond the sm in either top quark production or top quark decay .
however , with some measurements still being statistically limited , there is still room for surprises .
more detailed descriptions of the analyses presented here and many more interesting top quark physics results can be found online@xcite . continuously improving analysis methods and using the increasing integrated luminosity from a smoothly running tevatron that is expected to deliver more than 6 fb@xmath0 by the end of run ii , we are moving towards more precision measurements and hopefully discoveries within and outside the sm .
the author would like to thank the organisers for creating a very fruitful collaborative atmosphere at the rencontres de physique de la valle daoste , the cdf and d collaborations , the staffs at fermilab and collaborating institutions and also the alexander von humboldt foundation for their support .
99 cdf collaboration , f. abe _
et al . _ ,
lett . * 74 * , 2626 ( 1995 ) ; + d collaboration , s. abachi _
et al . _ ,
lett . * 74 * , 2632 ( 1995 ) .
tevatron electroweak working group for the cdf and d collaborations , fermilab - tm-2403-e ( 2008 ) , arxiv:0803.1683 .
n. kidonakis and r. vogt , phys .
d * 68 * , 114014 ( 2003 ) .
m. cacciari , s. frixione , m. l. mangano , p. nason , and g. ridolfi , jhep * 04 * , 068 ( 2004 ) .
d collaboration , v. m. abazov _
et al . _ , d conference note 5451 ( 2007 ) .
d collaboration , v. m. abazov _
, fermilab - pub-08 - 064-e ( 2008 ) , submitted to phys .
cdf collaboration , t. aaltonen _
, fermilab - pub-07 - 576-e ( 2007 ) , submitted to phys .
d , arxiv:0710.5335 .
d collaboration , v. m. abazov _
_ , d conference note 5600 ( 2008 ) .
cdf collaboration , t. aaltonen _
et al . _ , cdf conference note 9164 ( 2008 ) .
cdf collaboration , t. aaltonen _
et al . _ , cdf conference note 9157 ( 2008 ) . s. eidelman _
et al _ , phys .
b * 592 * , 1 ( 2004 ) .
d collaboration , v. m. abazov _
, fermilab - pub-08 - 010-e ( 2008 ) , submitted to phys .
, arxiv:0801.1326 .
cdf collaboration , t. aaltonen _
et al . _ , cdf conference note 9202 ( 2008 ) .
cdf collaboration , t. aaltonen _
et al . _ , cdf conference note 9114 ( 2007 ) .
d collaboration , v. m. abazov _
et al . _ ,
lett . * 100 * 062004 ( 2008 ) . ` http://www-d0.fnal.gov/run2physics/top/index.html ` ; ` http://www-cdf.fnal.gov/physics/new/top/top.html ` | the tevatron proton - antiproton collider at fermilab with its centre of mass energy of 1.96 tev is currently the only source for the production of top quarks .
this report reflects the current status of measurements of the top quark pair production cross section and properties performed by the cdf and d collaborations .
utilising datasets of up to two fb@xmath0 , these measurements allow unprecedented precision in probing the validity of the standard model .
= 11.6pt |
atomic - scale simulations have the capability to predict the properties of defect structures that are often inaccessible by experimental techniques.@xcite these predictions require accurate and efficient calculations of energies and forces on atoms in arrangements that sample a variety of atomic environments , and may represent even different binding configurations .
accurate quantum mechanical methods are difficult to scale to large systems and long simulation times , while empirical interatomic potentials offer increased computational efficiency at a lower level of accuracy .
maximizing the efficiency of computational material science studies requires the development of potentials that are transferrable , i.e. , capable of predicting properties outside their fitting range , and accurate for static and dynamic calculations
. however , without direct transferable derivations of interatomic potentials from quantum mechanical methods , empirical interatomic potentials require high - dimensional non - linear fitting .
many functional forms for empirical potentials have been proposed , including embedded - atom method ( eam)@xcite , modified embedded - atom method ( meam)@xcite and charged - optimized many - body potential ( comb).@xcite there have been multiple implementations of different potential functional forms for various materials.@xcite even for the same type of materials , such as cu@xcite and si@xcite , different empirical interatomic potential models are proposed for different applications with different transferabilities .
there are advanced techniques to optimize the potential parameters based on a weighted least - squares regression to a fitting database of experimental or quantum mechanical calculation data,@xcite including the force - matching method@xcite for empirical interatomic potential parameter optimization . in force - matching ,
a fitting database includes quantum mechanical force calculations for diverse atomic environments to obtain realistic empirical potential models . to study the transferability of the empirical potential model , frederiksen _ et al .
_ applied bayesian statistics to empirical interatomic potential models : instead of using the best fit , an ensemble of neighboring parameter sets reveal the flexibility of the model.@xcite they showed that the standard deviation of the potential prediction of structure property function is a good estimate of the true error . however , even with these advances , the determination of empirical interatomic potentials relies on the selection and weighting of a fitting database without a clear , quantitative guide for the impact on predictions . to address the issue of fitting database selection , we present an automated , quantitative fitting - database optimization algorithm based on prediction errors for a testing set using bayesian statistics . we construct an objective function of the prediction errors in the testing set to optimize the relative weights of a fitting database .
this includes the addition or removal of structures to a fitting database when weights change sign .
we demonstrate the viability of the optimization algorithm with a simple interatomic potential model : lennard - jones potential fitting of ti crystal structures .
we choose this example as a radial potential has difficulty describing the stability of different crystal structures of a transition metal .
the new algorithm also helps to understand the transferability of the empirical potential model for the structures in the testing set .
we start with a brief review of empirical potential models and parameter optimization using a fitting database .
next , we discuss bayesian error estimation as it applies to our problem .
then we define an objective function with a testing set , and use this quantitative measure to devise an algorithm to optimize a fitting database .
lastly , we demonstrate this new approach on an example system with clear limitations : lennard - jones potential for titanium .
the total energy and forces for the structure of interest are the most basic quantities to calculate since they determine the structural properties . in particular , we are interested in predictions that are derived from energies of atomic arrangements . in atomic - scale simulations ,
a structure @xmath0 is a set of atomic positions @xmath1 with chemical identities @xmath2 : @xmath3 .
the total energy of a structure @xmath0 is @xmath4 with forces @xmath5 .
density - functional theory ( dft)@xcite calculations can provide accurate structural energies and forces , but their computational cost limits them to simulation system with at most a few thousand atoms . other structural properties
are derived from energies and forces , and so without loss of generality , we develop our approach based on energies and forces .
parameterized empirical interatomic potentials offer a computationally efficient alternative to dft .
potentials provide approximate energies and forces for atomic configurations that are inaccessibly large for dft calculations .
generally , an empirical interatomic potential functional can be written as @xmath6 where @xmath7 are parameters , and @xmath8 is an interatomic potential function between @xmath9 atoms of chemical identity @xmath10 .
symmetries of the potential functional form , such as permutation symmetry , translational symmetry , rotational symmetry , etc .
, can simplify the functional form . a general empirical interatomic potential that reproduces all dft energy calculations
accurately is computational intractable , since it would require a large number of many - body terms .
rather , we are interested in simpler potentials that provide accurate results for a smaller range of atomic configurations including perfect crystals and defect structures under various thermodynamic conditions ; this includes potentials that may not be easily written in the form of eqn .
( [ eqn : pot ] ) , such as eam and meam potentials .
the optimal potential parameters @xmath7 derive from comparison to predictions of a database of dft calculations , and the performance of the potential is evaluated by a testing set of structure property predictions .
a fitting database @xmath11 is a set @xmath12 of structure property functions @xmath13 with an associated structure @xmath0 and positive ( relative ) weight @xmath14 .
while a single structure will often have multiple structure property functions with unique weights for fitting , we simplify our notation by indexing on the structure ; in what follows , sums over structures @xmath0 indicate sums over all members of the database @xmath11 .
the structure property function @xmath13 may be a scalar such as the energy ( relative to a reference structure ) , vectors such as forces on the atom of the structures , stress tensors and more complicated structure property functions such as lattice constant , bulk modulus or vacancy formation energy or anything that can be defined from the energy @xmath15 . in the fitting database
, we will compare the structure property functions evaluated using an empirical potential , with the values from dft , though other choices are possible , such as experimental data . in the weighted least - squares ( described later )
, we impose the trivial constraint @xmath16 , as only relative values of @xmath17 are important .
a testing set @xmath18 is a set @xmath19 of structures @xmath20 with structure property functions @xmath21 . in a testing set
, we will compare structure property functions evaluated using an empirical potential with _ either _ values from dft , or using bayesian sampling of the empirical potential , following frederiksen _
et al._,@xcite which we will discuss in section [ sec : bayes ] .
there are no relative weights for structures in a testing set ; rather , these represent a set of predictions whose errors we will evaluate . in order to assess the prediction errors of the structure property functions , we define the prediction error function @xmath22 as @xmath23 where @xmath24 is the structure property function from the empirical atomic potential with parameters @xmath7 , @xmath13 is the structure property function from dft , and @xmath25 denotes the 2-norm of a @xmath26-dimensional vector @xmath27 @xmath28 we will take the error evaluation of the energy differences between two structures and forces as examples . for energy calculations , the structure property function @xmath13 is @xmath29 where @xmath30 is the energy of a reference structure , @xmath31 .
the potential energy prediction error is @xmath32 the force predictions errors are @xmath33 then , the weighted summed squared error function for a fitting database @xmath11 is @xmath34
we introduce bayesian sampling to estimate the errors of structure property function predictions and quantitatively analyse the relative weight values in the fitting database .
given a fitting database , we calculate the prediction of structure property function @xmath35 and the error @xmath36 of the structure property function .
we then derive the analytical expression of the gradient of the bayesian errors with respect to the weights , @xmath37 .
these gradients provide quantitative information on how structure property functions in the fitting database influence the bayesian predictions of structure property functions in the testing set though weight change .
bayesian statistics treats model parameters as random variables with a probability distribution given by a posterior distribution.@xcite according to the bayes theorem , the posterior distribution of the parameters is a product of the prior distribution @xmath38 and the likelihood function @xmath39 , @xmath40 where the prior distribution @xmath38 includes the information about the potential model before the we take the fitting data into account .
here we use the maximally unbiased prior distribution of a uniform distribution over a measurable set @xmath41 of allowed parameters sets , @xmath42^{-1},\ ] ] though other choices are possible . assuming the errors are independent and identically normally distributed , the likelihood function is@xcite @xmath43 where @xmath44
the log - likelihood function is proportional to the squared error function , @xmath45 .
@xmath46 since the logarithm is a monotonically increasing function , minimizing @xmath45 is equivalent to maximizing the log - likelihood function .
the maximum likelihood estimate ( mle ) of the parameters @xmath47 is a function of the fitting database @xmath11 , and @xmath48 .
the bayesian prediction of a function @xmath49 is the mean @xmath50 all the averages are implicit functions of the relative weights in the fitting database .
the bayesian error is the mean squared error of the bayesian prediction : @xmath51 where @xmath52 is the variance of the bayesian prediction .
the covariance of two functions @xmath24 and @xmath53 represents the correlation between two functions : @xmath54 the derivative of a bayesian prediction with respect to weight is @xmath55 note that @xmath56 the derivative of @xmath57 with respect to weight is found using the chain rule , @xmath58 as @xmath47 is an extremum of @xmath59 . applying eqn .
( [ eqn : deriv1])eqn .
( [ eqn : deriv3 ] ) to the bayesian error @xmath36 yields @xmath60 \\ & = -\frac{1}{w}\left [ c^f_{\alpha\beta}-\frac{\epsilon_{\alpha}^2({\theta^\text{mle}})}{w}\sum_\gamma w_\gamma c^f_{\beta\gamma}\right ] , \end{split } \label{eqn : dedw}\ ] ] where @xmath61 we define the error of a structure property function in a testing set without dft calculations by approximating the unknown dft calculations of the structure property function with its bayesian prediction . based on eqn .
( [ eqn : error ] ) , @xmath62 is approximated by @xmath63 , and so a testing set can include structures _ in the absense of _ dft calculations .
we need to evaluate the integral in eqn .
( [ eqn : mean ] ) to calculate the bayesian predictions and bayesian error estimation . for complicated high - dimensional
, non - linear models such as empirical potentials , the integral can not be evaluated in closed form , and the high - dimensionality makes direct numerical quadrature converge slowly .
we instead use markov chain monte carlo ( mcmc ) to numerically integrate .
the chain of @xmath64 will contain a set of @xmath65 independent samples @xmath66 ( where @xmath67 is the autocorrelation length ) , and the numerical estimate of the mean is @xmath68 with a sampling error of @xmath69 .
hence , once fitting is complete , the `` best '' set of parameters @xmath47 defines the empirical potential for predictions , while the ensemble of parameters @xmath66 allows the estimation of errors on those predictions .
we define an optimal fitting database based on bayesian errors in the testing set .
an empirical potential model should reproduce dft calculations for a set of atomic environments described by structures in a testing set , and so the bayesian errors of structure property functions in the testing set are the quantities of interest . because different types of structure property functions have different units , different error magnitudes , and different degrees of freedom , we need an unbiased choice of objective function to evaluate different fitting database performances based on the bayesian errors for the same testing set . here
, we consider the difference of the logarithm of the bayesian errors for one structure property function for two different fitting databases , @xmath70 and @xmath71 @xmath72 if @xmath73 is small , then @xmath74 , and @xmath75 then the right side of the equation is a relative difference in errors .
we propose the objective function of a fitting database @xmath11 with testing set @xmath18 , @xmath76 so that @xmath77 is approximately the sum of relative differences in error .
then , the optimal fitting database minimizes the sum log bayesian errors for a testing set @xmath18 .
the objective function is implicitly dependent on the relative weights in the fitting database through the bayesian error .
the gradient of the objective function with respect to weight is analytically calculable ( c.f .
( [ eqn : dedw ] ) ) .
we obtain the optimal weights in the fitting database by minimizing the objective function .
hence we will be able to compare potentials fitted with different fitting databases with respect to the same testing set
. however , the minimum of the objective function can be trivial for pathological fitting databases and testing set combinations .
a pathological fitting database and testing set combination is an _ underdetermined _ fitting database , where the mle predictions can match the true values of a dft calculation in both the fitting database and testing set .
thus , if @xmath78 for any structure @xmath20 , then @xmath79 approaches @xmath80 logarithmically . in order to eliminate the trivial minimum of pathological databases ,
we introduce a threshold function @xmath81 , @xmath82 that creates a finite minimum of @xmath83 at @xmath84 .
we can choose different error tolerances @xmath84 for each testing set structure property function .
the objective function is then @xmath85 and the derivative of the objective function is @xmath86 where the derivative calculations are from eqn .
( [ eqn : dedw ] ) and eqn .
( [ eqn : derv_thres ] ) .
the derivative of the threshold function is @xmath87 finally , note that as our likelihood function is independent of @xmath88 , so @xmath89 the optimal weights are found by minimizing @xmath90 , and this includes the addition and removal of structures from the fitting database . according to the definition of the likelihood function , eqn .
( [ eqn : logl ] ) , the fitting database could include any structures with dft calculations with a non - negative relative weight value .
structures with positive weight values are structures to fit , and all the other structures that do not contribute to the fitting will have a weight of zero
. the optimal weight value can be determined _ even for structures not presently in the fitting database_. a structure is added to the fitting database if its optimal weight value is positive , as inclusion of that structure decreases the relative error in the testing set .
a structure is removed from the fitting database if its optimal weight value is zero or negative , since removing the structure decreases the relative error in the testing set . , and an initial fitting database @xmath11
, we find the best set of potential parameters , given by the maximum likelihood estimate @xmath47 .
we then use markov - chain monte carlo ( mcmc ) to generate an ensemble of @xmath66 independent parameters ; with this ensemble , we estimate the prediction errors @xmath91 and compute the gradient of the objective function @xmath90 .
if the gradients are nonzero , we determine optimal weights @xmath17 , as well as consider addition ( @xmath92 ) or removal ( @xmath93 ) of structures from the database , and reenter the loop .
once the gradients are zero , we can determine if the testing set errors are acceptable for use ; if not , either the range of transferability is lower indicating a smaller testing set @xmath18 is needed or the potential function requires additional flexibility to increase the transferability , and the algorithm is reentered . ] fig .
[ fig : flow_chart ] outlines the new interatomic potential development algorithm based on bayesian statistics .
it starts with a conventional interatomic potential fitting procedure by selecting a potential functional form and dft structural energies and forces forming a dft data set .
we build the fitting database with a set of structures from the dft data set and assign each structure with a relative weight .
the testing set also contains a set of structures from the dft data set that test the ability of the potential to model atomic environment outside of the fitting data .
we apply non - linear weighted least squares regression to obtain the mle of parameters of the empirical potential model , and use markov chain monte carlo ( mcmc)@xcite sampling of the posterior distribution to generate the ensemble of parameters around the mle .
we calculate the mean - squared errors of the testing set structures using the parameter ensemble , and construct the objective function and its gradients .
next , we apply a conjugate - gradient method to optimize the objective function and obtain the optimal weights of the fitting database ; we can also determine if structures should be added or removed from the fitting database
. this step can take advantage of structural searching methods,@xcite for example , to identify candidate structures , though we do not do so here .
we repeat the circuit with the modified relative weight set of the fitting database until the optimal weights converge .
the testing set is the key component of this approach not only because the objective function consists of the mean squared errors of the testing set structures , but also the empirical potential predictions for structures in the testing set should have small errors whether that is known from comparison with dft calculations or estimated from bayesian sampling _ without _ dft . with the relative errors in the
testing set minimized , any weight deviation from the optimal will result in an increase in relative errors .
this means that while we could choose weights to reduce the error of one or several testing set structure property function predictions , it will worsen the predictions of other structures and the trade - off is not worthwhile .
although we are able to optimize the fitting database of the empirical potential models , an optimal fitting database does not guarantee a reliable empirical potential model .
the optimization algorithm provides the best possible empirical interatomic potential for a given a fitting database and a given testing set , but it has no judgment on whether the optimal bayesian errors are acceptable ; they can , in fact , be quite large .
this can occur if the empirical potential model does not contain the relevant physics to describe the atomic environments in the testing set , which produces reduced transferability .
then , we must for predictive empirical potential methods decide to improve the potential model itself to increase transferability or remove structures from the testing set to optimize for reduced transferability .
we apply the database optimization algorithm to a simple empirical interatomic potential model , the lennard - jones potential .
the lennard - jones potential is a two - parameter pair potential : @xmath94 & \colon \mbox{$r \leq r_{\text{cutoff}}$}\\ 0 & \colon \mbox{$r > r_{\text{cutoff}}$}\end{array } \right.\\ & -v_2(r_\text{cutoff};r_0,e_\text{b } ) \end{split } \label{eqn : lj}\ ] ] where @xmath95 is the binding energy of a dimer with a separation of @xmath96{2}r_0 $ ] .
we choose the cutoff radius @xmath97 , and the allowable parameters are @xmath98 .
.dft energy calculations of ti crystal structures .
six different crystal structures were calculated using dft - gga .
the six crystal structures are hcp , bcc , fcc , simple hexagonal , a15 and @xmath99 .
the common low temperature phase is hcp , bcc is the high temperature phase , and @xmath99 is a high pressure phase nearly degenerate in energy with hcp . [
cols="^,^,^,^,^,^,^",options="header " , ] the dft data set contains six different crystal structures of ti and the energy versus volume data for all six structures .
the dft calculations are performed with vasp@xcite , a plane - wave density functional code .
we apply a ti ultrasoft vanderbilt type pseudopotential,@xcite , with a plane - wave cutoff energy of 400ev for energy convergence of 0.3mev / atom.@xcite the k - point meshes for different structures are , @xmath100 for hcp , @xmath101 for bcc , @xmath102 for fcc , @xmath103 for hexagonal , @xmath104 for a15 and @xmath105 for @xmath99 , with methfessel - paxton smearing parameter of 0.2ev to obtain an energy accuracy of 1mev / atom.@xcite the energy versus volume data includes four different structures with volume of the unit cell as @xmath106 and @xmath107 , where @xmath108 is the unit cell volume of the equilibrium structure .
the fitting databases are built from various energy differences and energy versus volume data combinations among the six crystal structures .
we generate the markov chain of the potential parameters using the metropolis - hasting algorithm.@xcite the ensemble of the potential parameters contains @xmath109 independent parameter sets from the mcmc simulation with @xmath110 attempted steps , with an auto - correlation length of approximately 100 .
we use a reweighting scheme discussed in appendix [ app : reweight ] to approximate the objective function values for all possible sets of weights with only one sampling run .
since a radial potential model does not describe the physics of metallic bonding , we expect that the lennard - jones potential will not be transferable for testing sets containing many different structures .
our goal is for the algorithm to identify this lack of transferability in the optimization .
we systematically consider different types of fitting databases and testing sets with this in mind .
we start with a simple fitting database that contains two structural energy differences , and a testing set with the same structures .
[ fig : binary ] shows two typical objective function behaviors considering all possible weight combinations of two different two - structured fitting database and testing set combinations .
[ fig : binary01 ] shows the behavior of the objective function of a fitting database with @xmath111 and @xmath112 .
the objective function has a unique minimum with a specific relative weight ratio of the two structures .
moreover , if we calculate the derivative of the objective function with respect to weight at endpoints ( where one weight is zero ) , we can see that each derivative of the objective function with respect to weights indicates that the other structure should be added to the fitting database .
therefore the optimal weight value for both fitting database structures are positive , and we refer this as a `` mixed '' fitting database . on the other hand , fig .
[ fig : binary02 ] shows a fitting database with @xmath113 and @xmath114 .
the objective function has minima at the endpoints , which means that a fitting database containing both @xmath114 and @xmath113 has higher relative errors for the testing set than a `` pathological '' fitting databases with only one structure .
this is due to the non - transferability between hcp and fcc structures .
we refer to these pathological cases as `` unmixed '' fitting databases .
[ fig : binary_all ] shows the result of optimizing all possible combinations of two - structured fitting databases with two energy differences sharing a common reference structure .
databases with physical mles with positive @xmath95 and @xmath115 are either mixed or unmixed two - structured fitting databases .
most mixed fitting databases include fcc , bcc , hex and a15 structures and most unmixed fitting databases includes hcp or @xmath99 energy differences . by exploring a wide phase space ( six crystal structures of ti ) of lennard - jones potential fitting
, we have shown that the database optimization algorithm offers an automated , systematic and quantitative way of analyzing empirical potential model fitting with different fitting and testing sets .
elements shows databases without a physical mle . ]
we next apply the database optimization algorithm to the three - structured fitting databases where the testing set contains the same structures as the fitting database .
[ fig : fhexa15 ] is the gibbs triangle ( so that @xmath116 ) contour plot for the objective function for a fitting database constructed from @xmath113 , @xmath117 and @xmath118 .
all three of the simpler two - structured fitting databases contained in the three - structured fitting database are mixed fitting databases , and there exists a global minimum in the interior of the gibbs triangle .
the optimal weights for all three fitting database structures are positive , with an optimal weight set @xmath119 , @xmath120 , @xmath121 .
if we start with any two - structured fitting database and consider the optimal weight value for the third structure , the gradients of the objective function for the third structure are negative .
this indicates that the inclusion of the third structure will decrease the relative errors .
[ fig : poly01 ] shows the comparison of the prediction distributions with equal weights and optimal weights , where the optimal weight set provides reasonable predictions for all three testing set structures .
note , however , that we are only testing energies at _ one volume _ for each structure ; we will consider the addition of volume changes in section [ sec : e_vs_vol ] .
, is obtained in the interior area with weights @xmath119 , @xmath120 , @xmath121 . ] , @xmath120 , @xmath121 . ]
[ fig : bhexa15 ] is a gibbs triangle for a three - structured fitting database constructed from @xmath111 , @xmath122 , and @xmath112 .
while all three of the two - structured fitting databases are mixed fitting databases , the minimum occurs between @xmath111 and @xmath122 .
the optimal weight values are @xmath123 and @xmath124 .
the gradient of @xmath125 from the two - structured fitting database @xmath111 and @xmath122 is positive meaning fitting to @xmath112 will increase the relative errors in the testing set . fig .
[ fig : poly02 ] shows comparison of the prediction distributions evaluated at equal weights and optimal weights .
the prediction distribution shows that the optimal testing errors for all three structures are all about 5mev .
therefore , it suggests that for lennard - jones potential , one can fit bcc and hex structures to predict a15 structure well .
, is on the edge of @xmath111 and @xmath122 for a two - structured fitting database with weights @xmath123 and @xmath124 . ] and @xmath124 . ]
[ fig : fhexa15hcp ] is a gibbs triangle for a three - structured fitting database constructed from @xmath126 , @xmath127 and @xmath128 .
now , all three of the two - structured fitting databases are unmixed fitting databases . based on the gibbs triangle contour plot of the objective function ,
the optimal database includes only one structure @xmath127 . adding any of our candidate structures to this will increase the relative errors and worsen the predictions .
[ fig : poly03 ] shows comparison of the prediction distributions evaluated at initial weight and optimal weight .
the potential yields a very good prediction for @xmath127 , but poor estimates for the other two structures . if we use this optimal lennard - jones potential to predict @xmath126 and @xmath128 ,
the optimal distributions show that the probability of getting the true values are very low . for @xmath126 ,
the true value is 58mev with bayesian errors of 93mev , and for @xmath128 , the true value is 192mev with bayesian error of 30.8mev .
it reveals that the optimal lennard - jones potential is not transferrable for this testing set , as we expect .
, is at the corner of @xmath127 for a single database with the optimal weights @xmath129 and @xmath130 . ] and @xmath130 . ]
next we apply the algorithm to larger fitting database and testing set combinations .
the testing set includes the energy differences between hcp and the other five structures and hcp energy versus volume data .
we use four hcp structures with unit - cell volumes of @xmath106 and @xmath107 , for the hcp equilibrium volumes @xmath108 .
the fitting database starts with the same set of structures and the hcp energy versus volume data .
for the four structures representing the hcp energy versus volume data , we constrain their weights to be equal during weight optimization .
the @xmath127 and hcp energy versus volume data have optimal weight values : @xmath131 and @xmath132 , and all other weights are zero .
[ fig : e_vol_hcp2 ] shows that after weight optimization , the inclusion of hcp energy versus volume data improves the prediction of the shape of the energy versus volume curve .
it also shows the automatic removal of structures from the database by the optimization algorithm .
, @xmath127 , @xmath128 and hcp energy versus volume data and testing set including all hcp energy differences and hcp energy versus volume data .
@xmath133 with ( a ) an equal weight set and ( b ) with the optimal weight set : @xmath131 , @xmath134 and all other weights 0 . ]
we next expand our testing set to include energy versus volume data for all six structures along with the structural energy differences .
our fitting database starts with all structural energy differences , and the hcp energy versus volume data but not the other energy versus volume data .
now , the @xmath127 energy difference and hcp energy versus volume data have optimal weight values @xmath135 and @xmath136 . in fig .
[ fig : e_vol_hcp3 ] , after weight optimization , the predictions for hcp energy versus volume data improve significantly compared to the initial equal weight guess . fig .
[ fig : bf_e_vol ] shows that the optimal fitting database offers a close prediction of the shape of fcc energy versus volume curve , which is expected since the fcc and hcp structures have the same first nearest neighbor atoms .
bayesian error of the four bcc energy differences is too large to have a good predictions for the bcc energy versus volume curve .
similarly , in fig . [ fig : hexomega_e_vol ] , sloppy predictions for hex and @xmath99 energy versus volume data are obtained from the optimized fitting database .
hence , the optimal lennard - jones potential with the given fitting database and testing set combination does not have enough flexibility to predict energy versus volume curve for bcc , hex and @xmath99 .
, @xmath127 , @xmath128 and hcp energy versus volume data and testing set including all hcp energy differences and all six energy versus volume data .
@xmath133 with ( a ) an equal weight set and ( b ) with the optimal weight set : @xmath135 , @xmath136 and and all other weights 0 . ] , @xmath127 , @xmath128 , and hcp energy versus volume data , and testing set including all hcp energy differences and all six energy versus volume data . ( a ) @xmath137 and ( b ) @xmath138 with the optimal weight set from fig .
[ fig : e_vol_hcp3 ] . ] energy versus volume curve with fitting database including @xmath126 , @xmath127 , @xmath128 , and hcp energy versus volume data , and testing set including all hcp energy differences and all six energy versus volume data .
( a ) @xmath139 and ( b ) @xmath140 with the optimal weight set from fig .
[ fig : e_vol_hcp3 ] . ]
finally , we demonstrate the inclusion of a structure without dft calculations in the testing set . as we explained in section
[ sec : bayes ] , we assign the structure property prediction as the mean value of the ensemble of the prediction .
the unknown " structure we add in the testing set is an unrelaxed hcp @xmath141 supercell containing one vacancy , and the structure property function is the vacancy formation energy . as we are not including dft data , the comparison is to the mean value of the vacancy formation energy from our empirical potential .
the fitting database includes @xmath126 , @xmath127 and @xmath128 and hcp energy versus volume data .
the testing set consists of all hcp energy difference , all six energy versus volume data and the single hcp vacancy configuration .
[ fig : vacancy ] shows the prediction distributions of the vacancy formation energy before optimization and after optimization . the @xmath127 energy difference and hcp energy versus volume data have optimal weight values : @xmath142 and @xmath143 .
the result shows a significant variance reduction for the vacancy formation energy prediction .
it suggests that the prediction will be accurate if the dft calculation locates within the high likelihood parameter neighborhood of the empirical potential prediction . , @xmath127 , @xmath128 and hcp energy versus volume data .
the bottom is an optimized fitting database where the testing set includes all hcp energy differences , all six energy versus volume data and vacancy configuration .
however , the vacancy formation energy from dft _ is not used as input _ , and instead is estimated from the potential prediction . ]
we combine conventional potential parameter optimization methods and the bayesian sampling technique to propose a new definition of optimal fitting database of empirical interatomic potential models .
we choose an objective function as a function of prediction errors for the testing set , and show that minimizing the objective function is equivalent to minimizing the sum of relative errors in the testing set .
we optimize the relative weights in the fitting database to minimize the objective function and quantitatively determine the inclusion and removal of structures in the fitting database .
moreover , the performances of two different fitting databases applied on the same testing set can be compared .
the algorithm is demonstrated by a simple classical potential model , lennard - jones potential fitting of ti .
we go through all possible combinations of two - structured and three - structured fitting databases and analyze the behavior of the objective function with respect to weight change .
the new algorithm leads to the best possible empirical interatomic potential model based on the current fitting database and testing set combination .
the authors thank yuguo chen , henry wu , and michael fellinger for useful comments .
this work was supported by the office of naval research through onr award no .
when weights are changed in the database , we need to reevaluate the mean in eqn .
( [ eqn : mean ] ) using eqn . ( [ eqn : mcmc ] ) . a change to the weights in the fitting database changes the likelihood function to @xmath144 .
we rewrite @xmath145 as @xmath146 thus a reweighting term is assigned to the original data , and provides new predictions without requiring a new sampling chain .
36ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevb.29.6443 [ * * , ( ) ] link:\doibase 10.1103/physrevb.46.2727 [ * * , ( ) ] link:\doibase 10.1103/physrevb.54.6941 [ * * , ( ) ] link:\doibase 10.1016/0920 - 2307(93)90001-u [ * * , ( ) ] link:\doibase 10.1016/s0022 - 3115(99)00166-x [ * * , ( ) ] link:\doibase 10.1016/s0022 - 5096(02)00119 - 9 [ * * , ( ) ] link:\doibase 10.1103/physrevb.63.224106 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.045507 [ * * , ( ) ] link:\doibase 10.1103/physrevb.67.125101 [ * * , ( ) ] link:\doibase 10.1103/physrevb.40.6085 [ * * , ( ) ] http://stacks.iop.org/0965-0393/2/i=1/a=011 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.085311 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.125328 [ * * , ( ) ] http://stacks.iop.org/0965-0393/8/i=6/a=305 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.054121 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.214121 [ * * , ( ) ] http://stacks.iop.org/0965-0393/4/i=3/a=004 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.144119 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.134118 [ * * , ( ) ] link:\doibase 10.1103/physrevb.33.7983 [ * * , ( ) ] link:\doibase 10.1016/s1359 - 6454(01)00287 - 7 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.59.2666 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.56.632 [ * * , ( ) ] link:\doibase 10.1103/physrevb.46.2250 [ * * , ( ) ] http://stacks.iop.org/0295-5075/26/i=8/a=005 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.165501 [ * * , ( ) ] link:\doibase 10.1103/physrev.136.b864 [ * * , ( ) ] link:\doibase 10.1103/physrev.140.a1133 [ * * , ( ) ] @noop _ _ , ed .
( , , ) @noop _ _ , ed .
( , , ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1103/physrevb.47.558 [ * * , ( ) ] link:\doibase 10.1103/physrevb.54.11169 [ * * , ( ) ] @noop * * , ( ) http://dx.doi.org/10.1038/nmat1292 [ * * , ( ) ] | weighted least squares fitting to a database of quantum mechanical calculations can determine the optimal parameters of empirical potential models .
while algorithms exist to provide optimal potential parameters for a given fitting database of structures and their structure property functions , and to estimate prediction errors using bayesian sampling , defining an optimal _ fitting database _ based on potential predictions remains elusive .
a testing set of structures and their structure property functions provides an empirical measure of potential transferability . here
, we propose an objective function for fitting databases based on testing set errors .
the objective function allows the optimization of the weights in a fitting database , the assessment of the inclusion or removal of structures in the fitting database , or the comparison of two different fitting databases . to showcase this technique , we consider an example lennard - jones potential for ti , where modeling multiple complicated crystal structures is difficult for a radial pair potential .
the algorithm finds different optimal fitting databases , depending on the objective function of potential prediction error for a testing set . |
the discrete potential kdv ( dkdv ) equation @xmath0 can be considered as a simplest representative of the integrable nonlinear discrete equations in two dimensions .
this field is widely studied and many other examples can be found in the literature , see e.g. [ 1 7 ] .
however , we restrict ourselves to this simplest model , since our aim here is to analyze the generalizations of the other kind .
namely , we will discuss the role of the _ support _ of the discrete equations , that is , roughly speaking , the set where the independent variables live .
of course , the most natural way to discretize 2-dimensional pde is to consider equations on the regular square grid @xmath1 as above , but recent results [ 8 15 ] demonstrate that possibly more general synonym of ` 2d ' in the discrete case is ` planar graphs ' .
it should be noted that non - standard planar graphs were considered in mathematical physics already quite a while ago ( see e.g. korepin s works @xcite where the solvable spin models on the quasi - crystallic tilings were investigated ) .
a systematic theory of the linear difference operators on the graphs in relation to soliton theory has been initiated by s.p .
novikov ( see [ 13 17 ] ) .
we start in section [ s : qeq ] with the necessary information about integrable equations on quad - graphs , which are planar graphs with quadrilateral faces . in this case
the integrability can be understood as the so called 3d consistency property , introduced and studied in [ 9 12 ] .
the discrete potential kdv equation on quad - graph reads locally exactly as on the square lattice : @xmath2 but now it is assumed that the fields @xmath3 are assigned to the vertices of the quad - graph and the parameters @xmath4 are assigned to the edges , as shown on the fig.[fig : quad ] below , and the relation ( [ dkdv ] ) must be fulfilled for each face of the graph .
section [ s : ivp ] is devoted to the general discussion of the possible settings of cauchy , or initial value problem ( ivp ) on quad - graphs . at the first sight
the question about the correctness of ivp is not related to integrability but in fact it does very much ( see e.g. fig . [
fig : defect10 ] , [ fig : aa ] below ) .
the 3d consistency property plays an essential role here .
section [ s : corr ] contains the main result of the paper : a criterion for the existence and uniqueness of the solution of cauchy problem for an integrable equation on a quad - graph ( theorem [ th : corr ] ) . in section 5
we discuss the effect of interaction of the solutions of the dkdv equation with the localized defects in the regular square lattice .
we introduce a notion of the _ weak defect _ and show that such a defect does not affect the dynamics outside it .
several examples of the kink / solitons solutions on various quad - graphs , including quasiperiodic tilings are presented in the section [ s : kink ] . in the last section
we discuss the solutions of the linear discrete wave equation @xmath5 on quad - graphs .
surprisingly enough , in contrast to the non - linear case these solutions have a non - trivial interaction with weak defects , although the effect is limited ( see theorem [ th : wave ] ) .
we start with the results of the paper @xcite where all the integrable cases were classified among the certain class of discrete equations on quad - graphs .
the basic building block of such equation is an equation on the elementary quadrilateral of the form @xmath6 where the field variables @xmath7 are assigned to the vertices and the parameters @xmath8 are assigned to the edges of quadrilateral , as shown on the fig.[fig : quad ] .
the following properties are assumed to be satisfied : \1 ) the parameters @xmath4 on the opposite edges of any face are equal ; \2 ) the equation ( [ qeq ] ) is invariant under the group @xmath9 of the square symmetries : @xmath10 ( 200,150)(-50,-20 ) ( 0,0)(100,0)(100,100)(0,100)(0,0 ) ( -6,50)(106,50 ) ( 50,-6)(50,106 ) ( -4,-4)(104,104 ) ( -4,104)(104,-4 ) ( -23,106)@xmath11 ( 106,106)@xmath12 ( -15,-13)@xmath3 ( 105,-13)@xmath13 ( -26 , 45)@xmath14 ( 113 , 45)@xmath14 ( 45,-17)@xmath15 ( 45,110)@xmath15 \3 ) the function @xmath16 is a polynomial of degree one with respect to each field variable . the first property means that the value of parameter is constant along any _ strip _ ( or , _ characteristic _ ) , which is a sequence of quadrilaterals adjacent by the opposite edges .
we will see in the next sections that this notion is very important in the theory of quad - graph equations and plays the role similar to the characteristics for the second - order hyperbolic eqautions . in the particular case of dkdv equation on the square lattice ( [ dkdvmn ] )
we have two families of the parameters @xmath17 and @xmath18 assigned to the vertical and horizontal straight strips , but in the general case the strips may bend and intersect in a very complicated way ( for instance , self - crossing , self - tangent and closed strips are possible ) .
some characteristics are shown at the pictures below by dashed lines crossing the adjacent edges of the faces in the strip .
the second property allows us to define equation on each face independently on its position in the quad - graph .
the third property means that the equation ( [ qeq ] ) can be solved with respect to a field at any vertex as linear fractional expression of other three fields .
this expression could be singular for some special values of the fields , which is not essential for many types of quad - graphs , since we are interested only in the generic initial data .
however , this becomes important for the graphs containing self - crossing strips , as we will see in example 5 .
the notion of integrability laid in the basis of the classification in @xcite is the _
3-dimensional consistency _ @xcite which means that the equation ( [ qeq ] ) may be consistently embedded into a 3-dimensional lattice , so that the similar equations hold for all six faces of any elementary cube , as on fig.[fig : cube ] .
it should be mentioned that this property is in a close relation to the _ set - theoretical solutions of the yang - baxter equation _
@xcite , or _ yang - baxter maps _
@xcite ( see the discussion of this in the conclusions of @xcite ) .
( 200,170)(-50,-20 ) ( 100 , 0 ) ( 0 , 100 ) ( 50 , 30 ) ( 150,130 ) ( 0 , 0 ) ( 100,100 ) ( 50,130 ) ( 150 , 30 ) ( 100,0)(0,0)(0,100)(100,100)(100,0)(150,30)(150,130)(50,130)(0,100 ) ( 100,100)(150,130 ) ( 0,0)(50,30)(50,130 ) ( 50,30)(150,30 ) ( -10,-13)@xmath13 ( 90,-13)@xmath12 ( 50,17)@xmath3 ( -13,110)@xmath19 ( 160,25)@xmath11 ( 45,140)@xmath20 ( 109,95)@xmath21 ( 157,135)@xmath22 ( 20,25)@xmath15 ( 80,17)@xmath14 ( 34,65)@xmath23 to describe more precisely what does the 3d consistency mean , consider the cauchy problem with the initial data @xmath24 .
the equations on the rear faces @xmath25 allow one to determine uniquely the values @xmath26 .
after that one has three different equations for @xmath21 , coming from the front faces , and consistency means that all three values thus obtained for @xmath21 coincide . for example , in the discrete kdv case we have the formula @xmath27 independently on the order of calculations .
it should be noted that the variable @xmath3 does not appear in this expression .
this property is very important in many aspects , in particulary in the classification problem which was solved in @xcite only under this assumption .
the role of this property for the cauchy problems is not very clear . although we will not use it in our analysis of the ivp below we
will see that some results for the linear wave equation ( which satisfies all the properties 1)3 ) but not the @xmath3-independence ) and for the equations from the list @xcite will be different .
of course , linear equation is exceptional from many points of view , so it does not mean that this is related to this particular property .
the following theorem is the immediate consequence of 3d consistency .
it defines the bcklund transformation for quad - graph equation , and we will make use of it in the next section . (
90,100)(0,5 ) ( 0,50 ) ( -10,20)(-10,70 ) ( 20,20)(20,70 ) ( 30,30)(30,80 ) ( 80,20)(80,70 ) ( 40,92)@xmath28 ( -23,72)@xmath29 ( 85,72)@xmath30 ( -20,40)@xmath4 ( -23,15)@xmath31 ( 85,15)@xmath32 ( 65,0)@xmath33 [ th : qeqbt ] let variables @xmath34 , @xmath35 satisfy 3d consistent quad - graph equation ( [ qeq ] ) on the quad - graph @xmath33 .
let us fix some vertex @xmath36 and choose arbitrarily the values @xmath29 and @xmath4 .
then the formulae @xmath37 define uniquely new solution @xmath38 of equation ( [ qeq ] ) .
let us consider some path from @xmath36 to the given vertex @xmath39 . along this path
we have the sequence of equations of the form @xmath40 which define recurrently @xmath41 the fact that the different choice of the path gives the same result and that the constructed @xmath38 solve ( [ qeq ] ) follows from the 3d consistency .
the discrete zero curvature representation for equation ( [ qeq ] ) @xmath42 where @xmath43 is @xmath44 matrix can be obtained by interpretation of the new solution on the upper copy of @xmath33 as the @xmath45-function @xcite .
to do this , the formula ( [ qeqbt ] ) at @xmath46 ( ` spectral parameter ' ) should be rewritten in the form of the discrete riccati equation @xmath47 where the function @xmath48 is linear fractional due to the property 3 ) , with the coefficients depending on @xmath49 , and the latter equation is linearized by the substitution @xmath50 . in the discrete kdv case this procedure brings to the matrices @xmath51
let us consider the ivp ( or cauchy ) problem for a discrete equation on a quad - graph @xmath33 with initial data on a connected subgraph @xmath52 in most of the examples we consider @xmath53 is a _ simple path _ which is a connected sequence of the edges of the graph without self - intersections . slightly abusing terminology
we say that such a cauchy problem is _ correct _ if for the generic initial data the solution exists and unique ( without investigation of the stability question ) . for finite graphs
this is certainly justified since the stability is automatic then .
let us first compare the number of equations and unknowns for the given quad - graph @xmath33 .
we assume that @xmath33 is simply connected and finite , and contains @xmath54 faces and @xmath55 vertices , from which @xmath56 ones belong to the boundary .
now apply the double counting of the angles in all faces ( we assume for simplicity that each edge is a straight line , which , of course , is not necessary ; the euler formula provides the universal method ) .
since the sum of angles in each quadrilateral is equal to @xmath57 , hence the total sum is equal to @xmath58 .
on the other hand , summation over the vertices gives @xmath59 for the interior nodes plus @xmath60 for the boundary , so that the following equality holds : @xmath61 this means that in order to balance the number of equations and unknowns one has to to assign initial data in some @xmath62 vertices .
these vertices are not necessarily lie on the boundary , as the simplest example of a staircase in rectangular domain of the square lattice demonstrates ( see fig.[fig : rect ] ; here and everywhere initial data are marked by solid line ) .
( 200,100)(-50,-10 ) ( 200,100)(-50,-10 ) [ [ remark . ] ] remark .
+ + + + + + + the same calculation in the case of planar graph with @xmath63-lateral faces yields the formula @xmath64 .
it means , under the natural assumption that @xmath65 for large domains , that graphs with triangular faces lead to the overdetermined systems , while systems on graphs with @xmath66 are underdetermined .
of course , the balance between equations and unknowns can be restored in many ways .
for example , one can consider systems with variable number of equations per face . in this way some integrable models were obtained on the triangular and hexagonal lattices ,
see e.g. @xcite .
other possibility is to consider graphs with various types of the faces .
of course , the above consideration is too rough and can not answer how to choose the correct settings of cauchy problem even in the simplest situations . in order to get some experience , let us consider several examples assuming that equation ( [ qeq ] ) satisfies the properties 1)3 ) from the previous section , but _ is not _
necessarily integrable at first .
a priori , several cases are possible : the ivp may be correct , or it may be underdetermined or overdetermined , and the calculation scheme for the correct ivp may be explicit ( need only solving of one basic equation ( [ qeq ] ) at each step ) or implicit ( need to solve the systems of such equations ) .
it turns out that all these possibilities can be easily realized on very simple quad - graphs .
( 100,80)(-30,-10 ) ( 5,-15)a ( 100,80)(-30,-10 ) ( 10,44)@xmath12 ( 43,44)@xmath11 ( -12,28)@xmath67 ( 32,32)@xmath68 ( 70,28)@xmath4 ( 12,13)@xmath13 ( 42,13)@xmath3 ( 5,-15)b [ [ example-1 . ] ] example 1 .
+ + + + + + + + + + it is easy to see that cauchy problem shown on the fig.[fig : defect2]a is correct , that is all fields are uniquely defined for arbitrary initial data ( except for possible singularities which may arize in dkdv equation due to zero denominator ) .
in contrast , the ivp on the right picture is not correct .
indeed , the values @xmath3 , @xmath13 , @xmath11 are uniquely defined , but for the values @xmath68 , @xmath12 we have the system of two equations ( use symmetry ( [ qsym ] ) ) : @xmath69 which gives the contradiction in the generic case @xmath70 .
( 100,80)(-30,-10 ) ( 5,-15)a ( 100,80)(-30,-10 ) ( 10,44)@xmath12 ( 43,44)@xmath11 ( 64,30)@xmath67 ( 21,25)@xmath71 ( 35,32)@xmath68 ( 23,-10)@xmath4 ( 12,13)@xmath13 ( 42,13)@xmath3 ( 5,-15)b [ [ example-2 . ] ] example 2 .
+ + + + + + + + + + both ivp shown on the fig.[fig : defect22 ] are correct .
for the left one this is obvious ; for the right one we have the system of three equations on the variables @xmath72 : @xmath73 this system implies @xmath74 , @xmath75 and for @xmath12 one has the same equation @xmath76 as in the case of lattice without defect . in other words , the defect shown on these pictures does not affect the dynamics . we shall call defects with such property _
transparent_. analogously , the graph shown on the fig.[fig : defect_s ] gives explicit and implicit schemes depending on the choice of initial data .
( 130,90)(0,-10 ) ( 20,-15)a ( 130,90)(0,-10 ) ( 20,-15)b ( 170,170)(-30,-10 ) ( 60,-20)a ( 170,170)(-30,-10 ) ( 60,-20)b ( 170,170)(-30,-10 ) ( 60,-20)c [ [ example-3 . ] ] example 3 .
+ + + + + + + + + + for the quad - graph shown on the fig.[fig : defect3 ] the first two ivp are incorrect : for a ) case one finds successfully the values @xmath3 inside the square bounded by initial data and the dashed characteristic , but encounters cotradiction outside it ; for b ) case the same square remains undetermined .
[ [ example-4 . ] ] example 4 .
+ + + + + + + + + + several examples of correct ivp on different quad - graphs are shown on the fig.[fig : qg ] .
inspired by these examples and analogies with continuous case one can make a conjecture that an ivp with initial data given on a simple path @xmath53 is correct iff each characteristic ( or strip ) in @xmath33 meets @xmath53 exactly at one edge .
( 120,170)(-30,-30 ) ( 0 , 0)(0,20)2(-5,0)(85,0 ) ( 0,60)(0,20)2(-5,0)(85,0 ) ( 0 , 0)(20,0)2(0,-5)(0,85 ) ( 60,0)(20,0)2(0,-5)(0,85 ) ( -5,40)(20,40)(40,30)(60,40)(40,50)(20,40 ) ( 40,-5)(40,30)(40,50)(40,85)(60,40)(85,40 ) ( 30,85)(30,45)(50,35)(50,-5 ) ( 50,85)(50,45)(30,35)(30,-5 ) ( 20,-20)a ( 200,170)(-30,-40 ) ( 20,-50)b ( 200,170)(-30,-40 ) ( 20,-50)c ( 200,180)(-70,-10 ) ( 0,0)(-20,30)4(0,0)(40,0)4 ( -60,60)(40,0)4(0,20)(20,10)(40,20 ) ( -10,5)(20,30)4(0,0)(0,20)(20,30 ) ( 50,5)(-20,30)4(0,0)(0,20)(-20,30 ) ( 20,-20)d ( 200,180)(-30,-10 ) ( 0,0)(40,0)3(0,0)(0,40)3 ( 7.5,65)(40,0)3(0 , 0)(25 , 0 ) ( 0,-10)(25,-10 ) ( 32.5,65)(40,0)2(0 , 0)(15,-10 ) ( 0,-10)(15 , 0 ) ( 20,-20)e we finish this section with two examples which show that for the general equation this conjecture is far from being true and emphasize the role of the 3d consistency condition in this problem .
( 270,120)(0,-20 ) ( 0,0)(20,0)3(0,0)(0,85 ) ( 220,0)(20,0)3(0,0)(0,85 ) ( -5,0)(0,20)5(0,0)(270,0 ) ( 80,10)(40,20)4(0,0)(20,10 ) ( 60,0 ) ( 60,10 ) ( 80,10 ) ( 80,0 ) ( 40,20 ) ( 60,10 ) ( 100,0)(100,30)(120,30)(120,0 ) ( 40,40)(100,30 ) ( 140,0)(140,50)(160,50)(160,0 ) ( 40,60)(140,50 ) ( 180,0)(180,70)(200,70)(200,0 ) ( 40,80)(180,70 ) ( 50,-5 ) ( 50,15 ) ( 90,15 ) ( 90,-5 ) ( 70,-5 ) ( 70,35)(130,35)(130,-5 ) ( 110,-5)(110,53)(170,55)(170,-5 ) ( 150,-5)(150,72)(210,75)(210,-5 ) ( -5,10)(40,10)(60,5)(80,5)(100,10)(265,10 ) ( 45,-15)@xmath4 ( 65,-15)@xmath67 ( 85,-15)@xmath4 ( -15,10)@xmath77 ( 38,-10)1 ( 58,-10)2 ( 78,-10)3 ( 98,-10)4 ( 34 , 23)5 ( 58 , 13)6 ( 78 , 13)7 ( 94 , 23)8 [ [ example-5 . ] ] example 5 .
+ + + + + + + + + + on the quad - graph on the fig.[fig : defect10 ] , the vertical characteristics come back and cross the initial data again , while the horizontal characteristics do not cross the initial data at all .
nevertheless , this ivp is correct in general case , by means of implicit calculation scheme which can be described as follows : + 1 ) the implicit step : unknowns @xmath78 , @xmath79 , @xmath80 , @xmath81 are found as functions on @xmath13 , @xmath11 , @xmath20 , @xmath82 by solving the system @xmath83 2 ) next , all other variables on the bottom strip are computed explicitly ; + 3 ) now we are in the same situation and can repeat procedure . of course , the question is if the above system is degenerate or not .
there is no reason for it to be degenerate for generic @xmath84 but it turns out to be degenerate for 3d consistent equations . indeed , in this case we can consider four quadrilaterals corresponding to the system ( [ sys8 ] ) , as the faces of 3d cube , and therefore the initial data @xmath85 must be related by equation @xmath86 in order to provide compatibility of this system .
moreover , even if this constraint is fulfilled , the values @xmath87 can not be found uniquely .
so , in the integrable case this ivp can be considered as overdetermined and underdetermined simultaneously , in accordance to our conjecture .
( 100,100)(-20,-20 ) ( 0,0)(60,0)(60,60)(0,60)(0,0)(20,20 ) ( 40,20)(60,0)(40,40)(60,60 ) ( 0,40)(40,40)(40,20)(20,20)(20,60 ) ( -5,30)(65,30 ) ( 30,-5)(30,65 ) ( 10,65)(10,10)(50,10)(50,50)(-5,50 ) ( -13,48)@xmath4 ( -13,28)@xmath67 ( 27,-13)@xmath77 ( 22,22)1(42,20)2(42,37)3(22,42)4(-5,38)5 ( -3,-8)6(58,-8)7(58,62)8(18,62)9(-5,62)10 [ [ example-6 .
] ] example 6 .
+ + + + + + + + + + the self - crossing strip on the fig.[fig : aa ] shows that we need to make some stipulation in the integrable case . in this example the path with initial data @xmath88 crosses 3 characteristics ; one computes consequently @xmath89 and @xmath90 automatically due to the 3d consistency . at this point
the dkdv turns into @xmath91 , so that the value @xmath92 remains arbitrary .
however , in this case we can eliminate the self - intersecting strips from consideration by splitting them into the pairs of _ independent _ strips ( one of them closed ) by identification of the vertices in which the fields coincide , as shown on the fig.[fig : split ] .
it should be noted that such splitting of the equation on the quadrilateral with the pair of equal parameters : @xmath93 takes place for all equations from the list obtained in the paper @xcite .
( 130,60)(0,-30 ) ( 0,0)(40,20)(80,10)(100,0)(110,-10)(100,-20)(40,10)(20,0 ) ( 40,-10)(100,20)(110,10)(100,0)(80,-10)(40,-20)(0,0)(20,0 ) ( 40,20)(40,10)(40,-20)(40,-10 ) ( 108,18)(90,5)(70,-5)(40,-15)(10,0 ) ( 40,15)(70,5)(90,-5)(108,-18 ) ( 80,10)(80,-10 ) ( 130,-3)@xmath94 ( 130,60)(-30,-30 ) ( 0,0)(40,20)(80,0)(40,-20)(0,0)(20,0)(40,10)(60,0)(40,-10)(20,0 ) ( 40,20)(40,10)(40,-20)(40,-10 ) ( 60,0)(100,0)(110,10)(100,20)(80,0 ) ( 100,-20)(110,-10)(100,0 ) ( 108,18)(90,0)(108,-18 ) ( 10,0)(40,15)(70,0)(40,-15)(10,0 ) ( 80,0 )
now we are ready to formulate our main result .
[ th : corr ] let equation ( [ qeq ] ) be 3d consistent , and let @xmath53 be a simple path in a finite simply connected planar quad - graph @xmath33 without self - intersecting strips
. consider cauchy problem for this equation with generic initial data on the path @xmath53 .
+ 1 ) if each characteristic in @xmath33 intersects @xmath53 exactly at one edge then the ivp is correct , that is the solution exists and unique .
+ 2 ) if some characteristic intersects @xmath53 more than once then the ivp is overdetermined ( no solution exists for generic initial data ) .
+ 3 ) if some characteristic does not intersect @xmath53 then the ivp is underdetermined ( if a solution exists , it is not unique ) .
we should clarify that when we consider the intersections of characteristics with a path we count only the internal edges of the characteristics .
for example the intersection of a vertical characteristic with a staircase is one edge ( but not two ) .
we will prove this theorem by use of some transformations of quad - graphs , based on the notion of bcklund transformation ( see theorem [ th : qeqbt ] ) .
however , the existence of solution can be proved also by use of quite different idea of immersion of @xmath33 into the multidimensional unit cube .
let the conditions of the part 1 ) of the theorem are fulfilled .
let us construct an immersion of @xmath33 into the @xmath95-dimensional unit cube , where @xmath95 is the length of @xmath53 .
to do this , assign the shift operators @xmath96 , @xmath97 to the edges of @xmath53 . by definition , @xmath96 acts on @xmath95-dimensional vectors by formula @xmath98 where @xmath99 is the vector with 1 at @xmath100-th place and 0 at the others .
after this , let us distribute these operators at all edges of @xmath33 , accordingly to the same rule as for the parameters @xmath4 , i.e. one operator per strip .
obviously , this is possible in unique way due to the condition 1 ) .
now , let us choose the first vertex @xmath101 of @xmath53 as the origin with coordinates @xmath102 and define the coordinates of any other vertex @xmath3 as the result of applying shift operators along any path @xmath103 from @xmath101 to @xmath104 the result does not depend on the choice of @xmath103
. indeed , any closed path has an even intersection index with any characteristic and therefore is equivalent to identity operator .
so , we have found an immersion of @xmath33 into @xmath95-dimensional unit cube , and each face of @xmath33 now is just a 2-dimensional face in this cube .
initial values are given on the path @xmath105 and allow to calculate the values in each vertex of the cube .
( it is not difficult to explain how to obtain the values for the coordinate vectors @xmath99 , and so on . )
of course , 3d consistency is necessary here in order to justify the correctness of the procedure . in this way
, we have found some solution of the cauchy problem .
[ [ remark.-1 ] ] remark .
+ + + + + + + it is worthy to note that all the fields at the vertices of the multidimensional cube can be calculated recursively using only one equation each time , so that on the cube the integration scheme becomes _
explicit_. however , this algorithm of solving our cauchy problem is not very effective , since the number @xmath106 of the vertices in the @xmath95-dimensional cube grows exponentially in @xmath95 , while the number of vertices in @xmath33 ( and therefore the size of the algebraic system we are solving ) is about @xmath107 .
we would like to mention also that a general question of embedding of a quad - graph into cubic lattice was investigated in details by dolbilin et al @xcite .
the absence of self - intersecting strips is crucial for the existence of such embedding .
the general proof presented below is based on the simple transformations of erasing and inserting of a strip .
these transformations bring , roughly speaking , to equivalent ivp and can be described as follows .
( 600,320)(0,-170 ) ( 170,35)@xmath108 ( 400,0 ) ( 160,10)(1,-1)50 ( 340,-40)(1,1)50 ( 180,-180)(0,-20 ) let us consider some strip @xmath108 ( with parameter @xmath4 ) .
it divides @xmath33 into two subgraphs , @xmath109 and @xmath110 .
let us denote @xmath111 the corresponding borders of the strip and let @xmath112 and @xmath113 be a pair of neighbour vertices separated by @xmath108 .
now , let us apply bt ( [ qeqbt ] ) to the part @xmath110 of the quad - graph , taking @xmath114 as the seed point , and @xmath115 .
this will give us some new solution in @xmath110 such that the variables @xmath3 along the border @xmath116 coincide with the corresponding variables @xmath3 along @xmath117 .
therefore we can construct the common solution on the graph @xmath118 obtained from @xmath33 by removing of the strip @xmath108 .
this solution coincides with the old one on @xmath109 .
of course , some information is lost after such an operation and in order to make it invertible one have to remember the value @xmath119 . on the fig.[fig : cutting ] these operations are shown for the case of simple strip , however the closed or self - tangent strips are allowed as well . for the self - tangent
strips the 3d consistency property implies the coincidence of the fields in some vertices , and these vertices can be merged during the above procedure ( fig.[fig : selftangent ] ) .
( 450,140)(0,-70 ) ( 160,-40)(120,-20)(80,-20)(60,-40)(20,-40)(0,0 ) ( 20,40)(60,40)(80,20)(120,20)(160,40 ) ( 60,0)(50,-20)(30,-20)(20,0)(30,20)(50,20)(60,0)(130,0)(170,20 ) ( 130,0)(170,-20 ) ( 80,-20)(60,0)(80,20 ) ( 80,20)(20,0)3 ( 100,-20)(100,20 ) ( 80,-20)(20,0)3 ( 120,-20)(130,0)(120,20 ) ( 90,-50)@xmath109 ( 100,50)@xmath110 ( 90,50)(-1,-1)50 ( 125,50)(1,-1)50 ( 200,-5)@xmath94 ( 300,-15)(0,0)(0,30)(20,40)(40,40)(50,30)(50,0 ) ( 70,0)(70,30)(90,0)(90,30)(150,50 ) ( 150,20)(135,15)(135,15)(90,0 ) ( 0,0)(0,30)2(0,0)(10,-10)(30,-10)(50,0)(90,0)(150,-20 ) ( 0,0)(20,10)(40,10)(50,0 ) ( 50,0)(20,0)3 \2 ) let some characteristic @xmath108 intersects the path @xmath53 with initial data twice .
let us enumerate the vertices of @xmath53 as shown on the fig.[fig : nec1 ] , so that the edges of intersection are @xmath120 and @xmath121 .
( 180,70)(-20,-20 ) ( 0,0)(-20,0)(-30,10)(-10,10 ) ( -30,10)(-20,30)(0,30 ) ( 70,0)(100,0)(110,10)(90,10 ) ( 110,10)(100,30)(80,20 ) ( -8,32)(-20,10)(-8,-2 ) ( 88,27)(100,10)(88,-2 ) ( -20,33)@xmath122(100,32)@xmath123 ( -10,-8)@xmath4(80,-8)@xmath4 ( 130,15)@xmath94 ( 100,70)(-20,-20 ) ( 0,0)(0,20)(-10,30)(0,50)(20,40)(20,20 ) ( -10,10)(-10,30 ) ( 70,0)(70,20)(90,30)(80,40)(60,40)(60,20 ) ( 90,10)(90,30 ) ( 2,8)(-10,20)(0,40)(20,30)(25,30 ) ( 68,9)(90,20)(80,30)(55,30 ) ( -5,51)@xmath122(82,41)@xmath123(40,30)@xmath4 suppose that ivp is correct , then solution of equation ( [ qeq ] ) exists on the subgraph @xmath109 bounded by @xmath108 and @xmath53 .
let us look at this picture as 3-dimensional one , assuming that the exterior border of the strip @xmath108 lies at the next floor , in particular the vertex @xmath122 is over @xmath124 and @xmath123 is over @xmath125 .
accordingly to the theorem [ th : qeqbt ] the value @xmath126 can be defined by solving the sequence of equation ( [ qeqbt ] ) along the path @xmath127 that is @xmath126 is some function on @xmath128 .
this means that the initial data are not free , that is the ivp is overdetermined .
\2 ) suppose that some strip @xmath108 never intersect the path @xmath53 with initial data .
let us cut off this strip , as shown on the fig.[fig : cutting ] ( we assume for defineteness that @xmath53 lies in the subgraph @xmath109 ) .
if a solution of the ivp exists in @xmath33 then this procedure gives us a solution with the same initial data on the reduced graph @xmath118 .
next , let us apply the inverse procedure to this solution , that is consider bt on @xmath110 , taking some arbitrary value for @xmath129 .
this gives us some solution on the graph @xmath33 which coincides with the original solution on @xmath109 but not on @xmath110 .
this means that the initial data are not sufficient to provide uniqueness of the solution .
this can be obtained by sequential erasing of the strips , as explained above , until the quad - graph is exhausted .
[ [ remark.-2 ] ] remark .
+ + + + + + + although we assumed that @xmath53 is a simple path ( which is probably the most natural case ) , it follows from the proof that the result is true for any connected subgraph .
we will call a connected subgraph @xmath53 _ cauchy subgraph _ if it has the property that each characteristic in @xmath33 meets @xmath53 exactly at one edge .
a natural question is whether such a subgraph does always exist or not .
the answer in general is negative .
the following picture gives of an example of a regular lattice on which no correct ivp is possible , at least with initial data on the connected subgraphs .
it is obtained from the kagome lattice by the standard process of merging with the dual graph ( which is the quad - graph shown on the fig.[fig : qg]d ) . for the equations on quad - graphs of such type the question arise how to describe the set of general solutions .
( 150,170)(-20,-20 ) ( 0,0)(40,0)4(0,0)(-20,30)4 ( 0,10)(20,0)(40,10)(40,30)(20,40)(0,30)(0,10 ) ( 10,5)(30,35)(30,5)(10,35)(0,20)(40,20 ) ( 0,60)(20,30)(40,60 )
consider the homogeneous dkdv equation @xmath130 on the regular square lattice , assuming that all horizontal strips carry one and the same parameter @xmath4 and analogously for the vertical strips .
let us replace now some @xmath131 rectangle by a finite quad - graph @xmath132 ( with the same vertices on the boundary ) .
what we will get is a _ regular square lattice with localized defect_. many examples of such localized defects can be found in the previous section , see e.g. figures [ fig : defect2 ] , [ fig : defect22 ] , [ fig : defect_s ] , [ fig : qg]a .
outside of the defect we have a regular picture of horizontal and vertical strips .
we will say that the defect is _ weak _ if all the characteristics entering the defect leave it in the same direction but possibly in a different order .
one can see the examples of weak defects on the figures [ fig : defect22 ] , [ fig : defect_s ] , [ fig : qg]a . on the fig.[fig : defect2 ] we have a non - weak defect since a pair of characteristics change their types .
it is clear that the homogeneous dkdv equation can be naturally defined on a regular lattice with a weak defect .
consider for the corresponding equation a cauchy problem with the initial data outside the defect ( e.g. on the staircases ) just like for the usual square lattice . by the results of the previous section for generic initial data it has a unique solution .
let us compare this solution with the solution of the same cauchy problem but on the regular square lattice without defects .
the result turns out to be surprising : outside of the defect they will be the same . [
th : weak ] for the homogeneous dkdv equation a weak defect does not affect the generic solution outside the defect . in other words for homogeneous dkdv solutions the weak defects are transparent .
this is especially remarkable since for the discrete wave equation this is not true ( see below section 7 ) .
the proof is based on the existence of discrete zero - curvature representation and follows from the following general considerations .
let @xmath53 be a closed path in a simply connected quad - graph @xmath33 then for any solution of dkdv equation on @xmath33 the ordered product of the transition matrices ( [ dkdvl ] ) is equal to a scalar matrix .
more precisely the formula @xmath133 is valid where the product in the right hand side is taken over all characteristics in @xmath33 , @xmath134 is the parameter corresponding to the strip @xmath108 and the _ index _ @xmath135 is the total number of intersections of @xmath108 and @xmath53 ( without signs ) .
this formula can be easily proved from the property ( [ zcr ] ) and the relation @xmath136 equivalently , if @xmath117 and @xmath116 are two paths from @xmath36 to @xmath137 then @xmath138 [ th : ll ] let the paths @xmath117 and @xmath116 have the common starting and end points and for any characteristic @xmath108 in @xmath33 we have @xmath139 then for any generic solution of dkdv equation on @xmath33 the fields @xmath3 on @xmath116 can be uniquely recovered from the values on @xmath117 .
let us consider the matrix product @xmath140 .
one can prove that for a generic solution @xmath141 for each strip @xmath108 intersecting @xmath117 ( for some special solutions this can fail : for example if we have @xmath142 and @xmath143 for three consequent vertices in the path , then @xmath144 ) .
now the result follows from the uniqueness of the refactorization of the same matrix as @xmath145 along the @xmath146 .
it is enough to show that there exists a unique matrix of the form @xmath147 such that @xmath148 and the matrix @xmath149 is polynomial in @xmath150 .
indeed , @xmath151 is uniquely defined from the condition @xmath152 since the polynomials @xmath153 and @xmath154 are divisible by @xmath155 the matrix @xmath156 is polynomial .
continuing in the same way we will reconstruct the solution on the path @xmath116 . now to derive the theorem [ th : weak ] one should consider two sides of the rectangle @xmath132 directed towards the cauchy subgraph as @xmath117 and other two sides as @xmath157 notice that in general the cauchy problem with the initial data on @xmath117 may be underdetermined , since some characteristics may not intersect it .
for example , some closed strips may occur inside the subgraph bounded by @xmath117 and @xmath116 .
nevertheless , the use of transition matrices @xmath158 will allow us to skip over such regions of underdeterminancy . for the correct ivp the theorem [ th : ll ]
gives an integration scheme which allows to obtain the solution of dkdv equation by refactorization of the @xmath159 matrices product along the path of initial data .
in this section we consider some explicit solutions of dkdv equation rather than solutions of cauchy problem . the hint is to search for solution compatible with the continuous dynamics .
the dkdv equation on @xmath1 lattice ( [ dkdvmn ] ) originates from the nonlinear superposition principle for the bcklund transformation @xmath160 of the potential kdv equation @xmath161 .
usage of equation ( [ kdvbt ] ) allows to obtain the profile of the kink solution , and then to find algebraically the dependence of the phase on @xmath162 and @xmath123 ( the dependence on @xmath163 can be easily recovered , but now this is unessential ) .
let us start from the the simplest situation of constant parameters @xmath164 , @xmath165 . in this case
the linear seed solution is obvious @xmath166 applying the bcklund transformation ( [ kdvbt ] ) one obtains @xmath167 let us search for the dependence of @xmath168 on @xmath169 in the form @xmath170 and denote @xmath171 then substitution into ( [ homdkdvmn ] ) yields @xmath172 collecting together and splitting over @xmath173 give the relations between the parameters which can written in the form @xmath174 so that finally 1-kink solution is given by the formulae @xmath175 where @xmath176 obviously , one has to choose @xmath177 in order to obtain real @xmath178 the dependence on the parameter @xmath179 can be neglected , but we may also make use of it in order to obtain the discrete dynamics of the soliton solution @xmath180 which corresponds to the usual kdv equation @xmath181 ( see fig.[fig : sol1 ] ) .
one can check using the formula ( [ kdvbt ] ) that the corresponding fields @xmath71 satisfy the following _ discrete kdv equation _
@xmath182 or , equivalently , @xmath183 in contrast to the discrete potential kdv equation this equation is quadratic with respect to each field , and thus is not covered by the classification @xcite .
, @xmath184 , @xmath185 , @xmath186 , @xmath187 two - kink solution can be obtained via nsp for bcklund transformation , which brings to the formula @xmath188 where the seed solution @xmath189 and phases @xmath190 , @xmath191 are defined as above . in order to obtain profile without singularity in @xmath179 one may choose @xmath192 and make a shift @xmath193 . in this way the general multikink solution can be constructed .
notice the different effect of the bcklund transform which may result either in the dressing of the solution or just in the phase shifts @xmath194 , @xmath195 . the seed solution ( [ seed ] ) is easily generalized for the case of variable parameters : @xmath196 ( notice that this is more general ansatz even in the case of constant @xmath197 . )
proceeding as above one obtains the kink moving with the variable velocity : @xmath198 where @xmath199 that is @xmath200 on the fig.[fig : solbend ] the graph of the soliton @xmath201 is plotted corresponding to the values @xmath202 moreover , the formula ( [ kink2 ] ) can be easily generalized for the case of multidimensional lattice @xmath203 in which any 2d sublattice is governed by dkdv equation : @xmath204 where @xmath205 this observation immediately gives us the kink solution for the quadgraphs which can be embedded into @xmath203 lattice .
for instance , the graph shown on the fig.[fig : qg]d can be obtained from the cubic lattice by taking all its faces which intersect the plane @xmath206 , so that the kink solution for this graph is given by the formula ( [ kinkd ] ) at @xmath207 restricted on the values @xmath208 in the case of the linear phase @xmath209 the corresponding soliton propagates in the direction @xmath210 ( see fig.[fig : solhex ] ) .
, @xmath211 , @xmath212 , @xmath185 , @xmath213),title="fig : " ] , @xmath211 , @xmath212 , @xmath185 , @xmath213),title="fig : " ] some subclass of the quasiperiodic tilings can be obtained if one choose the other planes and use de bruijn s projection method ( see e.g. @xcite ) . the formula ( [ kinkd ] ) works for such quad - graphs as well . on the fig.[fig : solquasihex ] the same 3d solution is presented , restricted on the plane @xmath214 two more figures fig.[fig : solquadrants3 ] and [ fig : solquadrants6 ] present the 3-dimensional solution restricted on the coordinate quadrants .
( @xmath215 , @xmath216 , @xmath217 , @xmath185 , @xmath186 , @xmath218),title="fig : " ] ( @xmath215 , @xmath216 , @xmath217 , @xmath185 , @xmath186 , @xmath218),title="fig : " ] ( @xmath215 , @xmath216 , @xmath217 , @xmath185 , @xmath213),title="fig : " ] ( @xmath215 , @xmath216 , @xmath217 , @xmath185 , @xmath213),title="fig : " ] an interesting problem is to investigate the interaction of the kink solutions with the localized defects in the square lattice .
according to the theorem [ th : weak ] a weak defect in the homogeneous lattice does not affect solutions , so that the ordinary kink ( [ kink ] ) does not change .
in contrast , it can be shown that the bended kink ( [ kink2 ] ) may acquire some phase shift after the interaction with such a defect .
it is interesting to compare the theory of the homogeneous dkdv equation we discussed above with the discrete linear wave equation @xmath219 on quad - graphs .
obviously , equation ( [ wave ] ) always can be solved with respect to any vertex and it does not contain singularities in contrast to the dkdv case .
therefore the splitting of self - intersecting strips is not valid in this case .
the most serious distinction is that since the discrete wave equation does not contain the parameters the corresponding matrices ( [ zcr ] ) @xmath220 contain no spectral parameter and give the fake zero curvature representation . in particular ,
the refactorization scheme from the section [ s : ll ] does not work , and the theorem [ th : weak ] is not valid , as the following simple examples demonstrate .
of course , for the wave equation it is sufficient to analyze the propagation of delta - like initial data ( black circle denotes 1 , white one denotes @xmath221 , no circle denotes 0 ) .
( 130,110)(-30,-10 ) ( 20,-20)a ( 20,0)(0,20)5 ( 40,45)(40,60 ) ( 40,80 ) ( 130,110)(-30,-10 ) ( 20,-20)b ( 40,0)(40,20)(40,30 ) ( 40,45)(40,60 ) ( 40,80 ) ( 60,50)(60,60 ) ( 60,80 ) ( 130,110)(-30,-10 ) ( 20,-20)c ( 60 , 0)(60,20)(60,35 ) ( 40,45)(40,60)(40,80 ) nevertheless we can claim that a weaker version of the theorem [ th : weak ] is valid also for the discrete wave equation .
let us consider the cauchy problem for this equation on a regular square lattice with a weak defect @xmath132 with initial data on a staircase ( or coordinate half - axes ) outside the defect .
again as in dkdv case according to our theorem [ th : corr ] the solution exists and unique and one can compare this solution with the one on the regular lattice without defects .
let us consider the _ cross
_ @xmath222 corresponding to the defect @xmath132 as the union of all the characterstics passing through @xmath132 .
besides the defect itself it contains also all the vertical and horizontal strips passing through the rectangle .
[ th : wave ] for the discrete wave equation a weak defect @xmath132 does not affect the solution outside the cross @xmath222 .
the proof easily follows from the fact that for the discrete wave equation the difference @xmath223 is preserved when we are moving along the corresponding charactersitic .
actually the affected area is only a half of this cross ( `` thick hook '' ; see the picture below ) . for a general equation one should expect the whole quadrant bounded by this hook to be affected .
( 200,160)(-30,-10 ) ( 185,60)(60,60)(60,100)(185,100 ) ( 60,145)(60,100)(120,100)(120,145 ) ( 80,70)(100,70)(100,90)(80,90)(80,70 ) ( 83,74)@xmath132 ( 60,0)(120,0)(120,60)(60,60)(60,0 ) ( 0,60)(60,60)(60,100)(0,100)(00,60 ) ( 0,20)(0,20)3(0,0)(185,0 ) ( 0,80)(60,80 ) ( 120,80)(185,80 ) ( 0,100)(0,20)3(0,0)(185,0 ) ( 20,0)(20,0)3(0,0)(0,145 ) ( 120,0)(20,0)4(0,0)(0,145 ) ( 80,0)(80,60)(80,100)(80,145 ) ( 100,0)(100,60)(100,100)(100,145 )
we have shown that for the integrable equations on the quad - graphs it is possible to formulate a criterion for the correctness of the cauchy problems .
the 3d consistency condition plays a crucial role in our analysis .
the question is whether such a criterion exists for a general eqaution on quad - graphs .
one can easily produce a lot of examples of graphs and ivp which are not sensitive to a particular form of the equation , but to describe all of them in some geometric way seems to be a hard problem .
we have not discussed also global geometrical aspects ( e.g. the quad - graphs on the surfaces of genus @xmath224 ) .
for some interesting results in this direction in the linear case we refer to novikov and dynnikov papers [ 13 16 ] .
novikov , i.a .
discrete spectral symmetries of low - dimensional differential operators and difference operators on regular lattices and two - manifolds .
_ uspekhi mat .
nauk * 52:5 * _ ( 1997 ) 175234 [ english translation : _ russian math .
surveys * 52:5 * _ ( 1997 ) 10571116 ] .
bobenko , t. hoffmann , yu.b .
hexagonal circle patterns and integrable systems : patterns with the multi - ratio property and lax equations on the regular triangular lattice .
notices * 2002:3 * _ 111164 .
dolbilin , a.g .
sedrakyan , m.a .
shtanko , m.i .
topology of a family of parametrizations of two - dimensional cycles arising in the three - dimensional ising model .
nauk sssr * 295:1 * _ ( 1987 ) 1923 [ english translation : _
soviet math .
dokl . * 36:1 * _ ( 1988 ) 1115 . | initial value problems for the integrable discrete equations on quad - graphs are investigated .
a geometric criterion of the correctness of such a problem is found .
the effects of the interaction of the solutions with the localized defects in the regular square lattice are discussed for the discrete potential kdv and linear wave equations .
the examples of kinks and solitons on various quad - graphs , including quasiperiodic tilings , are presented . |
our goal in this paper is to investigate the properties of a quantum mechanical wave function of very many variables known as the laughlin function @xcite .
it arises as an explanation of a curious and unexpected phenomenon known as the fractional quantum hall effect ( fqhe ) @xcite .
it is not our intention to discuss the physics of fqhe but rather to discuss the mathematical problems that arise when the laughlin function is appropriately modified for a two - dimensional cylindrical geometry @xcite .
the interesting point here is that although the setup is seemingly translation invariant ( in the direction of the cylinder axis ) the resulting one - particle density @xmath2 is not invariant but has a non - trivial periodicity of length @xmath1 ( in natural units ) , where @xmath1 is an integer related to the so - called `` filling fraction '' .
the main content of this paper is the rigorous verification of this periodicity when the cylinder radius is small compared to the natural length .
the periodicity has a natural interpretation in the context of classical coulomb systems and is manifest in the periodicity of a closely related `` quantum polymer '' @xcite .
the conjecture of periodicity for all radii remains open .
+ let us just say a few words about the function and why it was proposed .
imagine a hollow circular cylinder in @xmath3 of radius @xmath4 on whose surface @xmath5 electrons reside .
each electron has a coordinate @xmath6 , with @xmath7 denoting the distance parallel to the cylinder axis and @xmath8 denoting the angular coordinate .
we will often use @xmath9 to denote the coordinate instead of @xmath6 .
there is a magnetic field of magnitude @xmath10 perpendicular to the surface . since magnetic fields
must be divergence free we can think of this situation as one in which there is a ( unobserved ) flux through the hollow cylinder that leaks out through the surface with a constant flux through the cylindrical surface on which the electrons reside ( see also @xcite , sect .
if we pay attention only to the surface the physical situation is translation invariant but if we pay attention to the flux in the hollow core we can measure our @xmath11 coordinate by measuring the total flux in the circular cross section at @xmath11 .
the magnetic vector potential , @xmath12 , evaluated on the surface , is given by stokes theorem ( assuming that @xmath12 has no component in the axial direction ) as @xmath13 , which is obviously periodic in the @xmath14 direction .
our vector potential @xmath15 is determined only up to a gauge parametrized by @xmath16 . with this choice
the one particle schrdinger operator is @xmath17 in our units planck s constant @xmath18 , the mass and the charge of the electron are @xmath19 . to simplify notation @xmath20 actually denotes the magnetic field divided by the speed of light , and in the following we will choose units in such a way that @xmath21 .
our operator @xmath22 should be seen as the part of a three - dimensional hamiltonian which acts on the axial and the angular parts of the wave function .
the radial , angular and axial coordinates , the part of the laplacian that we are interested in is @xmath23 .
fixing @xmath24 and choosing @xmath25 , this becomes @xmath26 : the differential symbol is @xmath4-independent . ]
the vector potential @xmath12 is the restriction to the cylinder surface of some three - dimensional vector potential that gives rise to a magnetic field as sketched in @xcite , fig .
3 . in the chosen units ,
the `` magnetic length '' @xmath27 equals @xmath19 .
an important role will be played by the dimensionless parameter @xmath28 here we come to the crucial point that the usual gauge invariance ( meaning that we can replace @xmath15 by @xmath29 and a wave function @xmath30 by @xmath31 $ ] ) is not generally allowed .
gauge changes @xmath32 have to respect the @xmath33-periodicity in the @xmath14 direction .
consequently , the number @xmath16 in @xmath22 can not be changed by a gauge transformation unless the change is a multiple of @xmath34 .
this lack of complete gauge symmetry is intimately related to the breaking of translational symmetry .
if we translate in the axial direction by @xmath35 then @xmath36 , the hamiltonian ( [ laugh ] ) with this replacement is generally _ not _ gauge equivalent to the original one . that is , changing @xmath12 by @xmath37 with @xmath38 is not allowed except for the special cases @xmath39 with @xmath40 .
hence the system possesses a discrete translational symmetry .
the discrete translational invariance and the effect of changing @xmath16 can be nicely read off the eigenfunctions of @xmath22 .
the eigenfunctions belonging to the lowest eigenvalue are all of the form @xmath41\ ] ] where @xmath42 is an entire function
. to respect @xmath33 periodicity , the eigenfunction must be a superposition of functions @xmath43
\propto \exp({\mathrm{i}}n\gamma y ) \exp \bigl [ -\bigl(x - ( n-\frac{\phi}{2\pi})\gamma\bigr)^2/2 \bigr ] , \label{lowl}\ ] ] with @xmath44 .
these functions are essentially gaussians of @xmath11 ; their centers form a lattice with spacing @xmath45 , and changing @xmath16 amounts to a uniform shift of the centers . from now on , we set the parameter @xmath16 to @xmath46 .
if we now consider a finite cylinder of length @xmath47 then we impose the ` physical ' boundary condition that we restrict functions to the interval @xmath48 and consider only ground state eigenfunctions with centers inside that interval . ) with chiral boundary conditions @xcite . ]
thus , for a finite cylinder , the dimension of this lowest energy space - known as the lowest landau band - equals the `` number of fluxes '' @xmath49 , i.e. , the total flux @xmath50 divided by the flux quantum @xmath34 ( remember @xmath21 ) . if there are some number @xmath5 of electrons that repel each other via a coulomb interaction we have to construct a wave function of @xmath5 variables @xmath51 that minimizes the total energy .
the kinetic part is minimal if we choose sums of products of one - particle functions from the lowest landau band . in other words
our wave function must have the form @xmath52 where @xmath53 is a polynomial in @xmath5 variables of degree at most @xmath49 in each variable separately .
furthermore , the pauli exclusion principle for fermions demands that this polynomial be antisymmetric , i.e. , that it change sign if any two variables are interchanged .
if the particles are bosons , instead , then @xmath53 must be symmetric .
we are interested in the cases that @xmath54 with @xmath55 , in which case we say that the lowest landau band is filled to a fraction @xmath0 .
the function @xmath56 satisfies all the stated conditions for fermions ( when @xmath1 is odd ) and for bosons ( when @xmath1 is even ) .
later , it will be useful to fix the multiplicative factor @xmath57 as in eq .
( [ multconst : eq ] ) below .
the factor @xmath57 does not affect the one - particle density .
our choice of @xmath57 makes @xmath58 the power of a slater determinant of functions that resemble the lowest landau level basis functions ( see eq . ( [ powdet : eq ] ) ) , and considerably simplifies subsequent statements on normalization constants .
this function was invented by laughlin @xcite for the disk geometry and later adapted to the cylinder geometry @xcite , and is considered to be a good approximation to the true ground state .
its distinctive feature is that it is small when two particles @xmath59 and @xmath60 are close together , thereby making the interaction energy small .
+ in this article , we analyze the periodicity in the axial direction of the state @xmath61 in the limit @xmath62 for fixed radius @xmath4 and filling factor @xmath0 .
particular emphasis will be put on the one particle density @xmath63^{n-1 } } \bigl|\psi_n ( z , z_2, .. ,z_n)\bigr|^2 { \mathrm{d}}y_2 .. { \mathrm{d}}y_n { \mathrm{d}}x_2 .. { \mathrm{d}}x_n,\ ] ] where @xmath64 is the @xmath65-norm squared with respect to integration over @xmath66)^n$ ] . notice that we integrate over infinite cylinders but keep referring to @xmath58 as a finite cylinder state .
the reason is that the wave function vanishes exponentially fast outside the finite cylinder . for @xmath67 ,
the wave function is a simple slater determinant of basis functions ( [ lowl ] ) , the density is a sum of equally weighted gaussians @xmath68 $ ] ( see eq .
( [ density : eq ] ) ) , and changing the domain of integration to a finite cylinder does not affect the density in the middle of the cylinder .
the origin of the analysis of the density s periodicity is an observation for the filled lowest landau level : for @xmath67 , the density is periodic with minimal period @xmath45 .
this contrasts with the constant density of the filled landau level in the symmetric gauge and reflects the discrete translational symmetry of the hamiltonian ( [ laugh ] ) .
when @xmath69 , laughlin s state does not possess the full discrete symmetry of the hamiltonian ( [ laugh ] ) any longer , hence the symmetry is broken . however , the state is still periodic , with minimal period @xmath70 , i.e. , @xmath1 times the period of the filled landau level .
this is rigorously shown here for cylinders with sufficiently small radius ( large @xmath45 ) ; numerical results @xcite for @xmath71 and @xmath72 suggest that the result actually holds for all values of the cylinder radius .
related , finite @xmath5 results have been obtained in @xcite .
related considerations for laughlin type functions on tori , introduced in @xcite , can be found in @xcite . in @xcite , the amplitude of oscillations of the density for torus laughlin - type functions
is investigated numerically , as a function of the cylinder radius .
it follows from our results that the density and hence , the amplitude of @xmath73-oscillations , are analytic function of @xmath45 when @xmath45 is sufficiently large .
this implies that the only possibility for the oscillation amplitude to vanish on some small @xmath45 interval is that a phase transition occurs as @xmath45 is decreased , see @xcite for a similar argument .
( with `` phase transition ''
we mean that either there is no longer a well - defined unique thermodynamic limit or there is a unique thermodynamic limit but the density ceases to be analytic at some point . )
@xcite argue that no such phase transition occurs , but a rigorous proof is still lacking . due to the picture mentioned before , translation along the cylinder axis can be reinterpreted as a change of flux @xmath16 .
one flux unit ( @xmath34 ) corresponds to a translation by one natural length unit ( @xmath45 ) .
hence the many electron system is invariant under a change of @xmath1 flux units . in the picture of laughlin s original article on the integer hall effect @xcite
this corresponds to a transport of one electric charge from one edge to the other of the cylinder for every adiabatic change by @xmath1 flux units .
this is related to a heuristic argument given in @xcite , suggesting that the periodicity is needed in order to reconcile laughlin s iqhe article with fractional quantum hall conductances .
another interpretation of the periodicity is that the @xmath5 particles in a cylinder of length @xmath74 are not arranged randomly but , more like a one - dimensional crystal , they are separated by a distance of @xmath1 units .
this interpretation is of interest from the point of view of laughlin s plasma analogy @xcite : @xmath75 is proportional to the boltzmann weight of a classical jellium system placed on a cylinder ( see also @xcite , sect .
the factor @xmath76 in the wave function arises from the potential created by a neutralizing background while the polynomial in @xmath77 accounts for the repulsive interaction between charged particles .
one - dimensional jellium systems always display periodicity , i.e. , `` wigner crystallization '' , as was shown in @xcite . for jellium tubes ,
the symmetry breaking is well - known at coupling constant @xmath78 , corresponding to the filled landau level , see @xcite .
the periodicity has been conjectured to hold for other values of the coupling constant @xcite .
our results give a partial proof of this conjecture .
+ our result on the periodicity comes in several parts : a preliminary observation ( lemma [ foutrans : lem ] ) , holding for all values of @xmath45 , is that the one particle density can never be constant , and if it is periodic , the period must be a multiple of @xmath45 .
theorem [ thermolim : theo ] , valid for large @xmath45 , states that the laughlin state has a unique thermodynamic limit as @xmath79 , and the limiting state has @xmath70 as one of its periods .
theorem [ symbreak : theo ] says that @xmath70 is actually the smallest period .
we also prove , for large @xmath45 , that the state is mixing with respect to ( magnetic ) translations along the direction of the cylinder axis ( theorem [ cluster : theo ] ) .
the crucial idea for the proof of our results is the representation of laughlin s wave function as a `` quantum polymer '' @xcite .
this means that @xmath58 is written as a sum of functions , each associated with a partition of @xmath80 .
one of these functions is the tao - thouless function @xcite , which is the antisymmetrized tensor product of lowest landau band functions centered at multiples of @xmath70 .
the tao - thouless function can be interpreted as a simple monomer , see fig .
[ fig : intro ] .
the polymer expansion is explained in subsects .
[ model : sec ] and [ poly : sec ] .
the associated polymer system has a built - in form of translational invariance that is at the origin of the @xmath70-periodicity in the limit @xmath79 .
one consequence of the polymer representation is a recurrence relation for the normalization constant @xmath81 , known as a ( discrete ) renewal equation . in subsect .
[ norm : sec ] , we show that the associated renewal process has finite mean for sufficiently large @xmath45 .
we use this important technical result in subsects .
[ thermo : sec ] and [ symclu : sec ] to prove our results on laughlin s function . at this point , let us mention that plasma analogy sheds some light not only on the density s periodicity , but also on the recurrence relation : indeed , in the context of jellium systems , a renewal equation already appeared in @xcite , while a renewal inequality showed up in @xcite . in sect .
[ solv : sec ] , we investigate a solvable model that arises as a simplification of laughlin s function . in this model ,
only monomer - dimer systems occur .
this allows us to introduce in a simpler setting the ideas described above .
as a preparation to the examination of laughlin s wave function , we analyze a modified wave function . it is derived from laughlin s wave function by replacing gaussians with functions of compact support .
if the support is small , the normalization constants and the one - particle density can be computed explicitly ( prop .
[ modfunct : prop ] ) , and it is easy to check that the thermodynamic limit of the one - particle density is a periodic function ( corollary [ thermolim : cor ] ) .
the proof of these results relies on the representation of the modified wave function as a sum over monomer - dimer partitions ( lemma [ solvpoly : lem ] ) , leading to a recurrence relation of second order for the normalization constants , see eq .
( [ solvrecurr : eq ] ) .
the derivation of the modified wave function starts from the well known representation of laughlin s wave function as the @xmath82 power of a vandermonde determinant @xcite times a gaussian weight .
the gaussians can be absorbed into the determinant and we can complete the squares . for @xmath83 particles and @xmath72 ,
this gives @xmath84 & \ \
\exp\bigl[{\mathrm{i}}\gamma y_2- ( x_2 - 3\gamma)^2/6\bigr ] \end{pmatrix}^3 . \ ] ] more generally , with a suitable choice of @xmath57 ( see eq .
( [ multconst : eq ] ) below ) , laughlin s function becomes @xmath85 where @xmath86 notice that , up to the multiplicative constant , the quantity @xmath45 appears only in combination with the integer @xmath87 .
hence it suffices to consider the case @xmath88 .
the @xmath45-dependence can be restored by replacing in the forthcoming formulas integers @xmath89 by @xmath90 . to simplify the formulas further ,
we will limit ourselves to @xmath72 .
now we replace the gaussian in formula ( [ powdet : eq ] ) by an even measurable function @xmath42 of compact support .
let @xmath91 for @xmath92 , we define the modified wave function @xmath93 by @xmath94 it turns out that the one - particle density can be computed exactly when only neighbors overlap : @xmath95 if @xmath96 .
this can be achieved by choosing a function @xmath42 with a compact support contained in @xmath97 $ ] , so that @xmath98 the main idea in this section is to use the nearest neighbor overlap condition to simplify the expansion of the determinant power ( [ modfunct : eq ] ) . because of the third power , the expansion of eq .
( [ modfunct : eq ] ) can be written as a sum of wedge products . ] of functions @xmath99 .
the overlap condition restricts @xmath100 to @xmath101 , @xmath102 , @xmath103 , etc .
we define an orthonormal set of functions @xmath104 , labelled by their @xmath14 momentum @xmath105 , by : @xmath106 the wave function @xmath107 describes a particle localized around @xmath108 . the constants @xmath109 and @xmath110 are suitable positive normalization constants .
the normalization of @xmath93 and the one - particle density are easily expressed in terms of @xmath111 and @xmath104 : [ modfunct : prop ] let @xmath42 satisfy the nearest neighbor overlapping condition ( [ nearestneighbor : eq ] ) .
let @xmath112 .
let @xmath81 be the norm squared of @xmath93 , considered as a function in @xmath113)^n\bigr)$ ] , and @xmath114 the one - particle density .
then @xmath115 before we turn to the proof of the proposition , let us observe that it has the periodicity of the one - particle density in the thermodynamic limit as a simple consequence . to see this ,
observe that translating a function @xmath104 by a multiple of @xmath116 along the cylinder axis amounts , up to a phase factor , to increasing the index @xmath117 by a multiple of @xmath116 : @xmath118 the density @xmath114 of eq .
( [ modfunct : eq2 ] ) is a function of @xmath119 alone and vanishes when @xmath11 is far outside the interval @xmath120 $ ] .
we are interested in the limit of the density around the middle of this interval . for a given @xmath121 ,
only finitely many @xmath87 s contribute to the sum ( [ modfunct : eq2 ] ) .
when @xmath11 is near the middle of the cylinder , the @xmath87 s that contribute are of the order of @xmath122 .
but now , because of eq .
( [ modfunct : eq1 ] ) , @xmath123 when @xmath87 and @xmath124 both go to infinity .
a similar statement holds for the coefficient in the second line of eq .
( [ modfunct : eq2 ] ) .
hence in the limit @xmath79 , the shifted density @xmath125 converges to @xmath126 because of the covariance ( [ covarsolv : eq ] ) , this function is @xmath116-periodic .
thus we have obtained the following corollary from prop .
[ modfunct : prop ] : [ thermolim : cor ] there is a @xmath127-periodic function @xmath128 of the coordinate @xmath11 alone such that @xmath129 for all @xmath130 .
[ rem : constdens ] the corollary does not exclude periods smaller than @xmath127 . the density can even be constant .
for example , when @xmath42 is a step function taking only two values - one on the intersection of @xmath131 and the shifted supports @xmath132 and another one on the remainder of @xmath131 - a suitable choice of the two values gives a constant density . however , for most choices , the density will have a non - trivial periodicity .
now let us turn to the proof of prop .
[ modfunct : prop ] .
the proof of the proposition rests on a particular representation of @xmath93 in terms of `` monomer '' and `` dimer '' functions @xmath133 where @xmath134 . in the following lemma , the modified wave function @xmath93 is written as a sum over partitions @xmath135 of @xmath136 into monomers @xmath137 and dimers @xmath138 .
we will always assume that the partitions are _ ordered _
, i.e. , the elements of @xmath139 are smaller than those of @xmath140 , etc .
[ solvpoly : lem ] the modified wave function @xmath93 can be represented as a sum over ordered monomer - dimer partitions of @xmath80 : @xmath141 the function @xmath93 is defined as the power of a determinant .
this leads to the expression @xmath142 the sum is over permutations of @xmath143 . expanding the power of the sum
, we see that @xmath93 equals @xmath144 times
the antisymmetrization of @xmath145 suppose @xmath146 give a non - vanishing contribution to the sum above .
then because of the nearest neighbor overlapping condition , they must satisfy @xmath147 as a consequence , @xmath148 and @xmath149 are products of disjoint nearest neighbor transpositions @xmath150 .
the set of transpositions is uniquely determined by the permutations @xmath151 .
it can be represented by a monomer - dimer partition of @xmath136 , with @xmath152 dimers corresponding to the transpositions supports .
thus to each pair of permutations fulfilling eq .
( [ nnpermut : eq ] ) we can assign a monomer - dimer partition .
the proof of eq .
( [ solvpoly : eq ] ) is concluded by the following observation : if in eq .
( [ eq : sita ] ) we sum over permutations @xmath151 that give the _ same _ partition @xmath153 , and then antisymmetrize the resulting sum , we obtain @xmath154 . to see this , we consider the example @xmath155 , @xmath156 , @xmath157 and leave the general case to the reader .
there are three permutation pairs @xmath158 giving rise to the partition @xmath159 , namely all combinations of the identity @xmath160 and the transposition @xmath161 except @xmath162 .
their contribution to the sum ( [ eq : sita ] ) is @xmath163\ ] ] antisymmetrization and multiplication by @xmath164 gives @xmath165 .
lemma [ solvpoly : lem ] allows us to prove prop .
[ modfunct : prop ] , using the orthogonality of contributions from different monomer - dimer partitions .
the key observation is that each monomer - dimer partition @xmath135 is associated with a unique set of @xmath14-momenta @xmath166 : by the definition ( [ defpolyfunct : eq ] ) of @xmath167 , we have @xmath168 the @xmath14 momenta @xmath169 are obtained as follows : when @xmath135 is a monomer partition , @xmath170 equals @xmath171 .
the vector belonging to a partition containing the dimer @xmath138 is obtained from this reference vector by replacing @xmath172 with @xmath173 , see fig .
[ fig : mondim ] .
thus different partitions give rise to different sets of @xmath14-momenta . whence the orthogonality of the various contributions in eq .
( [ solvpoly : eq ] ) .
the normalization becomes @xmath174 as a consequence , @xmath81 satisfies the recurrence relation @xmath175 with the initial conditions @xmath176 , @xmath177 , whence eq .
( [ modfunct : eq1 ] ) .
( [ solvpoly : eq ] ) and ( [ eq : partym ] ) lead to @xmath178 with @xmath179 and @xmath180 . changing the order of summation and combining with eq .
( [ norm : solv : eq ] ) gives eq .
( [ modfunct : eq2 ] ) .
_ from eq .
( [ norm : solv : eq ] ) we see that @xmath81 is a monomer - dimer - partition function for a linear chain of length @xmath5 . the recurrence relation ( [ solvrecurr : eq ] )
was already given in @xcite .
similar results hold for @xmath181 and slightly different definitions of monomer and dimer functions .
_ if we relax from the nearest neighbor overlapping condition ( [ nearestneighbor : eq ] ) , lemma [ solvpoly : lem ] still holds true provided we allow polymers of arbitrary length instead of only monomers and dimers .
the normalization constants then satisfy a recurrence relation of infinite order .
_ one last remark concerns the comparison with the torus wave functions introduced by haldane and rezayi @xcite .
we can define the modified torus wave function @xmath182 by eq .
( [ modfunct : eq ] ) , replacing the functions @xmath183 on the right - hand side by periodified functions @xmath184 if in eq .
( [ rootbasefunct - mod : eq ] ) @xmath183 is defined with gaussians @xmath185 instead of a function with compact supports , the modified function @xmath182 is a haldane - rezayi torus function .
when the nearest neighbor overlapping condition is satisfied , the torus function @xmath182 has an explicitly computable normalization and one - particle density .
it is associated with a monomer - dimer system on a ring , with one long additional polymer covering the whole ring . in the limit of long tori
, we recover the density of the cylinder modified function .
thus for the solvable model , one can check that torus and cylinder functions are equivalent .
+ to summarize , we have seen how the representation ( [ solvpoly : eq ] ) of @xmath93 in terms of a `` quantum monomer - dimer '' system leads to simple formulas for the normalization constants and the one - particle density ; the periodicity of the one - particle density in the thermodynamic limits stems from the translational invariance of the corresponding monomer - dimer system . in the next section
, we will see that laughlin s wave function admits a similar representation in terms of a polymer system , and the mechanism leading to a periodic one - particle density is essentially the same
. however , the situation is complicated by the existence of polymers of arbitrary length and non - vanishing activity .
this section is devoted to the proof of the main results of this paper . in the first subsection , we prove general properties of laughlin s function . we proceed with a polymer representation ( subsect . [
poly : sec ] ) , leading to a recurrence relation of infinite order for the normalization constants .
this relation is exploited in subsect .
[ norm : sec ] where we prove an important technical result on the asymptotics of normalization constants .
the existence of the thermodynamic limit of laughlin s state and its periodicity are shown in subsect .
[ thermo : sec ] .
symmetry breaking and clustering properties are proved in subsect .
[ symclu : sec ] . throughout this section
, we will assume that @xmath1 is odd . however , most results hold for even @xmath1 , with minor modifications ( e.g. , replacing wedge products with symmetric tensor products ) .
the multiplicative constant in the definition ( [ laufunct : eq ] ) of @xmath58 will be fixed as @xmath186 laughlin s wave function has a few number of simple , but important properties that we present in this subsection .
they allow us to derive symmetries of laughlin s state and to show that in the limit of infinitely many particles , laughlin s state can not give rise to a constant density .
this holds regardless of the size of the radius , in contrast with the results proved in subsects .
[ thermo : sec ] and [ symclu : sec ] .
all properties rely on the expansion of @xmath58 as a sum of wedge products of lowest landau level basis functions @xmath187 they form an orthonormal set in @xmath188)$ ] .
one basis function can be transformed into another by a shift in the axial direction : @xmath189 where @xmath190 is the _ magnetic translation _ @xmath191 .
the expansion of @xmath58 can be obtained from the expansion of the @xmath1-th power of the vandermonde determinant into monomials .
define coefficients @xmath192 and @xmath193 by @xmath194 a straight - forward computation then gives @xmath195 this equation allows us to translate properties of the @xmath1-th power of the vandermonde determinant into properties of @xmath58 .
the power of the vandermonde determinant ( [ vandermonde : eq ] ) is a homogeneous polynomial of total degree @xmath196 .
therefore @xmath58 has definite @xmath14-momentum @xmath197 as a consequence , the one - particle density will be independent of the angular coordinate @xmath14 .
an additional symmetry comes from the `` reversal invariance '' @xcite @xmath198 laughlin s state is invariant with respect to a @xmath199 rotation around the middle of the cylinder , @xmath200 here @xmath201 refers to a `` magnetic '' rotation , @xmath202 .
notice that @xmath203 . because of eqs .
( [ lauexp : eq ] ) and ( [ yinvariance : eq ] ) , the one - particle density can be expressed as : @xmath204 @xmath205 refers to expectation values in the state @xmath206 and @xmath207 , @xmath208 are the fermionic creation and annihilation operators for the state @xmath209 , e.g. , @xmath210 .
notice that the expectation values @xmath211 , @xmath212 , vanish due to the @xmath14-invariance ( [ yinvariance : eq ] ) .
we are interested in limits of @xmath213 as @xmath214 .
( [ density : eq ] ) shows that any limit point of the shifted density is a sum of gaussians .
the following lemma deduces several statements from this observation . 1 .
@xmath128 can not be a constant .
2 . if @xmath128 is periodic , every period must be a multiple of @xmath45 .
@xmath128 is periodic with period @xmath70 if and only if @xmath218 is periodic with period @xmath1 .
@xmath219 is the convolution of the function @xmath220 and the measure @xmath221 . with @xmath2 and
@xmath222 we can associate tempered distributions @xmath223 and @xmath224 : @xmath225 their fourier transforms ( @xmath226 ) satisfy @xmath227 the product @xmath228 is the distribution @xmath229 .
it follows from eq .
( [ fouconv : eq ] ) that the mapping @xmath230 is injective .
therefore @xmath42 and @xmath222 must have the same periodicity .
since any period of @xmath222 is obviously a multiple of @xmath231 , the same must be true for periods of @xmath2 .
this excludes a constant density , since constant functions admit periods that are not multiples of @xmath45 .
notice that it is not completely trivial that the density can not be constant .
if instead of gaussians we added up more general functions , we could obtain a constant .
examples include functions of compact support , see the remark on p. , and log - concave functions , as was shown in @xcite , sect .
observe also that there is a relationship not only between the periods of @xmath218 and @xmath128 , but also between the amplitudes of oscillation .
if @xmath128 is @xmath70-periodic , the @xmath87-th fourier coefficients can be expressed in terms of occupation numbers : @xmath232 this relation generalizes poisson s summation formula , which shows that the fourier coefficients of @xmath128 in the filled landau level ( @xmath67 and @xmath233 ) are @xmath234 . in this subsection
we show that laughlin s wave function admits a representation as a sum over certain partitions of @xmath80 , analogous to lemma [ solvpoly : lem ] for the solvable model . instead of only monomers or dimers ,
the partition may contain longer polymers or `` rods '' @xmath235 .
the normalization becomes a discrete polymer partition function @xcite with translationally invariant activity and satisfies a discrete renewal equation , as explained at the end of this subsection .
the expansion ( [ polymer - function ] ) for @xmath155 particles and @xmath72 is explicitly written down in @xcite , eq .
( 10 ) , up to a small difference : in @xcite , the sum is over partitions of the set @xmath242 instead of @xmath136 , and rods @xmath237 have length @xmath243 instead of @xmath244 as is the case here .
we will see that the family of `` polymer functions '' ( @xmath167 ) has two important additional properties .
the first is a kind of localization : @xmath167 is a sum of wedge products of functions @xmath209 with indices @xmath87 in @xmath245 .
the second property is translational covariance : @xmath246 thus the shift of a rod results in the magnetic translation of the corresponding function .
this covariance is at the origin of @xmath70-periodicity in laughlin s state .
( [ polymer : eq2 ] ) says that the normalization @xmath81 is a polymer partition function with activity @xmath247 . as a consequence of the covariance ( [ covar : eq ] ) ,
the activity is translationally invariant , i.e. , the activity of a rod depends on its length @xmath236 only : we can define non - negative numbers @xmath248 by refers to integration over the _
cylinder @xmath249 $ ] . on finite cylinders , the activity is not completely translationally invariant due to boundary effects , although prop . [
polymer : prop ] and the covariance ( [ covar : eq ] ) stay true . ]
@xmath250 the numbers @xmath251 will appear as coefficients in a recurrence relation for the normalization @xmath81 .
* @xmath170 is * @xmath5-admissible * if @xmath254 for all @xmath255 with equality for @xmath256 .
* @xmath257 is a * renewal point * of @xmath170 if @xmath258 * @xmath170 is * reducible * if it admits a renewal point @xmath259 , and * irreducible * in the opposite case .
note that @xmath5-admissibility actually implies @xmath260 .
the reducibility of a sequence is most easily visualized with the help of the sequence of occupation numbers in laughlins state , see @xcite , sect .
there are various equivalent characterizations of @xmath5-admissibility @xcite . in all of them , the so - called tao - thouless configuration @xcite @xmath261 plays the role of a reference configuration . in the proof of the next lemma
, we will use the following : a vector @xmath170 is @xmath5-admissible if and only if it is _ majorized _ by @xmath262 , which is equivalent to the existence of a doubly stochastic matrix @xmath53 such that @xmath263 see e.g. @xcite , p. 49 .
this characterization allows us to visualize renewal points as block - diagonal matrices : if @xmath170 is given by eq , ( [ mpt ] ) for some block - diagonal , doubly stochastic matrix @xmath53 that has an upper left @xmath264 block , @xmath87 is a renewal point of @xmath170 .
conversely , if @xmath87 is a renewal point of the @xmath5-admissible vector @xmath170 , every doubly stochastic matrix fulfilling eq .
( [ mpt ] ) must be block - diagonal . by the birkhoff - von neumann theorem ,
a matrix is doubly stochastic if and only if it is a convex combination of permutation matrices .
interestingly , the only matrices that we shall need are equally weighted averages of permutation matrices , see eq .
( [ stochmat ] ) below . 1 .
suppose @xmath265 .
then @xmath170 is @xmath5-admissible .
2 . suppose @xmath266 is a renewal point of @xmath170 .
then : 1 .
@xmath170 is @xmath5-admissible if and only if @xmath267 and @xmath268 are @xmath87- resp .
the following * product rule * holds : @xmath270 the previous lemma holds true if @xmath274 is replaced with the coefficients @xmath275 defined in eq .
( [ vandermonde : eq ] ) .
the corresponding statements for @xmath275 have been proved in @xcite , and the statements for @xmath274 may be deduced with the help of eq .
( [ abrelation : eq ] ) .
the proof given in @xcite uses the polynomial character of the vandermonde determinant .
however , in view of the results on the solvable model ( sect .
[ solv : sec ] ) , we should like to stress the importance of the fact that we have the power of a determinant .
therefore we give a new proof .
for the sake of simplicity , we take @xmath72 ; the proof for other values of @xmath1 is similar . _ 1 . _
expanding the right - hand side of @xmath276 we find @xmath277 where the short - hand @xmath278 in the subscript stands for @xmath279 now suppose @xmath280 .
then there exist permutations @xmath281 adding up to @xmath170 in the sense of the previous equation . in terms of the associated permutation matrices , @xmath282 where @xmath262 is as in eq .
( [ taothoum ] ) with @xmath72 .
@xmath53 is doubly stochastic and therefore @xmath170 is @xmath5-admissible . since by eq .
( [ abrelation : eq ] ) @xmath283 if and only if @xmath280 , the first statement of the lemma follows .
_ let @xmath284 with @xmath253 . and
let @xmath87 be a renewal point of @xmath170 .
_ 2.(a ) _ is proven in @xcite . for _
2.(b ) _ , suppose @xmath281 are permutations such that eq .
( [ stochmat ] ) holds .
then the matrix @xmath53 from this equation must be block - diagonal with an upper left @xmath264 block .
this means that @xmath285 must leave @xmath286 and @xmath287 invariant . as a consequence ,
the sum ( [ sumbn ] ) factorizes into two sums @xmath288 the first sum is over permutations of @xmath286 such that @xmath289 and equals @xmath290 .
the second sum is over permutations of @xmath287 such that @xmath291 rewriting this equation with permutations of @xmath292 instead of @xmath287 , one finds that the second sum equals @xmath293 .
the product rule thus holds for @xmath193 .
a straight - forward computation on the exponential factor in ( [ abrelation : eq ] ) allows us to conclude that @xmath192 satisfies the product rule too . in the expansion ( [ lauexp : eq ] ) , we group together admissible @xmath170 s that have the same set of renewal points @xmath294 .
the product rule ( [ prodrule : eq ] ) gives @xmath295 the renewal points split @xmath170 into @xmath296 irreducible blocks .
this motivates the definition of polymer functions @xmath167 as sums over irreducible sequences : @xmath297 these functions fulfill the covariance ( [ covar : eq ] ) . combining the expansion ( [ lauexp : eq ] ) and eq .
( [ wfprod : eq ] ) , we obtain @xmath298 which concludes the proof .
the results presented in this article rely heavily on the relationship of @xmath58 and polymer systems presented in prop . [ polymer : prop ] .
let us recall some basic facts on polymer systems with translationally invariant activity supported by rods @xmath299 and the link to renewal equations .
in view of eq .
( [ activity : eq ] ) , the expression ( [ polymer : eq2 ] ) for the polymer partition function @xmath81 becomes @xmath300 it follows that @xmath301 .
this supermultiplicativity is a general property of polymer partition functions , see @xcite . as a consequence , we can define @xmath302 where we use the convention @xmath303 .
the left - hand side @xmath304 is the _ pressure _ of the polymer system ; note that @xmath152 is the radius of convergence of the power series @xmath305 . if we suppose in addition that there exists some @xmath306 such that the rescaled activity @xmath307 is _ stable _ , i.e. , @xmath308 the pressure @xmath304 is finite .
this follows from general theorems @xcite but can be shown here using a particularly simple argument : the stability condition ( [ stability : ineq ] ) is equivalent to @xmath309 .
it is fulfilled for some @xmath306 if and only if the power series @xmath310 has a nonvanishing radius of convergence . on the other hand , eq .
( [ normalization : eq ] ) leads to the formal power series identity @xmath311 therefore , if @xmath312 has a positive radius of convergence , so has @xmath305 , whence @xmath313 and @xmath314 .
the type of polymer systems envisaged here bears an interesting connection to renewal theory , as observed by @xcite ( see @xcite for an account on renewal theory ) .
the formal power series identity ( [ powerseries : eq ] ) is equivalent to the recurrence relation @xmath315 this recurrence relation is known in stochastics as a ( discrete ) _ renewal equation_. if we interpret @xmath316 as a discrete time axis and starting points of rods as events , the activity @xmath251 is related to a probability distribution on waiting times between events , and @xmath317 is related to the probability that an event occurs ( i.e. , that @xmath318 is a renewal point ) .
more precisely , we suppose that @xmath313 and set @xmath319 and @xmath320 .
[ assocrenew : page ]
suppose that the sequence of activities is _ aperiodic _ , i.e. , @xmath321 then one of the following cases necessarily holds : 1 .
@xmath322 defines a probability measure on @xmath323 with finite mean @xmath222 . in this case
( the associated renewal process is positive recurrent . )
@xmath322 defines a probability measure on @xmath323 with infinite mean @xmath325
. then @xmath326 ( but @xmath327 ) . ( the renewal process is null recurrent . )
@xmath322 defines a defective measure on @xmath323 ( i.e. , @xmath328 )
. then @xmath329 , whence @xmath330 .
( the renewal process is transient . )
see @xcite for a proof .
we will refer to the first case as a renewal process with _
finite mean_. thus in any case , @xmath331 , but @xmath332
if and only if the associated renewal process has finite mean . in this subsection , we exploit the relation of @xmath58 with polymer systems and renewal equation to investigate the asymptotics of the normalization constant @xmath64 .
the result comes in two parts : first , for all values of the radius , there exist a @xmath313 and @xmath333 such that @xmath334 ( lemma [ normalization : lem ] ) .
second , for sufficiently large @xmath45 ( thin cylinders ) , @xmath335 is strictly positive ( theorem [ laga : theo ] ) .
thus the associated polymer system , suitably rescaled , has a stable activity in the sense of eq .
( [ stability : ineq ] ) , and on sufficiently thin cylinders , the associated renewal process has finite mean .
in addition , we give lower and upper bounds on @xmath152 that are interesting in the context of laughlin s plasma analogy .
our bounds are consistent with results by @xcite on the free energy of a jellium system placed on a cylinder with large radius . we know
already that @xmath340 exists , see ( [ pressure : eq ] ) .
the strict positivity of @xmath152 will follow from ineq .
( [ normaliobo : eq ] ) .
the observation of the previous subsection yields the existence of @xmath341 ; note that @xmath342 so that the sequence of activities @xmath251 fulfills the aperiodicity condition ( [ aperiodicity : eq ] ) .
thus it remains to prove ineq .
( [ normaliobo : eq ] ) .
the idea is to use the representation ( [ powdet : eq ] ) of @xmath58 as the @xmath1-th power of a determinant times @xmath343 and to give lower and upper bounds on @xmath81 using hlder s and hadamard s inequalities .
these inequalities have already been used in @xcite , sect .
3.3 . , to derive bounds on the free energy of jellium on a sphere .
we start with the application of hadamard s inequality , which gives @xmath344 it follows that @xmath345 the integral from @xmath346 to @xmath347 and from @xmath348 to @xmath349 can be bounded by an @xmath5-independent constant .
the integral from @xmath347 to @xmath348 is bounded from above by @xmath350 representing @xmath42 as a fourier series via poisson s summation formula , we find that the first expression in ( [ midint : eq ] ) equals @xmath351 with @xmath352 thus we get @xmath353 , from which the lower bound on @xmath152 is easily obtained .
now we turn to a lower bound for @xmath81 . with hlder s inequality written as @xmath354 applied to the domain of integration @xmath355\times [ 0,2\pi \gamma^{-1}])^n$ ] , we find @xmath356 where @xmath357/2 $ ] and @xmath358 is the complementary error function .
the product over @xmath87 does not contribute to @xmath359 .
making use of stirling s formula , we obtain the desired upper bound to @xmath152 .
lemma [ normalization : lem ] leaves open the question whether @xmath332 or @xmath362 , i.e. , whether the associated renewal process has finite or infinite mean . in order to answer this question , it is useful to have a closer look at the activity @xmath251 .
[ actibo : lem ] monomers have activity @xmath363 .
the activity of a polymer of length @xmath364 is a polynomial of @xmath365 with minimal degree @xmath366 and coefficient in @xmath367 . in particular , @xmath368 the monomer functions are @xmath369 , whence @xmath363 . combining eqs .
( [ abrelation : eq ] ) and ( [ defpolyfunctions : eq ] ) , we find @xmath370 by eq .
( [ bnsigns : eq ] ) , the expansion coefficients @xmath193 are sums of signs of permutations and therefore integers .
the proof of the lemma is concluded by the following observation : if @xmath371 is @xmath5-admissible and irreducible , @xmath372 due to @xmath5-admissibility , we can write @xmath373 with @xmath374 and @xmath375 ( see also @xcite , property 3 ) .
@xmath376 is just @xmath377 $ ] .
because of irreducibility , @xmath378 must be strictly positive . in the left - hand side of ( [ sumsquares : ineq ] ) , we insert the expression of @xmath379 in terms of @xmath376 and perform a summation by parts , and obtain ineq .
( [ sumsquares : ineq ] ) .
now recall that @xmath332 if and only if @xmath380 and @xmath381 .
the crucial observation is that these two conditions are automatically fulfilled when the generating series of @xmath251 has a radius of convergence @xmath382 strictly larger than the radius of convergence @xmath152 of the power series with coefficients @xmath383 .
therefore we are going to compare domains of convergence . for a monomer system with @xmath363 ( and @xmath384 for @xmath385 ) , the quantities are trivial to compute : @xmath386 we will show that on sufficiently thin cylinders , the quantities @xmath387 take values close to the monomer values ( [ monomervalues ] ) . for this purpose
it is useful to keep track of the @xmath45-dependence in the notation . by lemma [ actibo : lem ] and eq .
( [ normalization : eq ] ) , the activity and the normalization constants are polynomials of @xmath388 with coefficients in @xmath323
. therefore we write @xmath389 , @xmath390 .
the power series @xmath391 are actually power series of _ two _ variables , @xmath392 and @xmath393 .
they have non - negative integer coefficients and are related through @xmath394 see eq .
( [ powerseries : eq ] ) .
the curves @xmath395 and @xmath396 delimit the domains of convergence of @xmath397 and @xmath398 , see fig .
[ domco_fig ] .
the following theorem states that the curve @xmath399 stays strictly below @xmath400 , at least for small @xmath401 ( large @xmath45 ) , and @xmath399 and @xmath402 converge to the monomer values @xmath403 in the limit of thin cylinders ( @xmath404 ) .
_ let @xmath412 . by lemma [ actibo : lem ] ,
@xmath413 with suitable non - negative integers @xmath414 .
it follows that @xmath415 .
we fix @xmath416 and let @xmath417 . since @xmath418 , we obtain that @xmath419 goes to infinity when @xmath404 , as expected from eq .
( [ monomervalues ] ) . on the other hand ,
we know that @xmath420 , hence @xmath421 .
thus for sufficiently small @xmath422 , @xmath423 .
the positivity of @xmath152 was proved in lemma [ normalization : lem ] .
the positivity of @xmath335 is a consequence of @xmath409 .
now , notice that the power series @xmath398 defines a holomorphic function of two complex variables in the domain @xmath424 .
for @xmath425 , @xmath399 is the unique solution of @xmath426 by lemma [ actibo : lem ] , @xmath427 .
thus the thin cylinder limit corresponds to the point @xmath428 .
we can apply an implicit function theorem for holomorphic functions to obtain the analyticity of @xmath152 .
the analyticity of @xmath335 follows from @xmath429 both @xmath399 and @xmath402 can be extended to holomorphic functions in a complex neighborhood of @xmath430 and take the value @xmath19 at @xmath46 , whence eq .
( [ laga : eq ] ) . in this subsection ,
we show that laughlin s state has a unique thermodynamic limit ( theorem [ thermolim : theo ] ) and show that the limiting state is periodic in the axial direction , with @xmath70 as one of its periods . the proof that @xmath70 is actually the smallest period is deferred to the next subsection .
the results presented here hold provided @xmath432 , i.e. , the associated renewal process has finite mean . from the previous subsection , we know that this condition is indeed fulfilled on sufficiently thin cylinders . in the following , @xmath433 is the @xmath434-algebra generated by the fermionic creation and annihilation operators @xmath435 , with @xmath436)$ ] .
the operators associated with the lowest landau level basis state @xmath209 are denoted @xmath437 .
[ existence of the thermodynamic limit ] [ thermolim : theo ] suppose @xmath438 .
there is a state @xmath439 such that for every sequence of integers @xmath440 such that @xmath441 and @xmath442 , the states associated with the shifted laughlin functions @xmath443 converge to @xmath439 : for all @xmath444 , @xmath445 for @xmath446 , let @xmath447 and @xmath448 .
it is enough to prove the convergence ( [ cstarlim ] ) for operators @xmath449 with @xmath450 .
the key idea of the proof is to show a formula similar to the one given in prop .
[ modfunct : prop ] for the solvable model , and then to use the asymptotics of the normalization constants , just as we did in the proof of corollary [ thermolim : cor ] .
let @xmath451 .
we will see that @xmath452 can be written as @xmath453 for a suitable @xmath5-independent family of functions @xmath454 .
we use the notation @xmath455 the @xmath456 s have finite support , @xmath457 and non - negative `` diagonal '' values : @xmath458 for fixed @xmath459 and @xmath318 , @xmath460 due to @xmath461 . thus formally , the right - hand side of eq .
( [ repo ] ) converges to @xmath462 provided the series converges .
we prove the theorem in three steps .
first , we define the auxiliary functions @xmath456 and prove the representation ( [ repo ] ) of the correlation functions .
second , we look at `` diagonal '' correlation functions ( @xmath463 ) and show that the series ( [ repolim ] ) is bounded and equals the limit of correlation functions : @xmath464 as a last step , we turn to off - diagonal values ( @xmath465 ) . we prove that the series ( [ repolim ] ) is absolutely convergent : @xmath466 and show that @xmath467 .
note that once we know that eq .
( [ repolim ] ) defines a state on @xmath433 , the inequality ( [ causs : ineq ] ) with absolute value bars _ outside _ the sum is just cauchy - schwarz for the state @xmath439 .
representation of correlation functions .
_ let @xmath468 with @xmath450 .
we start with the representation @xmath469 the sum ranges over @xmath5-admissible sequences @xmath470 .
suppose @xmath170 and @xmath471 have common renewal points @xmath472 such that @xmath473 . then @xmath474 let @xmath475 be the set of pairs @xmath476 such that by `` @xmath477 is below @xmath478 ''
we mean @xmath479 , and we say
@xmath477 is above @xmath480
when @xmath481 .
the set @xmath475 consists of the pairs @xmath476 for which no simplification of the type ( [ splitone ] ) is possible .
@xmath482 is defined by the sum ( [ sufo ] ) , except that the summation includes only @xmath476 from @xmath475 . with this definition , combining ( [ sufo ] ) and ( [ splitone ] ) we obtain ( [ repo ] ) .
+ _ 2 . `` diagonal '' correlation functions . _ for @xmath483 , the definition of @xmath456 gives @xmath484 where the sum ranges over @xmath5-admissible sequences that are @xmath47-irreducible .
in particular , @xmath485 .
moreover , if @xmath486 , there exists an @xmath5-admissible sequence @xmath170 such that @xmath487 , whence @xmath488 .
let @xmath489 . from ( [ repo ] ) and @xmath490 we get @xmath491 if @xmath492 we must have @xmath493 , thus only a finite , @xmath5-independent number of @xmath459 s contribute to the sum and we can take the limit @xmath79 , which gives @xmath494 letting @xmath495 , we obtain the bound @xmath496 . the proof of @xmath497
is then completed by an @xmath498 argument .
we leave the details to the reader and mention only a useful inequality on quotients of normalization constants . using the supermultiplicativity of @xmath499 , @xmath500 and @xmath501 , we get @xmath502 and @xmath503 _ 3 .
`` off - diagonal '' correlation functions ( @xmath504 ) . _
the procedure is similar to step _ 2 . _ but the analogue of the bound ( [ eq : step2bound ] ) is slightly more delicate to obtain . let @xmath489 .
then @xmath505 the notation @xmath506 refers to the amplitude of the increasing sequence obtained by rearranging the elements of @xmath507 .
letting first @xmath5 and then @xmath508 go to infinity , we obtain the bound ( [ causs : ineq ] ) .
the convergence @xmath509 can be shown with an @xmath498 argument .
[ rem : proba ] the representation ( [ repolim ] ) of correlation functions does not lend itself to a simple interpretation .
however , it leads to a very nice formula for the one particle density .
let @xmath510 be the number operator for the lowest landau level state @xmath209 .
the quantity @xmath511 equals @xmath512 with the polymer @xmath513 .
therefore eq .
( [ repolim ] ) may be rewritten as @xmath514 with @xmath515 and @xmath516 .
the sum is over all polymers @xmath235 , @xmath517 , @xmath518 .
this formula has an intuitive probabilistic interpretation : @xmath519 is the probability of finding a particle in the `` site '' @xmath87 , given that @xmath87 is contained in the polymer @xmath237 ( or , strictly speaking , in @xmath520 ) , which happens with probability @xmath521 refers to polymer correlation functions as defined in @xcite .
later , we shall use not only @xmath521 but also @xmath522 , see p. . ] . now let us turn to the symmetries of laughlin s state .
let @xmath524 be the automorphism of the algebra @xmath433 associated with the magnetic translation @xmath525 .
let @xmath526 be the morphism associated with the translation @xmath527 in the @xmath14-direction , and @xmath528 the morphism induced by the reversal @xmath529 .
[ symmetries : prop ] the state @xmath530 of the previous theorem is invariant with respect to reversal , translations in the @xmath11-direction by multiples of @xmath531 , and arbitrary translations in the @xmath14-direction : @xmath532 the invariance with respect to @xmath14-translations is a direct consequence of the fact that @xmath58 has a definite @xmath14-momentum , see eq .
( [ yinvariance : eq ] ) .
the reversal invariance follows from the invariance for finitely many particles ( [ reversalinv ] ) , see also @xcite .
the periodicity with respect to magnetic translations in the direction along the cylinder follows from the representation ( [ repolim ] ) of correlation functions .
[ cor : dens ] let @xmath114 be the one - particle density of laughlin s state @xmath58 . under the assumptions of theorem [ thermolim : theo ] ,
the shifted density converges pointwise to the one - particle density @xmath533 of the limiting state @xmath439 : @xmath534 the density is independent of the coordinate @xmath535 around the cylinder .
the density as well as the occupation numbers are periodic and reversal invariant : @xmath536 note that weak@xmath537-convergence of the state @xmath205 is replaced with pointwise convergence of the one - particle density .
this uses the representation of the density as a sum of gaussians with occupation numbers as coefficients as in eq .
( [ density : eq ] ) . due to the good localization of the gaussians , summation and limits
can be interchanged , whence eq .
( [ densitysum ] ) .
this subsection contains the second part of the main results of this paper : theorem [ symbreak : theo ] shows that on sufficiently thin cylinders , @xmath70 is actually the _ smallest _ period of the limiting state @xmath439 as well as the one - particle density @xmath2 of the previous subsection .
thus the state @xmath538 has a larger minimal period than the hamiltonian describing interacting electrons in a magnetic field , whose ground states it is supposed to approximate . in this sense , there is symmetry breaking .
[ symmetry breaking][symbreak : theo ] suppose @xmath539 .
let @xmath128 be the infinite cylinder density from cor .
[ cor : dens ] . then on sufficiently thin cylinders , @xmath70 is the smallest period of @xmath128 . due to
( [ densitysum ] ) and lemma [ foutrans : lem ] , it is enough to look at the occupation numbers . we will show that @xmath540 thus for sufficiently large @xmath45 , the sequence of occupation numbers has @xmath1 as the smallest period and @xmath70 is the smallest period of the one - particle density .
the idea behind ( [ densasy : eq ] ) is that the thin cylinder limit is at the same time a monomer limit , see lemma [ actibo : lem ] and p. .
the wave function corresponding to a pure monomer system , for @xmath5 particles , is @xmath541 in the limit @xmath79 , the corresponding monomer occupation numbers @xmath542 equal @xmath19 if @xmath87 is a multiple of @xmath1 , and @xmath46 otherwise .
( [ densasy : eq ] ) now is a consequence of the following observation : the occupation numbers @xmath543 are functions of @xmath544 that can be extended to holomorphic functions of @xmath416 in a complex neighborhood of the monomer point @xmath545 .
this can be shown with the representation @xmath546 see eq .
( [ repolim ] ) .
@xmath547 is a polynomial of @xmath365 , and @xmath335 and @xmath152 are analytic functions of @xmath365 .
we can adapt the procedure used in the proof of theorem [ laga : theo ] and deduce the analyticity of @xmath543 for small @xmath401 .
_ the monomer state ( [ taothoufunct ] ) is the tao - thouless state , corresponding to the reference configuration @xmath262 on p. .
the fact that laughlin s wave function for a fixed , _
finite _ number of particles on very thin cylinders approaches the tao - thouless state has been observed by rezayi and haldane @xcite .
the novelty here is twofold : first , the limits @xmath79 and @xmath548 can be interchanged ; second , the periodicity survives for small but non - vanishing cylinder radius .
_ if we assimilate orbitals @xmath209 with lattice sites @xmath134 , the restriction of the state @xmath549 to the algebra generated by the number operators @xmath550 can be described by a probability distribution @xmath53 on particle configurations on @xmath316 .
adapting techniques from @xcite , one can show that the probability measures corresponding to @xmath549 and the shifted states @xmath551, .. ,@xmath552 are mutually singular .
this result holds provided @xmath438 and the second moment @xmath553 is finite .
again , this condition is fulfilled when @xmath45 is large enough . as a consequence , the @xmath1 quantum - mechanical states @xmath549 , .. , @xmath552 are not only _ distinct _ , but also _ orthogonal _ in the sense of @xcite , def .
4.1.20 .
now we come to clustering properties . before we state our result in its general form ,
let us have a look at two - point correlations @xmath554 , where @xmath555 is the number operator for the state @xmath209 . in the spirit of the remark on p.
, @xmath554 may be interpreted as the probability of finding a particle in the site @xmath87 and another particle in the site @xmath556 . in fact
, we have a formula analogous to eq .
( [ eq : densproba ] ) for the one - particle density .
define @xmath521 and @xmath519 as on p. .
let @xmath557 and @xmath558 where @xmath237 is to the left hand side of @xmath559 , separated from @xmath559 by the distance @xmath560 @xmath561 is the probability of finding particles in the sites @xmath87 and @xmath556 given that @xmath87 and @xmath556 are contained in @xmath562 $ ] , while @xmath522 is the probability of finding the rods @xmath237 and @xmath559 .
suppose that @xmath563 .
then the diagonal two - point correlation equals @xmath564 again , this formula has an intuitive probabilistic interpretation .
the two sums correspond to the two different situations that @xmath87 and @xmath556 are contained in two distinct polymers ( first sum ) or in the same polymer ( second sum ) .
when @xmath87 and @xmath556 are far apart , the probability that they are in the same polymer is small . on the other hand ,
when @xmath237 and @xmath559 are far apart , we may write @xmath565 therefore we expect that @xmath566 is approximately the same as @xmath567 when @xmath87 and @xmath556 are far apart : laughlin s state inherits clustering properties from the polymer correlations @xmath568 .
we use the notation from the proof of theorem [ thermolim : theo ] .
it is enough to check ( [ clus ] ) for operators @xmath571 , @xmath572 with @xmath573 . because of particle number conservation and @xmath14-invariance , the only interesting case is @xmath574 in the following we will assume that ( [ clunotri : eq ] ) holds and show that @xmath575 is small when @xmath576 is far to the left of @xmath577 .
the main idea is to generalize the formula ( [ eq : probacorr ] ) for two - point correlations .
we will see that @xmath578 where @xmath579 and the functions @xmath580 , @xmath581 will be defined later .
similarly , @xmath582 the theorem is proved by making the following arguments precise : suppose @xmath576 and @xmath577 are far away .
then , intuitively , the main contributions to @xmath583 in ( [ fg : eq ] ) come from polymers @xmath584 separated by a large distance @xmath585 . since @xmath586 , we expect that @xmath583 is close to the righthand side of eq .
( [ cluprod : eq ] ) . the second contribution , @xmath587 , in eq .
( [ fg : eq ] ) will be bounded by the probability for finding a long polymer , which is small .
thus @xmath588 now we define @xmath580 and @xmath581 , and prove eqs .
( [ fg : eq ] ) and ( [ cluprod : eq ] ) .
the argument resembles step _ 1 .
_ in the proof of theorem [ thermolim : theo ] , therefore we shall only give the key elements .
let @xmath589 be the set of pairs @xmath476 such that : we say that @xmath477 is between @xmath478 and @xmath592 if @xmath576 is contained in @xmath593 and @xmath577 in @xmath594 .
the set @xmath589 is a subset of the set @xmath475 introduced on p. for the definition of @xmath595 .
let @xmath596 the product rule from lemma [ prodrule : lem ] gives @xmath597 for @xmath598 , it is convenient to define @xmath599 eqs .
( [ fg : eq ] ) and ( [ cluprod : eq ] ) hold with these definitions of @xmath580 , @xmath581 , as can be seen with eqs .
( [ repolim ] ) and ( [ gclus : eq ] ) .
next , we estimate @xmath600 . in view of eqs .
( [ fg : eq ] ) and ( [ cluprod : eq ] ) , this difference is a sum over polymers @xmath584 : @xmath601 with the convention @xmath602 for @xmath603 .
we split the sum in two parts .
first , suppose that @xmath237 is far to the left of @xmath559 , i.e. , @xmath604 for some fixed @xmath605 .
the sum over such @xmath606 is bounded by @xmath607 where we have used ineq .
( [ causs : ineq ] ) .
second , suppose that @xmath608 satisfy @xmath609 and @xmath610
. then @xmath237 or @xmath559 must be a long polymer , see fig .
[ fig : clust1 ] . more precisely , ( [ support ] ) can be used to show that @xmath611 must hold .
the sum ( [ fsum ] ) over pairs @xmath584 such that @xmath237 satisfies ineq . ( [ ineq : cond1 ] ) is bounded by @xmath612 where the prime at the sum refers to the constraint ( [ ineq : cond1 ] ) . by a procedure similar to step
_ of the proof of theorem [ thermolim : theo ] , one can show that @xmath613 the right - hand side represents the probability that a given point is in a polymer of length greater or equal to @xmath614 .
the sum ( [ fsum ] ) over pairs @xmath606 such that @xmath559 satisfies ineq . ( [ ineq : cond2 ] ) can be treated in a similar way . in the end , we obtain @xmath615 choosing @xmath616 of the order of @xmath617 , we see that @xmath618 goes to @xmath46 as @xmath508 goes to infinity .
it remains to give a bound on @xmath587 .
the idea is to have a closer look at the definition of @xmath619 and to show that if @xmath470 give a non - vanishing contribution to the sum , at least one of the vectors has no renewal point between @xmath478 and @xmath592 , see fig .
[ fig : clust2 ] .
this is stronger than the original requirement that @xmath170 and @xmath471 have no _ common _ renewal point between @xmath478 and @xmath592 . as a consequence , one
can bound @xmath587 by the square root of the probability that the space between @xmath592 and @xmath478 is covered by one single long polymer .
the square root comes from a cauchy - schwarz inequality and accounts for the fact that one of the sequences @xmath470 may have a renewal point .
we start by looking at the non - vanishing contributions to @xmath620 . without loss of generality
, we may assume @xmath621 ( otherwise , use @xmath622 to interchange @xmath623 and @xmath47 ) .
suppose that @xmath624 are @xmath5-admissible and @xmath625 we claim that if @xmath477 is a renewal point of @xmath471 such that @xmath626 @xmath477 is also a renewal point of @xmath170 . to see this , let @xmath627 , similarly for @xmath628 .
then we must have @xmath629 intersecting with @xmath593 , we obtain @xmath630 in particular , @xmath631 thus @xmath632 the inequality on the right - hand side must be an equality , showing that @xmath477 is a renewal point of @xmath170 . now suppose @xmath624 give a non - vanishing contribution to the sum defining @xmath619 .
they are by definition not allowed to have a common renewal point between @xmath478 and @xmath592 . as we have just shown , this means that @xmath471 can not have a renewal point between @xmath633 and @xmath592 .
thus if @xmath634 is the partition of @xmath80 determined by the renewal points of @xmath471 , there is a rod @xmath635 such that @xmath636 $ ] intersects both @xmath637 and @xmath638 .
it follows that @xmath639 in the spirit of step _ 3 .
_ in the proof of theorem [ thermolim : theo ] , we have @xmath640 where @xmath641 is the sum over sequences @xmath471 having no renewal point between @xmath478 and @xmath642 . taking the limit @xmath79 after shifting the origin to the middle of the cylinder , we get @xmath643 _ remark : rate of convergence . _ on sufficiently thin cylinders , we know from theorem [ laga : theo ] that @xmath310 has a radius of convergence @xmath382 strictly larger than @xmath152 . in this case @xmath644 and ineqs .
( [ ineq : fab ] ) and ( [ ineq : gab ] ) show that there is actually _ exponential _ clustering .
thanks the alexander von humboldt foundation for a research award , based at the technical university , berlin .
he also thanks the us national science foundation for partial support , grant phy-0652854 .
s.j . and r.s .
thank the dfg for support under grant no . se 456/7 - 1 . | we investigate laughlin s fractional quantum hall effect wave function on a cylinder .
we show that it displays translational symmetry breaking in the axial direction for sufficiently thin cylinders . at filling factor @xmath0 ,
the period is @xmath1 times the period of the filled lowest landau level .
the proof uses a connection with one - dimensional polymer systems and discrete renewal equations . |
the blazar 1156 + 295 ( @xmath0 , @xcite ) is among the most active of highly polarized ( hpq ) and optically violent variable ( ovv ) sources @xcite in its active phase , this source has shown flux - density fluctuations at optical wavelengths with an amplitude of @xmath1 57% on time scales of 0.5 hour . also , both the position angle and fraction of optical polarization vary dramatically . at cm wavelengths @xmath2 variations in total flux density have been observed on time scales less than several days .
finally , 1156 + 295 has flared in @xmath3 100 mev @xmath4-rays at least three times since 1992 , while the quiescent @xmath4-ray emission remains undetected @xcite .
high - resolution radio imaging has revealed arcsecond and milliarcsecond structure consistent with the activity described above as well as with the standard blazar model of a jet aligned nearly along the line of sight . on the arcsecond scale ,
the 1.5 ghz vla image shows a symmetrical structure in the north - south direction @xcite .
the merlin image at 1.6 ghz and the vla image at 5 ghz @xcite show extended emission on the arcsecond - scale and a knotty jet of length @xmath5 at p.a .
the northern end of the jet has a hotspot , at which point it turns through about @xmath7 to the east , ending in a diffuse region of size @xmath8 .
there is a region of similar diffuse emission to the south about @xmath9 from the core , at p.a .
@xmath10 . from the maps ,
1156 + 295 appears to resemble a classical double radio source seen end - on , with its northern jet relativistically beamed towards us , and hence dominating over any southern jet , while the diffuse emission is not beamed and appears symmetric . based on vlbi observations at three frequencies
, @xcite estimated the apparent superluminal velocity of @xmath11 .
this is much larger than any proper motion reliably found for any other source @xmath12 times greater than that of the next fastest object in its redshift bin on the @xmath13 diagram @xcite .
such a high apparent proper motion in principle could be a projection effect : when the jet axis is aligned near the line of sight , small changes in the direction of the jet axis can produce large changes in the apparent proper motion .
lower apparent velocities in the range of 3.5@xmath14 were reported on the basis of ten epochs of geodetic observations at 8.4 and 2.3 ghz @xcite .
arguments against highly beamed synchrotron emission from the milliarcsecond core ( thus , not necessarily a small angle between the very inner part of the jet and viewing directions ) have been reported on the basis of the vsop space vlbi observations at 1.6 ghz with the angular resolution of 4.4 @xmath15 1.4 mas @xcite . in this paper
, we discuss the observations of the blazar 1156 + 295 at 5 ghz with the evn + merlin ( february 1997 ) in comparison with the earlier vlba observation . throughout this paper , the values @xmath16 kms@xmath17mpc@xmath17 , @xmath18 , and @xmath19
will be used .
a full - track 12-hour evn + merlin observation of the blazar 1156 + 295 was carried out at 5 ghz in february 1997 . the merlin array comprised 6 antennas ( defford , cambridge , knockin , darnhall , mark 2 , and tabley ) .
the central wavelength was 4994 mhz and the observing bandwidth was 14 mhz .
the amplitude calibration was carried out at jodrell bank with the calibrator source oq208 .
the imaging was done using the nrao astronomical image processing system ( aips ) package . the evn array comprised effelsberg , wsrt , jodrell bank ( mark 2 ) , cambridge , onsala , medicina , torun , shanghai , urumqi , and hartebeesthoek .
the data were acquired with the mkiii vlbi recording system in mode b with an effective bandwidth of 28 mhz and correlated at the max - plank institut fr radioastronomie in bonn .
no fringes were found on the baselines to urumqi .
the evn data were calibrated and fringe - fitted using the nrao aips package .
the initial amplitude calibration was accomplished using the system temperature measurements made during the observations and the a priori station gain curves .
the snapshot - mode observation of 1156 + 295 with the vlba at 5 ghz was made in june 1996 as part of the vsop prelaunch vlba survey @xcite .
eight intermediate frequency channels , each 8mhz wide , were recorded for a total bandwidth of 64 mhz .
the data were correlated at the nrao vlba correlator ( socorro , nm , usa ) .
the fringe - fitted and calibrated vlba data were kindly made available to us by the vsop / vlba pre - launch survey group @xcite . the post - processing including editing , phase and amplitude self - calibration , imaging and model fitting
was performed within aips for both the evn and vlba data sets .
our merlin image shows traces of a straight jet at the position angle of @xmath20 and @xmath9 in length ( hong et al . , in preparation ) .
this is consistent with the lower - resolution vla and merlin images reported by @xcite .
there appear to be several regularly spaced knots within 1 arcsecond of the core , possibly indicating quasi - periodic activity that has propagated from the base of the jet .
( @xmath213 , 3 , 6 , 12 , 24 , 48 , 96 , 192 , 384 , 768 , 1536 ) .
the peak flux densities are 1.18 and 2.08 jy / beam , respectively.,title="fig : " ] ( @xmath213 , 3 , 6 , 12 , 24 , 48 , 96 , 192 , 384 , 768 , 1536 ) .
the peak flux densities are 1.18 and 2.08 jy / beam , respectively.,title="fig : " ] because of missing urumqi data , the u - v coverage of the evn data has a big gap between 25 and 90 m@xmath22 .
therefore , we imaged with the data of the european telescopes first .
the image ( fig . [ fig ] ) reveals a jet that is bending from north to north - east before turning back to the north , presumably to meet the arcsecond scale jet .
the image of vlba presents a similar structural pattern ( fig .
[ fig ] ) .
it shows more clearly that the jet is bent from the north to east and then back to the north - east .
a higher resolution evn image was obtained by using all the data , including data from hartebeesthoek and shanghai ( hong et al .
, in preparation ) .
the results show that a new jet component starts at about p.a .
@xmath23 before bending to the north within 2 milliarcseconds of the core , which is consistent with the 8 ghz images @xcite . in piner and kingham s images
, one can also note that the jet components start at the position angle of @xmath24 , then move clockwise to the north before bending to the east sharply .
we note that the peak flux density at two epochs varies from 2.09 jy / beam for the vlba map ( epoch 1996.43 ) to 1.2 jy / beam for the later evn map ( epoch 1997.14 ) , which is clear evidence of the decrease of the core brightness ( the vlba synthesized beam of fig .
[ fig ] is smaller than that of evn , fig .
[ fig ] ) . from the data of the university of michigan radio astronomy observatory ,
the total flux densities at 14.5 , 8.0 , and 4.8 ghz have all decreased from mid of 1996 to early 1997 @xcite . in 15ghz vlba maps ,
the peak flux density decreased from 1.95 jy / beam ( 16 may 1996 ) to 0.67 jy / beam ( 13 march 1997 ) @xcite .
all these indicate clearly that the flare of 1156 + 295 was due to the outburst of the central core component . at that time
a new jet component was ejected .
helical jets have been proposed to explain the bi - modal distribution of the difference between arcsecond and milliarcsecond structural axes observed for core - dominated radio sources @xcite .
the helical pattern could result from the precession of the base of the jet ( e.g. @xcite ) or fluid - dynamical instabilities in the interaction between the jet material and surrounding medium @xcite .
the structural oscillations at mas scales can be explained by the orbital motion of a binary black hole ( e.g. @xcite for the case of 1928 + 738 ) .
the jet of 1156 + 295 has a structural oscillation on the mas scale and a large @xmath25p.a . between its pc - scale and the kpc - scale directions , which may indicate that the oscillation of the jet is a projection effect of a helical jet .
the observed properties of the jet could be explained by two effects : the precession of the spin axis of the black hole emitting the jet and the orbital motion of a binary black hole . in this case , the jet of 1156 + 295 could be ejected in the direction very closely aligned to the line of sight .
the jet then would have a high doppler boosting but would demonstrate a relatively low apparent proper motion . as the direction of jet curves away from the line of sight , the doppler factor will decrease ( thus , flux density will decrease also ) , but the apparent proper motion velocity will increase . as the bright radio component moves
outward , it will reach the viewing angle of the maximum apparent transverse velocity , while its doppler boosting factor will continue to monotonically decrease .
this qualitative model can explain the evolution of the outburst in the core @xcite and the apparent initial acceleration of the proper motion of the jet components @xcite .
it is also generally consistent with the @xmath4-ray observation at the gev band : the source demonstrated several short active periods at the high - energy @xmath4-ray band but remained in a quiescent state most of the time between flares @xcite .
the model suggests that the @xmath4-ray flares may preceed radio outbursts since the former require higher doppler boosting
. * acknowledgements . *
this research was supported by the national science foundation and the pan deng plan of china .
xyh thanks jive for the hospitality during his visit in 1998 .
lig acknowledges partial support from the european comission tmr programme , access to large - scale facilities under contract erbfmgect950012 .
the authors are grateful to the staff of evn , merlin and vlba for support of the observing projects .
the authors express their gratitude to the team of the vsop / vlba pre - launch survey , particularly ed fomalont and phil edwards , for the permission to use their data .
this research has made use of data from the university of michigan radio astronomy observatory , which is supported by the national science foundation and by funds from the university of michigan .
the national radio astronomy observatory is operated by associated universities , inc . under a cooperative agreement with the national science foundation . | the blazar 1156 + 295 was observed by vlba and evn + merlin at 5 ghz in june 1996 and february 1997 respectively .
the results show that the jet of the source has structural oscillations on the milliarcsecond scale and turns through a large angle to the direction of the arcsecond - scale extension .
a helical jet model can explain most of the observed properties of the radio structure in 1156 + 295 .
radio continuum : blazar 95.75.kk , 98.54.cm |
relativistic jets occur in active galactic nuclei ( * agn * ) ( e.g. , urry & padovani 1995 ; ferrari 1998 ; meier et al .
2001 ) , occur in microquasars ( e.g. , mirabel & rodriguez 1999 ) , and are thought responsible for the gamma - ray bursts ( * grb * ) ( e.g. , zhang & mszros 2004 ; piran 2005 ; mszros 2006 ) .
the most promising mechanism for producing relativistic jets involves magnetohydrodynamic acceleration from an accretion disk around a black hole ( e.g. , blandford 1976 ; lovelace 1976 ; blandford & payne 1982 ; fukue 1990 ; meier 2005 ; narayan et al .
2007 ) , and/or involves the extraction of energy from a rotating black hole ( penrose 1969 ; blandford & znajek 1977 ) .
general relativistic magnetohydrodynamic ( * grmhd * ) codes have been used to study the extraction of rotational energy from a spinning black hole , i.e. , blandford - znajek mechanism ( koide 2003 ; komissarov 2005 ; mckinney 2005 ; komissarov & mckinney 2007 ; komissarov & barkov 2009 ) and from an accretion disk , i.e. , blandford - payne mechanism ( blandford & payne 1982 ) .
grmhd codes also probe the formation of grb jets in collapsars ( e.g. , mizuno et al . 2004a , 2004b ; liu et al . 2007 ; barkov & komissarov 2008 ;
stephens et al . 2008 ; nagataki 2009 ; komissarov & barkov 2009 ) and are used to study the propagation of jets injected from an unresolved keplerian disk ( e.g. , komissarov et al .
2007 , 2008 ; tchekhovskoy et al . 2008 , 2009 ) .
grmhd simulations with a spinning black hole indicate jet production with a magnetically dominated high lorentz factor spine with @xmath0 , and a matter dominated sheath with @xmath1 possibly embedded in a lower speed , @xmath2 , disk / coronal wind ( e.g. , de villiers et al . 2003 , 2005 ; hawley & krolik 2006 ; mckinney & gammie 2004 ; mckinney 2006 ; mckinney & narayan 2007a , 2007b ; mckinney & brandford 2008 ; beckwith et al .
2008 ; hardee et al .
circumstantial evidence such as the requirement for large lorentz factors suggested by the tev bl lacs when contrasted with much slower observed motions ( ghisellini et al.2005 ) suggests such a spine - sheath morphology , although alternative interpretations are also possible ( georganopoulos & kazanas 2003 ; bromberg & levinson 2008 ; stern & poutanen 2008 ; giannios et al .
2009 ) .
numerical simulations indicate that the jet spine and sheath are accelerated and , in part , collimated by strong magnetic fields twisted in the rotating black hole ergosphere and in the accretion disk , respectively .
the helically twisted magnetic fields are expected to be dominated by the toroidal component in the far zone as a result of jet expansion because the poloidal component falls off faster with expansion and distance . in configurations with strong toroidal magnetic field
the current driven ( * cd * ) kink mode is unstable .
this instability excites large - scale helical motions that can strongly distort or even disrupt the system . for static cylindrical force - free equilibria ,
the well - known kruskal - shafranov criterion states that the instability develops if the length of the column , @xmath3 , is long enough for the field lines to go around the cylinder at least once ( e.g. bateman 1978 ) : @xmath4 .
this criterion suggests that jets are unstable beyond the alfvn surface because of @xmath5 and @xmath6 at the alfvn surface , where @xmath7 is the cylindrical radius of the jet
. however , rotation and shear motions could significantly affect the instability criterion .
for relativistic force - free configurations , the linear instability criteria were studied by istomin & pariev ( 1994 , 1996 ) , begelman ( 1998 ) , lyubarskii ( 1999 ) , tomimatsu et al .
( 2001 ) and narayan et al .
( 2009 ) .
the linear mode analysis provides conditions for the instability but says little about the impact the instability has on the system .
the instability of the potentially disruptive kink mode found from a linear analysis must be followed into the non - linear regime .
the nonlinear development of the cd kink instability is different for external and internal kink modes ( e.g. , bateman 1978 ) .
external disruptive instability develops in a current carrying plasma column surrounded by a vacuum magnetic field . in astrophysical systems , some amount of plasma is present everywhere . in this case
an internal kink instability develops inside a resonant surface on which the helicity of the magnetic field lines matches that of the eigenmode . in dissipationless mhd , the internal instability is known to saturate .
finite resistivity makes the internal kink mode disruptive but the time - scale significantly exceeds the alfvn time - scale and may be too long to disrupt a fast outflow . in this case
helical structures may be formed in the flow . recently helical structures have been found in non - relativistic / relativistic simulations of magnetized jets ( e.g. , lery et al .
2000 ; ouyed et al . 2003
; nakamura & meier 2004 ; nakamura et al .
2007 ; moll et al . 2008 ; mckinney & blandford 2008 ; carey & sovinec 2009 ) .
twisted structures are observed in many agn jets on sub - parsec , parsec and kiloparsec scales ( e.g. , gmez et al .
2001 ; lobanov & zensus 2001 ) .
dissipation processes may result in relaxation to helical equilibrium as evoked by knigl & choudhuri ( 1985 ) for helical structures in agn jets . in the absence of cd kink instability and resistive relaxation ,
helical structures may be attributed to the kelvin - helmholtz ( * kh * ) helical instability driven by velocity shear at the boundary between the jet and the surrounding medium ( e.g. , hardee 2004 , 2007 ) or triggered by precession of the jet ejection axis ( begelman et al .
it is still not clear whether current driven , velocity shear driven or jet precession is responsible for the observed structures , or whether these different processes are responsible for the observed twisted structures at different spatial scales .
this is the first in a series of papers in which we study kink instability in relativistic systems . by relativistic we mean not only relativistically moving systems but any with magnetic energy density comparable to or greater than the plasma energy density , including the rest mass energy . in this paper
, we present 3d results of the cd kink instability of a static plasma column .
we start from static configurations because in the case of interest , the source of the free energy is the magnetic field , not the kinetic energy .
therefore static configurations ( or more generally rigidly moving flows considered in the proper reference frame ) are the simplest ones for studying the basic properties of the kink instability . at the next stage
, we will investigate the influence of shear motions and rotation on the stability and nonlinear behavior of jets .
this article is organized as follows .
we describe the numerical method and setup used for our simulations in 2 , present our results in 3 , discuss the astrophysical implications in 4 , and present some numerical tests in the appendix .
in order to study time evolution of the cd kink instability in the relativistic mhd ( * rmhd * ) regime , we use the 3d grmhd code `` raishin '' in cartesian coordinates .
raishin is based on a @xmath8 formalism of the general relativistic conservation laws of particle number and energy momentum , maxwell s equations , and ohm s law with no electrical resistance ( ideal mhd condition ) in a curved spacetime ( mizuno et al .
2006 ) . in the raishin code , a conservative , high - resolution shock - capturing scheme is employed .
the numerical fluxes are calculated using the hll approximate riemann solver , and flux - interpolated , constrained transport is used to maintain a divergence - free magnetic field .
the raishin code has proven to be accurate to second order and has passed a number of numerical tests including highly relativistic cases and highly magnetized cases in both special and general relativity ( mizuno et al .
the raishin code performs special relativistic calculations in minkowski spacetime by changing the metric . for our simulations we will choose a force - free helical magnetic field as the initial configuration .
a force - free configuration is a reasonable choice for the strong magnetic field cases that we study here . in general , the force - free equilibrium of a static cylinder is described by the equation @xmath9 in particular we choose a poloidal magnetic field component of the form @xmath10^{\alpha}}~,\ ] ] for which one finds a toroidal magnetic field component of the form @xmath11^{\alpha } } \sqrt { { [ 1 + ( r / a)^{2}]^{2 \alpha } -1 - 2 \alpha ( r / a)^2 \over 2 \alpha -1}}~,\ ] ] where @xmath12 is the cylindrical radial position normalized by a simulation scale unit @xmath13 , @xmath14 parameterizes the magnetic field amplitude , @xmath15 is the characteristic radius of the column , and @xmath16 is the pitch profile parameter .
the pitch profile parameter determines the radial profile of the magnetic pitch @xmath17 , and provides a measure of the twist of the magnetic field lines . with our choice for the force - free field
, the magnetic pitch can be written as @xmath18^{2 \alpha } -1 -2 \alpha ( r / a)^{2 } } } ~.\ ] ] if the pitch profile parameter @xmath19 , the magnetic pitch increases with radius . if @xmath20 , the magnetic pitch decreases . when @xmath21 , the magnetic pitch is constant .
this configuration is the same as that used in previous non - relativistic work ( appl et al .
2000 ; baty 2005 ) .
the simulation grid is periodic along the axial direction . as a consequence
the allowed axial wavelengths are restricted to @xmath22 , with @xmath23 a positive integer and @xmath24 is the grid length .
the grid is a cartesian ( @xmath25 ) box of size @xmath26 with grid resolution of @xmath27 . in simulations
we choose two different column radii , @xmath15 , relative to the column length , @xmath24 : ( case a ) @xmath28 and ( case b ) @xmath29 . for case a @xmath30 and for case b @xmath31 .
in terms of @xmath15 , the simulation box size is @xmath32 for case a and @xmath33 for case b. we impose outflow boundary conditions on the transverse boundaries at @xmath34 ( @xmath35 for case a and @xmath36 for case b ) .
we consider a low gas pressure medium with constant @xmath37 for the equilibrium state , and two different density profiles : ( case u ) uniform density @xmath38 and ( case n ) non - uniform density decreasing proportional to the magnetic field strength , @xmath39 with @xmath40 .
the equation of state is that of an ideal gas with @xmath41 , where @xmath42 is the specific internal energy density and the adiabatic index @xmath43 .
the specific enthalpy is @xmath44 .
the magnetic field amplitude is @xmath45 leading to a low plasma-@xmath46 near the axis .
the sound speed is @xmath47 and the alfvn speed is @xmath48^{1/2}$ ] . in order to investigate different radial pitch profiles ,
we perform simulations with : ( case 1 ) constant pitch , @xmath49 , ( case 2 ) increasing pitch , @xmath50 , and ( case 3 ) decreasing pitch @xmath51 . the radial profiles of the magnetic field components , the magnetic pitch , and the sound and alfvn speeds for the different cases are shown in figure 1 . the radial profile of the magnetic field components for cases a and b are the same when normalized by @xmath15 .
the initial mhd equilibrium configuration is perturbed by a small radial velocity with profile given by @xmath52 the amplitude of the perturbation is taken to be @xmath53 with radial width @xmath54 ( @xmath55 for case a and @xmath56 for case b ) .
we choose @xmath57 and @xmath58 in the above formula .
this is identical to imposing @xmath59 , because of the symmetry between @xmath60 and @xmath61 pairs .
the various different cases that we have considered are listed in table 1 .
lccccccc [ table1 ] a1u & 0.125 & 1.0 & uniform & constant & 2.0 & 16 & 0.4 + a1n & 0.125 & 1.0 & decrease & constant & 2.0 & 16 & 0.3 + b1u & 0.25 & 1.0 & uniform & constant & 3.0 & 12 & 0.4 + b1n & 0.25 & 1.0 & decrease & constant & 3.0 & 12 & 0.3 + a2u & 0.125 & 0.75 & uniform & increase & 2.0 & 16 & 0.4 + a3u & 0.125 & 2.0 & uniform & decrease & 2.0 & 16 & 0.4 + b2u & 0.25 & 0.75 & uniform & increase & 3.0 & 12 & 0.4 + b3u & 0.25 & 2.0 & uniform & decrease & 3.0 & 12 & 0.4 + a2n & 0.125 & 0.75 & decrease & increase & 2.0 & 16 & 0.3 + a3n & 0.125 & 2.0 & decrease & decrease & 2.0 & 16 & 0.3 + b2n & 0.25 & 0.75 & decrease & increase & 3.0 & 12 & 0.3 + b3n & 0.25 & 2.0 & decrease & decrease & 3.0 & 12 & 0.3 +
as an indicator of the growth of the cd kink instability we use the volume - averaged kinetic energy transverse to the @xmath62-axis within a cylinder of radius @xmath63 written as @xmath64 and normalized by the initial volume - averaged transverse magnetic energy .
the quantity @xmath65 allows determination of different evolutionary stages ( initial exponential ( linear growth phase ) growth , and subsequent non - linear evolution ) . in all cases ,
the initial growth regime is characterized by an exponential increase in @xmath65 by several orders of magnitude to a maximum amplitude followed by a slow decline in the non - linear regime .
the time evolution of the maximum radial velocity evaluated over half the grid length , @xmath66 , shows a growth trend similar to the growth of @xmath65 .
figure 2 shows the time evolution of @xmath65 and maximum radial velocity for cases a and b with constant pitch for uniform and decreasing radial density profiles .
the effect of changing the characteristic radius , @xmath15 , relative to the grid length , @xmath24 , is equivalent to changing the wavelength .
the wavelength in case a is @xmath67 , and in case b is @xmath68 .
note that according to the kruskal - shafranov criterion , the instability developes at @xmath69 .
the instability growth rate reaches a maximum at @xmath70 , the exact coefficient being dependent on the transverse distribution of the density and magnetic pitch .
specifically in the case of constant pitch and uniform density , appl et al .
( 2000 ) found @xmath71 and a corresponding growth rate of @xmath72 .
in general , one can use the estimate @xmath73 .
in uniform density ( solid and dashed lines in fig . 2 ) cases @xmath65 grows exponentially ( linear growth phase ) , reaches maximum amplitude at @xmath74 for @xmath75 and at @xmath76 for @xmath77 , and slowly declines in the non - linear regime .
the timescale @xmath78 is in units of the alfvn crossing time of the characteristic radius @xmath15 , i.e. , @xmath79 where the initial alfvn velocity on the axis is @xmath80 .
the difference in growth is a result of the different wavelengths , longer in case a and shorter in case b. both cases show similar amplitude of @xmath65 at transition to the non - linear stage .
here we find that the shorter wavelength , @xmath81 , grows slightly more rapidly than the longer wavelength , @xmath67 , as would be expected if @xmath82 . in decreasing density cases ( dotted lines and dash - dotted lines in fig .
2 ) , the initial growth in the linear growth phase is more rapid and reaches maximum amplitude at @xmath83 for @xmath75 and for @xmath77 , and slowly declines in the non - linear regime . here
the initial alfvn velocity on the axis is @xmath84 and it follows that @xmath85 .
this more rapid growth is a result of dependence of the growth rate on the alfvn velocity .
the density decline in the decreasing case leads to a more gradual radial decline in the alfvn velocity ( see fig .
1 g ) than for the uniform density cases , and on average a higher alfvn velocity .
the maximum amplitude of @xmath65 relative to the initial volume - averaged magnetic energy for the decreasing density cases is almost the same as the uniform density cases .
the maximum radial velocities normalized by the initial axial alfvn velocity are higher for the decreasing density cases but in absolute terms are almost the same .
we note that the normalized maximum radial velocity in the decreasing density cases appears larger than in the uniform density cases because the normalizing axial alfvn velocity is smaller for the decreasing density cases .
the effect of different radial pitch on the time evolution of @xmath86 and the maximum radial velocity for cases with different radial pitch are shown in figure 3 .
the three uniform density and three decreasing density cases show similar linear and non - linear evolution but with different time scales .
growth in the linear stage is more rapid for the decreasing density cases than for the uniform density cases .
this can be attributed to the higher average alfvn velocity in the decreasing density case . for both uniform and decreasing density cases the increasing pitch case ( @xmath87 )
grows more slowly and reaches maximum at a later time with a smaller value for @xmath65 than the constant pitch case ( @xmath88 ) . on the other hand ,
the decreasing pitch case ( @xmath89 ) grows more rapidly and reaches maximum at an earlier time with a larger value for @xmath65 than the constant pitch case .
although the transition time from linear to non - linear evolution is different for each pitch case , the maximum radial velocity is almost the same at transition .
the different growth rates as a function of the radial pitch profile are consistent with the non - relativistic linear analysis in appl et al .
( 2000 ) . the results of the previous subsection show that the instability excites relatively slow motions , @xmath90 , and that the overall energy of the system does not vary too much . nevertheless , the three - dimensional structure of the system is strongly distorted , as will be shown in this subsection .
we note that results for case b with the shorter wavelength are essentially the same as for case a. moreover , the evolution is similar in the uniform and decreasing density cases and only occurs faster in the decreasing density cases .
therefore we show the results of case a with decreasing density only .
figure 4 shows the time evolution of a density isosurface for the constant pitch case a1n .
displacement of the initial force - free helical magnetic field leads to a helically twisted magnetic filament around the density isosurface . at longer times
, the radial displacement of the high density region , best seen in transverse density slices at the grid midplane @xmath91 ( see right panels in fig . 4 ) , slows significantly .
continuing outwards radial motion is confined to a lower density sheath around the high density core .
the transverse growth of helical twisting is illustrated by the time evolution of magnetic field line displacement seen from the pole - on view shown in figure 5 .
here we follow field lines anchored initially at @xmath92 at the @xmath93 surface .
displacement of the magnetic field rapidly increases with time in the linear stage and the displacement continues to gradually increase throughout the non - linear phase .
thus , the magnetic field displacement indicates continued growth even though the high density region has ceased significant outwards motion by the terminal simulation time .
evolution of the helical perturbation in the magnetic field is accompanied by an appropriate evolution of the current distribution .
initially a large axial current is located near the axis . in the linear growth phase ,
the axial current is displaced and twists helically .
the axial current twist is cospatial with the high density twist . in the linear growth phase the axial current gradually decreases and in the non - linear phase
remains constant in magnitude .
figure 6 shows density isosurfaces and transverse density slices at the grid midplane , @xmath94 , for the cases a2n ( increasing pitch ) and a3n ( decreasing pitch ) that can be compared to figure 4 for case a1n with constant pitch .
the 3d density structure from a2n looks similar to that from a1n .
the cd kink instability grows exponentially initially . the transverse density slice at the grid midplane ( fig .
6c ) shows little outwards motion of the high density region in the non - linear stage similar to the constant pitch case a1n , however , there is little outwards motion in the low density sheath surrounding the high density region .
this result is somewhat different from the constant pitch case a1n and suggests a significant reduction in kink amplitude growth .
results from the decreasing pitch case a3n are very different .
figure 6b shows a more slender helical density structure wrapped by the magnetic field .
recall that transition from linear to non - linear evolution was reached at a much earlier time , @xmath95 ( see fig .
3 ) , than is shown here .
while the density cross section ( fig .
6d ) is similar to that of the constant pitch case a1n ( fig .
4f ) , radial motion continues in the non - linear stage to the end of the simulation .
figure 7 shows the axial current isosurface and the magnetic field line displacement seen from the pole - on view for cases a2n and a3n . like the constant pitch decreasing density case a1n , the axial current twist is cospatial with the high density twist . in the non - linear stage ,
the helically twisted axial current does not change significantly in case a2n .
a pole on view of field lines anchored initially at @xmath92 at the @xmath93 surface ( see fig .
7c ) shows that the magnetic field displacement does not increase significantly in the non - linear stage .
this along with the density results shown in figure 6a & c suggest significant non - linear stabilization of the kink mode in the presence of increasing pitch . on the other hand ,
the axial current isosurface for the decreasing pitch case a3n seen in figure 7b makes a slender helical structure coincident with the high density structure .
large displacement in the magnetic field can be seen in the pole on view ( fig .
we have investigated development of the cd kink instability of force - free helical magnetic equilibria in order to study the influence of the initial configuration on the linear and nonlinear evolution .
we followed the time evolution where the plasma column was perturbed to excite the @xmath96 azimuthal mode .
the kink developed initially as predicted by linear stability analysis . the rate at which the kink developed and , in particular , the non - linear behavior depended on the density and magnetic pitch radial profile . for a constant magnetic pitch profile a density decline with radius led to faster linear growth and made a transition to the non - linear stage sooner than occured for a uniform density as a result of the more gradual decrease in the alfvn velocity with radius .
while there were some differences in the kink density structure in these two different cases , amplitude growth of the kink continued to the terminal simulation time in both cases .
more profound differences in development accompanied different magnetic pitch profiles .
an increase in the magnetic pitch with radius led to a reduced growth rate in the linear ( exponential ) growth stage and made a transition to non - linear behavior at a later time when compared to the cases with constant magnetic pitch . in the nonlinear stage ,
amplitude growth of the kink appeared to have ceased by the terminal simulation time . on the other hand ,
a decrease in the magnetic pitch led to more rapid growth in the the linear growth stage and made a transition to non - linear behavior at an earlier time when compared to the cases with constant magnetic pitch . in the non - linear stage , a slender helical density and current carrying plasma column wrapped by magnetic field developed and amplitude growth continued to increase throughout the simulation .
thus , the linear growth and nonlinear evolution of the cd kink instability depended on the radial density profile and strongly depended on the magnetic pitch profile .
increasing magnetic pitch with radius was clearly stabilizing . in all cases ,
the initial axisymmetric structure is strongly distorted by the kink instability , even though not disrupted .
it is important to note that the instability develops relatively slowly , that is , the characteristic time for the instability to affect strongly the initial structure is roughly @xmath97 .
the growth rate of the kink instability is roughly @xmath98 , the exact coefficient being dependent on the structure of the undisturbed state . therefore it is not accidental that the full development of the instability takes a long time . in a jet context
, our simulations correspond to a perturbation that remains at rest in the flow frame .
therefore , in order to check whether the instability would affect a jet flow , one has to compare @xmath99 with the propagation time . in the relativistically moving case
, time dilation slows the instability further by the jet lorentz factor so that the condition for the instability to affect the jet structure may be written as @xmath100 in relativistic jets , one could take @xmath101 . in order to find the final criterion
, one has to know how the jet radius , @xmath15 , and lorentz factor , @xmath102 , vary with the distance , @xmath62 .
if the jet is narrow enough so that @xmath103 , where @xmath104 is the angular velocity at the base of the jet , one can use the scaling ( tschekhovskoy et al .
2008 ; komissarov et al . 2009
; lyubarsky 2009 ) @xmath105 . in this case one finds that the criterion for the kink instability is @xmath106 this means that the instability could affect the jet structure only if the jet expands slowly enough . assuming the parabolic shape for the jet , @xmath107 , where @xmath108 and @xmath109 are dimensionless numbers , one finds that the instability develops only if @xmath110 . even in this case
, the characteristic scale for the development of the instability is very large , @xmath111 the above estimates should be considered as preliminary because we assumed that the jet moves as a whole so that there is a frame of reference in which the plasma is at rest .
clearly , the criterion for the cd kink instability can be modified by the effects of relativistic rotation , gradual shear , a surrounding sheath , and sideways expansion .
for example , in relativistic poynting dominated jets the growth rate of the instability strongly depends on the transverse profile of the poloidal magnetic field so that @xmath99 goes to infinity when the poloidal field is uniform ( istomin & pariev 1994 , 1996 ; lyubarskii 1999 ) . on the other hand , poynting
dominated jets are accelerated such that the acceleration zone spans a large range of scales ( tschekhovskoy et al .
2008 ; komissarov et al.2008 ; lyubarsky 2009 ) .
therefore , a small kink growth rate does not necessarily preclude significant growth of helical perturbations .
we plan to perform 3d rmhd simulations including some of the missing physical effects such as jet flow and rotation in future work .
in particular , this will allow us to investigate the interaction between cd and kh driven structure in the relativistic regime .
, k.i.n . , and p.h .
acknowledge partial support by nsf ast-0506719 , ast-0506666 , nasa nng05gk73 g , nnx07aj88 g , and nnx08ag83 g .
this work is supported in part by us - israeli binational science foundation grant 2006170 .
the simulations were performed on the columbia supercomputer at nas division at nasa ames research center and the altix3700 bx2 at yitp in kyoto university .
in the simulations presented here numerical viscosity and resistivity are associated with the finite grid resolution .
it is generally expected that numerical viscosity affects the quantitative growth or damping of instabilities and numerical resistivity allows unexpected magnetic reconnection even when we assume the ideal mhd condition .
therefore we perform several numerical tests to check whether saturation is adequately captured and investigate how the results in the non - linear evolution stage depend on numerical resolution .
first , we repeat case a1u ( uniform density with constant pitch ) for four grid resolutions from 20 to 60 computational zones per simulation length unit @xmath112 .
figure 8 shows the time evolution of @xmath65 and the maximum radial velocity for different grid resolutions .
the results depend on grid resolution .
higher grid resolution shows faster growth and an earlier transition time to the non - linear stage at a larger amplitude .
this result is the same as that found in the resolution study of baty & keppens ( 2002 ) .
the lowest grid resolution with @xmath113 is clearly inadequate .
although growth does depend on grid resolution , the difference between our choice of @xmath27 and the highest grid resolution of @xmath114 is not significant .
therefore our simulation results adequately capture the cd kink instability .
second , we check the influence of the transverse boundary .
figure 9 shows the time evolution of @xmath65 and the maximum radial velocity within a box of length @xmath66 for a simulation box with transverse boundaries at @xmath115 ( @xmath116 ) and a simulation box with transverse boundaries at @xmath117 ( @xmath118 ) .
the results for the time evolution of @xmath65 and the maximum radial velocity are the same in both cases .
we conclude that behavior of the cd kink instability is not affected by our choice for the transverse boundary .
third , we check the influence of the initial gas pressure . in the force - free model we ignore the inertia and pressure of the plasma .
however , in the simulation the relativistic ideal mhd equations are solved numerically not the force - free equations .
figure 10 shows the time evolution of @xmath65 and the maximum radial velocity in a box of length @xmath66 for an initial pressure of @xmath119 and @xmath120 .
the results of the time evolution of @xmath65 and the maximum radial velocity show the same linear growth rate in both cases . however , the lower initial pressure case reaches saturation at a slightly later time with higher amplitude but with a larger decline in the non - linear stage .
thus , our required non - zero initial gas pressure does affect the saturation level and non - linear development of the cd kink instability somewhat . | we have investigated the development of current - driven ( cd ) kink instability through three - dimensional relativistic mhd simulations .
a static force - free equilibrium helical magnetic configuration is considered in order to study the influence of the initial configuration on the linear and nonlinear evolution of the instability .
we found that the initial configuration is strongly distorted but not disrupted by the kink instability .
the instability develops as predicted by linear theory . in the non - linear regime
the kink amplitude continues to increase up to the terminal simulation time , albeit at different rates , for all but one simulation .
the growth rate and nonlinear evolution of the cd kink instability depends moderately on the density profile and strongly on the magnetic pitch profile .
the growth rate of the kink mode is reduced in the linear regime by an increase in the magnetic pitch with radius and the non - linear regime is reached at a later time than for constant helical pitch . on the other hand , the growth rate of the kink mode
is increased in the linear regime by a decrease in the magnetic pitch with radius and reaches the non - linear regime sooner than the case with constant magnetic pitch .
kink amplitude growth in the non - linear regime for decreasing magnetic pitch leads to a slender helically twisted column wrapped by magnetic field . on the other hand , kink amplitude growth in the non - linear regime
nearly ceases for increasing magnetic pitch . |
galactic globular clusters ( gcs ) are dynamically active systems , where stellar interactions and collisions , especially those involving binaries , are quite frequent ( e.g. * ? ? ? * ; * ? ? ?
* ) and generate populations of exotic objects , like x - ray binaries , millisecond pulsars , and blue straggler stars ( bsss ; e.g. , @xcite , 2009;@xcite ) . among these objects ,
bsss are the most abundant and therefore may act as a crucial probe of the interplay between stellar evolution and stellar dynamics ( e.g. * ? ? ?
they are brighter and bluer ( hotter ) than main sequence ( ms ) turnoff ( to ) stars and are located along an extrapolation of the ms in the optical color - magnitude diagram ( cmd ) .
thus , they mimic a rejuvenated stellar population , with masses larger than those of normal cluster stars ( @xmath4 ; this is also confirmed by direct measurements ; @xcite ) .
this clearly indicates that some process which increases the initial mass of single stars must be at work
. the most probable formation mechanisms of these peculiar objects are thought to be either the mass transfer ( mt ) activity between binary companions , even up to the complete coalescence of the system @xcite , or the merger of two single or binary stars driven by stellar collisions ( col ; * ? ? ?
* ; * ? ? ?
unfortunately , it is still a quite hard task to distinguish bsss formed by either channel .
the most promising route seems to be high resolution spectroscopy , able to identify chemical anomalies ( as a significant depletion of carbon and oxygen ) on the bss surface ( see * ? ? ?
* ) , as it is expected in the mt scenario .
also the recent discovery of a double sequence of bsss in m30 @xcite seems to indicate that , at least in gcs that recently suffered the core collapse , col - bsss can be distinguished from mt - bsss on the basis of their position in the color - magnitude diagram ( cmd ) . in general , since collisions are most frequent in high density regions , bsss in different environments could have different origins : those in loose gcs might arise from the coalescence / mass - transfer of primordial binaries , while those in high density clusters might form mostly from stellar collisions ( e.g. , * ? ? ? * ; * ? ? ?
however , @xcite suggested that most bsss , even those observed in the cores of high density gcs , formed from binary systems , although no strong correlation between the number or specific frequency of bsss and that of binaries is found if all clusters ( including the post - core collapse ones ) are taken into account @xcite .
this is likely due to the fact that dynamical processes significantly modify the binary and the bss content of gcs during their evolution .
an exception seems to be represented by low density environments , where the efficiency of dynamical interactions is expected to be negligible . very interestingly , indeed , a clear correlation between the binary and the bss frequencies has been found in a sample of 13 low density gcs @xcite .
this remains the cleanest evidence that the unperturbed evolution of primordial binaries is the dominant bss formation process in low - density environments ( also consistently with the results obtained in open clusters ; e.g. * ? ? ?
one of the most notable effects of stellar dynamics on bsss is the shaping of their radial distribution within the host cluster . in most cases
a clear bimodality has been observed in the radial distribution of the @xmath5 ratio ( @xmath6 and @xmath7 being , respectively , the number of bsss and the number of stars belonging to a reference population , as red giant or horizontal branch stars ) : such a ratio shows a high peak at the cluster centre , a dip at intermediate radii and a rising branch in the external regions ( e.g. * ? ? ?
in a few other gcs ( see , for example , ngc 1904 and m75 ; * ? ? ?
* ; * ? ? ?
* respectively ) the radial distribution shows a clear central peak but no upturn in the external cluster regions . in the case of @xmath8centauri @xcite , ngc 2419 @xcite and palomar 14 @xcite
the bss radial distribution is flat all over the entire cluster extension .
since bsss are more massive than the average cluster stars ( and therefore suffer from the effects of dynamical friction ) , these different observed shapes have been interpreted as due to different `` dynamical ages '' of the host clusters ( * ? ? ? * hereafter f12 ) : a flat bss distribution is found in dynamically young gcs ( _ family i _ ) , where two - body relaxation has been ineffective in establishing energy - equipartition in a hubble time and dynamical friction has not segregated the bss population yet . in clusters of intermediate ages ( _ family ii _ ) , the innermost bsss have already sunk to the center , while the outermost ones are still evolving in isolation ( this would produce the observed bimodality ) ; finally , in the most evolved gcs ( _ family iii _ ) , dynamical friction has already segregated toward the centre even the most remote bsss , thus erasing the external rising branch of the distribution .
these results demonstrate that the bss radial distribution can be used as powerful dynamical clock . of course
, the same should hold for binary systems with comparable masses , but the identification of clean and statistically significant samples of binaries all over the entire cluster extension is a quite hard task and no investigations have been performed to date to confirm such an expectation .
ngc5466 is a high galactic latitude gc ( @xmath9 and @xmath10 ) , located in the constellation of botes at a distance of @xmath11 kpc @xcite .
the cluster has been recently found to be surrounded by huge tidal tails @xcite .
the bss stars in this cluster were first studied by @xcite , and , more recently , @xcite analyzed the bss population inside the first @xmath12 from the cluster center . since the cluster tidal radius is @xmath13 @xcite , a survey of the bss population over the whole body of the cluster is still missing . in the framework of the study of the origin of the bss , it is important to mention that ngc 5466 is the first gc where eclipsing binaries were found among the bss ( * ? ? ?
* see also kryachko et al.(2011 ) ) .
in their paper @xcite find indications that all the bss in ngc 5466 were formed from the merger of the components of primordial , detached binaries evolved into compact binary systems . in this paper
we intend to study the bss and binary populations of ngc 5466 , focussing in particular on their radial distribution within the whole cluster to investigate its dynamical age .
the paper is organized as follows . in section [ sec_obs ]
we describe the observations and data analysis .
section [ sec_bss ] is devoted to the definition of the bss population and the study of its radial distribution . in section [ sec_bin ]
we describe how the fraction of binaries and its variation with radius are investigated .
the summary and discussion of the obtained results are presented in section [ sec_concl ] .
the photometric data used here consist of a combination of shallow and deep images ( see table 1 ) acquired through the blue channel of the large binocular camera ( lbc - blue ) mounted on the large binocular telescope ( lbt ) , megacam on the canadian - france - hawaii telescope ( cfht ) and the advanced camera for survey ( acs ) on board the hubble space telescope ( hst ) . _ ( i)the lbc data - set _ :
both short and long exposures were secured in 2010 under excellent seeing conditions ( 05 - 07 ) with the lbc - blue ( @xcite ) mounted on the lbt @xcite at mount graham , arizona .
the lbc is a double wide - field imager which provides an effective @xmath14 field of view ( fov ) , sampled at @xmath15 arcsec / pixel over four chips of @xmath16 pixels each .
lbc - blue is optimized for the uv - blue wavelengths , from 320 to 500 nm , and is equipped with the @xmath17 , @xmath18 , @xmath19 , @xmath20 and @xmath21 filters ( hereafter @xmath22 , respectively ) .
a set of short exposures ( of 5 , 60 and 90 seconds ) was secured in the @xmath23 and @xmath24 filters with the cluster center positioned in the central chip ( namely # 2 ) of the lbc - blue ccd mosaic .
deep images ( of 400 and 200 s in the @xmath23 and @xmath24 bands , respectively ) were obtained with the lbc - blue fov positioned @xmath25 south from the cluster centre .
a dithering pattern was adopted in both cases thus to fill the gaps of the ccd mosaic .
the raw lbc images were corrected for bias and flat field , and the overscan region was trimmed using a pipeline specifically developed for lbc image pre - reduction from the lbc survey data center . _
( ii)the megacam data - set _ : @xmath26 and @xmath27 wide - field megacam images ( prop i d : 04ak03 ; pi : jang - hyun park ) , acquired with excellent seeing conditions ( 06 - 08 ) were retrieved from the canadian astronomy data centre ( cadc ) .
mounted at cfht ( hawaii ) , the wide - field imager megacam ( built by cea , france ) , consists of 36 @xmath28 pixel ccds ( a total of 340 megapixels ) , fully covering a 1 degree @xmath29 1 degree fov with a resolution of 0.187 arcsecond / pixel .
the data were preprocessed ( removal of the instrumental signature ) and calibrated ( photometry and astrometry ) by elixir pipeline . _
( iii)the acs data - set _ consists of a set of high - resolution , deep images obtained with the acs on board the hst through the f606w and f814w filters ( go-10775;pi : sarajedini ) .
the cluster was centered in the acs field and observed with one short and five long - exposures per filter , for a total of two orbits .
we retrieved and used only the deep exposures ( see table [ tab_obs ] ) .
all images were passed through the standard acs / wfc reduction pipeline .
as shown in the left panel of figure [ maps ] , the shallow lbc data - set is used to sample a region of @xmath30 around the cluster center , and for @xmath31 it is complemented with the megacam observations , which extend beyond the tidal radius , out to @xmath32 . in the following the combination of these two data - sets will be referred to as the `` shallow sample '' and adopted to study the bss population over the entire cluster extension ( see sect . [ sec_bss ] ) .
the acs observations sample the innermost @xmath33 of the cluster and they are complemented with the deep lbc observations , extending out to @xmath34 ( see right panel of figure [ maps ] ) : this will be called the `` deep sample '' and used to investigate the binary fraction of ngc 5466 ( see sect .
[ sec_bin ] ) . given the very low crowding conditions in the external cluster regions , we performed aperture photometry on the two @xmath26 and @xmath27 megacam images , by using sextractor @xcite with an aperture diameter of @xmath35 ( corresponding to a fwhm of 5 pixels ) .
each of the 36 chips was reduced separately , and we finally obtained a catalog listing the relative positions and magnitudes of all the stars in common between the @xmath26 and @xmath27 catalogs . the lbc and acs data - sets were reduced through a standard point spread function ( psf ) fitting procedure , by using daophotii / allstar @xcite .
the psf was independently modeled on each image using more than 50 isolated and well sampled stars in the field .
the photometric catalogs of each pass - band were combined , and the instrumental magnitudes were reported to a reference frame following the standard procedure ( see , e.g. , * ? ? ?
* ; * ? ? ?
the magnitudes of stars successfully measured in at least three images per filter were then averaged , and the error about the mean was adopted as the associated photometric uncertainty . in order to obtain an absolute astrometric solution for each of the 36 megacam chips , we used more than 15000 stars in common with the sloan digital sky survey ( sdss ) catalog covering an area of 1 square degree centered on the cluster .
the final r.m.s .
global accuracy is of @xmath36 , both in right ascension ( @xmath37 ) and declination ( @xmath38 ) .
the same technique applied to the lbc sample gave an astrometric solution with similar accuracy ( @xmath39 r.m.s . ) .
considering that no standard astrometric stars can be found in the very central regions of the cluster , we used the stars in the lbc catalog as _ secondary astrometric standards _ and we thus determined an absolute astrometric solution for the acs sample in the core region with the same accuracy obtained in the previous cases .
a sample of bright isolated stars in the acs data - set was used to transform the acs instrumental magnitudes to a fixed aperture of @xmath40 .
the extrapolation to infinite and the transformation into the vegamag photometric system was then performed using the updated values listed in tables 5 and 10 of @xcite . in order to transform the instrumental @xmath23 and @xmath24 magnitudes of the lbc sample into the johnson / kron - cousins standard system
, we used the stars in common with a photometric catalog previously published by @xcite .
these authors , in section 2.3 of their paper , provide a thorough comparison of their photometry with existing literature datasets , finding a satisfactory agreement . in particular they found an excellent agreement in the photometric zero - points ( within less than 0.02 mag ) and no residual trends with colors with the set of secondary standards by @xcite , that should be considered as the most reliable reference .
we verified , by direct comparison with stetson ( 2000 ) , that the quality of the agreement shown by @xcite is fully preserved in our final catalog . instead
, the @xmath26 and @xmath27 magnitudes in the megacam sample were calibrated using the stars in common with the sdss catalog and we then adopted the equation that robert lupton derived by matching dr4 photometry to peter stetson s published photometry for stars to transform the calibrated @xmath26 and @xmath27 into a @xmath24 magnitude .
we finally used more than 2000 stars in common between the lbc and the megacam data - sets to search for any possible off - sets between the @xmath24 magnitude in the two catalogs , finding an average systematic difference of 0.014 mag .
we therefore applied this small correction to the megacam catalog .
similarly , we used more than 200 stars in common between the acs data - set and the lbc catalog to calibrate the acs f606w magnitude to the johnson @xmath24 magnitude .
such a procedure therefore provided us with a homogenous @xmath24 magnitude scale in common to all the available data - sets .
the cmd obtained from the shallow lbc exposures for the innermost @xmath30 from the cluster center is shown in figure [ cmd_shallow ] .
typical photometric errors ( magnitudes and colors ) are indicated by black crosses .
the high quality of the images allows us to accurately sample all the bright evolutionary sequences , from the tip of the red giant branch ( rgb ) , down to @xmath41 magnitudes below the ms - to , with a large population of bsss well visible at @xmath42 and @xmath43 . in figure [ cmd_deep ]
we show the cmds obtained from the acs and the deep lbc observations . as in figure
[ cmd_shallow ] , the typical photometric errors ( magnitudes and colors ) for the two sample are indicated by black crosses .
a well populated ms is visible in the acs cmd , extending from the ms - to ( @xmath44 ) down to @xmath45 .
once again , the collecting power of the lbt combined with the lbc - blue imaging capabilities allowed us to sample the ms down to similar magnitudes in an area extending out to the cluster s tidal radius . for @xmath46
the plume of galactic m dwarfs is clearly visible , while for @xmath47 and @xmath48 a population of unresolved background ( possibly extended ) objects is apparent . in both
the cmds a well defined sequence of unresolved binary systems is observed on the red side of the ms .
the shallow sample was used to determine the cluster center of gravity ( @xmath49 ) and the projected density profile from accurate star counts , following the procedure already adopted and described , e.g. , in @xcite .
the resulting value of @xmath49 is @xmath50 , @xmath51 , in full agreement with that quoted by @xcite .
the best fit king model to the observed density profile has concentration parameter @xmath52 , core radius @xmath53 and tidal radius @xmath54 @xcite . ) . while the latter is formally more correct , we keep referring to the tidal radius for sake of continuity with our previous papers about the bss populations in gcs ( see f12 and references therein ) . ]
@xcite studied the radial distribution of the bss population in ngc 5466 , from the cluster centre out to @xmath12 ( see also * ? ? ? * ) .
they selected 89 bsss and showed that they are significantly more centrally concentrated than rgb and horizontal branch ( hb ) stars , but they did not find any evidence of upturn in the external regions .
however , that study did not sample the entire radial extension of the cluster , while our previous experience demonstrates that even a few bsss in the outer regions can generate the external rising branch .
hence , a detailed investigation even in the most remote regions is necessary before solidly excluding the presence of an upturn in the bss radial distribution and confirming the dynamical age of the cluster .
the selection of the bss population was primarily performed in the @xmath55 plane ( see open circles in figure [ cmd_bss ] , left panel ) , by adopting conservative limits in magnitude and color in order to avoid the possible contamination from stellar blends . the bss selection has been then `` exported '' to the megacam @xmath56 cmd by using the bsss in common between the two data - sets in the region at @xmath57 ( see right hand panel ) .
we finally count a population of 88 stars in the lbc sample and 9 stars in the megacam one , for a total of 97 bsss spanning the entire radial extent of the cluster .
the cross - correlation of our sample with that of @xcite confirms that the majority of the stars selected in the previous work are real bsss , but 18 of them are blended sources . in order to further test the completeness of our lbc shallow sample we performed a detailed comparison with the acs photometric catalogue of @xciteata / public_hstgc/ ] .
the completeness has been quantified as the fraction of stars in common between the two catalogs in a given magnitude interval : we find 93% , 95% , and 100% completness for stars with @xmath58 , 21.5 , and 21 , respectively . for a meaningful study of the bss population ,
a reference sample of `` normal '' cluster stars has to be defined .
here we choose a post - ms population composed of sub - giant ( sgb ) and red giant branch stars ( rgb ; see grey triangles in fig .
[ cmd_bss ] ) .
this has been selected in the same magnitude range ( @xmath59 ) of bsss , thus guaranteeing that the photometric completeness is the same in both samples .
we count 1358 and 96 stars in the lbc and megacam data - sets , respectively .
as extensively illustrated in literature ( see for example * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , the study of the bss radial distribution is a very powerful tool to shed light on the internal dynamical evolution of gcs and possibly gets some hints on the bss formation mechanisms . as a first step
, we compared the cumulative radial distribution of the selected bss population to that of reference stars ( see figure [ ks ] ) .
clearly , bsss are more centrally concentrated , and a kolmogorov - smirnov test yields a probability of @xmath60 that the two samples are not extracted from the same parent population .
this cumulative bss distribution closely resembles that already observed in this cluster by @xcite out to @xmath61 ( see their figure 19 ) , with the only difference that we sample the entire cluster extension .
moreover , the distributions shown in figure [ ks ] , closely resembles that observed in all gcs characterised by a bimodal radial distribution of bsss ( see , e.g. , the cases of m3 , @xcite , 47tuc , @xcite ; m53 , @xcite ; m5 , @xcite ) . following the same procedure described in @xcite , as a second step we divided the surveyed area in a series of concentric annuli centred on @xmath49 . in table 2
we report the values of the inner and outer radii of each annulus , together with the number of bsss ( @xmath6 ) and reference stars ( @xmath62 ) counted in each radial bin .
we also report the fraction of light sampled in the same area , with the luminosity being calculated as the sum of the luminosities of all the stars measured in the to region at @xmath63 . the contamination of the cluster bss and rgb populations due to background and foreground field stars can be severe , especially in the external regions where number counts are low .
we therefore exploited the wide area covered by megacam ( see figure [ maps ] ) to estimate the field contamination by simply counting the number of stars located beyond the cluster tidal radius ( at @xmath64 ) that populate the bss and rgb selection boxes in the cmd .
we count 2 bsss and 13 rgb stars in a @xmath65 deg@xmath66 area .
the corresponding densities are used to calculate the number of field stars expected to contaminate each population in every considered radial bin ( see numbers in parenthesis in table 2 ) . in the upper panel of figure [ bss_rad ]
we show the resulting trend of @xmath67 as a function of radius .
clearly it is bimodal , with a relatively low peak of bsss in the inner region , a distinct dip at intermediate radii , and a constant value slightly smaller than the central one , in the outskirts .
we identify the minimum of the radial distribution at @xmath68 .
this behavior is further confirmed by the radial distribution of the bss double normalized ratio , @xmath69 ( see the black circles in the lower panel of figure [ bss_rad ] ) . for each radial bin ,
@xmath69 is defined as the ratio between the fraction of bsss in the annulus and the fraction of light sampled by the observations in the same bin ( see * ? ? ?
the same quantity computed for the reference population turns out to be equal to one at any radial distance ( see grey regions in the figure ) .
this is indeed expected from the stellar evolution theory @xcite and it confirms that the distribution of rgb stars follows that of the cluster sampled light . instead , @xmath69 shows a bimodal behavior , confirming the result obtained above .
notice that ngc 5466 is one of the most sparsely populated galactic gcs , and so the number of cluster s stars and the fraction of sampled light in the external regions is quite low .
binaries play an important role in the formation and dynamical evolution of gcs @xcite . indeed these systems secure a enormous energy reservoir able to quench mass - segregation and delay or even prevent the collapse of the core . up to now
three main techniques have been used to measure the binary fraction : _ ( i ) _ radial velocity variability ( e.g. , * ? ? ?
* ) , _ ( ii ) _ searches for eclipsing binaries ( e.g. * ? ? ?
* ; * ? ? ?
* ) and _ ( iii ) _ studies of the distribution of stars along the cluster ms in cmd ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in this work
we adopt the latter approach , which has the advantage of being more efficient for a statistical investigation , and unbiased against the orbital parameters of the systems .
we followed the prescriptions extensively described in @xcite and ( * ? ? ?
* ; * ? ? ?
* see also dalessandro et al .
this method relies on the fact that the luminosity of a binary system is the result of the contributions of both ( unresolved ) companions . the total luminosity is therefore given by the luminosity of the primary ( having mass @xmath70 ) increased by that of the companion ( with mass @xmath71 ) .
along the ms , where stars obey a mass - luminosity relation , the magnitude increase depends on the mass ratio @xmath72 , reaching a maximum value of 0.75 mag for equal mass binaries ( @xmath73 ) . taking advantage of the large fov covered by our data - sets , we analyzed the binary fraction at different radial distances from the cluster center , out to @xmath74 . in order to estimate the fraction of binaries from the `` secondary '' ms in the cmd , it is crucial to have a realistic and robust measure of other possible sources of broadening of the ms ( e.g. blending , photometric errors ) .
these factors are related to the quality of the data and can be properly studied through artificial star experiments .
we therefore produced a catalog of artificial stars for the acs and lbc - deep data - sets , following the procedure described in @xcite .
as a first step , a list of input positions and magnitudes of artificial stars ( x@xmath75,y@xmath75,@xmath76 and @xmath77 ) is produced .
once these stars are added to the original frames at the x@xmath75 and y@xmath75 positions , we repeat the measure of stellar magnitudes in the same way as described in section [ photometry ] . at the end of this procedure
we obtain a list of stars including both real and artificial stars . since we know precisely the positions and magnitudes of the artificial stars , the comparison of the latter with the measured @xmath78 and @xmath79 magnitudes , allow us to estimate the capability of our data - reduction strategy to properly detect stars in the images , including the effect of blending . in practice , we first generated a catalog of simulated stars with a @xmath80 magnitude randomly extracted from a luminosity function modeled to reproduce the observed one .
then the @xmath81 and @xmath77 magnitudes were assigned to each sampled @xmath80 magnitude by interpolating the mean ridge line of the cluster ( solid grey lines in the left and right panels of figure [ cmd_deep ] , for the acs and the lbc - deep samples , respectively ) .
the artificial stars were then added on each frame using the same psf model calculated during the data reduction phase , and were spatially distributed on a grid of cells of fixed width ( 5 times larger than the mean fwhm of the stars in the frames ) .
only one star was randomly placed in such a box in each run , in order not to generate artificial stellar crowding from the simulated stars on the images .
the artificial stars were added to the real images using the daophot / addstar routine . at the end of this step
, we have a number of `` artificial '' frames ( which include the real stars and the artificial stars with know position in the frame ) equal to the number of observed frames .
the reduction process was repeated on the artificial images in exactly the same way as for the scientific ones .
a total of 100,000 stars were simulated on each data - set and photometric completeness ( @xmath82 ) was then calculated as the ratio between the number of stars recovered by the photometric reduction ( @xmath83 ) and the number of simulated stars ( @xmath84 ) . in figure [ compl ] , we show the photometric completeness estimated for the acs and the lbc - deep data - sets ( upper and lower panels , respectively ) as a function of the @xmath24 magnitude and for different radial regions around the cluster center .
the completeness of the acs sample is above 80% for all magnitudes down to @xmath85 ( i.e. , @xmath86 magnitudes below the cluster to ) , even in the innermost region . on the other hand ,
a completeness above 50% down to @xmath87 is guaranteed only at @xmath3 in the lbc - deep data - set .
instead , stellar crowding is so strong at lower radial distances ( region d ) that the completeness stays below @xmath88 at all magnitudes .
we therefore exclude this region ( @xmath89 ) from our study of the cluster binary fraction .
as mentioned at the beginning of section [ sec_bin ] , our estimation of the binary fraction in ngc 5466 is performed by determining the fraction of binaries populating the area between the ms and a @xmath90 ms , blue - ward shifted in colour by @xmath91 @xcite . as a first step of our analysis
, we measured the minimum binary fraction , i.e. the fraction of binaries with mass ratios large enough ( @xmath92 ) that they can be clearly distinguished from single ms stars in the cmd .
when @xmath93 approaches the null value , the shift in magnitude of the primary star ( i.e. the more massive one ) is negligible .
for this reason binary systems with small values of @xmath93 becomes indistinguishable from ms stars when photometric errors are present .
clearly , this is a lower limit to the total binary fraction of the cluster .
however , it has the advantage of being a pure observational quantity , since it is calculated without any assumption on the distribution of mass ratios @xcite .
the magnitude interval in which the minimum binary fraction has been estimated is @xmath94 for both the acs and the lbc samples .
according to a 12 gyr model from @xcite = -2.22 , distance modulus @xmath95 and adopting a reddening , @xmath96 @xcite . ]
, this corresponds to a mass range @xmath97 for a single star on the ms . in this interval
the photometric completeness @xmath98 is larger than @xmath99 for both samples . for the magnitude and radial intervals just defined
, we divided the entire fov in four annuli ( see table 3 for details ) .
then , the location of @xmath100 was estimated on the cmd as the value corresponding to a color difference of three times the photometric error from the ms ridge mean line ( grey solid lines in the cluster s cmds shown in figure [ cmd_deep ] ) .
this approach allows us to define an area on the cmd safe from contamination from single mass stars .
we notice that for each of the four annuli , this color difference corresponds to @xmath101 .
the contamination from blended sources was calculated on the same region of the cmd by adopting the catalogue of artificial stars produced with the completeness tests ( see section [ sec_compl ] ) .
a catalogue of stars from the galaxymodel of @xcite was used to estimate the field contamination .
we thus estimated @xmath102 independently for the four radial intervals using the same approach described in @xcite .
as shown in figure [ bin ] ( upper panel ) , we find that @xmath102 is almost constant and equal to @xmath103 for @xmath104 ( i.e. , within the acs sample ) , then it drops at @xmath105 at distances larger than @xmath106 ( in the lbc sample ) .
as described above , the minimum binary fraction has the advantage of being measured with no arbitrary assumptions about the mass ratio distribution @xmath107 .
it depends , however , on the photometric accuracy of the used data - sets , and caution must be exercised when comparing the minimum binary fractions estimated for different gcs and in different works .
an alternative approach consists in measuring the global fraction of binaries ( @xmath108 ) , that can be obtained by simulating a binary population which follows a given distribution @xmath107 , and by comparing the resulting synthetic cmd with the observed one .
we already adopted this technique in @xcite , @xcite and @xcite .
hence , we refer to these papers for details . in brief , in order to simulate the binary population we randomly extracted @xmath109 values of the mass of the primary component from a @xcite initial mass function , and @xmath109 values of the binary mass ratio from the @xmath107 distribution observed by @xcite in the solar neighborhood , thus also obtaining the mass of the secondary .
we considered the same magnitude limits and radial intervals used before .
for each adopted value of the input binary fraction @xmath110 ( assumed to range between 0 and 15% with a step of 1% ) , we produced 100 synthetic cmds .
then , for each value of @xmath110 the penalty function @xmath111 was computed and a probability @xmath112 was associated to it ( see * ? ? ?
the mean of the best fitting gaussian is the global binary fraction @xmath113 .
the values obtained in the four considered radial bins are listed in table 3 .
the radial distribution of @xmath113 well resembles that obtained for the minimum binary fraction , remaining flat at @xmath114 in the innermost @xmath115 and decreasing to @xmath116 in the two outermost annuli .
we have investigated the radial distribution of bsss and binary systems in the galactic gc ngc 5466 , sampling the entire cluster extension .
this has been possible thanks to a proper combination of lbc - blue and megacam observations , and deep acs and lbc - blue data , respectively .
the bss radial distribution has been found to be bimodal , as in the majority of gcs investigated to date ( see f12 and references therein ) .
in fact , our data - set allowed us to detect a `` rising branch '' in the external portion of the bss distribution that was not identified in previous investigations because of a too limited sample in radius ( @xmath117 * ? ? ? * ) .
following the f12 classification , ngc 5466 therefore belongs to _ family ii_. the minimum of the distribution is a quite prominent , purely observational feature , that , it in the light of the results published by f12 , can be used to investigate the dynamical age of ngc 5466 .
it is located at @xmath118 , corresponding to only @xmath119 .
this is the second most internal location of the minimum , after that of m53 .
hence , ngc 5466 is one of the dynamically youngest clusters of _ family ii _ , where bsss have just started to sink toward the center of the system .
the relatively low central peak of the bss distribution ( @xmath120 ; see fig .
[ bss_rad ] ) is also consistent with this interpretation . as a further check , we calculated the core relaxation time ( @xmath121 ) .
f12 shows that @xmath122 is proportional to @xmath123 ( see their figure 4 ) .
the value of @xmath121 has been computed by using equation ( 10 ) of @xcite , adopting the new cluster structural parameters reported in section [ sec_cmd ] . in figure
[ clock ] we show the position of ngc 5466 ( black solid circle ) on the `` dynamical clock plane '' ( @xmath124 as a function of @xmath125 ; figure 4 in f12 ) , where @xmath126 is the age of the universe ( @xmath126 = 13.7 gyr ) .
clearly , this cluster nicely follows the same relation defined by the sample analyzed in f12 ( see also @xcite ) , further confirming that the shape of the observed bss distribution is a good measure of gc dynamical ages .
this figure also confirms that ngc 5466 is one of the youngest cluster of _ family ii _ , meaning that only recently the action of dynamical friction started to segregate bsss ( and primordial binary systems of similar total mass ) toward the cluster centre . in this scenario ,
the most remote bsss are thought to be still evolving in isolation in the outer cluster regions .
the central values of the derived binary fractions ( @xmath127 and @xmath128 ) are slightly smaller than those quoted by @xcite .
this is due to the differences in the adopted data reduction procedures and the consequent completeness analysis results , as well as the assumed luminosity functions .
the value of @xmath102 is in very good agreement with the estimate obtained in the same radial range by @xcite from the same acs data - sets : @xmath129 ( see their figure 34 and the innermost open triangle in fig .
[ bin ] ) . instead
, our total binary fraction is significantly lower than that quoted by ( * ? ? ? * @xmath130 ) because of the different assumptions made about the shape of @xmath107 .
in fact , these authors assumed a constant mass - ratio distribution for all @xmath93 values , and their total binary fraction is simply twice the value of @xmath102 .
indeed , the measure of @xmath113 is quite sensitive to the assumed mass - ratio distribution , as also demonstrated by the value ( 11.7% ) obtained by simulating binaries formed by random associations between stars of different masses @xcite .
@xcite also provides the value of the minmum binary fraction abetween @xmath131 and @xmath132 from the cluster center : @xmath133% .
we emphasize , however , that this estimate has been obtained by using only the most external fragments of the acs fov , corresponding to less than 5% of the total sampled area
( this likely explains the large uncertainty quoted by the authors ) .
keeping this caveat in mind , in light of the good agreement between our central value of @xmath102 and that measured by @xcite , we include their estimate in our analysis , thus sampling the intermediate region which is not covered by our data .
very interestingly , the value computed by @xcite defines a minimum in the binary fraction radial distribution ( see the outer open triangle in fig .
[ bin ] ) , thus implying that also the radial distribution of binaries seems to be bimodal in ngc 5466 .
we stress that the investigated binary systems ( having primary star masses between 0.5 and @xmath134 ) are in a mass range consistent with that of bsss .
indeed , within the uncertainties , also the position of the minimum of the binary radial distribution is in good agreement with that of the bss population .
this result urges a confirmation through dedicated observations .
in fact , we emphasize that this would be the first time that a similar feature is observed in a gc
. it would further strengthen the interpretation proposed by f12 that the shape of the bss radial distribution ( and simialr mass objects ) is primarily sculpted by dynamical friction .
moreover , it adds further support to the conclusions that the unperturbed evolution of primordial binaries could be the dominant bss formation process in low - density environments @xcite . @ccccc@ data - set & number of exposures & filter & exposure time & date of observations + & & & ( s ) & + + lbc - blue & 1 & @xmath23 & 5 & 2007 - 06 - 17 + & 7 & @xmath23 & 90 & 2007 - 06 - 17 + & 1 & @xmath24 & 5 & 2007 - 06 - 17 + & 7 & @xmath24 & 60 & 2007 - 06 - 17 + megacam & 1 & @xmath26 & 90 & 2004 - 04 - 14 + & 1 & @xmath27 & 180 & 2004 - 04 - 14 + + lbc - blue & 11 & @xmath23 & 400 & 2010 - 04 - 11 + & 15 & @xmath24 & 200 & 2010 - 04 - 11 +
acs & 5 & @xmath135 & 340 & 2006 - 04 - 12 + & 5 & @xmath136 & 350 & 2006 - 04 - 12 + | we present a detailed investigation of the radial distribution of blue straggler star and binary populations in the galactic globular cluster ngc 5466 , over the entire extension of the system .
we used a combination of data acquired with the acs on board the hubble space telescope , the lbc - blue mounted on the large binocular telescope , and megacam on the canadian - france - hawaii telescope .
blue straggler stars show a bimodal distribution with a mild central peak and a quite internal minimum .
this feature is interpreted in terms of a relatively young dynamical age in the framework of the `` dynamical clock '' concept proposed by ferraro et al .
( 2012 ) .
the estimated fraction of binaries is @xmath0 in the central region ( @xmath1 ) and slightly lower ( @xmath2 ) in the outskirts , at @xmath3 . quite interestingly
, the comparison with the results of @xcite suggests that also binary systems may display a bimodal radial distribution , with the position of the minimum consistent with that of blue straggler stars .
if confirmed , this feature would give additional support to the scenario where the radial distribution of objects more massive than the average cluster stars is primarily shaped by the effect of dynamical friction .
moreover , this would also be consistent with the idea that the unperturbed evolution of primordial binaries could be the dominant bss formation process in low - density environments . |
one of the most successful attempt to resolve the problems of standard big bang cosmology such as homogeneity , isotropy , and flatness of the universe is the inflationary paradigm , characterised by an epoch of accelerated expansion , `` inflation '' , in the very early universe . during the inflationary epoch , quantum fluctuations
are highly amplified , their wavelengths are stretched to outside the hubble horizon and inevitably , superhorizon fluctuations are generated .
these fluctuations become classical after crossing the event horizon and are coherent on what appear to be superhorizon scales at decoupling . after the end of inflation , they re - enter the horizon , and seed the matter and the radiation fluctuations observed in the universe . these primordial fluctuations
are gaussian , adiabatic and nearly scale - invariant in the majority of inflation models and even a slight deviation from these properties can strongly constrain the assumptions in cosmological models . since the linearity of the cosmic microwave background ( cmb ) anisotropy preserves the basic properties of these primordial fluctuations , the cmb radiation anisotropy is a promising tool for testing these properties .
however , it is known that the observed quadrupole ( @xmath0 , which is the largest scale that can be observed ) amplitude has a lower value than the quadrupole expected from a best - fit @xmath1-dominated cold dark matter ( @xmath1cdm ) standard model to the entire power spectrum since the first data of the differential microwave radiometer ( cobe / dmr ) appeared in 1992 .
this anomaly was confirmed with the high resolution data provided by the first year ( 2003 ) and 3-year ( 2006 ) wilkinson microwave anisotropy probe ( wmap , wmap3 ) .
most recently , the 5-year wmap data did not improve on the quadrupole , it still seems to be outside of the fit .
today , this low value of the quadrupole seems inescapable and the main issue now is to elucidate possible explanations .
the proposed four possible candidates are that this anomaly is due to a systematic error , a pure statistical fluke , astrophysical ( i.e. unexpected foreground ) or cosmological reasons ( e.g. nontrivial spatial geometry of the universe ) .
although the correct explanation is still unknown , the first three candidates seem unlikely .
if it is indeed due to cosmological reasons , since the lowest multipoles represent especially the scale of the horizon at approximately dark energy ( de ) domination , it seems natural to associate the low value of the quadrupole with the nature of the de .
it is known that bianchi universe anisotropies give rise to cmb anisotropies depending on the model type .
recently , campanelli et al . showed that allowing large - scale spatial geometry of the universe to be plane - symmetric with eccentricity ( regardless of the origin ) at decoupling of order @xmath2 can bring the quadrupole amplitude in accordance with observations without affecting higher multipoles of the angular power spectrum of the temperature anisotropy .
although they are controversial , it should be mentioned here that there are also some independent indications of a symmetry axis in the large - scale geometry of the universe , coming from the analysis of spiral galaxies in the sloan digital sky survey ( sdss ) and from the analysis of polarization of electromagnetic radiation propagating over cosmological distances .
a spatially ellipsoidal geometry of the universe can be described with bianchi type metrics .
however , bianchi type i , v , vii models isotropize at late times even for ordinary matter , and the possible anisotropy of the bianchi metrics necessarily die away during the inflationary era .
in fact this isotropization of the bianchi metrics is due to the implicit assumption that the de is isotropic in nature . if the implicit assumption that the pressure of the de is direction independent is relaxed , the isotropization of the bianchi metrics can be fine tuned to generate arbitrary ellipsoidality ( eccentricity ) .
therefore , the cmb anisotropy can also be fine tuned , since the bianchi universe anisotropies determine the cmb anisotropies .
the price of this property of de is a violation of the null energy condition ( nec ) since the de crosses the phantom divide line ( pdl ) , in particular depending on the direction .
however , it has been known since the 1980s that such energy components might occur and their role as possible de candidates was raised by caldwell at the end 1990s ( see for further references ) . in theory , despite the observational constraints , extensions of general relativity are the prime candidate class of theories consistent with pdl crossing . on the other hand , while the current cosmological data from snia ( supernova legacy survey , gold sample of hubble space telescope ) , cmb ( wmap , boomerang ) and large scale structure ( sdss ) data rule out the @xmath3 , they mildly favor dynamically evolving de crossing the pdl ( see , for theoretical and observational status of crossing the pdl ) .
recently , rodrigues and koivisto & mota have investigated cosmological models with anisotropic equation of state ( eos ) .
rodrigues has constructed a bianchi type - i @xmath1cdm cosmological model with a de component which is non - dynamical but wields anisotropic vacuum pressure in two ways : i ) by implementing of anisotropic vacuum pressure consistent with energy - momentum tensor conservation ; ii ) by implementing a poisson structure deformation between canonical momenta such that rescaling of the scale factor is not violated .
he suggests to fine tune the de so as to not wipe out the anisotropic imprints in the inflationary epoch . on the other hand ,
koivisto & mota have proposed a different approach to resolve cmb anisotropy problem ; even if the cmb formed isotropically at early time , it could be distorted by the direction dependent acceleration of the later universe in such a way that it appears to us anomalous at the largest scales .
they have investigated a cosmological model containing a de component which has a non - dynamical anisotropic eos and interacts with the perfect fluid component .
they have also suggested that cosmological models with anisotropic eos can explain the quadrupole problem and can be tested by snia data . in reference mota
et al . have concluded that even though a perfect fluid representation might ultimately turn out to be a phenomenologically sufficient description of all the observational consequences of de , imperfectness in de can not be excluded .
although there is compelling evidence that the expansion of universe is speeding up , we are far from understanding of the nature of the de which is thought be the reason for this behaviour .
hence , we should examine models with anisotropic dark energy , in order to determine what possible new physical consequences they might give rise to , and if for no other reason than to rule such models out .
the above discussions lead us to examine the physical behaviours of the two - fluid , particularly @xmath1cdm , cosmological models with an anisotropic de component .
models of de are conveniently characterized by the eos parameter @xmath4 which is not necessarily constant , where @xmath5 is the energy density and @xmath6 is the pressure . however , while energy density is a scalar quantity , pressure is a vectorial quantity , and consequently the eos parameter of de may be determined separately on each spatial axis in a consistent way with the conservation of energy - momentum tensor .
hence , we consider a phenomenological parametrization of minimally interacting de in terms of its time - dependent deviation - free equation of state parameter @xmath7 and deviation parameters @xmath8 $ ] .
since such a parametrization yields an anisotropic expansion which is not compatible with the robertson - walker ( rw ) metric , in section 2 we have used the locally rotationally symmetric bianchi type i ( lrs bianchi - i ) metric which generalizes the flat rw metric in an axially symmetric way and is compatible with our parametrization .
the other component has assumed to be a perfect fluid ( dark matter or ordinary matter ) . with our assumptions ,
the conservation of the energy - momentum tensor implies dynamically anisotropic de except for very special solutions .
we have obtained exact solutions for the equations by assuming a special dynamic for the anisotropy of the dark the energy and a special law of variation for the mean hubble parameter in bianchi metrics , which yields a constant value of deceleration parameter .
the assumption on the mean hubble parameter allows us to determine the scale factors exactly , as well as to examine the behaviour of the anisotropy of de and other cosmological parameters of such a universe .
the spatially homogenous , anisotropic and lrs bianchi - i space - time is described by the line element @xmath9 where @xmath10 and @xmath11 are the scale factors ( metric tensors ) and functions of the cosmic time @xmath12 . in natural units ( @xmath13 and @xmath14 ) , the field equations , in the case of a mixture of the perfect fluid and the anisotropic de components , are @xmath15 with @xmath16\\ \nonumber = \mathrm{diag}[1,-w^{\mathrm{(m)}},-w^{\mathrm{(m)}},-w^{\mathrm{(m)}}]\rho^{\mathrm{(m)}}\end{aligned}\ ] ] and @xmath17\\ \nonumber = \mathrm{diag}[1,-{w_{x}}^{\mathrm{(de)}},-{w_{y}}^{\mathrm{(de)}},-{w_{z}}^{\mathrm{(de)}}]\rho^{\mathrm{(de)}}\\ \nonumber = \mathrm{diag}[1,-(w^{\mathrm{(de)}}+\delta),-(w^{\mathrm{(de)}}+\gamma),-(w^{\mathrm{(de)}}+\gamma)]\rho^{\mathrm{(de)}}\end{aligned}\ ] ] where @xmath18 ; @xmath19 is the four - velocity vector ; @xmath20 is the ricci tensor ; @xmath21 is the ricci scalar ; @xmath22 and @xmath23 are the energy densities of the perfect fluid and de components respectively ; @xmath24 is the eos parameter of the perfect fluid and we will call @xmath25 the deviation - free eos parameter of the de . here
, the perfect fluid represents the ordinary matter or cold dark matter , thus @xmath26 .
@xmath27 , @xmath28 and @xmath29 are the directional eos parameters of the de on @xmath30 , @xmath31 and @xmath32 axes respectively . @xmath28 and @xmath29 are set to be equal which is convenient with the metric given in ( 1 ) . @xmath33 and @xmath34 are deviations from the deviation - free eos parameter ( hence the deviation - free pressure ) of the de respectively on @xmath30 axis and @xmath31 and @xmath32 axes . in a comoving coordinate system , einstein s field equations ( 2 ) , for the anisotropic lrs bianchi - i space - time ( 1 ) , in case of ( 3 ) and
( 4 ) , read as @xmath35 @xmath36 @xmath37 where the over dot denotes derivation with respect to the cosmic time @xmath12 .
we have the following equation from the bianchi identity , @xmath38 which yields @xmath39 this equation , which is linearly dependent to einstein field equations , also represents the conservation of the total energy momentum tensor .
we have initially 8 variables ( @xmath40 , @xmath41 , @xmath22 , @xmath24 , @xmath23 , @xmath25 , @xmath33,@xmath34 ) and four equations , three of which are linearly independent , namely three einstein field equations ( 5 - 7 ) and the bianchi identity .
the system is thus initially undetermined and we need additional constraints to close the system .
we have assumed that the de is minimally interacting , @xmath42 ; thus due to the bianchi identity ( 8) the perfect fluid component is also minimally interacting , @xmath43 .
hence , the bianchi identity has been split into two separately additive conserved components ; namely , the conservation of the energy - momentum tensor of the de @xmath44 and the conservation of the energy - momentum tensor of the perfect fluid component @xmath45 once more we may split the conservation of the energy - momentum tensor of the dark - enery into two parts : @xmath46 where @xmath47 is the last term of the @xmath48 in ( 10 ) and arises due to the deviations from @xmath25 , and @xmath49 is the the deviation - free part of the @xmath48 in ( 10 ) .
now , we will do the following strong assumption , @xmath50 which also results in the deviation - free part of the @xmath48 to be null , that is , @xmath51 which looks like the conservation of the energy momentum tensor of a minimally interacting perfect fluid . according to ( 13 ) and ( 14 ) the behaviour of @xmath23 is controlled by the deviation - free part of the eos parameter of the de ( @xmath25 ) , but the deviations will affect @xmath23 indirectly , since , as can be seen later , they affect the value of @xmath25 . if the deviation parameters are assumed to be constants , to assure @xmath52 either @xmath33 and @xmath34 are trivially null or the ratio of the expansion rate on @xmath30 axis to the expansion rate on @xmath31 axis is equal to @xmath53 which is a very special case . on the other hand
, we may get more general solutions , if @xmath33 and @xmath34 are allowed to be function of the cosmic time @xmath12 and we constrained @xmath33 and @xmath34 by assuming a special dynamic which is consistent with ( 13 ) .
the dynamic of the deviation parameter on the @xmath30 axis , @xmath54 , is assumed to be @xmath55 and thus from ( 13 ) the deviation parameter on the @xmath31 and @xmath32 axes , @xmath56 , is found as @xmath57 in such an assumption @xmath54 and @xmath56 are dimensionless parameters , and @xmath58 is a dimensionless constant that parametrizes the amplitude of the deviation from @xmath25 and can be given real values .
the measure of the anisotropy of the de may be given by @xmath59 and it is null , which implies the de is isotropic , when @xmath60 .
the eos parameter of the perfect fluid has been assumed to be constant , @xmath61 while @xmath25 is allowed to be a function of the cosmic time , since the current cosmological data from snia , cmb and large scale structures mildly favor dynamically evolving de crossing the pdl as mentioned in the section 1 .
hence , since @xmath62 hasnt been constrained , we still need one more constraint to close the system .
we imposed a law of variation for the hubble parameter .
the law of variation for the hubble parameter which was initially proposed by berman for rw space - time and yields a constant value of deceleration parameter .
such a law of variation for hubble parameter given by berman is not inconsistent with observations and is also approximately valid for slowly time varying deceleration parameter @xcite . recently
singh and kumar proposed a similar law of variation for the hubble parameter in anisotropic space - time metrics that yields a constant value of the deceleration parameter , and generated solutions for bianchi type - i , lrs bianchi type - ii , bianchi type - v @xcite metrics in general relativity . according to the proposed law
, the variation of the mean hubble parameter for the lrs bianchi - i metric may be given by @xmath63 where @xmath64 and @xmath65 are constants .
the spatial volume is given by @xmath66 where @xmath67 is the mean scale factor .
the mean hubble parameter @xmath68 for lrs bianchi - i metric may be given by @xmath69 the directional hubble parameters in the directions of @xmath30 , @xmath31 and @xmath32 respectively may be defined as @xmath70 ( since @xmath71 , in the following they are represented by @xmath72 . )
the volumetric deceleration parameter is @xmath73 on integration , after equating ( 18 ) and ( 20 ) , we get @xmath74 and @xmath75 here @xmath76 and @xmath77 are positive constants of integration .
using ( 18 ) with ( 23 ) for @xmath78 , and with ( 24 ) for @xmath79 mean hubble paremeters are @xmath80 and @xmath81 using ( 23 - 24 ) and ( 19 ) in ( 22 ) we get constant values for the deceleration parameter for the mean scale factor as : @xmath82 and @xmath83 the sign of @xmath84 indicates whether the model accelerate or not .
the positive sign of @xmath84 ( i.e. @xmath85 ) corresponds to decelerating models whereas the negative sign @xmath86 for @xmath87 indicates acceleration and @xmath88 for @xmath89 correponds to expansion with constant velocity . using the deviation parameters ( 15 ) and ( 16 ) , and the mean hubble parameter ( 20 ) in the subtraction ( 6 ) from ( 7 )
we may get @xmath90 after little manipulation . on integration of ( 29 ) by considering ( 25 ) and ( 26 ) we obtain @xmath91 @xmath92 and @xmath93 where @xmath94 is the real constant of integration .
one can observe that @xmath94 and @xmath58 are two parameters that parametrize the difference between the directional hubble parameters .
now , we can find @xmath10 and @xmath11 explicitly for all @xmath95 values by using ( 30 - 32 ) .
using ( 30 ) we may get the ratios of the scale factors @xmath97 , and on manipulating the result by using ( 23 ) we get the following exact expressions for the scale factors : @xmath98 @xmath99 where @xmath100 is the positive constant of integration .
the spatial volume of the universe is found as @xmath101 the directional hubble parameters as defined in ( 21 ) are found as @xmath102 @xmath103 the anisotropy parameter of the expansion @xmath104 is defined as @xmath105 where @xmath106 ( i=1,2,3 ) represents the directional hubble parameters in the directions of @xmath30 , @xmath31 and @xmath32 respectively . by using equations ( 25),(36 ) and ( 37 ) in ( 38 ) we get @xmath107 for the anisotropy of the expansion .
the expansion scalar , defined by @xmath108 , is found as @xmath109 the shear scalar , defined by @xmath110 where @xmath111 is the shear tensor , is found as @xmath112 using the scale factors in ( 11 ) , the energy density of the perfect fluid is found as @xmath113 the energy density of the de may be found from ( 5 ) by using the scale factors and the energy density of the perfect fluid ( 42 ) , and may be written in terms of @xmath114 as @xmath115 using the scale factors and ( 43 ) in ( 14 ) we get @xmath116 for the deviation - free eos parameter of the de . using the scale factors and ( 43 ) in ( 15 ) and ( 16 ) we may get deviation parameters as following , @xmath117 @xmath118 for this model @xmath119 and @xmath120 , which implies the greatest value of the hubble parameter and the fastest rate expansion of the universe .
thus , this model may represent the inflationary era in the early universe and the very late times of the universe .
the spatial volume @xmath121 is finite at @xmath122 , expands exponentially as @xmath12 increases and becomes infinitely large at @xmath123 .
the directional hubble parameters @xmath124 and @xmath72 are finite at @xmath122 and @xmath125 .
they deviate from the mean hubble parameter @xmath68 due to @xmath94 and @xmath58 . while @xmath94 is supporting ( opposing ) the expansion on the @xmath30 axis , it opposes ( supports ) the expansion on @xmath31 and @xmath32 axes .
however it loses its effect exponentially by the cosmic time @xmath12 .
the anisotropy of the de has a similar effect on the directional hubble parameters , but its effect persists . while @xmath58 is supporting ( opposing ) the expansion on the @xmath30 axis , it opposes ( supports ) the expansion on @xmath31 and @xmath32 axes .
the anisotropy of the de does nt always act so as to increase the anisotropy of the expansion .
in fact , when the signs of @xmath94 and @xmath58 are opposite the overall effect is to lower the expansion anisotropy .
the expansion scalar is constant throughout the evolution of the universe .
the shear scalar is also finite at @xmath122 and tends to @xmath126 as @xmath12 increases . if @xmath127 all the axes will expand to infinitly large values as @xmath128 . on the other hand ,
the space - time exhibits a pancake type singularity for @xmath129 and a cigar type singularity for @xmath130 at @xmath131 . if @xmath132 while the @xmath30 axis converges to a constant , @xmath31 and @xmath32 axes expand to infinitly large values , and it is vice versa if @xmath133 . versus cosmic time @xmath12 for different values of @xmath58 in model @xmath78 . @xmath94 and @xmath134
have been chosen as @xmath135.,scaledwidth=100.0% ] versus cosmic time @xmath12 for different values of @xmath58 in model @xmath78 . @xmath94 and @xmath134
have been chosen as @xmath135.,scaledwidth=100.0% ] for different values of @xmath58 in model @xmath78 .
@xmath94 and @xmath134 have been chosen as @xmath135.,scaledwidth=100.0% ] for different values of @xmath58 in model @xmath78 .
@xmath94 and @xmath134 have been chosen as @xmath135.,scaledwidth=100.0% ] the energy density of the perfect fluid @xmath22 decreases exponentialy and converges to zero since @xmath136 by definition , while @xmath23 changes slightly at early times and converges to a non - zero value as @xmath12 increases .
thus , the ratio of @xmath137 converges to @xmath135 as @xmath12 increases , that is the de dominates the perfect fluid in the inflationary era as expected .
since de dominates the perfect fluid in the inflationary era as mentioned above , we may neglect the perfect fluid while examining the properties of the de in this model .
@xmath23 is always positive provided @xmath138 and @xmath139 , begins with a finite value at @xmath122 , exhibits different behaviours depending on the choice of the parameters at the earlier times and tends to @xmath140 for large values of @xmath12 ; see fig [ fig : rhom0 ] .
it can be observed that the bigger the anisotropy of the expansion the lower the @xmath23 in any given instant .
@xmath94 loses its effect exponentially as @xmath12 increases , thus we may say that the higher the @xmath141 the lower the @xmath23 in any given instant for relatively big @xmath12 values .
the eos parameter of the de @xmath25 may begin in phantom ( @xmath142 ) or quintessence ( @xmath143 ) region and tends to @xmath144 ( cosmological constant , @xmath145 ) by exhibiting various patterns as @xmath12 increases ; see fig [ fig : wm0 ] .
one can observe that @xmath23 increases when @xmath146 , decreases when @xmath147 and is constant when @xmath148 , as would be expected .
the anisotropy of the expansion decreases monotonically as @xmath12 increases when @xmath60 .
however , it exhibits nontrivial behaviour at the early times of the universe and converges to a non - zero value for the late times when @xmath149 ; see fig [ fig : deltam0 ] .
the deviation parameters @xmath33 and @xmath34 are finite at @xmath122 , and converge to @xmath150 and @xmath151 respectively as @xmath152 .
the anisotropy of the de which has been defined as @xmath153 tends to @xmath154 as @xmath12 increases ; see fig [ fig : adem0 ] .
the solutions in this subsection are valid for all possible values of @xmath95 except for @xmath156 and @xmath78 , thus the solutions for @xmath156 are given in the following subsection . from ( 24 )
one can see that the initial time of the universe is @xmath157 for @xmath158 . for brevity of the equations
, we may redefine the cosmic time as @xmath159 and by doing that the initial time of the universe has also been set to @xmath160 .
thus we may rewrite the metric as @xmath161 using ( 31 ) we may obtain the ratios of the scale factors @xmath97 , and on manipulating the result by using ( 24 ) we get the following exact expressions for the scale fators : @xmath162 @xmath163 where @xmath100 is the positive constant of integration . the spatial volume of the universe is found as @xmath164 the directional hubble parameters as defined in ( 21 ) are found as @xmath165 @xmath166 using ( 26 ) , ( 52 ) and ( 53 ) in ( 38 ) we get @xmath167 for the anisotropy parameter of the expansion .
the expansion and shear scalars are , respectively , found as @xmath168 @xmath169 using the scale factors ( 49 - 50 ) in ( 11 ) , the energy density of the perfect fluid is found as @xmath170 the energy density of the de can be found from ( 5 ) by using the scale factors ( 49 - 50 ) and the energy density of the perfet fluid ( 57 ) , and may be written in terms of @xmath114 as @xmath171 using ( 49 - 50 ) and ( 58 ) in ( 14 ) we get @xmath172 for the deviation - free eos parameter of the de . and finally using equations ( 49 - 50 ) and ( 58 ) in ( 15 ) and ( 16 ) we may get deviation parameters as following , @xmath173 @xmath174 the universe accelerates for @xmath175 , decelerates for @xmath85 and expands with constant velocity for @xmath89 .
this model may represent the radiation dominated era for @xmath176 and @xmath177 , and the matter dominated era for @xmath178 and @xmath179 .
the mean hubble parameter @xmath68 , expansion scalar @xmath180 and shear scalar @xmath181 are infinitly large at @xmath182 , and null at @xmath183 .
if @xmath184 , the space time exhibits a cigar type singularity at @xmath183 for @xmath185 and a pancake type singularity for @xmath186 . if @xmath187 , all the axes expand to infinitly large values provided @xmath188 as @xmath189 .
when @xmath190 , @xmath191 takes infinitly large values as @xmath189 , while @xmath192 converges to @xmath193 .
when @xmath194 , @xmath192 takes infinitly large values as @xmath189 , while @xmath191 converges to @xmath195 . when @xmath196 space time exhibits a cigar type singularity at @xmath183
when @xmath197 space time exhibits a pancake type singularity at @xmath183 .
@xmath104 diverges as @xmath198 , converges to a constant as @xmath199 for @xmath187 and vice versa for @xmath184 .
one can observe that the anisotropy of the expansion lowers the value of @xmath23 .
thus , @xmath104 must be considered while examining the dynamics of the energy density of the de , particularly since @xmath104 diverges in the above mentioned limits . in the model for @xmath187 , @xmath23 will eventually obtain negative values as @xmath200 goes to zero due to the divergence of @xmath104 . hence
, this model is not appropriate for representing the relatively earlier times of the universe .
on the other hand , this model can set to represent relatively later times of the universe since we can always set @xmath23 to be @xmath201 as @xmath199 by choosing the parameters in suitable intervals .
similarly , in the model for @xmath184 , @xmath23 will eventually obtain negative values as @xmath199 .
hence , this model is not appropriate for representing the late times of the universe , but for the earlier times of the universe by choosing the parameters in suitable intervals .
@xmath25 and @xmath153 , say the anisotropy of the de , exhibits various dynamics according to the choice of the parameters .
it is worth to mention here once more that the anisotropy of the de does nt always act so as to increase the anisotropy of the expansion , when the signs of @xmath94 and @xmath58 are opposite the overall effect is to lower the expansion anisotropy . versus cosmic time @xmath200 in the vicinity of @xmath202 in the model @xmath203 for different values of @xmath58 .
@xmath104 has been set to null at @xmath202 by choosing @xmath204 .
@xmath205,scaledwidth=100.0% ] versus cosmic time @xmath200 in the vicinity of @xmath202 in the model @xmath203 for different values of @xmath58 .
@xmath104 has been set to null at @xmath202 by choosing @xmath204 .
@xmath205,scaledwidth=100.0% ] versus cosmic time @xmath200 in the vicinity of @xmath202 in the model @xmath203 for different values of @xmath58 .
@xmath104 has been set to null at @xmath202 by choosing @xmath204 .
@xmath206,scaledwidth=100.0% ] versus cosmic time @xmath200 in the vicinity of @xmath202 in the model @xmath203 for different values of @xmath58 .
@xmath104 has been set to null at @xmath202 by choosing @xmath204 .
@xmath206,scaledwidth=100.0% ] in the vicinity of @xmath202 in the model @xmath203 for different values of @xmath58 .
@xmath104 has been set to null at @xmath202 by choosing @xmath204.,scaledwidth=100.0% ] in the vicinity of @xmath202 in the model @xmath203 for different values of @xmath58 .
@xmath104 has been set to null at @xmath202 by choosing @xmath204.,scaledwidth=100.0% ] we may try to set @xmath202 to the present universe by using the recent cosmological data which favor @xmath207 ( dimensionless hubble parameter ) , @xmath208 ( density parameter of the dark energy ) , @xmath209 , @xmath210 , @xmath179 and a space - time that can be represented by flat rw metric which implies @xmath211 for the present universe , where @xmath212 subscripts represent present ( @xmath202 ) values of the parameters . with above parameters we may obtain present critical energy density of the universe as @xmath213 , the deceleration parameter as @xmath214 which implies @xmath203 in our model , @xmath215 , @xmath216 , @xmath217 and @xmath211 by choosing @xmath204 . in this model with the above values of the parameters , which may correspond to the present universe in the vicinity of @xmath202 , @xmath22 is roughly proportional with @xmath218 , while @xmath23 is roughly proportional with @xmath219 and gets its highest value at @xmath202 where @xmath220 .
that is , the de eventually dominates the perfect fluid .
some graphs may be plotted with above values , since we can think that @xmath84 is changing only slightly for the present universe , for @xmath60 , @xmath221 and @xmath222 .
plots has been extended towards to the later times more than towards to the earlier times since , as mentioned above , models with @xmath187 are not appropriate for representing relatively earlier times of the universe .
@xmath23 increases when @xmath25 is in the phantom region , begins to decrease when @xmath25 passes into the quintessence region and converges to zero as @xmath199 ; see fig [ fig : rhodem ] .
@xmath25 begins in the phantom region , increases and becomes @xmath144 at @xmath202 then passes into the quintessence region and tends to a constant which is in the quintessence region ; see fig [ fig : wm ] .
@xmath22 decreases as @xmath200 increases independent of @xmath58 ; see fig [ fig : rhomm ] .
@xmath223 increases as @xmath200 increases , becomes @xmath224 at @xmath202 and converges to @xmath135 as @xmath200 keeps on increasing ; see fig [ fig : omegam ] .
@xmath104 decreases as @xmath200 increases and becomes null at @xmath202 then tends to a constant which depends on @xmath58 , see fig [ fig : deltam ] .
the bigger the @xmath104 the lower the @xmath23 , consequently lower the @xmath223 in any given instant .
the anisotropy of the de , @xmath153 , is null for @xmath60 , changes slightly and converges to a non - zero constant for @xmath225 ; see fig [ fig : adem ] . from ( 24 )
we can see that the initial time of the universe is @xmath227 for @xmath156 . for brevity of the equations
, we may redefine the cosmic time as @xmath228 and by doing that the initial time of the universe has also been set as @xmath160 .
thus we may rewrite the metric as @xmath229 using ( 32 ) we may get the ratios of the scale factors @xmath97 for @xmath156 , and manipulating the result by using ( 24 ) we get the following exact expressions for the scale fators ; @xmath230 @xmath231 where @xmath100 is the positive constant of integration .
the spatial volume of the universe is found as @xmath232 the directional hubble parameters as defined in ( 21 ) are found as @xmath233 @xmath234 using ( 26 ) and the scale factors in ( 38 ) we get @xmath235 for the anisotropy parameter of expansion .
the expansion and shear scalars are , respectively , found as @xmath168 @xmath236 using the scale factors in ( 11 ) , the energy density of the perfect fluid is found as @xmath237 the energy density of the de can be found from ( 5 ) by using the scale factors and the energy density of the perfet fluid ( 72 ) , @xmath238 using ( 64 - 65 ) and ( 73 ) in ( 14 ) we get @xmath239 for the deviation - free eos parameter of the de . and finally using equations ( 64 - 65 ) and ( 73 ) in ( 15 ) and ( 16 ) we may get deviation parameters as following , @xmath240 @xmath241 nothing physically special has been observed in this model for further investigation .
but we may mention that this model may only be valid for intermediate epochs of the universe . because @xmath104 diverges at @xmath182 and @xmath183 in this model
, thus @xmath23 will eventually get negative values as @xmath200 goes to extrem values , since the anisotropy of the expansion contribute the energy density of the de negatively .
locally rotationally symmetric bianchi - i cosmological models with dynamically anisotropic dark energy ( de ) and perfect fluid have been constructed in general relativity .
we assume that the dark energy ( de ) is minimally interacting , has dynamical energy density and anisotropic equation of state parameter ( eos ) .
the conservation of the energy - momentum tensor of the de has been assumed to consist of two separately additive conserved parts .
a special law has been assumed for the deviation from isotropic eos , which is consistent with the assumption on the conservation of the energy - momentum tensor of the de .
exact solutions of einstein s field equations have been obtained by assuming a special law of variation for the mean hubble parameter , which yields a constant value of the deceleration parameter and is not inconsistent with observations .
some basic geometrical and kinematical features of the models and the dynamics of the anisotropic de in these models have been examined .
a more general eos parameter has been introduced for the de , and the isotropic de can be recovered by choosing @xmath58 to be null , where @xmath58 parametrizes the amplitude of the deviation from isotropic eos parameter of the de .
in all the models , while the anisotropy of the de contributes to the expansion of one of the scale factors , it opposes to the expansion of the other .
two parameters that parametrize the difference between directional hubble parameters , @xmath94 and @xmath58 , give rise to non - trivial , dynamically anisotropic , expansion histories .
this allows for the possibility to fine tune the isotropization of the bianchi metric during an accelerating epoch of the universe in order to generate arbitrary ellipsoidality by choosing a suitable value of @xmath58 .
such a result provides also the possibility to fine tune the cmb anisotropy .
the anisotropy of the expansion can mildly or totally isotropize in relatively earlier times of the universe .
nevertheless , it converges to a nonzero constant value for the later times of the universe in all models .
it is also observed that the anisotropy of the de energy does nt always act so as to increase the anisotropy of the expansion , when signs of @xmath94 and @xmath58 are opposite the overall effect is to lower the expansion anisotropy in relatively earlier times of the universe .
one interesting observation is that the higher the anisotropy of the expansion the lower the energy density of the de in each given instant of the cosmic time .
@xmath25 is also dynamical and exhibits non - trivial behaviour in our model , but it eventually converges to a constant in the quintessence region for @xmath242 and converges to @xmath144 for @xmath78 .
such cosmological models are of interest because they give rise to an ellipsoidality of the universe in spite of the inflation , which is one of the promising proposals to the solution of the quadrupole problem and can also be checked by the direction dependency of the redshift - luminosity relation of the snia observations .
guth , a. h. , physical review d , 23 , 347 ( 1981 ) sato , k. , monthly notices of the royal astronomical society , 195 , 467 ( 1981 ) linde , a. d. , physics letters b , 108 , 389 ( 1982 ) albrecht , a. and steinhardt , p. j. , physical review letters , 48 , 1220 ( 1982 ) parker , l. , physical review , 183 , 1057 ( 1969 ) birrell , n. d. and davies , p. c. w. , quantum fields in curved space , cambridge : cambridge univ . press ( 1982 )
mukhanov , v. f. and chibisov , g. v. , soviet phys.journal of experimental and theoretical physics letters , 33 ( 1981 ) hawking , s. w. , physics letters b , 115 , 295 ( 1982 ) guth , a. h. , and pi , s. y. , physical review letters , 49 , 1110 ( 1982 ) starobinsky , a. a. , physics letters b , 117 , 175 ( 1982 ) bardeen , j. m. , et al . , physical review d , 28 , 679 ( 1983 ) mukhanov , v. f. , et al . ,
physics reports , 215 , 203 ( 1992 ) peiris , h.v .
, et al . , the astrophysical journal supplement series , 148 , 213231 ( 2003 ) [ arxiv : astro - ph/0302225v2 ] liddle , a. r. and lyth , d. h. , cosmological inflation and large - scale structure , cambridge : cambridge university press ( 2000 ) smoot , g.f . , et al .
, astrophysical journal , 396 , l1 ( 1992 ) bennett , c.l .
, et al . , astrophysical journal , 464 , l1l4 , ( 1996 ) [ arxiv : astro - ph/9601067 ] hinshaw , g. , et al . , the astrophysical journal supplement series , 148 , 135 ( 2003 ) [ arxiv : astro - ph/0302217v1 ] hinshaw , g. , et al . , the astrophysical journal supplement series , 170 , 288 ( 2007 ) [ arxiv : astro - ph/0603451v2 ] nolta , m.r . ,
et al . , arxiv : astro - ph/0803.0593v1 ( 2008 ) hinshaw , g. , et al .
, arxiv : astro - ph/0803.0732v1 ( 2008 ) eriksen , h.k . , et al . , astrophysical journal .
, 605 , 1420 , ( 2004 ) [ arxiv : astro - ph/0307507 ] copi , c.j .
, monthly notices of the royal astronomical society , 367 , 79 - 102 , ( 2006 ) [ arxiv : astro - ph/0508047 ] spergel , d.n . , et al . , the astrophysical journal supplement series , 170 , 377408 ( 2007 ) [ arxiv : astro - ph/0603449 ] gold , b. , et al .
, arxiv : astro - ph/0803.0715v1 hill , r.s . , et al .
, arxiv : astro - ph/0803.0570v1 ellis , g.f.r . , general relativity and gravitation , 38(6 ) , 1003 - 1015 ( 2006 ) campanelli , l , et al .
, physical review letters , 97 , 131302 ( 2006 ) campanelli , l , et al .
, physical review d , 76 , 063007 ( 2007 ) [ arxiv : astro - ph/0706.3802v2 ] longo , m. j. , arxiv : astro - ph:/0703325 , arxiv : astro - ph:/0703694 , arxiv : astro - ph:/0707.3793 , arxiv : astro - ph:/0708.4013 nodland , b. and ralston , j.p .
, physical review letters 78 , 3043 - 3046 ( 1997 ) [ arxiv : astro - ph/9704196 ] hutsemekers , et al . ,
astronomy and astrophysics , 441 , 915 - 930 ( 2005 ) [ arxiv : astro - ph/0507274 ] battye , r. a. , et al .
, monthly notices of the royal astronomical society , doi:10.1111/j.1365 - 2966.2007.12785.x ( 2008 ) borguet , b. , et al . ,
astronomy and astrophysics , 478 , 321 - 333 ( 2008 ) [ arxiv : astro - ph:/0710.4048v1 ] borguet , b. , et al .
, arxiv : astro - ph:/0710.4072v2 joshi , s.a . , et al .
, monthly notices of the royal astronomical society , 380 ( 1 ) , 162174 ( 2007 ) [ arxiv : astro - ph:/0705.2548v1 ] rakic , a. and schwarz d.j . , physical review d , 75 , 103002 ( 2007 ) [ arxiv : astro - ph/0703266v3 ] rakic , a. , et al . , arxiv : astro - ph:/0609188v2 bonometto , s. , et al .
, modern cosmology , 138 - 139 ( ellis , g.f.r . ) , institute of physics publishing , bristol and philadelphia ( 2002 ) rodrigues , d.c .
, physical review d , 77 , 023534 ( 2008 ) [ arxiv : astro - ph/0708.1168v2 ] carroll , s.m . and hoffman , m. , physical review d , 68 , 023509 ( 2003 ) [ arxiv : astro - ph/0301273 ] nesseris , s. and perivolaropoulos , l. , journal of cosmology and astroparticle physics , 0701 , 018 ( 2007 ) [ arxiv : astro - ph/0610092v3 ] riess , a. g. , et al . , astrophysical journal , 607 , 665 - 687 ( 2004 ) [ arxiv : astro - ph/0402512 ] astier , p. , et al . ,
astronomy and astrophysics , 447 , 31 - 48 ( 2006 ) [ arxiv : astro - ph/0510447 ] komatsu , e. , et al .
, arxiv : astro - ph/0803.0547v1 mactavish , c.j . ,
et al . , astrophysical journal , 647 , 799 ( 2006 ) [ arxiv : astro - ph/0507503v1 ] eisenstein , g. , et al . , the astrophysical journal , 633 , 560 - 574 ( 2005 ) [ arxiv : astro - ph/0501171 ] zhao , g.b . , et al . , physics letters b , 648 , 8 - 13 ( 2007 ) [ arxiv : astro - ph/0612728v2 ] copeland , e.j . , et al . , international journal of modern physics d , 15 , 1753 - 1936 ( 2006 ) [ arxiv : hep - th/0603057v3 ] koivisto t. , mota , d.f . , arxiv : astro - ph/0801.3676v2 ( 2008 ) koivisto t. , mota , d.f . , astrophysical journal , 679:1 ( 2008 ) [ arxiv : astro - ph/0707.0279v3 ] mota , d.f . , et al . , monthly notices of the royal astronomical society , 382 , 793 - 800 ( 2007 ) [ arxiv : astro - ph/0708.0830v2 ] burdyuzha , v. , vereshkov , g. , astrophysics and space science 305 , 235239 ( 2006 ) turner , m. , huterer , d. , journal of the physical society of japan , 76 , ( 2007 ) [ arxiv : astro - ph/0706.2186v2 ] singh , c.p . , and kumar , s. , , international journal of modern physics d , 15 , 419 - 438 ( 2006 ) kumar , s. and singh , c.p . ,
astrophsyics and space science , 312 , 57 - 62 ( 2007 ) singh , c.p .
, pramana - journal of physics , 68 , 707 - 720 ( 2007 ) singh , c.p .
, et al . ,
astrophysics and space science , 315 , 181 - 189 ( 2008 ) berman , m.s . , nuovo cimento b , 74 , 182 - 186 , ( 1983 ) berman , m.s . and gomide , f.m . ,
general relativity and gravitation , 20 , 191 - 198 ( 1988 ) | locally rotationally symmetric ( lrs ) bianchi type i cosmological models are examined in the presence of dynamically anisotropic dark energy and perfect fluid .
we assume that the dark energy ( de ) is minimally interacting , has dynamical energy density , anisotropic equation of state parameter ( eos ) .
the conservation of the energy - momentum tensor of the de is assumed to consist of two separately additive conserved parts .
a special law is assumed for the deviation from isotropic eos , which is consistent with the assumption on the conservation of the energy - momentum tensor of the de .
exact solutions of einstein s field equations are obtained by assuming a special law of variation for the mean hubble parameter , which yields a constant value of the deceleration parameter .
geometrical and kinematic properties of the models and the behaviour of the anisotropy of the dark energy has been carried out .
the models give dynamically anisotropic expansion history for the universe that allows to fine tune the isotropization of the bianchi metric , hence the cmb anisotropy .
-1 cm _ ege university , faculty of science , dept . of astronomy and space sciences , 35100 bornova , izmir / turkey .
_ * e - mail : * [email protected]_ ( _ [email protected]_ ) , [email protected]_ |
combined studies of the cosmic microwave background radiation , supernova cosmology and large galaxy redshift surveys provide nowadays compelling evidence of the existence of non - baryonic dark matter ( dm ) @xcite .
consistency with the observed structure of the universe , especially at large ( above galactic ) scales favors cold dark matter ( cdm ) , i.e. the particles constituting the cosmologically required dark matter have to be moving non - relativistically .
this property is always fulfilled by particles of mass in the gev to tev region that interact with the weak interaction strength , i.e. weakly interacting massive particles ( wimps ) .
wimps can annihilate or decay to detectable standard model particles , in particular gamma - rays , giving rise to the possibility of `` indirect detection '' @xcite .
a standard approach is to look for spectral signatures in the energy spectrum of gamma - rays from celestial objects , either targeting the continuum spectrum produced by decay of secondary pions or the more feeble signature of a line due to dm annihilating directly into photons .
another possible experimental signature is given by differences in the spatial distribution of gamma - ray signals induced by dm as compared to conventional astrophysical sources .
in particular , the small scale fluctuations in an otherwise isotropic gamma - ray background ( igrb ) might be different in the case of annihilating dm , since its flux scales as the squared density .
after the pioneering study of ando and komatsu @xcite , followed by two other more detailed works @xcite , the issue of dm annihilation and anisotropies in the isotropic gamma background has raised increasing interest . in connection to the extra - galactic gamma - ray background , anisotropy has been further studied in @xcite while predictions from the millenium ii simulation have recently been described in zavala et al .
it has also been realized that galactic substructures can produce an almost isotropic gamma background with similar fluctuation properties as the extragalactic background and with promising chances of detection @xcite ; see also @xcite ) .
other works have investigated the anisotropy pattern resulting from both the galactic and extragalactic contribution @xcite or from the the dm annihilation around intermediate mass black holes @xcite .
anisotropies from dm annihilation appearing in the radio sky have been investigated in @xcite .
finally, apart from the power spectrum an interesting approach is to investigate the probability distribution of the fluctuations in order to detect possible departures from gaussianity ( or , rather , from poisson statistics , in the common situation of low number of counts ) , thereby providing another handle on the separation of conventional astrophysical and dm contribution @xcite .
anisotropy properties of astrophysical processes other than dm contributing to the isotropic gamma background include resolved point sources @xcite , galaxy clusters and shocks @xcite and emission from normal galaxies @xcite .
the fermi large area telescope ( fermi - lat ) , which was launched june 11 , 2008 , is the most suitable tool for these kind of studies due to its large field of view ( @xmath6 2.4 sr ) , energy coverage from 100 mev to 300 gev @xcite and very good angular resolution .
we will thus focus in the following on the prospect to detect dm anisotropies with fermi - lat .
all of the above studies assume either that the impact of galactic foregrounds on the dm induced anisotropies is negligible or that the galactic foregrounds can be cleaned at a suitable level to allow dm anisotropies studies .
further , the effects of the fermi - lat detector are generally taken into account only in an approximate way , though a detailed representation of the fermi instrument is publicly available .
the aim of this work is to address the above issues in detail .
in particular our approach will be the following : we include the galactic foregrounds explicitly using a model based on the galprop code @xcite and we then use this model to build masks to select the most promising regions in the sky suitable for the analysis of the anisotropies . in these regions
we study the anisotropies in the scenario in which a dm contribution is present over the galactic foregrounds and the extra - galactic emission from blazars compared to the case without dm and we derive the sensitivity to the dm signal employing anisotropy studies with fermi - lat data .
both the cases of galactic dm in substructures and extra - galactic dm are considered separately .
the effect of the fermi - lat response , including the effects of the point spread function , the effective area , as well as non - uniformities in exposure , are included in all calculations at a realistic level by using the instrument response functions provided by the fermi science tools .
this approach is suitable for a sensitivity study like the one presented here , and does not aim to fully reproduce a real data analysis for which foreground cleaning ( not performed in this study ) most likely will be required .
one of the findings of our analysis is indeed that even with a very generous masking , foregrounds still have a sizeable contribution to the overall anisotropies and represent an important limitation to dm sensitivity .
this implies that foreground cleaning will be a crucial step in the data analysis pipeline . on the other hand ,
if cleaning can be achieved in an effective way , this also means that the sensitivity to dm can be further improved with respect to the results derived in this paper .
our results can thus be regarded as conservative from this point of view .
foreground cleaning will be the subject of forthcoming investigations .
the paper is structured as follows . in section
ii we introduce and discuss the various components contributing to the galactic and extragalactic gamma - ray emission . in sections
iii and iv we review their anisotropy properties and we describe the tools and methods for their analysis . in section v we briefly describe the fermi - lat instrument and the characteristics relevant for our analysis . in section
vi we discuss the results of our data simulation pipeline . finally ,
in section vii we evaluate fermi s sensitivity to distinguish a scenario in which the diffuse gamma - ray emission has a dm contribution ( either extragalactic or galactic in origin ) with respect to the case in which dm is absent .
conclusions are presented in section viii .
in the following , for clarity , we will refer to the isotropic experimentally detected emission as the isotropic gamma - ray background ( igrb ) given that in principle its origin may be not fully extra - galactic .
indeed , we will also consider the possibility of a nearly isotropic emission from dm annihilation in the milky way substructures .
we will use the term extra - galactic gamma background ( egb ) when referring to predictions from _ models _ of extra - galactic emission either from dm annihilation or from astrophysical processes .
a first detection of the igrb in the gev range was reported by the egret experiment @xcite ( see also the analysis by strong , moskalenko and reimer @xcite ) .
the recent measurement from the fermi collaboration @xcite finds that the igrb energy spectrum has an almost perfect power law with slope @xmath7 which is approximately what is expected from the contribution of unresolved blazars thus suggesting an extragalactic origin of the igrb . a ( possibly incomplete ) list of contributions to the igrb
is given by the emission in astrophysical processes from blazars , agns , starburst galaxies , star forming galaxies , galaxy clusters , clusters shocks , gamma - ray bursts @xcite .
a contribution is also expected from ultra high energy photons and electrons cascading to gev - tev energies through interactions with the cmb or the infrared - optical background @xcite .
each of these source classes is expected to produce anisotropies although all them should more or less trace the filamentary pattern of the cosmological large scale structures .
anisotropies will be discussed in more detail in the next two sections . apart from the astrophysical contributions , an exotic component arising from wimp annihilation could also be expected . in the case of
uniformly distributed dm , the annihilation flux for a thermal relic cross section ( @xmath8@xmath3s@xmath4 ) is many orders magnitude below the igrb .
however , given that the dm distribution on cosmological scales could be very clumpy , the enhancement factor ( denoted as @xmath9 ) due to the quadratic dependence of the annihilation signal on the dm density has to be taken into account .
this enhancement is typically several orders of magnitude , thus boosting the dm signal to a level comparable to the igrb @xcite .
the exact boost depends on the details of the modeling of the dm clustering , the dm halo profile and the structure formation history .
we fix the absolute normalization of the dm spectra following the dm haloes clustering model described in ullio et al.(2002 ) .
two versions of the model are employed : a conservative version assuming a navarro - frenk - white ( nfw ) profile @xcite for the haloes and no substructures , which gives a low dm normalization , and a more optimistic version where the haloes still follow a nfw profile but the dm signal is enhanced by the presence of substructures in the dm haloes .
it should be noted that the clustering properties of the dm haloes are still very uncertain , especially at small scales ( approximately below the typical galactic halo scale ) , and even larger normalizations are possible .
for example recent results from the millenium - ii simulation @xcite indicate that , depending on the extrapolation employed for the low mass haloes ( non resolved in the simulation ) , an order of magnitude enhancement with respect to the above `` optimistic '' case can be achieved .
overall , place the possible values of the boost @xmath10@xmath11@xmath12 in the range @xmath13 to @xmath14 ( see also @xcite , in particular fig.1 ) . the main contribution to the galactic foregrounds
is given by the decay of pions produced in the interaction of cosmic ray ( cr ) nucleons , by the inverse compton ( ic ) scattering of cr electrons on the interstellar radiation field ( isrf ) and by bremsstrahlung of cr electrons on the interstellar gas .
a further contribution from pulsars has also been considered recently @xcite . according to this analysis , the contribution from galactic millisecond pulsars can be quite isotropic ( i.e. can contribute to high galactic latitudes ) and constitute a large fraction of the igrb .
however , lacking more precise studies of the subject , we will neglect this contribution for the present analysis .
the normalization of the galactic foregrounds is clearly important to identify the regions where the igrb is dominant , i.e. the regions most promising for the anisotropy analysis .
we employ the so called `` conventional model '' @xcite which is derived under the assumption that the cr electron and nucleon spectra in the galaxy can be normalized by the local ( solar system ) measurements .
this model represents a nice fit of the fermi data , at least at galactic latitudes @xmath15 , where the emission is mostly local in origin ( coming within a few kpc s of the solar position ) @xcite .
we will , indeed , focus on this region of the sky in the following .
the galactic center region and the galactic plane are anyway strongly dominated by the galactic emission and they are thus not suitable for the anisotropy analysis .
a description of the diffuse emission in terms of the conventional model is thus accurate enough for the present analysis .
the number of resolved point sources has grown considerably in the fermi era relative to the roughly 270 in the 3rd egret catalogue @xcite .
the first year fermi catalogue includes 1451 sources with a significance larger than about 4 sigma @xcite .
we include explicitly these sources in our analysis .
as will be clear in the following , it is important to include and , in addition , to model this component correctly since it is a considerable source of anisotropy which needs to be reduced to gain sensitivity to the anisotropies of the diffuse component .
while we use the first year catalogue , in the following we produce a forecast over 5 years of data taking , thus should in principle be using an associated 5 years catalogue .
the effect of using a five year catalogue , however , is expected to be small : the masked area will increase , but on the other hand the poisson noise will be reduced .
both effects are small since increasing the data taking time for the catalogue will add sources on the faint end , requiring small masking area and producing a small decrease in poisson noise ( since the brightest sources are already removed ) .
on the other hand , modeling the 5 year catalogue requires an assumption on the nature of the unresolved sources , introducing a model uncertainty .
we therefore decided to use the 1 year catalogue only .
part of the igrb is made of _ unresolved point sources_. it is still a matter of debate if unresolved point sources can make up all the igrb or only part of it .
this issue is likely to become clearer when population studies of the fermi _ resolved point sources _ become available .
present studies indicate that unresolved blazars can only account for at most 30 % of the igrb measured by the fermi - lat @xcite .
unresolved point sources can also be a further source of anisotropy in the igrb itself , producing a poisson noise like contribution to the anisotropy spectrum .
it is worthwhile to point out that this poisson noise is different from the poisson shot noise in the angular power spectrum ( see below and appendix a ) , which is related to the finite number of counts .
the contribution of unresolved point sources depends both on the ability of the fermi - lat to detect point sources and remove them from the diffuse emission and on the intrinsic number density of the sources themselves .
both these terms , in turn , are quite model dependent but they are expected to be reasonably constrained from population studies of source classes with resolved members . for the moment , for the sake of studying the importance of the effect , we consider , beside the model described in the next section , a model fully made of unresolved point sources . for the poisson noise
we assume a value of @xmath16 which is in the middle between the estimate of @xmath17 @xcite ( see their fig.4 ) and the preliminary measurement of @xmath18 from the fermi collaboration @xcite .
it is also worth noting that the poisson noise of unresolved sources is a time varying term slowly decreasing in time .
the brighter unresolved sources , in fact , get finally resolved , as fermi continues to collect more statistics , and once they are removed , and/or masked , the `` new '' remaining unresolved signal has a slightly lower level of poisson noise .
a detailed study of this effect requires extrapolation from population studies which we leave for future work . in figure
[ fig : diffuse ] we show the energy spectrum of the various galactic and extra - galactic components for the model we employ .
in particular , we show the average spectrum in the regions @xmath19 and @xmath20 , where the igrb has the highest relative contribution . we show the power law fit to of the igrb from @xcite . also shown
is an example of extragalactic wimp emission normalized to the model of ullio et al .
( 2002 ) ( with nfw halo profile and no halo substructures ) .
the contribution from resolved point sources detected by fermi @xcite,@xcite is also present .
it can be seen that , combined , these sources constitute a significant fraction of the diffuse background . for simplicity
, the spectrum of each single source is modeled as a simple power law and thus the behavior at high energies ( @xmath21 gev ) may have some biases with respect to the real spectrum .
[ cols="^ " , ] we can see the results in fig . [
fig : limitsa ] , where the sensitivities are reported for various annihilation channels .
the absolute normalization of the dm spectra are obtained following the dm haloes clustering model of @xcite as described in the introduction .
two versions of the model are shown : a conservative version with a nfw profile for the haloes and no substructures which gives a low dm normalization , and thus weaker constraints , and a more optimistic version where the dm signal is enhanced by the presence of substructures in the dm haloes . in principle , using the recent results of@xcite from the millenium - ii simulation , an order of magnitude enhancement with respect to the above `` optimistic '' case can be achieved with a correspondingly better sensitivity ( see also @xcite , in particular fig.1 ) .
the various channels have generally similar behavior .
the @xmath22 channel produces quite smooth limits .
the @xmath23 and @xmath24 sensitivity curves in contrast have a steeper slope and exhibit more structure .
this is due to the fact that for these channels the gamma emission is concentrated at higher energies in a narrow peak near the energy corresponding to the dm mass .
the sensitivity is thus more sensitive to the coarse binning in energy chosen for this analysis .
the sensitivity in the @xmath24 channel and especially the @xmath23 channel at higher dm masses somewhat decreases due to the lower number of photons per annihilation ( and thus lower statistics ) available . including the cross - correlation spectra information does not seem to improve the sensitivity .
this probably has to be ascribed to the fact that the @xmath25 for the auto - correlation case already includes some cross - correlation information in the off - diagonal terms of the covariance matrix .
the level of anisotropies is also very low at low energy for this case ( see fig.[fig : anieni ] ) and this somewhat reduces the importance of the information contained in the cross - correlation .
the situation is different for the galactic dm scenario as discussed below .
we also checked the case in which only the highest energy bands are employed for the analysis .
we would expect an improvement in sensitivity given that very little dm signal is expected in the 0.4 - 0.8 gev band and thus scanning in this energy range introduces a statistical penalty factor .
we find that the sensitivity indeed improves , but only marginally . on the other hand ,
the lowest energy band is useful because being mostly clean of dm it represents a good calibration for the background anisotropies . in fig .
[ fig : limitsb ] we show the case in which the dm anisotropy pattern does not vary with energy .
this case is supposed to mimic the dm emission from unresolved substructures in the milky way .
we thus simulate a single anisotropic map which we use for all energies .
the angular spectrum for dm employed to simulate the map is normalized as in the case of substructures with a minimum mass of 10 @xmath26 as derived in @xcite and it roughly corresponds to the extragalactic anisotropies for an energy of about 300 gev ( see section 2 ) . for a direct comparison with the cosmological case we use for the galactic signal the same normalization employed for the extragalactic case , which is justified by the fact that more physically motivated models for the dm emission in the mw have a large degree of freedom @xcite and
can easily be tuned to match the extragalactic case .
the inclusion of the information from the cross - correlation spectra in this case , contrary to the cosmological one , gives a significant improvement , especially for higher dm masses .
the reason for this difference is likely to be found in the very different set of anisotropy maps employed for the galactic substructures case . in particular , the very high degree of anisotropy
present also at low energy for this case enhances the importance of a direct measurement of the cross - correlation between low and high energy bands .
in contrast , the very low dm anisotropy at low energy for the cosmological case ( lower than the astrophysical egb anisotropy ) makes this information almost irrelevant .
we also find in this case that excluding the lowest energy band does not improve the sensitivity significantly .
the sensitivity increases substantially with respect to the cosmological case , especially for high dm masses .
the slope of the sensitivity curves is indeed slightly less steep . assuming higher levels of anisotropy for the galactic case , which are possible , @xcite a correspondingly higher sensitivity could be achieved .
however , the scaling is not linear .
it is worth noting for example that although at about 10 gev the galactic anisotropies are about 2 orders of magnitudes larger than the extragalactic ones , the sensitivity is only a factor 2 - 3 better
. the scaling however is better for @xmath27 gev where the difference in sensitivity is roughly a factor of 10 .
this non linear scaling is due to the limited limited amount of detected photons ( as it is the case at the edges of the sensitivity curve ) . here
, it is very difficult to distinguish between the two ( galactic and extra - galactic ) levels of anisotropies even if the latter is a factor 100 larger , since the statistical error bars are also very large .
the much larger galactic anisotropy thus only modestly decrease the minimum amount of events required to detect the signal . finally in fig .
[ fig : limitcp ] we show the impact on the dm sensitivity of changing the underlying astrophysical egb model to the poisson dominated one . for illustration ,
only the extragalactic @xmath24 case with nfw profile and substructures is shown .
the interesting results ( perhaps somewhat counter intuitive ) is that even with this increased level of background in the anisotropy spectrum , still the sensitivity to the dm component remains roughly unchanged .
the reason is that a poisson noise at the level of 5.0e-5 is anyway clearly detected with a statistics of 5 years , so the dm component can still be easily detected as a modulation ( as usual , in multipole and in energy ) of this dominant poisson noise anisotropy spectrum .
the underlying assumption here , however , is that this new background is fully under control and well characterized . the extent to which this will be true is however not clear and a better understanding may require dedicated simulations and analyses which are beyond the scope of this paper . , extragalactic anisotropies and nfw halo profile with substructures , for the two cases of egb background anisotropies from the correlation term only ( same case of fig.[fig : limitsb ] , solid contours ) and from a pure poisson term only ( dashed contours).[fig : limitcp ] ]
in the present analysis we have calculated the sensitivity to dark matter annihilation exploiting the characteristic anisotropy properties that dm would imprint in the extra - galactic gamma background .
we performed a realistic simulation of the fermi - lat including the effect of the point spread function , effective area and of the anisotropic sky exposure .
more importantly , we investigated the impact of galactic foregrounds on the dm anisotropy sensitivity using a galprop model to produce the sky maps of the galactic gamma emission .
the contribution of galactic foregrounds is generally non negligible even above galactic latitudes of about @xmath28 .
foreground cleaning will thus be an important ( and likely challenging ) issue in the analysis of the real data .
for the present work we do not attempt foreground cleaning ( except for masking the galactic plane and point sources ) and we only require that the anisotropy from dm has to exceed the level of foregrounds anisotropy to be detectable . further , we only analyze a high galactic latitude region where the model dependence of the foreground is expected to be moderate . in this respect
our analysis can be regarded as a conservative one : in reality , if foreground cleaning can be effectively implemented , a sensitivity to a lower level of anisotropy can be achievable .
also , a larger fraction of the sky can be employed than the small conservative one used in the present analysis .
we leave a more careful investigation of the issue for future work . despite the difficulties involved in the foreground analysis
, we find that the prospects for detecting dm anisotropy signatures are promising .
for some wimp masses , sensitivity to @xmath29@xmath3s@xmath4(i.e . the thermal relic value ) can be achieved using the anisotropy information alone .
interestingly , these constraints are comparable to what is achievable using the energy spectrum alone @xcite .
the anisotropy information is thus complementary to the energy spectrum and provides a cross - check and an increase in the overall sensitivity to the dm signal .
we also investigated the recently proposed scenario of a population of dm substructures in the milky way halo as predicted by n - body simulations .
these subhaloes give rise to an annihilation signal which is almost uniform in the sky and which can be comparable in intensity to the observed isotropic gamma background .
for this case , the anisotropy signal is predicted to be even higher than for the extragalactic case due to the huge clumpiness of the dm distribution that correspondingly gives an enhanced sensitivity .
we find , indeed , that when increasing the anisotropy the sensitivity to the dm signal increases as well with an improvement of a factor of @xmath610 for @xmath27 gev . in summary , simulations indicate that , with the sensitivity of the fermi - lat , detailed anisotropies studies of the gamma - ray sky can be performed .
the next obvious step will be the analysis of the actual data .
anisotropy analyses will soon provide a complementary source of information which will give further insights into the dm problem and the properties of high - energy astrophysical sources .
the authors would like to thank toby burnett , jennifer siegal - gaskins and eiichiro komatsu for their numerous comments and suggestions for improvement , and chris savage for a carefully reading of the manuscript .
we also wish to thank gianfranco bertone for organizing the workshop `` gamma in zurich '' which stimulated many interesting discussions on gamma - ray anisotropies .
the _ fermi _ lat collaboration acknowledges generous ongoing support from a number of agencies and institutes that have supported both the development and the operation of the lat as well as scientific data analysis .
these include the national aeronautics and space administration and the department of energy in the united states , the commissariat lenergie atomique and the centre national de la recherche scientifique / institut national de physique nuclaire et de physique des particules in france , the agenzia spaziale italiana and the istituto nazionale di fisica nucleare in italy , the ministry of education , culture , sports , science and technology ( mext ) , high energy accelerator research organization ( kek ) and japan aerospace exploration agency ( jaxa ) in japan , and the k. a. wallenberg foundation , the swedish research council and the swedish national space board in sweden . additional support for science analysis during the operations phase is gratefully acknowledged from the istituto nazionale di astrofisica in italy and the centre national dtudes spatiales in france . 00 a. a. abdo _ et al . _
[ fermi lat collaboration ] , astrophys . j. suppl .
* 183 * ( 2009a ) 46 [ arxiv:0902.1340 [ astro-ph.he ] ] .
a. a. abdo _ et al .
_ [ fermi lat collaboration ] , arxiv:1002.3603 ( 2009b ) .
a. a. abdo _ et al .
_ [ fermi lat collaboration ] , phys .
lett . * 103 * ( 2009c ) 251101 [ arxiv:0912.0973 [ astro-ph.he ] ] .
a. a. abdo _ et al .
_ [ fermi lat collaboration ] , arxiv:1002.2280 ( 2009d ) .
a. a. abdo _ et al .
_ [ fermi - lat collaboration ] , arxiv:1002.4415 ( 2010a ) .
a. a. abdo _ et al .
_ [ fermi - lat collaboration ] , arxiv:1003.0895 ( 2010b ) .
m. ackermann , g. johannesson , s. digel , i. v. moskalenko , t. porter , o. reimer and a. strong [ fermi lat collaboration ] , aip conf .
* 1085 * ( 2009 ) 763 . s. ando , arxiv:0903.4685 [ astro-ph.co ] .
s. ando and e. komatsu , phys . rev .
d * 73 * ( 2006 ) 023521 [ arxiv : astro - ph/0512217 ] .
s. ando and v. pavlidou , mon . not .
* 400 * ( 2009 ) 2122 [ arxiv:0908.3890 [ astro-ph.he ] ] . s. ando , e. komatsu , t. narumoto and t. totani , mon . not .
roy . astron .
* 376 * ( 2007 ) 1635 [ arxiv : astro - ph/0610155 ] .
s. ando , e. komatsu , t. narumoto and t. totani , phys .
d * 75 * ( 2007 ) 063519 [ arxiv : astro - ph/0612467 ] .
w. b. atwood _ et al . _
[ lat collaboration ] , astrophys .
j. * 697 * ( 2009 ) 1071 [ arxiv:0902.1089 [ astro-ph.im ] ] .
j. ballet [ fermi collaboration ] , talk given at the washington fermi symposium 2009 e. a. baltz _ et al .
_ , jcap * 0807 * ( 2008 ) 013 [ arxiv:0806.2911 [ astro - ph ] ] . l. bergstrom , rept . prog
. phys .
* 63 * ( 2000 ) 793 [ arxiv : hep - ph/0002126 ] .
l. bergstrom , j. edsjo and p. ullio , phys . rev
. lett . *
87 * , 251301 ( 2001 ) [ astro - ph/0105048 ] .
g. bertone , d. hooper and j. silk , phys .
* 405 * ( 2005 ) 279 [ arxiv : hep - ph/0404175 ] .
a. cuoco , pos * idm2008 * ( 2008 ) 114 .
a. cuoco , s. hannestad , t. haugbolle , g. miele , p. d. serpico and h. tu , jcap * 0704 * ( 2007 ) 013 [ arxiv : astro - ph/0612559 ] .
a. cuoco , j. brandbyge , s. hannestad , t. haugboelle and g. miele , phys .
d * 77 * ( 2008 ) 123518 [ arxiv:0710.4136 [ astro - ph ] ] .
j. diemand , m. kuhlen and p. madau , halo , astrophys . j. * 657 * ( 2007 ) 262 [ arxiv : astro - ph/0611370 ] .
s. dodelson , a. v. belikov , d. hooper and p. serpico , arxiv:0903.2829 [ astro-ph.co ] . c. a. faucher - giguere and a. loeb , arxiv:0904.3102 [ astro-ph.he ] .
m. fornasa , l. pieri , g. bertone and e. branchini , arxiv:0901.2921 [ astro - ph ] . s. gabici and p. blasi , astropart .
* 19 * , 679 ( 2003 ) [ astro - ph/0211573 ] .
gondolo p. , edsjo j. , ullio p. , bergstrom l. , schelke m. and baltz e.a . ,
jcap * 0407 * ( 2004 ) 008 [ astro - ph/0406204 ] .
k. m. gorski , e. hivon , a. j. banday , b. d. wandelt , f. k. hansen , m. reinecke and m. bartelman , astrophys .
j. * 622 * ( 2005 ) 759 [ arxiv : astro - ph/0409513 ] .
hartman , r. c. , et al.1999 , vizier online data catalog , 212 , 30079 b. s. hensley , j. m. siegal - gaskins and v. pavlidou , arxiv:0912.1854 [ astro-ph.co ] .
e. hivon , k. m. gorski , c. b. netterfield , b. p. crill , s. prunet and f. hansen , astrophys .
j. * 567 * ( 2002 ) 2 , arxiv : astro - ph/0105302 .
d. hooper and p. d. serpico , jcap * 0706 * ( 2007 ) 013 [ arxiv : astro - ph/0702328 ] .
hunter s.d .
_ et al . _ , astrophys .
j. * 481 * ( 1997 ) 205 .
a. ibarra , d. tran and c. weniger , phys .
d * 81 * ( 2010 ) 023529 [ arxiv:0909.3514 [ hep - ph ] ] .
g. jungman , m. kamionkowski and k. griest , phys .
* 267 * ( 1996 ) 195 [ arxiv : hep - ph/9506380 ] .
o. e. kalashev , d. v. semikoz and g. sigl , phys .
d * 79 * ( 2009 ) 063005 [ arxiv:0704.2463 [ astro - ph ] ] .
u. keshet , e. waxman , a. loeb , v. springel and l. hernquist , astrophys .
j. * 585 * ( 2003 ) 128 [ arxiv : astro - ph/0202318 ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , arxiv:0803.0547 [ astro - ph ] .
s. k. lee , s. ando and m. kamionkowski , arxiv:0810.1284 [ astro - ph ] .
a. lewis , phys .
d * 78 * ( 2008 ) 023002 [ arxiv:0804.3865 [ astro - ph ] ] .
d. n. limber , astrophys .
j. * 117 * ( 1953 ) j. f. navarro , c. s. frenk and s. d. white , astrophys .
j. * 462 * ( 1996 ) 563 f. miniati , s. m. koushiappas and t. di matteo , astrophys .
j. * 667 * ( 2007 ) l1 [ arxiv : astro - ph/0702083 ] .
moskalenko i.w . and strong a.w .
, astrophys .
j. * 493 * ( 1998 ) 694 v. pavlidou and b. d. fields , astrophys .
j. * 575 * , l5 ( 2002 ) [ astro - ph/0207253 ] .
l. pieri , g. bertone and e. branchini , mon . not .
* 384 * ( 2008 ) 1627 [ arxiv:0706.2101 [ astro - ph ] ] .
l. pieri , j. lavalle , g. bertone and e. branchini , arxiv:0908.0195 [ astro-ph.he ] . t. a. porter [ fermi collaboration ] , arxiv:0907.0294 [ astro-ph.he ] .
t. a. porter [ fermi collaboration ] , talk given at tev particle astrophysics 2009 j. m. siegal - gaskins , jcap * 0810 * ( 2008 ) 040 [ arxiv:0807.1328 [ astro - ph ] ] . j. m. siegal - gaskins and v. pavlidou , arxiv:0901.3776 [ astro-ph.he ] . j. m. siegal - gaskins , for the fermi - lat collaboration and e. komatsu , [ arxiv:1012.1206 [ astro-ph.he ] ] .
sjostrand t. , mrenna s. and skands p. , comput .
commun .
* 178 * ( 2008 ) 852 [ arxiv:0710.3820 ] .
v. springel _ et al .
_ , arxiv:0809.0894 [ astro - ph ]
. v. springel _ et al .
_ , arxiv:0809.0898 [ astro - ph ] .
sreekumar p. _
[ egret collaboration ] , astrophys .
j. * 494 * ( 1998 ) 523 [ astro - ph/9709257 ] .
f. w. stecker and m. h. salamon , astrophys .
j. * 464 * , 600 ( 1996 ) [ astro - ph/9601120 ] .
f. w. stecker , m. a. malkan and s. t. scully , astrophys . j. * 648 * , 774 ( 2006 ) [ astro - ph/0510449 ] .
see also f. w. stecker , m. a. malkan and s. t. scully , astrophys .
j. * 658 * ( 2007 ) 1392 [ arxiv : astro - ph/0612048 ] . f. w. stecker , s. d. hunter and d. a. kniffen , astropart .
* 29 * ( 2008 ) 25 [ arxiv:0705.4311 [ astro - ph ] ] . strong a.w . and moskalenko i.w . ,
astrophys .
j. * 509 * ( 1998 ) 212 strong a.w . , moskalenko i.w . and reimer o. , astrophys .
j. * 613 * ( 2004a ) 956 [ astro - ph/0405441 ] .
strong a.w . , moskalenko i.w . and reimer o. , astrophys . j. * 613 * ( 2004b ) 962 [ astro - ph/0406254 ] .
m. taoso , s. ando , g. bertone and s. profumo , phys .
d * 79 * ( 2009 ) 043521 [ arxiv:0811.4493 [ astro - ph ] ] .
j. e. taylor and j. silk , `` the clumpiness of cold dark matter : implications for the annihilation signal , '' mon . not .
* 339 * ( 2003 ) 505 [ arxiv : astro - ph/0207299 ] .
m. tegmark _ et al .
_ , phys .
d * 74 * , 123507 ( 2006 ) [ arxiv : astro - ph/0608632 ] .
t. totani , astropart .
* 11 * , 451 ( 1999 ) [ astro - ph/9810207 ] .
a. a. abdo _ et al .
_ [ fermi lat collaboration ] , arxiv:1003.0895v1 [ astro-ph.co ] p. ullio , l. bergstrom , j. edsjo and c. g. lacey , phys . rev .
d * 66 * , 123502 ( 2002 ) [ arxiv : astro - ph/0207125 ] .
j. zavala , v. springel and m. boylan - kolchin , arxiv:0908.2428 [ astro-ph.co ] . l. zhang and g. sigl , jcap * 0809 * ( 2008 ) 027 [ arxiv:0807.3429 [ astro - ph ] ] .
we summarize in this appendix the conventions and notation used for the anisotropy analysis .
the anisotropy of a signal on the sky can be quantified in terms of the power spectrum i.e. the spherical harmonic transform on the sphere .
more specifically , a map on the sphere @xmath30 can be decomposed in spherical harmonics as @xmath31 and the angular power spectrum can then be defined as @xmath32 where @xmath33 indicates the statistical ensemble average .
the quantity @xmath34 is an unbiased estimator of the true power spectrum @xmath35 , i.e. @xmath36 . in the case of two maps , besides the two autocorrelation power spectra @xmath37 and @xmath38 , further information is contained in the _ cross - correlation power spectrum _ which can be estimated as @xmath39 where the equivalence of the two expressions comes from the equality : @xmath40 , @xmath41 .
the three spectra are of course not independent but are part of a covariance matrix : @xmath42 which has to be taken into account when evaluating the errors for the forecast .
this is described in more details in the next section . in practice , to calculate the power spectra , the harmonic transform @xmath43 is performed with healpix @xcite after binning the events into maps of the healpix format itself . due to the finite number @xmath44 of collected events beside the intrinsic anisotropies
the data also contain noise ( shot noise or white noise ) which appears in the power spectrum as a constant whose value is @xmath45 .
the noise sets the limit above which the intrinsic anisotropies can be detected .
it is standard to remove the shot noise from the spectrum showing only the intrinsic power spectrum .
it is also customary to show not the power spectrum @xmath35 but the quantity @xmath46 .
in the general case in which we include as observables the cross - correlation power spectra the covariance matrix can be calculated by propagation of the errors from the expressions [ pspectrum ] and [ cspectrum ] and then taking the ensemble average . the variance @xmath47 of the cross - correlation spectrum @xmath48 ( @xmath49 ) can be written as @xmath50 we then have the case of covariance @xmath51 between an auto - correlation spectrum @xmath52 and cross - correlation spectrum @xmath53 . for this case
we have @xmath54 for the case of covariance between two cross - correlation spectra we have to distinguish two cases . in the case of covariance @xmath55 between @xmath48 and @xmath56 where there is an energy band ( an index ) in common we have : @xmath57 if instead there is no energy band ( no index ) in common @xmath58 finally we have @xmath59 expression ( [ autocorrerror ] ) and ( [ crosscorrerror ] ) can also be rewritten in the more familiar form @xmath60 for the auto - correlation spectra and by @xmath61 for the cross - correlation spectra , where again @xmath62 and @xmath63 is the number of binned multipole in the band @xmath64 , and @xmath65 is the density of the counts per steradian in the map @xmath66 . | we perform a detailed study of the sensitivity to the anisotropies related to dark matter ( dm ) annihilation in the isotropic gamma - ray background ( igrb ) as measured by the fermi large area telescope ( fermi - lat ) . for the first time
, we take into account the effects of the galactic foregrounds and use a realistic representation of the fermi - lat . we implement an analysis pipeline which simulates fermi - lat data sets starting from model maps of the galactic foregrounds , the fermi resolved point sources , the extra - galactic diffuse emission and the signal from dm annihilation .
the effects of the detector are taken into account by convolving the model maps with the fermi - lat instrumental response .
we then use the angular power spectrum to characterize the anisotropy properties of the simulated data and to study the sensitivity to dm .
we consider dm anisotropies of extra - galactic origin and of galactic origin ( which can be generated through annihilation in the milky way sub - structures ) as opposed to a background of anisotropies generated by sources of astrophysical origin , blazars for example .
we find that with statistics from 5 years of observation fermi is sensitive to a dm contribution at the level of 1%-10% of the measured igrb depending on the dm mass @xmath0 and annihilation mode . in terms of the thermally averaged cross section @xmath1 , this corresponds to @xmath2@xmath3s@xmath4 , i.e. slightly above the typical expectations for a thermal relic , for low values of the dm mass @xmath5 gev .
the anisotropy method for dm searches has a sensitivity comparable to the usual methods based only on the energy spectrum and thus constitutes an independent and complementary piece of information in the dm puzzle . |
in @xcite-@xcite , korchagin stated three conjectures about ninth degree @xmath0-curves , two of them are completely proved ( or partly disproved ) , see@xcite , @xcite , @xcite .
the third one is still open .
let @xmath11 be a ninth degree @xmath0-curve with three nests , its real scheme is : @xmath12 , with @xmath13 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * conjecture * at least two of the @xmath3 are odd . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it is already known that at least one of the @xmath3 is odd , see @xcite .
we say that the curve has a jump in a non - empty oval @xmath4 if there exist four ovals @xmath5 , @xmath6 , @xmath7 , @xmath8 , with @xmath5 interior to some other non - empty oval @xmath9 , @xmath8 exterior , @xmath10 interior to @xmath4 , such that @xmath6 and @xmath7 are separated inside of @xmath4 by any line passing through @xmath5 and @xmath8 , see figure [ saut ]
. the main result of the present paper is the following : jump in @xmath4 ] let @xmath11 be an @xmath0-curve of degree @xmath15 with three nests and no jump .
then at least two of the @xmath3 are odd . in this section , we determine as much information as possible about the complex orientations and rigid isotopy types realizable by the ninth degree @xmath0-curves with three nests ( proposition 1 and tables [ orev1]-[orev16 ] ) . to this purpose , we apply lemmas and propositions stated in @xcite , in particular lemmas 19 and 20 , which are handy reformulations , for the specific case of the ninth degree @xmath0-curves with three nests , of orevkov s complex orientation formulas from @xcite .
let us first describe these curves , using the results from @xcite .
for the reader s convenience , we repeat here some definitions and notations introduced in that paper . for each empty oval @xmath16 of @xmath11
, we choose a point lying in it , denoted also by @xmath16 , it will be clear from the context whether we speak of the point or of the oval .
let @xmath17 be the three non - empty ovals .
for any pair of ovals @xmath18 , interior to @xmath19 and @xmath20 respectively , one denotes by @xmath21 $ ] and calls _ affine segment _ the segment of the line @xmath22 that does nt cut @xmath23 , see figure [ conj1 ] .
let @xmath21'$ ] be the other segment .
any pencil @xmath24 , based in an oval @xmath5 interior to @xmath20 or @xmath25 and sweeping out @xmath19 from the first oval interior to @xmath19 to the last one , meets the same sequence ( fiedler chain ) of empty ovals , each of them being either interior to @xmath19 or exterior .
the occurence of exterior ovals means that @xmath19 has some jump .
if @xmath19 has no jump , the sequence is called _ chain of @xmath19_. if @xmath3 is odd , the chain of @xmath19 is _ positive _ or _ negative _ , depending on whether it contributes by @xmath26 or @xmath27 to @xmath28 . the word _ chain _
, when used without further precision , has a very restrictive meaning .
a pencil of lines ( conics ) sweeping out a curve of degree @xmath29 is _ maximal _ if it has alternatively @xmath29 ( @xmath30 ) and @xmath31 ( @xmath32 ) real intersection points with the curve .
a _ chain _ is a fiedler chain of ovals arising from a maximal pencil , in such a way that the consecutive ovals are connected by vanishing cycles . in the pseudoholomorphic
setting , the curve may be perturbed so as to replace these cycles by real crossings .
it was proved in @xcite that at most one non - empty oval may have a jump , by convention it is @xmath33 .
moreover the sequence of @xmath33 ( or jump sequence ) splits into exactly three _ jump subsequences _
( int , ext , int ) , we say that there is only one jump .
if these subsequences have parities ( odd , odd , odd ) , we speak of an _
odd jump_. the two inner subsequences are the _ chains of @xmath33_. let @xmath19 be any one of the non - empty ovals , the lines passing through pairs of ovals interior to @xmath19 cut either all of the affine segments @xmath34 $ ] or all of the non - affine segments @xmath34'$ ] , where @xmath35 ( @xmath36 ) ranges over all of the ovals interior to @xmath20 ( @xmath25 ) . in the first case
, we say that @xmath19 is _ separating _ , otherwise @xmath19 is _ non - separating_. if @xmath33 has a jump , @xmath33 is non - separating .
let @xmath37 be three ovals distributed in the three successive jump subsequences .
up to a swap of @xmath6 and @xmath7 , the ovals @xmath38 ,
@xmath39 , @xmath7 , @xmath8 , @xmath6 lie in convex position in @xmath40 for any choice of @xmath41 , see @xcite and @xcite .
let @xmath33 have a jump and @xmath8 be an exterior oval of the jump sequence .
then @xmath33 cuts all of the segments @xmath42 $ ] or all of the segments @xmath42'$ ] , when @xmath5 ranges over all ovals interior to @xmath43 and @xmath44 . in the first case
, we say that @xmath8 is _ front _ , in the second case , @xmath8 is _ back _ , see figure [ conj2 ] .
if each exterior oval @xmath8 in the exterior jump subsequence is front ( back ) , we say that @xmath33 is _ crossing _ ( _ non - crossing _ ) , the exterior subsequence is a chain , the three jump subsequences will be called _
jump chains_. note that there may exist both front and back ovals , in this case , the exterior subsequence splits into two chains formed respectively by front and back ovals , see proposition 3 in @xcite .
@xmath0-curve of degree @xmath15 with three nests ] jump in @xmath33 , @xmath8 front , @xmath8 back ] let @xmath45 be three chosen ovals , distributed inside of the three non - empty ovals @xmath17 .
the six ovals @xmath46 will be called the _ principal ovals _ of @xmath11 .
the three lines @xmath47 , @xmath48 , @xmath49 , and @xmath23 divide the plane in four quadrangles and three triangles .
the contributions of the ovals in @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 to @xmath28 are denoted by @xmath57 , @xmath58 , see figure [ conj1 ] .
let @xmath59 be extremity of the chain of @xmath19 if @xmath19 has no jump ; if @xmath33 has a jump , @xmath60 is extremity of the jump sequence .
the interior of @xmath19 is divided in four parts by the lines @xmath22 , @xmath61 ( two in quadrangles , two in triangles ) , if @xmath19 has no jump , the chain of @xmath19 lies entirely in one of these four parts , if @xmath33 has a jump , the two chains of @xmath33 lie in the same part , contained in a quadrangles @xmath51 or @xmath52 ( by convention , we choose @xmath60 such that the quadrangle is @xmath51 ) . if @xmath19 is separating , the chain of @xmath19 lies in @xmath62 or in @xmath50 .
let @xmath19 be separating and even .
we say that @xmath19 is _ up _ if , when @xmath59 is chosen negative ( positive ) , the chain is in @xmath62 ( @xmath50 ) , otherwise @xmath19 is _ down _ , see figure [ updown ] .
@xmath43 up , @xmath43 down ] let us assign to each nest a code that gives the sign of the outer oval @xmath19 , and the number @xmath63 .
note that the values @xmath64 correspond to the nests with odd jump .
the codes are defined as explained in the example hereafter : @xmath65 stands for an even nest ( without odd jump ) with positive outer oval , @xmath66 stands for an odd nest with positive outer oval and negative inner chain , @xmath67 stands for a nest with odd jump , such that the outer oval @xmath33 is positive and the two chains of @xmath33 are negative .
the code @xmath68 stands for a nest that is either @xmath66 or @xmath69 .
the _ short complex scheme _ of @xmath11 is the set of three codes realized by the three nests .
the _ complex type _ of @xmath11 is the short complex scheme , enhanced with the information for the odd nests , whether they are separating or non - separating ; for the even nests , whether they are non - separating , up or down : @xmath70 and @xmath71 correspond respectively to a non - separating and a separating odd nest with positive outer oval @xmath19 and negative interior chain ; @xmath72 , @xmath73 , @xmath74 correspond respectively to a non - separating , up , and down even nest with positive outer oval @xmath19 . in the next subsection
, we will prove : the admissible complex types for @xmath11 with at least one even nest are : listed in table [ orev6 ] ( even , even , odd without jump ) ; listed in table [ orev10 ] ( even , even , odd with jump ) ; listed in table [ orev16 ] ( even , odd , odd without jump ) ; @xmath75 , @xmath75 , @xmath67 and @xmath75 , @xmath75 , @xmath76 ( even , odd , odd with jump ) .
the admissible short complex schemes for @xmath11 with three odd nests are listed in table [ orev18 ] .
the admissible complex types contradicting the conjecture are listed in table [ orev6 ] ( case without jump ) , and table [ orev10 ] ( case with jump ) .
theorem 1 is proved in section 2 .
let @xmath43 be even and non - separating .
we say that @xmath43 is _ left _
( _ right _ ) if , when @xmath38 was chosen with orientation opposite to that of @xmath43 , the chain of @xmath43 lies in @xmath52 ( @xmath77 . in section 3 , we will prove : let @xmath11 be a ninth degree @xmath0-curve even , even , odd with jump .
then @xmath11 is as shown in figure [ upcross ] or figure [ rightnocross ] .
the curve @xmath11 has complex type @xmath78 , with @xmath33 crossing and @xmath79 or @xmath80 ( figure [ upcross ] ) ; or @xmath11 has complex type @xmath81 , @xmath76 with @xmath33 non - crossing , @xmath43 right and @xmath79 ( figure [ rightnocross ] ) .
let us add a word of explanation to the figures .
the principal ovals , the base lines and @xmath23 divide the plane in zones .
we have indicated the contributions of the non - principal ovals to @xmath28 in each zone ( a single positive oval means @xmath26 ) . in a zone left blank ,
the contribution is @xmath82 , the zones that we know to be empty are hatched .
@xmath11 with complex type @xmath78 , @xmath79 or @xmath80 ] @xmath11 with complex type @xmath81 , @xmath76 ] all along the proof , we refer to lemmas and propositions from @xcite . for the reader
s convenience , we give here a quick summary of the main results . if @xmath33 is crossing , @xmath56 is empty ; if @xmath33 is non - crossing , @xmath83 is emtpy ; if the exterior jump sequence has both front and back ovals , @xmath84 is empty ( lemma 9 ) .
let @xmath85 .
one has @xmath86 ( lemma 10 ) .
let @xmath11 have a jump , then @xmath87 or @xmath88 .
if @xmath89 , then @xmath11 is even , odd , odd with odd jump . if @xmath87 , then @xmath33 is either crossing and positive , or non - crossing and negative ( lemma 18 ) .
note that if @xmath87 , the curve @xmath11 may be even , even , odd with odd jump , or odd , odd , odd with jump sequence either ( odd , odd , even ) or ( odd , even , even ) . in @xcite
, we defined parameters @xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 , @xmath95 associated to the code of each nest @xmath96 ( @xmath95 is the only one that depends on the data @xmath97 ) .
we defined also parameters @xmath98 depending on the three codes .
lemmas 19 and 20 are handy reformulations of orevkov s formulas : if @xmath62 contains some exterior ovals , then @xmath99 ( lemma 19 ) . if @xmath19 is separating , then @xmath100 ( lemma 20 ) .
if @xmath84 is empty , then the short complex scheme of @xmath11 is @xmath101 or @xmath102 ( lemma 21 ) .
finally , let us mention two propositions from @xcite : if @xmath62 contains some exterior ovals , then @xmath103 ; if @xmath104 , then @xmath105 and @xmath106 ( proposition 1 ) .
if @xmath50 contains only exterior ovals , then @xmath107 ( proposition 2 ) .
consider first the case even , even , odd .
the twelve admissible complex schemes without jump are displayed in table [ orev1 ] , along with : the numbers @xmath108 , the indices @xmath109 of the zones @xmath110 that may contain some exterior ovals ( row @xmath111 ) , and @xmath112 .
note that by lemma 21 , the first five complex schemes of table [ orev1 ] can not be realized with all of the @xmath19 non - separating .
adding to this the informations in tables [ orev2]-[orev4 ] allows to get the list of admissible complex types in table [ orev5 ] . in the tables ,
each row with @xmath113 nests @xmath68 corresponds actually to @xmath114 complex schemes or types .
choose for any separating even nest @xmath19 the base oval @xmath59 in such a way that @xmath115 form a positive pair , and @xmath59 is extremal in the chain of inner ovals .
the complex types in table [ orev5 ] are such that @xmath116 if @xmath33 is @xmath68 , and @xmath117 if @xmath33 is @xmath118 or @xmath119 .
this , combined with the data of @xmath111 , allows to find out the admissible values for the triangular parameters @xmath120 .
the six types with @xmath121 ( first four rows of table [ orev5 ] ) are such that the triangle @xmath54 contains only exterior ovals , as @xmath43 is either non - separating or @xmath74 . for these types ,
@xmath122 , this contradicts proposition 1 ( from @xcite ) .
the type with @xmath123 is such that the triangle @xmath56 contains only exterior ovals .
but @xmath124 , this contradicts also proposition 1 .
the types with @xmath125 are such that the triangle @xmath50 contains only exterior ovals .
but @xmath126 , this contradicts proposition 2 . the types @xmath127 and @xmath128 verify @xmath129 .
for @xmath33 non - separating , both @xmath50 and @xmath56 contain only exterior ovals , so that one may apply propositions 1 and 2 .
one must have @xmath130 , @xmath107 , this is a contradiction .
when @xmath33 is separating , one may choose the oval @xmath60 so that @xmath56 contains only exterior ovals , to apply proposition 1 .
then , one chooses the other extremal oval @xmath131 , so that @xmath132 contains only exterior ovals , to apply proposition 2.the values of the @xmath57 are the same for both choices , so one gets the same contradiction as for a non - separating @xmath33 . in the sequel ,
keep in mind that we only need to take care of the even nests to achieve the hypothesis needed for proposition 1 or 2 .
the types left are gathered in table [ orev6 ] , note that each of the three rows corresponds actually to four types , depending on whether the nest @xmath33 is @xmath66 or @xmath69 , separating or non - separating .
we get the admissible values of @xmath133 and @xmath134 applying propositions 1 and 2 .
as @xmath50 contains only exterior ovals , @xmath107 . if the even ovals are both non - separating , @xmath135 , hence @xmath136 or @xmath137 ; if @xmath43 is separating , @xmath138 , hence @xmath139 .
the @xmath0-curves with odd jump are either even , even , odd with @xmath87 or even , odd , odd with @xmath89 ( lemma 18 ) .
the list of admissible short complex schemes with odd jump is shown in table [ orev7 ] .
consider the case even , even , odd .
if @xmath33 is non - crossing , the ovals @xmath43 and @xmath44 are non - separating ( by lemma 9 ) .
if moreover @xmath43 is negative , then there are no exterior triangular ovals .
this is a contradiction as one should have @xmath140 . adding this to the informations in tables [ orev8]-[orev9 ] allows to get the list of admissible complex types even , even , odd with jump displayed in table [ orev10 ] , and to see that the ovals @xmath43
, @xmath44 of the even , odd , odd curves with jump must be non - separating . in table [ orev10 ] , we have also indicated the parameters @xmath141 , @xmath142 , @xmath143 , they are obtained from the identities @xmath144 , @xmath145 , @xmath146 .
( these identities follow from fiedler s theorem applied with the pencils of lines @xmath147 sweeping out @xmath148 , see first part of lemma 10 ) .
table [ orev11 ] shows the admissible short complex schemes even , odd , odd without jump . as there is at most one separating even nest ,
the contribution of the interior ovals to @xmath112 is @xmath149 or @xmath27 .
if @xmath43 is separating , choose the base oval @xmath38 so as to realize each time the required hypothesis to apply propositions 1 or 2 .
we get thus : if @xmath150 , then @xmath151 ; if @xmath125 , then @xmath107 hence @xmath152 ; if @xmath153 , then @xmath154 hence @xmath152 , or @xmath105 and @xmath106 . using the informations from tables [ orev12]-[orev15 ]
, we get the list of admissible complex types even , odd , odd without jump in table [ orev16 ] .
table [ orev17 ] shows the admissible short complex schemes odd , odd , odd . by propositions 1 and 2 : @xmath107 and @xmath154 .
the only admissible schemes left are shown in table [ orev18 ] . for @xmath155 ,
one has @xmath156 for both choices of @xmath33 , there is no contradiction ( see lemma 20 ) .
for @xmath157 , one has @xmath158 , contradiction .
any one of the nests @xmath68 may be separating , whereas the nest @xmath119 must be non - separating .
the curve may have a jump only if all three nests are @xmath68 ( lemma 18 ) .
.[orev1 ] admissible short complex schemes even , even , odd , without jump [ cols="^,^,^,^,^,^,^,^,^ " , ]
assume there exists @xmath11 contradicting the conjecture , @xmath11 realizes one of the three complex types in table [ orev6 ] . we will find a contradiction using bezout s theorem with auxiliary rational quartics .
the first step consists in studying a pencil of conics based at @xmath45 and a fourth point @xmath8 , chosen in a quadrangular oval .
it turns out that this pencil has a `` double jump '' : it sweeps out successively five ovals @xmath159 , such that @xmath160 are triangular , @xmath161 are quadrangular ( see lemma 3 ahead ) . for @xmath162 ,
let @xmath59 be the positive extreme oval in the chain of @xmath19 , and let @xmath60 be any one of the extreme ovals in the chain of @xmath33 , the pair of ovals @xmath60 ,
@xmath33 have opposite orientations . in figure [ types ] , we show all admissible distributions of ovals for @xmath11 .
the base lines , @xmath23 and the ovals @xmath163 divide the plane in zones .
the ovals that are not represented are distributed in such a way that the difference _ number of positive minus number of negative ones _
is @xmath82 in each zone .
let @xmath43 be non - separating ( @xmath11 has the first or the second type ) , we say that @xmath43 is _ left _ if the chain of @xmath43 is in @xmath52 , _ right _ if the chain of @xmath43 is in @xmath53 .
the oval @xmath44 is non - separating for all three types , we say that @xmath44 is _ left _ if the chain of @xmath44 is in @xmath53 , _ right _ if the chain of @xmath44 is in @xmath51 . perform the cremona transformation @xmath164 based at the points @xmath45 .
the respective images of the lines @xmath47 , @xmath49 , @xmath165 will be denoted by @xmath60,@xmath38 , @xmath39 ( note that after @xmath164 , the new base lines @xmath47 , @xmath49 , @xmath165 divide the plane in four triangles @xmath166 , @xmath167 , @xmath168 ) . for the points lying inside of non - principal ovals
, we use the same notation as before @xmath164 .
the curve @xmath11 is mapped onto a curve @xmath169 with ordinary @xmath15-fold singularities at the points @xmath170 , see figure [ conj3 ] .
an oval of @xmath11 is _ interior _ if it lies inside of one of the three nests , otherwise it is _
exterior_. an oval of @xmath11 is _ triangular ( quadrangular ) _ if it lies in @xmath84 ( @xmath171 ) .
an oval of @xmath169 is _ interior ( exterior , triangular , quadrangular ) _ if its preimage is .
the triangular ovals of @xmath169 are the ovals lying inside of @xmath172 .
the _ main part _ of @xmath169 is the union of @xmath173 and the images of the six principal ovals .
a conic through @xmath45 and two further points is mapped by @xmath164 onto the line passing through the images of the two points .
a rational quartic with nodes at @xmath45 and passing through five further points is mapped onto the conic passing through the images of these five points .
we shall always choose the two ( or the five ) points inside of non - principal ovals .
our goal is to find configurations of five ovals of @xmath11 leading to a contradiction : after @xmath164 , a conic passing through the images of these five ovals intersects the curve @xmath169 at @xmath174 points .
its preimage would be a rational quartic intersecting : each of the ovals @xmath175 at four points , each of the other five ovals at two points , and @xmath23 at four points .
a pencil of lines ( conics ) sweeping out a curve of degree @xmath29 is _ maximal _ if it has alternatively @xmath29 ( @xmath30 ) and @xmath31 ( @xmath32 ) real intersection points with the curve .
it is _ totally real _ if it has only real intersections with the curve .
the main part of @xmath169 divides the plane in @xmath176 zones .
the six zones corresponding to the interiors of the ovals @xmath45 contain no ovals . for @xmath168 , at most one of the four zones corresponding to the interior of @xmath19 may contain some ovals .
after @xmath164 , we will call _ chain _ a maximal fiedler chain of ovals lying in the same zone , swept out by some maximal pencil of lines .
the consecutive ovals are connected by vanishing cycles .
admissible distributions of ovals for the curves even , even , odd without jump ] the main part of @xmath169 and the base lines ] consider two non - principal ovals : @xmath177 exterior in @xmath51 and @xmath8 in @xmath52 , either exterior or extremal negative in the chain of @xmath43 ( in the latter case , @xmath43 is left ) .
note that such ovals exist , see figure [ types ] .
after @xmath164 , denote by @xmath178 the pencil of lines based at @xmath8 sweeping out @xmath173 from the first triangular oval to the last one , and meeting successively @xmath45 , this pencil is maximal .
we will say that an oval is interior to @xmath179 if its preimage is interior to @xmath19 .
if @xmath178 sweeps out some oval interior to @xmath179 , then it sweeps out all of the ovals interior to @xmath179 , consecutively ; the ordering of these ovals given by @xmath178 corresponds to the ordering of their preimages in the chain of @xmath19 .
the ovals interior to @xmath179 ( other than @xmath8 ) build a chain , formed by the images of the non - principal ovals interior to @xmath19 ( other than @xmath8 ) .
this follows from bezout s theorem , along with the fact that , in the pseudo - holomorphic setting , one may deform the curve , replacing the vanishing cycles connecting consecutive ovals in the chains by real double points , @xcite-@xcite .
the triangular ovals forming positive ( negative ) pairs with @xmath173 will be called _ plus - ovals ( minus - ovals)_. the plus - ovals ( minus - ovals ) are thus the negative ( positive ) ovals in @xmath166 and the positive ( negative ) ovals in @xmath180 .
for all three complex types of table [ orev6 ] , @xmath181 is empty ; for the first two complex types , @xmath182 is also empty .
there can not exist a configuration of three ovals @xmath183 met successively by @xmath178 , with @xmath16 in @xmath166 , @xmath184 in @xmath185 and @xmath186 quadrangular .
_ proof : _ the lines @xmath187 , @xmath188 and @xmath189 divide the sector @xmath190 , @xmath191 containing @xmath186 in four triangles , one of them is entirely contained in @xmath173 , see figure [ pqg ] .
the point @xmath186 may be chosen in any one of the other three triangles , denote by @xmath192 , @xmath193 and @xmath194 the three admissible choices of @xmath186 , the conic through @xmath195 is @xmath196 , @xmath197 or @xmath198 .
for each of the six lines determined by two of the points @xmath199 , consider the three intersection points with the base lines , we have thus five particular points on each line .
the ordering of these five points on the line is known only for @xmath190 , @xmath191 and @xmath188 . in the upper part of figure [ pqg ] , the position of the base lines with respect to the other three lines @xmath200 , @xmath187 ,
@xmath189 has been chosen arbitrarily among all admissible positions . in the lower part
, we have drawn in bold the boundaries of the zones containing each of the three base lines , each of these zones is a mbius band bounded by segments of lines : for @xmath201 , it is bounded by @xmath191 , @xmath200 , @xmath189 , @xmath16 lies outside ; for @xmath48 , it is bounded by @xmath190 , @xmath187 , @xmath200 , @xmath184 lies outside ; for @xmath47 , it is bounded by @xmath188 , @xmath187 , @xmath189 , @xmath8 lies outside .
( in the three pictures , the positions of @xmath186 are indicated only with the indices @xmath202 . )
the x assigned to a pair ( conic , base line @xmath22 ) means that the conic cuts @xmath22 ( twice ) and each of the objects @xmath203 , @xmath204 at four points .
we say that the conic is _ maximal _ with respect to the base line @xmath22 .
( we may say similarly that an arc of conic is maximal with respect to some base line . )
the three conics are all maximal with respect to the three base lines , and have each four intersection points with @xmath173 .
they cut @xmath169 at @xmath174 points , this is a contradiction .
@xmath205 there can not exist a configuration of four ovals @xmath16 , @xmath184 , @xmath206 , @xmath207 , with @xmath208 in @xmath185 , @xmath209 in @xmath166 and @xmath184 not in @xmath210 , such that @xmath178 meets successively @xmath211 or @xmath212 .
_ proof : _ assume there exist four such ovals , note that they are either all exterior or one unique oval among @xmath213 is interior to @xmath210 .
consider the pencil of conics @xmath214 , all of its conics have in total eight intersection points with the set of base ovals .
the pencil is divided in three portions by the double lines , in two of these portions , the conics cut @xmath173 four times and are maximal with respect to the three base lines .
these portions are totally real , see figure [ dpqrs ] , where the @xmath215 assigned to a pair ( portion , base line ) means that the portion is maximal with respect to the base line .
the remaining empty ovals are swept out by the third portion @xmath216 . depending on whether @xmath178 sweeps out @xmath211 or @xmath212 ,
the conic of @xmath214 passing through @xmath207 is @xmath217 or @xmath218 . taking account of the positions of the points in the various zones , and of the positions of the base lines given in the lower part of figure [ dpqrs ]
, we get the following arguments . in the first case ,
the arc @xmath219 cuts @xmath201 ( once ) , and does nt cut @xmath47 .
so it must cut @xmath48 twice , in such a way that it is maximal with respect to this line . in the second case ,
the arc @xmath220 cuts @xmath201 ( once ) , and does nt cut @xmath47 .
so it must cut @xmath48 twice , in such a way that it is maximal with respect to this line . in both cases ,
the conic is maximal with respect to the three base lines , and cuts @xmath173 four times , this is a contradiction .
@xmath205 for the first two complex types , the ovals swept out by @xmath178 are of three _ kinds _ : quadrangular , triangular in @xmath166 , triangular in @xmath185 .
for the third complex type , we have the same three kinds , plus supplementarily a chain of ovals in @xmath182 , interior to @xmath221 .
a triangular oval is exterior or interior to @xmath210 .
let us call _ sequence _ a maximal fiedler chain of ovals of the same kind , swept out consecutively by @xmath178 .
each sequence is formed of one or several chains .
a pair of odd chains of the same kind is _ inessential _ if the pencil sweeps out only even sequences between them . 1 .
there exist three exterior triangular plus - ovals @xmath160 , swept out successively by @xmath178 , each of them extremity of an odd chain , such that : for the first and third type , @xmath222 are in @xmath185 , @xmath7 is in @xmath166 ; for the second type , @xmath5 is in @xmath185 , @xmath10 are in @xmath166 . for any choice of @xmath160 ,
there exist two quadrangular ovals @xmath161 , each of them extremity of an odd chain , such that @xmath178 sweeps out successively @xmath159 with alternating orientations .
2 . let @xmath11 realize the third complex type , there exist five ovals @xmath223 , swept out successively by @xmath178 , extremities of odd chains , such that : @xmath222 are plus - ovals in @xmath185 , two of the ovals @xmath224 are plus - ovals in @xmath166 , the third one is a minus - oval in @xmath182 , interior to @xmath221 . if the minus - oval is @xmath184 or @xmath206 , let @xmath225 , if the minus - oval is @xmath16 , let @xmath226 , see figure [ abpqr ] . for any choice of @xmath223 , there exist furthermore two quadrangular ovals @xmath161 , each of them extremity of an odd chain , such that the pencil sweeps out successively : @xmath227 or @xmath228 with alternating orientations
3 . let @xmath11 realize the first type or the third type with @xmath229 , one may choose @xmath6 and @xmath7 such that the ovals in each zone @xmath166 and @xmath185 swept out beween @xmath6 and @xmath7 are distributed in even chains and disjoint pairs of inessential odd chains . _
note for the third type that 2 .
implies 1 .
let us `` delete '' all of the even sequences , and systematically disjoint inessential pairs of odd chains ( there are several ways to do that . ) in each odd sequence , delete all of the even chains , and then all of the ovals left but one extremity .
now , the pencil @xmath178 meets successively single ovals , such that any two consecutive ones are of different kinds , and each of them was originally extremity of an odd chain .
note that the deletion procedure has preserved the triangular parameters @xmath120 .
assume that @xmath178 meets successively @xmath0 in @xmath166 and @xmath230 in @xmath185 .
( note that for the third type , the ovals in @xmath182 will be swept out after @xmath230 . ) by lemma 1 , @xmath231 meets no quadrangular ovals .
so , this pencil must sweep out alternatively ovals in @xmath185 and @xmath166 ( the first one being in @xmath185 . by lemma 2 ,
there are no ovals at all between @xmath0 and @xmath230 . with the non - zero data @xmath133 , @xmath134 ( and @xmath232 for the third type )
, we deduce immediately the following property .
for the first ( second ) type , there exist three plus - oval @xmath160 swept out successively , such that @xmath222 are in @xmath185 , @xmath7 is in @xmath166 ( @xmath5 is in @xmath185 , @xmath10 are in @xmath166 ) .
one gets also the existence of @xmath224 for the third type .
the existence of ovals @xmath233 and @xmath234 as required is also clear by fiedler s theorem , we need yet to prove that they are quadrangular .
after the deletion procedure , the following obvious property holds : _ consider two ovals @xmath235 in the same zone @xmath236 ( @xmath237 or @xmath238 ) , such that @xmath178 meets no oval between them in @xmath239 .
then @xmath0 and @xmath230 have the same orientation and there is an odd number of ovals between them , alternatively quadrangular and in @xmath236 ( the first and the last are quadrangular ) .
_ let @xmath11 realize the first complex type , see upper part of figure [ efquad ] ( the quadrangular ovals have been placed arbitrarily in @xmath240 ) . by lemma 2 , there is no oval in @xmath166 between @xmath5 and @xmath6 , hence @xmath233 is quadrangular . if @xmath234 is in @xmath185 , then lemma 2 implies that there is no oval in @xmath166 between @xmath6 and @xmath234 , contradiction .
if @xmath234 is in @xmath166 , then lemma 2 implies that there is no oval in @xmath185 between @xmath234 and @xmath7 , contradiction again .
so , @xmath234 is quadrangular , point 1 .
is proved .
one may choose @xmath6 and @xmath7 such that @xmath6 is the last plus - ovals before @xmath234 in @xmath185 and @xmath7 is the first plus - oval after @xmath234 in @xmath166 .
we have already proved that there is no minus - oval between @xmath6 and @xmath7 in @xmath241 .
consider now again the whole set of ovals .
the pencil @xmath178 meets between @xmath6 and @xmath7 in @xmath241 only even chains and disjoint pairs of inessential odd chains , point 3 .
is proved .
caution : the argument does nt work if , instead of deleting ovals , we had chosen @xmath6 extremity of the last odd plus - chain in @xmath185 and @xmath7 extremity of the first odd plus - chain in @xmath166 .
the pencil @xmath178 could sweep out for example : @xmath6 , an even chain in @xmath240 , an odd chain in @xmath185 starting with a minus - oval @xmath242 and @xmath7 ( the odd chains containing @xmath6 and @xmath242 form an inessential pair ) . in this case
, we can not find @xmath234 quadrangular as required .
let @xmath11 have the second type , see lower part of figure [ efquad ] .
by lemma 2 , there is no oval in @xmath185 between @xmath6 and @xmath7 , so @xmath234 is quadrangular .
if @xmath233 is in @xmath185 , lemma 2 implies that there is no oval in @xmath166 between @xmath5 and @xmath233 , contradiction .
if @xmath233 is in @xmath166 , lemma 2 implies that there is no oval in @xmath185 between @xmath233 and @xmath6 , contradiction again .
hence , @xmath233 is quadrangular .
let @xmath11 have the third type .
for @xmath243 , there must be some oval @xmath234 between @xmath184 and @xmath206 by fiedler s theorem , and @xmath234 can be only quadrangular . apart from that , the same arguments as for the first type apply to prove that @xmath233 and @xmath234 are quadrangular . for @xmath229 ,
the same arguments as for the first type apply to prove point 3 .
@xmath205 @xmath16 in @xmath166 , @xmath184 in @xmath185 ] @xmath184 in @xmath166 exterior to @xmath210 , @xmath208 in @xmath185 ] third complex type , ovals @xmath223 ] quadrangular ovals @xmath233 , @xmath234 ] in the next lemmas 4 - 7 , @xmath8 and @xmath178 are as in the previous section , @xmath159 are five ovals swept out successively by @xmath178 , such that @xmath222 are exterior in @xmath185 , @xmath7 is exterior in @xmath166 , @xmath161 are quadrangular .
( these conditions are weaker than those in lemma 3 . )
denote by @xmath247 the sector swept out by the piece of pencil @xmath248 .
let @xmath249 $ ] be the segment of line @xmath250 that does not cut @xmath48 , and @xmath249'$ ] be the other segment .
we shall study the admissible distributions of the ovals in the various zones .
keep in mind that an oval in @xmath251 may be exterior , interior to @xmath252 or interior to @xmath203 .
if @xmath43 is left ( right ) , the ovals interior to @xmath221 are in @xmath253 ( @xmath240 ) . if @xmath43 is separating , the ovals interior to @xmath221 are in @xmath182 .
the intersection point @xmath254 lies on @xmath249 $ ] . _
proof : _ assume that @xmath254 lies on @xmath249'$ ] .
the lines @xmath255 , @xmath256 and @xmath257 divide the sector @xmath258 , @xmath259 containing @xmath233 in four triangles , one of these triangles is entirely contained in @xmath173 , see figure [ abfcd ] .
denote by @xmath260 the three choices of @xmath233 .
the conic through @xmath261 is @xmath262 , @xmath263 or @xmath264 , they cut all @xmath173 four times , and the first two are maximal with respect to the three base lines . hence , @xmath265 , the conic is @xmath266 . note that if @xmath48 does nt cut the triangle containing @xmath233 , then this triangle lies in @xmath267 , otherwise , @xmath165 divides this triangle in two pieces , one is in @xmath267 , the other in @xmath268 . as @xmath266 can not be maximal with respect to @xmath165 , the oval @xmath233 is in @xmath240 , or interior to @xmath269 in @xmath268 .
the lines @xmath256 , @xmath255 , @xmath257 , divide the sector @xmath259 , @xmath250 containing @xmath234 in four triangles , one of them is is entirely contained in @xmath173 , another is divided in two subtriangles by the line @xmath270 .
denote by @xmath271 the four admissible choices of @xmath234 , corresponding to @xmath272 , @xmath273 , @xmath274 , @xmath275 , @xmath276 , @xmath277 .
all of these conics cut @xmath173 four times .
the conics @xmath272 , @xmath273 and @xmath276 are maximal with respect to all three base lines , and the configuration of six points @xmath278 contradicts lemma 15 from @xcite , see figure [ case2 ] .
@xmath205 @xmath254 on @xmath249'$ ] , position of @xmath233 and @xmath234 , conics ] @xmath279 lie in convex position , contradiction ] the oval @xmath233 lies on a conic @xmath266 and is in @xmath240 ( interior to @xmath221 or exterior ) , or interior to @xmath221 in @xmath253 ( in this latter case , @xmath43 is left ) . _
proof : _ the lines @xmath255 , @xmath256 and @xmath257 divide the sector @xmath258 , @xmath259 containing @xmath233 in four triangles , one of these triangles is entirely contained in @xmath173 , see figure [ daebc ] .
the point @xmath233 may be chosen in any one of the other three triangles , denote by @xmath260 the three admissible choices of @xmath233 , the conic through @xmath261 is correspondingly @xmath280 , @xmath281 or @xmath282 , they cut all @xmath173 four times and the last two are maximal with respect to the three base lines .
the conic is @xmath266 , @xmath233 lies in a triangle bounded by the lines @xmath283 , @xmath257 , @xmath256 .
note that @xmath269 cuts twice each of the segment @xmath284 $ ] , @xmath285 $ ] with non - empty intersection with the mbius band corresponding to @xmath48 .
bezout s theorem between @xmath11 and the lines @xmath258 , @xmath259 implies thus that the triangle containing @xmath233 does nt intersect the interior of @xmath269 .
note also that if this triangle is not cut by @xmath201 , it lies entirely in @xmath267 .
if the triangle is cut by @xmath201 , one of the pieces obtained is in @xmath267 , the other is in @xmath253 .
assume @xmath233 is in @xmath253 , as @xmath266 can not be maximal with respect to @xmath49 , @xmath233 is interior to @xmath221 .
@xmath205 position of @xmath233 , conics ] the oval @xmath234 lies on a conic @xmath286 . _
proof : _ the six lines determined by @xmath287 divide the sector @xmath288 containing @xmath234 in seven zones , two of them lie entirely inside of @xmath173 .
denote by @xmath289 the remaining five admissible positions of @xmath234 .
each of them is characterized by the conic through @xmath290 , plus ( cases @xmath291 ) the position of @xmath8 inside ( @xmath292 ) or outside ( @xmath293 ) of this conic , see figure [ case1 ] , where @xmath234 lies in the hatched zones and the position of the base lines with respect to the dotted lines has been chosen arbitrarily among all possible positions .
all of the conics intersect @xmath173 at four points .
the zones containing the base lines are mbius bands bounded by segments of lines , they are shown in figure [ acebf ] , along with all five admissible positions of @xmath234 ( indicated with help of the indices @xmath294 ) .
the conics @xmath295 and @xmath296 are maximal with respect to all three base lines , contradiction .
the configuration of six points @xmath297 contradicts lemma 15 from @xcite , see second picture in figure [ case1 ] .
@xmath205 if an oval @xmath298 swept out between @xmath6 and @xmath7 is such that @xmath299 ( @xmath300 ) , we say that @xmath298 is in the zone @xmath301 ( @xmath302 ) , see the last two pictures in figure [ case1 ] .
note that the zone @xmath301 is a triangle bounded by the lines @xmath259 , @xmath250 , @xmath303 . in the case @xmath304
, we write shortly @xmath305 ( @xmath306 ) . from now on , we assume that if @xmath43 is left , then the oval @xmath8 of @xmath11 is the negative extremity of the chain of @xmath43 . 1 .
let @xmath43 be left .
the images by @xmath164 of the non - extreme ovals interior to @xmath43 are all together either in the zone @xmath301 , or outside of the sector @xmath247 .
the zone @xmath301 contains only these ovals , or is empty .
if @xmath306 , then @xmath234 may be : exterior or interior to @xmath210 in @xmath253 ; interior to @xmath210 in @xmath268 .
2 . let @xmath43 be separating .
@xmath305 may be : exterior or interior to @xmath210 in @xmath253 ; interior to @xmath210 in @xmath268
. @xmath306 may be : exterior or interior to @xmath210 in @xmath253 ; interior to @xmath210 in @xmath268 .
3 . let @xmath43 be right .
the admissible positions of @xmath234 are the same as for separating @xmath43 , plus one : @xmath305 interior to @xmath221 in @xmath240 _ proof : _ let @xmath305 and denote by @xmath307 $ ] the segment @xmath308 contained in the zone @xmath301 , @xmath307 $ ] does nt cut @xmath48 , and is interior to both conics @xmath309 and @xmath310 . each of these conics cuts @xmath173 four times , and
is maximal with respect to two base lines , see figure [ acebf ] .
if @xmath234 is in @xmath240 , the segment @xmath307 $ ] cuts @xmath201 .
if moreover @xmath234 is exterior to @xmath221 , the conic @xmath309 is maximal with respect to @xmath201 .
but this conic is also maximal with respect to the other two base lines , contradiction .
if @xmath234 is in @xmath268 , the segment @xmath307 $ ] cuts @xmath47 .
if moreover @xmath234 is exterior to @xmath210 , the conic @xmath310 is maximal with respect to @xmath47 .
but this conic is also maximal with respect to the other two base lines , contradiction .
hence , @xmath234 is in @xmath253 ; interior to @xmath210 in @xmath268 ; or interior to @xmath221 in @xmath240 .
let @xmath306 , the conic @xmath311 cuts @xmath173 four times and is maximal with respect to @xmath201 and @xmath48 .
the arc @xmath312 does nt cut @xmath201 , and cuts @xmath48 once , see figure [ acebf ] .
if @xmath234 is in @xmath240 or in @xmath268 exterior to @xmath210 , then @xmath312 is maximal with respect to @xmath47 , contradiction .
hence @xmath234 is in @xmath253 ; or interior to @xmath210 in @xmath268 .
note that if @xmath8 is exterior to @xmath221 ( @xmath43 is right or separating ) , then the zone @xmath302 does nt intersect the interior of @xmath221 .
points 2 . and 3 .
follow immediately . to get 1 .
, note that by bezout s theorem with @xmath221 , the line @xmath250 meets successively : @xmath8 , @xmath48 , @xmath7 , @xmath47 , @xmath49 .
the zone @xmath301 splits in two parts , one in @xmath240 , the other interior to @xmath221 in @xmath253 , see figure [ dfk ] where the boundary of the zone @xmath301 is drawn in bold , and the vanishing cycles connecting the images of the ovals interior to @xmath43 have been replaced by real double points .
@xmath205 consider five ovals @xmath159 satisfying with @xmath8 the conditions of lemma 3 . 1 .
let @xmath43 be left .
then @xmath306 , @xmath234 is positive exterior in @xmath253 .
2 . let @xmath43 be separating or right .
if @xmath305 , then @xmath234 is negative exterior in @xmath253 . if @xmath306 , then @xmath234 is positve exterior in @xmath253 .
_ proof : _ consider the intersections of the sector @xmath247 with the images of the quadrangles .
we see that @xmath313 has two connected components , @xmath314 and @xmath315 have each at most two connected components , see figure [ fq3 ] , where we have represented the set of objects under consideration in the upper part , and their preimages in the lower part .
if @xmath234 is in @xmath268 or @xmath253 , it is positive ( @xmath316 , @xmath317 ) or negative ( @xmath318 , @xmath319 ) depending on whether it lies in one connected component or in the other .
if @xmath234 is in @xmath240 , it is positive for both choices ( @xmath320 , @xmath321 ) of its position .
the case @xmath305 corresponds to the three positions @xmath316 , @xmath320 , @xmath319 , so if @xmath234 is in @xmath322 , @xmath234 is positive , if @xmath234 is in @xmath253 , @xmath234 is negative .
the case @xmath306 corresponds to the other three positions @xmath318 , @xmath317 , @xmath321 , so if @xmath234 is in @xmath323 , @xmath234 is positive , if @xmath234 is in @xmath268 , @xmath234 is negative . the ovals interior to @xmath210 form an even chain with respect to @xmath8 ,
so @xmath234 is not interior to @xmath210 .
the ovals interior to @xmath324 or @xmath325 bring a contribution of @xmath82 or @xmath27 to @xmath143 .
so , if @xmath234 is in @xmath240 or @xmath234 is in @xmath253 with the position @xmath317 , then @xmath234 is exterior .
@xmath205 let @xmath43 be right or separating .
there exist a base oval @xmath8 and five ovals @xmath159 ( seven ovals @xmath326 for the third type ) satisfying the conditions of lemma 3 , such that the zone @xmath301 is empty or contains only the even chain interior to @xmath210 ( so , @xmath306 ) .
_ proof : _ start with a set of five ( seven ) ovals satisfying the conditions of lemma 3 .
assume there exists @xmath327 exterior in @xmath253 , lying in the zone @xmath301 .
let @xmath328 be the pencil based at @xmath327 and sweeping out @xmath173 from the first triangular oval to the last one , in such a way that it meets successively @xmath45 .
the pencil @xmath328 meets successively @xmath329 . replacing the vanishing cycles in the chains by double real points
, we see that @xmath178 and @xmath328 give rise to the same sets of triangular chains , and of quadrangular chains that do not contain @xmath8 nor @xmath327 . using conics
, we can get more information .
let @xmath0 and @xmath230 be two triangular oval such that @xmath178 meets successively @xmath330 , then @xmath328 meets successively @xmath331 .
the pencil of conics @xmath332 has only one non - totally real portion , see figure [ emptyf4 ] .
any supplementary oval @xmath207 lies on a conic : @xmath333 , @xmath334 , @xmath335 , @xmath336 .
the pencils @xmath178 and @xmath328 sweep out the set of three ovals @xmath337 with the same ordering .
let @xmath338 and @xmath339 be the first and the last ( triangular ) oval met by @xmath178 , they are also the first and the last oval met by @xmath328 .
the two pencils sweep out the same set of ovals .
let @xmath340 be two ovals met successively by @xmath178 , one of them triangular , then @xmath340 are met successively by @xmath328 ( if @xmath215 is triangular , make @xmath341 , if @xmath342 is triangular , make @xmath343 ) . in particular , the orderings with which the triangular chains are swept out by @xmath178 and @xmath328 coincide .
let @xmath344 be two such chains , @xmath178 and @xmath328 sweep out the same set of quadrangular ovals between @xmath345 and @xmath346 .
we rename the ovals so that @xmath327 is now called @xmath8 , the new zone @xmath301 is contained in the old one . repeat the procedure as many times as necessary , in the end we have a new base oval @xmath8 such that @xmath347 ( @xmath348 ) are swept out successively by @xmath178 , these ovals are extremities of odd chains .
the ovals in @xmath253 swept out between @xmath6 and @xmath7 are all in the new zone @xmath302 .
finally , the pencil meets between @xmath6 and @xmath7 in @xmath241 only even chains and disjoint pairs of inessential odd chains .
we may choose now a new oval @xmath234 to fulfill the conditions of lemma 3 .
the pencil of conics @xmath349 is shown in figure [ pendfcb ] .
let us consider now the ovals lying in the new zone @xmath301 .
by lemma 7 , an oval @xmath298 in this zone may be : exterior or interior to @xmath210 in @xmath253 , interior to @xmath210 in @xmath268 or interior to @xmath221 in @xmath240 ( in this latter case , @xmath43 is right ) .
assume there exists @xmath298 interior to @xmath221 in @xmath240 , then @xmath301 contains the whole of the odd ( negative ) chain interior to @xmath221 .
let @xmath298 be extremity of this chain .
consider the pencil of conics @xmath350 , the portion @xmath351 is totally real , see figure [ dbmkc ] .
let @xmath352 be an oval swept out between @xmath6 and @xmath298 .
if @xmath352 is quadrangular , it must be in the zone @xmath302 , hence it is on a conic @xmath353 .
but this conic is in the totally real portion , contradiction .
if @xmath352 is in @xmath166 , @xmath352 lies on a conic @xmath353 or @xmath354 , in both cases , the conic is maximal with respect to all three base lines and cuts @xmath173 at four points , contradiction .
hence , @xmath352 is in @xmath185 .
all of the ovals swept out between @xmath6 and @xmath298 are in @xmath185 .
the orientations of @xmath6 and @xmath298 coincide , hence there should be an odd number of ovals between them .
this is a contradiction , as @xmath6 is by hypothesis extremity of an odd chain .
let us recap the new conditions on our set of ovals . if @xmath43 is right or separating , the ovals @xmath355 ( @xmath8 , @xmath5 , @xmath233 , @xmath6 , @xmath234 , @xmath16 , @xmath184 , @xmath206 ) are as in lemma 9 .
if @xmath43 is left , @xmath8 is the image of the negative extreme oval in the chain of @xmath43 , and @xmath355 satisfy the conditions of lemma 3 . in all cases ,
the zone @xmath301 contains no exterior ovals .
the five admissible positions of @xmath234 ] position of the base lines and of @xmath234 , conics ] @xmath43 left , two positions of the chain in @xmath221 : exterior to @xmath247 and in the zone @xmath301 ] @xmath316 , @xmath317 , @xmath320 , @xmath321 are positive , @xmath318 , @xmath319 are negative ] the pencil of conics @xmath332 ] let @xmath11 have the third complex type .
then , @xmath229 . _
proof : _ recall that the third complex type may realize the three cases shown in figure [ abpqr ] .
one has @xmath229 in the first two cases , and @xmath243 in the third case .
assume that @xmath243 , @xmath234 may be chosen in such a way that it is swept out after @xmath184 by @xmath178 ( see proof of lemma 3 ) .
the pencil of conics @xmath358 is divided in three portions by the double lines , in two of these portions , the conics have all @xmath359 intersection points with the main part of @xmath169 and the four base ovals , these portions are totally real , see figure [ pendfcb ] .
the remaining ovals are swept out by the third portion @xmath360 . as @xmath184 is swept out between @xmath6 and @xmath234 by @xmath178
, it is on a conic @xmath361 .
the arc @xmath362 cuts @xmath173 twice , each of the arcs @xmath363 and @xmath364 cuts @xmath173 once , hence @xmath365 must be entirely inside of @xmath173 .
the line @xmath47 separates @xmath6 from @xmath184 in @xmath173 , as moreover @xmath366 are not in @xmath210 , the conic @xmath361 is maximal with respect to @xmath47 , contradiction
. @xmath205 .
the pencil of conics @xmath349 ] @xmath8 is negative . _
proof : _ for @xmath43 left , this is already known .
let @xmath43 be right or separating .
by lemma 8 , @xmath368 is positive exterior in @xmath253 , it contributes @xmath26 to @xmath142 .
let @xmath298 be a quadrangular oval belonging to an odd chain , swept out between @xmath6 and @xmath7 by @xmath178 .
by lemma 9 , @xmath298 is in the zone @xmath302 .
the pencil of conics @xmath358 has only one non totally real portion , see figure [ pendfcb ] .
thus , @xmath298 is on a conic @xmath369 or @xmath370 .
let us choose @xmath234 to be the last quadrangular oval belonging to an odd chain that is met by @xmath178 before @xmath7 , the conic is @xmath369 .
let us call @xmath371 the triangular zone , bounded by the lines @xmath364 , @xmath372 , @xmath373 containing @xmath298 , this zone lies in @xmath374 ( see positions of the base lines given in figure [ pendfcb ] ) .
the zone @xmath371 contains either only exterior ovals in @xmath253 , or both exterior ovals in @xmath253 and the even chain interior to @xmath210 . in the latter case ,
the chain interior to @xmath210 lies either in @xmath268 or in @xmath253 .
the pencil @xmath375 sweeps out the ovals of @xmath371 plus ( only if @xmath43 is left ) an even chain interior to @xmath221 ( formed by the images of the non - extreme ovals interior to @xmath43 ) , fiedler s theorem with @xmath375 implies thus that the ovals in @xmath376 ( @xmath234 included , @xmath8 not ) contribute @xmath26 to @xmath142 .
the contribution of the ovals in @xmath371 ( @xmath8 now included ) to @xmath142 is @xmath82 or @xmath377 .
consider now the maximal pencil @xmath378 sweeping out @xmath371 .
any oval @xmath298 met by this pencil is on a conic @xmath369 , so @xmath298 is in @xmath371 .
the ovals @xmath8 and @xmath234 are connected via this pencil by an even number of ovals , their orientations are opposite .
@xmath205 there exists an oval @xmath379 exterior positive in @xmath253 , lying on conics @xmath380 , @xmath381 , @xmath382 .
_ proof : _ the ovals of @xmath253 in @xmath371 ( including now @xmath8 and @xmath234 ) bring a contribution @xmath82 to @xmath142 . as @xmath383 ( first complex type ) or @xmath325 ( third complex type ) , there must be some supplementary positive oval @xmath379 extremity of an odd chain in @xmath253 , that is not swept out by the pencil @xmath384 , the conic through @xmath385 is thus @xmath381 .
for @xmath43 left , the ovals interior to @xmath221 other than @xmath8 form an even chain , so we may assume that @xmath379 is not interior to @xmath221 .
the conic through @xmath386 is @xmath387 , see figure [ acebf ] with @xmath306 . in figure [ pendfcb ] one
may replace @xmath6 by @xmath5 .
the conic through @xmath388 is @xmath380 or @xmath389 . in the latter case , @xmath379 is swept out between @xmath5 and @xmath6 by @xmath178 , hence @xmath379 must verify the conditions required for @xmath233 , see lemma 5 . as @xmath379 is in @xmath253 , it should be in @xmath221 , contradiction .
so @xmath379 is on a conic @xmath380 . using the conics @xmath380 , @xmath381 , @xmath387
, we get the orderings @xmath390 and @xmath391 for the pencils of lines based at @xmath7 and @xmath8 respectively , hence the conic through @xmath392 is either @xmath382 or @xmath393 .
if the conic is @xmath393 , its arc @xmath394 cuts @xmath47 and @xmath48 ( see pencil @xmath395 in figure [ daebc ] ) , as @xmath396 are in @xmath253 and @xmath5 is in @xmath240 , the arc @xmath397 cuts @xmath47 and @xmath48 .
the arc @xmath397 of @xmath380 is homotopic with fixed extremities to the arc @xmath397 of @xmath393 .
the conic @xmath380 is maximal with respect to all three base lines and cuts @xmath173 four times , contradiction .
thus , the conic through @xmath392 is @xmath382 , see figure [ abdcj ] .
@xmath205 let @xmath177 be an exterior oval in @xmath268 .
then , @xmath177 lies on a conic @xmath398 .
_ proof : _ consider the pencil of conics @xmath399 , all of the conics in it cut @xmath173 four times , and two of the three portions determined by the double lines are totally real , see figure [ dageb ] . in this figure , we have indicated the positions of the base lines , distinguishing for @xmath201 the two cases @xmath233 in @xmath240 and @xmath233 in @xmath253 , interior to @xmath221 .
any remaining oval @xmath186 will be swept out by the third portion @xmath400 .
assume that the conic through @xmath186 is @xmath401 or @xmath402 , it can not be maximal with respect to @xmath49 hence @xmath186 is in @xmath240 , or in @xmath253 and interior to @xmath221 .
let @xmath177 be an exterior oval in @xmath268 and consider the conic of the pencil @xmath399 passing through @xmath177 , it is @xmath398 or @xmath403 . in the lower part of the figure , we have drawn the conic and indicated on it the two admissible positions @xmath404 ( conic @xmath405 ) and @xmath406 ( conic @xmath407 ) of @xmath177 .
we have also indicated two positions of the second base point : @xmath39 corresponding to the case @xmath233 in @xmath240 and @xmath408 corresponding to the case @xmath233 in @xmath253 , interior to @xmath221 .
the conic through @xmath409 is either @xmath403 or @xmath398 .
the conic through @xmath410 is in the portion @xmath411 of the pencil @xmath412 , it is @xmath413 ( otherwise , @xmath177 would be swept out by @xmath384 , but the exterior quadrangular ovals met by this pencil lie all in @xmath253 , see proof of lemma 11 ) similarly , we obtain the conic @xmath414 .
if @xmath177 lies on a conic @xmath403 , then the pencil of lines based at @xmath8 sweeps out successively : @xmath415 ( @xmath166 , hence @xmath7 is swept out in the portion @xmath416 ) .
we have thus the ordering @xmath417 , contradiction with the conic @xmath413 .
so , @xmath177 lies on a conic @xmath398 .
@xmath205 we will now find a contradiction using the configuration of eight ovals @xmath418 .
with help of the conics @xmath398 , @xmath380 , @xmath387 , @xmath266 and @xmath382 , we find the ordering for the pencil of lines based at @xmath419 .
the conic through @xmath420 can be @xmath421 , @xmath422 , @xmath423 or @xmath424 .
with help of @xmath398 , @xmath381 , @xmath387 , @xmath425 , @xmath266 , we find the ordering for the pencil based at @xmath6 : @xmath8 , @xmath234 , @xmath7 , @xmath379 , @xmath233 , @xmath5 , @xmath177 .
this rules out @xmath422 and @xmath424 .
the conic is @xmath423 or @xmath421 .
consider the pencil @xmath426 , see figure [ penabcj ] where we considered only the two relevant portions .
if the conic is @xmath421 , the arc @xmath427 does nt cut @xmath48 , and cuts @xmath47 ( once ) , as moreover , @xmath177 is exterior in @xmath268 , this arc must be maximal with respect to @xmath49 , contradiction .
if the conic is @xmath423 , the arc @xmath428 cuts both @xmath48 and @xmath49 ( once each ) , and must be maximal with respect to @xmath47 , contradiction .
the first and third complex type are not realizable .
@xmath298 interior to @xmath221 in @xmath240 swept out between @xmath6 and @xmath7 ] position of @xmath379 ] pencil of conics @xmath431 , position of @xmath177 ] pencil of conics @xmath433 let now @xmath11 realize the second complex type .
let @xmath249 $ ] be the segment of the line @xmath250 that does nt cut @xmath48 , and @xmath249'$ ] be the other segment .
we distinguish two cases , depending on whether the intersection @xmath254 lies on @xmath249 $ ] or on @xmath249'$ ] , see figure [ ef1 ] .
the oval @xmath234 is swept out between @xmath6 and @xmath7 by @xmath178 , the sector containing @xmath234 is divided in four triangles by the lines @xmath256 , @xmath255 , @xmath257 , one of them is entirely interior to @xmath173 so @xmath234 is in one of the other three . in both cases ,
the conic through @xmath434 is @xmath310 , @xmath435 or @xmath311 .
consider the pencil of conics @xmath395 . in both cases ,
the portion @xmath436 is totally real , hence the conic must be @xmath435 , it is in the portion @xmath437 .
the conics of this portion are maximal with respect to : @xmath48 and @xmath47 in the first case , @xmath47 , @xmath201 in the second case . in the first case ,
the triangle containing @xmath234 is divided in two pieces by @xmath49 , one in @xmath166 , the other in @xmath438 . as @xmath234 is quadrangular , @xmath234 is in @xmath268 , the arc @xmath439 of @xmath435 is maximal with respect to @xmath49 , and @xmath435 cuts @xmath173 four times , contradiction . in the second case ,
the triangle containing @xmath234 is divided in two pieces by @xmath48 , one in @xmath166 , the other in @xmath440 .
as @xmath234 is quadrangular , @xmath234 is in @xmath253 , the arc @xmath439 of @xmath435 is maximal with respect to @xmath48 , and @xmath435 cuts @xmath173 four times , contradiction .
the second complex type is not realizable .
this finishes the proof of theorem 1 .
@xmath205 the second complex type ]
assume there exists an @xmath0-curve @xmath11 even , even , odd with jump , it realizes one of the complex types listed in table [ orev10 ] . recall that if @xmath33 is positive ( negative ) , @xmath33 is crossing ( non - crossing ) , see section 1.2 .
let @xmath441 ( @xmath27 ) if @xmath33 is positive ( negative ) .
the six principal ovals , the base lines , and @xmath23 divide the plane in zones .
the contributions of the non - principal ovals to @xmath28 in each zone are indicated in figures [ xing02]-[crossnot ] .
let us perform again the cremona transformation @xmath164 based at @xmath45 , @xmath11 is mapped onto a curve @xmath169 with three @xmath15-fold singularities .
we use the same conventions as in the case without jump , see section 2.1 .
let us choose two ovals @xmath7 and @xmath8 of @xmath11 such that @xmath443 are extremities of the three successive jump chains . in figure [ lambdaquad ]
, we have represented the image @xmath173 of @xmath23 , plus the images of @xmath33 , @xmath7 and @xmath8 , the upper part corresponds to @xmath33 crossing , the lower part to @xmath33 non - crossing .
the pencil of conics @xmath444 sweeping out @xmath11 is mapped by @xmath164 onto a pencil of lines @xmath445 .
let @xmath446 be the tangency points of @xmath445 with @xmath210 and let @xmath447 be the points of tangency of @xmath445 with @xmath173 .
by lemma 17 from @xcite , the ovals in @xmath173 are all met consecutively by @xmath445 , they form a fiedler chain : _ triangular ovals _ @xmath448 ( @xmath33 crossing ) or @xmath449 _ triangular ovals _ ( @xmath33 non - crossing ) . for @xmath33 crossing , one has @xmath450 .
denote by @xmath233 the first oval of the triangular fiedler chain , in the figure , @xmath233 has been placed arbitrarily in @xmath181 , it could be as well in @xmath182 or @xmath166 . for @xmath33 non - crossing
, one has @xmath451 .
denote by @xmath234 the last oval of the triangular chain . applying fiedler s theorem with the pencil @xmath445
, we get the let @xmath452 , one has @xmath453 . for @xmath33 crossing , the quadrangular ovals are swept out by @xmath445 beween @xmath184 and @xmath233 . between @xmath184 and @xmath8 ,
there is one single fiedler chain , whose contribution to @xmath454 is @xmath80 , between @xmath8 and @xmath233 , there are two fiedler chains , starting at @xmath455 and ending at @xmath456 , they contribute @xmath82 to @xmath454 . for @xmath33
non - crossing , the quadrangular ovals are swept out by @xmath445 between @xmath234 and @xmath457 . between @xmath234 and @xmath8 , there are two fiedler chains starting at @xmath458 and ending at @xmath459 , they contribute @xmath82 to @xmath454 , between @xmath8 and @xmath457 there is one single fiedler chain whose contribution to @xmath454 is @xmath27 .
( in all these fiedler chains , the contributions of @xmath8 , @xmath233 and @xmath234 are not included by convention .
as @xmath8 and @xmath7 have opposite orientations , they contribute together @xmath82 to @xmath454 . )
the identity @xmath453 may actually be obtained directly from table [ orev10 ] . for @xmath33
crossing , there exists an oval @xmath177 , swept out between @xmath184 and @xmath8 , that is positive in @xmath460 or negative in @xmath240 , we denote by @xmath461 the three admissible choices of @xmath177 . for @xmath33 non - crossing
, there exists an oval @xmath177 swept out between @xmath8 and @xmath457 , that is negative in @xmath460 or positive in @xmath240 , we denote by @xmath461 the three admissible choices of @xmath177
. see figure [ lambdaquad ] .
the piece of pencil @xmath445 sweeping out the quadrangular ovals is divided in two portions by the line @xmath462 , the contribution of each portion to @xmath463 is given by lemma 14 . in the next lemmas
, we will find pencils of conics based at @xmath7 , @xmath8 and two other ovals , having only one non - totally real portion .
this allows to study the distribution of the quadrangular ovals between the two parts of @xmath445 and sometimes to find a contradiction .
to this purpose , we may ignore the even chains and treat each odd chain as if it consisted of a single oval .
so we may assume without loss of generality that @xmath44 contains only the one oval @xmath39 , and that each of the three jump chains consists of a single oval @xmath443 .
_ proof : _ let @xmath11 have this complex type , it has non - crossing jump , see figure [ crossnot ] .
let @xmath186 be the positive extremal oval in the chain of @xmath43 .
if @xmath43 is right , @xmath465 is in @xmath53 , if @xmath43 is left , @xmath466 is in @xmath52 .
after @xmath164 , consider the pencil of conics @xmath467 .
two portions are maximal with respect to the three base lines , see figures [ hfcd ] .
all conics of the pencil have supplementarily two intersection points with @xmath210 and two intersection points with @xmath173 .
so @xmath467 has only one non - totally real portion , @xmath468 , that sweeps out any empty non - principal oval @xmath469 of @xmath11 other than @xmath470 , see figures [ hfcd]-[noncross ] . in figure [ noncross ] , we have represented a conic of this portion for either case @xmath43 right and @xmath43 left .
let @xmath469 be a quadrangular oval .
if @xmath469 is in @xmath268 or @xmath469 is exterior to @xmath221 in @xmath240 , @xmath469 is on the arc @xmath372 of the conic , hence @xmath469 is swept out between @xmath234 and @xmath8 by @xmath445 . if @xmath469 is in @xmath253 or interior to @xmath221 in @xmath240 , @xmath469 is on the arc @xmath471 of the conic , hence @xmath469 is swept out between @xmath8 and @xmath457 . for @xmath43 left , one should have @xmath472 and @xmath473 ( by lemma 14 ) .
but actually @xmath474 and @xmath475 , contradiction .
let @xmath476 and @xmath477 be the respective contributions to @xmath143 of the exterior ovals and of the ovals interior to @xmath43 . for @xmath43 right , one must have @xmath478 and @xmath479 , there is no contradiction .
_ proof : _ let @xmath11 realize one of these complex types , @xmath43 may be non - separating , right or left , see figure [ crosses ] .
assume there is an exterior oval @xmath186 in @xmath182 ( this is true for @xmath43 non - separating ) .
assume there exists an ( exterior ) oval @xmath298 that is @xmath481 in @xmath268 , see figure [ ovalint1 ] .
consider the pencil of conics @xmath485 .
the three base lines are in the mbius band obtained cutting away the affine ( in the plane of the figure ) triangle @xmath486 from @xmath2 .
all conics of the pencil are maximal with respect to each base line .
moreover , the conics cut @xmath173 twice , and @xmath210 at two supplementary points , hence the pencil is totally real , contradiction .
the pencil of lines @xmath445 meets no oval in @xmath268 between @xmath184 and @xmath8 , hence there exists a positive oval @xmath487 or a negative oval @xmath488 that is @xmath481 in @xmath323 , see figure [ ovalint1 ] .
consider now the pencil of conics @xmath489 .
this pencil has only one non totally real portion : @xmath490 , see figure [ hgcd ] .
( if @xmath491 is not interior to @xmath221 , this portion is maximal also with respect to the line @xmath49 . )
let @xmath469 be a quadrangular oval .
if @xmath469 is in @xmath323 , @xmath469 is on the arc @xmath492 of the conic , hence @xmath469 is @xmath481 . if @xmath469 is in @xmath268
, @xmath469 is on the arc @xmath493 of the conic , hence @xmath469 is swept out between @xmath8 and @xmath186 by @xmath445 .
one must have thus @xmath494 ( lemma 14 ) .
but for all three cases @xmath43 left , right and up , @xmath495 , contradiction .
@xmath205 _
proof : _ for this complex type , @xmath496 , see lower part of figure [ crosses ] .
the triangular ovals are all interior to @xmath221 in @xmath166 .
let @xmath466 be @xmath481 in @xmath253 , see figure [ ovalint0 ] .
the pencil @xmath497 has only one non - totally real portion : @xmath498 , see figure [ ehcd ] .
let @xmath469 be a quadrangular oval .
if @xmath469 is in @xmath323 , @xmath469 is on the arc @xmath499 of the conic , hence @xmath469 is @xmath481 .
if @xmath469 is in @xmath268 , @xmath469 is on the arc @xmath500 , hence @xmath469 is @xmath482 .
one should have thus : @xmath494 and @xmath501 , contradiction .
let @xmath502 be @xmath481 in @xmath268 .
the pencil @xmath497 has only one non - totally real portion : @xmath503 .
if @xmath469 is in @xmath322 , @xmath469 is on the arc @xmath504 of the conic , hence @xmath469 is @xmath481 .
if @xmath469 is in @xmath253 , @xmath469 is on the arc @xmath505 , hence @xmath469 is @xmath482 .
one should have thus : @xmath506 and @xmath507 , contradiction .
see figures [ ovalint0]-[ehcd ] .
all of the ovals between @xmath184 and @xmath8 are in @xmath240 , they contribute @xmath27 to @xmath143 , between @xmath8 and @xmath233 , we find the ovals of @xmath374 , and the remaining ovals of @xmath240 , whose contribution to @xmath143 is @xmath377 .
let @xmath194 be an oval @xmath482 in @xmath240 .
the pencil of conics @xmath508 has only one non - totally real portion : @xmath509 .
if @xmath469 is in @xmath240 , @xmath469 is on the arc @xmath510 of the conic , hence @xmath469 is @xmath482 .
there are no ovals in @xmath240 between @xmath184 and @xmath8 , contradiction .
@xmath205 _ proof : _
hereafter , @xmath469 and @xmath186 stand for quadrangular ovals , let us write @xmath517 if @xmath186 is in @xmath251 .
assume there exists an oval @xmath518 that is exterior in @xmath166 or @xmath519 interior to @xmath221 in @xmath182 .
( note that if @xmath520 there exists @xmath518 , and there exists @xmath519 if and only if @xmath43 is up ) .
if all of the ovals in @xmath251 are @xmath481 ( @xmath482 ) , we say shortly that @xmath251 is @xmath481 ( @xmath482 ) .
assume there exists @xmath521 , the pencil of conics @xmath522 has only one non - totally real portion : @xmath523 , see upper part of figure [ xingplus ] and figure [ gcdh ] .
consider a conic of this portion , passing through a quadrangular oval @xmath469 .
if @xmath469 is in @xmath322 , then @xmath469 is on the arc @xmath524 of the conic , hence @xmath469 is @xmath481 .
if @xmath469 is in @xmath253 , then @xmath469 is on the arc @xmath200 , hence @xmath469 is @xmath482 .
one has thus @xmath506 and @xmath507 .
but for @xmath43 non - separating , @xmath525 and @xmath383 , contradiction . for @xmath43 up , @xmath526 , @xmath507 , @xmath527
, there is no contradiction . in summary : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(1 ) ( @xmath43 non - separating and @xmath528 ) @xmath529 ( @xmath530 ) .
+ ( 2 ) ( @xmath43 up and @xmath531 ) @xmath529 ( @xmath532 , @xmath533 ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ assume there exists @xmath534 , the pencil of conics @xmath535 has only one non - totally real portion : @xmath536 , see upper part of figure [ xingplus ] and figure [ gcdh ] . consider a conic of this portion , passing through a quadrangular oval @xmath469 .
if @xmath469 is in @xmath253 , then @xmath469 is on the arc @xmath537 of the conic , hence @xmath469 is @xmath481 .
if @xmath469 is in @xmath268 , then @xmath469 is on the arc @xmath538 of the conic , hence @xmath469 is @xmath482 .
let now @xmath518 .
if @xmath469 is in @xmath240 , then @xmath469 is on the arc @xmath539 of the conic , hence @xmath469 is @xmath481 . therefore , @xmath494 and @xmath501 . for @xmath43 non - separating , @xmath79 , and for @xmath43 up , @xmath540 . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(3 ) ( @xmath541 and @xmath528 ) @xmath529 ( @xmath542 , @xmath530 , @xmath79 for @xmath43 non - separating , @xmath540 for @xmath43 up ) .
+ ( 4 ) ( @xmath541 and @xmath43 is up ) @xmath543 ( @xmath544 , @xmath530 ) . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ assume there exists @xmath545 , the pencil of conics @xmath546 has only one non - totally real portion : @xmath547 , see upper part of figure [ xingplus ] and figure [ gcdh ] . consider a conic of this portion , passing through a quadrangular oval @xmath469 .
if @xmath518 , then @xmath469 is on the arc @xmath548 of the conic , hence @xmath469 is @xmath482 .
there are no quadrangular ovals @xmath481 , contradiction .
if @xmath519 and @xmath469 is in @xmath322 , then @xmath469 is on the arc @xmath548 of the conic , hence @xmath469 is @xmath482 .
the zones @xmath268 and @xmath240 are @xmath482 , hence there exists some oval @xmath549 , thus @xmath253 is @xmath481 ( 4 ) .
one has thus @xmath383 and @xmath550 .
but for @xmath43 up , @xmath526 , @xmath507 , @xmath551 , contradiction .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(5 ) ( @xmath552 ) @xmath529 ( @xmath553 ) . +
( 6 ) ( @xmath552 and @xmath541 ) @xmath529 ( @xmath542 and @xmath530 , @xmath79 for @xmath43 non - separating , @xmath540 for @xmath43 up ) .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ assume there exists an oval @xmath298 in @xmath181 .
let @xmath502 be @xmath481 , the pencil of conics @xmath554 has only one non - totally real portion : @xmath555 , see lower part of figure [ xingplus ] and figure [ cdhk ] .
let @xmath469 be in @xmath322 , @xmath469 is on the arc @xmath556 of the conic , hence @xmath469 is @xmath481 .
let @xmath469 be in @xmath253 , @xmath469 is on the arc @xmath557 of the conic , hence @xmath469 is @xmath482 .
one has thus @xmath506 and @xmath507 , contradiction for @xmath43 non - separating .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(7 ) ( @xmath558 and @xmath43 is non - separating ) @xmath529 ( @xmath530 ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(8 ) ( @xmath558 ) @xmath529 ( @xmath533 ) . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ let @xmath465 be @xmath481 , the pencil of conics @xmath560 has only one non - totally real portion : @xmath561 .
let @xmath469 be in @xmath322 , @xmath469 is on the arc @xmath562 of the conic , hence @xmath469 is @xmath481 .
let @xmath469 be in @xmath253 , then @xmath469 is on the arc @xmath557 of the conic , hence @xmath469 is @xmath482 .
one has thus @xmath506 and @xmath507 , contradiction for @xmath43 non - separating .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(9 ) ( @xmath558 and @xmath563 ) @xmath529 ( @xmath268 , @xmath553 , @xmath533 , @xmath43 is up ) + ( 10 ) ( @xmath558 and @xmath43 is non - separating ) @xmath529 ( @xmath268 , @xmath253 , @xmath564 ) , contradiction . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for @xmath43 non - separating , the zone @xmath181 must be empty ( 10 ) .
as @xmath565 , one has @xmath136 , so there exists @xmath518 in @xmath166 .
the ovals in @xmath268 are @xmath482 ( 1 ) . as @xmath520 ,
the ovals in @xmath253 are also @xmath482 ( 3 ) .
the ovals in @xmath240 are @xmath481 ( 5 ) .
one should have thus @xmath566 , this is a contradiction . the complex type with non - separating @xmath43 is not realizable . for @xmath43 up ,
the ovals in @xmath268 are all @xmath481 or @xmath482 ( 2 ) .
assume there exists @xmath534 , then on one hand @xmath540 ( 6 ) , on the other hand , the zone @xmath181 must be empty ( 8) .
but @xmath567 , contradiction .
therefore , @xmath253 is @xmath482 .
the zone @xmath240 is @xmath481 ( 5 ) .
if the ovals in @xmath268 are @xmath482 , we have @xmath568 hence there exist @xmath194 in @xmath240 and @xmath298 in @xmath181 . but
the existence of two ovals @xmath569 and @xmath298 implies that @xmath268 is @xmath481 ( 9 ) . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(11 ) @xmath43 is up @xmath529 @xmath532 , @xmath533 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ as @xmath506 , there exists an oval @xmath186 that is @xmath481 in @xmath268 or in @xmath240 .
let @xmath570 , recall that for any choice of @xmath177 in @xmath571 , the pencil of conics @xmath522 has only one non - totally real portion : @xmath572 .
an oval @xmath298 in @xmath181 will be on the arc @xmath200 of its conic , see figure [ gcdh ] .
let now @xmath465 and @xmath298 be in @xmath181 , recall that the pencil of conics @xmath573 has only one non - totally real portion : @xmath574 .
the oval @xmath177 is on the arc @xmath575 of its conic , see figure [ cdhk ] . in all cases ,
the pencil @xmath576 meets the ovals in @xmath181 before those in @xmath571 , the first oval met ( @xmath233 ) is positive , hence @xmath577 or @xmath26 .
@xmath205 s. orevkov : _ complex orientation formulas for @xmath0-curves of degree @xmath580 with 4 nests . _ http://picard.ups-tlse.fr/ orevkov , annales de la facult des sciences de toulouse , mathmatiques ( 6 ) , 19 ( 2010 ) , 13 - 26 | we consider the @xmath0-curve of degree nine with three nests @xmath1 in @xmath2 . after systematic constructions , korchagin conjectured that at least two of the @xmath3 must be odd .
it was later proved that there is always one odd @xmath3 .
we say that the curve has a jump in a non - empty oval @xmath4 if there exist four ovals @xmath5 , @xmath6 , @xmath7 , @xmath8 , with @xmath5 interior to some other non - empty oval @xmath9 , @xmath8 exterior , @xmath10 interior to @xmath4 , such that @xmath6 and @xmath7 are separated inside of @xmath4 by any line passing through @xmath5 and @xmath8 . in this paper , we prove the conjecture for the curves without jump , and we find restrictions on the complex orientations and rigid isotopy types admissible for the curves even , even , odd with jump . |
ultra cold gases in optical lattices offer an extraordinary controllable environment for the analysis of many - body phenomena , as highlighted by the realization of the mott insulator to superfluid phase transition in bose gases @xcite , and more recently by the observation of the mott insulator to metal transition in two component fermi systems @xcite .
in addition to a wealth of possible quantum phases @xcite , and due to its discrete nature , optical lattices offer as well novel possibilities for composite formation , including local on - site dimer ( or @xmath1-mer ) formation in attractive gases @xcite and fermionic composites in bose - fermi mixtures @xcite .
an even more striking effect of the lattice discreteness is given by the recent observation of repulsively bound atomic pairs ( doublons ) at a given lattice site @xcite , which are dynamically stable in the absence of dissipation .
note that contrary to standard condensed - matter systems , where e.g. interactions with phonons would lead to a rapid dissipation of the dimers , optical lattices provide to a large extent a dissipation - free environment .
indeed experiments on optical lattices provide an extraordinary scenario for the study of the intriguing physics of metastable bound states and other far - of - equilibrium phenomena @xcite .
spinor gases , formed by atoms with various available zeeman substates , present a very rich physics due to the interplay between internal and external degrees of freedom @xcite .
interestingly , interatomic interactions lead to spin - changing processes in which population is transferred between zeeman sublevels .
the corresponding spinor dynamics has attracted a large interest , mostly in the realm of spinor bose - einstein condensates @xcite .
spinor lattice gases offer fascinating novel physics , most relevantly in what concerns quantum magnetism , including antiferromagnetic order in spin-@xmath2 fermions @xcite and even more intriguing phases for higher spins @xcite .
repulsively on - site bound pairs in spin-@xmath2 fermions are formed since the large interaction energy of the doublon can not be accommodated by the maximal kinetic energy for two atoms in the lowest band ( proportional to the hopping energy ) @xcite . in this paper
we show that a completely different mechanism , based on the interplay between spin - changing and quadratic zeeman effect ( qze ) , may sustained novel types of composites for repulsive high - spin lattice fermions .
we illustrate this physics for the case of hard core 1d spin-@xmath0 fermions in a mott insulator initially polarized in @xmath3 .
we show that spin - flips into @xmath4 may lead to two types of composites formed by an exciton ( @xmath4 particle and @xmath3 hole ) and an antisymmetric biexciton ( @xmath5 particles and two @xmath3 holes ) ( fig .
[ fig:1 ] ) .
we show that the dynamics and stability of the exciton - biexciton composites exhibits a non - trivial dependence with spin - changing , qze and center - of - mass momentum .
finally , we show that inelastic composite - composite interactions may be very inefficient , opening exciting possibilities for many - body composite gases .
particle and a @xmath6 hole ; ( b ) biexciton of by two @xmath5 particles and two @xmath3 holes ; ( c ) broken biexciton due to @xmath7 processes ( see text).,scaledwidth=45.0% ] [ fig:1 ] the structure of the paper is as follows . in sec .
[ sec : model ] we introduce the spin-@xmath0 system under consideration as well as the possible exciton , and biexciton excitations occurring after an individual spin flip .
[ sec : bound ] introduces the idea of repulsively bound exciton - biexciton pairs , with a particular emphasis in their band - like dispersion and the conditions for their existence .
[ sec : dynamics ] is devoted to the intriguing dynamics of the repulsively bound pairs , and its dependence on the spatial delocalization of the initial spin - flip excitations . in sec .
[ sec : multiple ] we analyze the case of multiple spin - flip excitations . finally , in sec .
[ sec : conclusions ] we summarize our conclusions .
we consider spin-@xmath0 fermions ( @xmath8 ) in a deep 1d lattice , such that at low filling only the lowest band is relevant . in this regime ,
the physics is given by the interplay between inter - site hopping ( characterized by the constant @xmath9 ) , qze ( given by a constant @xmath10 , which may be externally controllable ) , and @xmath11-wave collisions .
the latter may occur in two different channels with total spin @xmath12 and @xmath13 , characterized by the coupling constants @xmath14 , with @xmath15 the scattering length and @xmath16 the atomic mass .
although typically @xmath17 and @xmath18 are similar , they may be varied by microwave dressing @xcite or optical feshbach resonances @xcite .
below we use @xmath19 , and @xmath20 .
the @xmath11-wave interactions preserve magnetization ( and hence the linear zeeman effect is irrelevant ) , but may induce spin - changing collisions ( characterized by @xmath21 ) , which redistribute the populations at different components .
for @xmath22 we may consider maximally one fermion per site ( hard - core case ) .
for a sufficiently large chemical potential the system enters into the mott insulator with one fermion per site . the spin physics within the mott insulator
is given by an effective hamiltonian : @xmath23 where @xmath24 denotes nearest neighbors and @xmath25 , \label{eq : hij}\end{aligned}\ ] ] describes the effects of the interatomic interactions , with @xmath26 where @xmath27 , @xmath28^{1/2}$ ] , and @xmath29^{1/2}+2q\right ) /gg$ ] @xcite .
the constant @xmath30 ( @xmath31 ) characterizes the super - exchange between neighboring sites with @xmath32 ( @xmath33 ) , whereas @xmath7 does the same for sites with @xmath34 and @xmath35 such that @xmath36 .
finally , @xmath37 characterizes the super - exchange leading to spin - changing from @xmath33 to @xmath32 or viceversa .
we consider a system initially polarized into @xmath3 , @xmath38 , where @xmath39 is the vacuum .
@xmath40 is stable since collisions preserve magnetization .
we are interested in excitations of the form @xmath41 , created by spin flips .
note that these excitations have actually an _ exciton - like character _
, since they are formed by a particle ( in @xmath42 ) and a hole ( in @xmath3 ) ( fig .
[ fig:1]a ) .
a single spin - flip into @xmath5 is not accompanied by spin - changing , and hence lead to isolated @xmath33 excitons which hop with an effective hopping @xmath7 .
spin - changing collisions lead to a much less trivial dynamics when spins are flipped into @xmath4 .
after the spin - flip the @xmath0 exciton may hop to the nearest neighbor via @xmath30 super - exchange ( second line in eq . ) .
in absence of spin - changing ( @xmath43 ) the @xmath0 exciton propagates freely with an energy @xmath44 , where @xmath45 is the center - of - mass momentum of the @xmath0 exciton .
spin - changing modifies this simple picture , since the @xmath4 spin may interact via @xmath37 super - exchange with a neighboring @xmath6 spin leading to two neighboring spins with @xmath5 , i.e. a biexciton of the form @xmath46 ( fig . [ fig:1]b ) .
after spin - changing , one of the newly created excitons , say @xmath47 , may move apart from the neighboring @xmath48 one by a @xmath7 superexchange ( fig .
[ fig:1]c ) .
further @xmath7 processes bring the pair even further away from each other .
since an isolated @xmath33 spin surrounded by @xmath6 neighbors remains stable against spin - changing , one naively expects all @xmath0 excitons to eventually dissolve into independent @xmath33 ones , which would freely move away with a hopping @xmath7 .
as shown below , contrary to the naive expectation , the @xmath0 excitons may become very robust .
we introduce the fourier transformed states @xcite @xmath49 and @xmath50 , where @xmath51 is the number of sites and @xmath45 the center of mass quasi - momentum either of the @xmath0 exciton or of the @xmath33 biexciton .
note that the biexciton states @xmath52 contain not only biexcitons of neighboring @xmath33 ( i.e. @xmath53 , as in fig .
[ fig:1]b ) , but also broken biexcitons ( i.e. @xmath54 , as in fig . [ fig:1]c ) .
this is true even for the bound solutions discussed before , which may contain as well a contribution of broken biexciton pairs .
it is particularly useful to introduce the symmetric and antisymmetric states @xmath55 .
note that the hamiltonian do not couple symmetric and antisymmetric states , and hence the hamiltonian in @xmath45-space splits into @xmath56 .
the physics of the symmetric biexcitons is given by @xmath57 with @xmath58 , @xmath59 and @xmath60 , whereas the coupling between @xmath0 excitons and antisymmetric @xmath33 biexcitons is given by @xmath61 where @xmath62 , @xmath44 , and @xmath63 .
, for @xmath64 and @xmath65 .
the circles indicate the results of eq .
.,scaledwidth=40.0% ] figure [ fig:2 ] shows a typical spectrum ( which resembles that of self - bound repulsive pairs of spin-@xmath2 fermions @xcite ) .
unbound @xmath33 biexcitons form a continuous band , due to their free relative motion with an effective hopping @xmath66 , which leads to a dispersion @xmath67 , where @xmath68 is the relative momentum of the pair .
we observe also the appearance of up to three bound states ( resembling the case of repulsive spin-@xmath2 fermions with nearest neighbor interactions @xcite ) .
one of the bound states ( upper state in fig .
[ fig:2 ] ) , with energy @xmath69 , belongs to the symmetric manifold , being hence decoupled from the @xmath0 exciton , and thus we do not consider it further .
the other two bound states result from the coupling of antisymmetric @xmath33 biexcitons and @xmath0 excitons . in order to understand these states
, we introduce the ansatz @xmath70 where @xmath71 is the localization length . imposing @xmath72 results in a transcendent equation for @xmath71 : @xmath73 \nonumber \\ & & \pm \frac{1}{2}\sqrt { \left [ e_{1a}+\omega_2 e^{1/\gamma}-e_{3/2 } \right ] ^2 + 4\omega_{sc}^2 } , \label{eq : trasc}\end{aligned}\ ] ] where the relation between energy and @xmath71 is given by @xmath74 in fig .
[ fig:2 ] we indicate the results obtained from eqs . and , which are in excellent agreement to those directly obtained from the diagonalization of .
. shows that the separation between the bound state and the lowest boundary @xmath75 of the sea of unbound @xmath33 pairs is crucial for its localization , and hence while the lower bound state is tightly bound , the upper one is much more loose . in particular , the upper bound state unbinds ( @xmath71 diverges ) if its energy approaches @xmath76 .
hence the existence of the bound exciton - biexciton states depends non - trivially on qze , spin - changing collisions , and exciton momentum .
[ fig:3 ] shows those values of @xmath21 and @xmath10 for which one or two bound states are expected for @xmath77 .
( black ) , @xmath78 ( grey ) and @xmath13 stable exciton - biexciton composites for @xmath77 , as a function of @xmath21 and @xmath79 for @xmath80.,scaledwidth=45.0% ]
the mixing angle @xmath81 fulfills @xmath82 the contribution of the @xmath0 exciton to the bound states ( @xmath83 ) is particularly important to understand the dynamics following a spin - flip into @xmath4 .
in particular , at @xmath77 ( when the center of mass of the spin - flip excitation is maximally delocalized ) , @xmath84 and hence the upper bound state is a pure antisymmetric @xmath33 biexciton with @xmath85 and @xmath86 , whereas the lower bound state is a pure @xmath4 exciton with energy @xmath87 . on the contrary , when the center of mass of the spin - flip excitation is strongly localized there is a strong coupling between the exciton and the antisymmetric biexciton which leads to a strongly distorted dynamics , as shown below . of the @xmath0 exciton wavepacket for @xmath80 , @xmath65 and @xmath88 with ( dotted ) and without ( solid ) spin - changing collisions , for an initial gaussian wavepacket with a width of @xmath89 sites.,scaledwidth=35.0% ]
if the initial spin - flip is delocalized in a spatial region much larger than the inter - site spacing the momentum distribution is strongly peaked at @xmath77 .
hence after the spin - flip there is no significant transfer into @xmath33 biexcitons , since the lower bound state is basically a @xmath0 exciton . without spin - changing the exciton tunnels with hopping @xmath30 , and
hence at @xmath77 the effective exciton mass is @xmath90 . in spite of the absence of any significant biexciton admixture , spin - changing strongly modifies the effective exciton mass into : @xmath91 note that @xmath92 differs significantly from @xmath93 ( it may even invert its sign ) , radically modifying the exciton wavepacket dynamics , as illustrated in fig .
[ fig:4 ] .
note that if the created initial exciton has a sharply defined but finite center - of - mass momentum @xmath94 , then the @xmath0 exciton is not any more a single bound state but a linear superposition of both bound states ( and possibly a slight contribution of broken biexcitons ) . since the bound states have different group velocities of opposite sign , the dynamics after the spin - flip is hence characterized by the appearance of two wavepackets moving in opposite directions formed by the two different types of exciton - biexciton composites ( fig .
[ fig:5 ] , middle , which should be compared with the single wavepacket spreading for @xmath95 in fig .
[ fig:5 ] , top ) . for an initial gaussian wavepacket for the flip excitation with width of @xmath89 sites and central momentum @xmath95 ( top ) , and @xmath96 ( middle ) , and for a localized initial flip with width of @xmath78 site and @xmath95 ( bottom),title="fig:",scaledwidth=45.0% ] for an initial gaussian wavepacket for the flip excitation with width of @xmath89 sites and central momentum @xmath95 ( top ) , and @xmath96 ( middle ) , and for a localized initial flip with width of @xmath78 site and @xmath95 ( bottom),title="fig:",scaledwidth=45.0% ] for an initial gaussian wavepacket for the flip excitation with width of @xmath89 sites and central momentum @xmath95 ( top ) , and @xmath96 ( middle ) , and for a localized initial flip with width of @xmath78 site and @xmath95 ( bottom),title="fig:",scaledwidth=45.0% ]
the situation is markedly different if the initial spin - flip is strongly localized , where all @xmath45 in the brillouin zone contribute . as a consequence , the initial @xmath0 excitation results in an admixture of both bound states , which result , as shown in fig .
[ fig:5 ] ( bottom ) , to a completely different wavepacket dynamics as that for initially extended spin - flips ( fig .
[ fig:5 ] , top ) .
note that , in addition , a localized initial spin - flip may become considerably more unstable against unbinding , especially if the upper antisymmetric-@xmath0 bound state unbinds for a given range of @xmath45 values . in that case
a significant part of the initial exciton excitation would dissolve into unbound biexciton pairs .
up to this point we have considered the dynamics after single spin - flips .
although single spin flips may be created using state of the art techniques for single - site addressing @xcite , in typical experiments more than one spin - flip will be induced . as a result more than one
composite will be simultaneously created , hence opening the possibility of inelastic composite - composite collisions which may induce a decay channel for the bound composites into broken biexciton pairs .
the various initial spin - flip excitations result in a complicated many - body non - equilibrium dynamics for the created bound composites , which we have studied by means of time - dependent matrix - product - state calculations @xcite for up to @xmath97 spin - flips delocalized in an initial extension of @xmath89 sites ( for a total of @xmath98 sites ) .
excitons ( bottom ) for @xmath80 , @xmath99 and @xmath88 , for the case of @xmath78 ( dotted ) , @xmath100 ( dashed ) and @xmath97 ( solid ) spin flips delocalized in an initial square wavepacket @xmath89 sites wide.,scaledwidth=45.0% ] figure [ fig:6 ] shows our results for the case of @xmath101 and @xmath99 . for this particular case
the bound states are well below the sea of broken pairs .
hence , as discussed above , it is expected that single spin - flip excitations do not lead to a significant population of unbound solutions .
two main features can be observed in fig .
[ fig:6 ] . on one side ,
there are marked rabi like oscillations of the @xmath0-exciton population due to the presence of the two bound states .
the oscillation amplitude is small since , as discussed in the previous section , the lowest bound state is to a large extent the @xmath0-exciton due to the initial delocalization of the spin - flip .
note that the rabi oscillations are however damped due to dephasing , since different momentum components close to @xmath95 have a slightly different rabi frequency .
note that a more localized initial spin - flip would result in a stronger frequency beating , due to the participation of momenta over all the brillouin zone .
note also that the presence of additional spin - flips leads to an even stronger damping of the rabi oscillations . on the other side ,
note that composite - composite collisions basically leave unaffected the total population in the bound states , which for this particular case basically reduces to the population of @xmath0 excitons ( fig .
[ fig:1]a ) and unbroken @xmath33 biexcitons ( fig .
[ fig:1]b ) .
hence , deeply bound states may be created , which do not undergo significant inelastic unbinding due to inelastic collisions .
thus , multiple spin - flips lead to an out - of - equilibrium ( but highly metastable ) many - body state , which may understood as a basically stable gas of bound composites , opening interesting perspectives for quantum composite gases . for the particular case considered in fig .
[ fig:6 ] we expect a quantum gas of almost pure @xmath0-excitons with `` biexciton dressed '' dynamics .
note that although the @xmath0-excitons present an effective attractive nearest neighbor interaction ( @xmath102 ) , this interaction becomes irrelevant at low momenta @xmath103 compared to the infinite on - site repulsion due to pauli exclusion .
hence the low @xmath45 properties of 1d @xmath0 excitons will be as those of a tonks - girardeau gas .
in summary , spin flips in a polarized repulsive high - spin fermi gas may lead to the formation of novel types of bound composites , which we illustrated for the case of 1d spin-@xmath0 hard - core fermions in the mott phase . in that case
the composites are formed by a @xmath0 exciton - like excitation and an antisymmetric @xmath33 biexciton - like one .
intriguing dynamics and stability properties of the composites result from a non trivial interplay between spin - changing , qze and exciton momentum .
the stability of the exciton gas against inelastic interactions opens exciting possibilities for the creation of intricate novel quantum composite phases . | we show that the interplay between spin - changing collisions and quadratic zeeman coupling provides a novel mechanism for the formation of repulsively bound composites in high - spin fermions , which we illustrate by considering spin flips in an initially polarized hard - core 1d mott insulator of spin-@xmath0 fermions .
we show that after the flips the dynamics is characterized by the creation of two types of exciton - biexciton composites .
we analyze the conditions for the existence of these bound states , and discuss their intriguing properties . in particular we show that the effective mass and stability of the composites depends non - trivially on spin - changing collisions , on the quadratic zeeman effect and on the initial exciton localization .
finally , we show that the composites may remain stable against inelastic collisions , opening the possibility of novel quantum composite phases . |
the mechanism responsible for high - temperature superconductivity remains unresolved despite intense study over the past two decades .
the observation of pseudogaps and a relatively universal phase diagram are thought to represent key aspects of the solution , but no theory describes all observations in a simple , unified way .
previously we introduced an su(4 ) model of competing antiferromagnetism ( af ) and @xmath0-wave superconductivity ( sc ) , used coherent states to provide a connection to landau - ginzburg and symmetry constrained hartree - fock - bogoliubov theory , and demonstrated that su(4 ) symmetry implies no double occupancy .
@xcite this has defined a new , unconventional method to simplify the rather complex many - body problem in cuprates which is otherwise very difficult to solve .
the su(4 ) coherent state solution represents a generalization of the bcs pairing theory having cooper pairs with non - zero momentum and angular momentum .
dynamics of these interacting cooper pairs at different hole - doping and temperature yields rich phase diagrams that may explain the main features of observations .
the occurrence of a pseudogap , a critical value of doping , and splitting of the sc gap at low doping region are all natural consequences of the dynamics . in this article
we extend our previous discussion and compare gap and phase diagrams computed using the formalism developed in ref.@xcite with available data .
we demonstrate that our analytically - solvable gap equations with doping - independent interactions already describe the basic features of experimental gap and phase diagrams for the cuprates .
with addition of simple doping dependence for interaction strengths in the low - doping region , our results can reproduce quantitatively the superconducting gap , the pseudogap , and the doping variation of both the superconducting transition temperature @xmath1 and the pseudogap transition temperature @xmath2 extracted from a range of cuprate data .
in the su(4 ) model the configuration space is built from _ coherent pairs _ formed from two electrons ( or holes ) centered on adjacent lattice sites : spin - singlet @xmath3 and spin - triplet @xmath4 pairs @xcite .
the 15 generators of su(4 ) then consist of two operators for singlet pairs ( @xmath3 and @xmath5 ) , six operators for triplet pairs ( @xmath6 and @xmath7 ) , three staggered magnetization operators @xmath8 , three spin operators @xmath9 , and a number operator @xmath10 .
an su(4 ) hamiltonian restricted to one and two - body interactions can be determined uniquely , @xmath11 where @xmath12 , @xmath13 , and @xmath14 are effective strengths of @xmath0-wave singlet pairing , triplet pairing , and af correlations , respectively .
we shall assume the total spin to be zero and ignore the last term for the present discussion .
the su(4 ) coherent state @xcite is @xmath15 where @xmath16 is the physical vacuum and the unitary operator @xmath17 is a 4-dimensional matrix parameterized in terms of variational parameters @xmath18 and @xmath19 , with @xmath20 .
it implements a quasiparticle transformation on the @xmath3@xmath4 pair space that preserves su(4 ) symmetry ( implying no double occupancy @xcite ) .
the corresponding matrix elements for one- and two - body operators are discussed in ref .
@xcite and matrix elements for the operators appearing in eq .
( [ hsu4 ] ) are easily evaluated in the @xmath21 limit @xcite .
introducing the `` gaps '' @xmath22 representing correlation energies for singlet pairing , triplet pairing , and af , respectively , the variational hamiltonian @xmath23 ( with chemical potential @xmath24 ) is @xmath25 variation of @xmath26 with respect to @xmath27 and @xmath28 yields @xmath29 which has the solution @xmath30 e_{\pm}=[(\varepsilon_{\pm}-\lambda)^2+{{\delta_\pm}}^2]^\frac12 \qquad { \delta_\pm}={\delta_d}\pm{\delta_\pi}\end{array}\ ] ] where we define @xmath31 from eq .
( [ uv ] ) , the definitions for @xmath32 , @xmath33 , and @xmath34 , and a theoretical doping fraction @xmath35 with @xmath36 the electron number , we obtain the temperature - dependent gap equations @xmath37 where @xmath38 and @xmath39 the quasiparticle number densities @xmath40 are assumed to be @xcite @xmath41 where @xmath42 and @xmath43 if @xmath44 . the gaps and @xmath24 follow from the algebraic equations ( [ 1.24])([1.27 ] ) , the energy @xmath45 can be calculated from ( [ hprimeeq ] ) and @xmath46 and @xmath47 , @xmath27 , and @xmath28 are determined by eqs.([uv ] ) .
this defines a complete formalism permitting calculation of general observables .
these results are similar to the bcs theory but with important differences . here
we have two pairing energy gaps and two kinds of quasiparticles , implying complex behavior ( ultimately tracing to the non - abelian commutator algebra for the su(4 ) operators @xcite ) .
the quantities @xmath48 correspond to two sets of single - particle energies @xmath49 , split by @xmath50 in the af background .
each level can be occupied by one electron of spin up or down .
the corresponding pairing gaps are @xmath51 and the probabilities for single - particle levels to be unoccupied or unoccupied are @xmath52 , and @xmath53 , respectively .
when @xmath54 , eq .
( [ hsu4 ] ) reduces to a pairing hamiltonian , @xmath55 , and eqs .
( [ c222])([uv ] ) reduce to the usual bcs equations .
we now describe the solution of the gap equations ( [ 1.24])([1.27 ] ) , first for temperature @xmath44 and then for finite @xmath56 .
data suggest that in hole - doped cuprates the interactions in ( [ 1.24])([1.27 ] ) are attractive , with af correlation strongest and triplet pairing weakest . assuming that @xmath57
, we find the general @xmath44 solution @xmath58^{1/2}\label{udopdq } \\ { \delta_d}&=&\tfrac12 g_0\omega [ x ( x_q^{-1 } - x)]^{1/2 } \label{udopdd } \\
\delta_\pi&=&\tfrac12 g_0\omega [ x ( x_q - x)]^{1/2 } \label{udoppi } \\ { \lambda^\prime}&=&-\tfrac12 \chi\omega x_q ( 1-x_q x)-\tfrac12 g_1 \omega x , \label{udopl}\end{aligned}\ ] ] which exists only for @xmath59 less than a critical doping @xmath60 the dependence of @xmath61 on three elementary interaction strengths shows explicitly that the critical doping point results from competition between correlations .
a trivial @xmath62 ( pure singlet ) solution exists also in the entire physical doping range @xmath63 , @xmath64^{1/2 } \qquad { \lambda^\prime}=-\tfrac12 g_0\omega x. \label{odopl}\ ] ] trivial solutions valid for @xmath63 that correspond to pure triplet pairing , pure af states , and metallic states ( all @xmath65 ) may be found also , but insertion of @xmath66 , @xmath67 and @xmath68 into eq .
( [ hprimeeq ] ) indicates that the nontrivial solution is always the ground state where it exists ( @xmath69 ) , and that for @xmath70 eq .
( [ odopl ] ) gives the ground state .
thus , when @xmath71 the ground state is a pure singlet pairing state given by ( [ odopl ] ) , but when @xmath72 the solution ( [ udopdq])([udopl ] ) ( which differs from the singlet pairing state in permitting finite values for @xmath32 , @xmath33 , and @xmath34 ) is the ground state and ( [ odopl ] ) defines an excited state .
the critical doping point @xmath61 separates these two qualitatively distinct ground states and marks a quantum phase transition .
we now use these results to construct gap diagrams for cuprate superconductors .
the parameters @xmath12 , @xmath13 , and @xmath14 are effective interaction strengths within a truncated space . in fig.[fig : fig]a , we first make the simplest possible approximation and assume that they are constants , independent of doping .
four energy scales are shown : 1 .
@xmath66 , which measures af correlations ; 2 . the singlet pairing gap @xmath67 [ eq . ( [ udopdd ] ) ] , which is the superconducting gap for @xmath73 ; 3 . the singlet pairing gap @xmath74 [ eq.([odopl ] ) ] , which is the superconducting gap for @xmath71 but is not the ground - state order parameter for @xmath73 ; 4 . the triplet pairing gap @xmath68 .
the af gap @xmath66 is maximal at @xmath75 , decreases rapidly with doping , crosses the pairing gaps , and vanishes at the critical doping @xmath61 . like @xmath66 ,
the triplet gap @xmath68 exists only within the doping range @xmath76 ; it peaks at @xmath77 . for @xmath78
the singlet gap is the curve @xmath79 but below @xmath61 it splits into two curves ( @xmath67 and @xmath79 ) having different doping behavior .
also shown in fig .
[ fig : fig]a are gap data from refs .
@xcite that support the complex gap structure suggested by the su(4 ) quasiparticle solutions .
in particular , the predicted splitting of the singlet pairing gap and termination of the @xmath34 gap at a critical doping @xmath80 are consistent with the data points plotted .
( color online ) energy gap diagrams at @xmath44 ( a , c ) and phase diagrams ( b , d ) for hole - doped cuprates .
the doping rate @xmath81 is defined as @xmath82 with @xmath83 and @xmath84 being the number of lattice sites and the maximum allowed number of holes , respectively . in ( a ) and ( b ) constant interaction strengths are assumed : @xmath85 , @xmath86 , and @xmath87 , while in ( c ) and ( d ) doping - dependent strengths indicated by the insets to the figures are used . in all plots @xmath88 , except for the dotted line in (
d ) marked @xmath89 , which illustrates how the phase boundary marked @xmath90 shifts if @xmath91 .
the green dashed curve in ( b ) and ( d ) is the empirical @xmath1 curve : @xmath92 $ ] .
data in ( a ) and ( c ) are taken from refs .
@xcite , and in ( b ) and ( d ) from refs .
@xcite , respectively . note that data in ( b ) and ( d ) marked with the legend ( dai ) taken from ref .
@xcite and include data taken from refs . @xcite . ]
figure [ fig : fig]a suggests qualitative agreement between predicted and observed gap structure at @xmath44 but there are two quantitative discrepancies : ( 1 ) the slope of @xmath34 is too small at low doping , and ( 2 ) the onset of a pairing gap at zero doping in the calculation contradicts data suggesting that this onset should occur closer to 5% hole doping .
these discrepancies indicate that the constant coupling strength assumption gives unrealistic strong pairing and too weak af strength in the low - doping region .
they can be removed by allowing a simple doping dependence for the coupling strengths according to the following physical considerations .
first , the su(4 ) symmetry requires that the pure af state can exist only when @xmath93 , which suggests that pairing strength must have an onset at @xmath94 .
second , to reproduce the experimental nel temperature , @xmath14 has to be strongest at @xmath95 .
these suggest a doping dependence for @xmath12 , @xmath13 , and @xmath14 given by the inset to fig . [
fig : fig]c , which implies an onset of pairing strength at @xmath96 ( in the simplest picture we assume the same onset for singlet and triplet pairing , but with different magnitudes ) , and a large af correlation at half - filling that decreases as doping increases , but finally is stabilized as pairing is established . with that change , the more quantitative agreement with data shown in fig .
[ fig : fig]c results .
in figs . [ fig : fig]a and [ fig : fig]c all energies are scaled by @xmath97 .
the solution ( [ udopdd ] ) indicates that @xmath67 depends only on @xmath61 , once @xmath12 is fixed .
if the critical doping @xmath61 has a universal value ( for example , see ref .
@xcite ) , the doping dependence for @xmath67 also should be universal if gaps are scaled by the maximum @xmath1 for each compound , since the maximum @xmath1 is proportional to @xmath12 ( see ref .
it is well known that cuprate sc pairing gaps exhibit such scaling .
the emergence of the critical doping @xmath61 and the splitting of the singlet gap for @xmath98 are a direct consequence of competing pairing and af correlations below @xmath61 . when doping is low , af correlations dominate pairing and a state with large af gap and suppressed pairing is the ground state .
therefore , the superconducting gap for the ground state is small ( @xmath99 ) and the large pairing gap @xmath79 is associated with an excited state when @xmath98 .
however , as doping increases the su(4 ) symmetry implies that pairing correlations decrease less quickly than the af correlations ( see ref .
@xcite ) and they eventually dominate .
the energy is then minimized by larger pairing and diminished af correlation .
the critical point @xmath61 is the hole - doping fraction where the ground - state af correlations vanish completely in the variational solution .
the competing - order picture @xcite assumes that the pseudogap ( pg ) is an energy scale for order competing with superconductivity that vanishes at a critical doping point where the competition is completely suppressed . from figs . 1a and 1c ,
@xmath66 ( an order parameter of the af phase ) has precisely these properties .
but the af operators are generators of su(4 ) , so @xmath66 _ also _ is the stabilization energy for a mixture of preformed singlet and triplet su(4 ) pairs that condense into a strong superconducting state only after af and triplet pairing fluctuations are suppressed by hole doping .
this represents a non - abelian generalization of the standard phase - fluctuation model @xcite for preformed pairs .
thus , the su(4 ) pseudogap state results from competing af and sc order expressed in a basis of singlet and triplet fermion pairs , which may be viewed as a unification of the competing order and preformed pair pictures
triplet pairs are an essential component of the su(4 ) many - body wavefunction ( for example , the su(4 ) algebra , which imposes the no double occupancy condition on the lattice , does not close in the absence of triplet pair operators )
. however , fig .
[ fig : fig ] indicates that the triplet pair correlation energy is small in the underdoped region and zero for doping larger than @xmath100 .
the primary role of triplet pairs in the hole - doped cuprates appears to lie in fluctuations mediating the af sc competition .
in figs . [ fig : fig]b and [ fig : fig]d we show phase diagrams resulting from finite - temperature calculations . figure
[ fig : fig]b assumes the coupling strengths of fig.[fig : fig]a and fig . [ fig : fig]d assumes the coupling strengths of fig . [ fig : fig]c , so the only adjustable parameter in either case is @xmath101 . the empirical factor @xmath102 [ see eq .
( [ density ] ) ] is introduced because actual single - particle energies due to thermal excitation are non - degenerate and realistic quasiparticle excitation should be easier than in our degenerate approximation .
there are four distinct phases in figs .
[ fig : fig]b and [ fig : fig]d : 1 .
an antiferromagnetic phase ( labeled af ) .
2 . a superconducting phase ( labeled sc ) .
3 . a transitional phase with all three correlations present ( labeled af+sc ) .
4 . a metallic phase .
the correlations involved in each phase are indicated in parentheses in fig .
[ fig : fig]b and the doping - dependent temperatures @xmath1 , @xmath2 , and @xmath90 ( or @xmath89 ) mark the phase boundaries .
data from refs .
@xcite are compared with the predicted su(4 ) phase structure in figs.[fig : fig]b and [ fig : fig]d .
the constant coupling approximation ( fig .
[ fig : fig]b ) reproduces data qualitatively , but the predicted superconductivity extends too low in doping and the upper boundary on the af phase does not rise steeply enough at low doping ( leading to a nel temperature a factor of 23 too low ) .
these discrepancies are removed in fig.[fig : fig]d , which allows the evolution of coupling with doping indicated in the inset .
it is commonly believed that in the cuprate phase diagram there is a boundary ( for example , the one labeled @xmath103 in ref.@xcite ) distinguishing an af long - range ordered state from a disordered one .
we do not discuss this possibility here but note that the potential to describe such a boundary is contained in the present model .
if ( contrary to our minimal assumptions here ) we permit onset of the triplet pairing strength at somewhat lower doping than for the singlet pairing , a new boundary will appear in the very - low doping region ( @xmath104 0.020.05 ) that separates a pure af phase from the pg region .
there are additional ( generally higher - temperature ) data that have been related to the pseudogap discussion beyond that displayed in fig .
1 . their physical interpretation is generally unsettled .
our theoretical @xmath2 is clearly defined as the temperature at which the pg in the su(4 ) model goes to zero .
once the interaction strengths are fitted to the gap data , there are no free parameters to modify the theoretical @xmath2 curve .
thus , any reproducible data that lie substantially above our predicted @xmath2 curve may be associated either with physics that goes beyond the minimal model that is described here , or with a definition for @xmath2 that is inconsistent with ours .
the results of fig . [ fig : fig ] are supported by various additional data .
we mention a few representative examples .
nmr measurements @xcite using strong magnetic fields to suppress sc in bi@xmath105sr@xmath106la@xmath107cuo@xmath108 concluded that the pg coexists with sc and terminates in a critical point near @xmath109 .
figure [ fig : fig ] could accommodate @xmath110 with small parameter changes , but also we note that magnetic fields could increase @xmath100 by altering the sc af competition . from specific heat , nmr , and transport data ,
tallon et al @xcite conclude that high-@xmath1 phase behavior is universal , with a critical doping @xmath111 where the pg vanishes and where large changes in quantities like the superfluid density indicate crossover from weak to strong superconductivity , and that this transition implies vanishing short - range magnetic order .
a key role for magnetic correlations in pg states also is suggested by the effect of impurities on @xmath112-axis optical conductivity @xcite .
in - plane resistivity measurements @xcite indicate that the phase diagram is surprisingly universal , with a pg terminating near optimal doping and different electronic states either side of the termination .
these measurements also indicate a crossover above @xmath1 between two kinds of pg behavior ( see also ref .
@xcite ) , which may be explicable in terms of the finite but decreasing pair fluctuations expected in the pg region , as we now discuss . figure [ fig : fig ] may explain the vortex - like nernst signal @xcite observed above @xmath1 .
the singlet pair gap vanishes there but the many - body su(4 ) wavefunction has finite pair content between the curves @xmath2 and @xmath1 that decreases with increasing @xmath56 and decreasing doping .
thus , contours for pair fluctuations above @xmath1 will be similar to observed contours for nernst signal strength , but a meissner effect is expected only below @xmath1 .
recently , ong et al @xcite concluded that a consistent explanation of pseudogap and nernst data requires pg and sc pairing states that are distinct but related by symmetry , as proposed here .
more detailed properties of fig .
[ fig : fig ] may be investigated in future work .
for example , the transition between the af+sc and sc phases is 2nd - order if @xmath100 lies at higher doping than the point _
b _ ( solid curve @xmath90 in fig.[fig : fig]d ) , but is 1st - order if it lies at lower doping ( dotted curve @xmath113 , starting from the point _ a _ ) .
this implies a wealth of testable consequences for the gap and phase structure depending both on global properties and on microscopic details .
in summary , finite - temperature su(4 ) coherent states have been used to construct a rich , universal gap and phase structure for hole - doped cuprates that may be expressed in a bcs - like formalism with two kinds of quasiparticles and competing superconducting and antiferromagnetic order .
we find a pseudogap of antiferromagnetic character terminating at a critical doping @xmath114 that is distinct from the superconducting state but related to it by a non - abelian symmetry .
the corresponding pseudogap states may be interpreted in terms of either sc af competition or preformed su(4 ) pairs that condense into a singlet @xmath0-wave superconductor as hole doping suppresses fluctuations .
our results represent a minimal variational solution of competing antiferromagnetism and @xmath0-wave superconductivity on a fermionic lattice with no double occupancy .
therefore , we believe that the general gap and phase structure presented here will be a necessary consequence of any realistic theory that takes a doped mott insulator with competing @xmath115wave pairing and antiferromagnetism as the basis for describing high - temperature superconductivity . | we use a symmetry - constrained variational procedure to construct a generalization of bcs to include cooper pairs with non - zero momentum and angular momentum .
the resulting gap equations are solved at zero and finite temperature , and the doping - dependent solutions are used to construct gap and phase diagrams .
we find a pseudogap terminating at a critical doping that may be interpreted in terms of both competing order and preformed pairs .
the strong similarity between observation and predicted gap and phase structure suggests that this approach may provide a unified description of the complex structure observed for cuprate superconductors . |
the reionization of the universe was one of the paramount milestones in the transition from the cosmic dark ages , the period before the formation of the first stars , to the present day ( see e.g. barkana & loeb 2001 ; ciardi & ferrara 2005 ) . as it is the first generations of luminous objects that were the sources of the high energy radiation which reionized the intergalactic medium ( igm ) , theoretical studies of how reionization took place ( e.g. gnedin & ostriker 1997 ; miralda - escud ' e et al .
2000 ; iliev et al .
2006 ; mcquinn et al .
2007 ; shin et al . 2008 ) in conjunction with observations of the radiation emitted during the epoch of reionization ( e.g. fan et al .
2006 ; robertson et al . 2010
; bowman & rogers 2010 ) stand to teach us a great deal about the nature of the first stars and galaxies ( see e.g. bromm & yoshida 2011 ) .
one of the outstanding questions pertaining to reionization is what were the sources of the ionizing radiation which effected it .
it appears unlikely that the progenitors of the galaxies observed to date at @xmath6 @xmath1 7 are responsible for reionizing the universe , as the escape fraction of ionizing photons from these galaxies would have to be very large or , otherwise , a low - metallicity stellar population or one with a relatively top - heavy mass function would be required ( see e.g. ouchi et al .
2009 ; bunker et al . 2010 ; wilkins et al . 2011 ; but see also wyithe et al . 2009 ; bouwens et al . 2011 ) . however , there are a number of possible sources which likely contributed to reionization at higher redshifts , where observations of the cosmic microwave background ( cmb ) suggest a large fraction of the igm was reionized ( komatsu et al .
2009 ) . among these
are stellar sources , including possibly massive population ( pop ) iii stars ( e.g. ciardi et al .
2003 ; yoshida et al .
2004 ) ; quasars powered by accretion onto black holes ( e.g. ricotti & ostiker 2004 ; alvarez & abel 2007 ; volonteri & gnedin 2009 ) ; early populations of high - mass x - ray binaries ( e.g. power et al .
2009 ; mirabel et al .
2011 ) ; high - velocity structure formation shocks ( miniati et al . 2004
; dopita et al . 2011 ) ; and dark matter decay and annihilation ( e.g. schleicher et al . 2008 ) .
one additional source of hydrogen ionizations which inevitably accompanies stellar sources are the supernovae ( sne ) which mark the end of the life of massive stars .
the impact that sne have on reionization is at least two - fold : not only are ionizing photons generated in sne explosions ( e.g. chevalier 1974 ; shull & silk 1979 ) , but the collisional ionization and evacuation of gas they cause lower the optical depth to ionizing radiation , thereby allowing for a higher escape fraction of ionizing photons into the igm ( e.g. clarke & oey 2002 ; yajima et al .
2009 ; paardekooper et al . 2011
; see also tegmark et al .
1993 ) . at the high redshifts ( @xmath6 @xmath5 20 ) at which the first stars formed , due to the high energy density of cmb photons ( which is @xmath7 ( 1+@xmath6)@xmath8 ) x - ray photons produced via compton scattering of the cmb inside sne remnants are likely to have contributed to the reionizization of the igm , as described by oh ( 2001 ) .
however , over the full redshift range over which reionization takes place ( i.e. @xmath6 @xmath1 6 ) sne generate ionizing photons predominantly through other processes , such as bremsstrahlung and resonance and recombination line emission ( e.g. cox 1972 ; chevalier 1974 ; shull & mckee 1979 ) ; indeed , these processes can also be dominant even for some pop iii sne in minihaloes at @xmath6 @xmath5 20 ( see kitayama & yoshida 2005 ; greif et al .
2007 ; whalen et al . 2008 ) .
while , in any case , the number of ionizing photons produced in sne shocks is in general well below the number produced during the pre - sn phases of stellar evolution , because they are on average more energetic than the photons emitted from stars , the cross section for their absorption by atomic hydrogen and helium is smaller ; this , in turn , allows for a higher fraction of sn shock - generated photons to escape into the igm and contribute to reionization . in the present work , we calculate the magnitude of this effect and derive an estimate of the contribution , relative to stellar sources , that ionizing photons generated by sne make to cosmic reionization . in the next section , we estimate the number of ionizing photons generated in sne shocks using recently published high velocity shock models . in section 3 , we calculate the average cross section for their absorption by neutral hydrogen and helium , and we use this to estimate the escape fraction of sne shock - generated ionizing photons as a function of the escape fractions of stellar photons found in the detailed numerical simulations of gnedin et al .
( 2008 ) and razoumov & sommer - larsen ( 2010 ) . in turn , we draw on these results to estimate the relative contributions of stars and sne to reionization in section 4 .
though of less overall importance , in section 5 we furthermore provide an upper limit for the relative contribution of cosmic rays generated in sne shocks to reionization . in section 6
we briefly discuss the implications of our findings for the population of galaxies responsible for reionization .
finally , we discuss our results and give our concluding remarks in section 7 .
we begin by estimating the rates at which hydrogen atoms are ionized due to both sne and stars , as a function of the star formation rate . in the following sections , we calculate the average cross section for the absorption of these photons by neutral hydrogen and helium , and we use these quantities to compare the relative contributions of sn - generated photons and stellar photons to the reionization of the igm .
to estimate the rate at which hydrogen atoms are ionized due to the ionizing radiation emitted from sn remnants , we first estimate the total energy emitted in ionizing photons from sn shocks during the early sedov - taylor phase of their evolution .
we then estimate the average energy of the sn shock - generated ionizing photons , and use this to estimate the rate at which atomic hydrogen is ionized , as a function of the star formation rate . in order to estimate the rate at which ionizing photons are produced in sn shocks we make use of the mappings iii high velocity shock models presented in allen et al .
we use an approximation of the convenient formula ( their equation 5 ) provided by these authors to calculate the flux @xmath9 of hydrogen - ionizing radiation , mostly composed of thermal bremsstrahlung and recombination and resonance lines , that is emitted from shocks with velocities 10@xmath10 km s@xmath11 @xmath12 @xmath13 @xmath12 10@xmath14 km s@xmath11 passing through a medium with a particle number density 0.01 @xmath15 @xmath12 @xmath16 @xmath12 10@xmath10 @xmath15 : @xmath17 next , we make the assumption , justified below in section 2.1.2 , that ionizing photons are produced by sne shocks while in the sedov - taylor phase of their expansion ( see e.g. truelove & mckee 1999 ) . using the sedov - taylor solution , we thus find the following expression for the the distance @xmath18 that the shock has traveled since the initial sn blast , as a function of its velocity @xmath13 : @xmath19 using equations ( 1 ) and ( 2 ) , we integrate over the whole of the spherical blast wave and over time , to find the total energy @xmath20 emitted in ionizing radiation : @xmath21 where in the second part of the equation we have expressed the integral in terms of @xmath13 using equation ( 2 ) . the lower limit of integration is set by the approximate minimum shock velocity ( @xmath5 200 km s@xmath11 ) for which ionizing radiation may propagate beyond the shock , assuming a largely neutral medium ahead of the shock ( see e.g. shull & mckee 1979 ; allen et al .
2008 ) ; for lower velocities , the ionizing photons produced in the shock are all absorbed by neutral gas as it is swept up by the shock .
if the ambient medium is instead largely ionized , then ionizing radiation produced in somewhat lower velocity shocks will propagate beyond the shock as well ; thus , our assumption of a neutral ambient medium is a conservative one ( see e.g. shull & silk 1979 ) .
the upper limit of integration in equation ( 3 ) is set by the maximum velocity for which the shock models of allen et al .
( 2008 ) are given ; however , for this reason , too , our estimate of the energy emitted in ionizing radiation is conservative . also , interestingly
, there is no dependence on the density @xmath16 of the ambient gas in equation ( 3 ) .
rather , it is simply that @xmath5 40 percent of the energy of the sn is emitted as ionizing radiation , which is roughly consistent with previous estimates ( see e.g. mansfield & salpeter 1974 ; chevalier 1977 ) , as well as being consistent with the amount of energy that is radiated during the sedov - taylor phase ( see e.g. draine & woods 1991 ; shu 1992 ) , the extent of which we shall turn to discuss in the next section . before moving on
, we must note that the shock models presented in allen et al .
( 2008 ) are , strictly speaking , valid in the steady - state approximation , which may be a poor description of the early stages of the sedov - taylor phase . nonetheless , we find agreement with previous works in part because the majority of both the total energy radiated in ionizing photons and the total number of ionizing photons ( see section 2.1.3 ) are emitted at the late stages of the sedov - taylor phase , when the steady state approximation is more robust .
furthermore , we note that even the gas shock - heated to very high temperatures @xmath22 10@xmath23 k at early times will eventually cool radiatively , albeit on timescales of up to @xmath5 10@xmath24 yr , considerably longer than the extent of the sedov - taylor phase ( see e.g. cioffi et al .
1988 ) . therefore , while it may take upwards of a million years , the post - shock gas will cool and the time - integrated spectrum of the radiation emitted from the sne remnants should converge , at least approximately , to the result that we have found under the assumption of steady - state ( radiative ) shocks . as this timescale is still shorter than the @xmath5 10@xmath23 yr timescale on which most massive stars emit ionizing photons and explode as sne , our use of steady - state shock models should not greatly impact our results in the end .
the radius @xmath25 at which a sn shock transitions from the sedov - taylor phase to the radiative phase depends on the rate at which the shocked gas cools ( see e.g. shull & silk 1979 ; blondin et al .
we shall draw on the results of draine & woods ( 1991 ) , who model the cooling of sn remnants .
these authors find the following fitting formula for @xmath25 , for the case of a gas with solar metallicity ( see their tables 1 & 2 ) : @xmath26 for gas with a metallicity different than solar , the time @xmath27 at which the transition to the radiative phase occurs can be estimated as the cooling time of the shocked gas , and is thus inversely proportional to the cooling rate @xmath28 of the gas . in turn , this affects the radius of the transition , as in the energy conserving phase @xmath25 @xmath7 @xmath27@xmath29 @xmath7 @xmath28@xmath30 ( see e.g. blondin et al .
1998 ) . to express @xmath25 as a function of metallicity , we can thus follow blondin et al .
( 1998 ) , and estimate the cooling function of the gas in the energy conserving phase , as a function of metallicity @xmath31 , by fitting to the cooling functions presented in sutherland & dopita ( 1993 ) .
we thus find an approximate fit to the cooling function as @xmath32 where @xmath33 is the temperature of the gas .
this fit is valid over the temperature range @xmath5 10@xmath34 k @xmath35 @xmath33 @xmath35 10@xmath24 k , in which the transition to the radiative phase takes place for @xmath16 @xmath35 10@xmath10 @xmath15 ( e.g. draine & woods 1991 ) , and over the metallicity range 0.05 @xmath36 @xmath35 @xmath31 @xmath35 @xmath36 , which is the same as that over which we model the stellar ionizing photon output below in section 2.2 . using the scaling of @xmath28 with metallicity
, we follow equations ( 4.15 ) and ( 4.16 ) of draine & woods ( 1991 ) to find the following expression for @xmath25 as a function of metallicity ( for @xmath16 @xmath35 10@xmath10 @xmath15 ) : @xmath37 thus , there is a very weak dependence of the transition radius on metallicity , and a somewhat stronger dependence on the density of the gas into which the sn explodes . following equation ( 2 ) , in the sedov - taylor phase the distance @xmath18
to which the shock has propagated , as a function of shock velocity @xmath13 , is @xmath38 thus , the radius at which the velocity drops to 200 km s@xmath11 , the minimum shock velocity for which we assume ionizing photons can escape the shock front ( see section 2.1.1 ) , is in general smaller than the radius at which the sedov - taylor phase ends ; therefore , our assumption in the previous section that the sedov - taylor phase lasts until @xmath13 = 200 km s@xmath11 is sound in most cases .
we note , however , that for @xmath39 @xmath1 8 @xmath40 10@xmath41 erg and/or @xmath16 @xmath1 40 @xmath15 , @xmath18 when @xmath13 @xmath4 200 km s@xmath11 will exceed @xmath25 and our simple treatment of the dynamical evolution of the sn remnant will have to be modified to account for ionizing photon production after the sedov - taylor phase .
we would like to emphasize that , while the sedov - taylor solution does well to describe the _ dynamical _ evolution of a sn remnant at early times under the assumption of conservation of mechanical energy , it is in fact the case that a portion of this energy is radiated away during this phase .
indeed , the sedov - taylor phase ends only when a substantial portion of the energy of the sn has been radiated away , with the result that the conservation of mechanical energy ceases to be an appropriate approximation for describing the dynamical evolution of the remnant , as has been found in myriad studies ( see e.g. chevalier 1974 ; draine & woods 1991 ; blondin et al .
1998 ; truelove & mckee 1999 ; shu 2002 ) .
as noted in section 2.1.1 , the amount of energy that we have found to be lost to ionizing radiation , especially in the late stages of the sedov - taylor phase , is in good agreement with these previous works .
we estimate the total number of ionizing photons emitted per sn using the tabulated ionization parameters given in table 3 of allen et al .
carrying out an integration similar to that shown above , approximated as a sum over shock velocities of @xmath13 = 200 to 10@xmath14 km s@xmath11 , we estimate the total number @xmath42 of ionizing photons emitted per sn to be @xmath43 dividing @xmath20 by @xmath42 we find the average energy of ionizing photons to be @xmath44 = 51 ev , which implies that the majority of the energy of the ionizing radiation is carried by photons with energies @xmath1 40 ev .
as shown by shull & van steenberg ( 1985 ) , ionizing photons with such high energies may ionize more than one hydrogen atom per photoionization , due to secondary ionizations caused by collisions with photoelectrons . in particular , these authors show that the fraction of the energy of such ionizing photons that go into ionizing hydrogen is @xmath45 @xmath5 0.3 .
we thus estimate the total number of hydrogen atoms that can be ionized by the radiation emitted from a single sn by taking it that 30 percent of the energy in ionizing radiation @xmath20 goes into ionizating hydrogen atoms , each of which requires @xmath46 = 13.6 ev . from this
, we find that the total number of ionizations per sn is @xmath45@xmath20/@xmath46 @xmath4 5.2 @xmath40 10@xmath47(@xmath39/10@xmath41 erg ) . in turn
, we can express the rate @xmath48 at which hydrogen ionizations take place for a given star formation rate @xmath49 and a stellar initial mass function ( imf ) for which @xmath50 supernovae are produced per unit mass in stars , as @xmath51 we have normalized to @xmath50 = 10@xmath52 m@xmath53 , which is approximately the value for a standard salpeter - like imf ( e.g. leitherer et al . 1999 ) .
we note that equation ( 9 ) takes into account only type ii sne produced by stars with masses @xmath1 9 m@xmath54 , and not type ia sne ; this is reasonable , however , because the number of type ia sne , which typically occur with a significant delay following a star formation episode , is in general much smaller than that of type ii sne ( see e.g. nagashima et al .
2005 ) .
next , we estimate the rate @xmath55 at which ionizing photons are produced by stars , as a function of @xmath49 , assuming a given number of ionizing photons @xmath56 emitted per baryon in stars .
normalizing @xmath56 to a value appropriate for pop ii stars formed with a standard salpeter - like imf ( e.g. haiman 2009 ) , we have @xmath57 the quantity @xmath56 is metallicity and imf dependent , and can , for instance , be upwards of 10 times higher than the fiducial value we have chosen here for a population of extremely low- or zero - metallicity stars ( e.g. bromm et al .
2001 ; schaerer 2002 ; wise & cen 2009 ) . in practice
, we shall use the values of @xmath55 given for a range of metallicities in the starburst99 database ( leitherer et al .
1999 ) ; specifically , we shall use the models from these authors for stellar populations formed with a salpeter imf , with minimum and maximum initial stellar masses of 1 and 100 m@xmath54 .
these choices will allow to facilitate a direct comparison with radiative transfer simulations from which the escape fractions of ionizing photons are calculated , as described in the next section .
comparing @xmath48 to @xmath55 , we see that the rate at which hydrogen atoms are ionized due to radiation from stars is in general more than an order of magnitude higher than that due to the emission of ionizing radiation from sn shocks .
nonetheless , the latter can contribute significantly to reionization due to a higher fraction of these photons escaping from galaxies , which we now turn to show .
in this section we first calculate the average photoionization cross section for sne shock - generated photons , and we then apply this to attain an estimate of the escape fraction of such photons as compared to that of stellar ionizing photons . in order to estimate the average cross section for photoionization of sn shock - generated photons , we make use of the high velocity shock spectra presented in allen et al .
as shown in section 2.1.3 , the average energy of ionizing photons from sne shocks is high enough that we need to account for multiple hydrogen ionizations per photon , as we have done in calculating @xmath48 .
as the number of ionizations is thus dependent on the total energy in ionizing radiation , it is the fraction of the energy in ionizing photons which escapes the host galaxy which determines the impact of sn shock - generated radiation on the reionization of the igm .
therefore , we estimate the cross section @xmath58 for photoionization as the _ photon energy_-averaged cross section : @xmath59 f(e_{\gamma } ) de_{\gamma}}{\int_{{\rm 13.6 ev}}^{\infty } f(e_{\gamma } ) de_{\gamma } } \mbox{\ , } \ ] ] where the spectrum is described by @xmath60(@xmath61 ) , the flux per unit photon energy .
we use the photoionization absorption cross sections for neutral hydrogen and helium provided by osterbrock & ferland ( 2006 ) .
in particular , for photon energies @xmath61 @xmath62 13.6 ev we take the cross section for absorption by neutral hydrogen to be @xmath63 = 6.3 @xmath40 10@xmath64 @xmath65 ( @xmath61/13.6 ev)@xmath66 , whereas for photon energies above 24.6 ev we compute the total cross section for absorption as the sum of the neutral hydrogen and neutral helium cross sections for absorption , the latter taken to be @xmath67 = 7.8 @xmath40 10@xmath64 @xmath65 ( @xmath61/24.6 ev)@xmath66 .
we have assumed a neutral helium to hydrogen ratio of 0.1 , such that our results are obtained as the average cross section to absorption per neutral hydrogen atom , which is a convenient form for the calculation of the escape fraction of sn shock - generated photons which follows .
we note that this assumption implies a negligible amount of absorption of photons by helium which is already once - ionized ( i.e. in the form of he ii ) .
given the relatively soft spectra of the stellar sources , the fraction of helium in the form of he ii is likely to be low ; given furthermore the low photoionization cross section of he ii compared to that of he i ( e.g. osterbrock & ferland 2006 ) , this is a conservative approximation . the average cross section for photoionization that we obtain following this procedure is @xmath58 @xmath5 0.7 @xmath40 10@xmath64 @xmath65 , for the solar metallicity shock spectra presented in allen et al .
as these authors note , the strength of the ionizing field , being largely generated by brehmsstrahlung , is only weakly dependent on the metallicity of the gas ; therefore , we shall use this single value for our calculations for simplicity , although we will express our results in terms of @xmath58 where possible , such that it is clear how our results will change for different average cross sections . next , in order to estimate the escape fraction of sn shock - generated photons relative to that of stellar photons
, we also need to calculate the average cross section @xmath68 for photoionization of stellar ionizing photons . for this
we adopt the model spectra presented in leitherer et al .
in particular , we use the spectra for continuous star formation at metallicities of @xmath31 = 0.05 , 0.2 , 0.4 , and 1 @xmath36 , and a salpeter imf over a mass range from 1 to 100 m@xmath54 .
we carry out the same exercise as described above for the calculation of @xmath58 , but we now instead calculate the
_ photon number_-averaged cross section for photoionization .
this is appropriate in this case , because for the softer spectra of ionizing radiation emitted from stars a single ionizing photon will on average lead to the ionization of a single hydrogen atom .
we thus calculate the cross section as @xmath69 \frac{f(e_{\gamma})}{e_{\gamma } } de_{\gamma}}{\int_{{\rm 13.6 ev}}^{\infty } \frac{f(e_{\gamma})}{e_{\gamma } } de_{\gamma } } \mbox{\ , } \ ] ] for each of the stellar spectra we consider , the values we find for @xmath68 range from 2.6 @xmath40 10@xmath64 cm@xmath10 to 4.8 @xmath40 10@xmath64 cm@xmath10 , as shown in table 1 .
due to the softer ionizing spectra of stellar populations at higher metallicity , for higher values of the metallicity the average cross section is found to be higher as well .
of particular use to us will be the average cross section for a stellar metallicity of @xmath31 = 0.05 @xmath36 , as this is the same population of stars used in the radiative transfer simulations which we shall draw on in the next section to estimate the escape fractions of sn shock - generated ionizing photons ( gnedin et al . 2008 ; razoumov & sommer - larsen 2010 ; see also yajima et al .
2011 ) . in order to compare the escape fractions of stellar and sn shock - generated ionizing photons
, we first estimate their ratio along a given line of sight out of a galaxy into the igm .
we then use this result , in conjunction with the results of detailed radiative transfer simulations , to estimate the overall , line of sight - averaged escape fractions of ionizing photons . to begin , we note that the escape fraction of ionizing photons along a given line of sight out of a galaxy can be approximately expressed as of e@xmath71 weighted by the source spectrum , is that we must extract the probability distribution function of the hydrogen column density d@xmath72/d@xmath0 from the escape fractions reported in gnedin et al .
( 2008 ) and razoumov & sommer - larsen ( 2010 ) , in order to obtain the relative escape fraction of sne ionizing photons .
this is made possible only by using the approximate formula given by equation ( 13 ) . ]
@xmath73 where @xmath74 is the optical depth along the line of sight , which is given in terms of the column density @xmath0 of neutral hydrogen and the average cross section @xmath75 for the absorption of ionizing photons by neutral hydrogen and helium .
assuming that the location of a population of stars in the galaxy is the same as that of the sne that mark their deaths ( i.e. that the spatial extent of radiating sne remnants is sufficiently small ) and that the column density @xmath0 along a given line of sight out of the galaxy does not change significantly between the time when a population of stars is shining and when they explode as sn ( i.e. within @xmath35 3 @xmath40 10@xmath23 yr ; e.g. leitherer et al .
1999 ) , the column density will be the roughly same for both cases . assuming that it is in fact the same , we have for the ratio of the escape fractions of sne and stellar ionizing photons @xmath76 where @xmath77 = @xmath0@xmath68 and @xmath78 = @xmath0@xmath58 . in turn , using once again @xmath79 = e@xmath80 , we solve for the escape fraction of sn photons : @xmath81 where for the final expression we have used the values of @xmath58 and @xmath68 found in section 3.1 for the case of star formation at a constant rate with a salpeter mass function from 1 - 100 m@xmath54 and a metallicity of @xmath31 = 0.05 @xmath36 , as presented in leitherer et al . ( 1999 ) .
we can see immediately from equation ( 15 ) that , especially for low escape fractions of stellar photons along a given line of sight , the escape fraction of sne photons can be much larger than that of stellar photons . for example , using the last expression in equation ( 15 ) , for @xmath79 = 0.1 we obtain @xmath82 = 0.54 , and for @xmath79 = 0.01 we obtain @xmath82 = 0.29 . taking the ratio of the number of ionizations caused by sne and stars found in section 2 ( @xmath48/@xmath55 @xmath5 0.02 ) , we see that more reionizations due to sne than to stellar photons will take place in the igm ( i.e. that @xmath79 @xmath55 @xmath35 @xmath82 @xmath48 ) when the escape fraction of stellar photons is @xmath79 @xmath35 0.005 ; this corresponds to a hydrogen column density along the line of sight out of a galaxy of @xmath0 @xmath4 10@xmath2 @xmath3 . for column densities greater than this ,
the contribution to reionization from sne is greater than that from stars .
while in deriving equation ( 15 ) we have assumed both a constant star formation rate and no evolution in the column density along a given line of sight over a timescale of @xmath35 3 @xmath40 10@xmath23 yr , roughly the time between the emission of the bulk of the ionizing photons from stars and the emission from sne remnants , in general these assumptions may lead us to underestimate the escape fraction of sne - generated photons .
the reason for this can be seen by considering a single stellar population formed instantaneously . in this case
, the high energy photons emitted by stars will act to heat and ionize the gas along a line of sight out of the galaxy , and the effect of this stellar feedback will be to increase the escape fraction of ionizing photons with time ( see e.g. johnson et al .
2009 ; wise & cen 2009 ) . because this feedback occurs _ before _ the ionizing photons from sne are emitted , it is likely that in general the column density along a line of sight out of a galaxy from a given cluster of stars is lower when sne occur than when the bulk of the ionizing photons from stars are emitted .
therefore , it is likely the case that the escape fractions of sn - generated photons are larger relative to the time - averaged stellar photon escape fractions than we have found here . to better estimate the relative numbers of hydrogen ionizations caused by sne and stars , we turn to the results of detailed radiative transfer simulations from which the escape fraction of stellar photons has been calculated . as our result for the ratio of escape fractions
is defined only along a given line of sight , we must use the results of numerical simulations which are likewise given along a representative collection of lines of sight . to this end
, we use the probability distributions for the escape fractions of stellar ionizing photons given in figure 2 of gnedin et al .
( 2008 ) and figure 9 of razoumov & sommer - larsen ( 2010 ) . as we would like to generalize our results for the various values of @xmath68 given in table 1
, we shall convert the probability distributions of escape fractions which are given by these authors to the probability distributions of neutral hydrogen column densities @xmath0 . to do this
, we first note that from equation ( 13 ) we have @xmath0 = -ln(@xmath79)/@xmath68 , where @xmath79 is the stellar photon escape fraction derived from the simulations and @xmath68 is the average cross section for photoionization that we have calculated for the same stellar spectra used in these simulations , the @xmath31 = 0.05 @xmath36 model from leitherer et al .
( 1999 ) . using this ,
we next convert from the probability distribution @xmath83/@xmath84 of stellar ionizing photons to the probability distribution of neutral hydrogen column densities @xmath83/@xmath85 , as follows : @xmath86 the resulting probability distributions @xmath83/@xmath85 that we find for the simulations presented in gnedin et al .
( 2008 ) and razoumov & sommer - larsen ( 2010 ) are shown in figure 1 .
these probability distributions contain the basic information that we would like to glean from these detailed hydrodynamics simulations . using these
, we can next apply different cross sections @xmath68 and @xmath58 to infer the probability distributions of the escape fractions of stellar and sn shock - generated ionizing photons , and from this the overall escape fractions averaged over all lines of sight .
we note that in our calculation of @xmath0 we have neglected the absorption of ionizing photons by dust ; however , as gnedin et al .
( 2008 ) show , this is a relatively minor effect and hence should not impact our results greatly ( see also yajima et al .
2011 ) . in order to directly compare the probability distributions of the two escape fractions @xmath87 and @xmath88
, we must convert from that for stellar photons @xmath83/@xmath89 given in gnedin et al .
( 2008 ) and razoumov & sommer - larsen ( 2010 ) to that for sne photons .
for the stellar spectra used in these works , the @xmath31 = 0.05 @xmath36 stellar population for which we have calculated the average cross section to photoionization , this is done as follows : @xmath90 where we have used equation ( 15 ) relating @xmath88 to @xmath87 . with this conversion from the stellar to the sne photon
escape fraction probability distribution , we can directly compare the two distributions , which we show together in figure 2 .
as the figure shows , there is a higher probability for larger escape fractions of sn shock - generated photons than for stellar photons . in figure 3
, we show the probability distributions of @xmath79 and @xmath82 for the case of a stellar population at solar metallicity , using the appropriate @xmath68 from table 1 , along with the corresponding transformation to the probability distribution @xmath83/@xmath89 for this average cross section . comparing the two figures it is evident that the probability distributions for @xmath88 and @xmath79 are much more different in the case of the solar metallicity stellar population than in the case of the lower metallicity stellar population .
this is simply due to the larger difference between the cross sections @xmath68 and @xmath58 for the more metal rich population ; because the spectra of ionizing radiation from more metal - rich stars is softer , the average cross section for photoionization is higher , and so the escape fraction of stellar ionizing photons is lower relative to the escape fraction of sn shock - generated photons .
cccccc & & & & & + metallicity & @xmath55 & @xmath50 & @xmath68 & ratio of escape fractions : & relative contribution of sne : + @xmath91@xmath36@xmath92 $ ] & @xmath9110@xmath93 s@xmath11@xmath92 $ ] & @xmath9110@xmath52 m@xmath53@xmath92 $ ] & @xmath9110@xmath64 @xmath65@xmath92 $ ] & @xmath94/@xmath95 & @xmath94@xmath48/@xmath95@xmath55 + & & & & + & & & & & + 1 & 2.2 & 1.91 & 4.8 & 6.8 & 0.12 + 0.4 & 2.5 & 1.86 & 2.9 & 4.2 & 0.062 + 0.2 & 2.8 & 1.91 & 2.8 & 4.1 & 0.056 + 0.05 & 3.2 & 1.95 & 2.6 & 4.0 & 0.048 + & & & & + using our results for the probability distributions of both @xmath87 and @xmath88 we calculate the overall escape fractions , averaged over all lines of sight , of sne and stellar photons . for the case of sne photons , this calculation is done as follows : @xmath96 where d@xmath72/d@xmath88 is given by equation ( 17 ) .
a similar calculation is carried out for the line of sight - averaged escape fraction of stellar photons @xmath95 , with the integral performed over @xmath87 instead of @xmath88 .
the ratios of the line of sight averaged escape fractions that we find for the stellar populations that we consider are similar for both of the simulations that we have drawn upon , and we present these ratios in the fifth column of table 1 . as shown there , depending on the metallicity of the stellar population , we find that the values of @xmath88/ @xmath87 vary between @xmath4 4 and @xmath4 7 .
we emphasize that , especially for the results of gnedin et al .
( 2008 ) , the sne photon escape fraction is significantly lower than what would be found by simply applying equation ( 15 ) to the overall stellar escape fraction directly .
this is likely due in part to the fact that the simulation results show relatively large probabilities for escape fractions of either @xmath79 @xmath97 1 or @xmath79 @xmath97 0 ; because these limiting values will yield the same escape fractions for sn generated photons , via equation ( 15 ) , the overall line of sight - averaged escape fractions of stellar and sne photons are not as different as they are along individual lines of sight for which 0 @xmath98 @xmath87 @xmath98 1 . finally , we note that our results are senstively dependent on the probability distribution functions that we have adopted from gnedin et al .
( 2008 ) and razoumov & sommer - larsen ( 2010 ) .
while the detailed simulations which these authors carry out in attaining these results are some of the most sophisticated and complete to date ( including radiative feedback from young stars and both mechanical and chemical feedback from sne ) , the resolution of the simulations is limited and the structure of the interstellar medium at sub - resolution scales will have an impact on the actual escape fractions of ionizing photons . despite this uncertainty which plagues all escape fraction calculations , the results presented by these authors allow us to make a first estimate of the relative role of sne in cosmic reionization .
future simulations at higher resolution will ideally include the ionizing radiation from both stars and sne explicitly , thereby allowing for improved estimates of the relative role of each of these sources .
working from our results from sections 2 and 3 , we may now compare the relative contributions of stars and sne to reionization .
the ratio of the number of hydrogen ionizations in the igm due to sne to that due to stars is given by @xmath99@xmath100/@xmath95@xmath101 .
using equations ( 9 ) and ( 10 ) , along with the values of @xmath55 and @xmath50 from leitherer et al .
( 1999 ) and the ratios of line of sight - averaged escape fractions , each given in table 1 , we estimate this ratio as @xmath102 the escape fractions appearing here are the overall line of sight - averaged values , normalized to typical values . in particular , we have normalized the ratio of escape fractions to @xmath94/@xmath95 = 4 , roughly the ratio found in the last section for a stellar population at sub - solar metallicity .
the right - most column of table 1 shows the values of this ratio for all of the stellar populations we consider .
the values that we find for the contribution of sne to reionization , relative to the contribution from stars , ranges from @xmath4 0.05 to @xmath4 0.12 , with higher relative contributions for the more metal - rich stellar populations . as discussed in section 3 , this is due largely to the higher ratios of escape fractions @xmath99/@xmath95 for the more metal - enriched populations , owing to their softer stellar spectra .
however , the different rates of sn production @xmath50 and stellar ionizing photon generation @xmath55 also have an effect .
importantly , we note that the relative contributions to reionization scale linearly with the energy per supernova @xmath39 ; while we have normalized our results to @xmath39 = 10@xmath41 erg , if sne are more energetic than this in the early universe , then the relative contribution of sne to reionization may be considerably higher .
finally , we note that our calculations of the quantities appearing in table 1 have been done drawing on numerous approximations .
we have nonetheless chosen to report the values of these quantities , in cases , to two or three decimal places , in order to highlight the differences between the stellar models at different metallicities that we have considered .
an additional process driven by sne which leads to ionization of hydrogen is the production of cosmic rays , and for completeness here we consider their contribution to reionization .
we may find an upper limit to the total number of ionizations that cosmic rays can cause in the igm , by making the simple assumption that all the energy injected into cosmic rays by sne is deposited into the igm via ionizations of hydrogen . following the discussion of cosmic ray production in the early universe given by stacy & bromm ( 2007 )
, we shall assume that 10 percent of the energy @xmath39 of sne goes into cosmic rays ( e.g. ruderman 1974 ) ; then , noting that each ionization removes 50 ev of energy from the impinging cosmic ray ( spitzer & tomasko 1968 ) , we find the maximum number of ionizations that can be produced per sn to be @xmath5 10@xmath47 .
relating this to the star formation rate , in order to compare the maximum rate ( assuming an escape fraction of cosmic rays of @xmath103 = 1 ) of ionizations @xmath104 by cosmic rays to the rate of ionizations due to stellar and sn shock - generated photons , we have @xmath105 we emphasize that this is an extremely hard upper limit , as in general the igm is optically thin to cosmic rays , especially to those with higher energies . to illustrate this
, we can make a simple conversion of the result presented in figure 1 of stacy & bromm ( 2007 ) to find the penetration depth @xmath106 of a cosmic ray with energy @xmath107 traveling through a completely neutral general igm at a redshift @xmath6 : @xmath108 where we have applied a simple power - law fit appropriate for low cosmic ray energies ( i.e. @xmath107 @xmath35 10@xmath109 ev ) , and the distance is in physical units . comparing this to the physical distance that a cosmic ray moving at close to the speed of light would travel from its generation at @xmath6 @xmath35 20 in a sn explosion to the end of reionization at @xmath6 @xmath5 6 , which is @xmath35 200 mpc , we see that only cosmic rays with @xmath107 @xmath35 10@xmath23 ev will have time to deposit a large portion of their energy into ionizing the igm .
although the energy distribution of cosmic rays in the early universe is not known , it is likely that the bulk of the total energy in cosmic rays generated before reionization is carried by cosmic rays at higher energies that this ( e.g. jasche et al .
2007 ; stacy & bromm 2007 ) . therefore , comparing @xmath104 to @xmath55 and @xmath48 , we conclude that the contribution to reionization from cosmic rays generated in sne is likely to be well below that from either stellar sources or from the radiation generated in sne shocks .
in the previous sections we have estimated the relative contributions of sne and stars to reionization .
we now consider the implications of our results for the population of galaxies responsible for reionization .
in particular , we highlight how constraints on the luminosity function of reionizing galaxies are strengthened by accounting for the contribution of sne . following wilkins et al .
( 2011 ) , we assume that the number of ionizing photons produced by a galaxy of luminosity @xmath110 at 1600 @xmath111 is @xmath112 , where the units of @xmath113 is erg s@xmath11 @xmath111@xmath11 . to facilitate a comparison with this work , we use for the fraction of ionizing photons @xmath114 .
this is appropriate for a stellar population with metallicity @xmath115 , with stars having formed with a salpeter imf at a constant rate for 100 myr ( leitherer et al .
1999 ) , as these authors consider .
the rate of ionizing photon production per unit comoving volume that is required to maintain reionization is given by ( madau et al .
1999 ; wilkins 2011 ) : @xmath116 here @xmath117 is the clumping factor of the igm . comparing the above expression for @xmath118 to the rate of photon production by both stars and sne , taking into account our result from table 1 that @xmath119 , we integrate over the galaxy luminosity function to find the following for the ratio of clumping factor to escape fraction required for the universe to remain reionized : @xmath120 for the luminosity function @xmath121 we use the @xmath122 best - fit values reported in wilkins et al .
given the uncertainty in the faint - end slope of the luminosity function ( e.g. ryan et al .
2007 ; khochfar et al .
2007 ) , we predict the ratio of @xmath123 for three different choices of faint - end slopes ( @xmath124 ) .
typical values for the clumping factor and escape fraction are @xmath125 ( e.g. pawlik et al . 2009 ) and @xmath126 ( e.g. ciardi & ferrara 2005 ) , respectively , which yields @xmath127 .
as shown in figure 4 , such high values are only reached if the luminosity function of galaxies is integrated down to a limiting magnitude of @xmath128 and if the faint - end slope of the luminosity function is as steep as @xmath129 . in the case of only stellar photons
the limiting magnitude is fainter by several magnitudes , showing that the contribution to reionization from stars in low luminous galaxies is comparable to the one from sne in massive galaxies .
finally , we note that recent work suggests that the star formation rates of high redshift galaxies are not constant , as assumed in the calculation of the ionization parameter @xmath130 used above , but instead are increasing with time ( e.g. finlator et al .
2011 ; khochfar & silk 2011 ) . in this event
, @xmath130 should be considerably higher ( wilkins et al .
2011 ) , and thus the contribution from low luminosity galaxies may be even lower than we find here .
we have used state of the art models of stellar populations and high velocity shocks , along with the results of detailed radiative transfer simulations , to evaluate the contribution of sne to the reionization of the universe .
we have found that , although the number of ionizing photons emitted from sn shocks is well below that emitted from stars , due to the harder spectrum of the ionizing radiation from sn shocks the fraction of such photons which escape galaxies is in general much larger than the escape fraction of stellar ionizing photons . related to this , owing principally to the softer spectrum of stars of higher metallicity , we find that the relative contribution of sne to reionization is metallicity dependent ; in particular , for stellar metallicities between 0.05 @xmath36 @xmath35 @xmath31 @xmath35 @xmath36 , sne are responsible for between @xmath4 5 and @xmath4 12 percent of the number of ionizations caused by stellar sources . also , while these numbers are for line of sight - averaged escape fractions , we have furthermore found that , along a given line of sight out of a galaxy from a star - forming region , for hydrogen column densities @xmath0 @xmath1 10@xmath2 @xmath3 the contribution to reionization from sne is larger than that from stars .
finally , we have shown that the contribution from sne allows reionization to be completed by galaxies a few magnitudes brighter than in the case in which only stellar sources contribute .
we emphasize that we have found a conservative estimate of the contribution from sne , for the following reasons .
firstly , we have included only the ionizing radiation emitted from sn shocks in the sedov - taylor phase ; while we find the radiated energy to already be @xmath4 40 percent of the total sn energy at this stage , there may be an additional contribution of ionizing photons from sne in the later , radiative phase if the upstream medium is already ionized by other sources ( shull & silk 1979 ; see also section 2.1.1 here ) .
secondly , in our modeling we have not accounted for the fact that the majority of sn shock - generated photons are emitted after the stellar ionizing radiation from a given population of stars ; as the radiation from the stars acts to ionize and rarify the gas along lines of sight out of a galaxy , when sne explode the column density of neutral hydrogen @xmath0 will likely be lower than the average column density seen by stellar photons .
we also note that , as we find that the number of ionizations caused by sn scales linearly with sn energy , if this is higher or lower than our fiducial value of @xmath39 = 10@xmath41 erg during the epoch of reionization , then sne may have contributed more or less , respectively , to reionization than we have found here . related to this , if the stellar imf is more top - heavy during reionization than it is in the galaxy today , as some recent studies would suggest ( e.g. van dokkum 2008 ; dav ' e 2008 ; gunawardhana et al .
2011 ) , then this may also lead to an enhancement in the contribution from sne , as in this case a higher fraction of stars may explode as sne and these sne may also be more energetic . beyond contributing to the ionization of hydrogen in the igm ,
the enhanced photoheating rate due to the contribution of ionizing photons from sne likely has additional effects .
for instance , a higher heating rate of the gas contributes to the suppression of low - mass galaxy formation in reionized regions of the universe ( e.g. dijkstra et al .
2004 ; hambrick et al .
it is also likely to lead to a reduction in the clumping factor @xmath117 of the igm , which has bearing on the rate at which reionization occurred ; however , we note that for heating rates comparable to what would be expected for sn - generated ionizations pawlik et al .
( 2009 ) find only a small decrease in the clumping factor .
while we have found sne to be of secondary importance compared to stars as contributors to reionization , they undeniably provide an enhancement in the rate at which reionization occurs above what is found in models accounting for only stellar sources ( e.g. sokasian et al .
2003 ; iliev et al .
2006 ; choudhury & ferrara 2007 ; trac & cen 2007 ; raicevi ' c et al . 2011 ) . in this , accounting for the effect of sne can aid in explaining how the universe was reionized by @xmath6 @xmath5 6 .
using the results we have presented in table 1 , it is straightforward to include the contribution of sne in reionization calculations that already include the contribution from stars .
the authors are thankful to jan - pieter paardekooper for many constructive comments on an early draft of this paper , as well as to stephen wilkins , umberto maio , fabrice durier , volker gaibler , bhaskar agarwal , eyal neistein , garrelt mellema , dan whalen and philipp podsiadlowski for helpful discussions .
we would also like to acknowledge constructive feedback from an anonymous reviewer . 1 allen , m. g. , groves , b. a. , dopita , m. a. , sutherland , r. s. , kewley , l. j. 2008 , apjs , 178 , 20 alvarez , m. , abel , t. 2007 , mnras , 380 , 30 barkana , r. , loeb , a. 2001 , phr , 349 , 125 blondin , j. m. , wright , e. b. , borkowski , k. j. , reynolds , s. p. 1998 , apj , 500 , 342 bouwens , r. j. , et al .
2011 , apj , submitted ( arxiv:1105.2038 ) bowman , j. d. , rogers , a. e. e. 2010 , nat , 468 , 796 bromm , v. , kudritzki , r. p. , loeb , a. 2001 , apj , 552 , 464 bromm , v. , yoshida , n. 2011 , ara&a , in press ( arxiv:1102.4638 ) bunker , a. j. , et al .
2010 , mnras , 409 , 855 chevalier , r. a. 1974 , 188 , 501 chevalier , r. a. 1977 , ara&a , 15,175 choudhury , t. r. , ferrara , a. 2007 , mnras , 380 , l6 ciardi , b. , ferrara , a. 2005 , ssrv , 116 , 625 ciardi , b. , ferrara , a. , white , s. d. m. 2003 , mnras , 344 , 7 cioffi , d. f. , mckee , c. f. , bertschinger , e. 1988 , apj , 334 , 252 clarke , c. , oey , m. s. 2002 , mnras , 337 , 1299 cox , d. p. 1972 ,
apj , 178 , 159 dav ' e , r. 2008 , mnras , 385 , 147 dijkstra , m. , haiman , z. , rees , m. j. , weinberg , d. h. 2004 , apj , 601 , 666 dopita , m. a. , krauss , l. m. , sutherland , r. s. , kobayashi , c. , lineweaver , c. h. 2011 , ap&ss , in press ( arxiv:1106.5546 ) draine , b. t. , woods , d. t. 1991 , apj , 383 , 621 fan , x. , carilli , c. l. , keating , b. 2006 , ara&a , 44 , 415 finlator , k. , oppenheimer , b. d. , dav ' e , r. 2011 , mnras , 410 , 1703 gnedin , n. y. , ostriker , j. p. 1997 , apj , 486 , 581 gnedin , n. y. , kravtsov , a. v. , chen , h .- w .
2008 , apj , 672 , 765 greif , t. h. , johnson , j. l. , bromm , v. , klessen , r. s. 2007 , apj , 670 , 1 gunawardhana , m. l. p. , et al .
2011 , mnras , accepted ( arxiv:1104.2379 ) haiman z. 2009 , astrophysics in the next decade : jwst and concurrent facilities , astrophysics & space science library , eds .
h. thronson , a. tielens , m. stiavelli ( arxiv:0809.3926 ) hambrick , d. c. , ostriker , j. p. , johansson , p. h. , naab , t. 2010 , mnras , submitted ( arxiv:1009.6005 ) iliev , i. t. , mellema , g. , pen , u .- l . , merz , h. , shapiro , p. r. , alvarez , m. a. 2006 , mnras , 369 , 1625 jasche , j. , ciardi , b. , enlin , t. a. 2007 , mnras , 380 , 417 johnson , j. l. , greif , t. h. , bromm , v. , klessen , r. s. ippolito , j. 2009 , mnras,399 , 37 khochfar , s. , silk , j. 2011 , mnras , 410 , l42 khochfar , s. , silk , j. , windhorst , r. a. , ryan , r. e. 2007 , apj , 668 , l115 kitayama , t. , yoshida , n. 2005 , apj , 630 , 675 komatsu , e. , et al .
2009 , apjs , 180 , 330 leitherer , c. , et al . 1999 ,
apjs , 123 , 3 madau , p. , haardt , f. , rees , m. j. 1999 , apj , 514 , 648 mansfield , v. n. , salpeter , e. e. 1974 , apj , 190 , 305 mckee , c. f. , tan , j. c. 2008 , apj , 681 , 771 mcquinn , m. , lidz , a. , zahn , o. , dutta , s. , hernquist , l. , zaldarriaga , m. 2007 , mnras , 377 , 1043 miniati , f. , ferrara , a. , white , s. d. m. , bianchi , s. 2004 , mnras , 348 , 964 mirabel , i. f. , dijkstra , m. , laurent , p. , loeb , a. , pritchard , j. r. 2011 , a&a , 528 , 149 miralda - escud ' e , j. , haehnelt , m. , rees , m. j. 2000 , apj , 530 , 1 nagashima , m. , lacey , c. g. , baugh , c. m. , frenk , c. s. , cole , s. 2005 , mnras , 358 , 1247 oh , s. p. 2001 , apj , 553 , 499 osterbrock , d. e. , ferland , g. j. 2006 , astrophysics of gaseous nebulae and active galactic nuclei , university science ouchi , m. , et al .
2009 , 706 , 1136 paardekooper , j .-
p . , pelupessy , f. i. , altay , g. , kruip , c. j. h. 2011 , a&a , submitted pawlik , a. h. , schaye , j. , van scherpenzeel , e. 2009 , mnras , 394 , 1812 power , c. , wynn , g. a. , combet , c. , wilkinson , m. i. 2009 , mnras , 395 , 1146 raicevi ' c , m. , theuns , t. , lacey , c. , 2011 , mnras , 410 , 775 razoumov , a. o. , sommer - larsen , j. 2010 , apj , 710 , 1239 ricotti , m. , ostriker , j. p. 2004 ,
mnras , 352 , 547 robertson , b. e. , ellis , r. s. , dunlop , j. s. , mclure , r. j. , stark , d. p. 2010
, nat , 468 , 49 ruderman , m. a. 1974 , sci , 184 , 1079 ryan , r. e. , et al .
2007 , apj , 668 , 839 schaerer , d. 2002 , a&a , 382 , 28 schleicher , d. r. g. , banerjee , r. , klessen , r. s. 2008 , phrvd , 78 , 3005 shin , m .- s . , trac , h. , cen , r. 2008 , apj , 681 , 756 shu , f. h. 1992 , the physics of astrophysics , university science shull , j. m. , mckee , c. f. 1979 , apj , 227 , 131 shull , j. m. , silk , j. 1979 , apj , 234 , 427 shull , j. m. , van steenberg , m. e. 1985 , apj , 298 , 268 sokasian , a. , abel , t. , hernquist , l. , springel , v. 2003 , mnras , 344 , 607 spitzer , l. , tomasko , m. g. 1968 , apj , 152 , 971 stacy , a. , bromm , v. 2007 , mnras , 382 , 229 sutherland , r. s. , dopita , m. a. 1993 , apjs , 88 , 253 tegmark , m. , silk , j. , evrard , a. 1993 , apj , 417 , 54 trac , h. , cen , r. apj , 671 , 1 truelove , j. k. , mckee , c. f. 1999 , apjs , 120 , 299 van dokkum , p. g. 2008 , apj , 674 , 29 volonteri , m. , gnedin , n. y. 2009 , apj , 703 , 2113 whalen , d. , van veelen , b. , oshea , b. w. , norman , m. l. 2008 , apj , 682 , 49 wilkins , s. m. , bunker , a. j. , lorenzoni , s. , caruana , j. 2011 , mnras , 411 , 23 wise , j. h. , cen , r. 2009 , apj , 693 , 984 wyithe , j. s. b. , hopkins , a. m. , kistler , m. d. , y " uksel , h. , beacom , j. f. 2010 , mnras , 401 , 2561 yajima , h. , choi , j .- h . , nagamine , k. 2011 , mnras , 412 , 411 yajima , h. , umemura , m. , mori , m. , nakamoto , t. 2009 , mnras , 398 , 715 yamada , m. , nishi , r. 1998 , apj , 505 , 148 yoshida , n. , bromm , v. , hernquist , l. 2004 , apj , 605 , 579
here we justify the expression that we have used for the fraction of ionizing photons which escape along a single line of sight , equation ( 13 ) .
the exact expression for the escape fraction of ionizing photons along a given line of sight with column density of hydrogen @xmath0 is given by where for concreteness we have chosen to treat the case of stellar photons ; a similar treatment to that given here can easily be given for the case of sne photons . in order to show that equation ( 13 ) is approximately equal to this expression , we expand the exponential in the numerator , which yields | while stars are widely discussed as the source of the high energy photons which reionized the universe , an additional source of ionizing photons that must also contribute to reionization in this scenario is the supernovae ( sne ) which mark the end of the life of massive stars .
here we estimate the relative contributions of sne and stars to reionization .
while the rate at which ionizing photons are produced in sn shocks is well below that at which they are produced by stars , the harder spectra of radiation emitted from sne leads to an enhanced escape fraction of sn generated photons relative to that of stellar photons . in particular , along a given line of sight out of a galaxy , we find that for neutral hydrogen column densities @xmath0 @xmath1 10@xmath2 @xmath3 the contribution to reionization from sne is greater than that from stars .
drawing on the results of simulations presented in the literature , we find that the overall ( line of sight - averaged ) sne shock - generated ionizing photon escape fraction is larger than the stellar photon escape fraction by a factor of @xmath4 4 to @xmath4 7 , depending on the metallicity of the stellar population .
overall , our results suggest that the effect of sne is an enhancement of up to @xmath5 10 percent in the fraction of hydrogen reionized by stellar sources .
we briefly discuss the implications of our results for the population of galaxies responsible for reionization . |
it has been observed that people addicted to watching streaming videos constitute more than half of the internet traffic . with the introduction of smartphones
, mobile networks are witnessing an exponential traffic growth every year .
this leads to scenario where internet and wireless networks are pushed to operate close to their performance limits , dictated by current architectural considerations .
though much effort has been expended and in turn significant progress has been made in recent years to increase the capacity of mobile networks , there is little progress on dealing with the user satisfaction , which is strongly related to the quality of experience ( qoe ) .
this is the big challenge that the operators face today because they have to look at both the server side and the client side to make a link between the quality of service ( qos ) of the network and the client satisfaction which depends on the qoe .
empirical studies in @xcite has identified critical metrics that affect the qoe through the user engagement : * starvation probability . denoting the probability that a streaming user sees frozen images . *
average bit - rate . denoting the mean video quality over the entire session .
* bit - rate stability . describing the jittering of video quality during the entire session .
* start - up delay . denoting the waiting duration between the time that the user requests streaming service and the time that media player starts to play .
paper @xcite pointed out that the buffering ratio is most critical across genres .
for example , 1% increase in buffering reduces 3 minutes for a 90-minutes live video streaming .
@xcite showed that the total time spent rebuffering and the frequency of rebuffering events have substantial impact on qoe . under this context ,
media servers and network operators face a crucial challenge on how to avoid the degradation of user perceived media quality based on these metrics . however , development of new models as function of these metrics can help operators and content publishers to better invest their network and server resources toward optimizing these metrics that really matter for qoe .
in this paper we focus on a setting in which a video is streamed over a wireless network which is subject to a lot of constraints like bandwidth limitation and rate fluctuations due to the frequent changes of channel states and mobility @xcite .
indeed , time varying network capacity is especially relevant when considering wireless networks where such variations can be caused by fast fading and slow fading due to shadowing , dynamic interference , and changing loads . to addresse this issue ,
we focus on performance modelling and analysis of a streaming video with long range dependence ( lrd ) traffic and variable service capacity . due to the inherent difficulty and complexity of modelling fractal - like lrd traffic , we assume that the arrival of packets at player buffer are characterized by a markov modulated fluid model , which accurately approximates the traffic exhibiting lrd behaviour and mimics the real behaviour of multimedia traffic with short - term and long - term correlation @xcite . in comparison to related works ,
our whole analysis is on transient regime .
we construct sets of partial differential equations ( pdes ) to derive the starvation probability generating function using the external environment , which is described by the continuous time markov chain ( ctmc ) .
this approach predicts the starvation probability as function of the file size as well as the prefetching threshold .
moreover we provide relevant results to understand on how the starvation probabilities are impacted by the variation of traffic load and prefetching threshold .
we do simulations to show the accuracy of our model using ns-3 . achieving this goal , we are able to identify through our model the dependencies between quality metrics . for example
, start - up delay can reduce rebuffing ratio .
similarly bitrate rate switching can reduce buffering . with the results developed in this work , we are able to answer the fundamental questions : how many frames should the media player prefetch to optimize the users quality of experience ? from what file size the adaptive coding is relevant to avoid the starvation ?
how bit - rate switching impacts the qoe metrics ?
knowing these answers enables the user to maximize his qoe realising the tradeoffs among different metrics incorporating user preferences on rebuffering ratio , start - up delay and quality @xcite .
we further introduce an optimization problem which takes these key factors in order to achieve the optimal tradeoff between them .
we adopt a more flexible method by defining an objective of qoe by associated a weight for each metric based on user preferences @xcite .
qoe analysis over wireless networks has been studied for many years . in @xcite , authors study the qoe in a shared fast - fading channel using an analytical framework based on takacs ballot theorem .
they use a gi / d/1 queue to model the system , so they assume that the arrival process is independent and identically distributed ( i.i.d ) . in @xcite ,
the analysis of buffer starvation using m / m/1 queue is performed .
they use a recursive method to compare the results with the ballot theorem method even if the recursive method did not offer explicit results .
they assume an i.i.d arrival process that is a rough model of streaming services over the wireless networks .
since the performance measures depend on the autocorrelation structure of the traffic , a consensus exists about the limitation of the poisson process to model the traffic behaviour . in @xcite , authors develop an analytical framework to investigate the impact of network dynamics on the user perceived video quality , they model the playback buffer by a g / g/1 queue and use the diffusion approximation method to compute the qoe .
the qoe of streaming from the perspective of the network flow dynamics is studied in @xcite .
the throughput of a tagged user is governed by the number of the other users in the network .
this study shows that the network flow dynamics is the fundamental reason for playback starvation . +
the rest of this paper is organized as follows : in section [ model ] , we describe the system model while section [ qsm ] presents the analysis of the queuing system model . section [ paq ] describes the performance analysis of the quality of experience and section [ 2mmpp ] presents explicit results for two states mmfm .
section [ na ] shows numerical results and section [ clc ] concludes this paper .
we consider a single user receiving a media file with finite size @xmath0 in streaming .
generally , media files are divided into blocks of frames . when a user makes a request the server segments this media into frames and transfers them to the user through the network ( wired and wireless links ) .
when frames traverse the internet , their arrivals are not deterministic due to the dynamics of the available bandwidth .
one of the main characteristics of wireless traffic and internet traffic in general is the rate fluctuation caused by fast fading and slow fading due to shadowing , dynamic interference , and changing load .
moreover , data packet arrivals in cellular networks are found to be correlated over both short and long - time scales .
this is generally due to the arrival of packets bursts of comparable size , often leading to high instantaneous arrival rates .
hence , video flows through the internet with fluctuating speed . in this paper
, we assume that frames arrive to the play - out buffer with a rate that can take values from finite set @xmath1 .
the rate of arrival frames is governed by a continuous - time markov chain ( ctmc ) @xmath2 with infinitesimal generator @xmath3 .
@xmath4 where @xmath5 .
the maximum buffer size is assumed to be large enough so that the whole file can be stored . at the user side ,
incoming frames are stocked in a buffer and from there they are played with a rate @xmath6 ( e.g. , 25 frames per second ( fps ) ) in the tv and movie - making business .
we quantify the user perceived media quality using two measures called start - up delay and starvation .
there is ongoing research on mapping these two measures on standard human evaluated qoe measures .
as explained earlier , the media player wants to avoid the starvation by prefetching packets .
however , this action might incur a long waiting time . in what follows
, we reveal the relationship between the start - up delay and the starvation behaviour , with the consideration of the file size .
we consider a fluid model that has been proven to be a powerful modeling paradigm in many applications and relevant to capture the key characteristics that determine the performance of networks .
let @xmath7 denote the effective input rate in state @xmath8 .
hence the matrix of the effective rates is @xmath9 , which is a diagonal matrix @xmath10 .
we denote by @xmath11 the length of playout buffer of playback at time t. let @xmath12 be the first time the buffer is empty before reaching the end of the file , i.e. , @xmath13 and @xmath14 be the start - up delay where @xmath15 is the prefetching threshold . in the next section we provide mathematical analysis to compute the distribution of the number of starvation and start - up delay for a general bursty arrival process .
we compute the laplace transform of the probability of starvation given the continuous - time markov chain @xmath16 . we define @xmath17
to be the probability of starvation in state @xmath18 before time @xmath19 , given the initial state @xmath8 and the initial queue length @xmath20 .
@xmath21 for @xmath22 and @xmath23 .
it is clear that the ctmc can not be in a state @xmath18 at time @xmath24 if @xmath25 . hence @xmath26
@xmath27 @xmath28 let @xmath29 be the steady state probability vector of the ctmc @xmath30 where @xmath31 is the probability to be in the state @xmath32 at the stationary regime .
the expected input and output rates are @xmath33 and @xmath34 respectively .
the buffer is stable if @xmath35 .
conditioning on the first transition from the state @xmath32 at time @xmath36 we have @xmath37 taking the limit @xmath38 and after some algebraic simplification we obtain the following partial differential equation @xmath39 with the initial conditions @xmath40
@xmath41 @xmath42 where @xmath43 $ ] . + the laplace stieljes transform ( lst ) of @xmath17 is @xmath44\end{aligned}\ ] ] for @xmath45 and @xmath46 $ ] . taking the lst of equation ( [ pde ] ) and using the fact that @xmath47 for all @xmath48 , we find @xmath49 for a fixed value of @xmath50 , we take @xmath51 as a solution to equation ( [ ode ] ) . substituting in ( [ ode ] ) we get @xmath52 where the scalar @xmath53 and the vector @xmath54 are to be determined .
the theorem @xmath55 from @xcite gives @xmath56 where the coefficients @xmath57 are obtained by solving @xmath58 @xmath59 are the roots with negative real parts of @xmath60 and @xmath61 are the corresponding eigenvectors satisfying the equation @xmath62 we consider the previous system during the prefetching process and we denote by @xmath63 the length playout buffer of playback at time @xmath64 .
let @xmath65 be the first time that the length playout buffer reaches @xmath20 .
@xmath66 is the time that the system will take to accumulate @xmath20 content in the buffer .
this distribution is difficult to solve directly , so we resort to the following duality problem : + * duality problem : * _ what is the starvation probability by time @xmath19 if the queue is depleted with rate @xmath67 and the duration of prefetching contents is @xmath20 ? _
+ this duality problem allows us to compute the prefetching delay as a probability of starvation .
we define @xmath68 to be the probability of starvation before time @xmath19 at the state @xmath18 , conditioning on the initial state @xmath8 and the initial prefetching content @xmath20 , i.e. , the start - up threshold .
@xmath69 for @xmath22 and @xmath23 .
conditioning on the first transition from the state at time @xmath36 , @xmath70 taking the limit @xmath71 and after some algebraic simplification we obtain the following partial differential equation @xmath72 with the same initial conditions as in section [ first_passage ] , where @xmath73 and @xmath74 $ ] .
+ @xmath75 where the coefficients @xmath57 are obtained by solving @xmath76 and @xmath59 are the roots of @xmath77 and @xmath61 are the corresponding eigenvectors . in what follows , we compute the probability that the prefetching ends at a given state @xcite .
for this purpose , we define @xmath78 to be the probability that the prefetching ends at state @xmath79 given the initial state i and the initial queue length @xmath80 where @xmath15 is the prefetching threshold . in the time interval
@xmath81 $ ] , conditioning on the first transition from the state at time @xmath36 , we have @xmath82 after some algebraic simplification and letting @xmath83 yields the differential equation @xmath84 with the boundary condition @xmath85 let @xmath86 eq .
( [ oded2 ] ) becomes : @xmath87 .
@xmath88 is given by @xmath89 using eq.([sol ] ) and the initial conditions , we get @xmath90 where @xmath91 , @xmath92 is the diagonal matrix containing all the eigenvalues of @xmath93 and @xmath94 is an invertible matrix . in the previous sections we derived explicit expressions for the laplace - stieltjes transform of the probability of starvation and the start - up delay . in this section ,
we present theoretical models to find the corresponding probability of starvation and start - up delay .
the laplace stieljes transform of @xmath17 is @xmath95\\ & = & \int_{0}^{\infty}{e^{-\omega t}dh_{ij}(x , t ) } = \int_{0}^{\infty}{e^{-\omega t}h_{ij}(x , t)}\end{aligned}\ ] ] where @xmath96 is the probability density function of @xmath97 . given the laplace transform @xmath98 , the function value @xmath99 can be recovered from the contour integral @xmath100 where @xmath101 is a real number to the right of all singularities of @xmath98 , @xmath102 , and the contour integral yields the value @xmath36 for @xmath103 .
it is shown in @xcite that for real value functions , @xmath104 has the following form @xmath105 according to the _ bromwich inversion integral _ , @xmath99 can be calculated from the transform @xmath98 by performing a numerical integration ( quadrature ) .
we use a specific algorithm based on the bromwich inversion integral .
it is based on a variant of the fourier - series method - the trapezoidal rule - which proves to be remarkably effective .
if we use a step size @xmath106 , then the trapezoidal rules gives @xmath107 where @xmath108 since @xmath101 is real .
replacing @xmath109 by @xmath110 which is the laplace transform of @xmath17 , we get the probability of starvation before time @xmath19 @xmath111\end{gathered}\ ] ] the infinite series in can simply be calculated by simple truncating because it converges , but more efficient algorithm can be obtained by applying a summation acceleration method .
an acceleration technique that has proven to be effective in our context is euler summation , after transforming the infinite sum into a nearly alternating series in which successive summands alternate in sign .
we convert into a nearly alternating series by letting @xmath112 and @xmath113 @xmath114 where @xmath115}\bigg)\end{gathered}\ ] ] let @xmath116 be the approximation @xmath117 with the infinite series truncated to @xmath118 terms , i.e. , @xmath119 where t is suppressed in the notation and @xmath120 .
we apply euler summation to @xmath121 terms after an initial @xmath118 , so that the euler sum approximation is @xmath122 euler summation can be very simply described as the weighted average of the last @xmath121 partial sums by a binomial probability with parameter @xmath121 and @xmath123 .
hence , is the binomial average of the terms @xmath124 .
the implementation of the algorithm takes into account the values of @xmath125 . as in @xcite
, we use @xmath126 where @xmath127 . after simplification and letting @xmath128
be 1 , we get the euler approximation @xmath129 of the inverse @xmath99 which seems to be a good approximation @xmath130\end{gathered}\ ] ] eq .
( [ eulerfct ] ) looks complicated , but it consists of only @xmath131 additions , that is a low computation level . to have the cdf @xmath132 , we just replace @xmath98 by @xmath133 .
the same formula holds for the start - up delay distribution in replacing @xmath98 by @xmath134 .
in this section we compute the qoe metrics based on the analysis derived in the previous section . we consider a single user receiving a media file with size @xmath0 . the necessary time to play the whole video
if there is no starvation is @xmath135 .
hence , using the first passage time distribution @xmath17 , the probability of starvation happened in state @xmath79 before reaching the end of file given the initial state @xmath32 , is given by @xmath136\end{gathered}\ ] ] the starvation of probability before time @xmath64 gives an idea of the severity of the starvation during the video session .
let @xmath137 $ ] be the mean continuous playback time if the initial state is @xmath8 , the prefetching threshold is @xmath20 and the starvation happens in state @xmath18 .
@xmath138 is an important measure for the severity of starvations .
a small @xmath138 means that the starvation events happen frequently .
we find @xmath138 by taking derivatives of @xmath139 in @xmath140 .
@xmath141 when the user starts the video session , the initial state is unknown to the system .
the video starts playing when the prefetching process is finished .
conditioning on the distribution of the entry states @xmath142 , the distribution of the states that the playback process begins ( or prefetching process ends ) is computed by @xmath143 . recalling that @xmath144 is the probability that the prefetching phase ends at state @xmath18 knowing that the video session starts at state @xmath8 .
then the starvation probability with the prefetching threshold @xmath20 is obtained by @xmath145 where @xmath146 is a column vector , @xmath147 and @xmath148 .
@xmath149 is called the overall starvation probability .
the probability of no starvation is @xmath150 .
the start - up delay is proportional to the start - up threshold .
but in the qoe literature , it is more practical to consider the delay rather than the threshold because the delay has a direct impact on the streaming user behaviour . using the results of the sections [ invert ] and [ delayup ]
, we derive the cumulative distribution function of the start - up delay @xmath151\end{gathered}\ ] ] where @xmath20 is the start - up threshold , @xmath152 is the file size and @xmath153 are the euler summation algorithm parameters .
when a starvation event happens , the media player pauses until @xmath20 contents are re - buffered .
we are interested in the probability distribution of the starvations , given the file size @xmath152 .
we define a _ path _ as a complete sequence of frames arrivals and departures .
we illustrate a typical path with @xmath18 starvations in fig .
[ starvl ] .
the path can be decomposed into three types of mutually exclusive events as follows : * event @xmath154 : the buffer becoming empty for the first time in the entire path .
* event @xmath155 : the empty buffer after the instant @xmath156 given that the previous empty buffer happens at @xmath157 . * event @xmath158 : the last empty buffer observed after the instant @xmath159 . obviously , a path with @xmath18 starvations is composed of a succession of events @xmath160 we let @xmath161 , @xmath162 and @xmath163 be the probabilities of events @xmath164 , @xmath165 and @xmath166 respectively .
the probability distribution of event @xmath164 is expressed as @xmath167 where @xmath168 and @xmath169 are @xmath170x@xmath170 and @xmath170x@xmath171 matrices respectively .
the first starvation can not happen at the departure of first @xmath172 contents because of the prefetching of @xmath20 contents .
it can not happen after all @xmath152 contents have been served because this empty buffer is not a starvation . for @xmath173
the starvation happens at time @xmath174 conditioned on the states that the playback process begins .
the probability distribution of event @xmath175 is given by @xmath176 where @xmath146 is a column vector .
@xmath159 is the time of the @xmath18-th starvation .
the extreme case is that these @xmath18 starvations take place consecutively.then @xmath177 should be greater than @xmath178 .
otherwise there can not have @xmath18 starvations .
if @xmath177 is no less than @xmath179 , the media player resumes until all the remaining content @xmath180 is stored in the buffer .
then , starvation will not appear afterwards . in the remaining case , it is the probability of having no starvation after time @xmath159 .
we denote by @xmath181 the probability of having @xmath18 starvations . the case with one starvation
is given by @xmath182 to compute the probability of having more than one starvation , we need to find the probability of event @xmath165 .
@xmath183 should not be less than @xmath184 in order to have @xmath128 starvations .
given that the buffer is empty just after time @xmath157 , the @xmath185 starvation can not happen at @xmath186 $ ] because of the prefetching process .
since there are @xmath18 starvations in total , the @xmath185 starvation must satisfy @xmath187 .
we next compute the remaining case that the @xmath188 and the @xmath185 starvations happen at time @xmath157 and @xmath156 respectively .
we compute this probability using the first passage time density when the starvation happens at time @xmath156 and the initial time was @xmath157 with a prefetching process .
@xmath162 is expressed as @xmath189 we use in this method a trick that concerns the time scale .
every time the player resumes for the prefetching process we resume also the time scale , that means if the starvation happens at time @xmath19 , the player will start playing at the same time @xmath19 with @xmath20 initial contents in the buffer .
the probability of having @xmath190 starvations is given by @xmath191 in the next section , we provide explicit expressions of qoe metrics where ctmc has two states .
in this section , we consider a special case in which the ctmc has two states : @xmath192 ( see fig .
[ 2mpfig ] ) with infinitesimal generator @xmath3 and rate matrix @xmath193 @xmath194 the two - state mmpp source ] our objective is to understand the interaction between the parameters of arrival process and the probability of starvation . using the results of section [ first_passage ] irrespective of the condition of stability of the queue @xmath195 , we get : @xmath196s + \omega ( \omega + \alpha + \beta ) \end{gathered}\ ] ] it is a polynomial of degree 2 in @xmath197 where the two zeros are given by : @xmath198 @xmath199 where @xmath200 .
equation contains terms with only @xmath201 .
so we have to determine the signs of @xmath202 and @xmath203 .
the next propositions give the placement of these two zeros in the complex plane .
let @xmath204 , * @xmath205 @xmath206 @xmath207 , so @xmath208 and @xmath209 then no starvation .
* @xmath210 and @xmath207 , so @xmath211 and @xmath208 then @xmath212 and @xmath213 .
* @xmath210 and @xmath214 , so @xmath211 and @xmath215 then @xmath216 and @xmath213 . * @xmath217 @xmath206 @xmath207 then no starvation . * @xmath218 @xmath206 @xmath210 , @xmath219 and @xmath220 .
let @xmath221 , * @xmath207 @xmath206 @xmath205 , so @xmath208 and @xmath209 then no starvation . * @xmath214 and @xmath205 , so @xmath215 and @xmath209 then @xmath212 and @xmath213 . *
@xmath214 and @xmath210 , so @xmath215 and @xmath211 then @xmath216 and @xmath213 .
* @xmath218 @xmath206 @xmath205 then no starvation . *
@xmath217 @xmath206 @xmath214 , @xmath222 and @xmath220 .
let @xmath223 , * @xmath224 , no starvation .
* @xmath225 , @xmath216 and @xmath213 . *
@xmath226 , no starvation because of the prefetching .
the lst @xmath227 of the distribution is given in the next theorem .
* when @xmath214 , @xmath228 , @xmath229 and @xmath230 * when @xmath210 , @xmath231 , @xmath232 and @xmath233 * when @xmath214 , @xmath210 , @xmath234 @xmath235 where @xmath236 the proof of this theorem can be find in the appendix of @xcite .
taking @xmath237 gives the first passage time distribution for the on - off source .
we use ns-3 simulator in order to compare the dynamics of the process with our model .
the simulation topology consists on a server and a client in order to simulate the queue model .
the server sends the traffic to the client following the continuous time markov chain .
the client holds a buffer where the traffic is stored .
the parameters of the traffic depend on the ctmc parameters .
then we analyze the behavior of the client buffer content which simulates the player .
we run 10 simulations and compute the 95% confidence interval on all observed metrics , but it is not shown on all the figures for improving readability because it is very narrow .
we first show the accuracy of the method that we use to invert the laplace transform . in fig .
[ testfig ] , we plot the known inverse laplace transform of the function @xmath238 that is @xmath239 and the inverse using formula ( [ eulerfct ] ) of section [ invert ] .
[ offinf ] shows the starvation probability for a two states mmfm source for @xmath240 , @xmath241 and @xmath242 , that means the buffer size increases on state @xmath243 and decreases on state @xmath244 .
this is done for states transitions @xmath245 and @xmath246 .
the accuracy of euler summation algorithm , height=196 ] , height=196 ] , height=196 ] , height=196 ] ] fig .
[ nostarvx ] illustrates the impact of the start - up threshold @xmath20 on the probability of no starvation .
[ onestarv ] shows the probability of having one starvation .
these simulation results validate the correctness of our analysis .
hence , in the following experiments , we only illustrate the analytical results . ,
height=196 ] , height=196 ] , height=196 ] fig .
[ startupdelay ] gives the cdf of the start - up delay for different values of start - up threshold @xmath20 .
we can see that the start - up delay increases with @xmath20 . on the other hand figure [ nostarvx ]
illustrates the impact of the start - up threshold @xmath20 on the probability of no starvation .
when @xmath20 is large enough ( near 300 pkts in the figure ) no starvation will happen until the end of the video . since , the curve grows sharply , it is clear that a slight increase in @xmath20 can greatly improve the starvation probability . in figure [ nstarvx ]
, we plot the probability of having no more than two starvations with @xmath247 , @xmath248 , @xmath249 irrespective to @xmath20 .
when @xmath20 is large enough , no starvation will happen until the end of the video session .
on the other side , figure [ nstarvn ] shows that the starvation happens for sure when the file size approaches infinity .
the curves of the probability of having one starvation or two starvations increase first , and then decrease to zero .
this means that the starvation can be avoided when @xmath20 is large enough .
the two curves have a maximum value at a given start - up threshold or a given file size .
so one can choose the threshold @xmath20 to have exactly one or two starvations .
this is an important measure because it allows one to achieve the buffer requirements in setting up the desired values .
indeed , a very small threshold do not help to reduce the starvation probability and very large thresholds do not further reduce the starvation probability .
hence the analytical model aims to predict the player buffer behavior in video streaming sessions .
the network parameters and the video size are the framework inputs that can be used to improve the qoe related to user preferences . in this section ,
we introduce an optimization problem of the qoe by including different metrics and incorporating user preferences by associated a weight to each metric .
we denote by @xmath250 the cost of a user watching the media stream , @xmath251 where @xmath252 is the number of starvation , @xmath14 is the start - up delay , @xmath253 is the lost on video quality and @xmath254 is the fraction of the total session time spent in low bit - rate .
@xmath255 , @xmath256 and @xmath257 are depending on the user preferences of the three metrics ( starvation , start - up delay and video quality ) .
based on the user preferences , we compare the cost of qoe for two scenarios . in scenario 1 ,
the adaptive bitrate streaming is not used and for the second scenario , the adaptive bitrate streaming is used in order to adjust the quality of a video stream according the available bandwidth .
we consider a network that throughput varies between 200kbps and 400kbps .
for the adaptive bitrate streaming , we have two coding rates depending on the throughput .
this leads to two different frame sizes ( 10kbits , 20kbits ) .
then we compute the cost @xmath258 for @xmath259 and @xmath260 .
[ progadap ] and [ progadap2 ] , we compare the cost for the two scenarios . for short video duration
, the adaptive bitrate streaming is not benefit because there is a less number of starvation and the quality of the video is degraded .
but , for the long video duration , the adaptive bitrate streaming becomes interesting because the low coding rate decreases the number of the starvation . in fig .
[ progadap ] and [ progadap2 ] , we can see that the value of the parameter @xmath257 changes the preference of the user for the quality of the video . for @xmath261 , @xmath262
, we use the adaptive bitrate streaming when the size of the file is more than 400 , 600 frames respectively .
otherwise , the adaptive bitrate streaming is not necessary . ,
height=196 ] , height=196 ]
in this paper , we have proposed a new analytical framework to compute the qoe of video streaming in the network modeled by the markov modulated fluid model .
we found the probability of starvation and the start - up delay in solving partial differential equations through the laplace transform method .
this allowed us to compute the number of starvation during the video session that is an important metric of the quality of experience of the user .
in addition , we have presented simulation results using ns3 to show the correctness of our model . we have proposed a method to optimize the quality of experience given a trade - off between the player starvation and the quality of the video .
these results show that the adaptive bitrate streaming could impact negatively on the quality of the short video duration . 1 y. xu , e. altman , r. el - azouzi , s. elayoubi and m. haddad , _ qoe analysis of media streaming in wireless data networks _ , networking springer berlin heidelberg , 2012 .
xu , y. zhou and d - m . chiu , _ analytical qoe models for bit - rate switching in dynamic adaptive streaming systems _
, ieee transactions on mobile computing , 2013 .
xu , s.e .
elayoubi , e. altman , r. elazouzi and y. yu _ flow level qoe of media streaming in wireless networks _ , eprint arxiv:1406.1255 , 06/2014 .
a. balachandran , v. sekar , a. akella , s. seshan , i. stoica and h. zhang , _ developing a predictive model of quality of experience for internet video _ , sigcomm 2013 , hong kong .
xu , s.e .
elayoubi , e. altman and r. elazouzi , _ impact of flow - level dynamics on qoe of video streaming in wireless networks _ , ieee infocom , 2013 .
a. narayanan , and v.g .
kulkarni , _ first passage times in fluid models with an application to two priority fluid systems _ , computer performance and dependability symposium , proceedings ieee international , 1996 .
f. dobrian , a. awan , i. stoica , v. sekar , a. ganjam , d. joseph , j. zhan , and h. zhang , _ understanding the impact of video quality on user engagement _ , proc . of acm sigcomm 2011 , vol.41 , no.4 , pp .
362 - 373 .
j. abate , g.l .
choudhury and w. whitt , _ computational probability _ , w. grassman ( ed . ) kluwer , boston 1999
. y.d .
xu , e. altman , r. el - azouzi , m. haddad , s. elayoubi and t. jimenez , _ analysis of buffer starvation with application to objective qoe optimization of streaming services _ , ieee infocom , 2012 . s. shah - heydari , t. le - ngoc , _ mmpp models for multimedia traffic _ , telecommunication systems 15 , march 1999 .
a. balachandran , v. sekar , a. akella , s. seshan , i. stoica , and h. zhang . a quest for an internet video quality - of - experience metric . in 11th acm workshop on hot topics in networks ( hotnets - ix ) , 2012 .
s. s. krishnan and r. k. sitaraman .
video stream quality impacts viewer behavior : inferring causality using quasi - experimental designs . in acm imc , 2012 . h. hu , x. zhu , y. wang , r. pan , j. zhu , and f. bonomi , `` qoebased multi - stream scalable video adaptation over wireless networks with proxy , '' in icc , 2012 d. bethanabhotla , g. caire , and m. j. neely , `` joint transmission scheduling and congestion control for adaptive video streaming in smallcell networks , '' arxiv:1304.8083 [ cs.ni ] , 2013 .
v. joseph , s. borst , and m. reiman , `` optimal rate allocation for adaptive wireless video streaming in networks with user dynamics , '' submitted 2013 .
vinay joseph and gustavo de vecian , nova : qoe - driven optimization of dash - based video delivery in networks , ieee infocom , april 2014 .
r. mok , e. chan , and r. chang , `` measuring the quality of experience of http video streaming , '' ifip / ieee international symposium on integrated network management , 2011 .
luan , l.x .
cai and x. shen , _ impact of network dynamics on user s video quality : analytical framework and qos provision _ , ieee transactions on multimedia , vol . 12 , january 2010 . _
when @xmath263 , @xmath264 , @xmath265 @xmath266 and @xmath267 is the eigenvector correspondind to @xmath59 according to section [ first_passage ] , then @xmath268 when @xmath214 , @xmath228 , we use only @xmath269 in computing the distribution , since @xmath270 .
thus the distribution @xmath271 becomes @xmath272 @xmath273 and @xmath274 because if we start without packets in the buffer in state 1 , we ll have starvation with probability 1 within the same state .
so @xmath275 , that yields the result of the theorem .
the same proof holds in the case @xmath210 , @xmath231 by interchanging @xmath276 and @xmath277 .
+ when @xmath214 , @xmath210 , we use both @xmath278 and @xmath269 . thus we have @xmath279 using the initial condition @xmath280 and @xmath281 , we solve the following system @xmath282 and get @xmath283 the same proof yiels for @xmath284 . | we take an analytical approach to study the quality of user experience ( qoe ) for video streaming applications .
our propose is to characterize buffer starvations for streaming video with long - range - dependent ( lrd ) input traffic .
specifically we develop a new analytical framework to investigate quality of user experience ( qoe ) for streaming by considering a markov modulated fluid model ( mmfm ) that accurately approximates the long range dependence ( lrd ) nature of network traffic .
we drive the close - form expressions for calculating the distribution of starvation as well as start - up delay using partial differential equations ( pdes ) and solve them using the laplace transform .
we illustrate the results with the cases of the two - state markov modulated fluid model that is commonly used in multimedia applications .
we compare our analytical model with simulation results using ns-3 under various operating parameters .
we further adopt the model to analyze the effect of bitrate switching on the starvation probability and start - up delay .
finally , we apply our analysis results to optimize the objective quality of experience ( qoe ) of media streaming realizing the tradeoff among different metrics incorporating user preferences on buffering ratio , startup delay and perceived quality . |
it is well established that gravitationally lensed quasars are unique natural rulers for measuring the universe and for deriving the cosmological parameters ( refsdal , 1964a , b ) . measuring
the time delay from the images of a lensed qso can provide an estimate of the hubble parameter @xmath5 , independent of any other classical method .
however , a good knowledge of the geometry of the lensed system is mandatory for the method to be effective ( e.g. schechter et al , 1997 ; keeton & kochanek , 1996 ; courbin et al , 1997 ) . in spite of this crucial requirement and
although the number of known gravitationally lensed quasars does not stop increasing ( see for a review keeton & kochanek , 1996 ) , the precise geometry of most lensed qsos remains poorly known . in most cases , even the matter responsible for the lensing , whether it be in the form of a single galaxy or several galaxies , is not detected .
the high redshifts of these galaxies ( hence their faint apparent magnitudes ) and the strong blending with the nearby much brighter qso images are the main reasons for their non - detection .
imaging in the near ir ( 1 to 2.5 microns ) has the advantage that the relative brightness between the lensed qso and any lensing galaxy decreases , making the galaxy easier to detect .
the disadvantage is that the ir sky is considerably brighter .
this forces one to take many images to avoid detector saturation ; however , this turns out to be an advantage ( see section 3 ) .
this paper presents ir observations of the quasar he 11041805 . the strong similarity between the optical spectra obtained for its two components ( wisotzki et al , 1993 ) makes he 11041805 a good gravitational lens candidate .
the high redshift of the object ( @xmath6 , smette et al , 1995 ) and the relatively wide angular separation between the lensed images ( 3.2 ) indicate that an large mass is involved in the lensing potential
. if the deflector is a high redshift galaxy or a galaxy cluster , deep ir observations should reveal it .
we used a recently developed image deconvolution algorithm ( magain , courbin & sohy , 1997 ; hereafter mcs ) to optimally combine the numerous ir frames and obtain deep , sharp images of he 11041805 .
the present paper describes how this powerful technique allows us to study the immediate environment of he 11041805 and detect the lensing galaxy .
the observations took place at the eso / mpi 2.2 m telescope situated at la silla observatory , chile , on the nights of april 14 and 15 , 1997 .
the ir camera irac2b was used at the cassegrain focus of the telescope .
the detector of irac2b is a 256@xmath7256 nicmos 3 hgcdte array and the instrument has a variety of optical lenses available for imaging at different pixel scales .
lens lb ( lidman , gredel & moneti , 1997 ) was chosen since it gives a good compromise between the pixel scale ( a small pixel is needed for the deconvolution ) and the size of the field . during the observations , pg 1115 + 080 was also observed .
a by - product of these observations is a more accurate estimate of the irac2b pixel size for images taken in lens lb .
the new scale , 0.2762/pixel , is based on the precise astrometry of this field given by courbin et al .
this results in a field of view of 71 .
numerous short exposures of he 11041805 were obtained in @xmath0 ( @xmath8=1.25 micron ) and @xmath1 ( @xmath8=2.15 micron ) under good meteorological conditions .
the mean seeing was 07 - 08 and the sky was photometric .
we set the detector integration time ( dit ) to 20 sec in @xmath1 and 60 sec in @xmath0 . each image taken in @xmath1 ( resp .
@xmath0 ) is the average of 3 ( resp .
2 ) such integrations .
the choice of the dit is dictated by detector saturation .
the number of dits is dictated by the frequency at which the sky intensity varies . since the field of he 11041805 is uncrowded , we took sequences of 9 science exposures , dithered in a semi random manner by 5 to 10 , always keeping the object and three psf stars in the field .
dome flat - fields were taken in order to correct for the pixel to pixel sensitivity variations of the detector
. however , dome flat - fields do not accurately represent the large scale sensitivity variations of the array , in the @xmath0 band .
fortunately , this variation does not show strong gradients and has a maximum amplitude of 3 - 4% over the whole field .
it was modeled by observing a bright star over a grid of 9 different positions across the array and by fitting a two dimensional , third order polynomial to the flux of the star .
the residuals of the fit were 1% .
this fit , which is commonly called an illumination correction , was multiplied by the dome flat to produce the final flat - field containing both the low and high frequency sensitivity variations of the array .
after subtraction of a dark frame , the flat - field correction is applied to all scientific frames . the background is removed from every exposure .
it is estimated for any particular exposure by averaging the 6 preceding and the 6 following exposures . thanks to the dithering between exposures , objects in these frames could be rejected before the average was taken .
this method allows us to accurately follow the background which varies on the time scale of a few minutes .
standard stars were observed every two hours .
the standard deviation in the zero points were 0.024 magnitudes in @xmath0 and 0.012 magnitudes in @xmath9 .
the magnitude and colours in this paper are in the @xmath10 system as defined by bessell and brett ( 1988 ) .
the frames were combined in two ways .
first , the standard reduction and image combination techniques implemented in the iraf package were used in order to average the frames .
the sigma - clipping algorithm was used for bad pixel rejection .
this leads to two deep @xmath0 and @xmath1 images over a field of 1 .
the total exposure times were 5040 sec in @xmath0 and 8100 sec in @xmath1 .
the resulting detection limit is 22 in @xmath0 and 20 in @xmath9 ( 3@xmath11 , integrated over the whole object ) .
1 presents the field in the @xmath0 band .
= 8.5 cm = 17.5 cm in order to study the immediate environment of he 11041805 , we used the new mcs deconvolution algorithm described in full detail by magain , courbin & sohy ( 1997 ) .
deconvolution of an image by the total observed point - spread - function ( psf ) leads to the so - called `` deconvolution artifacts '' or `` ringing effect '' around the point sources .
this results from the deconvolution algorithm attempting to recover spatial frequencies higher than the nyquist frequency , thus violating the sampling theorem .
instead , the mcs algorithm uses a narrower psf which ensures that the deconvolved image will not violate the sampling theorem .
additionally , the mcs algorithm takes advantage of important _ prior knowledge _ : in the deconvolved images , all the point sources have the same ( known ) shape .
this allows us to decompose the deconvolved image into a sum of analytical point sources plus a diffuse background which is smoothed to the final resolution _ chosen by the user_. most of the deconvolution artifacts are thus avoided .
this is of particular interest when one wishes , like in the present study , to discover faint objects embedded in the seeing disks of much brighter point sources .
if the deconvolution of a single image already yields very good results ( e.g. courbin et al .
1997 ) , the simultaneous deconvolution of numerous dithered exposures is even more efficient ( e.g. courbin & claeskens , 1997 ) . in particular , the mcs code allows the pixel size of the deconvolved image to be as small as desired .
this over - sampling possibility , already applicable to the deconvolution of a single frame , is of considerable interest when dealing with the spatial information contained in many dithered frames .
another advantage of the mcs algorithm is that the psf can vary from frame to frame .
for example , one can combine good quality images with trailed or even defocused images or , in a more reasonable way , frames of differing image quality and signal - to - noise ratios .
the resulting frame is an optimal combination of the whole data set , with _ improved resolution and sampling_. the seeing in the original irac2b images was of the order of 06 for the very best frames and up to 12 for the worse ones , in both @xmath0 and @xmath1 .
we adopted for the deconvolution a sampling step of 01381 , two times smaller than the original pixel size .
this allows us to reach a final resolution of 02762 which still samples well the resulting image ( 2 pixels full - width - half - maximum ( fwhm ) ) . since the signal - to - noise of individual images is very low and since we had to reject very numerous bad pixels , we first combined the images in groups of nine . thus , we obtained 6 intermediate images in @xmath0 and 12 images in @xmath1 .
a psf was derived for each of these images . in @xmath1 ,
only `` star 1 '' is bright enough , - i.e. comparable to the qso s luminosity - to compute an accurate psf ( see fig . 1 for the labeling of the stars used ) .
in @xmath0 , `` star 1 '' shows extended luminosity so we used both `` star 2 '' and `` star 3 '' .
the resulting total exposure time of the co - added images is different from the one of the images combined using iraf and the standard methods .
we rejected more frames with bad pixels falling right on the object or the psf stars . on the other hand
, we included in the different stacks more images with bad seeing .
thus , the total exposure time of the deconvolved images is 6480s in both @xmath0 and @xmath1 .
the program requires initial estimates for the positions and intensities of the point sources in the field .
this was done by choosing the central pixel of each qso image . during deconvolution
, the centres of the point sources are forced to be the same in all the images , only an image translation ( no rotation ) being allowed .
the data are never aligned or rebinned ; only the deconvolved model ( on which the highest spatial frequencies are modeled analytically ) is transformed .
the intensities of the point sources can be allowed to be different in each image so that even variable objects may be considered in the deconvolution .
the shape of irac2b psf shows significant variations across the field . in @xmath0 ,
the variation is still acceptable , mainly because we used `` star 2 '' and `` star 3 '' , which are closer to he 11041805 than `` star 1 '' , which is used for the psf computation in the @xmath1 band .
it is possible in our algorithm to let the psf depart slightly from its original shape , during the deconvolution process .
it is in fact re - determined directly from the point sources in the field that is deconvolved . in the present case of simultaneous deconvolution
, the correction on the psfs is well constrained by the numerous images considered .
the quality of the psf correction is even better if numerous stars are present in the field . with 2 point sources and 12 images in @xmath1 , it has been possible to correct rather well the psfs of the 12 images .
the deconvolution is first performed with variable psfs , and then repeated with the corrected psfs fixed .
= 9.0 cm the background component of the deconvolution is smoothed on the length scale of the final resolution .
the weight attributed to the smoothing ( see magain , courbin & sohy , 1997 for more detail ) is chosen so that the residual map between each data frame and the model image ( reconvolved with the psf ) in units of the photon noise , has the correct statistical distribution , i.e. is equal to unity all over the field . in other words , we chose the smoothing term by inspecting the local residual maps .
this ensures that the deconvolved image is compatible with the whole data set in any region of any of the data frames .
the deconvolution consists of a @xmath12 minimization between the deconvolved model image and _ the whole data set _ , using an algorithm derived from the conjugate gradient method .
again , the residual maps are used as a quality check of the result .
we stop the iteration process only when the residual maps show the correct statistical distribution all over the field so that we avoid local over or under - fitting .
the program produces the following outputs : a deconvolved image , the centre of the point sources , the shifts between the images , the intensities of the point sources for each of the individual frames and an image of the deconvolved galaxy , free of any contamination by the qsos .
figure 2 displays the result of the deconvolution for the @xmath0 band images .
six images were used to obtain this result .
the spatial resolution is 02762 , comparable to the resolution reached by the hst in the ir domain .
we chose the same final resolution for the simultaneous deconvolution of 12 @xmath1 images .
the lensing galaxy is clearly detected and displayed in fig .
it is also seen in @xmath1 , were it is in fact brighter .
.summary of the astrometry ( in the same orientation as fig . 1 and fig .
2 ) and photometry for he 11041805 and the lensing galaxy . the 1@xmath11 error bars are also indicated . [
cols="<,>,>,>",options="header " , ] figure 3 shows the position of the galaxy , relative to the qso images . a slight misalignment between the lens and the line joining qso a and qso b can be seen .
it is larger than our error bars and is apparent in both @xmath0 and @xmath1 .
in addition , the psf s shape does not show any significant variations across the deconvolved field ( only 8.8 ) so that any geometric distortion can be ruled out . the observed misalignment seems therefore real .
no obvious galaxy overdensity is detected in the immediate surrounds of the qso , although the detection limit of 22 in @xmath0 and 20 in @xmath9 would have allowed us to see any rich cluster up to @xmath13 .
the two nearest objects to the double qso are galaxies @xmath14 and @xmath15 ( see fig .
1 ) . table 2 gives their position relative to qso a and their photometry , both derived on the `` un - deconvolved '' image since they are outside the field used for the deconvolution .
the redshift of the lensing galaxy can be constrained by the @xmath16 colour . in figure 4
, we compare the colour of the candidate ( the shaded region ) with theoretical colours corresponding to five galaxy types .
these colours were obtained from the pegase _
`` projet detude des galaxies par synthse evolutive '' _ atlas ( leitherer et al .
1996 ; fioc and rocca - volmerange 1997 ) .
the five types correspond to e , sa , sb , sc and sd galaxies in a critical density universe with @xmath17 ( rocca - volmerange and fioc 1996 ) . the theoretical colours are a function of redshift .
the thick lines include the effect of galaxy evolution ; the thin lines do not .
= 9.0 cm the colour of the candidate constrains the object to be a galaxy with a redshift beyond @xmath18 , probably between z=1 and z=2 .
the redshifts of the two strongest metal absorption line systems , at @xmath19 and @xmath20 ( smette et al . 1995 and wisotzki et al .
1993 ) are marked with dotted arrows in fig .
4 and fig . 5 . in figure 5
, we plot the absolute k - band magnitude the galaxy would have today if it were at various redshifts .
the horizontal line represents the absolute k - band magnitude @xmath21 of a typical large galaxy .
the value we adopted for @xmath21 is the average between the value determined by glazebrook et al .
( 1995 ) and the one from mobasher et al .
( 1993 ) . if the galaxy is at a redshift between @xmath22 and @xmath13 ( as also suggested by smette et al .
1995 ) , then it is several times more luminous than an @xmath21 galaxy .
the main result of the present study is the detection of a red fuzzy object located between the two components of he 11041805 .
this result is one more very strong argument in favour of the lensed nature of this double quasar .
wisotzki et al ( 1993 ) do not detect the lensing galaxy in @xmath23 down to a limiting magnitude of 23 - 24 . in fig .
6 the tracks in the @xmath24 vs @xmath16 colour - colour diagram of two galaxy types , an elliptical and an sa galaxy , are plotted .
they are plotted with and without evolution and for the redshift range @xmath25 .
also plotted is the range allowed by the observations in this paper and the optical observations of wisotzki et al .
( 1993 ) .
if we include the effects of evolution ( thick lines in the figures ) , the ir colours are compatible with an elliptical galaxy ( as shown by fig .
6 ) between @xmath22 and @xmath13 ( fig . 4 and fig . 5 ) .
the ir - optical colours are less compatible with this ; however , the expected r magnitude of the lensing galaxy is @xmath26 , and this may have been difficult to see 11 away from the qso which is 5 to 6 magnitudes brighter .
in fact , a preliminary detection of the lensing galaxy by grundahl , hjorth & srensen ( 1995 ) allowed to measure an @xmath27-band magnitude of 20.6 , in better agreement with our findings .
one of the two metallic absorption line systems found at @xmath19 and @xmath20 ( smette et al 1995 ) could be produced by the lensing galaxy . in particular ,
the absorption system at @xmath20 is seen almost only in qso a. since the angular distance from the lens to qso a is much smaller than to qso b it seems reasonable to think that the lensing galaxy we detect is more likely to be at @xmath20 rather than 1.320 . despite the depth of our ir images , which enables us to detect @xmath21 galaxies up to a redshift of the qso , we do not detect any obvious overdensity of galaxies which could contribute significantly to the total gravitational potential involved in this system .
however , two faint galaxies ( g1 and g2 ) are detected close to the line of sight to the qso .
g1 has a @xmath16 colour of 1.1 , while g2 has a colour close to that of the lens galaxy .
these two objects could constitute an external source of shear , for example responsible for the misalignment between the lensing galaxy , qso a and qso b. we can infer from our deconvolutions that qso a is not exactly compatible with a single point source .
the deconvolution leaves significant residuals at the location of qso a , even in @xmath0 where the psf is rather stable across the field .
the signal - to - noise ratio and resolution of our observations do not allow to draw definite conclusions , but we suspect that image a is either not single or is super - imposed on a fuzzy faint light distribution .
however , one can not exclude the simpler ( but unlikely ) explanation that the light and mass centroids of the lensing galaxy do not coincide . from the geometry of the lensed system , given in table 1 , and assuming we see a galaxy at @xmath20 , the time delay we can expect between the two images of he 11041805 is of the order of 3.5 years .
we have assumed that the lens can be modeled as a singular isothermal sphere ( sis ) and that @xmath28 .
this large delay means that one measurement every second week would be enough to derive good light curves . assuming that an sis is appropriate for the lensing galaxy , we derive a mass of @xmath29 , not too far above the masses expected for big elliptical galaxies . fig .
5 supports that the lensing galaxy is bigger than a `` normal '' galaxy .
if it is actually at @xmath2 it is one magnitude brighter than an @xmath30 galaxy .
finally , the magnitude difference between the lensed images is @xmath31 and @xmath32 , where the magnitude difference is taken as @xmath33 .
the magnitude difference expected from the sis model is 0.75 magnitude , but with the deflector angularly closer to the faint image than to the brighter one .
this could indicate that component b is reddened relative to a , or that component a is preferentially amplified ( e.g. slight image splitting ) relative to b and that this preferential amplification is more efficient in the blue .
the latter hypothesis is more likely since the lens galaxy is angularly closer to qso a than to qso b and would therefore redden a more than b ( assuming the reddening is due to the lens galaxy ) . on the other hand , the lensing potential might be more complex than a sis ( for example elliptical + core ) , in particular if g1 and g2 introduce a significant source of shear .
wisotzki et al ( 1995 ) showed that microlensing was acting on qso a and that it was more efficient in the blue than in the red .
the magnitude difference they observed in @xmath34 in 1994 was @xmath35 , larger than our present values in @xmath0 and @xmath9 ( although the quasar has probably varied between 1994 november and 1997 april ) .
this suggests that microlensing is less efficient in the ir than in the visible .
if the source quasar is found to be variable in the ir domain , an ir photometric monitoring of he 11041805 may then minimize contamination by microlensing events and allow a better determination of the time delay than with optical data .
the authors would like to thank j .-
swings , p. schechter , j. hjorth and the referee y. mellier for helpful comments on the first version of this manuscript .
this work has been in part financially supported by the contract arc 94/99 - 178 `` action de recherche concerte de la communaut franaise '' ( belgium ) , and ple dattraction interuniversitaire p4/05 ( sstc , belgium ) . | we report on deep ir imaging of the double quasar he 11041805 .
a new image deconvolution technique has been applied to the data in order to optimally combine the numerous frames obtained .
the resulting @xmath0 and @xmath1 images allow us to detect and study the lensing galaxy between the two lensed qso images .
the near infrared images not only confirm the lensed nature of this double quasar , but also support the previous redshift estimate of @xmath2 for the lensing galaxy .
no obvious overdensity of galaxies is detected in the immediate region surrounding the lens , down to limiting magnitudes of @xmath3 and @xmath4 .
the geometry of the system , together with the time delays expected for this lensed quasar , make he 11041805 a remarkable target for future photometric monitoring programs , for the study of microlensing and for the determination of the cosmological parameters in the ir and optical domains .
# 1(ck ) # 1 |
the discovery of superconductivity in layered lafeaso@xmath8f@xmath9 four years ago has brought a new trend to synthesize and study new materials which are analogous to this compound.@xcite by now , most of the interest is focussed on the materials as @xmath10 `` 122''-type , @xmath4fe@xmath11 `` 111''-type and fe@xmath11 `` 11''-type which have high superconducting transition temperature @xmath12 and can be obtained as high - quality single crystals .
only few studies have been reported on low @xmath12 pnictide compounds or those that do not show superconductivity at all.@xcite among those is , that has the same tetragonal thcr@xmath5si@xmath5-type structure as @xmath13fe@xmath5as@xmath5 at room temperature and undergoes no structural transition upon cooling.@xcite the antiferromagnetic order appears at @xmath14 k ; i.e. at much higher temperature than for the @xmath13fe@xmath5as@xmath5 compounds.@xcite while most of the @xmath10 parent compounds are metallic , is a semiconductor with a small band gap @xmath15 27 mev according to @xmath16-resistivity measurement.@xcite an @xmath17 suggested that the difference in magnetic and electronic properties compared to those of @xmath13fe@xmath5as@xmath5 is due to the strong hund s coupling , the stability of the half - filled @xmath18-shell of the mn@xmath19 ( @xmath20 ion ) and strong spin - dependent mn - as hybridization.@xcite in this paper , we present mainly the infrared studies on single crystals with the electric field polarized within the @xmath0-plane and parallel to the @xmath1-axis .
the sample with a size of @xmath21 was grown out of sn fluxes .
the magnetic , transport , and thermal properties of our material have been reported recently.@xcite for the infrared reflectance measurements , we cleaved the crystal within the @xmath0-plane and finely polished the @xmath22-plane in order to get shiny surfaces .
the temperature dependent reflectivity spectra were measured in a wide frequency range from 40 to 37000 using fourier - transform infrared spectrometers ( @xmath23 ) and a variable - angle spectroscopic ellipsometer ( @xmath24 , restricted to room temperature ) .
raman spectra were measured using jobin - yvon t64000 spectrometer equipped with a microscope at room temperature .
we used 514.5 nm ar - laser excitation line with intensity below 1 mw .
spectra were measured in ( a , a ) and ( b , b ) polarizations .
several infrared and raman phonon modes are assigned and compared to the calculated values .
( color online ) temperature dependent optical reflectivity of measured in ( a ) the @xmath0-plane and ( b ) along the @xmath1-direction .
the inset shows the room - temperature reflectivity in the whole measured range ( 40 - 37000 ) . ] in the inset of fig .
[ fig : fig1 ] , we plot the room temperature reflectivity for both @xmath0-plane and @xmath1-axis polarization in the entire measured frequency range ( 40 to 37000 ) . as approaching the zero frequency , @xmath25 shows an upturn towards unity reflectance while @xmath26 tends to keep a constant value of approximately 40% .
the rather low overall reflectivity ( @xmath27 30% ) indicates that is a rather bad metal or a semiconductor . besides the pronounced phonon vibrations in far - infrared region which we will discuss in detail later in this paper
, two interband transition can be identified at approximately 7000 ( 0.87 ev ) and 26000 ( 3.2 ev).@xcite the temperature dependence of the reflectivity spectra is displayed in fig .
[ fig : fig1](a ) and ( b ) . in the @xmath0-plane , the reflectivity below 700 increases with decreasing temperature , indicating a metallic behavior .
this is consistent with the @xmath16-resistivity measurement reported in ref .
only a very small temperature dependence is observed along the @xmath1-axis due to the less conducting property .
.comparison of the phonon mode frequencies ( in ) found in different _
am@xmath5x@xmath5 _ compounds with the resonance frequencies of bamn@xmath5as@xmath5 .
[ table : modes ] [ cols="<,^,^,^,^,^",options="header " , ] due to the low reflection and weak absorption of in the @xmath0-polarization , a small fraction ( about 5% ) of the far - mid infrared light was found to be transmitted through the sample during the quality - check measurement .
this prevents us from a correct evaluation of conductivity spectra as commonly done by a kramers - kronig analysis.@xcite nevertheless , all the pronounced phonon vibration modes show lorentzian like shapes and thus are applicable for extracting the physics behind . according to the tetragonal thcr@xmath5si@xmath5-type structure ( with space group i4/mmm ) in which the @xmath10 usually crystallizes , a group theoretical analysis of the phonon modes in yields a@xmath28+b@xmath28 + 2e@xmath29 raman - active , 2a@xmath30 + 2e@xmath31 infrared - active and a@xmath30+e@xmath31 acoustic optical zone center phonons ,
where a / b and e modes correspond to an atomic motion perpendicular and parallel to mnas planes , respectively.@xcite in our infrared spectra , however interestingly , we observe at least four vibration modes both in ab - plane and along the c - axis , as shown in fig .
[ fig : fig1 ] .
thus we extract the imaginary part of the dielectric function @xmath32 and @xmath33 ( known as the energy loss function ) to identify the possible split transverse optical ( to ) and longitudinal optical ( lo ) modes , respectively.@xcite according to fig .
[ fig : fig5 ] , we can directly label the to / lo energies to be 74/92 and 210/238 for two e@xmath34 modes , 96/131 and 230/243 for two a@xmath30 modes . in fig .
[ fig : fig7 ] we show a room - temperature xx ( a@xmath28 and b@xmath28 modes are allowed ) raman spectrum of as obtained in ( a , a ) and ( b , b ) polarizations .
a pronounced peak can be observed at 174 .
we refer this contribution to the a@xmath28 mode which is corresponding to the as - as displacement along c - axis ( see also in fig .
[ fig : fig6 ] ) . in
table.i a comparison of phonon modes frequencies found in and several @xmath4fe@xmath5as@xmath5 ( @xmath4=ca , ba , sr and eu ) compounds is shown . due to the differences in lattice
constant , atomic mass , and possible magnetic interactions ( especially for eufe@xmath5as@xmath5 and ) , the vibration frequencies for the same phonon mode are slightly different in these materials .
( color online ) imaginary part of the dielectric function @xmath32 and @xmath33 at 300 k for single crystal with ( a , b ) e@xmath35ab and ( c , d ) e@xmath35c .
the frequencies of transverse and longitudinal optical modes are respectively identified as labeled in each panel .
@xcite ] ( color online ) xx - polarized raman spectrum of at 300 k. the peak at 174 is assigned as a@xmath28 mode . ] to compare with theoretical values , we also carried out phonon calculations ( in the _ nonmagnetic _ state ( nm ) ) on the density - functional perturbation theory level within the plane - wave method with the generalized gradient approximation ( gga ) of perdew - burke - enzerhof.@xcite we used the quantum espresso package for the electronic structure calculation with the implemented ultrasoft plane - wave basis set.@xcite the band occupation was fixed and the literature lattice parameters has been used.@xcite only the atomic positions in the unit cell were optimized .
the brillouin zone is sampled in the * k * space with a uniform monkhorst - pack grid ( 12@xmath3612@xmath3612).@xcite in fig .
[ fig : fig6 ] , we display the calculated atomic transverse optical displacements and the measured to / lo frequencies for all zone center phonons .
note that , in our calculations only a small lo - to splitting of less than 1 could be observed .
the reason for such a small splitting and the discrepancy in the frequencies bases mainly on the nonmagnetic state revealing no band gap .
an et al.@xcite show by using density functional theory that the ground state of bamn@xmath5as@xmath5 is antiferromagnetic ( afm ) with a small bandgap of about 0.2 ev in the case of gga calculations .
recent _ ab inito _ calculations of the phonon spectrum in the `` 122 '' -family including the magnetic interaction lead to a better agreement of the phonon spectrum .
@xcite the difference in the mode frequencies between the nm and the afm state can exceed 40 .@xcite it would be desirable to conduct further calculations including the afm state which should lead to a small band gap and consequently a bigger lo - to splitting.@xcite ( color online ) calculated atomic displacements and to / lo frequencies in for zone center phonons .
calculated frequencies are shown in brackets . ]
( color online ) temperature development of conductivity in the region of phonon modes of . ] in fig .
[ fig : fig4 ] , no change in the phonon shape is observed for these modes due to the absence of structural transition upon cooling .
but the e@xmath34 mode with higher vibrational frequency is enhanced in intensity and continuously shift to higher frequencies .
recent neutron diffraction measurement on the thermal contraction revealed that by cooling from 300 k to 10 k , @xmath37 distance decrease from 2.566(2 ) to 2.558(2).@xcite the lattice distortion with reducing bond lengths enhances the electronic interactions between atoms and therefore the phonon vibrating energy increases consistent with the observed blue shift of this e@xmath34 mode in our spectra .
there is an extra broad hump at the frequency around 400 in the in - plane reflectivity spectra , as shown in fig.1(a ) .
according to the phonon calculations , we can exclude it to be a phonon mode . as the optical conductivity spectra of
are phonon - rich ones , we will here analyze the optical data by looking at the relative change of the conductivity @xmath38 , as plotted fig .
[ fig : fig2 ] , to get rid of the strong phonon oscillations and see basically only the electronic background and consequently the spectral weight redistribution upon varying the temperature . as seen in fig .
[ fig : fig2 ] , around 300 @xmath39 drops to negative values with decreasing temperature , then it increases at higher frequency and crosses zero at 440 indicating that the spectral weight redistributes from low to high energies .
such a spectral weight rearrangement is the typical signature for an energy gap formation . from the minimum of @xmath40 we estimate the gap value 2@xmath41 ( 48 mev).@xcite interestingly , singh _ et al .
_ recently performed a gap analysis on the @xmath16-conductivity curve of this material.@xcite there two linear ranges were found in the @xmath42 , and according to the expression @xmath43 two activation energies @xmath44 mev and @xmath45 mev were derived respectively .
the gap value we yield from our optical experiments is in accord with the bigger one.@xcite ( color online ) @xmath46 at selected temperatures . ]
the @xmath1-axis conductivity ( fig .
[ fig : fig3 ] ) varies only slightly with temperatures .
when @xmath47 0 , no drude like behavior is observed and @xmath48 is quite low , indicating that the material is insulating along the @xmath1-axis .
no transfer of spectral weight is induced by the gap opening .
in contrast to iron - pnictides @xmath13fe@xmath5as@xmath5 , the present compound reveals distinctively different electronic properties for the in - plane and out - of - plane directions , i.e. the material is more two - dimensional . in previous optical studies on bafe@xmath5as@xmath5 by wang and
collaborators,@xcite the metallic responses were found for both @xmath49 and @xmath6 with an anisotropy - factor about three.@xcite while two spin - density - wave gaps were revealed in the @xmath0-plane , one of them also opens along the @xmath1-axis , giving evidence for a three - dimensional fermi surface in the electronic structure.@xcite furthermore , the low carrier mobility in infers more localized charge carriers , in the contrast to the itinerant nature of the electronic properties of @xmath13fe@xmath5as@xmath5 .
the reason is the antiferromagnetic arrangement of the mn@xmath19 spins both in the @xmath0-plane and along the @xmath1-axis ( g - type antiferromagnetic order ) , where at low temperature the electron hopping between nearest neighbor sites with opposite spin is suppressed and a strong scattering is induced by the spin disorder at high temperature.@xcite ( color online ) @xmath1-axis conductivity of as a function of frequency from 40 to 1000 measured at selected temperatures . ]
in summary , the optical properties of iron - pnictide analog have been studied in detail using light polarized in the @xmath0-plane and along the @xmath1-axis . within the @xmath0-plane
, the material behaves like a poor metal or semiconductor .
an energy gap can be defined at @xmath7=48 mev .
the @xmath1-axis spectra show a typical insulating behavior overall and no energy gap is found .
this indicates that the is a two - dimensional electron system rather than a quasi three - dimensional , as the fe - pnictides are considered to be .
the @xmath2 and @xmath3 phonon modes have clear to / lo splitting with frequencies differences of 10 - 40 .
a systematic comparison between experimental results and theoretical calculations are present for all the infrared and raman modes .
we thank b. gorshunov and v. i. torgashev for very helpful discussions .
we appreciate n. drichko at johns hopkins university , baltimore md , for performing the raman measurement .
the contributions of j. braun and d. rausch to the experiments are appreciated .
d.w acknowledges her fellowship by the alexander von humboldt foundation .
a.a . acknowledges the support of the german academic exchange service ( daad ) and also hungarian national research fund otka nn76727 . t.k . acknowledges his scholarship by the carl - zeiss - stiftung .
part of the work was funded by the deutsche forschungsgemeinschaft ( dfg ) .
e. kroumova , m. i. aroyo , j. m. perez - mato , a. kirov , c. capillas , s. ivantchev and h. wondratschek , phase transition * 76 * , 155 - 170 ( 2003 ) .
choi , d. wulferding , p. lemmens , n. ni , s. l. budko and p. c. canfield , phys .
b * 78 * , 212503 ( 2008 ) .
a. p. litvinchuk , v. g. hadjiev , m. n. iliev , bing lv , a. m. guloy and c. w. chu , phys .
b * 78 * , 060503(r ) ( 2008 ) . in fig .
[ fig : fig5](a ) and ( b ) , there are ranges ( for example , 170 - 210 ) where @xmath32 shows negative values .
this unphysical phenomenon is caused by the transparent property of sample as we discussed at the beginning of ii - a section .
r. mittal , s. rols , m. zibri , y. su , h. schober , l. chaplot , m. johnson , m. tegel , t. chatterji , s. matsuishi , h. hosono , d. johrendt and th .
brueckel , phys .
b * 79 * , 144516 ( 2009 ) .
m. zbiri , h. schober , m. r. johnson , s. rols , r. mittal , y. su , m. rotter and d. johrendt , phys .
b * 79 * , 064511 ( 2009 ) e. aktrk and s. ciraci , phy . rev .
b * 79 * , 184523 ( 2009 ) .
p. kumar , a. bera , d. v. s. muthu , a. kumar , u. v. waghmare , l. harnagea , c. hess , s. wurmehl , s. singh , b. bchner and a. k. sood , j. phys . : condens .
matter * 23 * ( 2011 ) 255403 . at around 100
we observed a similar but very small hump at 10 k and 50 k. it can be due to the formation of a smaller gap 2@xmath50=12 mev .
as this gap - like feature only appears below 75 k , it should be referred to the second gap @xmath50=6.5 mev measured in @xmath16-resistivity by singh _ et .
al _ in ref.5 .
z. g. chen , t. dong , r. h. ruan , b. f. hu , b. cheng , w. z. hu , p. zheng , z. fang , x. dai , and n. l. wang , phys .
105 * , 097003 ( 2010 ) .
in the case of eufe@xmath5as@xmath5 the anisotropy was about a factor of 8 ( ref . ) . | we have investigated the infrared and raman optical properties of in the @xmath0-plane and along the @xmath1-axis .
the most prominent features in the infrared spectra are the @xmath2 and @xmath3 phonon modes which show clear to - lo splitting from the energy loss function analysis .
all the phonon features we observed in infrared and raman spectra are consistent with the calculated values . compared to the iron - pnictide analog @xmath4fe@xmath5as@xmath5 ,
this compound is much more two - dimensional in its electronic properties . for @xmath6-axis ,
the overall infrared reflectivity is insulating like . within the @xmath0-plane
the material exhibits a semiconducting behavior .
an energy gap @xmath7=48 mev can be clearly identified below room temperature . |
detecting single neutral atoms state selectively is one of the essential ingredients for developing quantum atom optics , atomic physics based quantum technologies and a prerequisite for many quantum information experiments . the first single atom experiments relied on photo ionization for detection @xcite .
nowadays single atom observation is usually performed optically using fluorescence or absorption detection .
cavity assisted detection schemes are predominant and very successful for absorption detection @xcite , but require active stabilization , increasing the complexity and sensitivity to environmental disturbances .
fluorescence detection of single atoms can be very efficient if the atom remains localized , since long integration times allow collecting many fluorescence photons .
this is realized in single atom in a mot @xcite , dipole traps @xcite or trapped single ions @xcite .
fluorescence detection of moving atoms which have a short , finite interaction time with the detector is significantly harder and requires both : supreme background suppression and a high collection efficiency to reach single atom sensitivity . for freely falling atoms
high collection efficiencies have been achieved using a macroscopic mirror setup covering almost 4@xmath2 @xcite .
cavities can be used to enhance the fluorescence signal @xcite . a powerful technique to experiment with ultra cold neutral atoms are atom chips @xcite which employ micro - fabricated wires and electrodes to generate magnetic and electric fields for quantum manipulation of neutral atoms few micrometers above the chip surface @xcite .
many components of integrated matter wave technology have been demonstrated including , combined magnetic / electrostatic traps @xcite , motors and shift registers @xcite , atomic beam splitters @xcite the creation of bose - einstein condensation ( bec ) @xcite , atom interference @xcite , the the integration of optical lattices @xcite .
recently it was demonstrated that atoms in magnetic guides @xcite or on chips can be detected by field ionization near nano - structures @xcite , by mounted optical fibers in absorption @xcite , in fluorescence @xcite , by using integrated waveguides @xcite , or with the help of cavities @xcite .
the latter require active alignment and are technologically very challenging .
miniaturizing the different , highly sophisticated and efficient optics detection methods benefits from scaling with size @xcite and allows integration . in this manuscript
we describe in detail design , implementation and application of our very simple and robust integrated _ fluorescence _ detector @xcite that reaches a single atom detection efficiency of @xmath3 and a high signal to noise ratio ( snr ) of up to 100 without the need for either localization of the atoms or the assistance of a cavity .
the miniaturized detector is based on optical fibers fully integrated on an atom chip @xcite using lithographically fabricated su-8 mounting structures @xcite and allows to examine the statistical distribution of the atoms in a magnetic guide as well as observing non - classical fluorescence emission of single atoms @xcite . in the outlook
we discuss a straight forward improvement of our present set - up that will allow us to build atom counters with close to unit efficiency fully integrated on the atom chip .
details of the integrated atom detector : _ ( a ) _ the multi - mode fiber collects light from a cone ( blue ) determined by its na .
the overlap of this cone with the excitation light ( red ) from the tapered fiber defines the detection region ( yellow ) . _
( b ) _ image showing a cross - section illustrating the mounting of the optical fibers by the undercut su-8 structure which clamps the fiber to the chip surface . _ ( c ) _ schematic view of the lithographically defined su-8 holding structure .
the total length of 6 mm ensures excellent pointing stability , while segmentation of into 2 mm long sections allows for thermal expansion of the underlying chip . to ease insertion of the fibers during construction
a funnel has been included in the design .
, scaledwidth=80.0% ] a very simple way of detecting an atom is to observe its fluorescence .
the basic idea is to drive a ( closed ) transition of the atom to an excited state with an external light field and to detect the spontaneously emitted photons .
the sensitivity of the detection and its fidelity depend on the number of scattered photons , the detection efficiency of the scattered light and the suppression of background noise .
a well designed fluorescence detector collects as much light as possible and simultaneously has a negligible background . in the ideal case
a single detected photon implies that an atom is present in the detection region . in principle
, there is no fundamental limit to the efficiency of a fluorescence detector as long as the atom is not lost from the observation region .
long integration times allows the collecting of many fluorescence photons and efficient discrimination from background .
such a setting was realized for single ions in ion traps @xcite , for a single neutral atom in a dipole trap @xcite or for counting neutral atoms in a magneto - optical trap @xcite .
free neutral atoms are considerably harder to detect because the few scattered photons are difficult to distinguish from background light .
we achieve efficient light collection with small background by employing a fiber optics setup @xcite as shown in figure [ fig : setupdetector](a ) .
our fluorescence detector is built using a single mode tapered lensed fiber which delivers the excitation light and a multi - mode fiber to collect the fluorescence .
the multi - mode fiber is aligned at @xmath4 angle with respect to the tapered lensed fiber .
if no atom is present in the beam focus , very little light is scattered into the fiber .
as soon as an atom is present the multi - mode fiber collects the scattered photons .
such an arrangement is very reminiscent of a confocal microscope .
the excitation of the atoms is at a very small focal spot of the lensed fiber .
the collection fiber collects light preferably from the region around the focal spot .
altogether such an arrangement leads to efficient light collection and superb reduction of background .
for stable and reliable operation both the excitation and detection fibers should be fully integrated into the atom chip .
external mountings are poor alternatives since they might move relative to the chip structures and hence lead to non - reproducible results . the fibers can be glued to the chip by first employing fiber grippers for positioning and alignment .
while this method may result in good alignments it is cumbersome to employ and requires significant time and skill to achieve consistent results .
lithographically fabricated robust and precise fiber mounting structures @xcite are a great help in assembling the detector .
our atom chip holding the micro structures and the detector is fabricated on a @xmath5 m thick silicon substrate .
using optical lithography the micro structures to manipulate the atoms are patterned into an high quality evaporated gold layer @xcite of 2 thickness .
these wire micro structures support current densities exceeding @xmath6a/@xmath7 without the risk of wire destruction . to achieve the precise alignment of the fibers and robust integration on the atom chip we employ a lithographically patterned layer of su-8 on top of the chip structures .
su-8 @xcite is an epoxy based negative photo resist which is typically developed using uv - radiation in the range 365 - 436 nm .
su-8 has high mechanical , chemical and thermal stability .
its specific properties facilitate the production of thick structures with very smooth , nearly vertical sidewalls . with a single su-8 layer a coating film thickness of up to 300 is possible .
once fully developed its glass transition temperature is approximately @xmath8c with a degradation threshold around @xmath9c .
both temperatures are not reached during operation of the atom chip .
the detection fibers are held in su-8 trenches 90 deep and 125 wide , corresponding to the fiber diameter .
careful selection of the fabrication parameters produces slightly undercut sidewalls , as shown in figure [ fig : setupdetector](b ) .
the layer thickness of 90 is chosen to be larger than the fiber radius , which allows the undercut to effectively clamp the fibers to the chip surface .
the total length of the trenches was chosen to be 6 mm , sectioned into three 2 mm long substructures to reduce mechanical stress during thermal expansion ( figure [ fig : setupdetector](c ) ) .
the entry port is funneled to simplify the insertion of the fiber .
the total length guarantees supreme pointing stability , while allowing accurate positioning of the fibers with longitudinal alignment precision of a few ten nanometers , as well as excellent long term stability .
temperature changes and gradients up to @xmath10c resulted in no measurable misalignment of the fibers .
long term stability of more than three years under experimental conditions has been achieved .
in fabricating the fiber mounts a @xmath11 % shrinkage of the su-8 during development has to be taken into account @xcite , leading to a slight variability in the dimensions of the holding structures .
it should be noted here that the variations in fiber diameters are large in comparison .
the diameter of a fiber used in the experiment is specified with a precision of only 1 to a cladding diameter of @xmath12 .
therefore we fabricated a set of 3 parallel mounting structures with slightly different sizes ( 124.5 , 125 and 125.5 ) . to assess the quality and stability of the alignment structures and the glued fibers
, we used the su-8 structures to hold fiber optical resonators @xcite .
the finesse of the resonator strongly depends on losses introduced by misalignment . in our tests we measured a very small additional average loss in the spliced and mounted resonators of @xmath13% which , neglecting other additional losses , corresponds to a maximal transversal misalignment of less than @xmath14 nm .
details of the atom chip setup and detector : _ ( a ) _ schematic layout of the atom chip .
atoms are initially trapped in a magnetic trap generated by a z - shaped wire ( red ) .
a magnetic guide ( yellow ) transports the atoms to the focus of the tapered lensed fiber . _
( b ) _ basic layout of the detector experiment and the atom chip .
light from a frequency stabilized source is delivered via a tapered lensed fiber to the interaction region .
part of the light is split off in a commercial fiber beam splitter and directed to a pin - type photo diode ( pd ) for intensity stabilization ( pid ) using an acousto - optical modulator ( aom ) which can also be used as a fast light switch .
the polarization of the excitation light is adjusted using mechanical fiber paddles , optimizing for maximal fluorescence signal .
the light scattered by the atoms is collected by a multi - mode fiber that guides the light to one single photon counting module ( spcm ) for high efficiency atom detection or two spcms in hanbury brown and twiss configuration for correlation measurements . ] with a diameter of the mounted fibers of 125 , the detection region is situated 62.5 above the chip surface . to avoid stray light from the mounting structures when
loading the atoms in the mirror mot , the detector is built 5.5 mm away from the chip center . a magnetic guide on the chip transports the atoms to the detector ( see figure [ fig : setupgeneral ] ) .
the fibers mounted on the atom chip are then connected to the outside of the vacuum chamber via an ultra - high vacuum ( uhv ) compatible feed - through @xcite .
the whole setup can be baked to @xmath10 c and achieves uhv below @xmath15 mbar . to excite the atoms to fluorescence the tapered lensed fiber delivers the light to a focal spot in the guide potential minimum
our tapered fibers feature a focal length of @xmath16 with a focal diameter of @xmath17 . for the f=2@xmath18f=3 transition in @xmath19 saturation intensity
is reached in the focus at a power of @xmath20 pw .
the focus of the tapered fiber defines the interaction region as shown in figure [ fig : setupdetector](a ) and is located @xmath21 ( half a fiber diameter ) above the chip surface .
the excitation light is generated by an external cavity diode laser ( ecdl ) frequency stabilized to the f=2@xmath18f=3 transition with a precision better than 1 mhz , smaller than the natural line width of the transition .
the probe laser is frequency stabilized to a reference laser by means of a frequency offset ( beating ) lock .
this locking scheme enables us to shift the lock point by several hundred mhz during operation , allowing to probe the atomic ensemble over the full range of the rb - d2 transition . for additional intensity stabilization part of the light
is split off in a commercial fiber beam splitter and directed to a pin - type photo diode ( pd ) ( see figure [ fig : setupgeneral](b ) ) .
the pd signal is fed into custom built proportional - integral - differential lock electronics ( pid ) stabilizing the beam power with an uncertainty below 10 pw to the set value by controlling the output level of an acousto - optical modulator ( aom ) used as a fast light switch .
the polarization of the excitation light can be rotated employing mechanical fiber paddles . using a mechanical fiber
splice the stabilized excitation light is coupled into the single - mode tapered lensed fiber guiding the light into the vacuum chamber to the detection region .
the light scattered by the atoms is collected by a multi - mode fiber that guides the light to a single photon counting module ( spcm ) . to protect the spcm and to filter non - resonant background
light a narrow interference filter is included in the beam path ( 3 nm fwhm ) .
the multi - mode fiber has a limited field of view , given by its numerical aperture of na@xmath22 and the core diameter of @xmath23 . only light hitting the core at angles up to maximally @xmath24 will be collected .
consequently only emission from atoms within the dark blue cone marked in figure [ fig : setupdetector](a ) is collected with the full na .
hence the region of maximum collection efficiency for the atomic signal is given by the intersection of the excitation mode field ( red ) and the full na cone .
this defines the detection region , which is depicted yellow in figure [ fig : setupdetector](a ) .
the height of the detection region is automatically aligned to the excitation region since the outer diameters of the fibers match . by aligning the mm fiber such that the collection cone defined by the numerical aperture overlaps with the excitation region we reach both a highly selective excitation of the atoms and a matched collection region .
the angle between the tapered ( excitation ) and the multi - mode ( detection ) fiber has been chosen to be @xmath25 .
this leads to an excellent geometrical suppression of stray light . of each nano watt of excitation light
only about 30 photons per second are collected by the detection fiber , corresponding to a stray light suppression of better than @xmath26 .
the detection setup has been positioned such that both fibers are at @xmath27 to the guide to avoid blocking of the guide . for our collection fiber ( na=0.275 ) a maximum photon collection efficiency of @xmath28% can be determined . together with the finite efficiency @xmath29 of the spcm and optical losses this leads to a total photon detection efficiency of @xmath30 . in our experiments the atom chip @xcite serves as the experimental platform for efficient cooling and trapping of the atoms and for transporting them in a magnetic guide to the detection region .
our experimental procedure is described in @xcite .
the heart of our setup is a hybrid macroscopic - microscopic atom chip assembly .
it holds the macroscopic wire structures used to pre - cool and capture the atoms in the primary phase of the experiment as well as the appropriately designed wire micro structures on the atom chip needed for trapping and manipulating the atoms , including the atom detector .
the experimental procedure starts typically with more than @xmath31 @xmath32rb atoms accumulated in a mirror mot .
the atoms are subsequently optically pumped into the state and transferred to a ioffe - pritchard type magnetic trap on the atom chip generated by a z - shaped wire on the chip surface .
the procedure is optimized by a genetic algorithm @xcite .
the atoms in the magnetic trap are released into a magnetic guide ( l - shaped wire ) where the atomic cloud can expand toward the detector .
the position of the magnetic guide above the chip surface is aligned with the focus of the tapered lensed fiber at the detection region by adjusting the current through the chip wire and the strength of the external magnetic field @xcite .
when the atoms are released from the initial z trap into the magnetic guide they thermally expand towards the detector . in the focus of the lensed fiber
the atoms are excited by a laser operating near one of the the rubidium d2 transitions ( for example tuned to the @xmath33 to @xmath34 transition ) and the fluorescence photons are collected by the multi - mode fiber and detected by a single photon counting module ( spcm ) , their arrival times are recorded with 1 ns resolution .
the experiment is repeated several times to measure the photon statistics .
a typical signal is shown in fig .
[ fig : varandmean]a . in the current experiment
the phase space density in the magnetic trap and guide is always less than @xmath35 .
one observation in our experiments is that we see the full length of the atom pulse guided in the magnetic guide ( fig .
[ fig : varandmean](a ) ) even though the resonant excitation light of the detector is always on .
these measurements show that in our experiments effects of stray excitation light on the guided atoms can be neglected .
this is quite remarkable since magnetic traps are extremely sensitive to light scattering . on average , scattering of a little more than
a single photon is sufficient to pump the atom into a magnetically un - trapped state , removing it from the magnetic guide .
figure [ fig : varandmean](a ) shows the atomic signal together with its variance , calculated from 600 consecutive measurements .
the shape can be described by a 1d maxwell - boltzmann distribution given by the temperature of the atoms when the reflection of the part of the atom cloud that starts to expands towards the closed end of the guide is also taken into account .
the measured temperature of @xmath36 agrees with independent measurements via time of flight methods
. has been used .
inset : even after 2 s the signal is still significantly above the background level of 0.06 .
+ _ ( b ) _ ratio / : as soon as the atoms arrive , / exhibits a sudden jump to a super - poissonian value of @xmath37 .
even though @xmath38 remains constant , the ratio decreases with atomic density as background becomes more important .
the red line gives a fit according to equation [ eq : varmean1 ] over more than two orders of magnitude in atom density .
note that the initial overshoot is an artifact of the analysis because around peak the steep sloes can not be neglected during the 200 integration time .
+ _ ( c ) _ as a function of the mean recorded count rate .
an analysis according to equation [ eq : varmean1 ] ( red line ) yields a signal strength of @xmath39 counts per atom ( integration time @xmath40 ) .
the dashed line corresponds to poissonian statistics .
+ _ ( d ) _ the signal strength @xmath38 remains constant and is independent of atom density over three orders of magnitude . here
@xmath38 is evaluated for a shorter @xmath41 and shows a reduced value of @xmath42 ( compare figure [ fig : efficiency ] ) . + _ ( e ) _ @xmath38 as function of the probe beam detuning reveals an optimal blue detuning of 3 mhz .
the dashed red line is a model based on ref.@xcite . ]
the fiber - based detector presented here has exceptionally low background count rate of only @xmath43 counts per seconds ( cps ) at 1nw resonant probe light .
the dominating contribution are the dark counts ( @xmath44 cps ) of the employed photon counter ( perkin - elmer , spcm - aqr-12 ) . despite the proximity of the atom chip surface ( at a distance of 62.5 ) the influence of stray probe - light
is essentially eliminated with a suppression of better than @xmath26 .
1 nw excitation light contributes only @xmath45 cps to the background .
this excellent suppression of stray light is achieved by the fact that the a tapered lensed fiber delivers excitation light efficiently to a very small detection region and the multi - mode fiber , mounted at an angle of 90 degree , selectively collects the fluorescent photons from this small volume ( fig .
[ fig : setupdetector]a ) .
basically all the background light originating from scattering on the close by chip surface or the vacuum chamber is blocked very efficiently , because it is not matched to the guided modes in the collection fiber . in this sense
the setup has characteristics similar to a confocal microscope , with the difference that the involved point spread functions are different .
other background sources are stray light from the outside and the glow of the hot dispensers .
an interference filter efficiently suppresses this remaining background to below 10 cps .
( see fig .
[ fig : setupgeneral]b ) the excellent stray light suppression and overall small background level allows high fidelity detection of a single atom and accurate measurements of photon and atom statistics .
an important parameter of the detector is the total number of photons that can be scattered by the guided atoms before they are lost for the detection process .
we discuss this here for detecting @xmath19 atoms in f=2 using the f=2@xmath18f=3 transition .
the total number of scattered photons can be estimated from the ratio of the fluorescence counts when exciting on the closed f=2@xmath18f=3 transition to the fluorescence counts on the open transition f=2@xmath18f=1 .
exciting on the later an atom scatters slightly more than one photon before being optically pumped into the other hyperfine ground state , where it remains dark .
we find that that in our detector each atom scatters @xmath46 photons when excited on the f=2@xmath18f=3 transition .
combining this number with the overall photon detection efficiency given above , we estimate that our detector sees in average @xmath47 photon from each atom passing by .
this number was confirmed from independent global atom number measurements using absorption imaging .
the maximum signal per atom is generated not for resonant illumination , as might be expected , but for a blue detuning of 3 mhz of the exciting light field as shown in figure [ fig : varandmean](e ) .
looking more into the details , the atoms accumulate a time dependent detuning @xmath48 during interaction with the probe field due to photon recoil induced doppler shift .
an initial blue detuning allows pushing the atoms trough resonance to maximize the total number of scattered photons during interaction @xcite .
when an atom arrives at the detector , it absorbs and then re - emits photons .
a few of these photons are counted by the spcm .
this photon scattering strongly disturbs the atom . besides the photon recoil it will pump the atom into an un - trapped state and eventually into a different hyper fine ground state which is not excited by the probing light .
consequently after a time @xmath49 the atom will either leave the detection region and / or stop scattering photons . during detection
each atom emits a photon burst , and on average @xmath38 photons are detected per atom .
these photon bursts lead to a super - poissonian statistics for the detected photons ( fig .
[ fig : varandmean](b ) ) .
as we will show below we can relate the variance and the mean of the photon counts directly to the average number of photons detected per atom ( @xmath38 ) .
a coherent light source of constant mean photon number creates a poisson distributed photon stream with variance equal to the mean , @xmath50 and hence @xmath51 the superpoissonian photon statistics for the atom detection events can be seen in fig .
[ fig : varandmean](b ) by the jump of / @xmath1 1 as soon as atoms reach the detector .
the statistical atom number fluctuations lead to increase in the photon noise .
the atom number follows its own statistical distribution @xmath52 around a mean value and mandel s formula @xcite can be used to describe the photon count distribution in the presence of atoms .
the probability @xmath53 to measure @xmath54 counts is then expressed by folding the conditional probability @xmath55 to measure @xmath54 counts in the presence of @xmath56 atoms with the atom number probability distribution @xmath57 . @xmath58
the ratio / can then be calculated to @xmath59 hence the ratio of the variance of the photon counts to its mean value deviates from the poissonian value 1 by the number of photon counts per atom @xmath38 times the ratio of variance to mean of the atom number .
if in addition a poissonian background @xmath60 with @xmath61 is taken into account equation [ eq : varmean1 ] is modified and the photon statistics can be expressed as @xmath62 it is worth noting that this relation is independent of the actual atom number . for a poissonian atom number distribution , i.e. @xmath63 , the signal strength @xmath38 can be measured as the slope of the relation variance versus mean ( see figure [ fig : varandmean](c ) ) .
conversely , if @xmath38 is known equation [ eq : varmean2 ] allows examining the statistical distribution of the atoms by measuring the photon count distribution .
figures [ fig : varandmean](b ) and ( c ) show the measured ratio / respectively versus together with a fit according to equation [ eq : varmean2 ] . the signal strength is determined to @xmath64 counts per atom ( cpa ) , independent of atomic density over three orders of magnitude as demonstrated in figure [ fig : varandmean](d ) . a detailed analysis of the distributions
show that an integration time of @xmath65 is needed to collect the full signal .
specific applications will benefit from shorter integration times .
figure [ fig : efficiency](a ) illustrates that @xmath38 drops rather slowly when reducing the integration time , while the mean number of background counts is a linear function of @xmath66 .
for @xmath67 the integrated signal is @xmath68 cpa and even for @xmath69 we still find @xmath42 cpa .
.single atom detection efficiency @xmath70 and false detection probability @xmath71 as function of the integration time for the current system ( na 0.275 , @xmath30 , @xmath72 cps ) .
numbers in brackets denote the signal to noise ratio using a low noise spcm ( @xmath73 cps ) .
[ tab : detefficiency ] [ cols= " > , < , < , > , > " , ]
to conclude , we have built and evaluated an optical fiber - based atom detector which is fully integrated on an atom chip , alignment free by fabrication , and mechanically very robust and capable of state selective single atom detection at an efficiency of @xmath3 by counting fluorescence photons .
it enables spatially and spectral highly selective probing while at the same time offering extremely robust operation .
the total background is dominated by the dark count rate of the employed spcms .
high efficiency detection is possible as well as low noise , high bandwidth detection . at 20 integration time
@xmath74 single atom detection can be performed at an efficiency of @xmath75 .
the disadvantage of fluorescence detection is the destructive nature of the process , which our detector shares with photon detection .
our detector can therefore be best characterized as the atom optic equivalent of an avalanche photo diode type spcm .
low noise , high efficiency and insensitivity to stray light is achieved using fiber optics to create very selective excitation of the atoms in a small , matched observation volume .
the detection efficiency is currently limited by the numerical aperture of the multimode collection fiber . a straightforward substitution of the employed na=0.275 fiber by a commercially available fiber with na=0.53 increases the photon collection to @xmath38 = 4.5 counts / atom and the single atom detection efficiency to 95% at 50 khz bandwidth .
employing two collection fibers @xmath38 = 9 cpa and @xmath76 at integration times below @xmath77 can be achieved . with these improvements ,
full atom counting becomes feasible .
the high efficiency , signal - to - noise ratio and bandwidth of our integrated detector make it suitable for many physical systems where only a few photons can be scattered .
a prime example would be detecting trapped cold molecules . with its extremely low sensitivity to stray light ,
our detector is well suited for studies of correlated atomic systems and scalable quantum experiments on a single - atom or molecule level .
we thank a. haase , and m. schwarz for help in the early stages of the experiment , and k .- h .
brenner , and i. bar - joseph for help in the fabrication .
we gratefully acknowledge financial support from the landesstiftung baden - wrttemberg , the european union ( hip , chimono ) , the austrian nano - initiative ( platon - nap ) and the fwf . the atom chip shown in this work was fabricated at the the braun submicron center at weizmann institute of science , rehovot , il . | we describe a robust and reliable fluorescence detector for single atoms that is fully integrated into an atom chip .
the detector allows spectrally and spatially selective detection of atoms , reaching a single atom detection efficiency of 66 % .
it consists of a tapered lensed single - mode fiber for precise delivery of excitation light and a multi - mode fiber to collect the fluorescence .
the fibers are mounted in lithographically defined holding structures on the atom chip .
neutral @xmath0rb atoms propagating freely in a magnetic guide are detected and the noise of their fluorescence emission is analyzed .
the variance of the photon distribution allows to determine the number of detected photons / atom and from there the atom detection efficiency .
the second order intensity correlation function of the fluorescence shows near - perfect photon anti - bunching and signs of damped rabi - oscillations . with simple improvements one can boost the detection efficiency to @xmath1 95 % . |
inhomogeneities in the large - scale matter distribution can in many ways affect the light signals coming from very distant objects . these effects need to be understood well if we want to map the expansion history and determine the composition of the universe to a high precision from cosmolgical observations .
in particular the evidence for dark energy in the current cosmological concordance model is heavily based on the analysis of the apparent magnitudes of distant type ia supernovae ( sne ) @xcite .
inhomogeneities can affect the observed sne magnitude - redshift relation for example through gravitational lensing , in a way which essentially depends on size and composition of the structures through which light passes on its way from source to observer .
the fundamental quantity describing this statistical magnification is the lensing probability distribution function ( pdf ) .
it is not currently possible to extract the lensing pdf from the observational data and we have to resort to theoretical models .
two possible alternatives have been followed in the literature . a first approach ( e.g. ref .
@xcite ) relates a `` universal '' form of the lensing pdf to the variance of the convergence , which in turn is fixed by the amplitude of the power spectrum , @xmath0 .
moreover the coefficients of the proposed pdf are trained on some specific n - body simulations . a second approach ( e.g. ref .
@xcite ) is to build _ ab - initio _ a model for the inhomogeneous universe and directly compute the relative lensing pdf , usually through time - consuming ray - tracing techniques . the flexibility of this method is therefore penalized by the increased computational time . in ref .
@xcite we introduced a stochastic approach to cumulative weak lensing ( hereafter sgl method ) which combines the flexibility in modeling with a fast performance in obtaining the lensing pdf .
the speed gain is actually a sine - qua - non for likelihood approaches , in which one needs to scan many thousands different models ( see ref .
the sgl method is based on the weak lensing approximation and generating stochastic configurations of inhomogeneities along the line of sight .
the major improvements introduced here are the use of a realistic halo mass function to determine the halo mass spectrum and the incorporation of large - scale structures in the form of filaments .
the improved modeling together with the flexibility to include a wide array of systematic biases and selection effects makes the sgl method a powerful and comprehensive tool to study the impact of lensing on observations .
we show in particular that the sgl method , endowed with the new array of inhomogeneities , naturally and accurately reproduces the lensing pdf of the millenium simulation @xcite .
we also study a simple selection effect model and show that selection biases can reduce the variance of the observable pdf . such reduction could at least partly cancel the opposite effect coming from large scale inhomogeneities , masking their effect on the observable pdf .
we also show how a jdem - like survey could constrain the lensing pdf relative to a given cosmological model . along with this paper , we release an updated version of the package , which is a simple and very fast mathematica implementation of the sgl method @xcite . this paper is organized as follows . in section [ sup ]
we introduce the cosmological background , the generic layout of inhomogeneities and review the basic formalism needed to compute the weak lensing convergence . in section [ sec : inhoprop ]
we derive the halo mass function and the halo density profiles and define the precise modeling of filaments .
in section [ hmf ] we present the revised and extended sgl method .
the exact discretization of the model parameters , which is a crucial step in the sgl model building , is explained in section [ mbin ] and in section [ confi ] we explain how the realistic structures where halos are confined in filaments are modelled in the sgl method .
section [ results ] contains our numerical results including the comparison with the cosmology of the millennium simulation @xcite and , finally , in section [ conco ] we will give our conclusions .
we consider homogeneous and isotropic friedmann - lematre - robertson - walker ( flrw ) background solutions to einstein s equations , whose metric can be written as : @xmath1 \ , , \ ] ] where @xmath2 and @xmath3 & \ ; k<0 \\ \end{array}\right.\ ; , \ ] ] where @xmath4 is the spatial curvature of any @xmath5slice . in particular we will focus on @xmath6cdm models whose hubble expansion rate depends on redshift according to : @xmath7 where @xmath8 is given by : @xmath9 here @xmath10 could be taken to follow _
e.g. _ the parameterization @xcite : @xmath11 for which : @xmath12 for a constant equation of state @xmath13 , the latter reduces to @xmath14 .
@xmath15 , @xmath16 and @xmath17 are the present - day density parameters of dark energy , matter and radiation and @xmath18 represents the spatial - curvature contribution to the friedmann equation . introducing the hubble radius @xmath19 and the spatial - curvature radius @xmath20 , we have at any time the relation @xmath21 .
we will also need the matter and dark energy density parameters at a given time or redshift : @xmath22 throughout this paper , the subscript @xmath23 will denote the present - day values of the quantities .
for example the critical density today is @xmath24 where @xmath25 km s@xmath26 mpc@xmath26 , while @xmath27 is the critical density at any time . substituting in eq .
( [ hhh ] ) @xmath28 and @xmath29 we obtain the equation we have to solve in order to find the time evolution of the scale factor @xmath30 , the only dynamical variable in an flrw model .
we fix the radiation density parameter to @xmath31 .
moreover @xmath32 , that is , dark and baryonic matter contribute together to the total matter density .
the line - of - sight and transverse comoving distances are : @xmath33 from which we find the angular and luminosity distances together with the distance modulus : @xmath34 for scales smaller than the spatial - curvature radius @xmath35 , @xmath36 and we can use the euclidean geometry . the aim of this paper is to compute statistical weak lensing corrections to the measured light intensities , generated by the inhomogeneous matter distribution along the line of sight to a distant object .
the basic quantity we need to model is the matter density contrast @xmath37 : @xmath38 where the lowercase @xmath39 indicates the local and inhomogeneous matter field while @xmath40 is the time dependent average mass density .
the density contrast directly enters the expression for the weak lensing convergence @xcite : @xmath41 where @xmath42 is the co - moving position of the source and the integral is along an unperturbed light geodesic . the density
@xmath43 is the constant matter density in a co - moving volume and we defined the auxiliary function @xmath44 which gives the optical weight of an inhomogeneity at the comoving radius @xmath45 .
the convergence is related to the shift in the distance modulus by : @xmath46 where @xmath47 is the net magnification and the second - order contribution of the shear has been neglected @xcite .
it is obvious that an accurate statistical modeling of the magnification pdf calls for a detailed description of the inhomogeneous mass distribution . in this paper
we will significantly improve the modeling of the inhomogeneities from our previous work @xcite , where only single - mass spherical overdensities were considered .
first of all , we improve the modeling of these halos " by using a realistic halo mass function @xmath48 , which gives the fraction of the total mass in halos of mass @xmath49 at the redshift @xmath50 .
the function @xmath48 is related to the ( comoving ) number density @xmath51 by : @xmath52 where we defined @xmath53 as the number density of halos in the mass range @xmath54 .
the halo function is by definition normalized to unity @xmath55 the idea is of course that halos describe large virialized mass concentrations such as large galaxies , galaxy clusters and superclusters .
not all matter is confined into virialized halos however . moreover , only very large mass concentrations play a significant role in our weak lensing analysis : for example the stellar mass in galaxies affects the lensing pdf only at very large magnifications @xcite where the pdf is close to zero .
therefore the fraction of mass @xmath56 concentrated in large virialized halos can be defined by introducing a lower limit to the integral ( [ eq : fmnorm ] ) : @xmath57 only mass concentrations with @xmath58 are treated as halos . the remaining mass is divided into a family of large mass , but low density contrast objects with a fraction @xmath59 and a uniform component with a fraction @xmath60 , such that @xmath61 the low contrast objects are introduced to account for the filamentary structures observed in the large scale structures of the universe . in our analysis they will be modeled by elongated objects with random positions and orientations .
the mass in these objects can consist of a smooth unvirialized ( dark ) matter field and/or of a fine `` dust '' of small virialized objects with @xmath62 . in weak lensing this distinction does not matter because small halos act effectively as a mean field , with a sizeable contribution only at very large magnifications ( see comment above about stellar mass ) . for later use
we define @xmath63 which gives the total mass fraction not in large virialized halos .
if we only consider virialized masses larger than companion galaxies ( @xmath64 ) , then typical values for the concordance model are @xmath65 at @xmath66@xcite and @xmath67 at @xmath68 , with a weak dependence on the particular @xmath48 used . although @xmath69 is sensitive on @xmath70 , the lensing pdf depends only weakly on the cut mass .
moreover , halo functions obtained from n - body simulations are valid above a mass value imposed by the numerical resolution of the simulation itself and so the use of @xmath70 is also necessary in this case @xcite . to obtain an as accurate modeling as possible , we will use realistic mass functions and spatial density profiles for the halo distribution .
there is less theoretical and observational input to constrain the mass distribution of the filamentary structures or their internal density profiles .
we will therefore parametrize the filaments with reasonable assumptions for their lengths and widths and by employing cylindrical nonuniform density profiles .
mpc thick slice generated by our stochastic model .
the shaded disks represent clusters and the shaded cylinders the filamentary dark matter structures . only large clusters are displayed.,width=8 ] our modeling allows treating the two families of inhomogeneities independently .
both can be given random spatial distributions , or alternatively all halos can be confined to have random positions in the interiors of randomly distributed cylinders .
the latter configuration more closely resembles the observed large scale structures , and an illustration created by a numerical simulation using the turbogl package is shown in fig .
[ slice ] . in section [ discu ]
we will discuss the power spectrum and the large - scale correlations of our model universe .
we can now formally rewrite eq .
( [ contra1 ] ) for our model universe in which the density distribution is given by @xmath71 : @xmath72 where the index @xmath73 labels all the inhomogeneities , we used eq .
( [ eq : sumrule ] ) and we defined @xmath74 , so that both virialized halos and the unvirialized objects are described by a generic reduced density profile @xmath75 which satisfies the normalization @xmath76 .
the lensing convergence we are interested in can now be written as @xmath77 where the positive contribution due to the inhomogeneities ( halos and low contrast objects ) is @xmath78 the also positive contribution due to the uniformly distributed matter is @xmath79 and the negative empty beam convergence is : @xmath80 a light ray that misses all the inhomogeneities @xmath81 will experience a negative total convergence @xmath82 , which gives the maximum possible demagnification in a given model universe . in an exactly homogeneous flrw model the two contributions @xmath83 and @xmath84
cancel and there is no net lensing . in an inhomogeneous universe ,
on the other hand , a light ray encounters positive or negative density contrasts , and its intensity will be magnified or demagnified , respectively .
the essence of the sgl method is finding a simple statistical expression for the probability distribution of the quantity @xmath83 in the inhomogeneous universe described above .
for a discussion about the validity of the weak lensing approximation within our setup see ref .
@xcite , where it was shown that the error introduced is @xmath85% .
we stress again that the lensing caused by stellar mass in galaxies is negligible in the weak lensing regime @xcite and so we can focus on just modeling the dark matter distribution in the universe .
also , we will treat the inhomogeneities as perturbations over the background metric of eq .
( [ metric ] ) . in particular we will assume that redshifts can be related to comoving distances through the latter metric .
see ref .
@xcite for a discussion of redshift effects . in the next section
we will give the accurate modeling of the inhomogeneities .
we begin by introducing the halo mass function and the detailed halo profiles , and then move on to describe the precise modeling of the cylindrical filaments .
we begin by explaining our dark matter halo modeling .
the two main concepts here are the halo mass function @xmath48 giving the normalized distribution of the halos as a function of their mass and redshift , and the dark matter density profile within each individual halo .
both quantities are essential for an accurate modeling of the weak lensing effects by inhomogeneities
. we shall begin with the halo mass function introduced above in eq .
( [ eq : halomassf ] ) .
the halo mass function acquires an approximate universality when expressed with respect to the variance of the mass fluctuations on a comoving scale @xmath45 at a given time or redshift , @xmath86 . relating the comoving scale @xmath45 to the mass scale by @xmath87
, we can define the variance in a given mass scale by @xmath88 .
the variance @xmath86 can be computed from the power spectrum : @xmath89 where @xmath90 is the fourier transform of the chosen ( top - hat in this work ) window function and the dimensionless power spectrum extrapolated using linear theory to the redshift @xmath50 is : @xmath91 here @xmath92 is the spectral index and @xmath93 is the amplitude of perturbations on the horizon scale today , which we fix by requiring : @xmath94 where the value of @xmath0 is estimated by cluster abundance constraints @xcite .
@xmath95 is the linear growth function which describes the growth speed of the linear perturbations in the universe .
fit functions for @xmath96 could be found , _
e.g. _ , in ref .
@xcite , but it can also be easily solved numerically from the equation : @xmath97 \nonumber \\ & -&{3 \over 2 } { d(z ) \over ( 1+z)^{2 } } \ , \omega_{m}(z)=0 \,,\end{aligned}\ ] ] where we have neglected the radiation . as usual
, we normalize @xmath96 to unity at the present time , @xmath98 .
finally , for the transfer function @xmath99 we use the fit provided by the equations ( 28 - 31 ) of ref .
@xcite , which accurately reproduces the baryon - induced suppression on the intermediate scales but ignores the acoustic oscillations , which are not relevant for us here .
with @xmath100 given , we can now define our halo mass function .
several different mass functions have been introduced in the literature , but here we will consider the mass function given in eq .
( b3 ) of ref .
@xcite : @xmath101 which is valid in the range @xmath102 . because of the change of variable , @xmath103 is related to our original definition of @xmath104 by : @xmath105 the mass function of eq . ( [ jenk ] )
is defined relative to a spherical - overdensity ( so ) halo finder , and the overdensity used to identify a halo of mass @xmath49 at redshift @xmath50 is @xmath106 with respect to the mean matter density @xmath107 .
the so finder allows therefore a direct relation between halo mass @xmath49 and the radius @xmath108 : @xmath109 the subscript @xmath110 will denote the proper values of otherwise comoving quantities throughout this paper .
for example , the comoving halo radius @xmath111 is related to the proper value by @xmath112 .
a direct relation between the mass and the radius of a cluster is necessary for our sgl method .
this is why we prefer mass functions based on so finders over mass functions based on friends - of - friends halo finders ; for the latter an appropriate @xmath113 to be used in eq .
( [ mdel ] ) is not directly available .
moreover , as shown in ref .
@xcite , the so(180 ) halo finder gives a good degree of universality to the mass function .
another mass function which manifests approximate universality with the so(180 ) halo finder @xcite is provided by sheth & tormen in ref .
@xcite . with the halo mass function given
the only missing ingredient is the halo profile which , as said after eq .
( [ contra2 ] ) , we describe by means of the reduced halo profile @xmath114 .
we stress that our halo of a given mass @xmath49 and redshift @xmath50 is an avererage representative of the total ensemble of halos , which in reality have some scatter in the density profiles .
we will focus on the navarro - frenk - white ( nfw ) profile @xcite : @xmath115 here the scale radius is related to the halo radius by @xmath116 , where @xmath117 is the concentration parameter to be defined below . by integrating equation ( [ eq : nfwprofile ] )
one finds the total mass @xmath118 $ ] which , when combined with eq .
( [ mdel ] ) , gives : @xmath119 this relation fixes @xmath120 once the concentration parameter is given .
the reduced halo profile @xmath114 in comoving coordinates is then : @xmath121^{-1 } , \ ] ] where the @xmath122-dependence ( as well as @xmath49-dependence ) arises through the redshift dependence of @xmath111 ( eq . ( [ mdel ] ) ) and @xmath117 ( eq . ( [ duffy ] ) ) .
the last ingredient needed to fully specify the nfw profile is a relation between the concentration parameter and the halo mass at a given redshift .
for the @xmath123cdm model we will use the following fit obtained from numerical simulations @xcite satisfying wmap5 cosmology @xcite : @xmath124 where the pivot mass scale is @xmath125 .
apart from the nfw profile , alternative simple density profiles of interest include : @xmath126 \over ( 2 \pi)^{3/2 } ( r/3)^{3 } } \phantom{ggg } \textrm{gaussian , } \label{gau } \\
\varphi_{h } & = & { 1 \over 4 \pi r^{2 } r } \phantom{ggggggggggggggggggi } \textrm{sis , } \label{sis}\end{aligned}\ ] ] where @xmath111 is again given by eq .
( [ mdel ] ) .
the normalization factor in ( [ gau ] ) is necessary to have the correct volume integral of unity and is due to the restriction of the profile between 0 and @xmath111 . in these cases
there are no further parameters to be fixed .
we will model dark matter filaments by cylindrical , non - uniform low density objects .
let us stress that throughout this chapter by a
filament " we refer only to the smoothly distributed dark matter component of a real filament ; we shall later explain how virialized halos can be embedded within these objects to form a more realistic model of the true filamentary structures in the full sgl framework . keeping the low contrast component and
the halos as separate entities in our evaluation is convenient for the modeling and explicitly avoids double counting . using eqs .
( [ eq : halomassf ] ) and ( [ eq : sumrule ] ) we can obtain the following equation which relates the comoving filament density @xmath127 and the filament mass @xmath128 to the fraction @xmath69 of mass not confined in virialized halos : @xmath129 where the quantity @xmath130 defines the fraction of the total unvirialized mass which is confined in filaments : @xmath131 . because of the ongoing halo formation @xmath69 decreases with time and we shall simply assume that @xmath132 is a constant .
we will assume that all filaments have the same mass , but this assumption can easily be relaxed later by introducing a filament mass function @xcite .
we are also neglecting merging of smaller filaments into larger ones , whereby the comoving number density of filaments @xmath127 remains a constant .
this assumption seems to be supported by numerical simulations @xcite at least at low redshifts . from eq
( [ eq : fila ] ) it then follows that the filament mass @xmath128 is not constant , but it gradually decreases as more and more halos virialize .
( the total mass of the full filamentary structure which includes the halos remains on average constant . )
now we need to specify the precise geometry , dimensions and time - dependent density profile @xmath133 for our filaments .
first , we will relate the filament mass to its length and radius at present day according to : @xmath134 where @xmath135 and @xmath136 are the present - day filament proper radius and length and @xmath137 is the average overdensity of the filament with respect to the flrw matter density .
consistent with our assumptions of neglecting filament mergers and using a constant comoving filament density , we will assume that filaments have a constant comoving length .
this is again in accordance with the generic picture of a web of filaments of constant comoving length condensing with time seen in numerical simulations .
finally , for simplicity , we will choose our filaments to have a constant proper radius .
summarizing , we will use : @xmath138 the simplest assumption would be to take filaments to have uniform density . in this case
the reduced density profile is simply @xmath139 .
however , we will also consider a profile which is uniform along the length of the filament @xmath140 , but which has a gaussian profile in the radial direction . in this case
@xmath141 \over l \ , 2 \pi ( r/3)^{2 } } \,,\ ] ] where @xmath142 and @xmath143 , and @xmath144 otherwise . finally , let us compute the mass of the `` dressed '' filaments which comprises the `` bare '' filaments until now discussed together with the halos ( to be embedded in section [ confi ] ) . it is easy to find that : @xmath145 \,,\ ] ] where also the contribution from the uniformly distributed matter is included .
( [ dressed ] ) is valid if all halos are embedded within the filaments and has to be slightly modified if different levels of confinement are considered .
this completes our description of the _ objects _ in our model .
next we shall have to understand how to simulate their effect on the weak lensing properties of the complete inhomogeneous model universe .
in this section we will review and extend the stochastic gravitational lensing ( sgl ) method of ref .
@xcite for computing the probability distribution function of the lens convergence @xmath146 in the presence of inhomogeneities .
the basic quantity to evaluate is the inhomogeneity - induced part @xmath147 in eq .
( [ eq : kappaih ] ) : @xmath148 where @xmath149 .
we wish to obtain a probabilistic prediction for this quantity along a line of sight to a source located at @xmath150 , through a random distribution of halos and filaments .
the problem could be solved by constructing a large comoving volume of statistically distributed objects such as the one shown in fig .
[ slice ] , and then computing the distribution of @xmath83 along random directions in this space . however , an equivalent and computationally much more efficient approach is to construct random realizations of object locations along a fixed geodesic and compute the lensing pdf from a large sample of such realizations .
let @xmath151 refer to a particular _ realization _ of the integral in eq .
( [ eq : kappa3b ] ) and denote the resulting convergence by @xmath152 . because of the finite size of the halos and filaments , only a finite number @xmath153 of objects intercept the geodesic and contribute to the sum .
moreover , because all objects are small compared to a typical comoving distance to the source @xmath154 , the function @xmath155 is , to a good approximation , a constant @xmath156 for each individual object .
similarly , one can assume that @xmath157 with @xmath158 , and so one finds : @xmath159 where @xmath160 is the surface mass density of the @xmath73-th intercepting object in the realization .
next divide the geodesic into @xmath161 subintervals with centers at @xmath162 and with ( possibly variable ) length @xmath163 , such that @xmath164 holds within each interval . in this way we can write @xmath165 where @xmath166 labels all objects encountered by the geodesic within the comoving length bin @xmath167 , such that @xmath168 . at this level eq .
( [ eq : realization2 ] ) still consists of a sum over all _ objects _ along a given path .
next we categorize these objects into different classes depending on the parameters that define the surface density @xmath169 . for spherical halos of a given @xmath170-profile @xmath171
depends only on the mass of the halo and the impact parameter , and for filaments the relevant parameters are the angle between the main filament axis and the geodesic and the impact parameter in the cylindrical radius .
now assume all these variables are discretized into finite - length bins .
the precise way of the discretization will be discussed in detail in section [ mbin ] , but for now we simply let a generic index @xmath172 label the set of independent parameter cells created by the binning . if we now let @xmath173 denote the number of objects that fall into the cell labeled by indices @xmath174 and @xmath172 in the realization @xmath151 , we can replace the sum over the individual objects in eq .
( [ eq : realization2 ] ) by a weighted sum over the discrete parameter cells : @xmath175 where @xmath176 is the total number of independent parameter cells , and so the indexing @xmath172 and the number of different parameter cells @xmath177 could in fact depend on @xmath174 , but we suppress this notation for simplicity . ]
used in the binning , @xmath178 and @xmath179 is the convergence due to one object in the bin @xmath180 : @xmath181 note that the quantity @xmath179 is a function of the distance and internal variables , it is universal for arbitrary realizations : all the information specific to a particular realization @xmath151 is contained in the set of integers @xmath182 .
equation ( [ eq : realization3 ] ) is the starting point of the sgl analysis , because it can be easily turned into a probabilistic quantity .
indeed , instead of thinking of realizations @xmath151 along arbitrary lines of sight through a pre - created model universe , we can define _ a statistical distribution of convergences _ through eq .
( [ eq : realization3 ] ) .
indeed , we have shown in ref .
@xcite that , for random initial and final points of the geodesics , the integers @xmath173 are distributed as poisson random variables : @xmath183 where the parameter @xmath184 is the _ expected _ number of objects in the bin volume @xmath185 : @xmath186 here @xmath187 is the comoving density of objects corresponding to the parameters in the bin @xmath188 and @xmath189 is the corresponding cross sectional area of these objects in co - moving units .
we shall specify these quantities precisely in section [ mbin ] .
physically , the statistical distribution of convergences is equivalent to our original set of realizations @xmath151 averaged over the position of the observer .
this is a welcome feature because the statistical model explicitly incorporates the copernican principle .
the original realization @xmath151 has now been replaced by _ a configuration _ of random integers @xmath190 and the convergence equation eq .
( [ eq : realization3 ] ) by a statistical convergence : @xmath191 where the probability for a particular configuration to occur is just @xmath192 it is easy to show that this probability distribution is correctly normalized .
moreover , including the convergence @xmath193 due to the uniformly distributed matter and the empty space ( see eq .
( [ kem ] ) and below ) , one can show that the total convergence corresponding to the configuration @xmath190 becomes @xcite : @xmath194 for a configuration without any inhomogeneities we correctly recover @xmath195 , see section [ mbin ] .
moreover , eq . ( [ eq : fullkappa ] ) shows explicitly that the _ expected _ convergence vanishes consistently with the photon conservation in weak lensing , because @xmath196 for a random variable following eq .
( [ eq : poisson - i ] ) . the final convergence pdf can now be formally written as @xmath197 where @xmath198 is a discrete probability distribution .
note that the most likely configuration which maximises @xmath199 corresponds to the mode of the poisson distribution , which is the floor of its parameter : @xmath200 .
moreover , for a large @xmath184 the poisson distribution approximates a gaussian with mean and variance equal to @xmath184 and the most likely configuration therefore approaches the mean : @xmath201 . when this is the case the mode of the lensing pdf vanishes even for a single observation and the pdf tends to a gaussian .
one can similarly create the pdf in the shift in the distance modulus @xmath202 or in the magnification @xmath47 , using eq .
( [ eq : mukappa ] ) to compute these quantities for a given configuration and then replacing @xmath146 by the desired quantity in equations ( [ eq : pdf2a]-[eq : pdf1 ] ) . in practice ,
equations ( [ eq : pdf2a]-[eq : pdf1 ] ) are not useful for a direct evaluation of the pdf .
indeed , the configuration spaces are infinite and a direct evaluation of the sum of configurations would not be feasible even for relatively roughly discretized systems . instead , we compute an approximation for @xmath203 by statistically creating a large set of random configurations @xmath190 from the probability distribution @xmath199 , and by forming a discrete normalized histogram out from the corresponding set of convergences ( or distance moduli or magnifications ) .
that is , we define a histogram pdf using the function : @xmath204 where @xmath205 is the total number of realizations and @xmath206 is the number of convergences in the sample falling in the bin @xmath207 centered around the mean value @xmath208 .
it is easy to check that eq .
( [ eq : pdf2 ] ) is correctly normalized to unity .
moreover , the more configurations one creates in the simulation step , the more fine - detailed and accurate approximation for @xmath209 can be constructed . from the simulated data as a moving average weighted by some window function , but that would not give any advantage in the analysis with respect to using the simple histogram of eq .
( [ eq : pdf2 ] ) . ] to summarize , the sgl method consists of two steps : first the internal variables specifying the inhomogeneities are binned , reducing them to a finite number of different object classes labeled by the cell index @xmath172 , for which universal convergence functions @xmath179 can be computed .
second the occupation numbers @xmath210 of these objects within a given co - moving distance bin are shown to be poisson distributed with parameter @xmath184 .
the lensing pdf is then computed statistically , binning the convergence distribution obtained from a large set of random configurations @xmath190 on a given geodesic .
computing the lensing pdf this way is very efficient . finding the pdf for a given model universe using the turbogl package @xcite typically takes a few seconds in an ordinary desktop computer .
see the appendix [ perf ] for more details about its performance .
let us point out that eq .
( [ eq : pdf2 ] ) treats all objects on the same footing .
that is , all object families are independently randomly distributed throughout the space .
confining the halos to within filaments will require slight modifications which we will explain in the section [ confi ] .
for now , we shall study the effects that may cause the actual observed pdf deviate from the fundamental weak lensing pdf .
the lensing pdf ( [ eq : pdf2 ] ) is the fundamental quantity inherent to a given background model and a set of inhomogeneities .
however , in reality there are interferences that prevent us from observing @xmath203 directly .
firstly , the intrinsic magnitudes of the sources are not known with arbitrary accuracy , and so the _ observed _ lensing pdf is at best a convolution : @xmath211 where @xmath212 describes the source magnitude dispersion for an imperfect standard candle .
clearly the fundamental lensing pdf is recovered in the limit that @xmath212 becomes a delta function .
if there were no other sources of errors , the distribution @xmath213 could be used to contrast observations against different cosmological background models , sets of inhomogeneities and the models for the source magnitude distributions .
however , there can be several different types of selection effects caused by matter between the source and the observer or by the search strategies , that may bias the observed pdf from the one predicted by eq .
( [ eq : obspdf1 ] ) .
one of the strengths of the sgl method is that it can be easily extended to model the effects of a wide range of selection biases on the observable pdf .
let us begin with a simple class of biases that might affect the observed distance modulus beyond the weak lensing correction ; an obvious example would be an error made in the estimate of reddening .
what makes such correction nontrivial , is that one would expect the error to be _ correlated _ with the types of mass concentrations encountered by the photon beam on its way from the source to the observer .
the effects should thus be different along different lines of sight , affecting mostly those directions where light has to travel through volumes with higher mass densities . in the sgl method such correlated biases
are very easily modelled , as one needs merely to replace @xmath214 where the first part @xmath215 is the usual weak lensing correction and the second part the object - specific correction .
indeed , the bias function @xmath216 can be correlated with the redshifts @xmath217 and @xmath218 , and with all the internal characteristics of the objects , such as the halo and filament masses and impact parameters .
a quantitative analysis of the possible magnitude of the function @xmath219 would require detailed astrophysical information and it is beyond the scope of this paper .
however , once properly modelled , the sgl allows a very easy way of evaluating the outcome of biases on the observations .
other types of _ selection _ biases might not directly affect the magnitudes , but rather the relative likelihood of observing particular events , in a way that is correlated with the mass density .
this includes , for example , all sources leading to obscuration of the light beam , either alone or in combination with restrictions coming from imperfect search efficiencies and strategies .
for example selection by extinction effects or by outlier rejection mainly relate to high - magnification events which are clearly correlated with having high intervening mass concentrations ( halos with large @xmath49 and/or small @xmath220 ) along the light geodesic .
similarly , short duration events might be missed by search telescopes or be rejected from the data due to poor quality of the light curve , for example in cases where a supernova is not separable from the image of a bright foreground galaxy .
probability of such events would obviously be correlated with the brightness of the source ( @xmath217 and source dependent parameters ) , with the density and redshift of the intervening matter and of course with the search efficiencies .
the selection biases can be studied in the sgl framework by introducing a new , object - dependent _ survival probability function_. that is , we replace @xmath221 in the probability distribution for the magnification pdf ( [ eq : pdf1 ] ) .
here @xmath222 is a normalization constant that makes sure that the final pdf is normalized to unity .
the most generic form of the survival probability function is @xmath223 this allows the survival function to depend on the arbitrary local properties along the photon geodesic .
the physical interpretation of @xmath224 is of course clear : it gives the relative probability that an event whose signal intercepts an object with characteristics given by internal parameters @xmath172 at redshift @xmath218 would make it to the accepted observational sample .
given the form ( [ eq : piu ] ) the proper normalization is easily seen to be : @xmath225 where @xmath226 the correct expression for the convergence is still given by eq .
( [ eq : fullkappa ] ) , with the important difference that the random integers @xmath210 used to evaluate eq .
( [ eq : pdf2 ] ) are now drawn from the poisson distribution of eq .
( [ eq : survival2 ] ) with parameter the effective expected number of objects @xmath227 . as mentioned above ,
modeling @xmath224 quantitatively would need detailed input of the astrophysical properties of the intervening matter distributions and of the observational apparatus , which is beyond the scope of this work .
see ref @xcite for an analytical example giving @xmath228 , where @xmath229 is the effective filling factor of a partially - filled beam @xcite ( see @xcite for an extension of the partially - filled beam formalism ) .
we will discuss an illustrative numerical example in section [ results ] where we will consider as survival probability a simple step - function in the impact parameter . to summarise ,
let us note that it is easy to combine both types of observational biases discussed in this section .
the true observational pdf can be defined as @xmath230 where @xmath212 again is the intrinsic source magnitude dispersion and the effective lensing probability distribution function @xmath231 is computed from eq .
( [ eq : pdf2 ] ) using the effective probability distribution of eq .
( [ eq : survival2 ] ) , and computing the effective magnitudes for configurations from equations ( [ eq : deltameff ] ) and ( [ eq : mukappa ] ) .
this setting clearly allows modeling a very complex structure of observational biases .
finally note that the generic trend among all the effects discussed above is to suppress events with higher magnifications .
it could then be that a non - negligible overestimate of the observed magnitudes follows from neglecting the selection biases , even if the individual effects were relatively small .
moreover , while the present data can not yet constrain the lensing pdf itself , it can put bounds on its variance .
since the observational biases are suppressing the high magnification tail , they give @xmath232 a smaller effective variance which may open the possibility for a comparison with the experiments .
we will further discuss these topics in section [ results ] .
we conclude this section by deriving an effective distribution @xmath233 for a data sample @xmath234 , where the set @xmath234 refers to a given binning of the original data .
@xmath233 could be computed directly from the fundamental pdf @xcite , but it is much faster to create it directly during the initial simulation , bypassing the calculation of the fundamental @xmath203 altogether .
indeed , if we label the configurations within a given sample of observations by @xmath235 , the mean convergence after @xmath236 observations is : @xmath237 where in the second line we have used the fact that the quantities @xmath179 are independent of @xmath235 and in the last line we have used the property that the independent poisson variables @xmath238 ( with the same weight ) sum exactly ( in a statistical sense ) into the poisson variable @xmath239 of parameter given by the sum of the individual parameters , which is @xmath240 .
@xmath233 is then given by an expression similar to eqs .
( [ eq : pdf2a]-[eq : pdf1 ] ) or eq .
( [ eq : pdf2 ] ) .
note , however , that including the selection effects ( see section [ sbs ] ) in general does not commute with taking an average over the observations .
in other words , if the @xmath241 measurements are correlated , we can not use eq .
( [ equation : kappa3 ] ) , but we have to start from the fundamental pdf . on the other hand , eq .
( [ equation : kappa3 ] ) displays explicitly the effect of the size of the data sample : even if @xmath146 had a skewed pdf and a nonzero mode for @xmath242 , for large @xmath241 the distribution approaches a gaussian and eventually converges to a @xmath243-function with a zero convergence , as it is clear from the properties of poisson variables . from the general expression eq .
( [ equation : kappa3 ] ) , including a nontrivial survival probability , it follows that the expected value and variance of the convergence are : @xmath244 and @xmath245 where we assumed that the survival probability for a light ray going through the uniformly distributed matter density @xmath246 equals unity and that the random variables @xmath210 are uncorrelated .
( [ averagekappa]-[varkappa ] ) can then be put back in integral form giving the following _ exact _ results : @xmath247 and @xmath248 where the integral limits for the last two integrals are implicitly defined and the explicit form of surface area element @xmath249 depends on the geometrical properties of the object of mass @xmath49 . for example , for spherical halos it is @xmath250 .
we again stress that the quantity @xmath251 is a generic function that describes the probability that a light ray is observed when it hits a halo of mass @xmath49 at the redshift @xmath50 and impact parameter @xmath220 for a source at redshift @xmath252 .
( [ averagekappax]-[varkappax ] ) are a simple and direct prediction of the sgl method and allow to draw some general considerations .
first , if @xmath253 , the expected value of the convergence is correctly zero , showing the `` benevolent '' nature of weak lensing corrections for unbiased observations .
if however the survival probability is not trivial , selection biases persist even in very large datasets so that the average convergence over a large number of observations approaches a nonvanishing value .
second , eq . ( [ varkappax ] ) is a product of positive quantities and so _ a nontrivial survival probability always reduces the observed variance . _ we would like to stress that it is easy to numerically evaluate the integrals of eqs .
( [ averagekappax]-[varkappax ] ) , and their predictions can be straightforwardly implemented in @xmath254 analyses based on gaussian likelihoods , which may be a reasonable approximation if the observable pdf is not strongly skewed .
the key role played by the discretization of the model parameters in the sgl method is now quite obvious . of course
, the final lensing pdf should not depend on the discretization . indeed , as the numerical value of a riemann integral does not depend on the binning chosen
if the integrand does not significantly vary within the bins , the resulting pdf will not depend on the specific binning if the cosmological functions @xmath179 are approximately constant within the bin cells @xmath188 .
once this condition is well met a further refinement of binning does not increase the accuracy .
this statement can be proved formally as follows : imagine that we have a valid binning prescription @xmath255 and we want to move to a new prescription @xmath256 which is @xmath257 times finer . because @xmath258 is valid , the convergences @xmath259 are approximately constant within each @xmath255-bin .
the relative poisson variables can then be resummed as done in eq .
( [ equation : kappa3 ] ) , and we obtain the same formal expression in terms of @xmath258-bins we started from .
this theorem also explains why we do not need to bin any variable in which the surface mass density @xmath171 of a given object is constant , such as the rotational angle along the center of a spherical halo in the lens plane .
as the convergences @xmath179 are constants within angular bins , their contributions can be resummed to give a model without any angular binning .
that is , the fundamental halo objects are actually finite - width rings in the convergence plane . the first quantity to be binned is the co - moving distance to the source . here
the basic quantity that needs to be accurately modelled is the function @xmath260 , which gives the optical weight of a given convergence plane .
this is a rather smooth function with a shallow peak , and only slightly dependent on @xmath150 when expressed in the scaled variable @xmath261 .
simple linear bins @xmath162 have been found to give accurate results , with only @xmath262 redshift slices .
this turns out to be the typical number of bins needed for an accurate modeling of most of the internal variables as well .
we begin by discretizing the halo parameters , which are the mass and the impact parameter in the lens plane . binning the halo masses
involves two main considerations .
first , the halo mass function only carries information about high - contrast virialized halos within a specific mass range ( see eq .
( [ jenk ] ) ) .
moreover , as explained in section [ mati ] , only very large mass concentrations play a significant role in our weak lensing analysis .
consequently , we introduced the lower cut @xmath70 to the mass distribution so that only mass concentrations with @xmath58 are treated as halos , while the remaining mass has been divided into low contrast objects and uniform matter density ( see eq .
( [ eq : sumrule ] ) ) .
second , the fact that the concentration parameter of eq .
( [ duffy ] ) follows a power law suggests the use of a logarithmic binning in mass .
the mass function drops quickly at very large masses , so that we can combine all masses above a certain limit to a single bin .
we found it reasonable to define at any redshift .
one could , for example , use bins such that @xmath69 has always the present - day value .
see section [ mati ] for a discussion about @xmath69 .
] the cell boundaries by @xmath263 and the cell centers by @xmath264 , with @xmath265 . because @xmath266 to a very good approximation , it is sufficient to bin with an integer @xmath267 between 10 and 15 : @xmath268 the fraction of total mass in a bin is then @xmath269 which correctly gives the total halo mass fraction ( see eq .
( [ eq : def_of_fh ] ) ) when summed over all the halo bins : @xmath270 finally , the binned comoving halo densities are @xmath271 which also allows to relate the mass @xmath272 to the interhalo scale @xmath273 .
the halo impact parameter @xmath274 in the lens plane is restricted to the range @xmath275 $ ] , where @xmath111 is the comoving halo radius .
it should be noted that through eq .
( [ mdel ] ) the radius depends both on the halo mass and the redshift .
see fig .
[ sketch ] for an illustrative sketch .
it is efficient to discretize @xmath220 into bins of constant integrated surface density ( constant equal mass ) : @xmath276 where @xmath277 is the number of @xmath220-bins used and the surface density is given by ( see fig .
[ sketch ] ) : @xmath278 after the redshift and the halo mass is specified , the bin boundaries @xmath279 and the weighted centers of gravity @xmath280 can be computed using eq .
( [ binbounds ] ) .
after these are defined , the corresponding area functions @xmath281 , needed for the poisson parameters @xmath282 and the binned surface densities @xmath283 , can be computed .
overall , our halo model is described by the redshift index and two internal indices for the discretized mass and impact parameter . in terms of the generic parameter @xmath172
we can formally express this as : @xmath284 , where the label @xmath285 refers to the halo " family of the inhomogeneities .
.,width=8 ] of eq .
( [ ltt ] ) as shown in fig .
[ cylinder2].,width=8 ] the position and the orientation of a generic object in the three dimensional space is described by three position and three orientation degrees of freedom . since
our filaments objects are confined to within a given @xmath286 slice , they generically have only five degrees of freedom . special symmetries of a sphere reduced the number of relevant parameters for halos to a sigle one , the impact parameter @xmath274 . for filaments
the situation is slightly more complicated because a cylinder is invariant only for rotations around one symmetry axis .
this leaves us with four degrees of freedom .
the relevant parameters , however , are reduced by the symmetries to only two : the angle @xmath151 of the filament main axes with respect to the geodesic and the impact parameter @xmath287 . for illustration
see figs .
[ cylinder2]-[cylinder1 ] .
this can be understood by the fact that all the possible filament configurations with the same @xmath151 and @xmath287 have the same surface density @xmath288 and so can be resummed using the theorem given above . in particular , the projected surface density of our filaments does not depend on the coordinate @xmath289 along the projected main axes we move " the triangular section as shown in fig .
[ cylinder2 ] . strictly speaking our filaments
do not have a cylindrical symmetry , but a @xmath151-dependent tilted " rotation symmetry , such that the projected surface densities follow equation ( [ gammat0 ] ) for all @xmath290 .
this approximation should be good for long filaments with @xmath291 . ]
allowing to resum the @xmath289-bins into a single effective bin of the total @xmath151-dependent length of the projection : @xmath292 the angle @xmath151 is physically important because objects seen with a smaller @xmath151 have a smaller cross section and a higher surface density .
accounting for this geometrical effect leads to a more strongly skewed lensing pdf .
the expression for the surface density is given by @xmath293 for the uniform filament profile this gives simply @xmath294 , where @xmath295 .
for the gaussian profile ( [ filagau ] ) one instead finds @xmath296 \ , .
\label{gammat}\ ] ] the angle @xmath151 can be restricted by symmetry to the interval @xmath297 $ ] , which we divide into @xmath298 uniform length bins around centers @xmath299 .
the impact parameter is restricted to @xmath300 $ ] , and it will be discretized into @xmath301 bins using the equal mass criteria , as in the halo case defined in eq .
( [ binbounds ] ) .
that is , our filaments will be represented by a family of bars " of different surface densities @xmath302 and length @xmath303 .
the surface area of these bars to be used in the poisson parameters is @xmath304 .
overall , our filament model is described by the redshift index , and two internal indices for the discretized angle and impact parameter . in terms of the generic parameter @xmath172
we can again formally express this as : @xmath305 , where the family label @xmath104 refers to the filament " family of the inhomogeneity .
of course adding , _
e.g. _ , a mass distribution for filaments and/or an @xmath306-dependent density profile would increase the necessary number of indices in the discrete model .
until now we have implicitly assumed that the halo and the filament families are independent . we will now show how a more realistic model with halos confined to the filaments can be set up in the sgl approach .
the central observation is that the exact positions of halos and filaments within the co - moving distance bins are irrelevant in eq .
( [ eq : fullkappa ] ) for the convergence .
it then does not matter if we really confine the halos to filaments ; for the same effect it is sufficient to merely confine them into the equivalent volume occupied by the filaments in a given bin .
we can actually do even better than imposing a simple volume confinement .
indeed , it is natural to assume that the halo distribution follows the smooth density profile that defines the filament , which can be accounted for by weighting the volume elements by their reduced density profiles . because only the projected matter density is relevant for lensing ,
this weighting is accounted for by the effective surface densities of the objects .
we then introduce the effective co - moving thickness for the low density objects as : to refer to generic internal filamentary degrees of freedom . ]
@xmath307 where @xmath308 is the average mass density of the filament with volume @xmath309 .
the total effective length of a configuration of filaments @xmath310 along the line of sight is then : @xmath311 it is easy to show that the statistical average of @xmath312 over the configuration space is just the expected distance covered by filaments : @xmath313 where we defined the comoving average filament volume fraction : @xmath314 the confining simulation can now be performed in two steps : in the first step a random configuration of filaments is generated and eq .
( [ eq : master1 ] ) is used to compute the effective lengths @xmath315 at different co - moving slices .
these lengths define then the poisson parameters for the halos through eq .
( [ deltan ] ) : @xmath316 note that the halo densities @xmath317 are corrected to account for the reduced volume available due to confinement : @xmath318 where @xmath319 are the usual halo densities of eq .
( [ nbins ] ) when the halos are uniformly distributed through all space .
inserting eq .
( [ eq : haloweights ] ) back to ( [ deltanhalos ] ) and using eqs .
( [ eq : master1]-[eq : master2 ] ) we can rewrite eq .
( [ deltanhalos ] ) simply as @xmath320 where @xmath321 are the usual configuration independent poisson parameters computed without confinement .
note that the central quantities needed in the evaluation of eq .
( [ deltanf ] ) are the filamentary weights @xmath322 and surface densities @xmath323 . with
these given , all the information about the specific low contrast objects used is condensed into the single stochastic variable : @xmath324 which has an expectation value of unity and a mode smaller than unity . the lensing effects due to large - scale structures
are thus tied to the probability distribution of @xmath325 : its added skewness is the effect coming from confining halos within filaments .
this opens the possibility to investigate and compare different filamentary geometries by means of the @xmath325-pdf .
we will develop these thoughts further in a forthcoming paper @xcite .
we can further generalize the picture given in this section by considering different levels of confinement for halos of different masses . indeed , because halos of different mass are treated independently in eq .
( [ deltanf ] ) , one could have small halos populate also the voids and massive halos only the filamentary structure .
let us finally point out that in practical computations the convergence pdf is obtained by creating a large number of independent halo configurations for each master " filament configuration .
typically we use simulations with few hundred master configurations with a few hundred halo configurations each . in this section
we will discuss the power spectrum of our model universe and the importance of large - scale correlations for the lensing pdf .
we start by making a connection between the sgl method and the so - called `` halo model '' ( see , for example , @xcite and @xcite for a review ) , where inhomogeneities are approximated by a collection of different kinds of halos whose spatial distributions satisfies the linear power spectrum .
the idea behind the halo model is that on small scales ( large wavenumbers @xmath326 ) the statistics of matter correlations are dominated by the internal halo density profiles , while on large scales the halos are assumed to cluster according to linear theory .
the two components are then combined together .
the sgl modeling of the inhomogeneities can be thought as a two - step halo model where we first create the random filamentary structures and then place the halos randomly within these structures .
similarly to the halo model , one can then combine the linearly - evolved power spectrum with the nonlinear one coming from the filaments and the halos they contain . in this sense
our filamentary structures extend the halo model by introducing correlations in the intermediate scales between the halo substructures and the cosmological scale controlled by the linear power spectrum . in the halo model
the power spectrum can be computed analytically , but here the calculation has to be done numerically .
we are currently extending the sgl method such that the simulation will produce also the power spectrum @xcite in addition to the lensing pdf .
the power spectrum can indeed place useful constraints on the filament parameters which , as remarked in section [ mati ] , are not tightly constrained by observations .
while the power spectrum is relevant for understanding the correlations at the largest observable scales , the lensing pdf depends mainly on the much stronger inhomogeneities at smaller scales .
this can be understood from eq .
( [ eq : kappa1 ] ) which shows the direct dependence of the lensing convergence on the density contrast .
the small lensing variance from the large - scale correlations induced by the linear power spectrum was numerically computed , for example , in @xcite .
moreover , weak lensing and the power spectrum probe somewhat different aspects of the inhomogeneities .
the web - like structures of filaments and voids that affect weak lensing are mainly described by higher order correlation terms beyond the power spectrum , and so special care has to be put in designing the filamentary structure .
this is indeed the goal of the sgl method : to give an accurate and flexible modeling of the universe as far as its lensing properties are concerned .
our main goal in this section is to compare the sgl approach with the convergence pdf computed from large scale simulations .
the idea is that by achieving a good agreement with numerical simulations we are proving that the sgl method does provide a good and accurate description of the weak lensing phenomena .
as we shall see , this is indeed the case , as we can naturally reproduce the lensing pdf of the millenium simulation .
let us stress that while a comparison to simulations is a good benchmark test for the sgl approach , the simulations and underlying @xmath123cdm model do not necessarily describe nature .
indeed observations do not yet provide strong constraints to the lensing pdf , which leaves room for very different types of large scale structures , examples of which have been studied for example in ref .
@xcite . with the accuracy of the method tested
, one can reliably compute the effects of selection biases using sgl . while we lack the necessary expertise to quantitatively estimate the possible bias parameters , we will study a simple toy model for the survival probability function @xmath224 to show qualitatively how selection effets might bias the observable pdf .
finally , we will also show how a jdem - like survey could constrain the lensing pdf relative to a given cosmological model .
we shall begin , however , by the comparison with the simulations .
we shall now confront the sgl method with the cosmology described by the millennium simulation ( ms ) @xcite .
accordingly , we will fix the cosmological parameters to @xmath327 , @xmath328 , @xmath329 , @xmath330 , @xmath331 @xmath332 and @xmath333 .
moreover , the mass function of eq .
( [ jenk ] ) agrees with the ms results .
these parameters completely fix the background cosmology and the halos .
we are then left with the filament parameters to specify .
first , we fix @xmath334 , which means that half of the unvirialized mass forms filaments , while the other half is uniformly distributed .
this value determines the maximum possible demagnification @xmath335 in the lensing pdf , corresponding to light that misses all the inhomogeneities . for the background model described above this
implies @xmath336 mag . for each @xmath337
the sgl method evaluates the corresponding occupational number from the poisson statistics .
the program mathematica using one core of a cpu at @xmath338 ghz takes a time @xmath339 s to produce an array of @xmath340 poisson numbers .
other programs will likely have similar performances .
@xmath341 determines the statistics with which the lensing pdf is generated . using this information we
can then estimate the performance of the sgl method by evaluating @xmath342 .
the results are shown in table [ turbo ] where we fixed @xmath343 , @xmath344 and @xmath345 .
confining halos within filaments typically makes calculations about ten times longer .
p. valageas , astron .
astrophys . * 356 * , 771 ( 2000 ) .
d. munshi and b. jain , mon . not .
soc . * 318 * , 109 ( 2000 ) .
y. wang , d. e. holz and d. munshi , astrophys .
j. * 572 * , l15 ( 2002 ) .
s. das and j. p. ostriker , astrophys .
j. * 645 * , 1 ( 2006 ) .
l. amendola , k. kainulainen , v. marra and m. quartin , phys .
* 105 * , 121302 ( 2010 ) .
s. hilbert , s. d. m. white , j. hartlap and p. schneider , mon . not .
* 382 * , 121 ( 2007 ) .
s. hilbert , s. d. m. white , j. hartlap and p. schneider , mon . not .
* 386 * , 1845 ( 2008 ) .
v. springel _ et al .
_ , nature * 435 * , 629 ( 2005 ) . m. chevallier and d. polarski , int . j. mod
. phys .
d * 10 * , 213 ( 2001 ) ; e. v. linder , phys .
lett . * 90 * , 091301 ( 2003 ) . m. bartelmann and p. schneider , phys .
rept . *
340 * , 291 ( 2001 ) .
a. jenkins _ et al .
_ , mon . not .
soc . * 321 * , 372 ( 2001 ) .
r. kantowski , astrophys .
j. * 155 * , 89 ( 1969 ) .
v. marra , e. w. kolb , s. matarrese and a. riotto , phys .
d * 76 * , 123004 ( 2007 ) ; v. marra , e. w. kolb and s. matarrese , phys .
d * 77 * , 023003 ( 2008 ) ; v. marra , arxiv:0803.3152 [ astro - ph ] ; e. w. kolb , v. marra and s. matarrese , gen .
grav . * 42 * , 1399 ( 2010 ) .
e. rozo _ et al .
_ , astrophys . j. * 708 * , 645 ( 2010 ) . w. j. percival , astron .
astrophys . * 443 * , 819 ( 2005 ) ; see also : s. basilakos , astrophys .
j. * 590 * , 636 ( 2003 ) .
d. j. eisenstein and w. hu , astrophys .
j. * 496 * , 605 ( 1998 ) .
m. j. .
white , astrophys
. j. suppl .
* 143 * , 241 ( 2002 ) .
r. k. sheth and g. tormen , mon . not .
soc . * 308 * ( 1999 ) 119 .
j. f. navarro , c. s. frenk and s. d. m. white , astrophys . j. * 462 * , 563 ( 1996 ) .
a. r. duffy , j. schaye , s. t. kay and c. dalla vecchia , mon . not .
390 * , l64 ( 2008 ) .
e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys . j. suppl . * 180 * , 330 ( 2009 ) . | we revise and extend the stochastic approach to cumulative weak lensing ( hereafter the sgl method ) first introduced in ref .
@xcite . here
we include a realistic halo mass function and density profiles to model the distribution of mass between and within galaxies , galaxy groups and galaxy clusters .
we also introduce a modeling of the filamentary large - scale structures and a method to embed halos into these structures .
we show that the sgl method naturally reproduces the weak lensing results for the millennium simulation .
the strength of the sgl method is that a numerical code based on it can compute the lensing probability distribution function for a given inhomogeneous model universe in a few seconds .
this makes it a useful tool to study how lensing depends on cosmological parameters and its impact on observations .
the method can also be used to simulate the effect of a wide array of systematic biases on the observable pdf . as an example
we show how simple selection effects may reduce the variance of observed pdf , which could possibly mask opposite effects from very large scale structures .
we also show how a jdem - like survey could constrain the lensing pdf relative to a given cosmological model .
the updated code is available at . |
detailed studies of the infrared surface brightness of the galactic bulge ( matsumoto et al . 1982 ; blitz & spergel 1991 ) suggested a triaxial ellipsoid or bar with a major axis inclined to the line of sight and closer to us at positive galactic longitudes .
whitelock & catchpole ( 1992 ) see also ( whitelock 1992 , 1993 ) used the distribution of individual bulge mira variables , with distances from a near infrared period - luminosity relation , to show that these stars were in a bar with major axis at @xmath7 to the line of sight .
work on bulge structure to about 2009 was summarized in table 1 of vanhollebeke et al .
( 2009 ) ( also reproduced in rich 2013 ) .
this table , together with later work ( e.g. gonzalez et al .
2011 ; gerhard et al . 2012 ; dkny et al . 2013 ; pietrukowicz et al .
2012 , 2015 ; wegg et al . 2013 ; cao et al . 2013 ; vasquez et al . 2013 ; gran et al . 2015 ; nataf et al .
2015 ) shows that a large range of bar angles have been suggested .
these are generally in the range @xmath8 to @xmath9 but have ranged from @xmath10 ( robin et al .
2003 ) to a near spherical distribution ( dkny 2013 ) .
apart from observational uncertainties , one reason for this large range is the way the data have been handled .
the locus of mean distance along lines of sight to the bulge can have quite a different slope to that of the major axis of a bar , derived from the same data using a model ( e.g. wegg 2013 ; pietrukowicz et al .
for instance , for rr lyraes in a central bulge region pietrukowicz et al .
( 2015 ) find an angle of @xmath11 from the mean distances in different directions and @xmath12 from an elliptical model .
it is also now becoming realized that the observed bulge shape may depend on the region of the bulge being studied as well as on the type of tracers used .
whilst , as documented by rich ( 2013 ) , there is a large range in metallicities among bulge stars , there has not been full agreement on the age range in the bulge .
it seems clear that there is an old ( globular cluster age ) population in the bulge from the presence of a strong population of rr lyrae variables and it has often been maintained that the bulge is entirely or almost entirely composed of a very old population ( e.g. from colour - magnitude studies of a field at @xmath13 ( zoccali et al .
clarkson et al . ( 2011 ) put the fraction of the population younger than 5 gyr as at the most 3.4 percent in a field at @xmath14 .
however , a recent study of microlensed stars in the bulge suggests at least a small population of young stars scattered through the bulge ( bensby et al .
bulge mira variables , which show a wide range in periods and for which there is a period age relationship , have , in fact , long suggested a significantly younger population exists there .
nataf ( 2015 ) recently reviewed the evidence for and against the presence of a younger population and concluded that enhanced helium - enrichment would go some way to resolving the controversy . in the present paper
we report periods and @xmath0 photometry of 648 mira variables in the bulge , chosen to study bulge structure .
these observations are combined with 2mass data for 8057 bulge miras with ogle and/or macho periods , to study the structure of the bulge as a function of age as well as to estimate the distance to the galactic centre .
mira variables are asymptotic giant branch ( agb ) stars with long pulsation periods ( @xmath15 days ) and large amplitude variations ( @xmath16 , @xmath17 mag ) .
they are cool ( @xmath18k ) and are generally understood to be near the end of the agb evolutionary phase where they are rapidly losing mass ( typically @xmath19 ) .
kinematic and other studies ( e.g. feast & whitelock 2000 ; feast 2008 ) have shown that the period of a mira is a good indication of its age and/or initial mass .
this suggests that whilst the shorter period miras ( some of which are found in metal - rich globular clusters ) are very old with @xmath20 , the bulk of miras in the solar neighbourhood with @xmath21 are @xmath22gyr old .
an age of @xmath23 gyr has been estimated at @xmath24 and in the following we refer to these as of intermediate age " .
longer period miras ( including oh / ir stars which are generally long period miras ) are even younger .
as shown by wood ( 2000 ) , agb variables lie on several parallel sequences in a @xmath25 period - luminosity ( pl ) diagram , where each sequence is a different pulsation mode .
the mira variables , which have larger amplitudes than the semi - regular ( sr ) variables , fall mostly on the pl relation corresponding to fundamental pulsation ( sequence c in wood 2000 ) .
however , at the long periods ( @xmath26 days ) the mass - loss rates are sometimes sufficiently high that circumstellar extinction forces the apparent luminosity below the pl relation , unless the extinction is corrected for ( e.g. whitelock et al .
1991 ; glass et al . 1995 ; matsunaga et al .
2009 ) . in the large magellanic cloud
we also find a few miras above the pl relation ( e.g. feast et al .
1989 ) ; these are believed to be relatively high mass hot - bottom - burning ( hbb ) stars ( whitelock et al .
2003 , menzies et al .
2015 ) possibly pulsating in the first overtone ( feast 2009 ) .
we expect these higher mass objects to be concentrated toward the centre and close to the plane .
their numbers in our fields are expected to be small and should not affect our discussion .
the mira pl relation is potentially a very useful tool for distance scale studies , within the galaxy and at larger distances , as recently summarized by whitelock ( 2013 ) .
however , correction for both interstellar and circumstellar extinction must be made .
this is discussed in section [ ext ] .
the intrinsic colours of galactic long period variables were discussed by feast et al .
( 1982 ) and using these data glass & feast ( 1982 ) isolated an area of a @xmath27 two - colour diagram ( their fig .
1 ) in which dereddened miras are to be found .
this is the referred to as the mira box " in the following discussion .
the atmospheric abundance of carbon relative to oxygen determines whether miras are oxygen rich ( @xmath28 ) or carbon rich ( @xmath29 ) .
carbon enrichment happens during third dredge - up ( iben & renzini 1984 ) in stars which have sufficiently large initial masses , and the existence or otherwise of carbon stars has sometimes been taken as evidence for or against the existence of an intermediate mass population .
note that more carbon must be dredged - up in order to produce carbon stars in a population that has a high initial oxygen abundance , as does the bulge , and that dredge - up efficiency depends on the initial abundance as well as on the initial mass , being less efficient at high metallicity or high helium abundance ( karakas et al . 2014 and references therein ) .
the process of hot bottom burning ( e.g. boothroyd , sackmann & ahern 1993 ) prevents intermediate mass stars from becoming carbon rich , particularly at low metallicity , although this is probably not relevant to the galactic bulge .
it should be noted that we assume that the miras in the bulge are all or nearly all oxygen rich .
optical spectroscopy of giants suggests there are only a very few c - rich stars in the bulge ( blanco , mccarthy & blanco 1984 ) and those few are faint , blue and probably the products of binary evolution .
more recently soszynski et al . (
2013 ) have interpreted the ogle light - curves of agb stars as supporting an o - rich character for the majority of the variables .
relatively little is known about the chemistry of highly obscured variables towards the galactic centre , except for those from which oh - masers have been detected ( e.g. sevenster 1999 ) , which are obviously o - rich .
cole & weinberg ( 2002 ) are sometimes quoted as demonstrating the presence of numerous c - stars in the bar of the bulge , on the basis of 2mass colours ( and some unpublished spectra of the inner galaxy ) .
the interpretation of the 2mass colours is based on the assumption that c stars have large @xmath30 colours , as is found in lower metallicity environments such as the lmc . however , the @xmath27 colours of mass - losing o - rich stars are very similar to those of c - stars and these will dominate in the higher metallicity galactic bulge .
ojha et al .
( 2007 ) have questioned the cole & weinberg interpretation of the 2mass colours and ishihara et al .
( 2011 ) show , from a study of stars with both 2mass and akari photometry , that the majority of agb stars in the direction of the galactic centre are o - rich .
this is confirmed by uttenthaler et al .
( 2015 ) . ) of the bulge to the sun - centre line of sight .
compare this with fig .
[ allstar ] to see the distribution of the miras . , width=321 ]
the early optical survey of bulge miras by gaposchkin ( 1955 ) , on plates taken by baade and that of oosterhoff et al .
( 1967 , 1968 ) , on plates taken by thackeray and colleagues , as well as the work of lloyd evans ( 1976 ) , indicated a broad range in period with most miras having periods in the range 150 to 450 days in the ngc6522 , sgri and sgrii windows ( @xmath31 ) .
most of the bright iras sources delineating the overall structure of the bulge are mira variables ( feast 1986 ; glass 1986 ) and these tend to have even longer periods ( whitelock et al .
1991 ; glass et al .
1995 ) , with a few as long as @xmath32 days .
it is possible that some of the longest period stars may be due to binary mergers as suggested by renzini & greggio ( 1990 ) ( one probable example of such a process is a carbon rich mira with a period of 551 days in a globular cluster ( matsunaga 2006 ; feast et al .
however , it seems very unlikely that this could be the source of all the 300 to 400 day miras ( see also section [ sec2 ] ) .
thus , taking the presence of long period miras together with the discussion in section [ sec2 ] , indicates a substantial intermediate age population in the bulge .
as mentioned above , whitelock & catchpole ( 1992 ) and whitelock ( 1992 ) observed 141 iras miras in two strips on either side of the galactic centre at @xmath33 and @xmath34 and made the first attempt to use individual stars to probe the structure of the bulge in three dimensions .
a good fit to their data was given by a triaxial bulge inclined at @xmath35 to the line of sight .
the relatively high galactic latitude of the fields indicated a thick bar and there were suggestions of an x " structure .
sevenster ( 1999 ) discussed the distribution and velocities of 507 oh / ir stars within @xmath36 & @xmath37 .
she finds evidence for an inclined bar as well as other structures .
she also finds 9 stars from her young sample , lying along an hi feature identified with the 3 kpc arm .
groenewegen & blommaert ( 2005 ) use denis and 2mass colours to analyse 2691 mira variables identified in the ogle survey . from these data
they derive mean distances along various lines of sight and find evidence for a bar at an angle of @xmath38 to the line of sight .
matsunaga , fukushi & nakada ( 2005 ) find 1968 mira variables in the ogle ii data base .
they assign them 2mass colours and note that their surface density is similar to that of the 2.2@xmath39 m cobe image .
the radial distribution shows a tilt towards us in positive longitudes . the work of matsunaga et al .
( 2009 ) on miras close to the centre is discussed in section [ distance ] .
we selected two fields ( the saao fields ) , both at @xmath1 , centred at @xmath40 ( bulge east ) and @xmath41 ( bulge west ) and each covering about 25 square degrees ( see figs .
[ bulgemod ] & [ allstar ] ) . over a period of about six years
, the aao schmidt was used to obtain ivn ( @xmath42-band ) plates in each field , which are listed in table [ tab_dates ] .
the schmidt plates were scanned , using both the apm in cambridge and the mama at the paris observatory , to find the long period variable candidates and their periods .
the apm periods were found using phase dispersion minimization ( stellingwerf 1978 ) and the mama periods using the the method described by schwarzenberg - czerny ( 1996 ) .
the candidates were then observed with the saao 1.9-m telescope to obtain @xmath0 photometry in the saao system ( carter 1990 ) .
the error of individual observations is estimated to be 0.03 in @xmath27 and 0.05 in @xmath43 .
.schmidt survey plates [ cols="^,^,^,^ " , ] lcccccc & & + & & & & & & + + @xmath44 & @xmath45 & @xmath46 & 4671 & @xmath47 & @xmath48 & 1132 + @xmath49 & @xmath50 & @xmath51 & 4326 & @xmath52 & @xmath53 & 1038 + @xmath54 & @xmath55 & @xmath56 & 4607 & @xmath57 & @xmath58 & 1107 + no limit & @xmath59 & @xmath60 & 4883 & @xmath61 & @xmath58 & 1182 + + @xmath44 & @xmath62 & @xmath63 & 486 & @xmath64 & @xmath65 & 85 + @xmath49 & @xmath66 & @xmath67 & 464 & @xmath68 & @xmath69 & 84 + @xmath54 & @xmath70 & @xmath71 & 491 & @xmath72 & @xmath73 & 88 + no limit & @xmath74 & @xmath75 & 525 & @xmath76 & @xmath77 & 94 + table [ tabrt_red ] shows how the derived values of @xmath5 and @xmath78 vary using different methods of correcting for interstellar and circumstellar extinction . `
none ' indicates that no extinction correction was made .
the main significance of this table is to show that values of @xmath78 are insensitive to the extinction corrections used .
as expected , neglecting circumstellar extinction leads to an overestimate of @xmath5 , especially for the longer period stars . using the matsunaga " method ( section [ ext ] ) leads to @xmath79kpc for the short period group in table [ tabrt_red ] and @xmath80kpc for the long period group , in good agreement with the whitelock " results in that table .
the difference between these two values is probably a consequence of uncertainties in the reddening correction ( see section 6 ) and/or due to the 3-dimensional structure ( see section 8) of the bulge . since the whitelock " results are in the saao system and the matsunaga " results in the 2mass system , the agreement suggests that conversion between the systems has little effect on the results .
it is evident from , for example fig .
[ modlong ] , that there is a considerable apparent spread in distance moduli of the miras in any one direction .
part of this is no doubt real and part due to uncertainty in the derived moduli .
the effects of truncating the sample in the saao fields was already discussed in section [ largeb ] .
the first part of table 4 shows the effects of similar experiments with the lower latitude stars .
the effects are relatively small , especially considering that the uncertainties in the results are dominated by external errors ( see following paragraph ) .
truncation has the effect of slightly reducing the value of @xmath5 for the miras at low latitude and increasing them for the high latitude saao fields , presumably because of the different balance between foreground and background objects .
matsunaga et al . ( 2009 ) give some discussion of the real ( external ) errors of their value of @xmath5 derived from miras in the region of the centre .
they adopt 0.4kpc for this quantity .
our miras suffer less interstellar absorption than theirs but are further from the centre and are thus not fully model independent .
this is particularly true of the longer period stars which are in a bar structure .
the correction for circumstellar extinction is also more important for these latter stars .
we thus adopt as our best estimate of @xmath5 the value for miras with @xmath81 , restricted to @xmath82 and taking a mean of the last three estimates in table [ tabrt_red ] ( 8.9kpc ) and adopt matsunaga s error estimate ( 0.4kpc ) .
whether on not there is small scale structure in the bulge is of considerable significance for its formation and evolution . to illustrate possible structure , we have selected miras within a restricted range of @xmath83 and @xmath84 over which the coverage is fairly uniform , as can be judged by looking at fig .
[ allstar ] .
the miras are projected onto the galactic , @xmath85 plane , after correcting for whitelock " extinction .
the plane is divided into 0.15kpc square bins and the number of miras is counted in each bin .
each bin becomes an image pixel and the maximum number of stars in a pixel is 15 in the short period group and 7 in the long .
the maximum number of stars is used to scale the resulting image and normalise it to an intensity count of 255 . the image
is then smoothed and doubled in size to make it more legible .
images for the short and long period miras are shown in fig .
[ miraclump ] .
the restriction in @xmath83 and @xmath84 means that the images do not correspond to the subsets of data given in table [ tabrthet ] . and
@xmath86 , over which the areal coverage is nearly uniform .
the numbers are the total number of miras in the image.,width=321 ] in fig .
[ miraclump ] there is a strong contrast between the spatial distribution of old @xmath87 and intermediate age @xmath88 miras .
the short period group show a relatively smooth centrally concentrated distribution .
the long period group show a clumpy structure as well as evidence of the tilted bar to which the two main clumps at 6.8 and 8.6kpc seem to belong . and
@xmath86 and @xmath88 , overlain on fig . 6 ( grey circles ) of mcwilliam & zoccali ( 2010 ) , showing peaks in their red clump stars .
their distance scale is matched to ours.,width=321 ] in fig .
[ mirasmoothmcw ] the clumping in the long period miras is compared with the distribution of red clump stars found by mcwilliam & zoccali ( 2010 ) in the @xmath89 plane , shown in their fig .
they interpret these concentrations as the wings of an x shaped or triaxial bar , seen almost edge on to our line of sight , though this interpretation has been disputed ( lee , joo & chung 2015 ) .
the correspondence between some of their clumps and ours , shown in fig .
[ mirasmoothmcw ] , suggests that the red clump stars and the long period miras have similar distributions .
taken together with the results of the previous paragraph , this would indicate a more clumpy structure in a ( younger ) triaxial or x - shaped bar than in the older part of the bulge .
it has been proposed that the bulge formed from the infall and eventual merger of large massive lumps ( e.g. noguchi 1999 , immeli et al .
2004 , ferraro et al .
2009 , massari et al . 2015 ) whether they could be long enough lasting to explain the observed clumping is not clear . when looking at the structures , shown in fig .
[ miraclump ] and fig .
[ mirasmoothmcw ] , it is important to remember that they lie within a slice , which passes through the centre of the galaxy between 350pc and 600pc above the galactic plane .
we have obtained periods and @xmath0 photometry for 648 mira variables in two outer fields in the direction of the galactic bulge and combined it with 8057 miras found in the ogle and macho surveys of the inner bulge .
distances were derived from a period - luminosity relation together with corrections for interstellar and circumstellar extinction . in both our inner and outer regions
the longer period miras , tracers of an intermediate age population , give evidence of a tilted bar - like structure .
the older , shorter period miras do not show this .
the longer period miras also show a more clumpy structure than the shorter period ones , possibly part of an x - structure .
we estimate a distance to the galactic centre in satisfactory agreement with other estimates . the change of mira distribution with period , besides showing the complex structure of the bulge ,
also indicates that the longer period stars there , are fundamentally different from the shorter period ones .
it therefore supports the view that the longer period miras in the bulge belong to an intermediate age population , as such stars do elsewhere , and are not , for instance , the progeny of the mergers of old ( low mass ) stars .
the exact age of that intermediate age population may depend on abundance details and particularly on the level of helium enrichment .
we are grateful to the following people who took part in the infrared photometry programme at sutherland , brian carter , dave laney and enrico olivier .
mwf and paw gratefully acknowledge a research grant from the sa national research foundation .
we thank an anonymous referee for their comments and specifically for raising the question of the helium abundance . this paper uses observations made at the south african astronomical observatory ( saao ) .
bensby t. , yee j. c. , feltzing s. , et al .
, 2013 , apj , 763 , 141 bernhard a. , hmmerich s. , 2013 , oejv , 159 , 1 blanco v. m. , mccarthy m. f. , blanco b. m. , 1984 , aj , 89 , 636 blitz l. , spergel d. n. , 1991 , apj , 379 , 631 boothroyd a. i. , sackmann i .- j . , ahern s. c. , 1993 , apj , 416 , 762 buell j. f. , 2012 , mnras , 428 , 2577 cao l. , shude m. , nataf d. , rattenbury n. j. , gould a. , 2013 , mnras , 434 , 595 carpenter j. m. , 2001 , aj , 121 , 2851 carter b. s. c. , 1990 , mnras , 242 , 1 catchpole r. m. , 1992 , in : ( ed . ) warner b. , variable stars and galaxies , asp conference series , 30 , 295 clarkson w. i. , sahu k. c. , anderson j. , et al .
2011 , apj , 735 , 37 cole a. a. , weinberg m. d. , 2002 , apj , 574 , l43 dkny i. , minniti d. , catelan m. , zoccali m. , saito r. k. , hempel m. , gonzalez o. a. , 2013 , aj , 776 , l19 drimmel r. , cabrera - lavers a. , lopez - corredoira m. , 2003 , a&a , 409 , 205 feast m. w. , 1986 , in : light on dark matter , noordwijk , reidel publishing co. , p. 339 feast m. w. , 2008 , in : first middle east - africa , regional iau meeting , cairo , egypt .
feast m. w. 2009 , in : agb stars and related phenomena ( ueta t. , matsunaga , n. & ita , y. , eds . ) , agb stars and related phenomena , p. 48 feast m. w. , 2013 , in : ( eds . ) oswalt t. w. , gilmore g. , planets , stars and stellar systems , vol . 5 : galactic structure and stellar populations , springer science+business media dordrecht , p 829 feast m. w. , menzies j. w. , whitelock p. a. , 2013 , mnras , 428 , 36 feast m. w. , robertson b. s. c. , catchpole r. m. , lloyd evans t. , glass i. s. , carter b. s. 1982 , mnras , 201 , 439 feast , m. w. , glass , i. s. , whitelock , p. a. , catchpole , r. m. , 1989 , mnras , 241 , 375 feast m. w. , whitelock p. a. , 2000 , mnras , 317 , 460 ferraro f.r .
dalessandro e. , mucciarelli a. , et al . , 2009 , nature , 462 , 483 gaposchkin s. , 1955 , var .
ussr , 10 , 337 gerhard o. , martinez - valpuesta i. , 2012 , apj , 744 , l8 glass i. s. , 1986 , mnras , 221 , 879 glass i. s. , 1999 , handbook of infrared astronomy .
cambridge univ . press , cambridge glass i s. , feast m. w. , 1982 , mnras , 198 , 199 glass i. s. , whitelock p. a. , catchpole r. m. , & feast , m. w. , 1995 , mnras , 273 , 383 gonzalez o.a . , rejkuba m. , minniti d. , zoccali m. , valenti e. , , saito r. k. , 2011 , a&a , 534 , l14 gran f. , minniti d. , saito r. k. , navarrete c. , dkny i. , mcdonald i. , contreras ramos r. , catelan m. , 2015 , a&a , 575 , 114 groenewegen m. a. t. , blommaert j. a. d. l. , 2005 , a&a , 443 , 143 iben i. , renzini a. , 1983 , araa , 21 , 271 immeli a. , samland m. , gerhard o. , westera p. , 2004
, a&a , 413 , 547 ishihara d. , kaneda h. , onaka t. , ita y. , matsuura m. , matsunaga m. , 2011 , a&a , 534 , a79 karakas a. i. , 2014 , mnras , 445 , 347 lee y .- w . , joo s .- j . ,
chung c. , 2015 , mnras , 453 , 3906l ( arxiv:1508.05942 ) lloyd evans t. h. h. , 1976 , mnras , 174 , 169 lpez - corredoira m. , cabrera - lavers a. , mahoney t. j. , hammersley p. l. , garzn f. , gonzlez - fernndez c. , 2007 , aj , 133 , 154 massari d. , dalessandro e. , ferraro f.r . , 2015 , apj , 810 , 66 matsumoto t. , hayakawa s. , koizumi k. , murakami h. , uyama k. , yamagami t. , thomas j. a. , 1982 , aip conf .
proc . , 83 , 48 matsunaga n. , 2006 , phd thesis , university of tokyo matsunaga n. , fukushi h. , nakada y. , 2005 , mnras , 364,117 matsunaga n. , kawadu t. , nishiyama s. , nagayama t. , hatano h. , tamura m. , glass i. s. , nagata t. , 2009 , mnras , 399 , 1709 mcwilliam a. , zoccali m. , 2010 , apj , 724 , 1491 menzies j. w. , whitelock p. a. , feast m. w. , 2015 , mnras , 452 , 910 nataf d. m. , 2015 , pasa , in press , arxiv1509.00023v1 nataf d. m. , gould a. p. 2012 , apj , 751 , l39 nataf d. m. , udalski a. , skowron j. , et al .
, 2015 , mnras , 447 , 1535 nishiyama s. , nagata t. , baba d. , et al .
, 2005 , apj , 621 , l105 nishiyama s. , nagata t. , kusakabe n. , et al . , 2006 , apj , 638 , 839 nishiyama s. , tamura m. , hatano h. , kato d. , tanabe t. , sugitani k. , nagata t. , 2009 , apj , 696 , 1407 noguchi m. , 1999 , apj , 514 , 75 ojha , d. k. , tej , a. , schultheis , m. , omont , a. , schuller , f. , 2007 , mnras , 381 , 1219 oosterhoff p. th .
, ponsen j. , schuurman m. c. , 1967 , bull .
netherlands supp .
, 1 , 397 oosterhoff p. th .
, ponsen j. , 1968 , bull .
netherlands supp . , 3 , 79 pietrukowicz p. , et al .
, 2012 , apj , 750 , 169 pietrukowicz p. , et al . , 2015 , apj , 811 , 113 renzini a. , greggio l. , 1990 , in : eso / ctio workshop on bulges of galaxies , la serena , chile , p. 47
rich r. m. 2013 , in : ( eds . ) oswalt t. w. , gilmore g. , planets , stars and stellar systems , vol . 5 : galactic structure and stellar populations , springer science+business media dordrecht , p. 272
robin a. c. , reyl c. , derrire s. , picaud s. , 2003 , a&a , 409 , 523 schlafly e. f. , finkbeiner d. p. , 2011
, apj , 737 , 103 schlegel d. j. , finkbeiner d. p. , davis m. , 1998 , apj , 500 , 525 sevenster m. , 1999 , mnras , 310 , 629 schwarzenberg - czerny a. , 1996 , apj , 460 , l107 soszynski i. , udalski a. , szymanski m. k. , et al . , 2013 , acta astronomica , 63 , 21 stellingwerf r.f . , 1978 , apj , 224 , 953 uttenthaler s. , blommaert j. a. d. l. , wood p. r. , lebzelter t. , aringer b. , schultheis m. , ryde n. , 2015 , mnras , 451 , 1750 vanhollebeke e. , groenewegen m. a. t. , giradi l. , 2009 , a&a , 498 , 95 vsquez s. , zoccali m. , hill v. , et al . , 2013 ,
a&a , 555,91 wegg c. , gerhard o. , 2013 , mnras , 435 , 1874 whitelock p. a. , 1987 , pasp , 99 , 573 whitelock p. a. , 1992 , in : ( ed . ) warner b. , variable stars and galaxies , asp conference series 30 , p. 11
whitelock p. a. , 1993 , in : ( eds . ) de jonghe h. , habing h. j. , iaus 153 galactic bulges , kluwer academic publishers , p. 39 whitelock p. a. , 2013 , in : r de grijs ( ed . ) advancing the physics of cosmic distances , iaus , 289 , 209 whitelock p. a. , catchpole r. m. , 1992 , in : ( ed . )
blitz l. , the centre , bulge and disk of the milky way , kluwer academic publishers , p. 103
whitelock p. a. , feast m. w. , catchpole r. m. , 1991 , mnras , 248 , 276 whitelock p. a. , feast m. w. , van loon , j. t. h. , zijlstra , a. a. , 2003 , mnras , 342 , 86 whitelock p. a. , feast m. w. , van leeuwen f. , 2008 , 386 , 313 whitelock p. , marang f. , feast m. w. , 2000 mnras , 319 , 728 wood p. r. 2000 , pasa , 17 , 18 zoccali m. , renzini a. , ortolani s. , et al . , 2003 , a&a , 399 , 931 | we report periods and @xmath0 observations for 648 oxygen - rich mira variables found in two outer bulge fields at @xmath1 and @xmath2 and combine these with data on 8057 inner bulge miras from the ogle , macho and 2mass surveys , which are concentrated closer to the galactic centre .
distance moduli are estimated for all these stars .
evidence is given showing that the bulge structure is a function of age . the longer period miras ( @xmath3 , age @xmath4 gyr and younger ) show clear evidence of a bar structure inclined to the line of sight in both the inner and outer regions .
the distribution of the shorter period ( metal - rich globular cluster age ) miras , appears spheroidal in the outer bulge . in the inner region
these old stars are also distributed differently from the younger ones and possibly suggest a more complex structure .
these data suggest a distance to the galactic centre @xmath5 , of 8.9kpc with an estimated uncertainty of @xmath6kpc . the possible effect of helium enrichment on our conclusions is discussed .
stars : agb and post - agb ; galaxy : bulge ; stars : variables : general ; infrared : stars ; galaxy : structure . |
one of the primary goals of the large hadron collider ( lhc ) , a proton - proton collider with a center of mass energy @xmath2 tev , will be to investigate the sector responsible for the breaking of the electroweak symmetry . in the standard model ( sm ) ,
a single higgs doublet is responsible for the spontaneous breakdown of the @xmath3 gauge group to @xmath4 .
the coupling constants of the sole physical higgs scalar to the rest of the sm particles are completely determined by their masses , and consequently there is little guesswork involved in determining the most promising channels @xcite in which one might hope to discover such a scalar . for a relatively light ( @xmath5 ) sm higgs boson ,
those channels are @xmath6 and @xmath7 , while for an intermediate - mass ( @xmath8 ) higgs , the single most promising channel is the weak - boson fusion ( wbf ) @xcite process @xmath9 @xcite . for a heavier higgs , with @xmath10 , the most relevant channels are @xmath11 and @xmath12 , with the higgs produced via either gluon fusion or wbf @xcite . in models where the higgs sector differs significantly from that of the standard model , however , the situation can change dramatically .
this is true even in cases where the low - energy effective theory describing a given model at the weak scale contains only a single , light , @xmath0-even higgs scalar .
indeed , at low energies , many models with extended higgs sectors have effective descriptions that are `` standard - model - like '' in the sense that they contain a single light higgs boson , but one whose couplings to the standard model fermions and gauge bosons differ potentially significantly from those of a sm higgs .
such discrepancies , in turn , can translate into vast differences in lhc phenomenology : some ( or , in severe cases , even all ) of the standard detection channels for a sm higgs may disappear as a result of such modifications , while others , related to processes buried beneath background in the sm , may become crucial for discovery .
one set of channels which are not terribly significant for the discovery of a sm higgs , but could become so in models with modified higgs sectors , consists of those involving direct decays of the higgs boson to a pair of high-@xmath13 leptons . in the sm ,
a light higgs boson ( with mass @xmath14 gev ) decays predominantly into @xmath15 , and the ratio @xmath16 ( where @xmath17 ) is roughly proportional to @xmath18 , due to the fact that in the sm , the same higgs doublet is responsible for giving mass to both quarks and leptons .
consequently , attention has been focussed predominately on processes in which the higgs boson decays to a tau pair ( with a branching ratio of about 10% ) , and in particular on the weak - boson fusion process @xmath9 .
this is the only process particularly relevant for sm higgs discovery in which the higgs decays directly to leptons , though it is now regarded as one of the most promising discovery channels for a sm higgs in the intermediate mass region @xcite .
searching for the higgs in the @xmath19 and @xmath20 channels is more difficult , due to a combination of factors , including enhanced sm backgrounds and suppressed signal cross - sections .
by contrast , processes in which a sm higgs boson decays into first- or second - generation leptons are generally assumed to be irrelevant for discovery . this is because under the assumption of yukawa - coupling universality among the lepton generations ( an assumption we will be making throughout the present work ) , the small size of @xmath21 compared to @xmath22 results in @xmath23 being roughly two orders of magnitude smaller than @xmath24 , with @xmath25 nearly three orders of magnitude smaller still . consequently , the rates for processes involving @xmath26 and @xmath27 are extremely suppressed relative to those involving tau pairs , both in the sm and in most simple extensions of the higgs sector . on the other hand ,
there are strong motivations for considering processes of this sort at the lhc .
experimentally , a signal involving a pair of high-@xmath28 muons or electrons will be easy to identify , as the muon- and electron - identification efficiencies at each of the lhc detectors are each greater than 90%@xcite .
furthermore , once a higgs boson is discovered in these channels , its mass could be readily reconstructed with high precision .
such channels could also be of use in determining the higgs yukawa couplings to leptons .
two - higgs - doublet models ( 2hdm ) , which stand as perhaps the simplest , most tractable example of a non - minimal electroweak - symmetry - breaking sector , provide a useful context in which to study the role of leptonic higgs - decay processes .
these models arise in a number of beyond - the - standard - model contexts from supersymmetry to little higgs scenarios @xcite and have a rich phenomenology , many of whose consequences for lhc physics are still being uncovered . in general ,
2hdm can be categorized according to how the higgs doublets couple to the sm quarks and leptons . in what has become known as a type i 2hdm
, one doublet is responsible for the masses of both quarks and leptons , while the other decouples from the fermions entirely . in a type ii 2hdm , one higgs doublet couples to the up - type quark sector , while the other higgs doublet couples to both the down - type quark sector and the charged leptons as is the case , for example , in the minimal supersymmetric standard model ( mssm ) . in both of these standard scenarios , the leptonic branching ratios for
a light higgs do not differ much from their sm values throughout most of parameter space coupling is suppressed due to radiative corrections @xcite , and consequently @xmath29 becomes large .
] , since the same doublet gives masses to both the bottom quark and the charged leptons . one interesting alternative possibility , which will be the primary focus of the present work , is a 2hdm scenario in which one higgs doublet couples exclusively to ( both up- and down - type ) quarks , while the other couples exclusively to leptons a scenario which we will henceforth dub the leptophilic two - higgs - doublet model ( l2hdm ) .
this model has been discussed previously in the literature in relation to its effect on higgs branching fractions and decay widths @xcite , flavor physics @xcite , and potential implications for neutrino phenomenology @xcite and dark matter studies @xcite .
some analyses of the lhc phenomenology of the model were presented in ref .
@xcite , which focused on the non - decoupling region of the parameter space where additional physical higgs scalars are light . in this work ,
we discuss the leptonic decays of the lightest @xmath0-even higgs scalar in the l2hdm at the lhc .
in particular , we examine the discovery potential in a decoupling regime in which only one light scalar , which resembles the sm higgs , appears in the low - energy effective description of the model .
we begin in section [ sec : param ] by presenting the model and reviewing how the coupling structure of the lightest neutral higgs particle is modified from that of a sm higgs . in section [ sec : bounds ] , we discuss the applicable experimental constraints from flavor physics , direct searches , etc . and show that they still permit substantial deviations in the couplings between the higgs boson and the other sm fields away from their standard - model values . in section [ sec : brxsec ] , we discuss the implications of such modifications on the higgs branching ratios and production rates . in section [ sec : lhcsignatures ] , we discuss potential higgs discovery channels in which the higgs boson decays directly into a pair of charged leptons , and in section [ sec : combined ] , we calculate the discovery potential for a light , leptophilic higgs using the combined results from all of these leptonic channels . in section
[ sec : conclusion ] we conclude .
the l2hdm , as defined here , is a modification of the sm in which the higgs sector consists of two @xmath3 scalar doublets , both of which receive nonzero vacuum expectation values .
the first of these doublets , which we call @xmath30 , couples only to ( both up- and down - type ) quarks , while the other , which we call @xmath31 , couples only to leptons .
in other words , the yukawa interaction lagrangian is specified to be @xmath32 where @xmath33 , @xmath34 , and @xmath35 are @xmath36 yukawa matrixes , @xmath37 and @xmath38 respectively denote the left - handed quark and lepton fields , @xmath39 and @xmath40 respectively denote the right - handed up- and down - type quark fields , and @xmath41 denotes the right - handed lepton fields .
this coupling structure can be achieved by imposing a @xmath42 symmetry under which @xmath43 and @xmath41 are odd , while all the other fields in the model are even .
we will assume that this symmetry is broken only softly , by a term of the form @xmath44 in the scalar potential . in the l2hdm , that scalar potential takes the usual form common to all two - higgs doublet models .
assuming that there is no @xmath0-violation in the higgs sector , this potential can be parameterized as follows @xcite : @xmath45\label{eq : vstandard}\end{aligned}\ ] ] it is assumed that the parameters of the theory are assigned such that both @xmath30 and @xmath43 acquire nonzero vevs ( which we respectively denote @xmath46 and @xmath47 ) , and that @xmath48 .
we define @xmath49 as @xmath50 so that large @xmath49 corresponds to small @xmath47 , and therefore to large intrinsic lepton yukawa couplings . in the broken phase of the theory , the spectrum of the model includes the three massless goldstone modes which become the longitudinal modes of the @xmath51 and @xmath52 bosons , as well as five massive scalar degrees of freedom : two @xmath0-even fields @xmath53 and @xmath54 , a pseudoscalar @xmath55 , and a pair of charged fields @xmath56 .
the relationship between the physical @xmath0-even higgs scalars @xmath53 , @xmath54 and the real , neutral degrees of freedom in @xmath30 and @xmath43 is parameterized by the mixing angle @xmath57 : @xmath58\\ \mathrm{re}[\phi_q^0-v_q]\end{array}\right ) .
\label{eq : defalpha}\ ] ] in what follows , we will focus primarily on the physics of @xmath53 , the lightest of these two scalars .
since the potential given in eqn .
( [ eq : vstandard ] ) includes eight model parameters @xmath59 @xmath60 , @xmath61 , @xmath62 , and @xmath63 which are subject to the constraint @xmath64 , seven of these eight parameters may be considered free . in
what follows , it will be useful to work in a different , more physically meaningful basis for these parameters : @xmath65 where @xmath66 , @xmath67 , @xmath68 , and @xmath69 are the masses of the corresponding physical higgs scalars .
in order to study the collider phenomenology of the l2hdm , it will be necessary to characterize how the effective couplings between @xmath53 and the sm fields differ from their sm values .
( [ eq : yuks ] ) indicates that the effective couplings between the fermions and @xmath53 are given in terms of these mixing angles ) and ( [ eq : defalpha ] ) used in defining @xmath57 and @xmath70 , and hence frequently differ from source to source within the literature . ] by @xmath71 following @xcite , we can define a set of parameters @xmath72 which represent the ratios of these effective couplings to their sm values . at tree level , @xmath73 similarly , one can also define @xmath74-parameters for the trilinear couplings of @xmath53 with the electroweak gauge bosons , with the result that @xmath75 since a certain set of effective couplings whose leading contributions occur at one loop namely @xmath76 and @xmath77 are also relevant to the collider phenomenology of higgs bosons , it is worth deriving @xmath74-factors for them as well .
the effective operators that give rise to @xmath76 and @xmath77 are @xcite @xmath78 @xmath79
where @xmath80 , @xmath81 is the electric charge of quark @xmath82 , and @xmath83\label{eq : f12def } \\
f_1(\tau ) & = & 2 + 3\tau+3\tau(2-\tau)f(\tau)\end{aligned}\ ] ] and @xmath84 ^ 2 & \tau<1 \end{array}\right.\ ] ] with @xmath85 .
when @xmath86 and @xmath87 are complex ( which occurs when @xmath88 ) , it corresponds to internal lines going on shell .
this allows us to define a scaling factor for each of these effective vertices : @xmath89 since @xmath90 has an overall @xmath91 prefactor ( from the @xmath92 ) , the contribution from top quarks running in the loops will still dominate over the contribution from leptons unless @xmath93 ; thus the lepton loops generally can be neglected .
it is worth noting that since the effective higgs - gluon - gluon coupling receives contributions solely from quark loops , @xmath94 to leading order in @xmath95 , whereas @xmath96 depends on @xmath97 , @xmath98 , and @xmath99 in a nontrivial way .
the mixing angles @xmath57 and @xmath70 are constrained by several theoretical consistency conditions , as well as a number of experimental constraints .
we will put off discussion of the latter until section [ sec : bounds ] and focus on the former .
first of all , we require that the higgs sector not be strongly coupled , in the sense that all @xmath59 may be considered perturbatively small ( i.e. @xmath100 for all @xmath101 ) and that the @xmath102-matrix satisfies all relevant tree - unitarity constraints .
this implies that the quartic couplings @xmath59 appearing in eqn .
( [ eq : vstandard ] ) must satisfy @xcite @xmath103 perturbativity constraints also apply to the yukawa couplings @xmath104 , @xmath105 , and @xmath106 appearing in eqn .
( [ eq : yuks ] ) , which are modified from their sm values according to eqn .
( [ eq : hffcouplings ] ) .
however , since @xmath104 and @xmath105 are suppressed relative to their sm values rather than enhanced when @xmath107 ( the case of interest here ) , no stringent constraints arise on account of such modifications .
in addition to these perturbativity constraints , we must also require that the scalar potential given in eqn .
( [ eq : vstandard ] ) is finite at large field values and contains no flat directions .
these considerations translate into the bounds @xcite @xmath108 in this work , we will be primarily interested in examining situations in which the additional physical scalars @xmath56 , @xmath54 , and @xmath55 are heavy enough to `` decouple '' from the collider phenomenology of the theory in the sense that the only observable signals of beyond - the - standard - model physics at the lhc at low luminosity involve the light @xmath0-even scalar @xmath53 . for our present purposes , it will be sufficient to define our `` decoupling regime '' by the condition that @xmath109 , where @xmath110 is some high scale .
of course this regime includes the strict decoupling limit in which @xmath111 and the mixing angles satisfy the condition @xmath112 .
however , it also includes substantial regions of parameter space within which the values of @xmath57 and @xmath70 deviate significantly from this relationship .
3.5 truein the extent of parameter space allowed according to our definition of the decoupling regime is illustrated in fig .
[ fig : sinatanbexcl ] .
this figure shows the decoupling regions of @xmath113 - @xmath49 parameter space in which all of the aforementioned constraints are satisfied for a variety of different values of @xmath110 .
contours corresponding to @xmath114 gev , @xmath115 gev , and @xmath116 tev are displayed , along with a dash - dotted line representing the pure decoupling limit , where @xmath117 and @xmath112 .
the contours in fig .
[ fig : sinatanbexcl ] were obtained by fixing @xmath66 to a particular value ( @xmath118 gev ) and surveying over the remaining parameters . a given combination of @xmath113 and @xmath49 is considered to be `` allowed '' in this sense as long as there exists some combination of model parameters for which @xmath109 , and for which all of the constraints in eqs .
( [ eq : pertbounds ] ) and ( [ eq : vacstabbounds ] ) are simultaneously satisfied .
it is readily apparent from the figure that sizable regions of parameter space exist within which all constraints are satisfied , yet the masses of all scalars other than @xmath53 are large enough to effectively decouple from the low - energy effective description of the model .
it is also apparent that for @xmath119 tev , the decoupling region , as we have defined it , approaches the pure decoupling limit .
2.1 truein 2.1 truein 2.1 truein it is interesting to inquire to what extent the effective higgs couplings can be modified in the decoupling regime without running afoul of the aforementioned constraints . in the three panels shown in fig . [ fig : sinatanbetacont ] we plot a number of contours in @xmath113 - @xmath49 parameter space corresponding to different values of @xmath120 ( left ) , @xmath97 ( center ) and @xmath121 ( right ) . on each panel , we have also superimposed the @xmath122 gev contour from fig . [
fig : sinatanbexcl ] .
it is evident from these plots that while @xmath123 in the @xmath111 limit , large regions of parameter space are still allowed in the decoupling regime within these @xmath74-factors can deviate substantially from unity .
the message , then , is that the effective couplings of a light higgs boson in the decoupling regime do not have to approximate those which correspond to the pure decoupling limit , in which they approach those of a sm higgs . on the contrary , a wide variety of possibilities for the mixing angles @xmath57 and @xmath70 are still open in this regime , and as we shall soon see , many of these possibilities have a dramatic effect on in the collider phenomenology of the scalar @xmath53 .
in addition to the perturbativity and vacuum - stability bounds discussed in the previous section , the l2hdm is constrained by additional considerations related to flavor - physics experiment , direct searches , etc .
we now proceed to investigate these constraints in an effort to show that they can easily be satisfied in the decoupling regime even in the region of parameter space most interesting for collider phenomenology , where @xmath49 is large and @xmath113 deviates substantially from zero . let us begin with those bounds related to direct searches for beyond - the - standard - model scalars at lep . the current direct detection bounds
( at 95% cl ) for the masses of charged and neutral @xmath0-odd higgs bosons , as reported by the particle data group @xcite , are @xmath124 and @xmath125 . these clearly present no problem for the model in the decoupling limit considered here . far more stringent constraints on models with more than one higgs doublet can be derived , however , from experimental limits on flavor - violating processes that receive contributions at the one - loop level from diagrams involving charged higgs bosons .
let us first consider flavor violation in the lepton sector , which is constrained by analyses of @xmath126 , @xmath127 , @xmath128 , and @xmath129 conversion in nuclei . in the absence of neutrino masses ,
the matrix of effective @xmath130 couplings is proportional to the charged - lepton mass matrix ; hence there is no additional source of lepton - flavor violation ( lfv ) . in the presence of neutrino masses
this is no longer true , and nonzero contributions to lfv processes arise at one loop due to diagrams with charged higgs bosons running in the loop .
however , it has been shown @xcite that the resultant flavor - violating amplitudes are several orders of magnitude below current experimental bounds .
therefore , even in cases in which the effective @xmath131 coupling receives a substantial @xmath49-enhancement factor , such sources of lfv will not present any phenomenological difficulties .
now let us turn to consider the situation in the quark sector , where , by contrast , flavor - violation rates can be sizeable .
this is because the effective @xmath132 couplings in two - higgs - doublet models include flavor - violating terms proportional to the off - diagonal elements of the cabibbo - kobayashi - maskawa ( ckm ) matrix @xmath133 : @xmath134d_jh^+ -\frac{\tan\beta}{v}m_{e_i}\bar{\nu}_i p_r e_ih^+ + h.c.,\ ] ] as a result , such models are constrained by experimental bounds on @xmath135 , @xmath136 , @xmath137 , @xmath138 , rare kaon decays , etc . , which translate into bounds on the model parameters relevant to the charged - scalar sector : @xmath69 and @xmath49 .
since the flavor mixing in the charged higgs couplings to the quark sector is proportional to @xmath139 , it is the region where both @xmath49 and @xmath69 are small which is most tightly constrained by these bounds .
the most stringent constraints are those associated with @xmath140 and with mixing in the @xmath141 and @xmath142 systems . in the l2hdm , the same higgs doublet couples to both up- and down - type quarks , just as it does in type i 2hdm @xcite ; hence the bounds on @xmath69 and @xmath49 due to flavor mixing in the quark sector will be essentially identical to those applicable in type i models .
we now turn to review the bounds from each of these processes , updating the results obtained in @xcite .
2.3 truein 2.3 truein + 2.3 truein 2.3 truein the first bounds we consider are those associated with the observed branching ratio for the flavor - violating decay @xmath140 .
the combined result from the cleo and belle experiments @xcite is @xmath143 this is consistent with the expected standard model result @xmath144 . in models with a non - minimal higgs sector , additional contributions to the amplitude for @xmath140 arise at the loop level from diagrams involving virtual charged higgs bosons , as discussed above .
these diagrams are compiled in fig .
[ fig : btosgammadiagrams ] ( sm contributions to this amplitude come from diagrams of the same sort , but with @xmath145 in place of @xmath146 . )
the rate for the process can be calculated in the usual manner . after incorporating the effect of qcd corrections ( which can be quite large @xcite )
, one finds that @xcite @xmath147 where @xmath148 is the coefficient of the effective operator @xmath149 in the conventions of ref .
@xcite , evaluated at the scale @xmath150 .
this coefficient takes the form @xmath151-
\frac{3x}{28}\left [ \left(\frac{\alpha_{3}(m_{w})}{\alpha_{3}(m_{b})}\right)^{28/23}-1\right]\right],\end{aligned}\ ] ] where the weak - scale amplitude function @xmath152 in the l2hdm is given by @xmath153 .
\label{eq : c7weak}\ ] ] in these formulae , @xmath154 and @xmath155 , @xmath156 , @xmath157 , @xmath158 is the fermi constant , @xmath133 are elements in the ckm matrix , and @xmath159 . the functions @xmath160 , @xmath161 , and @xmath162 , which represent the loop integral contributions to the @xmath140 amplitude , are given in @xcite .
the constraints on @xmath163 and @xmath49 from @xmath140 are displayed in fig .
[ fig : bsgammagraph ] .
each curve therein represents the value of @xmath135 for a given choice of @xmath69 as a function of @xmath49 .
note that for the case under consideration here , in which @xmath164 , the experimental constraints are satisfied as long as @xmath165 .
4 truein 2.3 truein 2.3 truein constraints on @xmath163 and @xmath49 can also be obtained from limits on the observed mixing in the mesonic @xmath141 and @xmath142 systems .
the diagrammatic contributions to @xmath141 mixing are shown in fig .
[ fig : mesonboxdiagram ] , and these contributions translate into shift in the mass - splitting @xmath166 between @xmath167 and @xmath168 . in the l2hdm , this splitting , including sm contributions , is given by @xcite @xmath169,\label{eq : deltambb}\end{aligned}\ ] ] where the @xmath170 are defined as below eqn .
( [ eq : c7weak ] ) , @xmath171 is the @xmath172-meson decay constant , @xmath173 is the `` bag factor '' ( which encompasses all deviations from the vacuum saturation approximation ) .
expressions for the factor @xmath174 , which accounts for qcd effects , along with the functions @xmath175 , @xmath176 , and @xmath177 can be found in @xcite . as for @xmath171 and @xmath173 ,
there is a good deal of uncertainty as to their precise numerical values . since they appear in eqn .
( [ eq : deltambb ] ) in the combination @xmath178 , it is easier simply to deal with the uncertainty in this single quantity .
estimates of @xmath178 have been made using a variety of lattice qcd sum rules in conjunction with experimental evidence on heavy meson decays from slac , and the uncertainties in their values depend on the summation methods employed and the assumptions made . following @xcite , we take the range of uncertainty to be @xmath179 instead of dealing with @xmath166 directly , it is easier to deal with the combination @xmath180 , since the time - integrated mixing probability in the @xmath141 system depends on this combination of variables .
the accepted experimental value for @xmath181 , as reported by the heavy flavor averaging group , is @xmath182 @xcite . using the observed lifetime of the @xmath167 meson ( @xmath183 ) and the expression in eqn .
( [ eq : deltambb ] ) , one may obtains a theoretical value for @xmath181 , which can be compared to this experimental result . in fig .
[ fig : mixinggraphs ] , we show how the @xmath141 mixing bound constrains @xmath163 and @xmath49 . as there is a large uncertainty in @xmath178 [ eqn .
( [ eq : uncertaintyfbrootb0 ] ) ] , and in fact one far larger than that associated with the measured value of @xmath181 , the theoretical prediction for a given choice of @xmath163 translates into a broad band in @xmath184 space , rather than a thin line . in fig .
[ fig : mixinggraphs ] , the upper and lower bounds of each such band are demarcated by a pair of thick , dark lines of the same type ( solid , dotted and dot - dashed ) . the thin , shaded , horizontal stripe represents the experimentally - allowed window .
if any part of this stripe falls within the band corresponding to a given value of @xmath69 for a given @xmath49 , that parameter combination is permitted by the @xmath138 constraint .
we see from the plot that this constraint only becomes relevant for very small values of @xmath185 , and thus is not particularly stringent .
3.75 truein similar calculations to those outlined above for the @xmath141 system can also be performed for mixing in the @xmath142 and @xmath186 systems @xcite .
in addition , limits can also be derived on the @xmath0-violating parameters @xmath187 and @xmath188 . however , due to large theoretical uncertainties in the hadronic matrix elements , the resulting bounds on new physics from these considerations are not particularly stringent in the l2hdm , especially when @xmath107 @xcite .
experimental bounds on leptonic charged - meson decays @xmath189 , @xmath190 , @xmath191 , @xmath192 etc . can also be used to constrain 2hdm @xcite . in general , the partial width for the leptonic decay of a given meson is modified by a @xmath49-dependent factor @xmath193 , which in many scenarios ( e.g. type ii models ) can be quite sizeable when @xmath49 is large @xcite . in the l2hdm , however , the @xmath193 are independent of @xmath49 due to the cancellation of the @xmath49 factors between the quark and the lepton couplings . as a result
the model is essentially unconstrained by these considerations .
experimental limits on the rates for leptonic decays such as @xmath194 can also constrain models with enhanced higgs couplings to leptons @xcite . however , such constraints only become relevant when the charged - higgs mass is @xmath195 or lower , or when @xmath49 is extremely large , and thus have little bearing on the decoupling regime studied here .
the above analysis shows that in the decoupling region ( as we have defined it ) , where @xmath196 gev , all constraints from direct charged - higgs searches , neutral meson mixing , @xmath0-violation , charged - meson decay , etc . can be satisfied as long as @xmath49 is greater than @xmath197 .
this is mainly due to the fact that in the l2hdm , there is no new source of flavor violation except the sm ckm matrix .
the effective couplings between @xmath56 and the sm quarks are proportional to @xmath139 , which implies that the most stringent constraints become weaker as @xmath49 increases .
thus , we conclude that experimental constraints from flavor violation , direct searches , etc . do not pose any significant issues for the l2hdm as long as the charged higgs scalars are heavy .
( indeed , a relatively low value of @xmath198 and a charged higgs light enough to be discovered at the lhc are by no means incompatible .
) this is true even in the region of parameter space most interesting for collider physics , in which both @xmath113 and @xmath49 are large , and the effective couplings between the lightest @xmath0-even higgs and the sm leptons differ drastically from their sm values .
we now turn to examine the effect of these coupling modifications on the production cross - sections and decay widths of a light higgs boson .
since the overall amplitudes for higgs decays into any two - particle final state @xmath199 scale as @xmath200 ( i.e. , the appropriate @xmath74-factor for that final state ) , the associated branching ratios scale like @xmath201 in order to provide a concrete example of the effect such a modification can have on higgs phenomenology , let us focus on a particular benchmark point : @xmath202 , @xmath203 , which we have indicated by a dot in fig .
[ fig : sinatanbexcl ] .
we pick this particular point as a benchmark because it yields only a moderate deviation from the sm couplings and is consistent with the bounds ( [ eq : pertbounds ] ) and ( [ eq : vacstabbounds ] ) when @xmath204 .
the @xmath74-factors corresponding to this particular point are @xmath205 fig .
[ fig : brplot ] illustrates the effect of this coupling - constant modification on the branching ratios of a light , @xmath0-even higgs scalar . in the left - hand panel , we have plotted the sm branching ratios for a number of higgs decay processes as a function of @xmath66 .
all branching ratios used in the construction of the figure were calculated using hdecay @xcite . in the right - hand panel , we have plotted branching ratios for the same set of processes in the l2hdm at our chosen benchmark point . it is evident that even this moderate modification of the couplings has a dramatic effect on the decay behavior of a light higgs : for example , the rates for @xmath206 and @xmath207 are on the same order . since @xmath208 is the dominant decay channel for a higgs boson with a mass in the range @xmath209 , this clearly represents a substantial effect on higgs phenomenology .
it is also worth noting that @xmath210 and @xmath211 are also on the same order for this choice of parameters .
this suggests that processes involving direct decays of a light higgs boson to a pair of high-@xmath13 muons could play an important role in the collider phenomenology of the light higgs @xmath212 a suggestion we will explore further in section [ sec : lhcsignatures ] . the branching ratios for a number of other decay channels relevant to the study of a light sm higgs boson at the lhc , such as @xmath213 , are clearly suppressed here relative to their sm values .
2.5 truein 2.5 truein the effect of the coupling modifications on the total higgs width is shown in fig .
[ fig : gamtotplot ] . here , we have plotted the ratio of the total higgs width @xmath214 to its sm value for three different points in the allowed region of @xmath113 @xmath212 @xmath49 parameter space as a function of @xmath66 .
the first of these points is our chosen benchmark ( @xmath202 , @xmath203 ) , for which @xmath214 ( indicated by the solid line ) is slightly lower than its sm value due to the suppression of @xmath215 when @xmath66 is small .
when @xmath66 increases and decays to electroweak gauge bosons begin to dominate the higgs width , @xmath214 drops even further , since @xmath216 at this point [ see eqn .
( [ eq : etasforbm ] ) ] .
the second of these points , ( @xmath217 , @xmath218 ) , is located very near the pure decoupling " line in fig . [
fig : sinatanbexcl ] ; hence for this point @xmath214 ( indicated by the dotted line ) is essentially equal to @xmath219 . at the third point ,
( @xmath220 , @xmath221 ) , a substantial enhancement in @xmath222 overcomes the suppression factor in @xmath215 , and consequently @xmath223 for @xmath224 gev ( as indicated by the dash - dotted line ) . for larger values of @xmath66
, gauge - boson decays once again dominate the higgs width , which becomes suppressed relative to its sm value . even in the most extreme cases permitted by the model consistency constraints outlined in section [ sec : param ] , however , @xmath225
this implies that the narrow - width approximation remains valid over the higgs mass range @xmath209 , which will be the mass region of primary focus of the present work . 4.0 truein
since we have shown that the narrow - width approximation to be valid , we can proceed in a straightforward manner from the decay width calculations above to determine how the cross - sections for full collider processes are modified . in this approximation
, one assumes that essentially all the higgs bosons produced in any such process are produced on - shell .
this allows one to approximate the cross - section for any process of the form @xmath226 by @xmath227 furthermore , if the sm production cross - section @xmath228 for the process is known , one can use the fact that @xmath229 to obtain the relation @xmath230 which allows us to calculate the cross - sections for these overall processes in the modified model . for the benchmark point that we have chosen ( @xmath202 , @xmath49=3 ) ,
the cross - sections for most of the conventional higgs search modes at the lhc are suppressed relative to their sm values , due to the suppressed higgs couplings to quarks and to gauge bosons .
many of the processes in which the higgs decays directly to charged - lepton pairs , on the other hand , are substantially enhanced .
we will discuss the implications these modifications can have for higgs searches in detail in section [ sec : combined ] .
one of the most interesting aspects of the l2hdm is that in certain regions of parameter space , new channels for the discovery of a light higgs boson can open up .
in particular , when the effective coupling between @xmath53 and the sm leptons is substantially increased while its couplings to sm quarks and/or electroweak gauge bosons are not dramatically suppressed , a number of processes in which the higgs boson decays directly to a pair of high-@xmath13 leptons can become far more important for the discovery of a light higgs than they are in the sm . in our analyses
, we focus on the discovery of @xmath53 in the light - to - intermediate - mass region @xmath231 . for heavier higgs bosons
, @xmath232 dominates and leptonic higgs decays play a less important role . in the decoupling limit case studied here , in which the additional higgs scalars @xmath146 , @xmath54 , and @xmath55 are heavy , such processes might well constitute the only evidence for physics beyond the standard model accessible within the first @xmath233 of integrated luminosity at the lhc , and are therefore of crucial importance .
this situation is quite different from the one studied in ref .
@xcite , in which some of these additional scalars are light and play a significant role in collider phenomenology .
since the largest leptonic contribution to the higgs total width comes from @xmath234 , processes involving higgs decays directly to tau leptons will play a significant role in the collider phenomenology of the l2hdm .
however , the analysis of such processes is complicated by subtleties associated with tau decay .
each @xmath235 lepton can decay either leptonically or hadronically , with respective branching ratios @xcite , @xmath236 and @xmath237 .
we will henceforth denote a hadronically - decaying tau as @xmath238 and a leptonically decaying one as @xmath239 . for processes involving
@xmath0-even higgs boson decays into a @xmath240 pair , there are two final states which permit successful identification of both taus : @xmath241 and @xmath242 , where @xmath243 . final states resulting from fully hadronic decays have a large background from dijet processes with narrow jets misidentified as taus .
final states involving two leptons of like flavor ( @xmath244 and @xmath245 ) are also less useful due to the overwhelming sm background from @xmath246 processes .
a hadronically - decaying tau will decay into either a `` one - prong '' ( approximately 77% of the time ) or `` three - prong '' ( approximately 23% of the time ) final state .
these final states involve narrow , well - collimated jets including one or three charged pions , respectively .
the identification of a jet as coming from a hadronically - decaying @xmath235 , as opposed to some qcd process , is far from trivial .
one of the principal discrimination variables is jet radius @xmath247 ( see @xcite for more details regarding @xmath235 identification ) . at the tevatron ( run ii ) , @xmath248 for a @xmath249 gev cut . at the lhc , a @xmath235 identification efficiency of around @xmath250
is expected @xcite .
processes involving direct decays of @xmath53 to muon pairs can also be of interest for higgs discovery in the l2hdm .
the disadvantage of such channels for higgs searches , relative to those involving direct decays to taus , is the suppressed branching ratio . since yukawa coupling universality dictates that @xmath251 , both in the sm and in the l2hdm , @xmath252 .
however , this is compensated for to a great extent by the fact that the dimuon signal is exceptionally clean .
indeed , the muon identification efficiency at the lhc is more that 90% @xcite . in addition
, the measurement of muon momenta allows for a precise reconstruction of the higgs mass within @xmath253 gev .
this permits the implementation of an extremely efficient cut on @xmath254 , the invariant mass of the muon pair , and a substantial reduction in background levels for all channels involving direct higgs - boson decays to muon pairs .
we now turn to address the prospects for detecting a light sm - like @xmath0-even higgs boson at the lhc on a channel - by - channel basis . in the present work , as discussed in section [ sec : param ] , we will assume generation universality among the lepton yukawa couplings . therefore , we will ignore the @xmath255 channel and focus only on @xmath256 and @xmath26
. the channels of primary interest , then , are those in which the higgs is produced by gluon fusion , weak - boson fusion , or @xmath257 associated production and decays to either @xmath258 or @xmath259 .
associated @xmath51 and @xmath52 production processes generally have smaller rates , but may also potentially be of interest , and as such we briefly discuss them as well . bottom - quark - fusion processes with a leptonically - decaying higgs boson
@xcite , while potentially interesting for type ii 2hdm in which the @xmath260 vertex receives a large @xmath49-enhancement , are less useful in the l2hdm , since the effective down - type quark couplings are suppressed in that scenario rather than enhanced . in this section , we briefly summarize the results of the existing studies of the leptonic - higgs - decay channels at the lhc , with an eye toward their utility for the discovery of a leptophilic higgs .
we begin with a discussion of the weak - boson - fusion process @xmath261 , which is the only channel involving direct higgs - boson decay to a pair of charged leptons that contributes significantly to the higgs discovery potential in the sm .
indeed , it is a particularly promising channel for sm higgs discovery in the intermediate mass region ( @xmath8 ) @xcite . discriminating between signal and sm background
can be facilitated by requiring that events have two leading tagging jets in the forward - backward direction and imposing a minijet veto in the central region of the detector .
a great deal of attention has been devoted to this channel , with an emphasis on @xmath262 and @xmath263 final states . combining all channels , a statistical significance of more than @xmath264
can be reached for higgs masses around 120 @xmath212 130 gev with @xmath265 of integrated luminosity at atlas @xcite .
the detection prospects are similar at cms @xcite . the prospects for detecting a light , sm higgs boson produced by gluon fusion and decaying to @xmath259 at the tevatron were examined in ref .
@xcite . in order to effectively reconstruct the higgs mass from the various final - state particles produced during tau decay ,
it is necessary to focus on events in which the transverse momentum of the tau pair is nonzero ; hence the authors elected to focus on the process @xmath267 . taking into account both the @xmath268 and @xmath269 ratios
, @xmath262 turns out to be the most promising channel for signal identification , but that an integrated luminosity of @xmath270 would be needed at the tevatron in order to exclude a 120 gev sm higgs boson at the 95% c.l .
however , preliminary studies at atlas @xcite indicate that this will be a promising channel in which to look for a higgs boson with enhanced coupling to leptons at the lhc .
this process was examined in a standard model context in @xcite . in order to be able to reconstruct the two top quarks effectively , the authors restricted their analysis to cases in which one of the @xmath271 bosons produced during top decay decays leptonically , while the other decays hadronically .
only events with hadronic tau decays were considered , as reconstructing both tops proves to be slightly easier in this scenario .
thus the overall process of interest is @xmath272 .
since the production cross section drops quickly with increased higgs mass , this channel is only important when the higgs is light . for @xmath66 around 120 gev , a statistical significance of @xmath273
can be obtained with @xmath274 of integrated luminosity . in @xcite
, semileptonic tau decays were considered in particular , decays of the form @xmath275 , and it was found that such a process could provide evidence for a 120 gev higgs boson at the @xmath276 level for an integrated luminosity of @xmath233 .
the weak boson fusion process @xmath278 was analyzed in @xcite . after the appropriate cuts on the tagging jets are imposed , the leading sm background comes from irreducible @xmath279 or @xmath280 processes , with the @xmath281 decaying to muon pairs . due to the extremely suppressed sm @xmath282 branching ratio ,
an integrated luminosity of @xmath283 or more is generally required to claim a @xmath284 discovery for a higgs mass less than 140 gev . in the l2hdm
, however the detection prospects can be substantially improved if @xmath285 and @xmath286 .
the prospects for the detection of a light higgs boson of @xmath288 produced by gluon fusion and decaying directly into @xmath258 were discussed in @xcite .
the irreducible background for @xmath287 is dominated by the drell - yan processes @xmath289 .
the sharp invariant mass resolution of the muon pair allows for a substantial reduction in this background via a stringent cut on @xmath254 .
consequently , a significance level similar to that in the wbf channel as discussed in @xcite can be attained in this channel as well .
the prospects for discovering a standard model higgs boson in the @xmath290 channel were recently studied in @xcite .
this channel tends to be more important when the higgs mass is light ( around 120 gev ) since the production cross section drops quickly for a heavier higgs .
the primary irreducible backgrounds , which come from @xmath291 and @xmath292 production , with the @xmath281 decaying into a muon pair , can be reduced quite effectively by a cut on that muon pair s invariant mass .
additional , reducible backgrounds such as @xmath293 can be effectively eliminated by reconstructing the masses of both top quarks , which is possible in the case where the tops decay either fully hadronically or semileptonically .
the statistical significances for the @xmath290 channel are of roughly the same order as those in the @xmath287 and @xmath278 channels , and hence could contribute significantly to the discovery potential for a light higgs scalar with enhanced couplings to leptons .
higgs production via the processes @xmath296 and @xmath297 could also potentially play a role in the discovery of a leptophilic higgs , though the prospects in these channels are not as favorable as the other , aforementioned ones .
sm cross - sections for these processes , taking into account the leptonic decay of the higgs boson , are given in table [ table : whzh ] for the case in which @xmath298 gev .
these were determined from leading - order results obtained using madgraph @xcite and modified by the appropriate @xmath299-factors : @xmath300 for signal @xcite , @xmath301 for background @xcite . for processes in which the higgs decays to @xmath258 ,
the signal is clearly too small to be of any use .
however , for processes involving decays to @xmath259 , the signal is only about a factor of @xmath302 smaller than the background . by optimizing cuts to eliminate the sm background
, this channel might potentially be of use particularly if @xmath303 is enhanced , as in the l2hdm .
little analysis of these processes exists in the literature , and we leave the detailed study of these channels for future work .
.sm production cross sections at the lhc for @xmath304 , @xmath305 associated production , with @xmath53 decays into muon pair or tau pair .
the higgs mass is taken to be 120 gev .
also shown are the sm background @xmath306 , @xmath307 with one @xmath52 decays into muons or taus .
the numbers are obtained using madgraph @xcite . [ cols="^,^,^,^,^,^",options="header " , ]
now that we have discussed the channels in which one might look for a leptonically - decaying higgs boson at the lhc , let us investigate the prospects for the discovery of such a higgs boson in the l2hdm , using the combined results from all channels discussed above ( excepting the @xmath304 , @xmath305 channels , which we have shown do not contribute significantly to the discovery potential ) .
in particular , we focus on the region of @xmath113 - @xmath49 parameter space in which @xmath120 is large and @xmath308 . in this case , the cross - sections for processes involving a @xmath309 coupling are substantially increased , while those for processes involving @xmath310 , @xmath311 , or @xmath76 are only slightly reduced .
as before , for purposes of illustration , we will focus on the benchmark point ( @xmath202 , @xmath203 ) , which exemplifies this situation nicely . in fig .
[ fig : sigplot ] , we show the effect of the coupling - constant modifications on the discovery potential of a light higgs boson for this particular benchmark point . in the right - hand panel ,
the statistical significance associated with each of the relevant leptonic channels discussed in section [ sec : lhcsignatures ] is displayed as a function of higgs mass for our chosen benchmark point in the l2hdm .
the sm results for the same processes are shown in the left - panel for comparison .
the results in each panel correspond to an integrated luminosity of @xmath312 .
it is apparent from fig .
[ fig : sigplot ] that @xmath313 is one of the most promising detection channels for the chosen benchmark point in the l2hdm , as in the sm . for this particular choice of parameters ,
@xmath314 and @xmath214 does not deviate drastically from @xmath219 ( see fig .
[ fig : gamtotplot ] ) , and consequently the overall significance level in this channel is essentially unchanged from its sm value .
however , in other regions of parameter space , drastic amplifications can occur : for example , the choice ( @xmath315 , @xmath221 ) results in a amplification of the statistical significance for the same process by a factor of @xmath316 .
it should also be noted that in the ( @xmath317 , @xmath203 ) case , the significance levels for both @xmath266 and @xmath318 also exceed @xmath264 .
the processes in which the higgs decays to muons are statistically less significant , but also provide strong evidence at the @xmath284 level with @xmath319 100 @xmath320 of integrated luminosity . indeed , the evidence for such a higgs boson would be dramatic and unmistakable . furthermore , once the higgs is observed in any of the muonic channels , the excellent invariant - mass resolution of the muon pairs can be used to determine the value of @xmath66 with a very high degree of precision . 2.75 truein 2.75 truein 2.75 truein 2.75 truein while the significances in those channels which involve a leptonically - decaying higgs can potentially be amplified in l2hdm , those in other channels useful for the detection of a sm higgs may be substantially suppressed .
this is illustrated in fig .
[ fig : sigplotother ] , which shows the significance of discovery in each individual channel which contributes meaningfully to the discovery potential of a sm higgs boson in the low to intermediate - mass region , both in the sm ( left - hand panel ) and in the l2hdm at the benchmark point ( @xmath202 , @xmath203 ) ( right - hand panel ) . in the latter case
, there is no single , non - leptonic channel in which evidence for the higgs boson can be obtained at the @xmath264 level . to further illustrate the point , in fig .
[ fig : sigscomb ] , we display the combined statistical significances for the leptonic channels discussed in section [ sec : lhcsignatures ] , as well as the combined significances for all other relevant channels for higgs discovery , both in the sm and in the l2hdm at the benchmark point ( @xmath317 , @xmath203 ) .
indeed , for this particular parameter choice , all relevant non - leptonic channels are suppressed relative to their standard - model to such an extent that , for most of the @xmath321 mass window displayed in the plot , their combined significance does not even provide @xmath284 evidence for much less a @xmath264 discovery of a light higgs boson . on the other hand , statistical significance for leptonic higgs decay channels
are enhanced , therefore becoming the dominant discovery channels for the light @xmath0-even higgs in the l2hdm model .
this clearly illustrates the crucial role leptonic channels can play in the lhc phenomenology of models with extended ( and particularly leptophilic ) higgs sectors . 3.5 truein we emphasize that these plots represent the results for a single benchmark point , and one in which the @xmath74-factors are not particularly extreme .
there exist other points in the parameter space of the model allowed by all constraints for which the deviations of the effective couplings of @xmath53 to the other fields in the theory are even more severe . as an example , consider the case in which @xmath322 and @xmath323 , for which @xmath324 , @xmath325 , and @xmath326 .
for this choice of parameters , most of the standard higgs discovery channels those involving @xmath327 and @xmath328 , as well as all weak - boson - fusion processes not involving direct higgs decays to leptons are strongly suppressed ; furthermore , other contributing channels such as @xmath329 and @xmath7 are also moderately suppressed .
in such a case , the leptonic channels discussed in section [ sec : lhcsignatures ] especially ones such as @xmath330 , which do not involve a direct coupling between @xmath53 and the electroweak gauge bosons may well constitute the only observable evidence of the higgs boson , and would thus be crucial for its discovery at the lhc .
the phenomenology of a light higgs boson in two - higgs - doublet models can differ drastically from that of a sm higgs . in this work
, we have focussed on one particularly interesting example : a leptophilic 2hdm , in which different higgs bosons are responsible for giving masses to the quark and lepton sectors .
we have examined the effect of such a modification on the collider phenomenology of a light higgs boson in a decoupling regime in which the only light scalar is a standard - model - like higgs boson , and have shown that a number of collider processes involving the direct decay of the higgs to a pair of charged leptons can play a crucial role in its discovery . in particular , we have shown that there are regions of parameter space in which the higgs - boson couplings to leptons can be greatly enhanced .
this can have a potentially dramatic effect on the higgs discovery potential , as signals involving direct , leptonic decays of the higgs can be substantially amplified . at the same time , signals in some ( or in some cases , even all ) of the other conventional channels useful for the detection of a standard model higgs boson can suffer a dramatic suppression . even when coupling modifications are not severe , leptonic decay processes will also play an important rule in differentiating between the higgs sector of the standard model and that of other , more complicated scenarios .
we would like to thank keith dienes , tao han , chung kao and lisa randall for fruitful discussions and correspondence .
we would also like to thank the kavli institute of theoretical physics at santa barbara for its hospitality during the completion of this work .
this work was supported in part by the department of energy under grant de - fg02 - 04er-41298 .
note added : after the completion of the work reported in this paper , a number of papers @xcite appeared which discuss the phenomenology of the l2hdm .
@xcite gives a brief presentation on the effect of effective - coupling modification in the decoupling regime .
their results agree with ours .
@xcite and @xcite focussed on the situation in which @xmath54 , @xmath55 and @xmath56 are light .
s. asai _ et al .
_ , eur .
j. c * 32s2 * , 19 ( 2004 ) [ arxiv : hep - ph/0402254 ] . s. abdullin _ et al .
_ , eur .
j. c * 39s2 * , 41 ( 2005 ) .
d. l. rainwater and d. zeppenfeld , jhep * 9712 * , 005 ( 1997 ) [ arxiv : hep - ph/9712271 ] .
d. l. rainwater , d. zeppenfeld and k. hagiwara , phys .
d * 59 * , 014037 ( 1999 ) [ arxiv : hep - ph/9808468 ] .
[ atlas collaboration ] , cern - lhcc-99 - 15 , atlas - tdr-15 .
g. l. bayatian _ et al . _
[ cms collaboration ] , j. phys .
g * 34 * , 995 ( 2007 ) .
n. arkani - hamed , a. g. cohen and h. georgi , phys .
b * 513 * , 232 ( 2001 ) [ arxiv : hep - ph/0105239 ] .
h. e. haber , m. j. herrero , h. e. logan , s. penaranda , s. rigolin and d. temes , phys .
d * 63 * , 055004 ( 2001 ) [ arxiv : hep - ph/0007006 ] ; m. s. carena , h. e. haber , h. e. logan and s. mrenna , phys .
d * 65 * , 055005 ( 2002 ) [ erratum - ibid .
d * 65 * , 099902 ( 2002 ) ] [ arxiv : hep - ph/0106116 ] .
v. barger , h. e. logan and g. shaughnessy , arxiv:0902.0170 [ hep - ph ] .
h. s. goh , l. j. hall and p. kumar , arxiv:0902.0814 [ hep - ph ] .
r. m. barnett , g. senjanovic , l. wolfenstein and d. wyler , phys .
b * 136 * , 191 ( 1984 ) .
r. m. barnett , g. senjanovic and d. wyler , phys . rev .
d * 30 * , 1529 ( 1984 ) .
v. d. barger , j. l. hewett and r. j. n. phillips , phys .
d * 41 * , 3421 ( 1990 ) .
y. grossman , nucl .
b * 426 * , 355 ( 1994 ) [ arxiv : hep - ph/9401311 ] .
m. aoki , s. kanemura and o. seto , arxiv:0807.0361 [ hep - ph ] .
j. f. gunion and h. e. haber , phys .
d * 67 * , 075019 ( 2003 ) [ arxiv : hep - ph/0207010 ] .
d. phalen , b. thomas and j. d. wells , phys . rev .
d * 75 * , 117702 ( 2007 ) [ arxiv : hep - ph/0612219 ] . j. f. gunion , h. e. haber , g. l. kane and s. dawson , _ the higgs hunter s guide _ , addison - wesley : redwood city , ca , 1990 .
a. g. akeroyd , a. arhrib and e. m. naimi , phys .
b * 490 * , 119 ( 2000 ) [ arxiv : hep - ph/0006035 ] ; a. arhrib , arxiv : hep - ph/0012353 ; i. f. ginzburg and i. p. ivanov , arxiv : hep - ph/0312374 .
w. m. yao _ et al .
_ [ particle data group ] , j. phys .
g * 33 * , 1 ( 2006 ) .
w. grimus and l. lavoura , phys .
d * 66 * , 014016 ( 2002 ) [ arxiv : hep - ph/0204070 ] . h. e. haber , g. l. kane and t. sterling , nucl . phys .
b * 161 * , 493 ( 1979 ) .
n. g. deshpande , p. lo , j. trampetic , g. eilam and p. singer , phys .
* 59 * , 183 ( 1987 ) .
s. bertolini , f. borzumati and a. masiero , phys .
lett . * 59 * , 180 ( 1987 ) ; b. grinstein and m. b. wise , phys .
b * 201 * , 274 ( 1988 ) ; b. grinstein , r. p. springer and m. b. wise , phys .
b * 202 * , 138 ( 1988 ) .
w. s. hou and r. s. willey , phys .
b * 202 * , 591 ( 1988 ) ; t. g. rizzo , phys .
d * 38 * , 820 ( 1988 ) ; j. l. hewett , s. nandi and t. g. rizzo , phys . rev .
d * 39 * , 250 ( 1989 ) .
b. grinstein and m. b. wise , phys .
b * 201 * , 274 ( 1988 ) .
j. franzini , phys .
* 173 * , 1 ( 1989 ) .
w. s. hou , phys .
d * 48 * , 2342 ( 1993 ) .
a. g. akeroyd and f. mahmoudi , arxiv:0902.2393 [ hep - ph ] . m. krawczyk and d. temes , eur .
j. c * 44 * , 435 ( 2005 ) [ arxiv : hep - ph/0410248 ] .
a. djouadi , j. kalinowski and m. spira , comput .
commun .
* 108 * , 56 ( 1998 ) [ arxiv : hep - ph/9704448 ] .
f. tarrade [ atlas collaboration ] , _ prepared for hadron collider physics symposium 2005 , les diablerets , switzerland , 4 - 9 jul 2005 _ ; a. f. saavedra [ atlas collaboration ] , atl - phys - proc-2009 - 007 .
v. d. barger and c. kao , phys .
b * 424 * , 69 ( 1998 ) [ arxiv : hep - ph/9711328 ] ; s. dawson , d. dicus and c. kao , phys .
b * 545 * , 132 ( 2002 ) [ arxiv : hep - ph/0208063 ] ; s. dawson , d. dicus , c. kao and r. malhotra , phys .
lett . * 92 * , 241801 ( 2004 ) [ arxiv : hep - ph/0402172 ] .
a. belyaev , t. han and r. rosenfeld , jhep * 0307 * , 021 ( 2003 ) [ arxiv : hep - ph/0204210 ] .
e. richter - was and t. szymocha , atl - phys-2004 - 012 .
t. plehn and d. l. rainwater , phys .
b * 520 * , 108 ( 2001 ) [ arxiv : hep - ph/0107180 ] .
t. han and b. mcelrath , phys .
b * 528 * , 81 ( 2002 ) [ arxiv : hep - ph/0201023 ] .
s. su and b. thomas , arxiv:0812.1798 [ hep - ph ] .
j. alwall _ et al .
_ , jhep * 0709 * , 028 ( 2007 ) [ arxiv:0706.2334 [ hep - ph ] ] .
o. brein , a. djouadi and r. harlander , phys .
b * 579 * , 149 ( 2004 ) [ arxiv : hep - ph/0307206 ] .
s. frixione , p. nason and g. ridolfi , nucl .
b * 383 * , 3 ( 1992 ) .
m. aoki , s. kanemura , k. tsumura and k. yagyu , arxiv:0902.4665 [ hep - ph ] . | in this work , we examine a two - higgs - doublet extension of the standard model in which one higgs doublet is responsible for giving mass to both up- and down - type quarks , while a separate doublet is responsible for giving mass to leptons .
we examine both the theoretical and experimental constraints on the model and show that large regions of parameter space are allowed by these constraints in which the effective couplings between the lightest neutral higgs scalar and the standard - model leptons are substantially enhanced .
we investigate the collider phenomenology of such a `` leptophilic '' two - higgs - doublet model and show that in cases where the low - energy spectrum contains only one light , @xmath0-even scalar , a variety of collider processes essentially irrelevant for the discovery of a standard model higgs boson ( specifically those in which the higgs boson decays directly into a charged - lepton pair ) can contribute significantly to the discovery potential of a light - to - intermediate - mass ( @xmath1 gev ) higgs boson at the lhc . |
probability distributions that take correlations into account , can be in some cases constructed by deformation of mathematical independent laws @xcite . along this line of approach , a variety of generalizations of the binomial distribution have been recently proposed @xcite .
the generalization consists in replacing the sequence of natural numbers that correspond to the variable of the original binomial distribution , by an arbitrary sequence of non - negative numbers .
consequently , their factorial and combinatorial numbers are redefined , in such a way that the simple powers of `` win '' ( `` loose '' ) probability must be replaced by characteristic polynomials whose degree is the number of wins ( loose ) .
these polynomials are obtained through generating functions that force the generalized distributions to still satisfy the conditions of normalization and non - negativeness .
resulting probabilities @xmath25 can be symmetrical or asymmetrical , and represent the probability of having @xmath6 wins and @xmath26 losses , in a sequence of @xmath7 _ correlated _ trials .
we shall be concerned with a particular set of _ symmetrically _ generalized binomial distributions , so as to preserve the win - loss symmetry , which is an essential prerequisite for the distribution to be used in the present analysis of strongly correlated systems and their entropic behavior .
the generating functions of this particular set are @xmath27-exponentials ( @xmath4^{1/(1-q^{gen})};\,e_{1}^z = e^z)$ ] , where _ gen _ stands for _ generating _ , and @xmath28 ( @xmath29 ) , @xmath2 being a parameter to be soon defined .
these generating functions are the only ones , besides the ordinary binomial case ( which corresponds @xmath30 ) , that yield probabilities which obey the leibnitz triangle rule ( see for instance @xcite ) .
the family of the generated probabilities depends on two parameters ( @xmath31 , @xmath2 ) , where @xmath31 is the `` win '' probability , @xmath32 is the loss " probability , and @xmath2 characterizes the generating function .
a variety of @xmath3-statistical functions @xcite related to the probabilities @xmath0 emerges , all of them univocally defined by @xmath24 , and whose respective @xmath33 indices appear to obey an algebra that reminds the underlying algebra in @xcite .
in fact , such probabilities provide a probability distribution function of the scaled quantity @xmath34 ( @xmath35 ) that is likely to achieve any of the complete set @xcite of bounded support @xmath8-gaussian @xmath36 ( where _ disc _ stands for _ discrete _ , and @xmath37 ) , where @xmath38 is a generalized inverse temperature ( @xmath39 ) .
the @xmath3-gaussian form corresponds to strongly correlated random variables , and arises from the extremization of the nonadditive entropy @xmath40 ( @xmath41 , @xmath42 , where bg stands for boltzmann - gibbs ) @xcite under appropriate constraints @xcite . since the proposal in @xcite , several statistical models which provide , in the @xmath22 limit
, @xmath3-gaussian attractors have been constructed @xcite .
some of them exhibit extensivity of the boltzmann - gibbs entropy @xmath43 , and at least one of them exhibits extensivity of the @xmath44 entropy for @xmath45 @xcite . in particular , the @xmath27-exponentially generated probabilities @xmath0 exhibit an extensive @xmath43 @xcite , and they appear to provide @xmath11-gaussian attractors ( where @xmath46 stands for _ attractor _ ) . an outstanding fact is that these probabilities have been rigorously deduced a priori , imposing a particular structure of the generating function ( see also @xcite ) under non - negativeness and normalization probability conditions .
in addition to these properties , these probabilities can be shown to correspond to the plya urn model @xcite .
the paper is organized as follows .
section [ sec2 ] presents the symmetric generalized binomial distribution introduced by h. bergeron et al .
@xcite , based on the @xmath27-exponential generating functions .
section [ sec3 ] is devoted to the characterization of the involved @xmath3-gaussian distributions ( where @xmath3 refers to @xmath8 for finite @xmath7 , and to @xmath11 in the @xmath22 limit ) in the generalized probability distribution of the ratio ( number of wins)/(number of realizations ) .
section [ sec4 ] describes large - deviation - like properties of the distribution of the scaled quantity @xmath47 ( @xmath48 ) .
section [ sec5 ] deals with the behavior of the large - deviation probabilities with regard to the @xmath10 limit distribution , which violates the law of large numbers .
we conclude in section [ conclusions ] .
in a sequence of @xmath49 independent trials with two possible outcomes , `` win '' and `` loss '' , the probability of obtaining @xmath6 wins is given by the binomial distribution : @xmath50 where the parameter @xmath31 ( @xmath51 ) is the probability of having the outcome `` win '' , @xmath32 corresponding to the outcome `` loss '' .
therefore , the bernoulli binomial distribution above preserves the symmetry win - loss .
let us now consider an strictly increasing infinite sequence of _ nonnegative real _
numbers @xmath52 . with each sequence @xmath53
defined above , a bernoulli - like distribution is constructed : @xmath54 where the factorials are defined as @xmath55 , @xmath56 , @xmath31 is a running parameter on the interval @xmath57 $ ] , and @xmath58 are polynomials of degree @xmath6 .
observe that the symmetry win - loss of binomial - like distribution ( [ binlike ] ) is preserved , as invariance under @xmath59 \to [ ( n - k),(1-\eta)]$ ] is verified .
that means that no bias can exist favoring either win or loss when @xmath9 .
the polynomials @xmath58 are to be defined , is such a way that quantities @xmath60 represent the probabilities of having @xmath6 wins and @xmath26 losses in a sequence of _ correlated @xmath7 trials_. consequently , @xmath60 must be constrained by the normalization equation : @xmath61,\ ] ] and the non - negativeness condition : @xmath62.\ ] ] different sets of polynomials can be associated with ( [ normal ] ) and ( [ nnegatv ] ) . with this aim ,
some generating functions of polynomials can be considered .
let us make use of a @xmath27-exponential generating function @xmath63^{1/(1-q^{gen})}$ ] .
such a generating function can be written as @xmath64 ( @xmath65 ) , where the @xmath3-exponential parameter that characterizes the generating function is @xmath66 .
the following probability distributions are consequently obtained @xcite : @xmath67 where @xmath68 is the pochhammer symbol , and @xmath69 .
observe that the @xmath70 limit recovers the ordinary binomial case ( @xmath30 ) .
the expectation value and the variance of ( [ distribevaldo ] ) are @xmath71 and @xmath72 , respectively @xcite . in the particular cases @xmath9 and @xmath73 ( i.e. , @xmath74 ) , we can also use the following equivalent expression : @xmath75\times\left[\displaystyle\prod_{j=1}^{\frac{\alpha}{2}-1 } \left(\frac{k+j}{j}\right ) \left(\frac{n - k+j}{j}\right ) \right ] & ( k\ne 0 , n ) \ , .
\end{array}\right . \label{improvenum}\ ] ] this expression is computationally very convenient .
the histograms @xmath76 ( @xmath77 ) are numerically obtained for fixed values of @xmath7 , @xmath31 and @xmath2 . fig .
[ fig1 ] shows that , for @xmath9 and @xmath70 , @xmath78 approaches the unbiased binomial distribution for all values of @xmath7 .
( @xmath79 ) in a sequence of @xmath80 correlated trials that follow the symmetric generalized binomial distribution .
the parameters of the model are @xmath9 , @xmath81 .
observe that the limit distribution tends to a binomial distribution as @xmath70 .
if , in addition to this limit , we consider @xmath22 , the distribution shrinks onto a dirac delta one .
[ fig1],width=302 ] , @xmath82 and @xmath83 .
_ left panel _ : the corresponding @xmath8-gaussian fitting function is superimposed .
_ right panel _ : the @xmath8-logarithmic representation exhibits that indeed the discrete ( i.e. , @xmath84 ) distribution is extremely close to a @xmath3-gaussian , the linear regression coefficient being @xmath85 .
[ fig2],title="fig:",width=264 ] , @xmath82 and @xmath83 .
_ left panel _ : the corresponding @xmath8-gaussian fitting function is superimposed .
_ right panel _ : the @xmath8-logarithmic representation exhibits that indeed the discrete ( i.e. , @xmath84 ) distribution is extremely close to a @xmath3-gaussian , the linear regression coefficient being @xmath85 .
[ fig2],title="fig:",width=245 ] fig .
[ fig2 ] illustrates , for @xmath9 and a typical choice of @xmath2 and @xmath7 , the distribution @xmath78 normalized to its maximum @xmath86 .
our numerical results strongly suggest that @xmath78 can be fitted by a @xmath8-gaussian distribution with @xmath87 .
other values of @xmath2 and @xmath7 have been studied as well and , in all cases , numerical results strongly suggest @xmath8-gaussian distributions ( with @xmath88 ) , i.e. , @xmath89_+^{\frac{1}{1-q^{disc } } } \label{qgaussian}\ ] ] where @xmath90_+=x$ ] if @xmath91 , and @xmath90_+=0 $ ] otherwise .
table [ table1 ] shows the values of @xmath8 for typical values of @xmath2 and @xmath7 ( @xmath9 in all cases ) . for a fixed value of @xmath7
, @xmath8 is a monotonous function of @xmath2 .
similarly , for a fixed value of @xmath2 , @xmath8 is a monotonous function of @xmath7 .
increasing the value of @xmath7 , a sequence of @xmath78 distributions is obtained ( see fig . [ fig3 ] ) .
their corresponding @xmath92-logarithmic representation show that @xmath8-gaussian distributions fit very well the data .
( @xmath93 ) for typical values of @xmath2 and @xmath7 . _ right panels _ :
the corresponding @xmath92-logarithmic representations show that the discrete distributions are very close to @xmath3-gaussians , except for the last point of the tail .
similar results are obtained for other parameter values .
we notice that for @xmath94 ( @xmath95 ) , which corresponds to @xmath96 ( @xmath97 ) , the terminal derivative in the linear - linear representation vanishes ( diverges ) .
this derivative is finite for @xmath98 , which corresponds to @xmath99 .
[ fig3],title="fig:",width=264 ] ( @xmath93 ) for typical values of @xmath2 and @xmath7 .
_ right panels _
: the corresponding @xmath92-logarithmic representations show that the discrete distributions are very close to @xmath3-gaussians , except for the last point of the tail .
similar results are obtained for other parameter values .
we notice that for @xmath94 ( @xmath95 ) , which corresponds to @xmath96 ( @xmath97 ) , the terminal derivative in the linear - linear representation vanishes ( diverges ) .
this derivative is finite for @xmath98 , which corresponds to @xmath99 .
[ fig3],title="fig:",width=264 ] + ( @xmath93 ) for typical values of @xmath2 and @xmath7 .
_ right panels _ : the corresponding @xmath92-logarithmic representations show that the discrete distributions are very close to @xmath3-gaussians , except for the last point of the tail .
similar results are obtained for other parameter values .
we notice that for @xmath94 ( @xmath95 ) , which corresponds to @xmath96 ( @xmath97 ) , the terminal derivative in the linear - linear representation vanishes ( diverges ) .
this derivative is finite for @xmath98 , which corresponds to @xmath99 .
[ fig3],title="fig:",width=264 ] ( @xmath93 ) for typical values of @xmath2 and @xmath7 .
_ right panels _ : the corresponding @xmath92-logarithmic representations show that the discrete distributions are very close to @xmath3-gaussians , except for the last point of the tail .
similar results are obtained for other parameter values .
we notice that for @xmath94 ( @xmath95 ) , which corresponds to @xmath96 ( @xmath97 ) , the terminal derivative in the linear - linear representation vanishes ( diverges ) .
this derivative is finite for @xmath98 , which corresponds to @xmath99 .
[ fig3],title="fig:",width=264 ] in the limit @xmath100 , we define @xmath101 and the corresponding @xmath11-gaussian ( @xmath10 ) limit distributions are characterized by @xmath102 .
this result that can in fact be rigorously proved @xcite .
[ deltaqeff ] shows the @xmath10 convergence of the index @xmath8 to the index @xmath11 .
index to the @xmath11 index .
[ deltaqeff],width=264 ] the value of @xmath11 depends only on the parameter @xmath2 , and some values of the ( @xmath103 ) pair are shown in table [ table2 ] .
we heuristically conclude ( see fig .
[ qalphaatt ] ) that @xmath104 -dependence of @xmath11.[qalphaatt],width=264 ] c|c|c|c|c|c|c|c|c|c|c|c|c @xmath2 & @xmath105 & @xmath106 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 + @xmath11&@xmath117 & @xmath118&@xmath119&@xmath120 & @xmath121&@xmath122&@xmath123&@xmath124&@xmath125 & @xmath126 & @xmath127 & @xmath128 + fig .
[ betaatt ] shows the generalized temperature @xmath129 of non - normalized @xmath11-gaussian distribution , as they are numerically obtained from @xmath2 .
the @xmath130 values correspond to positive values of @xmath38 , and the @xmath131 values correspond to negative values of @xmath38 . in both cases , the following equation is satisfied : @xmath132 this corresponds to the cut - offs of the distributions .
we can infer , from ( [ qalfa ] ) and ( [ bq ] ) , the following simple relation : @xmath133 that can also be checked in fig .
[ betaatt ] ( right panel ) . and @xmath134 dependence for the @xmath11-gaussian limit distribution .
left panel : @xmath130 provides positive generalized temperature of the @xmath11-gaussian .
central panel : @xmath131 provides negative generalized temperature of the @xmath11-gaussian .
right panel : @xmath2-dependence of the inverse temperature @xmath38 of @xmath11-gaussian .
[ betaatt],title="fig:",width=226 ] and @xmath134 dependence for the @xmath11-gaussian limit distribution .
left panel : @xmath130 provides positive generalized temperature of the @xmath11-gaussian .
central panel : @xmath131 provides negative generalized temperature of the @xmath11-gaussian .
right panel : @xmath2-dependence of the inverse temperature @xmath38 of @xmath11-gaussian .
[ betaatt],title="fig:",width=226 ] and @xmath134 dependence for the @xmath11-gaussian limit distribution . left panel : @xmath130 provides positive generalized temperature of the @xmath11-gaussian .
central panel : @xmath131 provides negative generalized temperature of the @xmath11-gaussian .
right panel : @xmath2-dependence of the inverse temperature @xmath38 of @xmath11-gaussian .
[ betaatt],title="fig:",width=226 ] summarizing , fig .
[ htempnegtva1 ] shows that @xmath135 provides bell - shaped and compact support @xmath11-gaussian distributions ( @xmath136 , @xmath137 ) , and @xmath138 provides convex and bounded but non compact support @xmath11-gaussian distributions ( @xmath139 , @xmath140 ) . in the @xmath141 limit , an uniform distribution is obtained .
this distribution is the limit of a @xmath11-gaussian distribution whose @xmath142 with @xmath137 , as well as @xmath143 with @xmath140 . in the @xmath144
limit a double peaked delta distribution emerges , i.e. , the distribution is the limit of a @xmath145-gaussian distribution whose @xmath146 with @xmath140 .
this description illustrates the diagram presented in @xcite , where negative generalized temperatures @xmath129 of @xmath3-gaussian distributions are also considered .
( @xmath10 ) limit distributions , are concave with bounded and compact support . _
right panel _ : the @xmath131 ( @xmath10 ) limit distributions , are convex with bounded but non compact support .
the limiting case @xmath147 corresponds to the uniform distribution .
[ htempnegtva1],title="fig:",width=268 ] ( @xmath10 ) limit distributions , are concave with bounded and compact support .
_ right panel _ : the @xmath131 ( @xmath10 ) limit distributions , are convex with bounded but non compact support .
the limiting case @xmath147 corresponds to the uniform distribution .
[ htempnegtva1],title="fig:",width=264 ] the indexes @xmath11 and @xmath27 are related by @xmath148 equation that reminds the equations of the algebra indicated in @xcite . making use of the expression of a normalized @xmath3-gaussian @xcite , it can be consequently stated that the generalized distribution @xmath78 defined in eq . corresponds , for win - loss equiprobability ( i.e. , @xmath9 ) , to the following @xmath11-gaussian limit distribution : @xmath149_+^{\frac{1}{1-q^{att } } } \label{qgausianeq}\ ] ] where @xmath29 , @xmath13 , and @xmath150 .
let us now consider that each of the @xmath7 single variables takes values @xmath151 or @xmath121 ( i.e.,``win '' or `` loose '' ) .
consequently , the value of @xmath6 corresponds to the sum of all @xmath7 binary random variables , and , after centering and re - scaling @xmath6 , the attractors that emerge correspond to the abscissa currently associated with central limit theorems . in the case of the unbiased symmetric generalized probability distribution ( i.e. ,
@xmath9 ) defined in eq . , the abscissa measured from its central value scales as @xmath152 with @xmath153 , and emerging attractors of @xmath78 are the @xmath11-gaussians defined in eq . .
this fact precludes the vanishing limit of the probability @xmath154 of a deviation of @xmath34 from its central value @xmath155 , i.e. , @xmath156 where @xmath157 is the minimum deviation of @xmath34 with respect to central value @xmath34 , and @xmath154 can be evaluated adding up the weight of the possible values of @xmath6 which do not fall inside @xmath158 .
taking into account the symmetry of the distribution , and denoting @xmath159 , we can write @xmath160 and conclude that , @xmath161 , @xmath162 where @xmath163 is the largest integer number that @xmath164 , and @xmath165 is defined in eq . .
states that these correlated models do _ not _ satisfy the classical version of the weak law of large numbers . due to this fact ,
let us consider instead the scaled quantity @xmath15 , @xmath166 , in analogy with the anomalous diffusion coefficient introduced in nonlinear fokker - plank equations @xcite .
more precisely , the case @xmath166 is similar to anomalous diffusion , where the square space is nonlinear with time . in the case
@xmath16 , we have numerically found a @xmath17-exponential ( where @xmath167 stands for _ large - deviation - like _ ) decaying behavior of the probability @xmath168 , for typical values of @xmath31 and @xmath2 , when @xmath7 increases .
in fact ( see fig .
[ histgama ] ) it can be straightforwardly verified that , @xmath169 , for @xmath9 , @xmath170 , and typical sequences of @xmath7 .
_ left panel : _ scaling factor @xmath16 makes @xmath171 , @xmath172 .
_ right panel : _ scaling factor @xmath173 makes @xmath174 .
[ histgama],width=566 ] @xmath175 and the zero limit is attained as @xmath176^{1/(1-q^{ldl})},\ ] ] with @xmath177 .
this result can be checked in fig .
[ qlog ] , where @xmath92-logarithmic representation of @xmath178 as a function of @xmath7 is shown , for typical values of @xmath2 and @xmath179 .
straight lines are obtained for all values of @xmath180 ( @xmath181 ) , @xmath2 ( @xmath182 ) and @xmath179 ( @xmath183 ) that have been considered .
distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] + distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] + distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] distribution , for typical values of @xmath2 ( @xmath9 ) , in semi-@xmath17-logarithmic representation .
the values of @xmath184 , related to the @xmath17-exponential decay , are @xmath185 , @xmath186 and @xmath187 , no matter the value of @xmath179 ( @xmath188 , @xmath189 and @xmath190 , in figure ) .
the central panel shows the significative deviations from a straight line , for @xmath191 deviations of @xmath17 .
[ qlog],title="fig:",width=215 ] + -dependence of the @xmath17-exponential index of the decaying probability @xmath192 , for @xmath9 and @xmath193 models .
[ qalpha],width=264 ] -exponential indices appear to be independent of @xmath179 .
[ qgamma],width=283 ] we have heuristically obtained ( see fig . [ qalpha ] ) that , for all the values of @xmath194 for which we have checked , the @xmath17 index satisfies @xmath195 notice that it does not depend on @xmath179 or @xmath194 ( see fig .
[ qgamma ] ) .
consequently , from ( [ qalfa ] ) and ( [ qldtetaalfa ] ) , we infer that , for @xmath9 and for all values of @xmath29 , the @xmath184-exponential decay index and the @xmath145-gaussian attractor index are related as follows : @xmath196 let us now focus on the values of the slopes of the semi-@xmath17-logarithmic representation of @xmath197 .
[ qlog ] shows that the @xmath17-exponential decaying rate does not only depend on the value of @xmath194 , i.e. , @xmath198 .
the mechanism that precludes a simple dependence on @xmath194 can be understood by fixing a particular value of @xmath194 and @xmath179 , as shown in fig .
[ qexpx ] .
-exponential decaying rate of @xmath199 for @xmath9 and @xmath200 shows that the sequence of slopes that correspond to a value of @xmath194 , are associated to @xmath201 } involved in @xmath202.[qexpx],width=283 ] for a fixed value of @xmath194 , the sequence of slope values are associated to the maximum value of @xmath6 involved in @xmath203 , i.e. , @xmath204 .
summarizing , the deviation re - scaled probability behavior presents a @xmath17-exponential decay when @xmath205 , that can be written as @xmath206 thus exhibiting a non - trivial dependence of the rate function @xmath207 .
let us now analyze the @xmath10 evolution behavior of the probability of @xmath15 for @xmath21 , i.e. , how @xmath78 approaches its attractor @xmath208 .
the probability left deviations of @xmath34 from @xmath194 ( @xmath209 ) , @xmath210 , with respect to the attractor , can be written as @xmath211 where @xmath212 . from eq .
, we obtain that the probability left deviations of @xmath34 from @xmath194 , for @xmath9 and even values of @xmath213 ( i.e. @xmath73 ) , can be writen as @xmath214 let as analytically study the simplest model , i.e. , @xmath9 and @xmath98 .
we verify @xmath215 the corresponding @xmath10 limit distribution is a @xmath11-gaussian , with @xmath99 .
consequently , the corresponding asymptotic left deviation of @xmath34 from a fixed value @xmath194 is given by @xmath216dx= 3x^2 - 2x^3.\label{attdevn}\ ] ] from and we have that , for a certain value of @xmath7 , the probability left deviations of @xmath34 from @xmath194 , with respect to the probability left deviation of the @xmath10 limit distribution , is analytically obtained as @xmath217 the upper bound @xmath218 of can be obtained in the case that @xmath219 , as shown in fig . [ ldpatt ]
. a lower bound @xmath220 can be also considered as @xmath221 , but it is never attained and @xmath222 .
we can consider , instead of @xmath220 , the maximum of all lower bounds @xmath223 , for each value of @xmath194 .
all these bounds verify the relation @xmath224 -logarithmic representation of @xmath223 .
observe that , for @xmath225 , a bias from the linear behavior exists .
_ right panel : _ the linear behavior of the @xmath226-logarithmic representation of the upper bound @xmath218 could reflect its @xmath23-exponential decay with @xmath7 .
[ qexpupp],title="fig:",width=268]-logarithmic representation of @xmath223 .
observe that , for @xmath225 , a bias from the linear behavior exists .
_ right panel : _ the linear behavior of the @xmath226-logarithmic representation of the upper bound @xmath218 could reflect its @xmath23-exponential decay with @xmath7 .
[ qexpupp],title="fig:",width=260 ] fig .
[ qexpupp ] exhibits the @xmath226-logarithmic representation of the upper and the minimum bound deviations , @xmath218 and @xmath223 .
@xmath218 appears to @xmath23-exponentially decay with @xmath7 ( were @xmath227 stands for _ large deviation _ ) , as conjectured in @xcite .
it is not the case of @xmath223 for @xmath228 , as it is shown in fig .
[ ldpatt ] . from a fixed value of @xmath194 , with respect to the deviation in @xmath10 limit distribution , i.e.
, @xmath229 for @xmath9 and @xmath98 . from left to right , @xmath230 , @xmath228 and @xmath231 .
the upper and minimum bounding sets , i.e. , @xmath218 and @xmath223 , are dashed in red and blue , respectively .
bottom figures show the same results in semi-@xmath23-logarithmic representation , where @xmath232 in all cases.[ldpatt],title="fig:",width=222 ] from a fixed value of @xmath194 , with respect to the deviation in @xmath10 limit distribution , i.e. , @xmath229 for @xmath9 and @xmath98 . from left to right , @xmath230 , @xmath228 and @xmath231 .
the upper and minimum bounding sets , i.e. , @xmath218 and @xmath223 , are dashed in red and blue , respectively .
bottom figures show the same results in semi-@xmath23-logarithmic representation , where @xmath232 in all cases.[ldpatt],title="fig:",width=222 ] from a fixed value of @xmath194 , with respect to the deviation in @xmath10 limit distribution , i.e. , @xmath229 for @xmath9 and @xmath98 . from left to right , @xmath230 , @xmath228 and @xmath231 .
the upper and minimum bounding sets , i.e. , @xmath218 and @xmath223 , are dashed in red and blue , respectively .
bottom figures show the same results in semi-@xmath23-logarithmic representation , where @xmath232 in all cases.[ldpatt],title="fig:",width=222 ] + from a fixed value of @xmath194 , with respect to the deviation in @xmath10 limit distribution , i.e. , @xmath229 for @xmath9 and @xmath98 . from left to right , @xmath230 , @xmath228 and @xmath231 .
the upper and minimum bounding sets , i.e. , @xmath218 and @xmath223 , are dashed in red and blue , respectively .
bottom figures show the same results in semi-@xmath23-logarithmic representation , where @xmath232 in all cases.[ldpatt],title="fig:",width=222 ] from a fixed value of @xmath194 , with respect to the deviation in @xmath10 limit distribution , i.e. , @xmath229 for @xmath9 and @xmath98 . from left to right , @xmath230 , @xmath228 and @xmath231 .
the upper and minimum bounding sets , i.e. , @xmath218 and @xmath223 , are dashed in red and blue , respectively .
bottom figures show the same results in semi-@xmath23-logarithmic representation , where @xmath232 in all cases.[ldpatt],title="fig:",width=222 ] from a fixed value of @xmath194 , with respect to the deviation in @xmath10 limit distribution , i.e. , @xmath229 for @xmath9 and @xmath98 . from left to right , @xmath230 , @xmath228 and @xmath231 .
the upper and minimum bounding sets , i.e. , @xmath218 and @xmath223 , are dashed in red and blue , respectively .
bottom figures show the same results in semi-@xmath23-logarithmic representation , where @xmath232 in all cases.[ldpatt],title="fig:",width=222 ] the hypothesis of @xmath218 @xmath23-exponentially decaying behavior can be verified by using the asymptotic expansion of the analytical expressions of the bounding values @xmath218 , @xmath233 : @xmath234 \ , .
\label{aimpt}\ ] ] we can compare the respective terms with the asymptotic expansion of a @xmath3-exponential function @xcite , namely @xmath235^{\frac{1}{q-1}}}\nonumber\\ & \times & \left [ 1-\frac{1}{(q-1)^2r_q(x)n}+\sum_{m=2}^{\infty}(-1)^m\frac{q(2q-1)\dots[(m-1)q-(m-2)]}{m!(q-1)^{2m}(r_q(x)n)^m } \right]\ , .
\label{qexpesp}\end{aligned}\ ] ] equations ( [ aimpt ] ) and ( [ qexpesp ] ) would provide , by neglecting higher - order terms , an index @xmath232 .
in such a situation , and identifying the two first terms of expansions , the best @xmath23-generalized rate function and the corresponding @xmath23-exponential factor @xmath236 would be @xmath237 but , in fact , the third term of eq .
is not negligible for some values of @xmath194 and , in such cases , eq . and
eq . are not compatible .
the @xmath23-exponential decay of the upper bound of the large deviation to the attractor is precluded , as obtained within a different context in @xcite .
other values of @xmath2 have been tested and , in all cases , the large deviation probability with respect the attractor presents a power - law decay .
the large - deviation probability does not in fact @xmath3-exponentially decay to the attractor , even though , in some cases , eq . roughly describes the large - deviation behavior
a generalized binomial distribution based on @xmath3-exponential generating functions is characterized by eqs .
( [ distribevaldo ] ) and ( [ improvenum ] ) .
its probability function @xmath0 depends on two parameters @xmath24 , and can be considered as the following urn scheme : from a set of @xmath238 black balls and @xmath239 red balls contained in an urn , one extracts one ball and returns it to the urn , together with @xmath240 balls of the same color . in that case
, @xmath0 represents the probability to have @xmath6 black balls in the urn after the _ n_-th trial , and it can can be written as a function of @xmath241 , as @xmath242 and @xmath243 @xcite .
if no bias exists , i.e. for @xmath9 , the probability to find a relative number of black balls @xmath34 after the _ n_-th trial , closely approaches a @xmath92-gaussian distribution .
the @xmath10 limit probability distribution is in fact a @xmath11-gaussian @xcite whose @xmath11 index and @xmath129 generalized temperature , can be obtained from @xmath241 as @xmath244 and @xmath245 . in other words ,
the numerical discrete distributions appear to be very close to a set of @xmath92-gaussian distributions that , increasing @xmath7 , evolve towards to a @xmath11-gaussian attractor .
this urn scheme provides a procedure to attain a @xmath11-gaussian @xmath10 limit distribution that verifies , in all cases ( @xmath246 ) , the relation @xmath247 .
when the number of reposition balls verifies @xmath248 , such attractors are concave @xmath11-gaussian distributions with a bounded and compact support ( where index @xmath249 and generalized temperature @xmath250 ) ; on the contrary , when @xmath251 , such attractors are convex @xmath11-gaussian distributions with bounded but non compact support ( where index @xmath252 and generalized temperature @xmath253 ) .
these generalized binomial distributions violate the law of large numbers , but nevertheless present a large - deviation - like property . indeed , by using , instead of the variable @xmath34 , the rescaled variable @xmath47 ( @xmath16 )
, the left deviation probability behaves as @xmath254 .
moreover , an interesting result is that , when no bias exists ( i.e. , @xmath255 ) , the @xmath11-gaussian index , the @xmath17 index , and the @xmath27 index ( characterizing the generating function ) , are univocally defined by the @xmath256 ratio , and simple mathematical relations exist that remind the algebra indicated in @xcite
we acknowledge useful conversations with j. p. gazeau and e. m. f. curado , and partial financial support from universidad politcnica de madrid , centro brasileiro de pesquisas fsicas , cnpq , faperj , and capes ( brazilian agencies ) .
one of us ( gr ) acknowledges warm hospitality at cbpf .
the other one ( ct ) acknowledges partial financial support from the john templeton foundation .
99 v. v. dodonov , j. opt .
b : quantum semiclass . opt . * 4 * , 1 ( 2002 ) , and references therein . j .-
gazeau , _ coherent states in quantum physics _ , wiley - vch , 2009 .
r. l. de matos filho and w. vogel phys rev .
a * 54 * , 4560 ( 1996 ) .
z. kis , w. vogel , and l. davidovich , phys rev . a * 64 * , 0033401 ( 2001 ) .
e. m. f. curado , j. p. gazeau , and l. m. c. s. rodrigues , phys .
scr . * 82 * , 038108 ( 2010 ) .
h. bergeron , e.m.f .
curado , j.p .
gazeau , l. m. c. s. rodrigues , j. math . phys .
* 146 * , 103304 - 1 - 22 ( 2012 ) .
h. bergeron , e. m. f. curado , j. p. gazeau and ligia m.c.s .
rodrigues , j. math . phys . * 54 * , 123301 ( 2013 ) . c. tsallis , m. gell - mann and y. sato , proc .
acad . sc .
usa * 102 * , 15377 ( 2005 ) . c. tsallis , j. stat .
phys . * 52 * , 479 ( 1988 ) .
c. tsallis , _ introduction to nonextensive statistical mechanics - approaching a complex world _ , springer , new york , 2009 .
s. umarov , c. tsallis , m. gell - mann and s. steinberg , j. math . phys . * 51 * , 033502 ( 2010 ) .
a. rodriguez and c. tsallis , private communication .
s. umarov and c. tsallis , american institute of physics conference proceedings * 965 * , 34 - 42 ( 2007 ) .
d. prato and c. tsallis , phys .
e * 60 * , 2398 ( 1999 ) .
r. hanel , s. thurner and c. tsallis , eur .
j. b * 72 * , 263 ( 2009 ) .
a. rodrigues and c. tsallis , j. math .
phys . * 53 * , 023302 ( 2012 ) .
j. ruseckas , [ cond-mat.stat-mech 1408.0088v1 ] .
h. bergeron , e. m. f. curado , j. p. gazeau , and l. m. c. s. rodrigues , private communication .
a. carati , physica a * 348 * , 110 ( 2005 ) .
a. carati , physica a * 387 * , 1491 ( 2008 ) .
plya distribution , encyclopedia of mathematics : http://www.encyclopediaofmath.org/ t.d .
frank , _ nonlinear fokker - planck equations : fundamentals and applications _ , springer , berlin - new york , 2005 . g. ruiz and c. tsallis , phys .
a * 377 * , 491 ( 2013 ) .
j. max and c. tsallis , private communication . | we study a symmetric generalization @xmath0 of the binomial distribution recently introduced by bergeron et al , where @xmath1 $ ] denotes the win probability , and @xmath2 is a positive parameter .
this generalization is based on @xmath3-exponential generating functions ( @xmath4^{1/(1-q^{gen})};\,e_{1}^z = e^z)$ ] where @xmath5 .
the numerical calculation of the probability distribution function of the number of wins @xmath6 , related to the number of realizations @xmath7 , strongly approaches a discrete @xmath8-gaussian distribution , for win - loss equiprobability ( i.e. , @xmath9 ) and all values of @xmath2 .
asymptotic @xmath10 distribution is in fact a @xmath11-gaussian @xmath12 , where @xmath13 and @xmath14 .
the behavior of the scaled quantity @xmath15 is discussed as well . for @xmath16 , a large - deviation - like property showing a @xmath17-exponential decay
is found , where @xmath18 . for @xmath9 , @xmath17 and @xmath11
are related through @xmath19 , @xmath20 . for @xmath21 ,
the law of large numbers is violated , and we consistently study the large - deviations with respect to the probability of the @xmath22 limit distribution , yielding a power law , although not exactly a @xmath23-exponential decay .
all @xmath3-statistical parameters which emerge are univocally defined by @xmath24 .
finally we discuss the analytical connection with the plya urn problem . |
heat conduction in low dimensional anharmonic systems has attracted considerable interest in recent years @xcite .
phonon , as the predominant heat carrier in insulating materials , undoubtedly lies in the heart of heat conduction .
however , phonon bears a solid basis only in harmonic systems .
the intrinsic nonlinearity in anharmonic systems will inevitably affect the behavior of phonon .
therefore , understanding the phonon properties in anharmonic systems represents an important question in the heat conduction field .
a renormalized phonon ( r - ph ) picture was then put forward independently by several groups using varying techniques @xcite . within the scope of this picture
, one can successfully interpret and understand a wide range of physical phenomena , including a theoretical description of the sound velocity @xcite as well as scaling laws of thermal conductivity @xmath0 with temperature @xmath1 @xcite .
however , the generality of existing , state - of - the - art quasi - harmonic theories is limited . they were found to provide inaccurate predictions on the sound velocity of a nonlinear lattice with an asymmetric inter - particle potential @xcite .
recently , a numerical method which aims to justify the validity of phonon concept in nonlinear lattices was proposed @xcite .
the existence of phonon modes in nonlinear lattices with asymmetric potentials is confirmed for a wide range of parameters .
hence , a unified quasi - harmonic theory which extend to cover the general cases beyond harmonic is desirable . in statistical mechanics ,
variational approaches are often used to obtain approximate information of a nonlinear system via a reference system and an associated variational principle @xcite .
the reference system whose properties can be easily obtained contains several variational parameters . in a variational scheme
, an optimal reference system can be obtained by varying those parameters such that bounds of the variational principle go to a relative minimum or maximum . in this paper
, we develop a variational approach to study phonons in nonlinear lattices .
we choose a general harmonic system as the reference system for this purpose and regard the so - obtained optimal harmonic system as an approximation to the original system from which we identify the properties of r - phs .
we consider applications to one dimensional ( 1d ) nonlinear lattices with either symmetric or asymmetric inter - particle fermi - pasta - ulam ( fpu ) potentials @xcite and demonstrate the power of our approach by comparing with molecule dynamics ( md ) results .
this paper is organized as follows . in sec.ii , we introduce our variational approach . in sec.iii , we present numerical details . in sec.iv and sec.v
, we study nonlinear lattices with symmetric and asymmetric fpu potentials , respectively .
finally , we briefly summarize the work in sec.vi .
our current investigation aims to develop a quasi - harmonic theory applicable for nonlinear lattices with asymmetric potentials .
one of the attractive features of such systems is the existence of nonzero internal pressure due to the asymmetry of the potential .
then from a statistical mechanics point of view , it s convenient to use the language of the isothermal - isobaric ensemble which maintains constant temperature @xmath1 , constant particle number @xmath2 and constant pressure @xmath3 ( so - called @xmath4 ensemble ) to describe them @xcite .
their equilibrium thermodynamic properties can thus be well determined by the gibbs free energy @xmath5,\ ] ] where @xmath6 , @xmath3 and @xmath7 denote the hamiltonian , pressure and volume ( or length in 1d cases ) of the system , respectively , @xmath8 is the inverse temperature ( we set @xmath9 ) , @xmath10 and @xmath11 are short for the products of all the coordinates and momenta of the system , respectively .
the volume @xmath7 is a function of @xmath10 only .
introducing @xmath12 whose ensemble average is just the enthalpy of the system @xcite .
the corresponding probability measure in phase space reads @xmath13 a variational principle is an inequality satisfied by the physical quantity in which we may be interested @xcite .
hence we should look for inequalities for the gibbs free energy @xmath14 . to do this ,
we introduce a reference system with a hamiltonian @xmath15 and prepare the original system and the reference system at same pressure . then following this spirit ,
we integrate both sides of the following equality @xmath16 over the whole phase space to yield @xmath17 where @xmath18 is the gibbs free energy of the reference system and @xmath19 , @xmath20 denotes the ensemble average under the probability measure @xmath21 . taking logarithm over both sides of the above equation and using the jensen s inequality for exponential functions , i.e. , @xmath22(following @xcite )
, we have @xmath23 by switching the role of the nonlinear system and the reference system in eq .
( [ eq : rho_e ] ) , we obtain @xmath24 with @xmath25 the ensemble average with respect to the probability measure @xmath26 [ c.f .
( [ eq : npt - measure ] ) ] .
these two inequalities [ eqs .
( [ eq : up ] ) and ( [ eq : low ] ) ] give the upper and lower bound of @xmath14 , respectively . if the pressure vanishes , i.e. , @xmath27 , the gibbs free energies @xmath14 and @xmath18 reduce to the helmholtz free energies @xmath28 and @xmath29 , respectively , the enthalpy goes to the energy , the probability measure eq .
( [ eq : npt - measure ] ) should also be replaced by the canonical measure @xmath30 then we found that the inequality eq .
( [ eq : up ] ) recovers the well known gibbs - bogoliubov ( gb ) inequality @xcite @xmath31 where @xmath29 is the helmholtz free energy of the reference system and @xmath32 stands for the ensemble average with respect to the canonical measure @xmath33 $ ] .
and the inequality eq .
( [ eq : low ] ) goes to @xmath34 with @xmath35 the ensemble average with respect to the probability measure @xmath36 [ c.f .
( [ eq : ca_dis ] ) ] . as a lower bound of the helmholtz free energy @xmath28 @xcite , it s little applied since the ensemble average in it can not be evaluated analytically .
however , it is more accurate than the upper bound in determining the free energy of solids @xcite .
as can be seen later , in our case for determining the sound velocity for anharmonic lattices , it is still the lower bound that gives better prediction comparing to the upper bound .
we consider 1d momentum conserving nonlinear lattices described by the general hamiltonian @xcite @xmath37,\ ] ] where @xmath2 is the particle number , @xmath38 denotes the momentum of @xmath39-th particle , @xmath40 denotes the displacement of @xmath39-th particle from its equilibrium position @xmath41 with @xmath42 the absolute position and @xmath43 the equilibrium distance for the interaction bond , and @xmath44 represents the inter - particle potential . for brevity and without loss of generality , we take @xmath45 and @xmath46 as the unit of mass and length , respectively . the average lattice spacing @xmath47
is then given by @xmath48 , and the average lattice length @xmath49 equals @xmath50 for an @xmath2-particle lattice .
we say that the lattice is at its natural length if @xmath47 equals @xmath43 [ = 1 ] , namely , @xmath51 for an @xmath2-particle lattice . the average length can be changed to other values by applying pressure .
furthermore , we introduce @xmath52 .
if the potential satisfies @xmath53 , we refer it to a symmetric potential , otherwise we call it an asymmetric one . in terms of @xmath54 and @xmath38 , the equations of motion ( eoms ) read @xmath55 where the dot and the prime denote the time and space derivative , respectively . specifically , in this work , we focus on the fpu lattices with the inter - particle potential @xcite @xmath56 which has become an archetype 1d nonlinear system in statistical mechanics @xcite .
we call the lattice with @xmath57 the fpu-@xmath58 lattice .
otherwise , we refer it to the fpu-@xmath59 lattice . for simplicity but without loss of generality , we only study the case of @xmath60 in this paper .
it is evident that the fpu-@xmath58 lattice has a symmetric potential while the fpu-@xmath59 lattice has an asymmetric one .
therefore , an fpu-@xmath58 lattice with natural length has zero pressure , while an fpu-@xmath59 lattice with natural length has a nonvanishing pressure due to the asymmetry of the potential .
nevertheless , our variational principles [ eqs . ( [ eq : up ] ) and ( [ eq : low ] ) ] enable us to study them within the same theoretical scheme .
phonons bear a solid basis only in harmonic systems , in order to develop an effective phonon theory for nonlinear lattices , we should choose a quadratic @xmath15 as a reference system .
moreover , the present work investigates nonlinear lattices with asymmetric inter - particle potentials . taking those into consideration
, we consider a harmonic reference system in the following form @xmath61;\quad v_0(\delta_n ) ~=~\frac{k}{2}(\delta_n - d)^2,\ ] ] in which @xmath62 and @xmath63 are variational parameters with @xmath62 being the effective elastic constant and @xmath63 quantifying the degree of asymmetry of the potentials .
note that by setting @xmath64 , we can use @xmath15 to deal with nonlinear systems with symmetric potentials as well .
the dispersion relation of this reference system is given by @xmath65 where @xmath66 is the wavenumber .
it can be regarded as an approximate dispersion relation for the r - phs in nonlinear lattices @xcite .
the corresponding sound velocity reads @xmath67 the variational method is then to select the optimal reference systems with parameterized hamiltonian @xmath15 that either minimize the right hand side ( r.h.s . ) of eq .
( [ eq : up ] ) or maximize the r.h.s . of eq .
( [ eq : low ] ) .
this strategy then provides approximations for the gibbs free energy and allows us to find optimal @xmath15 for the system @xmath6 .
the process is going to be detailed in the next subsection .
note that @xmath68 in the large @xmath2 limit ( the jacobian is unity ) , then for a system with the probability measure eq .
( [ eq : npt - measure ] ) , we have @xmath69 where @xmath70 denotes the marginal phase space distribution for the single site variables @xmath54 and @xmath38 of @xmath39-th particle @xcite @xmath71\right\}\ ] ] with @xmath72 the corresponding partition function @xmath73\right\ } d\delta d p.\ ] ] such a result can be readily tested using md simulations , see sec .
[ sec : md ] . for a system with sufficiently large particle number @xmath2
, the gibbs free energy can be expressed as @xmath74 where @xmath75 is the gibbs free energy per particle .
so the gibbs free energy of the reference system @xmath15 reads @xmath76 with @xmath77 the partition function determined by eqs .
( [ eq : h0 ] ) and ( [ eq : z ] ) .
it can be evaluated analytically , which gives @xmath78.\ ] ] using the single - site distribution eq .
( [ eq : single_rho ] ) , and noticing that the length @xmath7 and @xmath79 have a same expression , namely , @xmath80 , the averages in eqs .
( [ eq : up ] ) and ( [ eq : low ] ) read @xmath81 where @xmath82 and @xmath83 denote single - site averages with respect to the probability measure eq .
( [ eq : single_rho ] ) with the potential being @xmath84 and @xmath44 , respectively .
we will use these symbols in the rest of the paper .
[ [ section ] ] firstly , we focus on the upper bound of @xmath14 [ eq . ( [ eq : up ] ) ] .
minimizing it with respect to the variational parameters @xmath62 and @xmath63 , i.e. , @xmath85 & = & 0,\\ \frac{\partial}{\partial d}[g_0+\langle \mathcal{h}-\mathcal{h}_0\rangle_{\rho_0 } ] & = & 0,\end{aligned}\ ] ] we obtain the following coupled equations : @xmath86 it can be readily tested that the second order derivatives are positive at @xmath87 so that this set of solution indeed gives the minimum of the upper bound .
thus @xmath88 and @xmath89 determine the optimal upper bound of the gibbs free energy @xmath14 and the corresponding optimal harmonic reference system ( denoted as the uh system ) .
[ [ section-1 ] ] now we turn to the lower bound of @xmath14 [ eq . ( [ eq : low ] ) ] . similarly , we maximize it with respect to the variational parameters @xmath62 and @xmath63 , the optimal solution reads @xmath90}.\label{eq : k1}\end{aligned}\ ] ] we have also checked that @xmath91 and @xmath92 indeed give the maximum of the lower bound , thus them uniquely determine the optimal lower bound of the gibbs free energy @xmath14 and also the corresponding optimal harmonic reference system ( denoted as the lh system ) .
note that @xmath91 does not rely on @xmath92 and is completely determined by eq .
( [ eq : k1 ] ) . the lh system and
the uh system can be regarded as effective descriptions of the original nonlinear system .
if we use these two optimal systems to construct the effective theory of r - phs in the original nonlinear system , then the dispersion relation should follow eq .
( [ ome ] ) with @xmath62 given by @xmath91 or @xmath88 .
the accuracy of the results can be evaluated by comparing the calculated sound velocity with the predictions using eq .
( [ eq : vs ] ) . in the following ,
we present numerical details and then apply this strategy to two 1d nonlinear lattices , namely , the fpu-@xmath58 lattice and the fpu-@xmath59 lattice .
in this work , we consider systems either with or without pressure @xmath3 .
the average lattice length @xmath93 can be controlled by adjusting the pressure , as demonstrated in fig .
[ fig : pl ] . as a function of pressure @xmath3 .
the dashed - dotted line indicates that for the 1d fpu-@xmath58 lattice with @xmath94 .
the solid line denotes that for the 1d fpu-@xmath59 lattice with @xmath94 and @xmath95 . ]
it can be clearly seen that for the fpu-@xmath58 lattice once the pressure is nonzero , the average length of the lattice is away from the natural length , i.e. , @xmath96 .
it further indicates that we can apply a positive or negative pressure to this lattice system in order to compress or elongate its total length .
but for systems with asymmetric potentials , i.e. , the fpu-@xmath59 lattice , the pressure would be nonzero at its natural length .
note that the average length @xmath93 equals @xmath97 , so a stressless system with an asymmetric potential corresponds to an average length @xmath98 , where @xmath99 means the canonical single - site distribution , i.e. , @xmath70 [ eq .
( [ eq : single_rho ] ) ] with @xmath27 .
particularly , for the fpu-@xmath59 lattice with zero pressure , the average length is smaller than @xmath2 as can be seen from the figure .
an @xmath4 system with pressure @xmath3 can be prepared either by applying an external pressure @xmath3 to the end particles with free boundary condition or by fixing the total length @xmath7 to @xmath100 , where @xmath100 as a function of @xmath3 is just depicted in fig .
[ fig : pl ] . in the present work
we adopt the second approach .
specifically , a modified periodic boundary condition is thus used to fix the total length at a certain value , namely , @xmath101 . for such systems ,
a symplectic integrator @xmath102 with a corrector @xmath103 @xcite is adopted to integrate the eoms with a time step @xmath104 to ensures the conservation of the total system energy and momentum up to high accuracy @xmath105 in the whole run time , a transient time of order @xmath106 is used to equilibrate the system .
we first check the single - site distributions in eq .
( [ eq : single_rho ] ) .
after the system reaches its thermal equilibrium state , we can obtain time series of @xmath107 and @xmath54 of the @xmath39-th particle ( @xmath39 can be an arbitrary integer from 1 to @xmath2 ) .
we then plot their histograms and compare the envelopes against @xmath108 with @xmath109,\label{eq : single_q}\end{aligned}\ ] ] where @xmath110 $ ] . to illustrate the comparison
, we simulate the fpu-@xmath58 lattice with @xmath111 , @xmath94 , and @xmath51 or @xmath112 , the corresponding pressure is 0 or -0.4309 according to fig . [
fig : pl ] , respectively .
a similar simulation is carried out on the fpu-@xmath59 lattice with @xmath111 , @xmath94 , @xmath95 , and @xmath113 or @xmath51 , with the pressure equals 0 or -0.3904 , respectively .
the results are depicted in figs .
[ fig : fpudis ] and [ fig : afpudis ] .
@xmath114 and @xmath115 are shown as green and red solid curves in the figures , respectively .
perfect agreements between theoretical curves and the md results are presented , which indicates the validity of the single - site distribution in eq .
( [ eq : single_rho ] ) . lattice with @xmath27 ( upper panel ) or @xmath116 ( lower panel ) .
the blue parts in the left and right column denote the histogram of @xmath117 and @xmath118 , respectively .
the green and red solid lines are @xmath119 and @xmath115 [ c.f .
( [ eq : single_v ] ) and ( [ eq : single_q ] ) ] , respectively . ] lattice with with @xmath27 ( upper panel ) or @xmath116 ( lower panel ) .
the blue parts in the left and right column denote the histogram of @xmath117 and @xmath118 , respectively .
the green and red solid lines are @xmath119 and @xmath115 [ c.f .
( [ eq : single_v ] ) and ( [ eq : single_q ] ) ] , respectively . ]
now we turn to the calculation of the sound velocity .
so far , there are mainly three numerical methods which can obtain the sound velocity of the nonlinear lattices .
the first method regards the moving velocity of front peaks of the equilibrium spatiotemporal correlation function as the sound velocity @xcite .
however , a broadening of front peaks at high temperature or strong anharmonicity will affect its accuracy .
the second one is to look at the frequency of the lowest phonon peak of the power spectrum of an @xmath2-particle lattice @xcite @xmath120 where @xmath121 is the sound velocity .
this lowest phonon peak can always be detected in the power spectrum with high resolution ( see fig . [
fig : ps ] ) , thus the sound velocity can be well determined by @xmath122 .
the third one obtains the sound velocity from the dispersion relation @xcite .
this method , although follows the definition of the sound velocity , is most computationally expensive comparing with the other two , because the calculation of the dispersion relation is time consuming . therefore ,
in this study , we choose the second method to obtain md results of the sound velocity of nonlinear lattices . according to the wiener - khinchin theorem @xcite
, the power spectrum @xmath123 can be obtained by doing fourier transform of the velocity autocorrelation of a single particle @xmath124 with @xmath125 . in order to get smooth phonon peaks in the power spectrum ,
the maximum correlation time @xmath126 whose inverse @xmath127 determines the frequency resolution is set to be @xmath128 under the condition that @xmath129 .
[ fig : ps ] presents the power spectrum of the fpu-@xmath58 lattice and the fpu-@xmath59 lattice with @xmath95 at their natural length .
@xmath94 is used for both models .
it is apparent that phonon peaks with the lowest frequencies are well distinguished from the power spectrum .
the nonlinearity renders its effect in the phonon peak broadening .
interestingly , we found that the phonon peak of the fpu-@xmath59 lattice is broader than the one of the fpu-@xmath58 lattice , which indicates that the asymmetric potentials provide stronger phonon - phonon interaction compared to the symmetric ones . in the low frequency regime .
the dashed - dotted line stands for the fpu-@xmath58 lattice at its natural length .
the solid line represents the fpu-@xmath59 lattice with @xmath95 and natural length . for comparison ,
the lowest harmonic frequency is depicted as black solid line . ]
in the following , we will give a detailed study of the fpu-@xmath58 lattice by using our variational approach .
being a special case of the fpu-@xmath59 lattice , i.e. , @xmath130 , it has a symmetric inter - particle potential .
the existing quasi - harmonic theories of this model all focus on this particular case @xcite . in this part , we not only present the comparison between our approach and md results , but also point out the connections between our approach and some of the existing quasi - harmonic theories . the first observation is @xmath131 [ c.f . , eqs .
( [ eq : d2 ] ) and ( [ eq : d1 ] ) ] by noting that the potential of the fpu-@xmath58 lattice is an even function of @xmath117 .
then in the following , we only need to concern about the parameter @xmath62 .
firstly , we focus on the uh system . the corresponding parameter @xmath88 [ eq .
( [ eq : k2 ] ) ] can be calculated to yield @xmath132 which has a positive solution @xmath133 this result coincides with the prediction of the so called self - consistent phonon theory ( scpt ) @xcite for the fpu-@xmath58 lattice , since the scpt based on the gb inequality eq .
( [ eq : gb ] ) as well .
then we turn to the lh system .
the parameter @xmath91 [ eq .
( [ eq : k1 ] ) ] reduces to the simple form @xmath134 the corresponding sound velocity @xmath135 is exactly the same as that predicted by the effective phonon theory ( ept ) @xcite based on the generalized equipartition theorem @xcite as well as the nonlinear fluctuating hydrodynamics ( nfh ) using the hydrodynamic approximation @xcite , although the nfh is not an effective theory for phonons .
results of the sound velocity are illustrated in fig .
[ fig : bfpuv ] .
the excellent agreement between the lh system s predictions and md results is obvious . while for the uh system , the deviation from md results becomes larger and larger as the temperature increases .
the reason behind this fact is clear , since high order nonlinear terms ignored by the uh system become important at high temperature .
hence the lh system can describe r - phs in the fpu-@xmath58 lattices without pressure .
lattice with zero pressure .
the circles are md results obtained from the power spectrum , the solid line the prediction of the lh system ( @xmath135 [ eq . ( [ eq : b_lower ] ) ] ) , and the dashed - dotted line the prediction of the uh system ( @xmath136 [ eq . ( [ eq : b_upper ] ) ] ) .
the inset presents the relative deviation of the sound velocity : @xmath137 with @xmath138 and @xmath139 the theoretical result and the md result of the sound velocity , respectively .
the squares denote @xmath140 [ eq . ( [ eq : b_lower ] ) ] and @xmath136 [ eq . ( [ eq : b_upper ] ) ] in red and blue , respectively . ]
now we begin to apply our variational approach to the fpu-@xmath58 lattice with pressure . the method used to obtain nonzero pressure has been discussed in sec . [
sec : md ] . for simplicity , but without loss of generality , here we only study a specific situation , i.e. , the lattice length is fixed to be @xmath141 . since @xmath142
, we have @xmath143 .
therefore , the pressure @xmath3 can be obtained by solving @xmath144 the pressure is temperature dependent , which is plotted in the inset ( b ) of fig .
[ fig : bfpuvp ] .
since the lattice is slightly elongated , the pressure should be negative . with those values of pressure
, we can calculate the averages in eq .
( [ eq : k1 ] ) for @xmath91 numerically and thus obtain the value of @xmath91 .
as for @xmath88 , we insert the values of pressure into eq .
( [ eq : d2 ] ) and eq .
( [ eq : k2 ] ) , then solve the coupled equations together to get the value of @xmath88 .
the sound velocity of the lh system and the uh system can be obtained by @xmath145 and @xmath146 , respectively .
we present results of the sound velocity in fig .
[ fig : bfpuvp ] .
it is clearly shows that the lh system still gives better results than the uh system , a significant deviation exists between the latter and md results .
so the lh system can also describe r - phs in the fpu-@xmath58 lattices with pressure .
lattice with nonzero pressure .
the circles are md results obtained from the power spectrum , the solid line the prediction of lh system ( @xmath135 [ eq . ( [ eq : k1 ] ) ] ) , the dashed line the prediction of the nfh ( @xmath147 [ eq . ( [ eq : spohn1 ] ) ] ) , and the dashed - dotted line the prediction of the uh system ( @xmath136 [ eq . ( [ eq : k2 ] ) ] ) .
inset ( a ) presents the relative deviation of the sound velocity : @xmath137 with @xmath138 and @xmath139 the theoretical result and the md result of the sound velocity , respectively .
the squares denote @xmath140 [ eq . ( [ eq : k1 ] ) ] , @xmath136 [ eq . ( [ eq : k1 ] ) ] , and @xmath147 [ eq . ( [ eq : spohn1 ] ) ] in red , blue , and cyan , respectively . inset ( b ) shows pressure @xmath3 as a function of temperature @xmath1 for 1d fpu-@xmath58 lattice with @xmath112 . ] the prediction of the sound velocity given by the nfh
is also shown in fig .
[ fig : bfpuvp ] .
it takes a complex form @xmath148 + \frac{1}{2\beta_t}\langle \delta ; \delta\rangle}\right)^{1/2}\nonumber\\ & \equiv & \sqrt{k_{nfh}},\end{aligned}\ ] ] in which all the @xmath149 is short for @xmath150 , i.e. , averages under the single site probability measure eq .
( [ eq : single_rho ] ) , and @xmath151 . noticing the ept and the scpt are unable to deal with systems with pressure , they become invalid in this case
when a cubic term is added to the fpu-@xmath58 potential , we get the asymmetric fpu-@xmath59 potential [ eq .
( [ eq : abp ] ) ] . the potential with different @xmath152
is plotted in fig .
[ fig : afpupo ] .
it is clearly seen that the potential becomes more asymmetric as @xmath152 increases .
when @xmath152 exceeds 2 , @xmath153 begins to have two solutions , the potential tends to become a double - well one which is out of the scope of the present investigation , we can see that from the form of the potential with @xmath154 in the figure . of the 1d fpu-@xmath59 lattice with varying @xmath152 .
] the inherent asymmetry can induce thermal expansion , which has been shown to play a significant role in heat conduction @xcite . in the following
, we will study this lattice by using our variational approach .
we first apply the uh system to the fpu-@xmath59 lattice .
the corresponding parameters @xmath89 and @xmath88 now satisfy @xmath155 we note that @xmath89 is no longer zero due to the asymmetric potential @xmath44 .
the two equations are coupled , we should solve them together to get the value of @xmath88 .
we then utilize the lh system to investigate the lattice .
the parameters of the lh system are given by @xmath156}.\label{eq : ablk}\end{aligned}\ ] ] similarly , @xmath92 is nonzero for an asymmetric potential .
note that the ept still predicts the effective force constant as @xcite @xmath157 with @xmath44 now the fpu-@xmath59 potential eq .
( [ eq : abp ] ) .
it is evident that the denominator of @xmath91 and @xmath158 take the form of the variance and the second moment of @xmath117 , respectively .
this discrepancy results from the asymmetry of the potential , since @xmath159 vanishes for symmetric potentials .
as can be seen later , such a correction improves the accuracy of the results significantly .
we consider lattices at a fixed temperature @xmath160 and vary @xmath152 from 0.2 to 2 .
results of the sound velocity are illustrated in fig .
[ fig : abfpuv ] . for the 1d fpu-@xmath59 lattice with @xmath160 and zero pressure .
the circles are md results obtained from the power spectrum , the dashed line the prediction of the nfh ( @xmath147 with @xmath27 [ eq . ( [ eq : spohn1 ] ) ] ) , the dashed - dotted line the prediction of the uh system ( @xmath136 [ eq . ( [ eq : abuk ] ) ] ) , the red solid line the prediction of the lh system ( @xmath135 [ eq . ( [ eq : ablk ] ) ] ) , and the maroon solid line the prediction of the ept ( @xmath161 [ eq . ( [ eq : ept ] ) ] ) .
the inset presents the relative deviation of the sound velocity : @xmath137 with @xmath138 and @xmath139 the theoretical result and the md result of the sound velocity , respectively .
the squares denote @xmath140 [ eq . ( [ eq : ablk ] ) ] , @xmath136 [ eq . ( [ eq : abuk ] ) ] , @xmath147 with @xmath27 [ eq . ( [ eq : spohn1 ] ) ] , and @xmath161 [ eq . ( [ eq : ept ] ) ] in red , blue , cyan , and maroon , respectively . ] the deviation of predictions of the ept from the measured value
is clearly revealed as @xmath152 increasing .
this demonstrates the invalidity of the ept for the strong asymmetric cases . while the lh system gives accurate results no matter how large the asymmetry is .
it thus shows the significance of the correction term in the denominator of @xmath91 [ eq .
( [ eq : ablk ] ) ] .
still , we observe that the lh system gives much better results than the uh system .
so we can regard the lh system as an effective description of r - phs in the fpu-@xmath59 lattice without pressure .
meanwhile , an agreement between the nfh and the lh system renders a fact that hydrodynamic approximations can also be applied to long - wavelength r - phs in lattices with asymmetric potentials .
now we consider the fpu-@xmath59 lattice with pressure . for simplicity , we only investigate the system at its natural length .
the value of the pressure can be obtained by solving @xmath163 results of pressure as a function of temperature are plotted in the inset ( b ) of fig .
[ fig : abfpuvp ] .
the absolute value of pressure increases as the temperature increases because of the inter - particle interaction is stronger . at the low temperature regime where thermal excitations dwell the very bottom of the potential
, particles can not feel the asymmetry of the potential strongly , so the pressure approaches zero as the temperature tends to zero . with these values
, @xmath91 can be easily calculated from eq .
( [ eq : k1 ] ) . the value of @xmath88 can be obtained by solving eqs .
( [ eq : d2 ] ) and ( [ eq : k2 ] ) together .
since the existing quasi - harmonic theories fail in this model , only our results and the nfh s will be shown .
the result for the case @xmath95 is illustrated in fig .
[ fig : abfpuvp ] . lattice with @xmath95 and nonzero pressure .
the circles are md results obtained from the power spectrum , the solid line the prediction of lh system ( @xmath135 [ eq . ( [ eq : k1 ] ) ] ) , the dashed line the prediction of the nfh ( @xmath147 [ eq . ( [ eq : spohn1 ] ) ] ) , and the dashed - dotted line the prediction of the uh system ( @xmath136 [ eq . ( [ eq : k2 ] ) ] ) .
inset ( a ) presents the relative deviation of the sound velocity : @xmath137 with @xmath138 and @xmath139 the theoretical result and the md result of the sound velocity , respectively .
the squares denote @xmath140 [ eq . ( [ eq : k1 ] ) ] , @xmath136 [ eq . ( [ eq : k1 ] ) ] , and @xmath147 [ eq . ( [ eq : spohn1 ] ) ] in red , blue , and cyan , respectively .
inset ( b ) shows pressure @xmath3 as a function of the temperature @xmath1 for the fpu-@xmath59 lattice with @xmath95 at its natural length . ] from this figure , we see that the lh system works better than the uh system as usual .
we further investigate the fpu-@xmath59 lattices with varying @xmath152 by using the lh system .
the temperature @xmath1 is fixed to be @xmath164 . from the fig .
[ fig : abfpuva ] , it is apparent that the discrepancy between the theoretical prediction and md results tends to increase as @xmath152 increases .
for the 1d fpu-@xmath59 lattice with @xmath160 and nonzero pressure .
the circles are md results obtained from the power spectrum , the solid line the prediction of @xmath135 [ eq .
( [ eq : k1 ] ) ] .
the inset presents the gibbs free energy per particle .
the dashed line is the exact gibbs free energy @xmath165 according to eq .
( [ eq : g_exact ] ) , the solid line is the prediction of the lower bound according to eq .
( [ eq : low ] ) . ]
one possible reason is that the gibbs free energy given by the lower bound depart from the exact result significantly at large @xmath152 ( we can see that from the inset in fig . [
fig : abfpuva ] ) .
this fact indicates that the first cumulant approximation we adopted in deriving the inequalities is not enough .
higher order contributions to the gibbs free energy must be considered in order to improve this method @xcite .
in summary , we have presented a variational approach to study renormalized phonons in momentum conserving nonlinear lattices .
specifically , we obtain two optimal harmonic reference systems , namely , the lh system and the uh system , by optimizing the two bounds of the gibbs free energy of the nonlinear system .
these optimal systems which determine the properties of renormalized phonons can be regarded as optimal harmonic approximations of the nonlinear system .
this method has been applied to lattices with either symmetric or asymmetric potentials , with or without pressure . for lattices with symmetric potentials , such as the fpu-@xmath58 lattice ,
our variational approach gives two optimal and symmetrical reference harmonic lattices .
in the zero pressure case , it is very interesting to find that the sound velocity derived from the lh system reproduce the previous theoretical results from the effective phonon theory and nonlinear fluctuating hydrodynamic theory , and the sound velocity derived from the uh system recover the previous theoretical results from a self - consistent approach . for lattices with asymmetric potentials , such as the fpu-@xmath59 lattice where the existing effective phonon theory fails to predict the sound velocity
, our variational approach can also give accurate predictions .
in particular , our approach reveals the reason why the effective phonon theory can not predict the correct sound velocity . in all the cases , the lh system works better than the uh system ,
since the ensemble averages in the former are evaluated under the probability measure of the nonlinear system , it simultaneously takes nonlinear contributions into account compared with the latter .
therefore , the lh system can be treated as an effective theory of renormalized phonons in various momentum conserving nonlinear lattices . a deviation between the prediction of our variational approach and the md results has also been found for stronger asymmetry and nonzero pressure situation where the underlying mechanism needs further investigations .
compared with the existing quasi - harmonic theories , this approach can be used to investigate nonlinear systems with asymmetric potentials .
we owe this ability to two aspects .
one is the choice of an parameterized asymmetric harmonic potential with a parameter @xmath63 which can quantifies the degree of asymmetry of the potentials .
the other is the consideration of the gibbs free energy instead of the helmholtz free energy , which enables us to deal with systems with pressure .
we also found that the nfh gives satisfactory predictions of sound velocity in all those cases .
the theory focuses on a hydrodynamic description of nonlinear lattices , not a construction of an effective theory of the renormalized phonons in nonlinear lattices .
those agreements means that the hydrodynamic approximation really captures essential features of the systems in the long - wavelength limit .
however , for systems without momentum conservation , the concept of sound velocity is invalid , the nfh may not be able to give the information of the renormalized phonons , although it can still describe correlation functions of those systems @xcite . while our variational approach can be uesd to investigate phonon band gap and phonon dispersion of those systems @xcite .
we acknowledge the financial supports from the national nature science foundation of china with grant no .
11334007 ( n.li and b.li ) , the national basic research program of china(2012cb921401 ) and the program for new century excellent talents of the ministry of education of china with grant no .
ncet-12 - 0409 .
60 s. lepri , r. livi and a. politi , phys .
* 377 * , 1 ( 2003 ) .
a. dhar , adv . phys . * 57 * , 457 ( 2008 )
. s. liu , x. xu , r. xie , g. zhang , and b. li , eur .
j. b * 85 * , 337 ( 2012 ) . c. alabiso , m. casartelli , and p. marenzoni , j. stat
* 79 * , 451 ( 1995 )
. s. lepri , phys .
e * 58 * , 7165 ( 1998 ) . c. alabiso and m. casartelli , j. phys .
a * 34 * , 1223 ( 2001 ) .
b. gershgorin , y. v. lvov , and d. cai , phys .
lett . * 95 * , 264302 ( 2005 ) .
n. li , p. tong , and b. li , europhys .
lett . * 75 * , 49 ( 2006 ) .
b. gershgorin , y. v. lvov , and d. cai , phys .
e * 75 * , 046603(2007 ) .
d. he , s. buyukdagli , and b. hu , phys .
e * 78 * , 061103(2008 ) .
n. li and b. li , phys .
e * 87 * , 042125 ( 2013 ) .
n. li , b. li , and s. flach , phys .
lett . * 105 * , 054102 ( 2010 ) .
n. li and b. li , europhys . lett . *
78 * , 34001 ( 2007 ) .
y. zhang , s. chen , j. wang , and h. zhao , arxiv:1301.2838(2013 ) .
s. liu , j. liu , p. hnggi , c. wu , and b. li , phys .
b * 90*,174304(2014 ) .
m. d. girardeau and r. m. mazo , adv .
* 24 * , 187(1973 ) .
f. fermi , j. pasta , s. ulam , and m. tsingou , studies of nonlinear problems i , los alamos preprint la-1940 , 1955 .
y. zhong , y. zhang , j. wang , and h. zhao , phys .
e * 85 * , 060102 ( 2012 ) .
l. wang , b. hu , and b. li , phys . rev .
e * 88 * , 052112 ( 2013 ) .
a. v. savin and y. a. kosevich , phys .
e * 89 * , 032102(2014 ) .
s. das , a. dhar , and o. narayan , j. stat .
* 154 * , 204 ( 2014 ) .
g. a. mansoori and f. b. canfield , j. chem .
phys . * 51 * , 4958(1969 ) . g. a. mansoori and f. b. canfield , j. chem . phys .
* 53 * , 1618(1970 ) . | we propose a variational approach to study renormalized phonons in momentum conserving nonlinear lattices with either symmetric or asymmetric potentials . to investigate the influence of pressure to phonon properties
, we derive an inequality which provides both the lower and upper bound of the gibbs free energy as the associated variational principle .
this inequality is a direct extension to the gibbs - bogoliubov inequality . taking the symmetry effect into account ,
the reference system for the variational approach is chosen to be harmonic with an asymmetric quadratic potential which contains variational parameters .
we demonstrate the power of this approach by applying it to one dimensional nonlinear lattices with a symmetric or asymmetric fermi - pasta - ulam type potential . for a system with a symmetric potential and zero pressure ,
we recover existing results . for other systems which beyond the scope of existing theories , including those having the symmetric potential and pressure , and those having the asymmetric potential with or without pressure , we also obtain accurate sound velocity . |
shannon s source - channel separation theorem essentially states that asymptotically there is no loss from optimum by decoupling the source coding component and channel coding component in a point - to - point communication system @xcite .
this separation result tremendously simplifies the concept and design of communication systems , and it is also the main reason for the division between research in source coding and channel coding .
however , it is also well known that in many multi - user settings , such a separation indeed incurs certain performance loss ; see , _
e.g. _ , @xcite . for this reason ,
joint source - channel coding has attracted an increasing amount of attention as the communication systems become more and more complex .
one of the most intriguing problems in this area is joint source - channel coding of a gaussian source on a gaussian broadcast channel with @xmath0 users under an average power constraint .
it was observed by goblick @xcite that when the source bandwidth and the channel bandwidth are matched , _
i.e. _ , one channel use per source sample , directly sending the source samples on the channel after a simple scaling is in fact optimal , but the separation - based scheme suffers a performance loss @xcite .
however , when the source bandwidth and the channel bandwidth are not matched , such a simple scheme is no longer optimal .
many researchers have considered this problem , and significant progress has been made toward finding better coding schemes based on hybrid digital and analog signaling ; see , _
e.g. _ , @xcite and the references therein . in spite of the progress on the achievability schemes , our overall understanding on this problem is still quite limited . as pointed out by caire @xcite , the key difficulty appears to be finding meaningful outer bounds .
such outer bounds not only can provide a concrete basis to evaluate various achievability schemes , but also may provide insights into the structure of good or even optimal codes , and may further suggest simplification of the possibly quite complex optimal schemes in certain distortion regimes . in this regard , the result by reznic _
@xcite is particularly important , where they derived a non - trivial outer bound for the achievable distortion region for the two - user system .
this outer bound relies on a technique previously used in the multiple description problem by ozarow @xcite , where one additional random variable beyond those in the original problem is introduced .
the bound given in @xcite is however rather complicated , and was only shown to be asymptotically tight for certain high signal to noise ratio regime .
in this work , we derive an outer bound for the @xmath0-user problem using a similar technique as that used in @xcite , however , more than one additional random variable is introduced .
the technique used here also bears some similarity to that used in @xcite .
the outer bound has a more concise form than the one given in @xcite , but for the @xmath1 case , it can be shown that they are equivalent . this outer bound is in fact a set of outer bounds parametrized by @xmath2 non - negative variables .
though one can optimize over these variables to find the tightest one , this optimization problem appears difficult .
thus we take an approach similar to the one taken in @xcite , and choose some specific values for the @xmath2 variables which gives specific outer bounds . moreover , by combining these specific outer bounds with the simple achievability scheme based on source - channel separation , we provide approximate characterizations of the achievable distortion region within some universal constant multiplicative factors , independent of the signal to noise ratio and the bandwidth mismatch factor . in one of the approximations , the multiplicative factor is roughly of form @xmath3 for the distortion at the @xmath4-th user , while in the other , the factor is @xmath0 for all the distortions .
thus although shannon s source - channel separation result does not hold strictly in this problem , it indeed holds in an approximate manner .
in fact , this set of results is extremely flexible , and can be applied in the case with an infinite number of users but the minimum achievable distortion is bounded away from zero , for which we can conclude that the source - channel separation based approach is also within certain finite constant multiplicative factors of the optimum .
in this case , these constants can be upper bounded by factors related to the disparity between the best and worse distortions , which is not influenced by the number of users being infinite .
though the outer bound is derived using techniques that have some precedents in the information theory literature , the difficulty lies in determining which terms to bound . in contrast to pure source coding problems or pure channel coding problems , where we can usually meaningfully bound a linear combination of rates , in a joint source - channel coding problem the notion of rates does not exist . in @xcite ,
the lower bound on one distortion is given in terms of a function of the other distortion in the two - user problem .
it is clear that such a proof approach becomes unwieldy for the general @xmath0-user case . in this work ,
we instead derive bounds for some quantity which at the first sight may seem even unrelated to the problem , but eventually serves as an interface between the source and channel coding components , thus replacing the role of `` rates '' in traditional shannon theory proofs . inspired by a recent work of avestimehr , caire and tse @xcite , where source - channel separation in more general networks is considered ,
we further show that our technique can be conveniently extended to general broadcast channels , and the source - channel separation based scheme is within the same multiplicative constants of the optimum as for the gaussian channel case .
the rest of the paper is organized as follows .
section [ sec : definition ] gives the necessary notation and reviews an important lemma useful in deriving the outer bound .
the main results are presented in section [ sec : mainresult ] , and the proofs for these results are given in section [ sec : proof ] .
the extension to general broadcast channels is given in section [ sec : general ] , and section [ sec : conclusion ] concludes the paper .
in this section , we give a formal definition of the gaussian source broadcast problem in the context of gaussian broadcast channels ; the notation will be generalized in section [ sec : general ] when other broadcast channels are considered .
let @xmath5 be a stationary and memoryless gaussian source with zero - mean and unit - variance .
the vector @xmath6 will be denoted as @xmath7 .
we use @xmath8 to denote the domain of reals , and @xmath9 to denote the domain of non - negative reals .
the gaussian memoryless broadcast channel is given by the model @xmath10 where @xmath11 is the channel output observed by the @xmath4-th receiver , and @xmath12 is the zero - mean additive gaussian noise on the channel input @xmath13 .
thus the channel is memoryless in the sense that @xmath14 is a stationary and memoryless process .
the variance of @xmath12 is denoted as @xmath15 , and without loss of generality , we shall assume @xmath16 the mean squared error distortion measure is used , which is given by @xmath17
. the encoder maps a source sample block of length @xmath18 into a channel input block of length @xmath19 , and each decoder maps the corresponding channel output block of length @xmath19 into a source reconstruction block of length @xmath18 .
the bandwidth mismatch factor is thus defined as @xmath20 which is essentially the ( possibly fractional ) channel uses per source sample ; see fig . [
fig : systemdiag ] .
the channel input is subject to an average power constraint .
, width=302 ] we can make the codes in consideration more precise by introducing the following definition . an @xmath21 gaussian source - channel broadcast code is given by an encoding function @xmath22 such that @xmath23 and @xmath0 decoding functions @xmath24 and their induced distortions @xmath25 where @xmath26 is the expectation operation .
note that there are two kinds of independent randomness in the system , the first of which is by the source , and the second is by the channel noises ; the expectation operation in ( [ eqn : expectation ] ) is taken over both of them . in the definition , @xmath27 in the expression @xmath28 is understood as the length-@xmath19 vector addition . from the above definition ,
it is clear that the performance of any gaussian joint source - channel code depends only on the marginal distribution of @xmath29 , but not the joint distribution @xmath30 .
this implies that physical degradedness does not differ from statistical degradedness in terms of the system performance .
since the gaussian broadcast channel is always statistically degraded , we shall assume physical degradedness from here on without loss of generality .
the channel noises can thus be written as @xmath31 where @xmath32 is a zero - mean gaussian random variable with variance @xmath33 , which is independent of everything else ; for convenience , we define @xmath34 , and it follows that @xmath35 and @xmath36 .
[ def : distortionvector ] a distortion vector @xmath37 , where @xmath38 is achievable under power constraint @xmath39 and bandwidth mismatch factor @xmath40 , if for any @xmath41 and sufficiently large @xmath18 , there exist an integer @xmath42 and an @xmath21 gaussian source - channel broadcast code such that @xmath43 note that the constraint @xmath38 is without loss of generality , because otherwise the problem can be reduced to an alternative one with fewer users due to the assumed physical degradedness . the collection of all the achievable distortion vectors under power constraint @xmath39 and bandwidth mismatch factor @xmath40 is denoted by @xmath44 , and this is the region that we are interested in .
one important result we need in this work is the following lemma , which is a slightly different version of the one given in @xcite .
[ lemma : difference ] let @xmath45 be a random variable jointly distributed with the gaussian source vector @xmath7 in the alphabet @xmath46 , such that there exists a deterministic mapping @xmath47 satisfying @xmath48 let @xmath49 and @xmath50 , where @xmath51 and @xmath52 are mutually independent gaussian random variables independent of the gaussian source @xmath53 and the random variable @xmath45 , with variance @xmath54 and @xmath55 , respectively . then with @xmath56 and @xmath57 , we have 1 . * mutual information bound * @xmath58 2 . * bound on mutual information difference *
@xmath59 the proof of this lemma is almost identical to the one given in @xcite .
the only difference between the two versions is that in @xcite the random variable @xmath45 is in fact a deterministic function of @xmath7 , however it is rather straightforward to verify that this condition was never used in the proof given in @xcite ; we include the proof of this lemma in the appendix for completeness .
our main results for gaussian source broadcast on gaussian broadcast channels are summarized in theorem [ theorem : maintheorem ] , corollary [ corollary : firstcorollary ] , proposition [ proposition : firstcorollary ] , corollary [ corollary : infiniteusers ] and corollary [ corollary : secondcorollary ] , the proofs of which are given in the next section ; extensions of these results to general broadcast channels are given in section [ sec : general ] .
define the region in ( [ eqn : innerbound ] ) on the top of next page , which is in fact the inner bound via source - channel separation .
next define the regions in ( [ eqn : firstoutbound ] ) and ( [ eqn : secondouterbound ] ) also on the top of next page , which are in fact outer bounds to the achievable distortion region .
we have the the following theorem .
@xmath60 [ theorem : maintheorem ] @xmath61 theorem [ theorem : maintheorem ] is stated as inner and outer bounds to the achievable distortion region , however it can be observed that the bounds have similar forms , and their difference , in terms of distortions , can be bounded by some multiplicative constants .
the following corollary follows directly from theorem [ theorem : maintheorem ] , by comparing ( [ eqn : innerbound ] ) and ( [ eqn : firstoutbound ] ) .
[ corollary : firstcorollary ] if @xmath62 , and if @xmath63 for @xmath64 , then @xmath65 .
the condition @xmath63 in corollary [ corollary : firstcorollary ] is to ensure that the distortion vector @xmath66 satisfies the monotonicity requirement in definition [ def : distortionvector ] and ( [ eqn : innerbound ] ) .
this result has the following intuitive interpretation if the condition indeed holds that @xmath67 for all @xmath64 : if a genie helps the separation - based scheme by giving each individual user half a bit information per source sample , and at the same time all the better users also receive this half a bit information for free , then the separation - based scheme is as good as the optimal scheme
. this approximation can in fact be refined , and for this purpose , the following additional definition is needed . for any @xmath68
, we associate with it a _ relaxed distortion vector _ @xmath69 and a binary labeling vector @xmath70 in a recursive manner @xmath71 for @xmath72 , and we have defined @xmath73 for convenience . it is easily verified that @xmath74 for @xmath64 , and moreover @xmath75 .
[ proposition : firstcorollary ] let @xmath69 be the relaxed distortion vector of @xmath76 . if @xmath62 , then @xmath77 .
the notion of relaxed distortion vector essentially removes the rather artificial condition @xmath63 in corollary [ corollary : firstcorollary ] .
when this condition does not hold for some @xmath4 , the relaxed distortion vector is introduced to replace @xmath66 , which in this case does not satisfy the monotonicity requirement in definition [ def : distortionvector ] and thus is not a valid choice of a distortion vector ; nevertheless , in this case , the difference between the original distortion vector and its relaxed version is in fact smaller , being @xmath78 , instead of @xmath3 for @xmath79 as in the case already considered in corollary [ corollary : firstcorollary ] .
proposition [ proposition : firstcorollary ] can be used in the situation where there are an infinite number of users such as in a fading channel .
let the set of users indexed by @xmath80 and their associated distortions be denoted as @xmath81 , since there may be an uncountably infinite many of them .
if we apply the construction given in ( [ eqn : enhanceddistortion ] ) , with @xmath82 replaced by @xmath83 , @xmath84 taking the role of @xmath85 and @xmath86 taking the role of @xmath87 , then the following lemma is straightforward by observing that @xmath88 and @xmath89 .
[ lemma : maximumgroups ] the sequence @xmath83 specified by ( [ eqn : enhanceddistortion ] ) satisfies @xmath90 . it is clear that the maximum multiplicative constant is less than @xmath91 in the statement of proposition [ proposition : firstcorollary ] . if there exists a lower bound on the achievable distortion for the best user , denoted as @xmath92 , which is strictly positive , _
i.e. _ , @xmath93 , then since @xmath94 , the multiplicative factor can be bounded as @xmath95 thus even when the number of users is infinite , as long as the lower bound @xmath92 is bounded away from zero , the multiplicative factors are in fact finite .
more formally , we have the following corollary , however a more rigorous approach is to derive the outer bounds for this case and show the result holds .
this can indeed be done either along the line of the proof given in section [ sec : proof ] with careful replacement of summation by integral , or more straightforwardly along the line of proof given in section [ sec : general ] . ] .
[ corollary : infiniteusers ] for an infinite number of users indexed by @xmath80 with @xmath96 , let @xmath97 be the relaxed distortion vector of @xmath98 . if @xmath99 , then @xmath100 , and furthermore , @xmath101 .
the next corollary gives another version of the approximation , essentially stating that for any achievable distortion vector , its @xmath0-fold multiple is achievable using the separation approach . in terms of the genie - aided interpretation
, the genie only needs to provide @xmath102 bits common information to the users in the separation - based scheme , then it is as good as the optimal scheme .
more formally , the following corollary follows directly from theorem [ theorem : maintheorem ] .
[ corollary : secondcorollary ] if @xmath62 , then @xmath103 , where @xmath104 .
theorem [ theorem : maintheorem ] , proposition [ proposition : firstcorollary ] and the corollaries provide approximate characterizations of the achievable distortion region , essentially stating that the loss of the source - channel separation approach is bounded by constants .
the bound on the gap is chosen to be ( largely ) independent of a specific distortion tuple on the boundary of @xmath44 , but it will become clear in the next section that such a choice is not necessary .
the proofs of theorem [ theorem : maintheorem ] and proposition [ proposition : firstcorollary ] rely heavily on the following outer bound , which is one of the main contributions of this work . [
theorem : outerbound ] let @xmath105 be any @xmath2 non - negative real values , and @xmath106 .
if @xmath62 , then @xmath107^{\frac{1}{b}}\leq p+n_1.\label{eqn : outerbound}\end{aligned}\ ] ] with the above theorem in mind , let us denote the set of distortion vectors satisfying ( [ eqn : outerbound ] ) for a specific choice of @xmath105 as @xmath108 , _
i.e. _ , ( [ eqn : defineoutbounds ] ) as given on the top of next page .
@xmath109^{\frac{1}{b}}\leq p+n_1,\right.\nonumber\\ & \qquad\qquad\qquad\qquad\qquad\qquad\left.\phantom{\sum_{k=1}^k \delta n_k \left[\frac{(1+\tau_k)\prod_{j=2}^k(d_j+\tau_{j-1})}{\prod_{j=1}^k ( d_j+\tau_j)}\right]^{\frac{1}{b } } } 1\geq d_1\geq d_2 \geq\ldots\geq d_k\geq 0\right\}.\label{eqn : defineoutbounds}\end{aligned}\ ] ] thus theorem [ theorem : outerbound ] essentially states that @xmath110 for any valid choice of @xmath111 .
the following corollary is then immediate .
[ corollary : outerbound ] @xmath112 to illustrate corollary [ corollary : outerbound ] , let us consider the case @xmath1 for which the bound involves only one parameter @xmath113 .
for this case , it can be shown through some algebra that this outer bound is equivalent to the one given in @xcite . in fig .
[ fig : outerbound ] , we illustrate the outer bounds for several specific choices of @xmath114 . for comparison , the achievable region using the proposed scheme in @xcite
is also given .
note that although the inner bound given by this scheme is extremely close to the outer bound , it appears they do not match exactly .
it is worth emphasizing that we view this outer bound differently from the authors in @xcite : for each possible value of @xmath113 we view the condition ( [ eqn : outerbound ] ) as specifying an outer bound for the distortion region @xmath115 ; in contrast , the authors of @xcite viewed the distortion @xmath116 as being lower bounded by a function of @xmath85 , and the parameter @xmath113 was viewed as an additional variable that is subject to optimization , and consequently only the optimal choice of @xmath113 value was of interest .
these two views are complementary , however the former view appears to be more natural for the @xmath0-user problem , which also readily leads to the approximate characterizations . in certain cases , the second view may be more convenient , such as when we are given a specific achievable distortion tuple , and wish to determine how much further improvement is possible or impossible . for @xmath1 , the properties of the outer bound were thoroughly investigated in @xcite . in certain regimes ,
this outer bound in fact degenerates for the case of bandwidth compression , and it is looser than the trivial outer bound with each user being optimal in the point - to - point setting .
due to its non - linear form , the optimization of this bound is rather difficult , and it also appears difficult to determine whether it is always looser than the trivial outer bound in all distortion regimes with bandwidth compression .
nevertheless , it is clear that this outer bound always holds whether the bandwidth is expanded or compressed , and the approximate characterizations are valid in either case .
a different and simpler approximate characterization may in fact be more useful for the bandwidth compression case .
consider a different genie who helps the separation - based scheme by giving each individual user half a bit information _ per channel use _ , and at the same time all the better users also receive this half a bit information for free , then the genie - aided separation - based scheme is as good as the optimal scheme , and moreover each user can in fact achieve the optimal point - to - point distortion . to see this approximation holds , first observe that the following broadcast channel rates are achievable by using the gaussian broadcast channel capacity region characterization @xcite ( it is particularly easy by using the alternative gaussian broadcast channel capacity characterization given in ( [ eqn : capacity ] ) ) @xmath117 the @xmath4-th user can thus utilize a total rate of @xmath118 per channel use on this broadcast channel ; together with the genie - provided rates , it will have at least a total rate of @xmath119 per channel use , _
i.e. _ , the optimal point to point channel rate .
since the gaussian source is successively refinable @xcite , it is now clear that each user can achieve the optimal point - to - point distortion with this genie - aided separation - based scheme .
note that though this approximation is good for bandwidth compression , it can be rather loose when the bandwidth expansion factor is large .
in contrast , the approximations given in theorem [ theorem : maintheorem ] and proposition [ proposition : firstcorollary ] are independent of the bandwidth mismatch factor ( the genie provides information in terms of _ per source sample _ ) ; another difference is that the approximations given in theorem [ theorem : maintheorem ] and proposition [ proposition : firstcorollary ] rely on the new outer bound , instead of the simple point - to - point distortion outer bound .
it is clear from the above discussion that the outer bound in theorem [ theorem : maintheorem ] may be further improved by taking its intersection with the trivial point - to - point outer bound . in the remainder of this paper , we do not pursue such possible improvements , but instead focus on the proofs for the results stated in theorem [ theorem : maintheorem ] and proposition [ proposition : firstcorollary ] .
the proofs of the main results for gaussian source broadcast on gaussian broadcast channels are given in this section .
we start by establishing a simple inner bound for the distortion region @xmath44 based on source - channel separation , and then focus on deriving an outer bound , or more precisely a set of outer bounds .
the approximate characterizations are then rather straightforward by combining these two bounds . from here on , we shall use natural logarithm for concreteness , though choosing logarithm of a different base does not make any essential difference .
the source - channel separation based coding scheme we consider is extremely simple , which is the combination of a gaussian successive refinement source code and a gaussian broadcast channel code ; this scheme was thoroughly investigated in @xcite , and a solution for the optimal power allocation was given to minimize the expected end - user distortion . since gaussian broadcast channel is degraded , a better user can always decode completely the messages sent to the worse users , and thus a successive refinement source code is a perfect match for this channel .
note that such a source - channel separation approach is not optimal in general for this joint source - channel coding problem ; see for example @xcite .
the gaussian broadcast channel capacity region is well known @xcite , which is usually given in a parametric form in terms of the power allocation . in this work , we will use an alternative representation , which first appeared in @xcite and was instrumental for deriving the optimal power allocation solution in @xcite .
the gaussian broadcast channel capacity region ( per channel use ) can be written in the form in ( [ eqn : capacity ] ) as given on the top of next page .
@xmath120 the rate @xmath121 is the individual message rate intended only to the @xmath4-th user , however due to the degradedness , all the better users can also decode this message . since the gaussian source is successively refinable @xcite , by combining an optimal gaussian successive refinement source code with a gaussian broadcast code that ( asymptotically ) achieves ( [ eqn : capacity ] )
, we have the following theorem .
[ theorem : innerbound ]
@xmath122 we wish to show that any @xmath123 is indeed achievable . using the separation scheme ,
we only need to show the channel rates @xmath124 specified by @xmath125 are achievable on this gaussian broadcast channel .
the non - negative vector @xmath124 is uniquely determined by @xmath76 , and it is straightforwardly seen that it indeed satisfies the inequality in ( [ eqn : capacity ] ) .
the proof is thus complete .
next we derive a set of conditions that any achievable distortion vector @xmath76 has to satisfy , _
i.e. _ , theorem [ theorem : outerbound ] .
let us first introduce a set of auxiliary random variables , defined as @xmath126 where @xmath127 s are zero gaussian random variables , with variance @xmath128 , and furthermore @xmath129 where @xmath130 is a zero - mean gaussian random variable , independent of everything else , with variance @xmath131 . for convenience ,
we define @xmath132 , which implies @xmath133 ; furthermore , define @xmath134 , _
i.e. _ , being a constant .
this technique of introducing auxiliary random variables beyond those in the original problem was previously used in @xcite to derive outer bounds , and specifically in @xcite more than one random variable was introduced , whereas in @xcite only one was introduced . for any encoding and decoding functions ,
we consider a quantity which bears some similarity to the expression for the gaussian broadcast channel capacity ( [ eqn : capacity ] ) , and we denote this quantity as @xmath135 due to its sum exponential form @xmath136.\end{aligned}\ ] ] the subscript @xmath137 makes it clear that this quantity depends on the specific encoding and decoding functions .
next we shall derive universal upper and lower bounds for this quantity regardless the specific choice of functions @xmath137 , which eventually yield an outer bound for @xmath44 .
let @xmath137 be any encoding and decoding functions that ( asymptotically ) achieve the distortions @xmath76 .
we first derive a lower bound for @xmath135 . observe that for @xmath138 , @xmath139 where the equality is due to the markov string @xmath140 , and the inequality is by lemma [ lemma : difference ] .
moreover , also by lemma [ lemma : difference ] , we have @xmath141 it follows that @xmath142 summarizing the above bounds , we have @xmath143.\end{aligned}\ ] ] next we turn to upper - bounding @xmath135 , and first write the following .
@xmath144\nonumber\\ & = \frac{2}{n}\sum_{j=1}^k \left[h(y^n_j|u^m_{j-1})-h(y^n_j|u^m_j)\right]\nonumber\\ & = \frac{2}{n}\sum_{j=1}^k h(y^n_j|u^m_{j-1})-\frac{2}{n}\sum_{j=1}^k h(y^n_j|u^m_j).\end{aligned}\ ] ] applying the entropy power inequality @xcite for @xmath145 , we have @xmath146\nonumber\\ & \geq \exp\left[\frac{2}{n}h(y^n_{j+1}|u^m_j)\right]+\exp\left[\log(2\pi e\delta n_j)\right]\nonumber\\ & = \exp\left[\frac{2}{n}h(y^n_{j+1}|u^m_j)\right]+2 \pi e\delta n_j .
\label{eqn : applyentropypower}\end{aligned}\ ] ] for @xmath147 ,
it is clear that @xmath148=\exp\left[\frac{2}{n}h(y^n_k|s^m)\right]\nonumber\\ & \qquad\qquad\qquad\qquad\qquad= 2 \pi en_k=2 \pi e \delta n_k.\end{aligned}\ ] ] by defining @xmath149\triangleq 0 $ ] , it now follows that @xmath150\nonumber\\ & \leq \sum_{k=1}^k \delta n_k \frac{\exp\left[\frac{2}{n}\sum_{j=1}^k h(y^n_j|u^m_{j-1})\right]}{\prod_{j=1}^k\left[\exp(\frac{2}{n}h(y^n_{j+1}|u^m_j))+2 \pi e\delta n_j\right]}.\end{aligned}\ ] ] we bound this summation , by considering the summands in the reversed order , _
i.e. _ , @xmath151 . starting with the summands when @xmath152 and @xmath153 , we have ( [ eqn : firststep ] ) as given on the top of next page @xmath154}{\prod_{j=1}^{k-1}\left[\exp(\frac{2}{n}h(y^n_{j+1}|u^m_j))+2 \pi e\delta n_j\right]}+\delta n_{k } \frac{\exp\left[\frac{2}{n}\sum_{j=1}^{k } h(y^n_j|u^m_{j-1})\right]}{\prod_{j=1}^{k}\left[\exp(\frac{2}{n}h(y^n_{j+1}|u^m_j))+2 \pi e\delta n_j\right]}\nonumber\\ & = \frac{\exp\left[\frac{2}{n}\sum_{j=1}^{k-1 } h(y^n_j|u^m_{j-1})\right]}{\prod_{j=1}^{k-1}\left[\exp(\frac{2}{n}h(y^n_{j+1}|u^m_j))+2 \pi
e\delta n_j\right]}\left[\delta n_{k-1}+\delta n_{k } \frac{\exp\left[\frac{2}{n}h(y^n_k|u^m_{k-1})\right]}{2 \pi e\delta n_k}\right]\nonumber\\ & = \frac{1}{2 \pi e}\frac{\exp\left[\frac{2}{n}\sum_{j=1}^{k-1 } h(y^n_j|u^m_{j-1})\right]}{\pi_{j=1}^{k-2}\left[\exp(\frac{2}{n}h(y^n_{j+1}|u^m_j))+2 \pi e\delta n_j\right]}.\label{eqn : firststep}\end{aligned}\ ] ] continuing this line of reduction , we finally arrive at ( [ eqm : upperbounde ] ) when @xmath155 @xmath156}{\exp\left[\frac{2}{n}h(y^n_{2}|u^m_1))\right]+2 \pi e\delta
n_1}+\frac{1}{2 \pi e}\frac{\exp\left[\frac{2}{n}\sum_{j=1}^{2 } h(y^n_j|u^m_{j-1})\right]}{\left[\exp(\frac{2}{n}h(y^n_{2}|u^m_1))+2 \pi
e\delta n_1\right]}\nonumber\\ & = \frac{\exp\left[\frac{2}{n } h(y^n_1)\right]}{\exp\left[\frac{2}{n}h(y^n_{2}|u^m_1))\right]+2 \pi e\delta n_1}\left[\delta n_1+\frac{\exp \left[\frac{2}{n}h(y^n_2|u^m_1)\right]}{2\pi e}\right]\nonumber\\ & = \frac{\exp\left[\frac{2}{n } h(y^n_1)\right]}{2\pi e}\leq p+n_1,\label{eqm : upperbounde}\end{aligned}\ ] ] where the last inequality is by the concavity of the @xmath157 function and the given power constraint . combining ( [ eqn : lowerbounde ] ) and ( [ eqm : upperbounde ] ) , it is clear that for any encoding and decoding functions @xmath137 @xmath158,\end{aligned}\ ] ] which completes the proof .
the meaning of the newly introduced random variable @xmath159 can be roughly understood as the message meant for the @xmath4-th user . under this interpretation
, the term @xmath160 in the quantity @xmath135 essentially represents the individual rate intended for the @xmath4-th user in the gaussian broadcast channel ; this informal understanding provides the rationale for bounding @xmath135 .
this interpretation is nevertheless not completely accurate , and thus the outer bound is likely to be not tight in general , but suffices to provide approximate characterizations .
now we are ready to prove theorem [ theorem : maintheorem ] and proposition [ proposition : firstcorollary ] .
the first inclusion @xmath161 in theorem [ theorem : maintheorem ] is simply theorem [ theorem : innerbound ] , and thus we focus on the other inclusion @xmath162 , for which we prove @xmath163 and @xmath164 separately . from theorem [ theorem :
outerbound ] , it is clear that if @xmath62 , then ( [ eqn : outerbound ] ) holds with any @xmath165 , and thus ( [ eqn : outerbound ] ) holds when we choose @xmath166 for @xmath64 .
it follows that the following condition has to be satisfied by any achievable distortion vector @xmath167^{\frac{1}{b}}\leq p+n_1.\end{aligned}\ ] ] however , notice that @xmath168^{\frac{1}{b}}\nonumber\\ & \qquad\geq \sum_{k=1}^k \delta n_k \left[\frac{\prod_{j=2}^kd_{j-1}}{\prod_{j=1}^k 2d_j}\right]^{\frac{1}{b}}=\sum_{k=1}^k \delta n_k ( 2^kd_k)^{-\frac{1}{b}}.\end{aligned}\ ] ] it now follows straightforwardly that any achievable distortion vector has to satisfy @xmath169 and @xmath163 is proved . to prove @xmath164 , note again that any achievable distortion vector has to satisfy theorem [ theorem : outerbound ] , and because of the similarity between the forms given in ( [ eqn : secondouterbound ] ) and ( [ eqn : outerbound ] ) , we only need to prove @xmath170 for some specific choice of @xmath171 .
we first consider the case that @xmath172 ; the case that @xmath173 needs to be treated in a slightly different manner , as we shall discuss later .
we take an induction approach , for which the following auxiliary quantities are needed @xmath174 we claim that there exist @xmath171 such that ( [ eqn : kfactor ] ) holds with equality for @xmath64 , and holds for @xmath153 with equality or inequality ; moreover with these @xmath128 s we have @xmath175 .
first consider the case @xmath155 , since the function @xmath176 is monotonically decreasing and continuous in the range @xmath177 $ ] , as long as @xmath178 there exists a unique solution of @xmath113 such that ( [ eqn : kfactor ] ) holds with equality .
this is indeed true for @xmath172 , which gives @xmath179 it follows that @xmath180 and thus our claim holds for @xmath155 .
next suppose the claim is true for @xmath181 and we prove it is also true @xmath182 , for which we wish to find @xmath183 such that @xmath184 where the last equality is by the supposition that the claim holds true for @xmath185 .
again by the monotonicity and continuity of @xmath186 , as long as the choice of @xmath187 satisfies @xmath188 there exists a valid solution @xmath183 in @xmath189 $ ] for ( [ eqn : kstarequal ] ) to hold .
the second inequality in ( [ eqn : inductioninequality ] ) is clearly true , and thus we only need to consider the first inequality .
however , notice that @xmath190 and we thus conclude that there indeed exists a valid solution of @xmath183 for ( [ eqn : kstarequal ] ) , or more precisely @xmath191 to bound @xmath192 , we write @xmath193 where the last inequality is by the monotonicity of @xmath194 in @xmath195 , and the fact @xmath196 .
the induction is thus complete .
it only remains to check that when @xmath153 the inequality ( [ eqn : kfactor ] ) still holds , for which we have @xmath197 where the last inequality is by the fact @xmath198 .
we have thus proved that @xmath164 for the case @xmath172 .
next we briefly discuss the case @xmath199 , and we shall prove that ( [ eqn : kfactor ] ) holds for some @xmath171 .
notice that by choosing @xmath200 sufficiently large , we can make @xmath201 because of the strict inequality given in @xmath202 , and the fact that the left - hand - side of ( [ eqn : lhsrhs ] ) goes to @xmath203 when we send @xmath200 to infinity .
we have also @xmath204 , and thus there exist @xmath205 such that @xmath206 holds with equality for @xmath207 and with either equality or inequality for @xmath153 , by applying the result for the previously discussed case in a system with @xmath208 users .
since we can choose @xmath200 sufficiently large , it is clear that indeed this set of @xmath128 s makes ( [ eqn : kfactor ] ) hold for @xmath209 .
this completes the proof for @xmath164 . to prove proposition [ proposition : firstcorollary ]
, we need to choose different values for @xmath210 .
essentially , when the condition @xmath211 is not satisfied for certain @xmath4 , we will choose to ignore the contribution of @xmath212 in the outer bound of theorem [ theorem : outerbound ] by taking an appropriate value of @xmath213 .
for convenience , define the set @xmath214 , where @xmath215 is the labeling function given before proposition [ proposition : firstcorollary ] ; denote the member of @xmath216 in an increasing order as @xmath217 , where @xmath218 is the cardinality of the set @xmath216 .
the value of @xmath128 s are set by the following recursive formula @xmath219 where we have defined @xmath220 for convenience .
note that @xmath221 for @xmath222 .
it follows from theorem [ theorem : outerbound ] that this achievable distortion vector has to satisfy ( [ eqn : conditions ] ) on the top of next page , @xmath223^{\frac{1}{b}}\nonumber\\ & \qquad\qquad\qquad = n_1-n_{k_1}+\sum_{i=1}^{|\mathcal{b}| } ( n_{k_i}-n_{k_{i+1 } } ) \left[\frac{(1+d_{k_i})\prod_{j=2}^{i}(d_{k_j}+d_{k_{j-1}})}{\prod_{j=1}^{i } ( d_{k_j}+d_{k_j})}\right]^{\frac{1}{b}}\leq p+n_1,\label{eqn : conditions}\end{aligned}\ ] ] where for convenience we define @xmath224 ; to see the equality holds , note that the terms for @xmath225 are combined with the terms of @xmath226 , because this choice of @xmath210 cancels out some of terms in the numerator and the denominator .
this implies that the distortion vector has to satisfy @xmath227 where the last two equalities are by the definition of @xmath228 .
the proof can now be completed by applying theorem [ theorem : innerbound ] .
in this section , we show that the results for sending gaussian sources on gaussian broadcast channels can be conveniently extended to general broadcast channels , which was inspired by a recent work by avestimehr , caire and tse @xcite . the broadcast channel is now given by an arbitrary conditional distribution @xmath229 , in the alphabets @xmath230 .
for brevity , we omit repeating the definition of the codes here . to distinguish from the gaussian channel case , we denote the achievable distortion region as @xmath231 , with a bandwidth mismatch factor @xmath40 . for the separation - based scheme
, we shall consider combining successive refinement source codes with broadcast codes with degraded message set @xcite .
particularly , for an arbitrary permutation @xmath232 , the degraded message set requirement implies that there are a total of @xmath0 independent messages @xmath233 , such that the user @xmath234 should decode the messages @xmath235 . for an arbitrary permutation @xmath236 ,
let us denote the achievable distortion by the separation - based approach of combining successive refinement source code with the broadcast code with degraded message set as @xmath237 , and the overall achievable distortion region using this separation - based approach is given by @xmath238 clearly the convex closure of the above region is also achievable , however such generality is not required .
it is worth noting that since the characterization of the broadcast channel capacity region with degraded message set is still an open problem for @xmath239 , we do not have a characterization for the region @xmath240 .
however , if the broadcast channel is degraded , then only one permutation needs to be considered ; moreover , if the capacity region of the broadcast channel ( with degraded message set or it is a degraded broadcast channel ) is known , such as for the gaussian broadcast channel case , the region @xmath240 can then be completely characterized .
now we present the counterpart of corollary [ corollary : firstcorollary ] , proposition [ proposition : firstcorollary ] , corollary [ corollary : infiniteusers ] and corollary [ corollary : secondcorollary ] for general broadcast channels .
[ corollary : zerocorollaryprime ] if @xmath241 , and if @xmath63 for @xmath64 , then @xmath242 .
[ proposition : firstcorollaryprime ] let @xmath69 be the relaxed distortion vector of @xmath76 as defined in ( [ eqn : enhanceddistortion ] ) . if @xmath243 , then @xmath244 . for
an infinite number of users indexed by @xmath80 with @xmath245 , let @xmath97 be the relaxed distortion vector of @xmath98 as defined in ( [ eqn : enhanceddistortion ] ) . if @xmath246 , then @xmath247 , and furthermore , @xmath101 .
[ corollary : firstcorollaryprime ] [ corollary : secondcorollaryprime ] if @xmath241 , then @xmath248 .
we only prove corollary [ corollary : secondcorollaryprime ] here , since the proofs of corollary [ corollary : zerocorollaryprime ] , proposition [ proposition : firstcorollaryprime ] and corollary [ corollary : firstcorollaryprime ] are quite similar .
assume a distortion vector @xmath76 is indeed achievable with certain joint source - channel coding scheme .
let us view the induced random mapping from @xmath249 to @xmath250 , @xmath251 , by this joint source - channel code as a super - broadcast - channel ; without loss of generality , let us assume @xmath252 .
we pick up the story from ( [ eqn : difference ] ) and ( [ eqn : firstone ] ) , and claim that there exists a degraded message set broadcast code on the super - broadcast - channel with asymptotic rate per each length-@xmath18 block as follows @xmath253 it is not difficult to see that the distribution @xmath254 can be used to construct a well - known super - position code with the above rates @xcite by ( [ eqn : difference ] ) and ( [ eqn : firstone ] ) .
this code needs to span over @xmath255 blocks , and note that the induced super - broadcast - channel is block - wise memoryless .
we only need to confirm that receiver @xmath4 can decode all the messages up to the @xmath4-th layer by successive decoding .
observe that since @xmath256 indeed receiver @xmath4 can decoder the first layer code , and thus recover the codewords based on @xmath257 .
the above inequality essentially shows that the channel to @xmath258 is more powerful than that to @xmath259 , with the given channel input distribution , although the broadcast channel itself is not necessarily degraded . for the @xmath260-th layer
where @xmath261 , we have @xmath262 because of the monotonicity of the function @xmath263 when @xmath264 , and the fact @xmath265 .
thus our claim is indeed true . using this set of degraded message set broadcast channel codes , we can achieve ( or more precisely , approach arbitrarily close to ) the following distortion @xmath228 for any @xmath165 , and @xmath266 , @xmath267 however , in the proof of theorem [ theorem : maintheorem ] , we have already showed that there exist @xmath268 s such that ( [ eqn : kfactor ] ) holds , _
i.e. _ , @xmath269 the proof is thus complete .
we derived a new outer bound to the achievable distortion region for the joint source - channel coding problem of sending a gaussian source on gaussian broadcast channels with bandwidth mismatch .
when combined with a simple separation - based achievability result , this new bound leads to approximate characterizations of the achievable distortion region within some universal constant multiplicative factors .
these results are further extended to the case of gaussian source broadcast on general broadcast channels .
the outer bound was not fully optimized , which seems to be a difficult problem by itself .
it may be beneficial to investigate the outer bound more thoroughly when more powerful achievability schemes become available . in the current work ,
we only considered the separation - based scheme that yields approximate characterizations .
the technique used in this work can be applied to the problem of multi - source broadcast on more complex communication networks under certain conditions .
we believe that similar results hold for many classes of suitably well - behaved networks . in a follow - up work to this paper , we have shown approximate separation for a class of such networks along with other results on source - channel separation @xcite .
the outer bounding technique of introducing auxiliary random variables used in this work is inspired by those used in @xcite .
we believe this technique is also promising for other multi - user information theoretic problems .
the role of the auxiliary random variable introduced in @xcite was not quite well understood or interpreted previously , and may even appear mysterious to many researchers less familiar with the specific problems being treated .
the current work ( see also @xcite ) has made the meaning of the introduced random variables explicit .
more specifically , they are introduced to either substitute the messages in the channel coding problem , or to substitute the source reconstructions in the source coding problem . by this substitution ,
the quantities representing rates are replaced by the corresponding information quantities . with the interpretation made clear in a general manner
, it is our hope that this technique can inspire other meaningful results in the future .
[ appendix : lemmadifference ] define @xmath270 and @xmath271 , and thus @xmath272 and @xmath273 . to prove the first statement ,
we consider the following chain of inequalities @xmath274\\ & \stackrel{(c)}{\geq } mh(u')\\ & \qquad-\sum_{i=1}^m\frac{1}{2}\log\left\{(2\pi e){\mbox{${\mathbb e}$ } } [ ( s(i)+z'(i)-\hat{s}(i))^2 ] \right\}\\ & = mh(u')-\sum_{i=1}^m\frac{1}{2}\log\left[(2\pi e)({\mbox{${\mathbb e}$ } } d(s(i),\hat{s}(i))+\tau ' ) \right],\end{aligned}\ ] ] where @xmath275 is the reconstruction with @xmath45 , and its @xmath276-th position is denoted as @xmath277 .
the inequality ( a ) is because conditioning reduces entropy , ( b ) is because of the chain rule for differential entropy and the fact that conditioning reduces entropy , and ( c ) is because gaussian distribution maximizes the differential entropy for a given second moment .
since @xmath157 is a concave function , we have @xmath278\nonumber\\ & \qquad\leq \frac{m}{2}\log\left[2\pi e \left({\mbox{${\mathbb e}$ } } d(s^n,\hat{s}^n)+\tau'\right)\right].\end{aligned}\ ] ] it follows @xmath279\\ & \geq mh(u')-\frac{m}{2}\log\left[2\pi e(d+\tau')\right]\\ & = \frac{m}{2}\log\frac{1+\tau'}{d+\tau'},\end{aligned}\ ] ] which is the first claim in the lemma . to prove the second claim , we write the following @xmath280 for the latter two terms , we have @xmath281 where ( a ) is because @xmath282 is independent of @xmath283 and @xmath45 ; ( b ) is by the definition of @xmath284 .
continuing the chain of inequalities , we have @xmath285}\\ & = \sum_{i=1}^m i\big{(}v'(i);s(i)-\hat{s}(i)+v(i)+v'(i)\big{)}\\ & \stackrel{(d)}{\geq}\sum_{i=1}^m \frac{1}{2}\log \frac{{\mbox{${\mathbb e}$ } } d(s(i),\hat{s}(i))+\tau'}{{\mbox{${\mathbb e}$ } } d(s(i),\hat{s}(i))+\tau}\\ & \stackrel{(e)}{\geq } \frac{m}{2}\log \frac{d+\tau'}{d+\tau},\end{aligned}\ ] ] where ( a ) is because @xmath52 is independent of @xmath45 ; ( b ) is because conditioning reduces entropy ; ( c ) is by applying the chain rule , and the facts that @xmath282 is an i.i.d . sequence and conditioning reduces entropy ;
( d ) is by applying the mutual information game result that gaussian noise is the worst additive noise under a variance constraint @xcite ( pg .
1 ) , and taking @xmath286 as channel input ; finally ( e ) is due to the convexity and monotonicity of @xmath287 in @xmath288 when @xmath289 . this completes the proof for the second claim .
the authors would like to thank david tse for the stimulating discussions at several occasions as well as his insightful comments .
the authors are also grateful to the anonymous reviewers for their comments .
u. mittal and n. phamdo , `` hybrid digital - analog ( hda ) joint source - channel codes for broadcasting and robust communications , '' _ ieee transactions on information theory _ , vol .
48 , no . 5 , pp .
10821102 , may 2002 .
m. skoglund , n. phamdo , and f. alajaji , `` hybrid digital - analog source - channel coding for bandwidth compression / expansion , '' _ ieee transactions on information theory _ , vol .
52 , no . 8 , pp . 37573763 , aug .
2006 . c. tian , a. steiner , s. shamai , and s. diggavi , `` successive refinement via broadcast : optimizing expected distortion of a gaussian source over a gaussian fading channel , '' , vol .
54 , no . 7 , pp .
29032918 , jul .
2008 .
d. n. c. tse , `` optimal power allocation over parallel gaussian broadcast channels , '' in ; available at http://www.eecs.berkeley.edu/pubs/techrpts/1999/3578.html.[http://www.eecs.berkeley.edu/pubs/techrpts/1999/3578.html ] c. tian , j. chen , s. diggavi and s. shamai , `` optimality and approximate optimality of source - channel separation in networks , '' , austin , tx , usa , jun .
2010 , pp . 495499 .
see also http://arxiv.org/abs/1004.2648 chao tian(s00 , m05 ) received the b.e .
degree in electronic engineering from tsinghua university , beijing , china , in 2000 and the m.s . and
d. degrees in electrical and computer engineering from cornell university , ithaca , ny in 2003 and 2005 , respectively .
tian was a postdoctoral researcher at ecole polytechnique federale de lausanne ( epfl ) from 2005 to 2007 .
he joined at&t labs research , florham park , new jersey in 2007 , where he is now a senior member of technical staff .
his research interests include multi - user information theory , joint source - channel coding , quantization design and analysis , as well as image / video coding and processing .
suhas n. diggavi ( s93 , m99 ) received the b. tech .
degree in electrical engineering from the indian institute of technology , delhi , india , and the ph.d .
degree in electrical engineering from stanford university , stanford , ca , in 1998 . after completing his ph.d . , he was a principal member technical staff in the information sciences center , at&t shannon laboratories , florham park , nj .
since then he had been in the faculty of the school of computer and communication sciences , epfl , where he directed the laboratory for information and communication systems ( licos ) .
he is currently a professor , in the department of electrical engineering , at the university of california , los angeles .
his research interests include wireless communications networks , information theory , network data compression and network algorithms .
he is a recipient of the 2006 ieee donald fink prize paper award , 2005 ieee vehicular technology conference best paper award and the okawa foundation research award .
he is currently an editor for acm / ieee transactions on networking and ieee transactions on information theory .
he has 8 issued patents .
shlomo shamai ( shitz)(s80 , m82 , sm89 , f94 ) received the b.sc .
, m.sc . , and
degrees in electrical engineering from the technion
israel institute of technology , in 1975 , 1981 and 1986 respectively . during 1975 - 1985
he was with the communications research labs in the capacity of a senior research engineer .
since 1986 he is with the department of electrical engineering , technion
israel institute of technology , where he is now the william fondiller professor of telecommunications .
his research interests encompasses a wide spectrum of topics in information theory and statistical communications .
shamai ( shitz ) is an ieee fellow , and the recipient of the 2011 claude e. shannon award .
he is the recipient of the 1999 van der pol gold medal of the union radio scientifique internationale ( ursi ) , and a co - recipient of the 2000 ieee donald g. fink prize paper award , the 2003 , and the 2004 joint it / com societies paper award , the 2007 ieee information theory society paper award , the 2009 european commission fp7 , network of excellence in wireless communications ( newcom++ ) best paper award , and the 2010 thomson reuters award for international excellence in scientific research .
he is he is also the recipient of 1985 alon grant for distinguished young scientists and the 2000 technion henry taub prize for excellence in research .
he has served as associate editor for the shannon theory of the ieee transactions on information theory , and has also served on the board of governors of the information theory society . | we consider the joint source - channel coding problem of sending a gaussian source on a @xmath0-user gaussian broadcast channel with bandwidth mismatch . a new outer bound to the achievable distortion region
is derived using the technique of introducing more than one additional auxiliary random variable , which was previously used to derive sum - rate lower bound for the symmetric gaussian multiple description problem . by combining this outer bound with the achievability result based on source - channel separation , we provide approximate characterizations of the achievable distortion region within constant multiplicative factors .
furthermore , we show that the results can be extended to general broadcast channels , and the performance of the source - channel separation based approach is also within the same constant multiplicative factors of the optimum .
gaussian source , joint source - channel coding , squared error distortion . |
this note offers a contribution to the study of equational characterizations of the parallel composition operation modulo ( variations on ) the classic notion of bisimulation equivalence @xcite .
in particular , we provide a finite equational axiomatization of _
split-2 bisimulation equivalence_a notion of bisimulation equivalence based on the assumption that actions have observable beginnings and endings @xcite over the recursion , relabelling and restriction free fragment of milner s ccs @xcite enriched with an auxiliary operator proposed by hennessy in a 1981 preprint entitled _
`` on the relationship between time and interleaving '' _ and its published version @xcite .
to put this contribution , and its significance , in its research context , we find it appropriate to recall briefly some of the key results in the history of the study of equational axiomatizations of parallel composition in process algebra .
research on equational axiomatizations of behavioural equivalences over process algebras incorporating a notion of parallel composition can be traced at least as far back as the seminal paper @xcite , where hennessy and milner offered , amongst a wealth of other classic results , a complete equational axiomatization of bisimulation equivalence over the recursion free fragment of ccs .
( see the paper @xcite for a more detailed historical account highlighting , e.g. , hans beki s early contributions to this field of research . )
the axiomatization given by hennessy and milner in that paper dealt with parallel composition using the so - called _ expansion law_an axiom schema with a countably infinite number of instances that is essentially an equational formulation of the plotkin - style rules describing the operational semantics of parallel composition .
this raised the question of whether the parallel composition operator could be axiomatized in bisimulation semantics by means of a finite collection of equations .
this question was answered positively by bergstra and klop , who gave in @xcite a finite equational axiomatization of the merge operator in terms of the auxiliary left merge and communication merge operators .
moller clarified the key role played by the expansion law in the axiomatization of parallel composition over ccs by showing in @xcite that strong bisimulation equivalence is _ not _ finitely based over ccs and pa without the left merge operator .
( the process algebra pa @xcite contains a parallel composition operator based on pure interleaving without communication and the left merge operator . )
thus auxiliary operators like the ones used by bergstra and klop are indeed necessary to obtain a finite axiomatization of parallel composition .
moreover , moller proved in @xcite that his negative result holds true for each `` reasonable congruence '' that is included in standard bisimulation equivalence .
in particular , this theorem of moller s applies to split-2 bisimulation equivalence since that equivalence is `` reasonable '' in moller s technical sense . in his paper
@xcite , hennessy proposed an axiomatization of observation congruence @xcite ( also known as rooted weak bisimulation equivalence ) and timed congruence ( essentially rooted weak split-2 bisimulation equivalence ) over a ccs - like recursion , relabelling and restriction free process language .
those axiomatizations used an auxiliary operator , denoted @xmath0 by hennessy , that is essentially a combination of bergstra and klop s left and communication merge operators . apart from having soundness problems ( see the reference @xcite for a general discussion of this problem , and corrected proofs of hennessy s results ) , the proposed axiomatization of observation congruence is _ infinite _ , as it used a variant of the expansion theorem from @xcite . confirming a conjecture by bergstra and klop in @xcite , and answering problem 8 in @xcite , we showed in @xcite that the language obtained by adding hennessy s merge to ccs does _ not _ afford a finite equational axiomatization modulo bisimulation equivalence .
this is due to the fact that , in strong bisimulation semantics , no finite collection of equations can express the interplay between interleaving and communication that underlies the semantics of hennessy s merge .
technically , this is captured in our proof of the main result in @xcite by showing that no finite collection of axioms that are valid in bisimulation semantics can prove all of the equations in the following family : @xmath1 in split-2 semantics , however , these equations are not sound , since they express some form of interleaving .
indeed , we prove that , in sharp contrast to the situation in standard bisimulation semantics , the language with hennessy s merge _ can be finitely axiomatized _ modulo split-2 bisimulation equivalence , and its use suffices to yield a finite axiomatization of the parallel composition operation .
this shows that , in contrast to the results offered in @xcite , `` reasonable congruences '' finer than standard bisimulation equivalence can be finitely axiomatized over ccs using hennessy s merge as the single auxiliary operation
compare with the non - finite axiomatizability results for these congruences offered in @xcite .
the paper is organized as follows .
we begin by presenting preliminaries on the language @xmath2the extension of ccs with hennessy s merge operator and split-2 bisimulation equivalence in sect .
[ sect : hccs ] .
we then offer a finite equational axiom system for split-2 bisimulation equivalence over @xmath2 , and prove that it is sound and complete ( sect .
[ sect : ax ] ) .
this is a companion paper to @xcite , where the interested readers may find further motivation and more references to related literature .
however , we have striven to make it readable independently of that paper . some familiarity with @xcite and the basic notions on process algebras and bisimulation
equivalence will be helpful , but is not necessary , in reading this study .
the uninitiated reader is referred to the textbooks @xcite for extensive motivation and background on process algebras .
precise pointers to material in @xcite will be given whenever necessary .
the language for processes we shall consider in this paper , henceforth referred to as @xmath2 , is obtained by adding hennessy s merge operator from @xcite to the recursion , restriction and relabelling free subset of milner s ccs @xcite .
this language is given by the following grammar : @xmath3 where @xmath4 ranges over a set of _ actions _
we assume that @xmath5 has the form @xmath6 , where @xmath7 is a given set of _ names _ , @xmath8 is the set of _ complement names _ , and @xmath9 is a distinguished action . following milner @xcite , the action @xmath9 will result from the synchronized occurrence of the complementary actions @xmath10 and @xmath11 .
we let @xmath12 range over the set of _ visible actions _ @xmath13 . as usual
, we postulate that @xmath14 for each name @xmath15 .
we shall use @xmath16 to range over process terms .
the _ size _ of a term is the number of operation symbols in it . following standard practice in the literature on ccs and related languages , trailing * 0 * s will often be omitted from terms .
the structural operational semantics for the language @xmath2 given by hennessy in sect .
2.1 of @xcite is based upon the idea that visible actions have a beginning and an ending .
moreover , for each visible action @xmath10 , these distinct events may be observed , and are denoted by @xmath17 and @xmath18 , respectively .
we define @xmath19 in the terminology of @xcite , this is the set of _ events _ , and we shall use @xmath20 to range over it . as usual , we write @xmath21 for the collection of finite sequences of events . the operational semantics for the language @xmath2 is given in terms of binary next - state relations @xmath22 , one for each event @xmath23 . as explained in @xcite , the relations @xmath22 are defined over the set of _ states _
@xmath24 , an extension of @xmath2 obtained by adding new prefixing operations @xmath25 ( @xmath26 ) to the signature for @xmath2 .
more formally , the set of states is given by the following grammar : @xmath27 where @xmath28 ranges over @xmath2 .
intuitively , a state of the form @xmath29 is one in which the execution of action @xmath10 has started , but has not terminated yet .
we shall use @xmath30 to range over the set of states @xmath24 .
@xmath31 @xmath32 @xmath33 @xmath34 [ tab : sos - rules ] the plotkin style rules for the language @xmath24 are given in table [ tab : sos - rules ] ; comments on these rules may be found in ( * ? ? ?
[ def : traces ] for a sequence of events @xmath35 ( @xmath36 ) , and states @xmath37 , we write @xmath38 iff there exists a sequence of transitions @xmath39 if @xmath38 holds for some state @xmath40 , then @xmath41 is a _ trace _ of @xmath42 .
the _ depth _ of a state @xmath42 , written @xmath43 , is the length of the longest trace it affords . in this paper
, we shall consider the language @xmath2 , and more generally the set of states @xmath24 , modulo split-2 bisimulation equivalence @xcite .
( the weak variant of this relation is called _
t - observational equivalence _ by hennessy in @xcite .
later on , this relation has been called _ timed equivalence _ in @xcite . here
we adopt the terminology introduced by van glabbeek and vaandrager in @xcite . )
[ def : bisimulation ] _
split-2 bisimulation equivalence _ , denoted by @xmath44 , is the largest symmetric relation over @xmath24 such that whenever @xmath45 and @xmath46 , then there is a transition @xmath47 with @xmath48 .
we shall also sometimes refer to @xmath44 as _
bisimilarity_. if @xmath45 , then we say that @xmath42 and @xmath49 are _ split-2 bisimilar_. in what follows , we shall mainly be interested in @xmath44 as it applies to the language @xmath2 .
the interested reader is referred to ( * ? ? ?
2.1 ) for examples of ( in)equivalent terms with respect to @xmath44 . here , we limit ourselves to remarking that @xmath44 is a non - interleaving equivalence .
for example , the reader can easily check that the three terms @xmath50 , @xmath51 and @xmath52 are pairwise inequivalent .
it is well - known that split-2 bisimulation equivalence is indeed an equivalence relation .
moreover , two split-2 bisimulation equivalent states afford the same finite non - empty set of traces , and have therefore the same depth .
the following result can be shown following standard lines
see , e.g. , @xcite .
[ fact : congruence ] split-2 bisimilarity is a congruence over the language @xmath2 .
moreover , for all states @xmath53 , if @xmath54 and @xmath55 , then @xmath56 .
a standard question a process algebraist would ask at this point , and the one that we shall address in the remainder of this paper , is whether split-2 bisimulation equivalence affords a finite equational axiomatization over the language @xmath2 . as we showed in @xcite , standard bisimulation equivalence is not finitely based over the language @xmath2 . in particular , we argued there that no finite collection of equations over @xmath2 that is sound with respect to bisimulation equivalence can prove all of the equations @xmath57 where @xmath58 denotes * 0 * , @xmath59 denotes @xmath60 , and the terms @xmath61 are defined thus : @xmath62 note , however , that none of the equations @xmath63 holds with respect to @xmath44 .
in fact , for each @xmath64 , the transition @xmath65 can not be matched , modulo @xmath44 , by the term @xmath66 .
indeed , the only state reachable from @xmath66 via an @xmath17-labelled transition is @xmath67 .
this state is not split-2 bisimilar to @xmath68 because it can perform the transition @xmath69 whereas the only initial event @xmath68 can embark in is @xmath18 .
thus the family of equations on which our proof of the main result from @xcite was based is unsound with respect to split-2 bisimilarity . indeed , as we shall show in what follows , split-2 bisimilarity affords a finite equational axiomatization over the language @xmath2 , assuming that the set of actions @xmath5 is finite .
hence it is possible to finitely axiomatize split-2 bisimilarity over ccs using a single auxiliary binary operation , viz .
hennessy s merge .
let @xmath70 denote the collection of equations in table [ tab : hccs ] . in those equations the symbols @xmath71 are variables .
equation hm6 is an axiom schema describing one equation per visible action @xmath10 .
note that @xmath70 is finite , if so is @xmath5 .
@xmath72 we write @xmath73 , where @xmath74 are terms in the language @xmath2 that may possibly contain occurrences of variables , if the equation @xmath75 can be proven from those in @xmath70 using the standard rules of equational logic . for example , using axioms a1 , a2 , a4 , m , hm1 , hm2 , hm3 and hm4 , it is possible to derive the equations : @xmath76 that state that , modulo @xmath44 , the language @xmath2 is a commutative monoid with respect to parallel composition with @xmath77 as unit element .
( in light of the provability of ( [ eqn : ass ] ) , we have taken the liberty of omitting parentheses in the second summand of the term at the right - hand side of equation hm6 in table [ tab : hccs ] . )
moreover , it is easy to see that : [ fact : nil ] for each @xmath2 term @xmath28 , if @xmath78 , then the equation @xmath79 is provable using a4 , hm4 and m. all of the equations in the axiom system @xmath70 may be found in the axiomatization of t - observational congruence proposed by hennessy in @xcite .
however , the abstraction from @xmath9-labelled transitions underlying t - observational congruence renders axiom hm2 above unsound modulo that congruence .
( see the discussion in ( * ? ? ? * page 854 and sect .
3 ) . ) indeed , to the best of our knowledge , it is yet unknown whether ( t-)observational congruence affords a finite equational axiomatization over ccs , with or without hennessy s merge . our aim , in the remainder of this note , will be to show that , in the presence of a finite collection of actions @xmath5 , split-2 bisimilarity _ is _ finitely axiomatizable over the language @xmath2 .
this is the import of the following : [ thm : main - result ] for all @xmath2 terms @xmath74 not containing occurrences of variables , @xmath80 if , and only if , @xmath81 .
we now proceed to prove the above theorem by establishing separately that the axiom system @xmath70 is sound and complete .
[ prop : soundness ] for all @xmath2 terms @xmath74 , if @xmath81 , then @xmath80 . since @xmath82 is a congruence over the language @xmath2 ( fact [ fact : congruence ] ) , it suffices only to check that each of the equations in @xmath70 is sound .
the verification is tedious , but not hard , and we omit the details .
[ rem : states ] for later use in the proof of proposition [ prop : unique - factorization ] , we note that equations ( [ eqn : identity1])([eqn : ass ] ) also hold modulo @xmath44 when the variables @xmath83 are allowed to range over the set of states @xmath24 .
the proof of the completeness of the equations in @xmath70 with respect to @xmath82 follows the general outline of that of ( * ? ? ?
* theorem 2.1.2 ) . as usual , we rely upon the existence of normal forms for @xmath2 terms . in the remainder of this paper , process terms are considered modulo associativity and commutativity of @xmath84 . in other words
, we do not distinguish @xmath85 and @xmath86 , nor @xmath87 and @xmath88 .
this is justified because , as previously observed , split-2 bisimulation equivalence satisfies axioms a1 , a2 in table [ tab : hccs ] . in what follows , the symbol @xmath89 will denote equality modulo axioms a1 , a2 .
we use a _ summation _ @xmath90 to denote @xmath91 , where the empty sum represents * 0*. [ def : nf ] the set @xmath92 of _ normal forms _ is the least subset of @xmath2 such that @xmath93 where @xmath94 are finite index sets , if the following conditions hold : 1 .
the terms @xmath95 ( @xmath96 ) and @xmath97 ( @xmath98 ) are contained in @xmath92 and 2 .
[ tau - closure ] if @xmath99 for some @xmath100 , then @xmath101 for some @xmath98 .
[ prop : nf ] for each @xmath2 term @xmath28 , there is a term @xmath102 such that @xmath103 .
define the relation @xmath104 on @xmath2 terms thus : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath105 if , and only if , * @xmath106 or * @xmath107 and the size of @xmath28 is smaller than that of @xmath100 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ note that @xmath104 is a well - founded relation , so we may use @xmath104-induction .
the remainder of the proof consists of a case analysis on the syntactic form of @xmath28 .
we only provide the details for the case @xmath108 .
( the cases @xmath109 , @xmath110 and @xmath111 are trivial
the last owing to the fact that @xmath112 is an instance of axiom hm3 , and the case @xmath113 follows from the case that is treated in detail using axiom m. ) assume therefore that @xmath108 .
then @xmath114 and the size of @xmath100 is smaller than that of @xmath28 , so @xmath115 .
hence , by the induction hypothesis there exists @xmath116 such that @xmath117 , say @xmath118 by axioms @xmath119 , @xmath120 , @xmath121 and @xmath122 it follows that @xmath123 since @xmath124 for each @xmath96 and @xmath98 , it follows that @xmath125 and hence @xmath126 and @xmath127 . by the induction hypothesis
there are normal forms @xmath128 , @xmath129 such that @xmath130 .
so @xmath70 proves the equation @xmath131 finally , using equation hm6 , it is now a simple matter to add summands to the right - hand side of the above equation in order to meet requirement [ tau - closure ] in definition [ def : nf ] .
in fact , let @xmath96 and @xmath132 using a4 , we have that @xmath133 is provable from @xmath70 .
then , using hm6 and the induction hypothesis repeatedly , we can prove the equation @xmath134 using this equation as a rewrite rule from left to right in ( [ eqn : temp ] ) for each @xmath96 produces a term meeting requirement [ tau - closure ] in definition [ def : nf ] that is the desired normal form for @xmath135 .
the key to the proof of the promised completeness theorem is an important cancellation result that has its roots in one proven by hennessy for his t - observational equivalence in @xcite .
[ thm : decomposition ] let @xmath136 be @xmath2 terms , and let @xmath10 be a visible action .
assume that @xmath137 then @xmath138 and @xmath139 .
for the moment , we postpone the proof of this result , and use it to establish the following statement , to the effect that the axiom system @xmath70 is complete with respect to @xmath82 over @xmath2 .
[ thm : completeness ] let @xmath74 be @xmath2 terms such that @xmath80
. then @xmath140 .
by induction on the depth of @xmath28 and @xmath100 .
( recall that , since @xmath141 , the terms @xmath28 and @xmath100 have the same depth . ) in light of proposition [ prop : nf ] , we may assume without loss of generality that @xmath28 and @xmath100 are contained in @xmath92 .
let @xmath142 we prove that @xmath143 , from which the statement of the theorem follows by symmetry and transitivity . to this end
, we argue that each summand of @xmath100 can be absorbed into @xmath28 using the equations in @xmath70 , i.e. , that 1 .
[ tauabs ] @xmath144 for each @xmath145 , and 2 .
[ actabs ] @xmath146 for each @xmath147 .
we prove these two statements in turn .
* proof of statement [ tauabs ] .
let @xmath145 . then @xmath148 .
since @xmath138 , there is a term @xmath149 such that @xmath150 and @xmath151 . since @xmath152 , condition [ tau - closure ] in definition [ def : nf ] yields that @xmath153 for some @xmath98 . the induction hypothesis together with closure with respect to @xmath9-prefixing now yields that @xmath154 therefore , using a1a3 , we have that @xmath155 which was to be shown . * proof of statement [ actabs ] .
let @xmath147 .
then @xmath156 .
since @xmath138 , there is a state @xmath42 such that @xmath157 and @xmath158 .
because of the form of @xmath28 , it follows that @xmath159 for some @xmath96 such that @xmath160 . by theorem [ thm :
decomposition ] , we have that @xmath161 since the depth of all of these terms is smaller than that of @xmath28 , we may apply the induction hypothesis twice to obtain that @xmath162 therefore , using a1a3 and @xmath163 , we have that @xmath164 which was to be shown .
the proof of the theorem is now complete .
to finish the proof of the completeness theorem , and therefore of theorem [ thm : main - result ] , we are left to show theorem [ thm : decomposition ] .
our proof of that result relies on a unique decomposition property with respect to parallel composition for states modulo @xmath44 . in order to formulate this decomposition property ,
we shall make use of some notions from @xcite .
these we now proceed to introduce for the sake of completeness and readability .
[ def : prime ] a state @xmath42 is _ irreducible _ if @xmath165 implies @xmath166 or @xmath167 , for all states @xmath168 .
we say that @xmath42 is _ prime _ if it is irreducible and is not split-2 bisimilar to @xmath77 . for example
, each state @xmath42 of depth @xmath169 is prime because every state of the form @xmath170 , where @xmath171 and @xmath172 are not split-2 bisimilar to @xmath77 , has depth at least @xmath173 , and thus can not be split-2 bisimilar to @xmath42 .
[ fact : asp - prime ] the state @xmath174 is prime , for each @xmath2 term @xmath28 and action @xmath10 .
since @xmath174 is not split-2 bisimilar to @xmath77 , it suffices only to show that it is irreducible . to this end ,
assume , towards a contradiction , that @xmath175 for some states @xmath176 that are not split-2 bisimilar to @xmath77 .
then , since @xmath177 , we have that @xmath178 and @xmath179 , for some @xmath180 .
but then it follows that @xmath181 whereas the term @xmath174 can not perform two subsequent @xmath18-transitions . we may therefore conclude that such states @xmath171 and @xmath172 can not exist , and hence that the term @xmath174 is irreducible , which was to be shown . the following result is the counterpart for the language @xmath2 of the unique decomposition theorems presented for various languages in , e.g. , @xcite .
[ prop : unique - factorization ] each state is split-2 bisimilar to a parallel composition of primes , uniquely determined up to split-2 bisimilarity and the order of the primes .
( we adopt the convention that @xmath77 denotes the empty parallel composition . )
we shall obtain this result as a consequence of a general unique decomposition result , obtained by the fourth author in @xcite .
let @xmath182}}$ ] denote the set of states modulo split-2 bisimilarity , and , for a state @xmath183 , denote by @xmath184}}$ ] the equivalence class in @xmath182}}$ ] that contains @xmath42 .
by fact [ fact : congruence ] we can define on @xmath182}}$ ] a binary operation @xmath185 by @xmath186}}{\ensuremath{\mathbin{|}}}{\ensuremath{[t ] } } = { \ensuremath{[s{\ensuremath{\mathbin{|}}}t ] } } \enskip.\ ] ] by remark [ rem : states ] , the set @xmath182}}$ ] with the binary operation @xmath185 and the distinguished element @xmath187}}$ ] is a commutative monoid .
next , we define on @xmath182}}$ ] a partial order @xmath188 by @xmath189}}{\ensuremath{\mathrel{\preccurlyeq}}}{\ensuremath{[s]}}\ \text{iff}\ \text{there exist $ s''\in s$ and $ \sigma\in e^*$ such that $ s{\stackrel{\sigma}{\longrightarrow}}s''{~{\underline{\leftrightarrow}_{\it 2s}}~}s'$}.\ ] ] note that @xmath188 is indeed a partial order ( to establish antisymmetry use that transitions decrease depth , and that split-2 bisimilar states have the same depth ) . for each state @xmath42 , there are a sequence of events @xmath41 and a state @xmath40 such that @xmath190 so @xmath187}}$ ] is the least element of @xmath182}}$ ] with respect to @xmath188 .
furthermore , if @xmath191}}{\ensuremath{\mathrel{\preccurlyeq}}}{\ensuremath{[s]}}$ ] , then @xmath192 , for some @xmath193 and state @xmath194 .
so , using the sos rules for @xmath24 and fact [ fact : congruence ] , it follows that @xmath195 and hence @xmath189}}{\ensuremath{\mathbin{|}}}{\ensuremath{[t ] } } = { \ensuremath{[s'{\ensuremath{\mathbin{|}}}t ] } } { \ensuremath{\mathrel{\preccurlyeq}}}{\ensuremath{[s{\ensuremath{\mathbin{|}}}t ] } } = { \ensuremath{[s]}}{\ensuremath{\mathbin{|}}}{\ensuremath{[t ] } } \enspace .\ ] ] thereby , we have now established that @xmath182}}$ ] with @xmath185 , @xmath187}}$ ] and @xmath188 is a positively ordered commutative monoid in the sense of @xcite . from the sos rules for @xmath24
it easily follows that this positively ordered commutative monoid is _ precompositional _ ( see @xcite ) ,
i.e. , that @xmath196}}{\ensuremath{\mathrel{\preccurlyeq}}}{\ensuremath{[s_1]}}{\ensuremath{\mathbin{|}}}{\ensuremath{[s_2]}}$ , then there are $ { \ensuremath{[s_1']}}{\ensuremath{\mathrel{\preccurlyeq}}}{\ensuremath{[s_1]}}$ , $ { \ensuremath{[s_2']}}{\ensuremath{\mathrel{\preccurlyeq}}}{\ensuremath{[s_2]}}$ s.t.\ $ { \ensuremath{[s]}}={\ensuremath{[s_1']}}{\ensuremath{\mathbin{|}}}{\ensuremath{[s_2']}}$}.\ ] ] consider the mapping @xmath197}}\rightarrow\mathbf{n}$ ] into the positively ordered monoid of natural numbers with addition , @xmath198 and the standard less - than - or - equal relation , defined by @xmath186}}\mapsto{{\mathit{depth}}}(s ) \enskip.\ ] ] it is straightforward to verify that @xmath199 is a _ stratification _
( see @xcite ) , i.e. , that 1 . @xmath200}}{\ensuremath{\mathbin{|}}}{\ensuremath{[t]}}|=|{\ensuremath{[s]}}|+|{\ensuremath{[t]}}|$ ] ; and 2 .
if @xmath184}}\prec{\ensuremath{[t]}}$ ] , then @xmath200}}|<|{\ensuremath{[t]}}|$ ] .
we conclude that @xmath182}}$ ] with @xmath185 , @xmath187}}$ ] and @xmath188 is a stratified and precompositional positively ordered commutative monoid , and hence , by theorem 13 in @xcite , it has unique decomposition .
this completes the proof of the proposition . using the above unique decomposition result , we are now in a position to complete the proof of theorem [ thm : decomposition ] .
thm : decomposition assume that @xmath201 .
using proposition [ prop : unique - factorization ] , we have that @xmath202 and @xmath203 can be expressed uniquely as parallel compositions of primes .
say that @xmath204 for some @xmath205 and primes @xmath206 ( @xmath207 ) and @xmath97 ( @xmath208 ) in the language @xmath2 .
since @xmath174 and @xmath209 are prime ( fact [ fact : asp - prime ] ) and @xmath44 is a congruence ( fact [ fact : congruence ] ) , the unique prime decompositions of @xmath210 and @xmath211 given by proposition [ prop : unique - factorization ] are @xmath212 respectively . in light of our assumption that @xmath213 , these two prime decompositions coincide by proposition [ prop : unique - factorization ] .
hence , as for each @xmath208@xmath214 we have that 1 .
@xmath215 , 2 .
@xmath216 and , without loss of generality , 3 .
@xmath217 for each @xmath207 .
it is now immediate to see that @xmath80 and @xmath218 , which was to be shown .
we thank the referees for suggestions that led to improvements in the presentation of the paper . the work reported in this paper
was carried out while luca aceto was on leave at reykjavk university , wan fokkink was at cwi , and anna inglfsdttir was at decode genetics .
they thank these institutions for their hospitality and excellent working conditions .
luca aceto s work was partially supported by the statens naturvidenskabelige forskningsrd ( danish natural science research council ) , project `` the equational logic of parallel processes '' , nr . 21 - 03 - 0342 .
luca aceto . some of my favourite results in classic process algebra . in vladimiro
sassone , editor , _ bulletin of the european association for theoretical computer science _ , volume 81 , pages 89108 .
luca aceto , wan fokkink , anna ingolfsdottir , and bas luttik . with hennessy
s merge has no finite equational axiomatization .
research report rs-03 - 34 , brics , november 2003 . to appear in _ theoretical computer science_. r. van glabbeek and f. vaandrager .
petri net models for algebraic theories of concurrency . in j. de bakker , a.j .
nijman , and p.c .
treleaven , editors , _ proceedings parle conference , eindhoven , vol .
ii ( parallel languages ) _ , volume 259 of _ lecture notes in computer science _ , pages 224242 .
springer - verlag , 1987 .
bas luttik . a unique decomposition theorem for ordered monoids with applications in process theory . in branislav rovan and peter vojts , editors , _ proceedings of mathematical foundations of computer science 2003 , 28th international symposium , mfcs 2003 , bratislava , slovakia , august 2529 , 2003 _ , volume 2747 of _ lecture notes in computer science _
, pages 562571 .
springer - verlag , 2003 .
f. moller .
the importance of the left merge operator in process algebras . in m.
paterson , editor , _
proceedings @xmath219 icalp , warwick _ , volume 443 of _ lecture notes in computer science _ , pages 752764 .
springer - verlag , july 1990 .
f. moller .
the nonexistence of finite axiomatisations for ccs congruences . in
_ proceedings @xmath220 annual symposium on logic in computer science , philadelphia , usa _ , pages 142153 .
ieee computer society press , 1990 .
d.m.r . park .
concurrency and automata on infinite sequences . in p. deussen , editor , _ @xmath220 gi conference , karlsruhe , germany _ ,
volume 104 of _ lecture notes in computer science _ , pages 167183 .
springer - verlag , 1981 . | this note shows that split-2 bisimulation equivalence ( also known as timed equivalence ) affords a finite equational axiomatization over the process algebra obtained by adding an auxiliary operation proposed by hennessy in 1981 to the recursion , relabelling and restriction free fragment of milner s calculus of communicating systems .
thus the addition of a single binary operation , viz .
hennessy s merge , is sufficient for the finite equational axiomatization of parallel composition modulo this non - interleaving equivalence .
this result is in sharp contrast to a theorem previously obtained by the same authors to the effect that the same language is not finitely based modulo bisimulation equivalence . |
in this paper , our goal is to design new decoding algorithms that can enhance techniques known to date for rm codes .
in general , rm codes can be designed from the set @xmath6 of all @xmath7-variate boolean polynomials of degree @xmath8 or less . here
each polynomial @xmath9 @xmath10 @xmath6 is defined on the @xmath7-dimensional space @xmath11 for any @xmath12 we consider the sequence of binary values @xmath13 obtained as argument @xmath14 runs through @xmath15 .
these sequences - codewords @xmath16 - form an rm code , which is below denoted @xmath17 and has length @xmath18 dimension @xmath19 and distance @xmath20 as follows:@xmath21 the decoding algorithms discussed in this paper ( including the new algorithms ) can be applied to any rm code .
however , we will mostly focus on their asymptotic performance obtained for long rm codes of _ fixed order _ @xmath1 to define their error - correcting performance , we use the following definition . given an infinite sequence of codes @xmath22 we say that a decoding algorithm @xmath23 has a sequence @xmath24 and a sequence @xmath25 if for @xmath26 @xmath27 @xmath23 correctly decodes all but a vanishing fraction of error patterns of weight @xmath28 or less ; @xmath27 @xmath23 fails to decode a nonvanishing fraction of error patterns of weight @xmath24 or less can satisfy the same definition@xmath29@xmath30 nonexponential decoding algorithms known for rm codes can be loosely separated into three groups .
first , _ _
_ was developed in the seminal paper @xcite .
the algorithm requires complexity of order @xmath31 or less . for rm codes of fixed order @xmath32 it was proven in @xcite that majority decoding achieves maximum possible threshold @xmath33 ( here and below we omit index @xmath34 with a residual @xmath35 where @xmath36 is a constant that does not depend on @xmath7 and @xmath1 the second type of decoding algorithms makes use of the symmetry group of rm codes .
one very efficient algorithm is presented in @xcite . for long rm codes
@xmath37 this algorithm reduces the residual term @xmath38 from ( [ eps - maj ] ) to its square @xmath39 where @xmath40 . on the other hand ,
the complexity order of @xmath41 of majority decoding is also increased in algorithm @xcite to almost its square @xmath42 the corresponding thresholds for higher orders @xmath43 are yet unknown .
another result of @xcite concerns maximum - likelihood ( ml ) decoding .
it is shown that ml decoding of rm codes of fixed order @xmath8 yields a substantially lower residual @xmath44 where @xmath45 however , even the best known algorithm of ml decoding designed by the multilevel trellis structure in @xcite has yet complexity that is exponent in @xmath0 . finally , various techniques were introduced in @xcite , @xcite , @xcite , and @xcite .
all these algorithms use different recalculation rules but rely on the same code design based on the _ _
plotkin construction _ _
@xmath46 the construction allows to decompose rm codes @xmath17 onto shorter codes , by taking subblocks * * * * @xmath47 and @xmath48 from codes @xmath49 and @xmath50 the results from @xcite , @xcite , and @xcite show that this recursive structure enables both encoding and bounded distance decoding with the lowest complexity order of @xmath51 known for rm codes of an arbitrary order @xmath8 . in the same vein , below
we also employ plotkin construction .
the basic recursive procedure will split rm code @xmath17 of length @xmath0 into two rm codes of length @xmath52 .
decoding is then relegated further to the shorter codes of length @xmath53 and so on , until we reach basic codes of order @xmath54 or @xmath55 at these points , we use maximum likelihood decoding or the variants derived therefrom .
by contrast , in all intermediate steps , we shall only recalculate the newly defined symbols . here
our goal is to find efficient _ recalculation rules _ that can _ provably _ improve the performance of rm codes .
our results presented below in theorems [ th:1 - 2 ] and [ th:1 - 1 ] show that recursive techniques indeed outperform other polynomial algorithms known for rm codes .
these results also show how decoding complexity can be traded for a higher threshold.@xmath56 [ th:1 - 2 ] long rm codes @xmath17 of fixed order @xmath8 can be decoded with linear complexity @xmath57 and decoding threshold@xmath58 @xmath56 [ th:1 - 1 ] long rm codes @xmath17 of fixed order @xmath8 can be decoded with quasi - linear complexity @xmath59 and decoding threshold@xmath60
@xmath56 rephrasing theorems [ th:1 - 2 ] and [ th:1 - 1 ] , we obtain the following@xmath61 long rm codes @xmath17 of fixed order @xmath8 can be decoded with vanishing output error probability and linear complexity @xmath57 ( or quasi - linear complexity @xmath62 on a binary channel with crossover error probability @xmath63 ( correspondingly , @xmath64 as @xmath65 note that theorem [ th:1 - 2 ] increases decoding threshold of the recursive techniques introduced in @xcite and @xcite from the order of @xmath66 to @xmath52 while keeping linear decoding complexity .
th:1 - 1 ] improves both the complexity and residual of majority decoding of rm codes .
when compared with the algorithm of @xcite , this theorem reduces the quadratic complexity @xmath67 to a quasi - linear complexity @xmath59 and also extends this algorithm to an arbitrary order @xmath68 of rm codes .
the algorithms designed below differ from the former algorithms of @xcite , @xcite , and @xcite in both the intermediate recalculations and the stopping rules .
firstly , we employ new _ intermediate recalculations _ , which yield the exact decoding thresholds , as opposed to the bounded distance threshold @xmath66 established in @xcite and @xcite .
this leads us to theorem [ th:1 - 2 ] .
secondly , by analyzing the results of theorem [ th:1 - 2 ] , we also change the former _ stopping rules _
, all of which terminate decoding at the repetition codes .
now we terminate decoding earlier , once we achieve the biorthogonal codes .
this change yields theorem [ th:1 - 1 ] and substantially improves decoding performance ( this is discussed in section 7 ) .
finally , we employ a new probabilistic analysis of recursive algorithms . in section 7
, we will see that this analysis not only gives the actual thresholds but also shows how the algorithms can be advanced further .
below in section 2 we consider recursive structure of rm codes in more detail . in section 3
, we proceed with decoding techniques and design two different recursive algorithms @xmath69 and @xmath70 these algorithms are analyzed in sections 4 , 5 , and 6 , which are concluded with theorems [ th:1 - 2 ] and [ th:1 - 1 ] . in section 7
, we briefly discuss extensions that include decoding lists , subcodes of rm codes , and soft decision channels . for the latter case
, we will relate the noise power to the quantity @xmath71 thus , the residual @xmath4 will serve as a measure of the highest noise power that can be withstood by a specific low - rate code .
consider any @xmath7-variate boolean polynomial @xmath72 and the corresponding codeword @xmath16 with symbols @xmath73 below we assume that positions @xmath74 are ordered lexicographically , with @xmath75 being the senior digit .
note that any polynomial @xmath9 can be split as @xmath76 where we use the new polynomials @xmath77 and @xmath78 these polynomials are defined over @xmath79 variables and have degrees at most @xmath8 and @xmath80 respectively .
correspondingly , one can consider two codewords @xmath81 and @xmath82 that belong to the codes @xmath49 and @xmath83
. then representation ( [ poly ] ) converts any codeword @xmath84 @xmath17 to the form @xmath46 this is the well known by continuing this process on codes @xmath49 and @xmath85 we obtain rm codes of length @xmath86 and so on . finally , we arrive at the end nodes , which are repetition codes @xmath87 for any @xmath88 and full spaces @xmath89 for any @xmath90 this is schematically shown in fig . 1 for rm codes of length 8
. now let @xmath91 be a block of @xmath92information bits @xmath93 that encode a vector @xmath94 by decomposing this vector into @xmath47 and @xmath95 we also split @xmath96 into two information subblocks @xmath97 and @xmath98 that encode vectors @xmath47 and @xmath95 respectively .
in the following steps , information subblocks are split further , until we arrive at the end nodes @xmath87 or @xmath89 .
this is shown in fig .
2 . note that only one information bit is assigned to the left - end ( repetition ) code @xmath99 while the right - end code @xmath89 includes @xmath100 bits .
below these @xmath100 bits will be encoded using the unit generator matrix . summarizing , we see that any codeword can be encoded from the information strings assigned to the end nodes @xmath87 or @xmath89 , by repeatedly combining codewords @xmath101 and @xmath48 in the @xmath102-construction .
0,0 @xmath103 1,0 1,1 @xmath104 @xmath105 @xmath106 @xmath107 @xmath108 @xmath109 2,0
2,1 2,2 @xmath107 @xmath108 @xmath107 @xmath108 @xmath107 @xmath108 3,0 3,1 3,2 3,3 fig .
1@xmath110 decomposition of rm codes of length 8 . given any algorithm @xmath111 in the sequel we use notation @xmath112 for its complexity .
let @xmath113 denote the encoding described above for the code @xmath114 taking a complexity estimate from @xcite and its enhancement from @xcite , we arrive at the following lemma .
@xmath115 @xmath116 @xmath117 @xmath118 @xmath119 @xmath120 @xmath121 @xmath122 @xmath123 @xmath124 @xmath125 @xmath126 @xmath121 @xmath121 @xmath121 @xmath121 fig .
2 . decomposition of information paths rm codes @xmath17 can be recursively encoded with complexity @xmath127 _ proof .
_ first , note that the end nodes @xmath87 and @xmath89 require no encoding and therefore satisfy the complexity bound ( [ encoding ] ) .
second , we verify that code @xmath17 satisfies ( [ encoding ] ) if the two constituent codes do .
let the codewords @xmath128 and @xmath129 have encoding complexity @xmath130 and @xmath131 that satisfies ( [ encoding ] ) .
then their @xmath132-combination requires complexity @xmath133 where @xmath52 extra additions ( @xmath134 were included to find the right half @xmath135 .
now we substitute estimates ( [ encoding ] ) for quantities @xmath131 and @xmath136 . if @xmath137 then @xmath138 the two other cases , namely @xmath139 and @xmath140 can be treated similarly .
@xmath141 now consider an information bit @xmath93 associated with a left node @xmath99 where @xmath142.$ ] we will map @xmath93 onto a specific `` binary path''@xmath143 of length @xmath7 leading from the origin @xmath17 to the end node @xmath144 .
to do so , we first define the senior bit @xmath145{ll}% 0 , & \text{if}\;a_{j}\in\mathbf{a}_{\,\,r-1}^{m-1},\smallskip\smallskip\\ 1 , & \text{if}\;a_{j}\in\mathbf{a}_{\,\,\,\,\,\,r}^{m-1}. \end{array } \right.\ ] ] next , we take @xmath146 if @xmath93 encodes the left descendant subcode on the following step .
otherwise , @xmath147 similar procedures are then repeated at the steps @xmath148 and give some @xmath149 that arrives at the node @xmath150 we then add @xmath151 right - hand steps and obtain a full path @xmath152 of length @xmath7 that arrives at the node @xmath153 using notation @xmath154 for the sequence of @xmath151 ones , we write @xmath155 now consider any right - end node @xmath89 , where @xmath156 $ ] and let @xmath149 be any right - end path that ends at this node . then @xmath149 is associated with @xmath100 information bits
. therefore we extend @xmath149 to the full length @xmath7 by adding any binary suffix @xmath157 this allows us to consider separately all @xmath100 information bits and use common notation @xmath158 .
when all left- and right - end paths are considered together , we obtain all paths of length @xmath7 and binary weight @xmath159 or more .
this gives one - to - one mapping between @xmath160 information bits and extended paths @xmath161 below all @xmath152 are ordered lexicographically , as @xmath7-digital binary numbers .
now we turn to recursive decoding algorithms .
we map any binary symbol @xmath162 onto @xmath163 and assume that all code vectors belong to @xmath164 obviously , the sum @xmath165 of two binary symbols is being mapped onto the product of their images .
then we consider any codeword @xmath166 transmitted over a binary symmetric channel with crossover probability @xmath167 the received block @xmath168 consists of two halves @xmath169 and @xmath170 , which are the corrupted images of vectors @xmath47 and @xmath171 .
we start with a basic algorithm @xmath172 that will be later used in recursive decoding . in our decoding
, vector @xmath173 will be replaced by the vectors whose components take on real values from the interval @xmath174.$ ] therefore we take a more general approach and assume that @xmath175 * step 1*. we first try to find the codeword @xmath48 from @xmath176 in the absence of noise , we have the equality @xmath177 ( which gives the binary sum of vectors @xmath169 and @xmath170 in the former notation ) . on a noisy channel ,
we first find the `` channel estimate''@xmath178 of @xmath179 next , we employ ( any ) decoding @xmath180 which will be specified later .
the output is some vector @xmath181 and its information block @xmath182.@xmath183 * step 2*. we try to find the block @xmath184 given @xmath185 from step 1 . here
we take two corrupted versions of vector @xmath47 , namely @xmath169 in the left half and @xmath186 in the right half .
these two _ real _ vectors are added and combined in their `` midpoint '' @xmath187 then we use some decoding @xmath188 which is also specified later .
the output is some vector @xmath189 and its information block @xmath190 so , decoding @xmath172 is performed as follows.@xmath183 @xmath191{l}% \text{algorithm } \psi_{\text{rec}}(\mathbf{y)}.\medskip\\ \text{1 .
calculate vector } \mathbf{y}^{v}=\mathbf{y}^{\prime}\mathbf{y}% ^{\prime\prime}\text{.}\\ \text{find } \mathbf{\hat{v}=}\psi(\mathbf{y}^{v})\text { and } \hat { \mathbf{a}}^{v}.\medskip\\ \text{2 . calculate vector } \mathbf{y}^{u}=(\mathbf{y}^{\prime}+\mathbf{y}% ^{\prime\prime}\hat{\mathbf{v}})/2\text{.}\\ \text{find } \mathbf{\hat{u}=}\psi(\mathbf{y}^{u})\text { and } \hat { \mathbf{a}}^{u}.\medskip\\ \text{3 . output decoded components:}\\ \hat{\mathbf{a}}:=(\hat{\mathbf{a}}^{v}\mid\hat{\mathbf{a}}% ^{u});\quad\mathbf{\hat{c}}:=(\mathbf{\hat{u}}\mid\mathbf{\hat{u}\hat{v } } ) . \end{array } $ } % \ ] ] in a more general scheme @xmath69 , we repeat this recursion by decomposing subblocks @xmath192 and @xmath193 further . on each intermediate step
, we only recalculate the newly defined vectors @xmath192 and @xmath193 using ( [ 1 ] ) when decoder moves left and ( [ 3 ] ) when it goes right . finally , vectors @xmath192 and @xmath193 are decoded , once we reach the end nodes @xmath87 and @xmath89 . given any end code @xmath194 of length @xmath195 and any estimate @xmath196
we employ the ( soft decision ) ( md ) decoding @xmath197 that outputs a codeword @xmath198 closest to @xmath199 in the .
equivalently , @xmath198 maximizes the inner product @xmath200 the algorithm is described below .
@xmath191{l}% \text{algorithm } \psi_{\,r}^{m}(\mathbf{y)}.\medskip\\ \text{1 . if } 0<r < m\text { , perform } \psi_{\text{rec}}(\mathbf{y)}\text { } \\ \text{using } \psi(\mathbf{y}^{v})=\psi_{\,\,r-1}^{m-1}\text { and } % \psi(\mathbf{y}^{u})=\psi_{\,\,\,\,\,r}^{m-1}\text{.}\medskip\\ \text{2 . if } r=0,\text { perform md decoding } \\
\psi(\mathbf{y}^{v})\text { for code } \left\ { % tcimacro{\qatop{r}{0}}% % beginexpansion \genfrac{}{}{0pt}{}{r}{0}% % endexpansion \right\ } .\medskip\\ \text{3 .
if } r = m,\text { perform md decoding } \\ \psi(\mathbf{y}^{u})\text { for code } \left\ { % tcimacro{\qatop{r}{r}}% % beginexpansion \genfrac{}{}{0pt}{}{r}{r}% % endexpansion \right\ } .
\end{array } $ } % \ ] ] in the following algorithm @xmath201 , we refine algorithm @xmath202 by terminating decoding @xmath23 at the biorthogonal codes @xmath203 .
@xmath191{l}% \text{algorithm } \phi_{\,r}^{m}(\mathbf{y)}.\medskip\\ \text{1 . if }
1<r < m\text { , perform } \psi_{\text{rec}}(\mathbf{y)}\text { } \\ \text{using } \psi(\mathbf{y}^{v})=\phi_{\,\,r-1}^{m-1}\text { and } % \psi(\mathbf{y}^{u})=\phi_{\,\,\,\,\,r}^{m-1}\text{.}\medskip\\ \text{2 .
if } r=1,\text { perform md decoding } \\
\phi(\mathbf{y}^{v})\text { for code } \left\ { % tcimacro{\qatop{r}{1}}% % beginexpansion \genfrac{}{}{0pt}{}{r}{1}% % endexpansion \right\ } .\medskip\\ \text{3 .
if } r = m,\text { perform md decoding } \\ \phi(\mathbf{y}^{u})\text { for code } \left\ { % tcimacro{\qatop{r}{r}}% % beginexpansion \genfrac{}{}{0pt}{}{r}{r}% % endexpansion \right\ } .
\end{array } $ } % \ ] ] thus , procedures @xmath69 and @xmath201 have a recursive structure that calls itself until md decoding is applied on the end nodes . now
the complexity estimate follows.@xmath183 for any rm code @xmath204 algorithms @xmath69 and @xmath201 have decoding complexity@xmath205 and @xmath206 md decoding can be executed in @xmath0 operations and satisfies the bound ( [ comp - fi ] ) ( here we assume that finding the sign of a real value requires one operation ) . for biorthogonal codes , their md decoding @xmath207 can be executed in @xmath208 operations using the green machine or @xmath209 operations using fast hadamard transform ( see @xcite or @xcite , section 14.4 ) .
obviously , this decoding satisfies the upper bound ( [ comp - f ] ) .
second , for both algorithms @xmath23 and @xmath210 , vector @xmath192 in ( [ 1 ] ) can be calculated in @xmath52 operations while vector @xmath193 in ( [ 3 ] ) requires @xmath211 operations .
therefore our decoding complexity satisfies the same recursion @xmath212 finally , we verify that ( [ comp - fi ] ) and ( [ comp - f ] ) satisfy the above recursion , similarly to the derivation of ( [ encoding ] ) .
@xmath141 _ discussion .
_ both algorithms @xmath69 and @xmath213 admit bounded distance decoding .
this fact can be derived by adjusting the arguments of @xcite for our recalculation rules ( [ 1 ] ) and ( [ 3 ] ) .
algorithm @xmath214 is also similar to decoding algorithms of @xcite and @xcite .
however , our _ recalculation rules _ are different from those used in the above papers . for example
, the algorithm of @xcite performs the so - called `` min - sum '' recalculation@xmath215 instead of ( [ 1 ] ) .
this ( simpler ) recalculation ( [ 1 ] ) will allow us to substantially expand the _ `` provable '' _ decoding domain versus the bounded - distance domain established in @xcite and @xcite .
we then further extend this domain in theorem [ th:1 - 1 ] , also using the new _ stopping rule _ that replaces @xmath216 in @xmath69 with @xmath217 in @xmath218 however , it is yet an open problem to find the decoding domain using any other recalculation rule , say those from @xcite , @xcite , @xcite , or @xcite . finally , note that the scaling factor @xmath219 in recalculation rule ( [ 3 ] ) brings any component @xmath220 back to the interval @xmath174 $ ] used before this recalculation .
this scaling will also allow us to simplify some proofs , in particular that of lemma [ lm : nei1 ] .
however , replacing ( [ 3 ] ) by the simpler rule @xmath221 does not change any decoding results . though being equivalent , the new rule also reduces complexity estimates ( [ comp - fi ] ) and ( [ comp - f ] ) to@xmath222 ) for the only reason to simplify our proofs .
we begin with the algorithm @xmath69 and later will use a similar analysis for the algorithm @xmath70 note that @xmath69 enters each end node multiple times , by taking all paths leading to this node .
it turns out that the output bit error rate ( ber ) significantly varies on different nodes and even on different paths leading to the same node .
therefore our first problem is to fix a path @xmath152 and estimate the output ber for the corresponding information symbol @xmath223 in particular , we will define the most error - prone paths .
consider any ( sub)path of some length @xmath224 let @xmath149 be its prefix of length @xmath225 so that @xmath226 where@xmath227 . \label{sub}%\ ] ] first , note that algorithm @xmath202 repeatedly recalculates its input @xmath173 , by taking either an estimate @xmath192 from ( [ 1 ] ) when a path @xmath152 turns left or @xmath228 from ( [ 3 ] ) otherwise .
the following lemma shows that recursive decoding follows lexicographic order of our paths @xmath229
@xmath230 for two paths @xmath231 and @xmath232 the bit @xmath158 is decoded after @xmath233 if @xmath234 _ proof .
_ given two paths @xmath152 and @xmath235 let @xmath236 be the first ( senior ) position where they disagree . if @xmath237 @xmath235 then @xmath238 and @xmath239 thus , after @xmath236 steps , @xmath231 moves left while @xmath152 moves right .
correspondingly , @xmath231 proceeds first.@xmath240 on any subpath @xmath152 of length @xmath241 algorithm @xmath202 outputs some vector @xmath242 of length @xmath243 next , we derive a recursive expression for @xmath242 using formulas ( [ 1 ] ) and ( [ 3 ] ) . in any step @xmath236 ,
the algorithm first splits @xmath244 into halves @xmath245 and @xmath246 for @xmath247 @xmath242 is given by recursion ( [ 1 ] ) and is rewritten below in the upper line of ( [ main ] ) . if @xmath248 then @xmath242 is obtained from ( [ 3 ] ) .
here we also need the vector @xmath249 decoded on the preceding subpath @xmath250 .
the corresponding output is written in the second line of ( [ main ] ) : @xmath251{ll}% \mathbf{y}^{\prime}(\,\underline{\xi}\,)\mathbf{y}^{\prime\prime}% ( \,\underline{\xi}\ , ) , & \text{if}\;\xi_{s}=0,\\ \mathbf{y}^{\prime}(\,\underline{\xi}\,)/2+\hat{\mathbf{v}}% ( \xi)\mathbf{y}^{\prime\prime}(\,\underline{\xi}\,)/2 , & \text{if}\;\xi_{s}=1 .
\end{array } \right .
\label{main}%\ ] ] finally , consider any left - end path @xmath252 that passes some repetition code @xmath87 .
note that no preceding decodings are used after @xmath152 reaches the repetition code @xmath150 here we define the end result on the path @xmath152 as @xmath253 by taking @xmath254 in the last @xmath151 steps of recursion ( [ main ] ) .
note that md decoding also makes its decision on the entire sum of symbols @xmath255 and outputs the symbol @xmath256 for any right - end code @xmath257 the output is some vector @xmath258 of length @xmath259 again , md decoding takes every symbol @xmath260 on the full path @xmath152 and converts it into the information bit @xmath261 making bit - by - bit decision ( [ md1])@xmath262 this is summarized as @xmath183 for any end path @xmath232 the algorithm @xmath69 decodes the outputs @xmath260 into the information bits @xmath263 using the rule ( [ md1 ] ) .
next , we consider the decoding error probability @xmath264 for any information bit @xmath223 on an additive binary symmetric channel , @xmath264 does not depend on the transmitted codeword @xmath265 and we can assume that @xmath266 .
according to our decoding rule ( [ md1 ] ) , an error event @xmath267 has probability @xmath268 ( here we follow footnote 2 and assume that @xmath269 with probability 1/2 if @xmath270 ) note , however , that recursive output @xmath260 depends on the outputs @xmath271 obtained on all preceding paths @xmath272 to simplify our calculations , we wish to consider the above event @xmath273 conditioned that all preceding decodings are correct@xmath262 this implies that any path @xmath274 gives an information bit @xmath275 and a codeword @xmath271 as follows : @xmath276 this assumption also allows us to simplify our recalculations ( [ 1 ] ) , ( [ 3 ] ) , and ( [ main ] ) by removing all vectors @xmath277 : @xmath278@xmath251{ll}% \mathbf{y}^{\prime}(\,\underline{\xi}\,)\mathbf{y}^{\prime\prime}% ( \,\underline{\xi}\ , ) , & \text{if}\;\xi_{s}=0,\\ \mathbf{y}^{\prime}(\,\underline{\xi}\,)/2+\mathbf{y}^{\prime\prime } ( \,\underline{\xi}\,)/2 , & \text{if}\;\xi_{s}=1 .
\end{array } \right .
\smallskip\label{main2}%\ ] ] therefore our first goal is to find how much unconditional probabilities @xmath264 change given that preceding decodings are correct .
first , let @xmath279 be the leftmost path that begins with @xmath8 zeros followed by @xmath159 ones .
for any path @xmath152 , let @xmath280 and @xmath281 denote the events @xmath282 which include all error vectors that are correctly decoded on the paths @xmath283 or @xmath284 respectively .
we define the complete ensemble of all error vectors by @xmath285 in the sequel , we replace each probability @xmath264 by the probability @xmath286 conditioned that all previous decodings are correct .
the following upper bound ( [ tot2 ] ) conservatively assumes that an information symbol @xmath287 is _ always _ incorrect whenever a failure occurs in any step @xmath283 .
similarly , the upper bound in ( [ tot3 ] ) uses the formula of total probability and adds up probabilities @xmath288 over all paths @xmath152 .
by contrast , the lower bound takes into account that the block is always incorrect given the decoding failure on the first step @xmath289 .
[ lm : prob0]for any path @xmath290 its bit error rate @xmath264 satisfies inequality@xmath291 block error probability @xmath292 satisfies inequalities @xmath293 @xmath183 _ proof .
_ the probability @xmath264 can be estimated as@xmath294 similarly , the total probability @xmath292 is bounded as @xmath295 @xmath240 given any path @xmath232 we now assume that decoder gives correct solutions @xmath296 on all previous paths @xmath272 our next goal is to estimate the decoding error probability@xmath297 where @xmath260 is a random variable ( rv ) , which satisfies simplified recalculations ( [ main2 ] ) . here
we begin with the original probability distribution @xmath298{ll}% 1-p , & \text{if}\;y_{i}=+1,\\ p , & \text{if}\;y_{i}=-1 , \end{array } \right .
\label{init}%\ ] ] where @xmath299 are @xmath0 independent identically distributed ( i.i.d . )
rv that form the received vector @xmath300 _ remark . _
note that the above problem is somewhat similar to that of `` probability density evolution '' researched in iterative algorithms .
namely , in both algorithms the original rv @xmath299 undergo two different transformations , similar to ( [ main2 ] ) .
however , in our setting these transformations can also be mixed in an arbitrary ( irregular ) order that only depends on a particular path @xmath232 in general , and on its current symbol @xmath301 in particular .
@xmath183 to simplify this problem , below we estimate @xmath288 using only the first two moments of variables @xmath299 and their descendants
. this will be done as follows .
first , note that the blocks @xmath169 and @xmath302 used in ( [ main1 ] ) always include different channel bits@xmath262 consequently , their descendants @xmath303 and @xmath304 used in ( [ main2 ] ) are also obtained from different channel bits .
these bits are combined in the same operations .
therefore all symbols @xmath305 of the vector @xmath242 are i.i.d .
this allows us to use the common notation @xmath260 for any random variable @xmath305 obtained on the subpath @xmath161 \2 .
let @xmath306 @xmath260 denote the expectation of any rv @xmath307 below we study the normalized random variables @xmath308 all of which have expectation 1 .
our goal is to estimate their variances@xmath309 then decoding error probability always satisfies chebyshev s inequality @xmath310 \3 . to prove theorem 1 ,
we first consider those left - end paths @xmath152 that pass through the nodes @xmath87 with growing @xmath311 for any such path , we show that the corresponding rv @xmath260 satisfies the central limit theorem as @xmath312 this will allow us to replace chebyshev s inequality ( [ cheb1 ] ) by ( a stronger ) gaussian approximation .
we will also see that the variance @xmath313 rapidly declines as decoding progresses over the new paths @xmath152 .
for this reason , we shall still use chebyshev s inequality ( [ cheb1 ] ) on the remaining paths with @xmath314 which will only slightly increase the block error probability @xmath292 defined in ( [ tot3 ] ) .
our next goal is to recalculate the variances @xmath313 defined in ( [ cheb])@xmath262 let the channel residual @xmath315 be fixed . according to ( [ init ] ) , original channel outputs @xmath299 have the means @xmath316 in which case rv @xmath317 @xmath318 have the variance @xmath319 for any path @xmath320 the variance @xmath313 satisfies the recursions@xmath321{ll}% $ \mu(\xi)+1=(\mu(\,\underline{\xi}\,)+1)^{2},$ & $ \text{if}\;\xi_{s}=0,$% \end{tabular } \ \label{t12}%\]]@xmath321{ll}% $ \qquad\qquad\mu(\xi)=\mu(\,\underline{\xi}\,)/2,$ & $ \text{if}\;\xi _ { s}=1.\medskip$% \end{tabular } \
\label{t1}%\ ] ] _ proof . _
first , we need to find the means @xmath322 of rv @xmath260 to proceed with new variables @xmath323 here we simply replace all three rv used in ( [ main2 ] ) by their expectations . then for any @xmath226 the means @xmath322 satisfy the recursion@xmath324{ll}% e^{2}(\underline{\xi } ) , & \text{if}\;\xi_{s}=0,\medskip\\ e(\underline{\xi } ) , & \text{if}\;\xi_{s}=1.\medskip \end{array } \right . \label{e10}%\ ] ] here we also use
the fact that vectors @xmath325 and @xmath326 are independent and have symbols with the same expectation @xmath327 now we see that the normalized rv @xmath328 satisfy the recursion @xmath329{ll}% z^{\prime}(\underline{\xi})\cdot z^{\prime\prime}(\underline{\xi } ) , & \text{if}\;\xi_{s}=0,\medskip\\ z^{\prime}(\underline{\xi})/2+z^{\prime\prime}(\underline{\xi})/2 , & \text{if}\;\xi_{s}=1,\medskip \end{array } \medskip\medskip\right .
\label{main5}%\ ] ] similarly to ( [ main2 ] ) . by taking @xmath330
we immediately obtain ( [ t12 ] ) and ( [ t1]).@xmath331 _ discussion . _ note that the means @xmath322 only depend on the hamming weight @xmath332 of a binary subpath @xmath161 indeed , a subpath @xmath152 has @xmath333 zero symbols @xmath334 according to ( [ e10 ] ) , the original expectation @xmath335 of rv @xmath299 is squared @xmath333 times and is left unchanged @xmath332 times
. therefore @xmath336 by contrast , equalities ( [ t12 ] ) and ( [ t1 ] ) show that variance @xmath337 depends on positions of all ones in vector @xmath152 .
thus , direct ( nonrecurrent ) calculations of @xmath313 become more involved . in lemma
[ lm : mu ] , we will see that even the simplest paths give rather bulky expressions ( [ mu1 ] ) for @xmath338 for this reason , we use a different approach .
namely , in the next section we find the paths @xmath152 that maximize @xmath313 .
_ _ preliminary discussion .
_ _ consider a channel with crossover error probability _ _ _ _ @xmath339 and residual @xmath340 initially , rv @xmath328 have the variance @xmath341 and always satisfy inequality @xmath342 by definition ( [ cheb ] ) .
according to ( [ t12 ] ) , @xmath343 is always squared when a path @xmath149 is appended by @xmath344 thus , moving from a code @xmath17 to its left descendant @xmath83 is equivalent to the replacement of the original residual @xmath4 by its square @xmath345 in other words , any left - hand movement makes the descendant channel noisier . for small @xmath346 ( very high quality channel ) , squaring @xmath347 is almost insignificant .
however , it becomes more substantial as @xmath348 grows .
by contrast , @xmath348 is always cut in half@xmath349 when @xmath350 in general , any right - hand movement makes the descendant channel less noisy .
for example , we obtain @xmath351 on ( bad ) channels with small residual @xmath340 then performing the right step , the recursion replaces this residual @xmath4 with the quantity almost equal to @xmath352 therefore our first conclusion is that variance @xmath313 increases if @xmath238 is replaced by @xmath353 .
_ neighboring paths .
_ our next step is to consider two `` equally balanced '' movements .
namely , in ( [ loop ] ) below , we consider two subpaths @xmath354 and @xmath355 of length @xmath236 that have the same prefix @xmath356of length @xmath357 but diverge in the last two positions as follows @xmath358{l}% \xi_{-}=(\underline{\xi},0,1),\ \ \ \ \ \ \ \ \ \,\ \,\nearrow\nwarrow\ \\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \,\ \,\
^{0}% \nwarrow\nearrow\,^{1}\ \\ \xi_{+}=(\underline{\xi},1,0).\ \ \ \ \ \underline{\xi}=\xi_{1}, ... ,\xi_{s-2}% \end{array } \right . \label{loop}%\ ] ] we say that @xmath354 and @xmath355 are left and right , correspondingly .
[ lm : nei1 ] any two neighbors @xmath354 and @xmath355 satisfy inequality@xmath359 _ proof .
_ let @xmath360
then we use recursive equations ( [ t12 ] ) and ( [ t1 ] ) , which give @xmath361 therefore ( [ in1 ] ) holds@xmath262@xmath331 _ the weakest paths .
_ now we see that any path @xmath152 that includes two adjacent symbols @xmath362 increases its @xmath313 after permutation ( @xmath363 in this case , we say that this path @xmath152 becomes . from now on , let @xmath364 be the complete set of @xmath160 extended paths @xmath161 also , let @xmath365 be the subset of all left - end paths @xmath152 that enter the node @xmath87 and @xmath366 be the subset of the right - end paths .
given any subset @xmath367 we now say that @xmath368 is the weakest path in @xmath369 if @xmath370 then we have the following.@xmath371 [ lm : path]the weakest path on the full set @xmath364 of all @xmath160 paths is the leftmost path ( [ w1 ] ) .
more generally , for any @xmath142,$ ] the weakest path on the subset @xmath365 is its leftmost path@xmath372 @xmath183 _ proof . _
first , note that on all left - end paths @xmath152 , the variances @xmath373 are calculated after @xmath7 steps , at the same node @xmath153 by contrast , all right - end paths @xmath152 end at different nodes @xmath374 therefore their variances @xmath313 are found after @xmath375 steps . to use lemma [ lm : nei1 ] , we consider an extended right - end path @xmath376 obtained by adding @xmath377 zeros .
then we have inequality @xmath378 since the variance @xmath379 increases after zeros are added . despite this fact ,
below we prove that @xmath289 from ( [ w1 ] ) and @xmath380 from ( [ w2 ] ) still represent the weakest paths , even after this extension .
indeed , now all paths have the same length @xmath7 and the same weight @xmath381 so we can apply lemma [ lm : nei1 ] . recall that each path @xmath382 ends with the same suffix @xmath383 . in this case , @xmath380 is the leftmost path on @xmath365 . by lemma [ lm : nei1 ] , @xmath380 maximizes the variance @xmath313 over all @xmath384 finally , note that @xmath289 is the leftmost path on the total set @xmath364 since all @xmath8 zeros form its prefix @xmath385 .
thus , @xmath289 is the weakest path .
now we find the variances @xmath386 and @xmath387 for the weakest paths @xmath289 and @xmath388 [ lm : mu ] for crossover error probability @xmath339 , the weakest paths @xmath289 and @xmath380 give the variances@xmath389@xmath390 @xmath183 _ proof .
_ consider the weakest path @xmath289 from ( [ w1 ] ) .
the recursion ( [ t12 ] ) begins with the original quantity @xmath391 after completing @xmath8 left steps @xmath392 the result is @xmath393 then we proceed with @xmath159 right steps , each of which cuts @xmath394 in half according to ( [ t1 ] ) .
thus , we obtain equality ( [ t0 ] ) .
formula ( [ mu1 ] ) follows from representation ( [ w2 ] ) in a similar ( though slightly longer ) way .
@xmath240 lemma [ lm : mu ] allows us to use chebyshev s inequality @xmath395 for any path @xmath161 however , this bound is rather loose and insufficient to prove theorem 1 .
therefore we improve this estimate , separating all paths into two different sets .
namely , let @xmath396 be the subset of all left - end paths that enter the node @xmath87 with @xmath311 we will use the fact that any path @xmath397 satisfies the central limit theorem as @xmath7 grows .
however , we still use chebyshev s inequality on the complementary subset @xmath398 in doing so , we take @xmath4 equal to the @xmath399 from theorem 1 : @xmath400 [ lm : fi]for rm codes with @xmath401 and fixed order @xmath8 used on a binary channel with crossover error probability @xmath402 algorithm @xmath69 gives on a path @xmath152 the asymptotic bit error rate @xmath403 with asymptotic equality on the weakest path @xmath289 . _ proof .
_ according to ( [ md2 ] ) , any left - end path @xmath152 gives the rv @xmath404 which is the sum of @xmath405 i.i.d .
limited rv @xmath406 for @xmath407 this number grows as @xmath408 or faster as @xmath312 in this case , the normalized rv @xmath328 satisfies the central limit theorem and its probability density function ( pdf ) tends to the gaussian distribution @xmath409 according to lemmas [ lm : path ] and [ lm : mu ] , the weakest path @xmath410 gives the maximum variance @xmath411 in particular , for @xmath412 equality ( [ t0 ] ) gives @xmath413 using gaussian distribution @xmath414 to approximate @xmath415 we take @xmath416 standard deviations and obtain ( see also remark 1 following the proof ) @xmath417 here we also use the asymptotic@xmath321{ll}% $ q(x)$ & $ \overset{\text{def}}{=}\int_{x}^{\infty}\exp\{-x^{2}/2\}dx/\sqrt { 2\pi}\medskip$\\ & $ \sim\exp\{-x^{2}/2\}/(x\sqrt{2\pi})$% \end{tabular } \ \ \\ ] ] valid for large @xmath418 this yields asymptotic equality for @xmath419 in ( [ gauss1 ] ) .
for any other path @xmath407 @xmath328 is approximated by the normal rv with a smaller variance @xmath420 therefore we use inequality in ( [ gauss1 ] ) : @xmath421 finally , consider any path @xmath152 with @xmath422 in this case , we directly estimate the asymptotics of @xmath423 namely , we use the substitution @xmath412 in ( [ mu1 ] ) , which gives a useful estimate:@xmath424{ll}% 2^{-m - r+g}(2r\ln m)^{-1 } , & \text{if}\;g>\frac{m - r}{2}+\ln m,\smallskip\\ 2^{-(m - r-2)/2}(2r\ln m)^{-1/2 } , & \text{if}\;g<\frac{m - r}{2}-\ln m , \end{array } \right .
\label{mu - low}%\ ] ] @xmath183 thus , we see that for all @xmath314 variances @xmath425 have the same asymptotics and decline exponentially in @xmath426 as opposed to the weakest estimate ( [ small ] )
. then we have @xmath427 which also satisfies ( [ gauss1 ] ) as @xmath312@xmath428 _ discussion .
considering approximation ( [ gauss ] ) for a general path @xmath232 we arrive at the estimate @xmath429 according to theorem xvi.7.1 from @xcite , this approximation is valid if the number of standard deviations @xmath430 is small relative to the number @xmath431 of rv @xmath432 in the sum @xmath433 @xmath434 in particular , we can use ( [ gauss ] ) for the path @xmath435 since ( [ small ] ) gives @xmath436 @xmath183 2 .
note that for @xmath437 variance @xmath438 in ( [ mu - low ] ) declines exponentially as @xmath151 moves away from @xmath159 . on the other hand
, we can satisfy asymptotic condition ( [ gauss3 ] ) for any path @xmath407 if @xmath439 in ( [ gauss3 ] ) is replaced with parameter of a fixed degree as @xmath312 ] @xmath441 as @xmath312 we then use inequality @xmath442 valid for any @xmath407 instead of the weaker inequality ( [ gauss2 ] ) .
thus , the bounds on probabilities @xmath288 rapidly decline as @xmath151 moves away from @xmath381 and the total block error rate @xmath292 also satisfies the same asymptotic bound ( [ gauss1])@xmath30 \3 .
note that the same minimum residual @xmath399 can also be used for majority decoding .
indeed , both the majority and the recursive algorithms are identical on the weakest path @xmath289 .
namely , both algorithms first estimate the product of @xmath443 channel symbols and then combine @xmath444 different estimates in ( [ md2 ] ) .
however , a substantial difference between the two algorithms is that recursive decoding uses the previous estimates to process any other path @xmath152 . because of this , the algorithm outperforms majority decoding in both the complexity and ber @xmath288 for any @xmath445 \4 .
theorem [ lm : fi ] almost entirely carries over to any @xmath446 .
namely , we use the normal pdf @xmath447 for any @xmath397 . here
any variance @xmath373 declines as @xmath4 grows
. therefore we can always employ inequality ( [ gauss2 ] ) , by taking the _ maximum possible _ variance @xmath386 obtained in ( [ small ] ) . on the other hand ,
asymptotic equality ( [ gauss ] ) becomes invalid as @xmath4 grows . in this case ,
tighter bounds ( say , the chernoff bound ) must be used on @xmath448 however , in this case , we also need to extend the second - order analysis of lemma [ lm : nei1 ] to exponential moments .
such an approach can also give the asymptotic error probability @xmath288 for any @xmath449 @xmath450 however , finding the bounds on @xmath288 is an important issue still open to date .
@xmath183 \5 .
it can be readily proven that for sufficiently large @xmath451 the variance @xmath438 becomes independent of @xmath452 similar to the estimates obtained in the second line of ( [ mu - low ] ) .
more generally , more and more paths yield almost equal contributions to the block error rate as @xmath4 grows .
this is due to the fact that the neighboring paths exhibit similar performance on sufficiently good channels .
now theorem 1 directly follows from theorem [ lm : fi ] .
_ proof of theorem 1 .
_ consider a channel with crossover probability @xmath453 for @xmath312 the output block error probability @xmath292 of the algorithm @xmath69 has the order @xmath106at most @xmath454 where @xmath160 is the number of information symbols .
this number has polynomial order of @xmath455 on the other hand , formula ( [ gauss1 ] ) shows that @xmath419 declines faster than @xmath456 for any @xmath457 as a result , @xmath458 next , we note that the error patterns of weight @xmath459 or less occur with a probability @xmath460 since @xmath461 the above argument shows that decoder fails to decode only a vanishing fraction of error patterns of weight @xmath459 or less@xmath262 next , we need to prove that @xmath69 fails to correct nonvanishing fraction of errors of weight @xmath52 or less . in proving this , consider a higher crossover probability @xmath462 where@xmath463 for this @xmath464 our estimates ( [ t0 ] ) and ( [ gauss ] ) show that @xmath465 and @xmath466 also , according to ( [ tot3 ] ) , @xmath467 on the other hand , the central limit theorem shows that the errors of weight @xmath52 or more still occur with a vanishing probability @xmath468 thus , we see that @xmath69 necessarily fails on the weights @xmath52 or less , since the weights over @xmath52 still give a vanishing contribution to the nonvanishing error rate @xmath419 . @xmath240
before proceeding with a proof of theorem [ th:1 - 1 ] , we summarize three important points that will be used below to evaluate the threshold of the algorithm @xmath201 . * 1 . *
the received rv @xmath299 , all intermediate recalculations ( [ main ] ) , and end decodings on the right - end paths @xmath152 are identical in both algorithms @xmath201 and @xmath469 by contrast , any left - end path @xmath470 first arrives at some biorthogonal code @xmath471 of length @xmath472 and is then followed by the suffix @xmath473 also , let @xmath474 denote the @xmath236th codeword of @xmath475 where @xmath476 here we also assume that the first two codewords form the repetition code : @xmath477 for each @xmath478 define its support @xmath479 as the subset of positions that have symbols @xmath480 . here
@xmath481 codewords with @xmath482 have support of the same size @xmath483 whereas @xmath484 also , below @xmath158 denotes any information symbol associated with a path @xmath152 .
let the all - one codeword @xmath485 be transmitted and @xmath173 be received .
consider the vector @xmath486 obtained on some left - end path @xmath149 that ends at the node @xmath487 by definition of md decoding , @xmath486 is incorrectly decoded into any @xmath488 with probability @xmath489 where @xmath490 in our probabilistic setting , each event @xmath491 is completely defined by the symbols @xmath492 which are i.i.d .
recall that lemma [ lm : prob0 ] is `` algorithm - independent '' and therefore is left intact in @xmath201 .
namely , we again consider the events @xmath280 and @xmath281 from ( [ events ] ) .
similarly to ( [ tot4 ] ) , we assume that all preceding decodings are correct and replace the unconditional error probability @xmath264 with its conditional counterpart @xmath493 this probability satisfies the bounds@xmath494 here we take the probability of incorrect decoding into any single codeword @xmath474 as a lower bound ( in fact , below we choose @xmath482 ) , and the union bound as its upper counterpart .
now we take parameters @xmath495 for rm codes with @xmath401 and fixed order @xmath8 used on a binary channel with crossover error probability @xmath496 algorithm @xmath201 gives for any path @xmath152 a vanishing bit error rate@xmath497 @xmath56 _ proof .
_ consider any left - end path @xmath149 that ends at the node @xmath487 for any vector @xmath486 and any subset @xmath369 of @xmath498 positions , define the sum @xmath499 here @xmath500 form @xmath498 i.i.d .
thus , the sum @xmath501 has the same pdf for any @xmath369 . in turn , this allows us to remove index @xmath369 from @xmath502 and use common notation @xmath503 then we rewrite bounds ( [ b - fi ] ) as@xmath504 equivalently , we use the normalized rv @xmath505 with expectation 1 and rewrite the latter bounds as@xmath506 similarly to the proof of theorem [ lm : fi ] , note that the sum @xmath328 also satisfies the central limit theorem for any @xmath507 and has pdf that tends to @xmath508 as @xmath312 thus , we see that @xmath288 depends only on the variance @xmath509 obtained on the sum @xmath328 of i.i.d .
this variance can be found using calculations identical to those performed in lemmas [ lm : nei1 ] to [ lm : mu ] . in particular , for any @xmath151 we can invoke the proof of lemma [ lm : path ] , which shows that @xmath313 achieves its maximum on the leftmost path@xmath510 similarly to ( [ t0 ] )
, we then find@xmath511 direct substitution @xmath512 in ( [ t14 ] ) gives @xmath513 now we almost repeat the proof of theorem [ lm : fi ] . for the first path @xmath435 we employ gaussian approximation @xmath514 as @xmath312 for maximum @xmath515 the latter inequality and ( [ main6 ] ) give the upper bound @xmath516 also , @xmath517 for any other path @xmath152 with @xmath518 we can use the same estimates in ( [ main6 ] ) due to the inequalities @xmath519 and @xmath520 finally , consider any path @xmath152 with @xmath521 in this case , we use chebyshev s inequality instead of gaussian approximation .
again , for any node @xmath475 we can consider its leftmost path @xmath522 similarly to our previous calculations in ( [ t0 ] ) and ( [ mu1 ] ) , it can be verified that@xmath523 then for small @xmath524 , substitution @xmath525 gives the equality @xmath526 thus , we obtain chebyshev s inequality in the form @xmath527 and complete the proof , since bound ( [ log2 ] ) combines both estimates ( [ t15 ] ) and ( [ cheb3]).@xmath528 _ proof of theorem 2 .
_ we repeat the proof of theorem 1 almost entirely .
consider a channel with crossover probability @xmath529 as @xmath312 the output block error probability @xmath292 of the algorithm @xmath201 satisfies the estimate @xmath530 where the number @xmath160 of different paths @xmath152 is bounded by @xmath455 formula ( [ log2 ] ) shows that all @xmath288 decline exponentially in @xmath531 as a result , we obtain asymptotic estimate @xmath458 on the other hand , the error patterns of weight @xmath459 or less occur with a total probability that tends to @xmath532 so decoder fails to decode only a vanishing fraction of these error patterns .
now we take a smaller residual@xmath533 and consider a channel with crossover probability @xmath534 then direct substitution of @xmath4 in ( [ t14 ] ) gives @xmath535 .
then formula ( [ lo1 ] ) shows that the decoding block error rate is @xmath536 note also that errors of weight @xmath52 or more occur with vanishing probability .
thus , @xmath69 fails on errors of weight @xmath52 or less .
@xmath537 _ discussion .
_ the proofs of theorems 1 and 2 also reveal the main shortcoming of our probabilistic technique , which employs rather loose estimates for probabilities @xmath538 indeed , the first two moments of the random variables @xmath260 give tight approximation only for gaussian rv .
by contrast , error probabilities @xmath288 slowly decline as @xmath539 whenever chebyshev s inequality ( [ cheb3 ] ) is applied for small parameters @xmath422 as a result , we can obtain a vanishing block error rate only if @xmath540 this is the case of rm codes of fixed order @xmath1 by contrast , the number of information symbols @xmath160 is linear in @xmath0 for rm codes of fixed rate @xmath541 this fact does not allow us to extend theorems 1 and 2 for nonvanishing code rates .
more sophisticated arguments - that include the moments @xmath542@xmath543 of an arbitrary order @xmath236 - can be developed in this case .
the end result of this study is that recursive decoding of rm codes @xmath17 of fixed rate @xmath544 achieves the error - correcting threshold @xmath545 this increases @xmath546 times the threshold of bounded distance decoding .
however , the overall analysis becomes more involved and is beyond the scope of this paper .
now consider an infinite sequence of optimal binary codes of a low code rate @xmath544 used on a channel with high crossover error probability @xmath547 . according to the shannon coding theorem , ml decoding of such a sequence gives a vanishing block error probability if @xmath548 where @xmath549 is the inverse ( binary ) entropy function .
note that for @xmath550@xmath551 correspondingly , a vanishing block error probability is obtained for any residual @xmath552 next , recall that rm codes @xmath553 of fixed order @xmath8 have code rate @xmath554 for this rate , ml decoding of optimal codes gives @xmath555 thus , we see that optimal codes give approximately the same residual order ( [ ord - ml ] ) as the former order ( [ ml ] ) derived in @xcite for rm codes @xmath553 .
in other words , rm codes of low rate @xmath544 can achieve nearly optimum performance for ml decoding .
by contrast , low - complexity algorithm @xmath201 has a substantially higher residual that has the order of @xmath556 .
this performance gap shows that further advances are needed for the algorithm @xmath70 the main problem here is whether possible improvements can be coupled with low complexity order of @xmath557 the performance gap becomes even more noticeable , if a binary symmetric channel is considered as a `` hard - decision '' image of an awgn channel .
indeed , let the input symbols @xmath558 be transmitted over a channel with the additive white gaussian noise @xmath559 for code sequences of rate @xmath560 we wish to obtain a vanishing block error probability when @xmath561 in this case , the transmitted symbols @xmath558 are interchanged with very high crossover probability @xmath562 which gives residual @xmath563 thus , @xmath564 serves ( up to a small factor of @xmath565 as a measure of noise power @xmath566 in particular , ml decoding operates at the above residual @xmath567 from ( [ ord - ml ] ) and can withstand noise power @xmath568 of order up to @xmath569 by contrast , algorithm @xmath201 can successfully operate only when noise power @xmath568 has the lower order of @xmath570 similarly , algorithm @xmath69 is efficient when @xmath568 is further reduced to the order of @xmath571 . therefore for long rm codes , algorithm @xmath201 can increase @xmath571 times the noise power that can be sustained using the algorithm @xmath69 or majority decoding .
however , performance of @xmath201 also degrades for longer blocks when compared to optimum decoding , though this effect is slower in @xmath201 than in other low - complexity algorithms known for rm codes . for moderate lengths ,
this relative degradation is less pronounced , and algorithm @xmath201 achieves better performance . in particular , some simulation results are presented in fig .
[ fig2 ] to fig .
[ fig4 ] for rm codes @xmath572 , @xmath573 , and @xmath574 , respectively . on the horizontal axis ,
we plot both input parameters - the signal - to noise ratio @xmath575 of an awgn channel and the crossover error probability @xmath576 of the corresponding binary channel
. the output code word error rates ( wer ) of algorithms @xmath69 and @xmath201 represent the first two ( rightmost ) curves@xmath262 decoding is performed on a binary channel , without using any soft - decision information .
these simulation results show that @xmath201 gains about @xmath577 db over @xmath69 on the code @xmath573 and about 0.5 db on the code @xmath574 even for high wer . a subsequent improvement can be obtained if we consider _ soft - decision _ decoding@xmath349 which recursively recalculates the _ posterior probabilities _ of the new variables obtained in both steps 1 and 2 .
these modifications of algorithms @xmath69 and @xmath201 - called below @xmath578 and @xmath579 - are designed along these lines in @xcite .
the simulation results for the algorithm @xmath580 are also presented in fig .
[ fig2 ] to fig .
[ fig4 ] , where these results are given by the third curve .
this extra gain can be further increased if a few most plausible code candidates are recursively retrieved and updated in all intermediate steps .
we note that the list decoding algorithms have been of substantial interest not only in the area of error control but also in the learning theory . for _ long _ _ biorthogonal codes _
@xmath581 the pioneering randomized algorithm is presented in @xcite . for @xmath401 and
@xmath582 @xmath583 this algorithm outputs a complete list of codewords located within the distance @xmath584 from any received vector , while taking only a polynomial time @xmath585 to complete this task with high probability @xmath586 substantial further advances are obtained for some low - rate @xmath587-ary rm codes in @xcite and the papers cited therein . for binary rm codes of _ any order _ @xmath32 we mention three different soft decision list decoding techniques , all of which reduce the output wer at the expense of higher complexity .
the algorithm of @xcite and @xcite reevaluates the most probable information subblocks on a _
single run_. for eachpath @xmath232 the decoder - called below @xmath588 - updates the list of @xmath589 most probable information subblocks @xmath590 obtained on the previous paths @xmath591 this algorithm has overall complexity of order @xmath592 the technique of @xcite proceeds recursively at any intermediate node , by choosing @xmath589 codewords closest to the input vector processed at this node .
these lists are updated in _
multiple recursive runs .
_ finally , the third novel technique @xcite executes sequential decoding using the main stack , but also utilizes the complementary stack in this process .
the idea here is to lower - bound the minimum distance between the received vector and the closest codewords that will be obtained in the _ future steps_. computer simulations show that the algorithm of @xcite achieves the best complexity - performance trade - off known to date for rm codes of moderate lengths 128 to 512 . in fig .
[ fig2 ] to fig .
[ fig4 ] , this algorithm @xmath588 is represented by the fourth curve , which shows a gain of about 2 db over @xmath593 .
here we take @xmath594 in fig .
[ fig2 ] and @xmath595 in fig .
[ fig3 ] and [ fig4 ] .
finally , complexity estimates ( given by the overall number of floating point operations ) are presented for all three codes in table 1 .
[ c]|c|c|c|c|c|code & @xmath596 & @xmath597 & @xmath598 & @xmath599 + @xmath600 & 857 & 1264 & 6778 & 29602 , @xmath601 + @xmath602 & 1753 & 2800 & 16052 & 220285 , @xmath595 + @xmath603 & 2313 & 2944 & 12874 & 351657 , @xmath595 + recall also that different information bits - even those retrieved in consecutive steps - become much better protected as recursion progresses .
this allows one to improve code performance by considering a subcode of the original code , obtained after a few least protected information bits are removed .
the corresponding simulation results can be found in @xcite and @xcite . summarizing this discussion ,
we outline a few important open problems .
recall that the above algorithms @xmath69 and @xmath201 use two simple recalculation rules@xmath607 therefore the first important issue is to define whether any asymptotic gain can be obtained : the second important problem is to obtain tight bounds on the decoding error probability in addition to the decoding threshold derived above .
this is an open problem even for the simplest recalculations ( [ rule1 ] ) utilized in this paper , let alone other rules , such as ( [ boss ] ) or those outlined in @xcite . from the practical perspective ,
recursive algorithms show substantial promise at the moderate lengths up to 256 , on which they efficiently operate at signal - to - noise ratios below 3 db .
it is also interesting to extend these algorithms for low - rate subcodes of rm codes , such as the duals of the bch codes and other sequences with good auto - correlation . in summary ,
the main result of the paper is a new probabilistic technique that allows one to derive exact asymptotic thresholds of recursive algorithms .
firstly , we disintegrate decoding process into a sequence of recursive steps .
secondly , these dependent steps are estimated by independent events , which occur when all preceding decodings are correct .
lastly , we develop a second - order analysis that defines a few weakest paths over the whole sequence of consecutive steps .
i. dumer and k. shabunov , `` recursive constructions and their maximum likelihood decoding , '' _ _ proc .
38__@xmath610 _ allerton conf . on commun .
, cont . , and comp . , _ ,
monticello , il , usa , 2000 , pp .
71 - 80 .
r. lucas , m. bossert , and a. dammann , `` improved soft - decision decoding of reed - muller codes as generalized multiple concatenated codes , '' _ proc .
itg conf . on source and channel coding , _
aahen , germany , 1998 , pp .
137 - 141 .
n. stolte and u. sorger , `` soft - decision stack decoding of binary reed - muller codes with `` look - ahead '' technique , '' _ _ proc .
7__@xmath610 _ int .
workshop `` algebr . and comb .
coding theory '' _ , bansko , bulgaria , 2000 , pp .
293 - 298 . | recursive decoding techniques are considered for reed - muller ( rm ) codes of growing length @xmath0 and fixed order @xmath1 an algorithm is designed that has complexity of order @xmath2 and corrects most error patterns of weight up to @xmath3 given that @xmath4 exceeds @xmath5 this improves the asymptotic bounds known for decoding rm codes with nonexponential complexity . to evaluate decoding capability ,
we develop a probabilistic technique that disintegrates decoding into a sequence of recursive steps .
although dependent , subsequent outputs can be tightly evaluated under the assumption that all preceding decodings are correct . in turn , this allows us to employ the second - order analysis and find the error weights for which the decoding error probability vanishes on the entire sequence of decoding steps as the code length @xmath0 grows .
* keywords * - decoding threshold , plotkin construction , recursive decoding , reed - muller codes . |
nonlinear evolution equations play an important role in the modern mathematics and physics .
there are several `` universal '' equations that are widely physical applicable and have remarkable mathematical properties ( see e.g. , refs .
@xcite ) . in this set of equations
one can include the korteweg de vries equation , the burgers equation , the nonlinear shrodinger equation .
such `` universal '' equations are very often obtained with the help of some asymptotic approach , for example , with the help of the reductive perturbation method ( see e.g. , refs .
if we consider the first order approximation in the asymptotic expansion we can obtain famous nonlinear evolution equations such as the burgers and the korteweg de vries equations . on the other hand , taking into account high order corrections in the reductive perturbation method one can obtain generalizations of the above mentioned equations .
these equations can be considered as `` universal '' equations with high order corrections . in this work
we study the burgers equation with high order corrections : the perturbed burgers equation .
this equation was obtained for the description of viscous gas dynamics and nonlinear acoustic waves in refs .
@xcite . in refs .
@xcite it was shown that the perturbed burgers equation can be used for the description of nonlinear waves in a liquid with gas bubbles .
this equation has the form ( see refs .
@xcite ) : @xmath0 where @xmath1 is non dimensional time , @xmath2 is a non dimensional cartesian coordinate , @xmath3 is non dimensional perturbation of gas liquid mixture density , @xmath4 , @xmath5 , @xmath6 , @xmath7 are non dimensional physical parameters , @xmath8 , @xmath9 are arbitrary parameters introduced by the near identity transformations ( see .
@xcite ) . let us note that in fact eq . is a two parametric family of equations parameterized by @xmath8 and @xmath9 .
it is known that at the derivation of the perturbed burgers equation appears an obstacle to the integrability ( see refs .
the perturbed burgers equation is integrable under a certain condition on the physical parameters in contrast to the burgers equation . in the integrable case the perturbed burgers equation coincides with the sarmo
olver equation which was investigated ( see refs .
however , the general case of the perturbed burgers equation was not studied previously .
thus it is an interesting problem to perform an analytical and numerical study of the perturbed burgers equation without imposing any conditions on the physical parameters . below
we show that the perturbed burgers equation can be transformed to a canonical form with a single parameter .
this parameter can be considered as the bifurcation parameter , because there is only one value of this parameter corresponding to the integrable case of the perturbed burgers equation .
the aim of this work is to analytically and numerically study the perturbed burgers equation at all values of the bifurcation parameter . using the painlev approach for partial differential equations
we illustrate that the perturbed burgers equation is integrable only at a certain value of the bifurcation parameter .
we construct classical and nonclassical symmetries of the perturbed burgers equations .
we show that in the general case the perturbed burgers equation admits three classical lie groups . in the integrable case
the perturbed burgers equations admits an infinite dimensional lie group .
both regular and singular cases of nonclassical symmetries are considered .
the regular nonclassical symmetries coincide with the classical lie symmetries . in the singular case
we find symmetries which do not correspond to classical lie symmetries .
we construct symmetry reductions of the perturbed burgers equation and obtain their exact solutions .
we numerically study the stability of a certain periodic solution of the integrable case of the perturbed burgers equation under the variations of the bifurcation parameter .
we show that this solution is stable under variations of the bifurcation parameter .
we also find that initial solitary waves governed by the perturbed burgers equation evolve to a weak shock wave with oscillating structure .
we construct analytical solution of the integrable case of the perturbed burgers equation that is similar to this weak shock wave . to the best of our knowledge
classical and nonclassical symmetries of the perturbed burgers equation were not obtained previously .
we believe that all exact solutions constructed in this work are new .
the stability of the exact solutions of the perturbed burgers equation under the variations of the bifurcation parameter is studied for the first time as well .
this work is organized as follows . in sec .
[ sec:2 ] we transform the perturbed burgers equation to the canonical form .
the painleve analysis of this equation is carried out as well .
symmetries of the perturbed burgers equation are studied in sec .
[ sec:3 ] .
we investigate both classical and nonclassical symmetries of this equation .
symmetry reductions of the perturbed burgers equation are obtained in sec .
[ sec:4 ] .
we construct exact solutions for the symmetry reductions of the perturbed burgers equation in sec .
[ sec:4 ] as well .
[ sec:5 ] is devoted to the numerical investigation of the perturbed burgers equation .
we give final remarks and conclusions in sec .
[ sec:6 ] .
in this work we will use the perturbed burgers equation in the form obtained in refs .
@xcite . eq . can be simplified with the help of the scaling transformations and using arbitrariness of @xmath8 , @xmath9 .
indeed , let @xmath8 and @xmath9 have the form @xmath10 substituting the transformations @xmath11 into equation we have ( primes are omitted ) @xmath12 where @xmath13 further we study eq . .
let us investigate whether eq . possesses the painelve property for partial differential equations . in accordance with the algorithm by weiss , tabor and carnevale@xcite we present a solution of eq .
in the form @xmath14 the necessary condition for eq . to possess
the painelve property is that expansion contains three arbitrary functions .
substituting into the leading order terms of eq .
we get @xmath15 thus eq . admits two expansions in the form . substituting the expression @xmath16 into the leading order terms of eq .
we find the fuchs indices @xmath17 at this step the necessary condition for eq . to possess the painlev property is that all fuchs indices are real integers .
consequently @xmath18 and @xmath19 have to be real integers . and
the expression @xmath20 has to be an integer as well .
expression in the right hand side of is an integer only at the following values of @xmath21 @xmath22 we recall that both of indices @xmath18 , @xmath19 must be integers .
this is true only for three values of @xmath21 : @xmath23 , @xmath24 , @xmath25 . at @xmath26 from we obtain @xmath27 in this case eq .
does not possess the painlev property since the coefficient @xmath28 is not arbitrary function .
in the case of @xmath29 fuchs indices have the form @xmath30 we see that there are multiple fuchs indices . it is indicated that expansion contains logarithmic terms .
. does not possess the painlev property at @xmath29 . at @xmath31 fuchs indices
have the form @xmath32 so in this case the necessary condition of the painlev test is held .
indeed , at @xmath31 eq . is the sharmo
olver equation ( see refs .
@xcite ) which is linearizable by the cole hopf transformation .
the following cases of parameter @xmath21 are also interesting : @xmath33 . for @xmath34
the fuchs indices in the second expansion are integers . in the case of @xmath35 equation
admits one expansion of form with fuchs indices that are integers .
however , this expansion does not correspond to the general solution . in our opinion
the presence of the fuchs indices that are integers indicates that eq
. can admit meromorphic solutions .
in this section we have found that eq .
passes the painleve test only in one case : @xmath36 . at any other value of @xmath21 the equation
does not have the painleve property .
this fact indicates that eq .
is integrable only at @xmath36 and this is really the case .
let us study symmetries and symmetry reductions of eq . .
here we repeat eq . for the convenience @xmath37 below we consider the application of classical and nonclassical methods for finding symmetries of .
first we apply the classical lie method ( see e.g. , refs .
@xcite ) to eq . .
in accordance with this method eq .
is invariant under transformations @xmath38 if the determining equations are satisfied on the solutions of @xmath39 here @xmath40 is the third prolongation of the infinitesimal generator @xmath41 that has the form @xmath42 substituting expressions for @xmath43 into we obtain an overdetermined system of partial differential equations for @xmath44 , @xmath45 , @xmath46 .
this system of equations has the general solution @xmath47 where @xmath48 , @xmath49 , @xmath50 are arbitrary constants .
consequently eq .
is invariant under three one parameter lie groups with the following generators @xmath51 the first generator corresponds to space translation invariance , the second generator corresponds to time translation invariance and the third one corresponds to scaling invariance .
thus there are two classical symmetry reductions of eq . .
the first one is the traveling wave reduction corresponding to @xmath52 and the second one is the self
similar reduction corresponding to @xmath53 . in the next section
we will study these symmetry reductions .
let us note that we can not indicate the integrable case of using symmetry analysis presented above .
one can find the integrable case of studying classical symmetries for the potential form of this equation . indeed , using the variable @xmath54 in we find that at @xmath36 this equation admits the generator @xmath55 where @xmath56 is any solution of the third order linear evolution equation @xmath57 thus one can find the integrable case of using the classical lie symmetry analysis for the potential form of .
it is known that there exist symmetries which do not coincide with the classical symmetries .
these symmetries are called nonclassical symmetries @xcite .
the method for finding nonclassical symmetries for partial differential equations was proposed by bluman and cole@xcite .
this method was applied for finding symmetry reductions and exact solutions of various partial differential equations ( see e.g. , refs .
@xcite ) . here
we apply the method by bluman and cole to study nonclassical symmetries of eq . .
according to the method by bluman and cole@xcite we considered the additional auxiliary equation @xmath58 eq .
is the invariant surface condition associated with operator .
then we look for classical symmetries admitted by eqs . and simultaneously .
one can show that the invariance condition for eq .
satisfies without imposing any conditions on @xmath44 , @xmath45 and @xmath46 ( see refs .
@xcite ) . let us note that if @xmath59 is a nonclassical symmetry generator then @xmath60 is a nonclassical symmetry generator as well for any function @xmath61 ( see refs .
@xcite ) . thus further we have to consider two cases of nonclassical operators @xmath59 .
the first one is the case of @xmath62 , where without loss of generality we can assume that @xmath63 .
the other one is the case of @xmath64 , where without loss of generality we can assume that @xmath65 .
the case of @xmath62 is called the regular case where we have an overdetermined system of equations for infinitesimals . in the singular case @xmath64
we obtain a single partial differential equation for infinitesimal @xmath46 .
as it was shown by kunzinger and popovych@xcite every solution of this equation generates family of solutions of a considered equation .
let us study the case of @xmath62 ( we assume that @xmath63 ) . applying the classical lie method to and and using and its differential consequences for eliminating derivatives involving time and using for eliminating higher order derivative with respect to @xmath2 we obtain the system of equations for @xmath44 and @xmath46 : @xmath66 solving this system of equations we find @xmath67 where @xmath68 are arbitrary constants .
similarity variables connected with have the form @xmath69 we can see that reduction corresponding to coincides with the reduction generated by the linear combination of classical generators @xmath70 , @xmath71 and @xmath53 . let us consider the case of @xmath64 . in this case without loss of generality
we assume @xmath65 . applying the classical lie approach and excluding space derivatives using and its differential consequence and we obtain the determining equation for infinitesimal @xmath46 : @xmath72\eta^{2}+\vspace{0.1cm}\\+ 3\left[\eta_{x}(\eta_{uu}-3)+(\eta_{u}-2u)\eta_{xu}+3\eta_{xxu}\right]\eta+\vspace{0.1cm}\\ + 3\bar{\alpha } u^{2 } \eta_{x}+3\eta_{x}\eta_{xu}-3 u\eta_{xx}+\eta_{t}+\eta_{xxx}=0 .
\label{eq : bc_9 } \end{gathered}\ ] ] we can not find the general solution of eq . .
kunzinger and popovych@xcite showed that the problem for construction all solutions of is completely equivalent to the problem of the construction of all one - parametric solutions families of eq . .
however every particular solution of generates nonclassical symmetry reduction of . using this fact we can find exact solutions families of that are not invariant under classical lie groups admitted by .
.nonclassical infinitesimals admitted by [ cols="^,^,^",options="header " , ] we seek for solution of assuming that @xmath46 is polynomial in @xmath3 . in this way we find infinitesimals which are presented in table [
t : tab1 ] .
parameters @xmath73 , @xmath74 from table [ t : tab1 ] are arbitrary .
infinitesimal generator @xmath75 corresponds to the rational solution of eq . at @xmath76 . using generator @xmath77 we can obtain exact solution of eq .
expressed via the airy functions .
below we show that this solution is the generalization of the self similar solution of eq . .
the third generator @xmath78 corresponds to rational and solitary wave solutions of eq . .
using generators @xmath79 , @xmath80 we can obtain rational , solitary wave and airy function solutions of eq . at @xmath81 .
we believe that at least generators @xmath75 ,
@xmath77 , @xmath80 correspond to the solutions of eq . that can not be found with the help of the classical lie approach . below
we construct exact solutions of eq .
corresponding to generators @xmath75 , @xmath77 , @xmath80 .
also one can find additional infinitesimal generator at @xmath36 : @xmath82 where function @xmath83 satisfies eq . at @xmath36 .
using this infinitesimal generator we can construct auto
backlund transformations for the integrable case of . indeed , because @xmath83 satisfies eq .
at @xmath36 and using the invariant surface condition at @xmath84 we obtain @xmath85_{x } , \label{eq : auto_baclund } \end{gathered}\ ] ] where @xmath3 and @xmath83 both satisfies eq . at @xmath36 . using these auto
backlund transformations we can generate infinite series of solutions of the sharma tasso
olver equation .
let us note that auto
backlund transformations were found in ref .
@xcite . here
we recover these transformations using the symmetry approach .
it is worth to note that for the first time the nonclassical symmetries were used for constructing auto
backlund transformations by nucci@xcite . in this section
we have found the classical symmetry generators admitted by eq . .
further we use these generators for constructing traveling wave and self similar solutions of eq . .
we also obtain several nonclassical generators admitted by eq . .
below we use some of these generators to construct solutions of eq . which are not invariant under classical lie groups .
in this section we study symmetry reductions of eq . .
we construct several families of exact solutions of this equation .
both classical and nonclassical symmetry reductions are considered .
let us consider reduction of eq .
corresponding to the linear combination of generators @xmath70 and @xmath71 . using the traveling wave variable @xmath86 and integrating with respect to @xmath87 from eq .
we get @xmath88 where @xmath89 is an integration constant . introducing new variable @xmath90
we obtain the abel equation of the second kind @xmath91 eq .
can be transformed to the canonical form with the help of the substitution @xmath92 : @xmath93 where @xmath94 , @xmath95 .
list of the abel equations which can be solved is presented in handbook by zaitsev and polyanin@xcite .
general solution of at @xmath96 can be found only in the case of @xmath36 ( see ref .
@xcite ) . ,
height=264 ] for constructing particular solutions of we use the simplest equation method proposed in ref .
we find that eq . at @xmath97
has the following solution @xmath98 one can see that solution can not be used at @xmath81 . in this case
admits stationary solution ( @xmath99 ) at @xmath97 @xmath100 where @xmath101 is an arbitrary constant .
let us note that solutions , can be obtained by the direct integration of .
the plots of solution for several values of @xmath21 are presented in fig.[fig1 ] .
we can see that solution is a weak shock wave .
let us consider the self similar reduction of eq . .
using the variables @xmath102 from eq .
at @xmath103 we get @xmath104 integrating once we have @xmath105 where @xmath106 is an integration constant .
it is known that general solution of eq .
can be obtained in the case of @xmath36 ( see ref .
@xcite ) . however at the arbitrary value of @xmath21 eq .
admits particular solution which is expressed via the airy functions .
indeed , if the following relation holds @xmath107 eq . has the solution in the form @xmath108 where @xmath109 and @xmath110 denotes the derivative with respect to the function argument , @xmath111 is an arbitrary constant . and @xmath112.,height=264 ] in the case of @xmath81 we can not find self similar solution of eq .
however , one can obtain stationary solution of eq . at @xmath81 that is expressed via the airy functions .
this solution can be obtained using generator @xmath79 or by the direct integration of the stationary case of eq . at @xmath81 .
let us note that using the cole
hopf transformation one can obtain more general self
similar solution of eq . at @xmath36 .
this solution has the form @xmath113 where @xmath114 and @xmath115 are arbitrary constants . using and
we can find dependence of on @xmath2 and @xmath1 .
the plot of this solution at various values of @xmath1 is presented in fig.[fig1a ] . from fig.[fig1a ] we see that this solution is a weak shock wave with oscillations on a wave crest . when time @xmath1 increases the number of oscillations and the amplitude of waves decrease and the width of wave front increases .
waves of this type were observed experimentally in a liquid with gas bubbles .
let us study nonclassical symmetry reductions of eq . .
first we consider reduction corresponding to generator @xmath75 from table [ t : tab1 ] .
using @xmath75 we find similarity variables for eq . at @xmath116 @xmath117 where function h(t ) satisfies the equation @xmath118 solving this equation and using we obtain the rational solution of eq . at @xmath116 @xmath119 where @xmath120 is an arbitrary constant .
we see that we obtain the exact solution of eq .
that is not invariant under classical lie symmetries of this equation .
let us consider the generator @xmath80 from table [ t : tab1 ] .
in this case symmetry reduction has the form @xmath121 where @xmath122 satisfies the equation @xmath123 using , we find the following solution of eq . at @xmath81 @xmath124 where @xmath125 is an integration constant .
symmetry generators @xmath78 and @xmath79 correspond to the traveling wave and stationary solutions that were obtained above .
it is interesting to consider generator @xmath77 , generator @xmath78 at @xmath126 and generator @xmath79 at @xmath127 .
generator @xmath77 corresponds to the exact solution of eq .
that is expressed via the airy functions .
this solution does not coincide with self similar solution and has the form @xmath128^{2}}\right)^{1/3}\frac{c_{9}\mbox{ai}^{'}(\theta)+\mbox{bi}^{'}(\theta ) } { c_{9}\mbox{ai}(\theta)+\mbox{bi}(\theta)},\vspace{0.1cm}\\ \theta=\left(x+c_{7}\right)\left(\frac{3t(-3+\sqrt{9 - 8\bar{\alpha}})+2c_{8 } } { \sqrt{c_{8}(3+\sqrt{9 - 8\bar{\alpha}})-12\bar{\alpha } t}}\right)^{-\frac{2}{3 } } , \label{eq : nc_reduction_9 } \end{gathered}\ ] ] where @xmath110 denotes the derivative with respect to function argument , @xmath129 is an arbitrary constant .
solution is the generalization of self
similar solution and coincides with this solution at @xmath130 .
using generator @xmath78 at @xmath126 and generator @xmath79 at @xmath127 we can obtain the following rational solutions of eq .
@xmath131 where @xmath132 is an arbitrary constant .
thus we find one parametric families of stationary rational solutions of eq . . in this section
we have found several families of exact solutions of eq . .
two families of exact solutions correspond to the classical similarity reductions of eq . .
the other families of solutions correspond to the nonclassical similarity reductions of eq . .
we believe that all exact solutions found in this section are new .
as it was mentioned above the perturbed burgers equation is integrable only at one value of the bifurcation parameter ( @xmath36 ) .
so at this value of @xmath21 one can obtain a plenty of exact solutions .
there is an interesting question .
what happens with the solutions of the integrable case of eq . when the value of @xmath21 is being varied ?
in other words , it is interesting to study the stability of solutions for eq . with respect to perturbations of the parameter @xmath21 .
for this purpose we will use the numerical approach .
let us consider the boundary value problem for eq . with periodic boundary conditions and some initial condition .
we use the integrating factor with the fourth order runge
kutta approximation method for the numerical calculations ( see refs .
@xcite ) . we can present the perturbed burgers equation in the form @xmath133 where @xmath134 and @xmath135 .
the boundary conditions and the initial condition have the form @xmath136 where @xmath137 and @xmath138 is a period .
we discretize the spatial part of using the fourier transform . as a result
we obtain the system of ordinary differential equations @xmath139 where @xmath140 , @xmath141 and @xmath142 are the fourier transforms of @xmath3 , @xmath143 and @xmath144 respectively . applying the integrating factor in @xmath145 we have @xmath146 we solve this system of ordinary differential equations using the fourth - order runge kutta method .
.,height=226 ] to verify our numerical algorithm we use one of the simplest periodic exact solutions of eq . at @xmath36 .
this solution has the form @xmath147 where @xmath148 , @xmath149 and @xmath150 are arbitrary constants .
we calculate the average error of our numerical algorithm by the formula @xmath151 where @xmath144 is the number of approximation points on the @xmath2 axis . versus @xmath152,width=453 ]
we use solution at @xmath153 as the initial condition for problem .
we calculated the average error by formula for @xmath154 and @xmath155 .
total time of calculations was chosen equal to @xmath156 .
we obtain that the magnitude of the average error in both cases is very small and equals to @xmath157 and to @xmath158 respectively .
thus we can use @xmath154 in all other calculations .
we have also tested our numerical algorithm for the long time of calculations .
the dependence of average error on time @xmath1 is presented in fig.[fff ] .
let us investigate the stability of solution under the perturbations of parameter @xmath21 .
we again use this solution as the initial condition and perform calculations for @xmath159 .
the calculations are performed until numerical solution does not break up . in this way for each value of @xmath21
we can obtain the value of time @xmath160 at which the solution breaks up .
the values of @xmath21 versus @xmath152 is presented in fig .
[ fig3 ] . from fig.[fig3 ] we see that there is exist a critical value of parameter @xmath161 at which time @xmath152 becomes very small .
this critical value @xmath161 is located between @xmath162 and @xmath163 .
we performed analogous calculations for @xmath164 .
we have found that solution remains stable until @xmath165 for all @xmath164 , thus there is no critical value of @xmath21 on this interval .
.,width=529 ] in fig.[fig4 ] we demonstrate evolution of solution governed by eq . at @xmath166 .
we see that the solution remains close to the exact solution until time @xmath167 . on the next stage of evolution
the numerical solution is slightly deformed in comparison to the exact solution . at @xmath168
the numerical solution significantly changes the form .
however , it remains stable . from figs.[fig3],[fig4
] we see that periodic exact solution is stable under the variations of parameter @xmath21 .
thus under small variations of @xmath21 we can use solution for the description of waves governed by the nonintegrable case of eq . .
.,width=529 ] let us study the propagation of solitary waves governed by eq . at values of the parameter @xmath21 typical for a real liquid with gas bubbles .
we consider waves propagated in water with carbon dioxide bubbles with equilibrium bubbles radius equal to @xmath169 m . in this case
the parameter @xmath21 has value @xmath170 .
.,width=529 ] in fig.[f5 ] we demonstrate the evolution of initial solitary wave @xmath171 governed by eq . .
we see that the solitary wave is transformed to the weak shock wave with the oscillating tail after the wave crest .
let us consider as the initial condition the wave with compact support : @xmath172 \hfill \\ 0 , \quad \quad \quad \quad\,\,\,\ , x
\not\in [ 3\pi,12\pi ] \hfill \end{gathered } \right .
\label{eq : alpha}\ ] ] from fig.[f6 ] we see that this wave is transformed to a weak shock waves with the oscillating tail after the wave crest as well .
we perform analogous calculations for initial gaussian and rectangular pulses .
we obtained that these initial pulses evolve to the same weak shock wave . , @xmath173 , @xmath174 , @xmath175 .
, width=377 ] we can see that several types of solitary wave initial conditions evolve to the same asymptotic profile .
we found that solutions of eq .
have the same asymptotic behavior at large values of @xmath1 .
thus there is a stable solution of eq .
that describes dynamics of solitary waves at large values of @xmath1 .
let us note that the integrable case of eq
. admits analytical solution that is very similar to this asymptotic profile .
this solution has the form @xmath176 where @xmath177 and @xmath178 are arbitrary constants .
the plot of solution at @xmath174 is presented in fig.[f7 ] . from fig.[f7 ] we see that solution is very similar to numerical solutions obtained above .
the perturbed burgers equation was investigated .
we have shown this equation passes the painleve test only in the case of @xmath36 .
this case correspond to the integrable sharma
olver equation .
we have applied the classical lie method to the perturbed burgers equation .
three classical generators admitted by this equation have been found : translations in @xmath2 and @xmath1 and scaling in @xmath2 , @xmath1 and @xmath3 .
some nonclassical symmetries of the perturbed burgers equation have been obtained using the method by bluman and cole .
exact solutions of classical and nonclassical symmetry reductions have been constructed .
we have obtained traveling wave solutions , self similar solutions and three families of solutions that are invariant under nonclassical symmetries .
we believe that all found exact solutions are new .
the stability of exact solutions of the perturbed burgers equation was investigated numerically .
we have shown that periodic solution is stable under the variation of parameter @xmath21 .
we have seen that dynamics of waves described by nonintegrable case of eq .
remains close to the dynamics of waves described by integrable case of eq . at small variations of parameter @xmath21 .
it can be conjectured that other periodic solutions of the integrable case of eq .
are also stable under variations of @xmath21 .
we have found that the dynamics of waves governed by eq . at large values of time
is the same for various solitary wave initial conditions .
the initial disturbances evolve to the same weak shock wave with oscillations after wave crest .
so for large values of time there is a stable solution of eq . .
we have shown that integrable case of eq .
admits analytical solution in the from of the weak shock wave with oscillations after wave crest .
this research was partially supported by grant for scientific schools 2296.2014.1 .
, by grant for the state support of young russian scientists 3694.2014.1 and by russian science foundation 14 - 11 - 01 - 0258 . | the perturbed burgers equation is considered .
this equation is used for the description of one dimensional viscous gas dynamics , nonlinear waves in a liquid with gas bubbles and nonlinear acoustic waves .
the integrability of this equation is investigated using the painleve approach .
the condition for parameters for the integrability of the perturbed burgers equation is established .
new classical and nonclassical symmetries admitted by this equation are found .
all corresponding symmetry reductions are obtained .
new exact solutions of these reductions are constructed .
they are expressed via trigonometric and airy functions .
stability of the exact solutions of the perturbed burgers equation is investigated numerically . |
the title of this short review could well be `` lepton flavour and cp violations without neutrinos for neutrino physicists '' since the aim here is to emphasize the impact of the present and near future experiments on charged lepton transitions on ( _ a _ ) the flavour and cp pattern of the new lepton - like states in theories beyond the sm and ( _ b _ ) the neutrino mass models .
indeed , fcnc and cpv are generic problems for theories that postulate new states at the tev scales which is typically the case in string inspired solutions to the hierarchy problem : squark and sleptons in supersymmetric models and kaluza - klein states in models with large compact dimensions .
basically , the problems arise because particles get mass from different , generically unrelated , mechanisms : higgs couplings for the usual fermions , supersymmetry breaking or compactification for the new heavy states .
any misalignment in flavour space or in cp phases between the resulting mass matrices yields fcnc or cpv effective operators from radiative corrections involving the heavy states .
the present experiments already put either strong limits on these misalignments or lower bounds above the tev for the new state masses .
these constraints open a new framework to investigate both the flavour problem and the structure of the new theories , in particular , in the lepton sector .
let us first summarize the basic facts .
- neutrino oscillations require large lepton flavour ( as defined by the charged leptons ) mixing in the effective neutrino mass matrix , but this does not necessarily imply large mixing in the charged sector . - leptogenesis is viable within the seesaw model for neutrino masses if the yukawa couplings of the neutrinos to the higgs ( _ i.e. , _ the dirac masses ) have sizeable cp phases . - in the somewhat minimal framework where only the seesaw mechanism to generate neutrino masses is added to the standard model , the lfv&cpv effects in the charged lepton physics are extremely tiny because of the analogue of the gim mechanism : they depend on factors of @xmath2 . - therefore , if observed in the current or planned experiments , lfv in charged decays ( @xmath3 ) or cpv electric dipole moments ( @xmath4 ) would be signals of new physics beyond the sm and seesaw models around the tev region ( _ e.g. , _ supersymmetry ) .
- conversely , the present and future bounds on these lfv&cpv transitions constrain new physics around the tev region ( _ e.g. , _ supersymmetry ) and require a lfv&cpv inhibition mechanism in the corresponding theories . -
some of these experiments provide already relevant constraints on radiative corrections from new lfv&cpv couplings in theories at unattainable scales : gut s , seesaw , flavour models ... we shall emphasize here the last point to demonstrate the importance of these experiments . indeed
they provide already restrictions on the masses and couplings of very heavy states in seesaw and/or gut models from their virtual contributions in radiative corrections to the effective low energy parameters in any new theory in the tev region ( _ e.g. , _ slepton masses in supersymmetry ) .
future experiments will make these constraints even more impressive .
the phenomenological information gathered from lfv&cpv in the charged sector is quite complementary to those provided by measurements of neutrino oscillations or from the assumption of leptogenesis for the generation of the baryon asymmetry in the universe .
this has been discussed in several papers and reviews at a more technical level @xcite and , here , we shall rather illustrate this aspect by comparing the explicit predictions of simple basic models with the available set of experimental data .
the variety of these data are reviewed by m. aoki in these proceedings @xcite , which we refer to for more detailed description of different experiments and prospects and for the relevant bibliography .
it is enough for our discussion to concentrate on the more relevant ones and to just keep the orders of magnitude of the experimental bounds , as shown in table [ table:1 ] .
notice that electric dipole moments are displayed in units of fm rather than cm - this is kind of more natural in comparing with the anomalous magnetic moments . [ cols="<,<,<,<,<",options="header " , ] + the value of the new mass scale @xmath5 can be as low as o(@xmath6gev ) without violating the limits on @xmath7 . on the contrary , bounds on @xmath1 already require a suppression of two orders of magnitude in the cp violating effective coupling if @xmath8 o(tev ) .
it is a well known fact that new physics is expected when experiments will go beyond the present tev frontier in order to explain the success of the sm as a very predictive effective theory . even on general grounds ,
the experimental data on lfv&cpv are already providing relevant restrictions on the flavour pattern of this new physics .
supersymmetry is one of the best candidates for the new physics framework and it has many sources of lfv&cpv , in particular , in the slepton mass matrices .
it is worth looking in more detail for the experimental restrictions in the case of supersymmetry since we expect the sparticle masses to remain below the tev scale to avoid excessive fine - tuning in the model .
let us consider msugra models where , by assumption , the slepton masses have no flavour structure and no relevant cp violating phases at the tree - level .
we denote @xmath9 the masses of the scalar leptons associated to the @xmath10 doublets and the @xmath11 singlet respectively and we skip the so - called @xmath12terms for simplicity . in order to estimate the allowed deviations in the alignment in flavour space between @xmath9 and the charged lepton mass matrix , @xmath13
, one usually defines the ratios @xmath14 with @xmath15 and determine the limits on these misalignment ratios as a function of the sfermion and slepton masses from the contributions they induce in lfv decays , and in cpv e.d.m .
if they are complex @xcite .
two examples of these limits are shown in figures [ fig:1 ] and [ fig:2 ] @xcite in terms of three most relevant parameters : the slepton mass @xmath16 , the gaugino ( bino ) mass @xmath17 and the higgs v.e.v .
angle , @xmath18 .
( -50,-5)@xmath19 ( gev ) ( -220,227)@xmath20 ( gev ) the plots show the allowed flavour misalignment @xmath21 for sparticle masses below @xmath22tev to be already o(@xmath23 ) it will become 30 times smaller with the planned experiments for the case of @xmath24 , actually a stronger limit than the generic order of magnitude in table [ table:2 ] ( instead the limits on @xmath25 are much more model dependent due to destructive interference between different contributions ! ) .
this is the lepton counterpart of the supersymmetric flavour problem : the supersymmetric scalar mass generation must be endowed with such a small misalignment between slepton and lepton masses .
however , since the orders of magnitude are close to those expected in radiative corrections , the alignment must be preserved by quantum contributions .
this fact will be exploited in the next section .
( -50,3)@xmath26 ( gev ) ( -227,235)@xmath20 ( gev ) the limits on the phases in slepton mass matrices obtained from the experimental bounds on @xmath1 are even stronger as seen in figure [ fig:2 ] for the product @xmath27 . indeed , for large @xmath18
their imaginary parts are at most of o(@xmath28 ) , to go down to @xmath29 in the next future !
also notice ( _ cf .
_ table [ table:2 ] ) that the future measurement of @xmath30 will provide limits comparable with the ones presently extracted from @xmath1 .
it is a remarkable fact that the e.d.m .
experiments are reaching the accuracy needed to put constraints on new physics around the tev scale at the level of radiative corrections .
we now turn to discuss how this fact may tell us of properties of theories at a scale much above the tev region , such as the seesaw models and gut s .
very heavy states that couple to sleptons leave their traces in the slepton mass matrices through their radiative corrections until they decouple from the effective supersymmetric theory . thus , in the ( type i ) seesaw model , the heavy singlet neutrinos are given very large masses , @xmath31 , and couple to the higgs and the light leptons and their sleptons through a yukawa matrix @xmath32 .
correspondingly , they produce a correction to the lfv part of @xmath33 with a misalignment @xmath34 , where the planck mass @xmath35 is the cut - off for effective supergravity .
this misalignment is bounded by the lfv decays @xcite , but only the bound on @xmath21 is really relevant .
the phases generated only by the seesaw radiative loops could produce lepton e.d.m . at observable rates only if the @xmath31 eigenvalues are strongly hierarchical @xcite .
on the other hand , in the simplest versions of su(5 ) gut s , the colour triplet partners of the higgs bosons , must get large masses , @xmath36 , through the doublet - triplet mechanism , while they couple to the right - handed leptons through a yukawa matrix which is approximately @xmath37 , the usual couplings of the up quarks .
correspondingly , radiative corrections are generated to @xmath38 .
they induce and are constrained by lfv decays as well @xcite .
what about cpv corrections to the slepton masses from these heavy neutrinos and colour triplets ?
it has been recently shown @xcite that with su(5 ) gut @xmath39 seesaw one obtains interesting constraints on @xmath32 matrix elements from the @xmath1 experiments , this is shown in figure [ fig:3 ] @xcite for @xmath40 , where @xmath41 is the phase in the quark ckm matrix .
remember that these limits will be improved by two orders of magnitude . another way to see the relevance of the restrictions from @xmath1 on new physics is to show that they are competitive with the well - known constraints on the colour triplets masses from proton life - time , @xmath42 @xcite .
let us take for definiteness @xmath43 and `` typical '' sparticle masses , @xmath44 and , _
e.g. _ , @xmath45 , with msugra mass conditions and let us calculate @xmath1 and @xmath42 in the simplest models ( which are not meant to be more theoretically justified , see _
@xcite ) : ( i ) su(5 ) with seesaw and the simplifying assumption , @xmath46 ; ( ii ) so(10 ) with two colour triplets and @xmath47 ; ( iii ) so(10 ) with pseudo - dirac like colour triplets , @xmath48 with @xmath49 .
the results are summarized in table [ table:3 ] .
@lllll & @xmath50)/ ( @xmath51 & @xmath52 + & @xmath53 & @xmath54 + ( i ) su(5 ) w/ @xmath55 & @xmath56 & @xmath57 + ( ii ) so(10 ) w/ @xmath58 & @xmath56 & @xmath59 + ( iii ) so(10 ) w/ @xmath60 & @xmath61 & @xmath62 + + while models i and ii predict @xmath42 much below the experimental value , the pseudo - dirac trick introduces large negative interference between the two massive colour triplets increasing @xmath42 by almost two orders of magnitude ( depending on the various yukawa phases ; note that the quark ckm phase in @xmath37 alone is enough unless it is compensated by the unknown phases ) .
but the supersymmetric so(10 ) models violate the limits on @xmath1 , at least in the simpler versions .
while @xmath42 could be reduced by interference , the different contributions add up in the wave function renormalisation that leads to @xmath1 .
in order to explicitly see the impact of the @xmath24 and @xmath1 experiments on flavour models to explain the fermion masses , including the neutrinos , let us consider the froggatt - nielsen simplest model with an @xmath63 flavour group @xcite .
it is assumed to be broken by field whose v.e.v .
defines a small number @xmath64 with respect to the cut - off scale of the model .
charges are associated to each lepton field as follows : @xmath65 to the doublets @xmath66 , @xmath67 to @xmath68 , @xmath69 to @xmath70 , with @xmath71 as the flavour index , @xmath72 to the two higgses . in this way one obtains , after the @xmath63 breaking , effective yukawa couplings endowed with some kind of hierarchy in the eigenvalues and the mixings in terms of powers of @xmath64 defined by the flavour charges . ordering the heavy right - handed neutrinos by @xmath73
, the present neutrino oscillation data ( mass differences and mixing angles for atmospheric and solar neutrinos ) basically fix the charges : @xmath74 , together with @xmath75 , close to the cabibbo angle that plays a similar role in the construction of quark mass models .
the masses @xmath76 remain arbitrary , but leptogenesis is possible only if @xmath77 . because the @xmath65 are now given , the present experimental limit on @xmath24 requires @xmath78 , for @xmath79 .
the charges @xmath69 are now more or less fixed so that the orders of magnitude of all other lfv&cpv predictions , in particular @xmath1 , come out much below the present experiments . with the expected improvement by a factor 30 in b.r.@xmath80
, one would get @xmath81 and this very popular model is close to be excluded .
what is nice in this exercise is that the constraints from different observables apply to different parameters and complement each other in an explicit way .
the lfv&cpv experiments , mainly @xmath24 and @xmath1 but hopefully also @xmath30 in the near future are a selective tool for new physics model building .
they provide indirect tests which are already sensitive to physics in the tev region at the level of accuracy of the radiative corrections .
they are pure signals of new phenomena around the tev scale since the sm contributions are definitely negligible .
it is quite remarkable that they also give effective tests of the lfv&cpv in seesaw couplings and in grand - unified theories . in particular , the limits on @xmath1
nicely complement the proton decay bounds in selecting o(10 ) models !
the devised improvements in these experiments are important and theoretically welcome .
the current direct measurement of @xmath30 , improving the present indirect bound , motivates a simple question @xcite : to what extent can the naive relation @xmath82 be violated ?
needless to say , the fact that they are getting into the level of accuracy corresponding to where we do expect the new physics to show up means they are on the way to discover lfv&cpv in charged lepton transitions .
9 s. davidson and a. ibarra , jhep 0109 ( 2001 ) 013 ; i. masina , in proceedings of susy02 , desy , hamburg , 2002 .
m. aoki , these proceedings .
f. gabbiani , e. gabrielli , a. masiero and l. silvestrini , nucl .
b 477 ( 1996 ) 321 ; j. hisano and d. nomura , phys . rev .
d 59 ( 1999 ) 116005 .
i. masina and c.a .
savoy , nucl .
b 661 ( 2003 ) 365 .
f. borzumati and a. masiero , phys .
57 ( 1986 ) 961
. j. ellis , et al .
b 621 ( 2002 ) 208 ; phys .
b528 ( 2002 ) 86 .
i. masina , nucl .
b 671 ( 2003 ) 432 . r. barbieri and l. hall , phys .
b338 ( 1994 ) 212 .
a. romanino and a. strumia , nucl .
b 622 ( 2002 ) 73 , i. masina and c.a .
savoy , phys .
b579 ( 2004 ) 99 .
b. bajc , p.f .
perez and g. senjanovi , phys .
d66 ( 2002 ) 075005 .
w. buchmuller and t. yanagida , phys .
b445 ( 1999 ) 399 ; w. buchmuller , d. delepine and f. vissani , phys . lett .
( 1999 ) 171 ; j. sato and k. tobe , phys .
rev.d 63 ( 2001 ) 116010 ; k.s .
babu , ts .
enkhbat , i. gogoladze , nucl .
b 678 ( 2004 ) 233 . k.s .
babu , b. dutta and r.n .
mohapatra , phys .
85 ( 2000 ) 5054 . | we shortly review and emphasize how @xmath0experiments and the searches for lepton e.d.m . are constraining new physics model building .
they are pure signals of new phenomena around the tev scale since the sm contributions are definitely negligible .
it is quite remarkable that they also give effective tests of the lfv&cpv in seesaw couplings and in grand - unified theories . in particular , the limits on @xmath1
nicely complement the proton decay bounds in selecting o(10 ) models . |
recent multi - waveband observations of active galactic nuclei ( agns ) have established significant deviations in the spectral shape of big blue bump from a blackbody one .
a number of authors have tried to distort the accretion - disk spectrum toward the high energy regime so that the disk can emit substantial soft x - ray radiation as is observed .
one promising idea is comptonization within the disk in the vertical direction ( e.g. , [ 1 , 2 , 3 ] ) .
the effect of comptonization is more prominent at higher accretion rates
. however , there still remain discrepancies between models and observations .
i ) although comptonization tends to increase @xmath0 , the far - uv ( fuv ) spectrum of a comptonized accretion disk has @xmath0 @xmath2 1 at best ( [ 1 , 3 ] ) , whereas observed fuv spectra of distant quasars seem to be steeper ( @xmath5 1.82.2 ; [ 4 ] ) .
ii ) the observed spectral index in soft x - rays ( @xmath6@xmath7 ; [ 5 ] ) is not achieved by any disk models since they produce wien bumps at high energy , thus exhibiting exponential roll - over .
iii ) comptonized accretion disks can not reproduce the hard x - rays .
thus , hard x - ray emission should be treated as an additional component in these models .
we therefore propose a new model , aiming to produce the overall sed by a disk - corona model .
the observed composite spectrum is taken from [ 4 , 5 ] ( see [ 6 ] for issues to be kept in mind ) .
the numerical code used in this study is basically the same as that of [ 2 , 3 ] except for some modifications . a constant fraction , @xmath8 , of mass accretion ( @xmath9 ) is assumed to be dissipated in the corona with a thomson optical depth of @xmath10 , where we consider advective energy transport of protons as well as radiative cooling of electrons ( figure 1 ) . a remaining fraction , @xmath11 , dissipates within the disk layer . the advective cooling rate in the corona is taken from the expression of an optically thin advection dominated accretion flow ( adaf ; [ 7 ] ) the efficiency of the advection is controlled by the viscosity parameter in the corona , @xmath12 . the heating rate in the disk and corona , @xmath13 and @xmath14 , are proportional to and @xmath15 , respectively , and the advective cooling rate in the corona per unit volume , @xmath16 , is assumed to be proportional to the matter density .
the energy balance in each layer is as follows : @xmath17 where @xmath18 is the energy exchange rate due to coulomb collisions and @xmath19 is the radiative cooling rate .
we consider free - free emission / absorption and thomson / compton scattering as radiation mechanisms .
the comptonization is described by the kompaneets equation .
we divide the disk - corona from @xmath20 to @xmath21 into 20 rings so that each ring radiates approximately the same luminosity ( cf .
[ 1 , 3 ] ) . in total , the input parameters required for the calculations are @xmath22 , @xmath23 , @xmath8 , @xmath10 , @xmath24 , and @xmath12 .
the number of input parameters is similar to that of the relevant observed parameters which we aim to reproduce simultaneously ; e.g. , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 .
the spectrum of the whole disk - corona system is obtained by summing up the emergent spectra of all the rings .
we first show the most successful case with @xmath31 and @xmath32 that corresponds to a luminosity of about 5% of the eddington luminosity , @xmath33 .
the thick line in figure 2 shows an example of the resultant broad - band spectra .
a significant fraction @xmath34 of mass accretion occurs via the corona .
it turns out that the height of the disk - corona boundary measured from the mid - plane at @xmath35 = 5@xmath36 is 0.03 @xmath36 , and that of the surface of the corona is 0.3 @xmath36 .
then , the disk - corona system is indeed geometrically thin .
the presence of multiple spectral components is the most noteworthy feature of the present model .
this is because different radiative mechanisms play roles in different wave - bands ; thermal radiation of the disk in opt./uv , unsaturated comptonization in fuv / soft x - rays , and a combination of unsaturated comptonization , bremsstrahlung , and reflection in hard x - rays .
we wish to stress that the underlying radiative processes in soft hard x - rays are distinct from the traditional explanation , in which the uv soft x - ray component is due to blackbody radiation whereas the hard power - law component is due to comptonization .
note that the hard x - ray emission in our model looks only apparently like a power law with @xmath3 0.51.0 .
the accretion - rate dependence of the emergent spectra is shown in figure 3 . as the accretion rate increases , the cut - off frequency of hard x - rays ( i.e. , coronal electron temperature ) decreases due to the increasing efficiency of compton cooling .
then , the spectral slope at @xmath2 0.031 kev gets steeper ; @xmath37 = 1.4 , 1.7 and 2.1 for @xmath38 = 0.1 , 0.5 and 0.7 , respectively .
the qualitative trend that higher @xmath39 leads to larger @xmath29 is quite reminiscent of the cases of narrow - line seyfert 1s . | to explain the broad - band spectral energy distributions ( sed ) of seyfert nuclei and qsos , we study the emission spectrum emerging from a vertical disk - corona structure composed of a two - temperature plasma by solving hydrostatic equilibrium and radiative transfer self - consistently . our model can nicely reproduce the soft x - ray excess with @xmath0 ( @xmath1 ) of about 1.5 and the hard tail extending to @xmath2 50 kev with @xmath3 0.5 . the different spectral slopes ( @xmath3 1.5 below 2 kev and @xmath4 0.5 above ) are the results of different emission mechanisms : unsaturated comptonization in the former and a combination of comptonization , bremsstrahlung , and reflection of the coronal radiation at the disk - corona boundary in the latter . accretion , accretion disks , radiation mechanisms : miscellaneous |
during the last three decades , the two - dimensional ( 2d ) spin-1/2 heisenberg model and its generalizations have been investigated in great detail both analytically and numerically @xcite .
this is because these models are regarded as the relevant models for the 2d quantum antiferromagnets .
furthermore , although the phase diagram of high temperature cuprate superconductors is not well understood , it is generally believed that these cuprate superconductors may be obtained by doping the quantum antiferromagnetic insulators with charge carriers . as a result ,
research related to these models is still very active even today .
in addition to the spin-1/2 heisenberg model , higher spin antiferromagnets , in particular the spin-1 heisenberg model , are of theoretical interest due to the fact that they are relevant in explaining experimental results of real materials as well @xcite .
for example , ndmaz , nenp , and pbni@xmath8v@xmath8o@xmath9 are found to be spin-1 quasi - one - dimensional antiferromagnets . besides the simplest type of these models which have spatially isotropic couplings , the spatially anisotropic heisenberg models are also studied thoroughly @xcite . in particular , these generalized models are frequently used as a route for studying quantum phase transitions . moreover , the spatially anisotropic models are important in understanding experimental data .
two notable examples are the 2d spin-1/2 heisenberg model with antiferromagnetic couplings @xmath3 and @xmath4 refers to its value in the spatial @xmath10-direction . ] on the square and rectangular lattices as depicted in fig .
[ fig0 ] , and the three - dimensional ( 3d ) quantum antiferromagnet with a ladder pattern of spatial anisotropy on a cubic lattice .
the former is argued to be relevant to the newly discovered pinning effects of the electronic liquid crystal in the underdoped cuprate superconductor yba@xmath8cu@xmath11o@xmath12 @xcite , and the latter is considered to be the right model for explaining the phase diagram of tlcucl@xmath11 under pressure @xcite . to conclude , despite their simplicity ,
the spatially anisotropic heisenberg models are among the most important and frequently studied systems in condensed matter physics . among the spatially anisotropic heisenberg models with quantum spin , the one shown in fig .
1 is particularly special . for this model
one sees clearly that as the magnitude of the ratio of couplings @xmath5 decreases , the system will eventually become decoupled one - dimensional ( 1d ) antiferromagnetic chains ( this takes place when @xmath13 ) .
one intriguing physics to explore for this spin-1/2 model is to examine whether a phase transition , between the antiferromagnetic and dimerized phases , occurs before one reaches the extreme case @xmath13 .
analytic ( and some numerical ) evidence indicates that for the model of fig . 1 with quantum spin , the long - range antiferromagnetic order is destroyed only for infinitesimal @xmath5 @xcite . as a result , to study whether a phase transition appears at a particular value of @xmath14 using unbiased quantum monte carlo simulations is subtle . indeed , as suggested in refs .
@xcite , square lattice is not the appropriate lattice geometry for studying this model and rectangular lattice should be used instead .
furthermore , to capture the 2d characters of the model , the ratio of linear lattice sizes @xmath6 needs to be adjusted individually for each considered @xmath5 .
even so , one must carry out careful investigation for the relevant observables so that the correct physics is obtained .
for example , to make sure that the extrapolating results are reliable , for every @xmath5 several ratios of @xmath6 may be needed in the calculations . for the spin-1/2 model depicted in fig . 1 , one proposed quantitative approach to study the 2d ground state properties for every considered @xmath5 is to adjust the corresponding ratio @xmath6 so that the spatial winding number squared in the 1- and 2-directions , namely @xmath15 and @xmath16 take the same values . with such a method the 2d characters of the model
will not be lost and can be obtained unbiasedly .
this idea and similar ones have been used to study the spin-1/2 model ( of fig .
1 ) @xcite as well as quantum phase transitions of 2d dimerized quantum spin models @xcite .
the motivation of our study presented here is to provide a more convincing numerical result to support the quantitative correctness of the method of requiring @xmath17 in the simulations . indeed , for the model in fig . 1 with quantum spin
, monte carlo and series expansion results of the observable spin stiffness in the 1-direction agree with each other only for @xmath18 @xcite .
therefore it is desirable to carry out a more detailed investigation so that the validity of this method can be justified .
notice it is known that the quantum phase transition associated with dimerization for the spin-1 model with the same spatial anisotropy takes place at a finite value of @xmath5 @xcite . hence simulating the spin-1 model with the same spatial anisotropy shown in fig .
1 provides a great opportunity to further examine the validity of this method .
indeed as we will demonstrate later , our monte carlo results of @xmath7 for the spin-1 model agree quantitatively with those determined by the series expansion method from @xmath19 down to @xmath20 @xcite .
consequently our investigation gives convincing evidence for the correctness of this method .
notice the spatially anisotropic quantum heisenberg model of fig .
[ fig0 ] and its 1d limit are two completely different systems @xcite .
therefore an unconventional behavior is likely to appear as one approaches the 1d limit from the 2d model .
our study paves a way to investigate the novel phenomena of dimension crossover from a 2d system to its 1d limit .
we would like to point out that even with the method of adjusting the aspect ratio @xmath6 in the calculations so that the two spatial winding numbers squared reach the same numbers , the determination of ground state bulk properties requires the simulations being conducted at very low temperatures .
such zero temperature calculations are computationally demanding . therefore instead of performing the simulations at very low temperatures and using the conventional approach of fitting the data with polynomials of the relevant parameters , we carry out the calculations at finite temperatures and employ the relevant predictions from magnon chiral perturbation theory ( m@xmath21pt ) to extract the bulk properties of the considered model .
later we will briefly argue that under certain circumstances , this approach seems to be more efficient than the conventional one .
indeed , for each considered @xmath5 , a few tens of the corresponding data points can be described quantitatively by two equations with only three unknown parameters , and particularly , we are able to arrive at high precision results with moderate computing resources .
finally , another remarkable finding in our investigation is that , the m@xmath21pt can also be used to examine unambiguously the presence of the long - range antiferromagnetic order .
this will be explained in more detail in the relevant section .
this paper is organized as follows .
first , after the introduction , the spatially anisotropic spin-1 heisenberg model and the observables considered in this work are detailed .
furthermore , the m@xmath21pt is briefly introduced and some of its predictions relevant to our study are also listed .
we then present the data as well as the resulting numerical results based on these data .
finally a section is devoted to conclude our investigation .
the 2d spin-1 heisenberg model we consider in this study is defined by the hamilton operator @xmath22,\end{aligned}\ ] ] and is depicted by fig . 1 .
( 1 ) , @xmath23 and @xmath24 refer to the two spatial unit - vectors and @xmath25 is the lattice spacing . in addition ,
@xmath26 and @xmath27 are the antiferromagnetic couplings in the @xmath28- and @xmath29-direction , respectively .
finally the @xmath30 shown above is a spin-1 operator at site @xmath10 .
a physical quantity measured in our simulations is the staggered susceptibility @xmath31 , which is given by @xmath32.\end{aligned}\ ] ] here @xmath33 is the inverse temperature , @xmath34 and @xmath35 are the spatial box sizes in the @xmath28- and @xmath29-direction , respectively , and @xmath36 is the partition function .
furthermore , @xmath37 appearing in eq .
( 2 ) is the third component of the staggered magnetization @xmath38 .
another observables considered here is the uniform susceptibility @xmath39 , which takes the form @xmath40.\end{aligned}\ ] ] here @xmath41 is the uniform magnetization . both @xmath31 and @xmath39 can be measured very efficiently with the loop - cluster algorithm using improved estimators @xcite .
notice both of these two observables can be expressed in quantities related to the clusters .
in particular , @xmath39 is associated with the temporal winding number @xmath42 which is the sum of winding numbers @xmath43 of the loop - clusters @xmath44 around the euclidean time direction .
similarly , the spatial winding numbers @xmath45 for @xmath46 are defined in the same manner . with the convention employed here ,
the spin stiffness @xmath0 ( for @xmath19 ) can be obtained directly from the standard relation @xmath47 in the zero - temperature and infinite - volume limits .
in addition , the temporal winding number squared @xmath48 calculated in this study is exactly the susceptibility @xmath21 and is related to @xmath39 by @xmath49 . finally the spinwave velocity @xmath2 can be estimated from these winding numbers squared by the method detailed in refs .
in this section we summarize the relevant theoretical predictions , namely the finite - volume and -temperature expressions of @xmath31 and @xmath39 from m@xmath21pt @xcite .
these predictions are used to calculate the desired low - energy constants . due to the spontaneous breaking of the @xmath50 spin symmetry down to its @xmath51 subgroup , the low - energy physics of antiferromagnets
is governed by two massless goldstone bosons , the magnons .
the systematic low - energy effective field theory for magnons is formulated in term of the staggered magnetization .
the staggered magnetization of an antiferromagnet is described by a unit - vector field @xmath52 in the coset space @xmath53 , i.e. @xmath54 with @xmath55 .
here @xmath56 denotes a point in ( 2 + 1)-dimensional space - time . to leading order
, the euclidean magnon low - energy effective action takes the form @xmath57 & = & \int^{l_1}_{0 } dx_1 \int^{l_2}_{0 } dx_2 \int^{\beta}_{0 } \ dt \ \left(\frac{\rho_{s1}}{2}{\partial}_1 \vec e \cdot { \partial}_1 \vec e \right . \nonumber \\ & & + \left .
\frac{\rho_{s2}}{2}{\partial}_2 \vec e \cdot { \partial}_2 \vec e + \frac{\rho_s}{2c^2 } { \partial}_t \vec e \cdot { \partial}_t \vec e\right),\end{aligned}\ ] ] where the index @xmath58 labels the two spatial directions and @xmath59 refers to the euclidean time - direction .
the temporal spin stiffness @xmath0 is given by @xmath60 , where @xmath7 and @xmath61 are the spin stiffness in the spatial directions . finally , the parameter @xmath2 in eq .
( [ action ] ) is the spinwave velocity .
notice the physical quantities , namely @xmath0 , @xmath7 , @xmath61 , and c appearing inside eq .
( [ action ] ) are the bulk ones @xcite . by introducing @xmath62 and @xmath63 , eq .
( [ action ] ) can be rewritten as @xmath64 & = & \int^{l'_{1}}_{0 } dx'_1 \int^{l'_{2}}_{0 } dx'_2 \int^{\beta}_{0 } dt \
\frac{\rho_s}{2}\big({\partial}'_i \vec{e } \cdot
{ \partial}'_i \vec{e } \nonumber \\ & & + \frac{1}{c^2 } { \partial}_t \vec{e } \cdot { \partial}_t \vec{e}\big).\end{aligned}\ ] ] if one additionally requires @xmath65 , then the condition of square area is obtained . with the euclidean action eq .
( [ action1 ] ) , the finite - volume and -temperature expressions of @xmath31 and @xmath39 in the cubical regime , where the condition @xmath66 is met , are calculated in ref .
@xcite and take the following forms @xmath67 + o\left(\frac{1}{l^3}\right ) \right\}\end{aligned}\ ] ] and @xmath68 + o\left(\frac{1}{l^3}\right ) \right\},\end{aligned}\ ] ] respectively . in eqs .
( [ chiscube ] ) and ( [ chiucube ] ) , the functions @xmath69 , @xmath70 , and @xmath71 , which only depend on @xmath72 , are shape coefficients of the space - time box .
the explicit formulas of these shape coefficients can be found in ref .
in order to determine the low - energy constants as functions of @xmath5 for the 2d spatially anisotropic spin-1 heisenberg model given by eq .
( [ hamilton ] ) ( and depicted in fig . 1 )
, we have performed simulations for @xmath73 with various box sizes using the loop algorithm @xcite .
the value of @xmath74 is included in our consideration since it is slightly below the critical point @xmath75 determined in ref .
the results at @xmath74 provide an opportunity to examine whether our monte carlo data at @xmath74 can be captured quantitatively by the relevant predictions of m@xmath21pt . without loss of generality
, we have set @xmath76 in our monte carlo simulations .
the cubical regime is determined by the condition @xmath77 .
notice eqs .
( [ chiscube ] ) and ( [ chiucube ] ) are obtained for a ( 2 + 1)-dimensional box with equal extent in the two spatial directions . since @xmath78 in our calculations ,
the box sizes @xmath79 and @xmath80 used in the simulations must satisfy @xmath81 so that the condition @xmath82 can be fulfilled .
finally interpolation of the data points is necessary as well in order to use eqs .
( [ chiscube ] ) and ( [ chiucube ] ) . after employing all these requirements in our calculations ,
the low - energy constants can be extracted by fitting the monte carlo data to the effective field theory predictions .
first of all , in the next subsection we focus on our monte carlo results of the isotropic situation @xmath83 . as a first step toward a high accuracy determination of the desired low - energy constants ,
the spinwave velocity @xmath2 is calculated through the square of winding numbers as suggested in refs .
specifically , for a given square lattice with linear box size @xmath79 , one tunes the inverse temperature @xmath33 to a value @xmath84 so that @xmath77
. then the @xmath2 corresponding to this finite lattice is estimated to be @xmath85 .
the numerical values of @xmath2 obtained by this method for several box sizes @xmath79 are given in table 1 , and a weighted average of these values of @xmath2 leads to @xmath86 .
the result of @xmath2 , namely @xmath86 reached here has ( much ) better precision than its early known monte carlo results @xcite , and is in nice agreement with the spinwave theory estimate @xmath87 @xcite .
furthermore , since for the isotropic case one can reach the condition @xmath88
= @xmath89 without performing any interpolation , we employ eqs .
( [ chiscube ] ) and ( [ chiucube ] ) directly with a fixed @xmath86 to simultaneously fit the uninterpolated data of @xmath31 and @xmath90 .
the numerical values of @xmath1 and @xmath0 obtained from the fit are given by @xmath91 and @xmath92 , respectively .
in addition , the @xmath93 of this fit is around 1.0 .
notice the @xmath1 and @xmath0 calculated here also match excellently with @xmath94 and @xmath95 determined by the spinwave theory .
the results of the fit are shown in fig .
[ fig1 ] . besides the method of employing the predictions of m@xmath21pt
, @xmath0 can be determined directly from the spatial winding numbers squared .
such a calculation of @xmath0 through @xmath96 and @xmath97 provides a good opportunity to verify the quantitative correctness of m@xmath21pt .
hence we have carried out new simulations with @xmath98 for several box sizes @xmath99 at low temperatures .
the newly obtained @xmath0 data as a function of the box sizes @xmath99 is shown in fig .
[ fig1.5 ] .
remarkably , the @xmath0 determined from the spatial winding numbers squared is in quantitative agreement with that calculated using the predictions of m@xmath21pt .
for instance , by applying polynomials up to third ( fourth ) order in @xmath100 to the relevant data of @xmath101 ( @xmath101 ) , we arrive at @xmath102 ( @xmath103 ) .
these values of @xmath0 match nicely with those we calculated earlier using the related formulas from m@xmath21pt . to extract @xmath0 directly from the squares of spatial winding numbers
, one has to obtain the zero - temperature values of the relevant observables .
this is computationally demanding .
in addition , it is also crucial to include as many data points as possible in the fits so that one can reach a accurate result of @xmath0 .
hence when obtaining the ground state properties of a system is challenging , our approach , i.e. fitting the data using the relevant equations from m@xmath21pt , seems to be a more efficient way of calculating these low - energy constants since the related simulations are performed at finite - temperature and cubical regime .
[ tab1 ]
.the numerical values of @xmath2 at finite lattices for @xmath19 .
these results are obtained from the squares of spatial and temporal winding numbers . [ cols="^,^,^,^",options="header " , ] 1.0 cm , @xmath61 , and @xmath0 as functions of @xmath5 .
the series expansion results shown in the figure are estimated from ref .
the solid lines are added to guide the eyes.,scaledwidth=35.0% ]
in this study , we have calculated the low - energy constants , namely the spin stiffness @xmath0 , the staggered magnetization density @xmath1 per area , and the spinwave velocity @xmath2 of the spin-1 heisenberg model with antiferromagnetic couplings @xmath26 and @xmath27 on the rectangular lattices using the quantum monte carlo simulations .
the relevant finite - volume and -temperature predictions of m@xmath21pt are employed in extracting the numerical values of these low - energy constants .
such an approach is of computational efficiency as well , since the related simulations are conducted at finite temperatures .
the precision of @xmath1 , @xmath0 , and @xmath2 obtained here for @xmath98 is improved .
furthermore , the anisotropy @xmath5 dependence of @xmath1 , @xmath104 ( @xmath7 ) , and @xmath2 are investigated in detail as well .
in particular , the results of @xmath1 and @xmath2 determined here for the spatially anisotropic models are new and can serve as benchmarks for future related studies .
our monte carlo and the series expansion results of @xmath7 are in nice agreement for @xmath105 @xcite .
consequently the quantitative correctness of our approach is justified .
it is remarkable that the series expansion method leads to consistent values of @xmath7 with those from monte carlo simulations in such strong anisotropic regime @xmath105 .
we also confirm that the m@xmath21pt can be used efficiently to study the breaking down of long - range ( antiferromagnetic ) order .
specifically , for a considered relevant parameter , one can conclude that a phase transition from the long - range antiferromagnetic phase to a disordered phase already takes place before reaching that given parameter , if the fits using predictions of m@xmath21pt lead to poor fitting quality .
indeed , the outcomes of the fit using the data of @xmath74 is consistent with the conclusion obtained in ref .
@xcite that the critical point @xmath106 is given by @xmath75 . considering the subtlety of quantitatively capturing the 2d characters of the spin-1/2 quantum heisenberg model with the same spatial anisotropy as the one considered here
, our study paves a way to unbiasedly investigate the novel phenomena of 2d to 1d dimension crossover of the related spin-1/2 model .
simulations were done based on the loop algorithms in alps @xcite .
f .- j . j. is supported partially by most of taiwan . | the low - energy constants , namely the spin stiffness @xmath0 , the staggered magnetization density @xmath1 per area , and the spinwave velocity @xmath2 of the two - dimensional ( 2d ) spin-1 heisenberg model on the square and rectangular lattices are determined using the first principles monte carlo method .
in particular , the studied models have antiferromagnetic couplings @xmath3 and @xmath4 in the spatial 1- and 2-directions , respectively . for each considered @xmath5 , the aspect ratio of the corresponding linear box sizes @xmath6 used in the simulations
is adjusted so that the squares of the two spatial winding numbers take the same values .
in addition , the relevant finite - volume and -temperature predictions from magnon chiral perturbation theory are employed in extracting the numerical values of these low - energy constants .
our results of @xmath7 are in quantitative agreement with those obtained by the series expansion method over a broad range of @xmath5 .
this in turn provides convincing numerical evidence for the quantitative correctness of our approach .
the @xmath1 and @xmath2 presented here for the spatially anisotropic models are new and can be used as benchmarks for future related studies . -0.5 |
a representation of closed @xmath0-manifolds by @xmath1-colored graphs has been independently introduced in the seventies by m. pezzana and in the eighties by s. lins ( see @xcite and @xcite ) , by using dual constructions .
a @xmath1-colored graph is a regular graph of degree four , endowed with a proper edge - coloration ( i.e. , adjacent edges has different colors ) with four colors , and it represents a closed @xmath0-manifold if and only if a certain combinatorial condition is satisfied .
this type of representation has been very fruitful with regard to the catalogation of closed @xmath0-manifolds in terms of increasing complexity , which in this context means number of vertices of the representing graph ( see @xcite , , @xcite and @xcite ) .
the extension of the representation by @xmath1-colored graphs to @xmath0-manifolds with boundary has been performed in @xcite , where any @xmath1-colored graph is associated to a compact @xmath0-manifold with ( possibly empty ) boundary without spherical components , this correspondence being surjective on the whole class of such manifolds . as a consequence ,
an efficient computer aided catalogation of @xmath0-manifolds with boundary can be performed by this tool .
some computational results have been already obtained in @xcite , namely the classification of all manifolds representable by a graph of order @xmath2 in the orientable case and @xmath3 in the non - orientable one . in the present paper
our main goals are the creation of : * the census of all non - isomorphic contracted bipartite @xmath1-colored graphs ( without @xmath4-dipoles ) of order @xmath5 , representing compact orientable @xmath0-manifolds with ( possibly disconnected ) toric boundary ; * the census of the prime @xmath0-manifolds represented by the above graphs ; * the catalogue of all manifolds in the above census which are complements of knots and links in the @xmath0-sphere , also giving , for each of such manifolds , the diagram of a corresponding knot / link .
for ( a ) , we wrote a script that , for each positive integer @xmath6 , enumerates all non - isomorphic contracted @xmath1-colored graphs with @xmath7 vertices and no @xmath4-dipoles , representing compact bordered @xmath0-manifolds , possibly fixing the topological type of their boundary . in section [ sec3 ] , after describing the enumeration process , we give computational results for @xmath8 with no restriction on the orientability and the topological type of the boundary ( see table 1 ) , and for @xmath9 in the case of graphs representing orientable manifolds with toric boundary ( see table 2 ) . in order to get the census of ( b ) from the one of ( a )
, three steps were necessary .
the first one consisted in removing certain graphs which were not minimal among those representing the same manifold ( see lemma [ rho2 ] ) or which represented non - prime manifolds ( see lemma [ rho3 ] ) . in the second step
, for each contracted bipartite @xmath1-colored graph @xmath10 of this `` reduced '' census we tried to identify the corresponding manifold @xmath11 by means of ` 3-manifold recognizer ` @xcite , which is a computer program for studying @xmath0-manifolds written by v. tarkaev according to an algorithm elaborated by s. matveev and other members of the chelyabinsk topology group . as a first result
, we obtained a census of compact orientable @xmath0-manifolds with toric boundary not containing duplicates and non - prime manifolds .
furthermore , all seifert manifolds and graph manifolds of waldhausen ( see @xcite ) were identified by means of their structure ( see subsections [ seifert ] and [ graph - manifolds ] ) . for the manifolds @xmath11 that were not identified by ` 3-manifold recognizer ` , the program returned the isomorphism signature of the triangulation corresponding to @xmath10 . in the third step the classification
was completed by using ` snappy ` , a computer program for studying the topology of @xmath0-manifolds , written by m. culler and n. dunfield using the snappea kernel by j. weeks , with contributions from many others .
more precisely , the isomorphism signatures were entered into ` snappy ` , which identified all still unrecognized manifolds , except three , as hyperbolic manifolds in its census ( see section [ sec4 ] ) .
for the census in ( c ) , we used a catalogue of links ( up to @xmath12 crossings ) included in ` snappy ` and in some cases the recognized structure of graph manifolds .
it turned out that all @xmath0-manifolds in our census except two were complements of links in the @xmath0-sphere ( see table 3 ) . for each manifold of this kind
we also displayed the diagram of a corresponding link ( see figures [ links1 ] and [ links2 ] ) .
when more than one link was available a prime one was preferred ( identified by its name in the hoste - thistlethwaite table ) .
let @xmath10 be a finite graph which is @xmath1-regular ( i.e. , each vertex has degree four ) , possibly with multiple edges but with no loops .
a map @xmath13 is called a _ @xmath1-coloration _ of @xmath10 if adjacent edges have different colors .
an edge of @xmath10 colored by @xmath14 is also called a _ @xmath15-edge_. a _ @xmath1-colored graph _ is a connected @xmath1-regular graph equipped with a @xmath1-coloration .
it is easy to see that any @xmath1-colored graph has even order . if @xmath16 , a _ @xmath17-residue _ ( as well as a _ @xmath18-residue _ ) of @xmath10 is any connected component of the subgraph @xmath19 of @xmath10 containing exactly all @xmath15-edges , for each @xmath20
of course , @xmath21-residues are vertices , @xmath22-residues are edges and @xmath4-residues are bicolored cycles . for each @xmath14 , we denote by @xmath23 the number of @xmath0-residues corresponding to the colors of @xmath24 . a @xmath1-colored graph @xmath10 induces a compact connected @xmath0-manifold @xmath11 without spherical boundary components via the following construction . by attaching a disk to each @xmath4-residue of @xmath10 ( considered as a @xmath22-dimensional cellular complex )
we obtain a @xmath4-dimensional polyhedron @xmath25 .
each @xmath0-residue of @xmath10 , with its associated disks , forms a closed connected surface @xmath26 embedded in @xmath25 .
if @xmath26 is a @xmath4-sphere the residue is called _ ordinary _ , otherwise it is called _
singular_. then , for any @xmath0-residue @xmath26 , fill it with a @xmath0-ball if it is ordinary , and just thicken it by attaching @xmath27 along @xmath28 if it is singular . at the end of this process ,
a compact connected @xmath0-manifold @xmath11 with non - spherical boundary components is obtained .
we will say that @xmath10 _ represents _ @xmath11 , as well as that @xmath11 is _ associated _ to @xmath10 .
obviously isomorphic @xmath1-colored graphs - colored graphs @xmath29 and @xmath30 , with coloration @xmath31 and @xmath32 respectively , are isomorphic if there exist a graph isomorphism @xmath33 between @xmath29 and @xmath30 and a permutation @xmath34 of @xmath35 such that @xmath36 . ]
represent homeomorphic @xmath0-manifolds . for closed @xmath0-manifolds
the construction reduces to the one introduced by lins ( see @xcite ) , and it is dual to the one introduced by pezzana ( see @xcite ) .
so , the novelty of this construction is that it works also in case of @xmath0-manifolds with boundary , and @xmath37 is a surjective map from the whole class of @xmath1-colored graphs to the whole class of compact @xmath0-manifolds with ( possibly empty ) boundary without spherical components .
in fact , the following result is proved in @xcite .
if @xmath38 is a compact @xmath0-manifold with ( possibly empty ) boundary without spherical components , then there exists a @xmath1-colored graph @xmath10 such that @xmath38 is homeomorphic to @xmath11 .
moreover , @xmath11 is orientable if and only if @xmath39 is bipartite .
since two compact @xmath0-manifolds are homeomorphic if and only if ( i ) they have the same number of spherical boundary components and ( ii ) they are homeomorphic after capping off by balls these components , there is no loss of generality in studying compact @xmath0-manifolds without spherical boundary components . in the following with the term `` manifold
'' we will always refer to such type of @xmath0-manifolds .
-colored graphs @xmath40 and and @xmath41 , representing the genus one orientable and non - orientable handlebody respectively . ] in @xcite , @xmath1-colored graphs representing important families of compact @xmath0-manifolds , such as handlebodies ( both orientable and non - orientable ) , and @xmath27 , where @xmath26 is any closed surface ( either orientable or non - orientable ) are shown .
as usually happens in any representation technique of @xmath0-manifolds ( triangulations , spines , branched coverings , surgery on links , etc ... ) , any compact @xmath0-manifold admits representation by infinitely many @xmath1-colored graphs .
so , moves connecting non - isomorphic graphs representing the same @xmath0-manifolds are important tools in the theory .
let @xmath42 and @xmath43 be adjacent vertices of a @xmath1-colored graph @xmath10 of order @xmath44 , and let @xmath45 be the subgraph of @xmath10 consisting of @xmath46 and all the @xmath47 edges connecting them .
then @xmath45 is called an _ @xmath47-dipole _ if @xmath43 and @xmath48 belong to different @xmath49-residues of @xmath10 , where @xmath50 are the colors of the @xmath47 edges of @xmath45 . by _ cancelling _
@xmath45 _ from _ @xmath10 , we mean to remove @xmath45 and to paste together the hanging edges according to their colors , thus obtaining a new @xmath1-colored graph @xmath51 .
conversely , @xmath10 is said to be obtained from @xmath51 by _ adding _ @xmath45 .
an @xmath47-dipole @xmath45 is called _ proper _ when @xmath10 and @xmath51 represent the same manifold .
[ proper ] @xcite an @xmath47-dipole @xmath45 in a @xmath1-colored graph @xmath10 is proper if and only if one of the following conditions holds : * @xmath52 ; * @xmath53 and at least one of the @xmath54-residues containing the two vertices of @xmath45 is ordinary . in the closed case
all dipoles are proper and casali proved in that dipole moves are sufficient to connect different @xmath1-colored graphs representing the same manifold .
a similar fact is no longer true in our context ; in fact dipole moves do not change the singular colors - colored graph @xmath10 , we mean a color @xmath14 such that at least one of the @xmath55-residues of @xmath10 is singular . ] of the involved graph , and for any @xmath0-manifold with disconnected boundary it is easy to find two different graphs representing it , the first one with only a singular color and the second one with ( at least ) two singular colors .
a @xmath1-colored graph @xmath10 is said to be _ contracted _ if either @xmath56 or all @xmath55-residues of @xmath10 are singular , for each @xmath14 . note that a @xmath1-colored graph is contracted if and only if it does not contain any proper @xmath22-dipole , and a contracted @xmath1-colored graph contains no @xmath0-dipoles .
moreover , by cancelling proper @xmath22-dipoles , any @xmath1-colored graph can be reduced to a contracted one representing the same manifold .
so , any compact @xmath0-manifold admits representation by contracted @xmath1-colored graphs . a @xmath1-colored graph @xmath10 is called _ minimal _ if there exists no graph with less vertices representing the same manifold .
of course , a graph containing a proper dipole is not minimal . in particular ,
any minimal graph is contracted .
the combinatorial nature of colored graphs makes them particularly suitable for computer manipulation .
in particular , it is possible to generate catalogues of @xmath1-colored graphs for increasing number of vertices , in order to analyze the represented compact @xmath0-manifolds .
we focus our attention to @xmath0-manifolds with non - empty boundary , since the catalogation of graphs associated to closed manifolds ( also called _ gems _ , and _ crystallizations _ in the contracted case ) , has been already performed up to 30 vertices ( see and @xcite ) . by the results of section [ sec2 ] , without loss of generality we can restrict the catalogues to contracted graphs with no @xmath4-dipoles .
therefore , given a positive integer @xmath6 , the catalogue of contracted @xmath1-colored graphs with @xmath7 vertices and no @xmath4-dipoles is generated algorithmically in the following way .
first of all consider the set of all ( possibly disconnected ) @xmath4-colored graphs ( i.e. , disjoint union of @xmath57-colored cycles ) with @xmath7 vertices . for
any element of this set , add the @xmath4-edges in all possible way in order to obtain , by checking bipartition and computing the euler characteristic of the connected components , the set @xmath58 of all ( possibly disconnected ) @xmath0-colored graphs @xmath10 with @xmath7 vertices , such that each connected component of @xmath10 represents a surface of positive genus .
for the values of @xmath6 from @xmath22 to @xmath59 , the cardinality of @xmath58 is @xmath60 and @xmath61 , respectively .
afterwards , the @xmath0-edges are added to each graph of @xmath58 in all possible ways giving rise to a contracted @xmath1-colored graph without @xmath4-dipoles . in order to avoid the presence of isomorphic graphs in the catalogue
, the concept of code of a @xmath1-colored graph is an useful tool .
the _ code _ of a @xmath62-bipartite @xmath1-colored graph - colored graph @xmath10 is called @xmath62-bipartite , if @xmath62 is the largest integer such that all @xmath62-residues of @xmath10 are bipartite . ]
@xmath10 with @xmath7 vertices is a `` string '' of length @xmath63 , which completely describes both combinatorial structure and coloration of @xmath10 ( see @xcite for details ) .
the importance of the code for representing @xmath1-colored graphs relies on the fact that two @xmath1-colored graphs are isomorphic if and only if they have the same code . as a consequence , by using codes we can produce catalogues containing only graphs which are pairwise non - isomorphic .
table 1 shows the number of non - isomorphic contracted @xmath1-colored graphs without @xmath4-dipoles and inducing manifolds with non - empty boundary , where @xmath64 ( resp .
@xmath65 ) refers to bipartite ( resp .
non - bipartite ) graphs of order @xmath7 , with @xmath66 .
with regard to the classification of the manifolds represented by the above graphs , some results concerning cases with small number of vertices have been already established : as it is clear from table 1 , the number of generated graphs quickly increases with the number of vertices and becomes very large even in the initial segment of the catalogues .
so we carried on our research taking care of the cases with ( possibly disconnected ) toric boundary , which includes the important cases of complements of knots or links in the @xmath0-sphere . in this context ,
orientable manifolds are of particular interest .
notice that the five orientable manifolds of the previous proposition are in fact knot / link complements .
namely , the two manifolds represented by graphs with six vertices are the complements of the trivial knot and the hopf link respectively , whereas the three manifolds represented by graphs with eight vertices are the complements of the links l6n1 , l8n8 and l8n7 respectively ( see table 3 and figure [ links1 ] ) , according to the notation of thistlethwaite link table . at the end of this section , table 2 lists in the first two rows the number of non - isomorphic bipartite contracted @xmath1-colored graphs without @xmath4-dipoles representing manifolds with toric boundary , indicated by @xmath67 and the special cases of connected toric boundary , indicated by @xmath68 . when a @xmath1-colored graph is bipartite , special configurations of its edges gives some information about the minimality of the graph and the reducibility of the associated orientable manifold . a
_ @xmath69-pair _ of a bipartite @xmath1-colored graph @xmath10 , for @xmath70 , is a pair of two equally colored edges @xmath71 sharing exactly @xmath72 @xmath4-residues of @xmath10 .
the _ switching _ of a @xmath69-pair @xmath71 is the operation of replacing the edges of the pair , by two new edges @xmath73 and @xmath74 , with the same color as the old ones , connecting the endpoints of @xmath75 and @xmath76 in the only way which preserves the bipartition of @xmath10 ( see figure [ fig : fig2 ] for the case @xmath77 ) . each @xmath4-residue shared by the two edges of the pair splits into two cycles and the new bipartite @xmath1-colored graph @xmath29 obtained after the switching may be either connected or disconnected with two components . in particular , when @xmath78 the new graph @xmath29 is always connected .
-pair . ]
first consider the effect of @xmath79-pair switchings .
let @xmath71 be a @xmath79-pair of @xmath10 , and let @xmath80 be the three colors of the @xmath4-residues shared by @xmath75 and @xmath76 , we call @xmath71 _ good _ if after the switching the @xmath81-residue containing @xmath75 and @xmath76 splits into two connected components and at least one of them is ordinary .
the definition relies on the following lemma which extends to manifolds with non - empty boundary the corresponding result established by lins for the closed case ( @xcite ) : [ rho2 ] let @xmath10 be a bipartite @xmath1-colored graph with a @xmath79-pair , and denote by @xmath29 the @xmath1-colored graph obtained from @xmath10 by switching the pair .
then @xmath10 and @xmath29 represent the same @xmath0-manifold if and only if the pair is good . in this case
@xmath10 is not minimal .
let @xmath82 be the edges of the pair , @xmath83 be the new edges created by the switching , and let @xmath15 be their color .
the @xmath81-residue involved in the switching , considered with its associated disks , is a closed orientable surface @xmath26 .
the switching operates on @xmath26 as follows : ( i ) split each one of the two disks containing both @xmath75 and @xmath76 by cutting them along a properly embedded arc @xmath84 , with @xmath85 , connecting an internal point @xmath86 of @xmath75 with an internal point @xmath87 of @xmath76 ; ( ii ) push the edges @xmath88 and @xmath89 obtained from the cutting until both @xmath75 and @xmath76 disappear ; ( iii ) then indentify @xmath90 ( resp .
@xmath91 ) to a single @xmath15-colored edge @xmath73 ( resp .
@xmath74 ) .
this operation is topologically equivalent to cutting @xmath26 along a suitable closed curve , and filling with two disks the two circles arising from the cutting .
consequently , if @xmath71 is not good then @xmath92 , since either @xmath26 decreases its genus or it splits into two components of positive genus .
conversely , suppose that @xmath71 is good , then the switching can be factorized by a sequence of proper dipole moves as shown in figure [ fig : ro2 ] .
note that @xmath93 is a @xmath4-dipole , hence it is always proper and its cancellation does not change the represented manifold , while @xmath94 is a proper @xmath22-dipole if and only if @xmath71 is good . in this case
, @xmath29 is not contracted since it has at least two @xmath81-residues which are not both singular . as a consequence , @xmath29 is not minimal and therefore @xmath10 is not minimal too , since it has the same order as @xmath29 .
-pair switching ] [ minimal ] any minimal @xmath1-colored graph representing a compact orientable @xmath0-manifold contains no good @xmath79-pair . by corollary [ minimal ] , graphs containing good @xmath79-pairs can be ignored in a process of catalogation of compact orientable @xmath0-manifolds via @xmath1-colored graphs .
observe that a graph without good @xmath79-pairs is automatically without @xmath4-dipoles .
let us consider now the effect of switching @xmath95-pairs .
it is easy to see that the two edges of a @xmath95-pair belong to the same three @xmath0-residues ; the number @xmath96 of the ordinary ones is called the _
index _ of the @xmath95-pair . in the closed case
all @xmath97-pairs are of index three , and in the case of connected boundary it is @xmath98 .
when the switching is performed on a @xmath95-pair of index @xmath98 , there is a strong topological relation between the manifold associated to @xmath10 and the manifold(s ) associated to @xmath29 , as stated by the following lemma .
[ rho3 ] let @xmath10 be a bipartite @xmath1-colored graph with a @xmath95-pair of index @xmath98 , and denote by @xmath29 the ( possibly disconnected ) @xmath1-colored graph obtained from @xmath10 by switching the pair .
* if @xmath99 and @xmath29 is disconnected with components @xmath100 , then @xmath101 .
* if @xmath99 and @xmath29 is connected , then @xmath102 . *
if @xmath103 and @xmath29 is disconnected with components @xmath100 , then either @xmath101 or @xmath104 . * if @xmath103 , @xmath29 is connected and @xmath105 has the same number of components as @xmath106 , then either @xmath102 or @xmath107 . * if @xmath103 , @xmath29 is connected and @xmath105 has less components than @xmath106 , then @xmath108 . in particular , if @xmath103 and @xmath106 is a union of tori , then either @xmath101 or @xmath102 or @xmath108 .
let @xmath82 be the edges of the pair , @xmath83 be the new edges created by the switching , and let @xmath15 be their color .
in term of @xmath4-skeleton the switching operates as follows : ( i ) split each one of the three disks containing both @xmath75 and @xmath76 by cutting them along a properly embedded arc @xmath84 , with @xmath109 , connecting an internal point @xmath86 of @xmath75 with an internal point @xmath87 of @xmath76 ; ( ii ) push the edges @xmath88 and @xmath89 obtained from the cutting until both @xmath75 and @xmath76 disappear ; ( iii ) then indentify @xmath110 ( resp .
@xmath111 ) to a single @xmath15-colored edge @xmath73 ( resp .
@xmath74 ) , where @xmath112 .
the cuttings of the disks extend to @xmath11 in the following way .
for each @xmath113 , if the @xmath114-residue containing @xmath75 and @xmath76 is ordinary , the attached @xmath0-ball is cut along a properly embedded disk with boundary @xmath115 , otherwise the attached @xmath27 is cut along an annulus @xmath116 , where @xmath117 .
( i , ii ) if @xmath99 , the cutting is performed along three embedded disks whose union is an embedded sphere .
so ( i ) and ( ii ) are proved .
if @xmath103 , the cutting is performed along two disks and an annulus , whose union is an embedded disk @xmath94 with boundary belonging to a boundary component of @xmath11 , which is cut along @xmath118 .
then the arguments of the proof of corollary 2.4.4 of @xcite apply , considering that in our case @xmath119 may be greater than one .
\(iii ) if @xmath29 is disconnected and @xmath26 is non - trivially disconnected by the cutting into two components @xmath120 then @xmath104 , where the boundary connected sum is performed along @xmath121 and @xmath122 ( of course this case can not occurs if the genus of @xmath26 is 1 ) .
if @xmath29 is disconnected and @xmath26 trivially disconnects into two components @xmath120 , where @xmath122 is a disk , then @xmath101 , since @xmath123 is a @xmath4-sphere ( for @xmath124 this is case 1 of @xcite ) .
\(iv ) if @xmath29 is connected and @xmath105 has the same number of components as @xmath106 then either @xmath26 is trivially disconnected by the cutting or @xmath125 and @xmath26 is not disconnected by the cutting . in the first case @xmath102 ( for @xmath124 this is case 2 of @xcite ) and in the second one @xmath107 .
\(v ) if @xmath29 is connected and @xmath105 has less components than @xmath106 , then @xmath124 and @xmath26 is not disconnected by the cutting .
so @xmath108 ( this is case 3 of @xcite ) . because of the previous result , we call _ good _ a @xmath95-pair satisfying one of the conditions ( i)-(v ) of lemma [ rho3 ] .
so , any @xmath95-pair of index three is good .
moreover , a @xmath95-pair of index two is good if and only if the switching either disconnects the graph or the singular residue associated to the pair does not split into two components which are both singular .
observe that in case of toric boundary all @xmath95-pairs of index @xmath126 are good .
let @xmath10 be a bipartite @xmath1-colored graph with a good @xmath95-pair .
then either ( i ) @xmath11 is not prime or ( ii ) @xmath11 is not boundary - prime or ( iii ) @xmath10 is not minimal or ( iv ) @xmath127 or ( v ) @xmath128 .
suppose the switching of the @xmath95-pair disconnects @xmath10 into two components @xmath40 and @xmath129 , then either @xmath101 or @xmath104 . if the connected sum is not trivial , then either ( i ) or ( ii ) holds . if the connected sum is trivial then ( iii ) holds .
otherwise , suppose the switching does not disconnect @xmath10 . if @xmath29 is the graph obtained after the switching , then either @xmath102 or @xmath107 or @xmath108
. if the connected sum is not trivial then either ( i ) or ( ii ) holds , otherwise either ( iv ) or ( v ) holds .
a @xmath1-colored graph without good @xmath79- and @xmath95-pairs will be called _ rigid_. [ rigid ]
any minimal @xmath1-colored graph representing a compact orientable prime and boundary - prime @xmath0-manifold different from @xmath130 and @xmath131 is rigid .
given in table 3 is minimal , but it is not rigid since it contains a good @xmath95-pair . ] therefore , for the study of compact prime and boundary - prime orientable @xmath0-manifolds ( except for the trivial cases of @xmath130 and @xmath131 ) , it is enough to consider only contracted rigid @xmath1-colored bipartite graphs .
when the boundary of the represented manifolds consists of tori , rigid graphs are the ones without good @xmath79-pairs and without @xmath95-pairs of index @xmath126 .
we focus on this case since it contains complements of knots / links in the @xmath0-sphere , which are the most important @xmath0-manifolds with boundary
. table 2 lists in the last two rows the number of non - isomorphic contracted rigid @xmath1-colored graphs representing compact prime orientable @xmath0-manifolds with toric boundary , indicated by @xmath132 , and the special cases of connected toric boundary , indicated by @xmath133 , showing the wide simplification due to the property of rigidity , expecially in the case of connected boundary .
* there exist exactly @xmath134 non - isomorphic contracted bipartite @xmath1-colored graphs ( without @xmath4-dipoles ) of order @xmath5 , representing compact @xmath0-manifolds with ( possibly disconnected ) toric boundary , and @xmath135 of them are rigid .
* there exist exactly @xmath136 non - isomorphic contracted bipartite @xmath1-colored graphs ( without @xmath4-dipoles ) of order @xmath137 , representing compact @xmath0-manifolds with connected toric boundary , and only @xmath0 of them are rigid .
the next theorem summarizes the results of computer classification of prime orientable @xmath0-manifolds with toric boundary represented by @xmath1-colored graphs of order @xmath5 .
* there exist exactly @xmath138 non - homeomorphic compact orientable prime @xmath0-manifolds with ( possibly disconnected ) toric boundary , admitting a representation by a @xmath1-colored graph of order @xmath5 , and exactly @xmath139 of them are complements of links in the @xmath0-sphere ( see table 3 ) .
* there exist exactly @xmath1 non - homeomorphic compact orientable prime @xmath0-manifolds with connected toric boundary , admitting a representation by a @xmath1-colored graph of order @xmath137 , and exactly @xmath4 of them are complements of knots in the @xmath0-sphere ( the trivial and the trefoil knot ) . if @xmath87 is a compact surface with non - empty boundary , @xmath140 and @xmath141 are coprime pairs of integers , with @xmath142 , we define an orientable seifert manifold @xmath143 as follows .
consider the surface @xmath144 obtained from @xmath87 by removing the interiors of @xmath145 disjoint discs .
the boundary circles of these discs are denoted by @xmath146 .
let @xmath147 be all the remaining circles of @xmath148 .
let @xmath149 be the orientable @xmath150-bundle over @xmath144 .
in other words , @xmath151 or @xmath152 , depending on whether @xmath144 is orientable or not .
we choose an orientation of @xmath149 and a section @xmath153 of the projection map @xmath154 .
each torus @xmath155 , for @xmath156 , carries an orientation induced by the orientation of @xmath149 . on each @xmath157
we choose a coordinate system , taking @xmath158 as the meridian @xmath159 and a fiber @xmath160 as the longitude @xmath161 .
the orientations of the coordinate curves must satisfy the following conditions : 1 . if @xmath151 , the orientation of each @xmath161 must be induced by a fixed orientation of @xmath150 . if @xmath152 , then the orientation of each @xmath161 can be chosen arbitrarily .
the intersection number of @xmath159 with @xmath161 in @xmath157 is @xmath22 .
now , for each @xmath162 , let us attach a solid torus @xmath163 to @xmath149 via a homeomorphism @xmath164 taking the meridian @xmath165 of @xmath166 into a curve of type @xmath167 .
the resulting manifold is denoted by @xmath143 , and the cores @xmath168 of the solid tori @xmath166 in @xmath38 are called _ exceptional fibers _ of the seifert manifold .
we emphasize that the remaining boundary tori @xmath157 of @xmath38 , for @xmath169 , still possess coordinate systems @xmath170 .
there are seven seifert manifolds in our census : @xmath171 , @xmath172 , @xmath173 , @xmath174 , @xmath175 , @xmath176 and @xmath177 .
we denote here by @xmath178 and @xmath179 respectively the disc and the mbius strip with @xmath72 holes .
manifolds which can be obtained from seifert manifolds by gluing their boundary tori are known as graph manifolds of waldhausen .
the structure of the seven graph manifolds @xmath180 , @xmath181 , @xmath182 , @xmath183 , @xmath184 , @xmath185 and @xmath186 arising in our census is very simple : each of them is obtained by gluing together either two or three seifert manifolds as follows . 1 .
let @xmath187 be two seifert manifolds with fixed coordinate systems on @xmath188 , where @xmath189 .
choose arbitrary tori @xmath190 and @xmath191 of @xmath192 and @xmath193 , respectively .
we say that a homeomorphism @xmath194 corresponds to a matrix @xmath195 if @xmath196 takes any curve of type @xmath197 on @xmath190 to a curve of type @xmath198 on @xmath191 .
we can then define @xmath199 as @xmath200 .
+ in particular , in our census @xmath38 and @xmath201 can only be either @xmath202 , for @xmath70 , or @xmath203 .
2 . let @xmath204 be three seifert manifolds with fixed coordinate systems on @xmath205 , where @xmath206 .
choose arbitrary tori : @xmath190 of @xmath192 , @xmath207 and @xmath208 of @xmath193 and @xmath209 of @xmath210 .
let @xmath211 , @xmath212 be homeomorphisms corresponding to the matrices @xmath213 , then we can define @xmath214 as @xmath215 .
+ in particular , in our census each one of @xmath216 and @xmath217 is fibered over a disc @xmath178 , for @xmath218 , with @xmath72 holes and it has at most one exceptional fiber .
let @xmath38 be a compact @xmath0-manifold with toric boundary .
we say that @xmath38 is hyperbolic if , after removing the boundary , we get a cusped @xmath0-manifold , i.e. a noncompact hyperbolic manifold @xmath219 of finite volume .
we call a hyperbolic @xmath0-manifold _ platonic _ if it can be decomposed into ideal platonic solids
tetrahedral _ or _ octahedral _
manifold is a hyperbolic platonic manifold made up of the corresponding platonic solids .
the census of tetrahedral manifolds with at most @xmath220 ( orientable case ) and @xmath221 ( non - orientable case ) tetrahedra was provided in @xcite .
the tetrahedral census and a small version of the octahedral census @xcite have been incorporated into snappy @xcite .
there are nine hyperbolic manifolds in our census : @xmath222 , @xmath223 , @xmath224 @xmath225 and @xmath226 .
all of them , except @xmath227 , are contained in snappy . among these manifolds
there are tetrahedral and octahedral ones .
the classification results up to 12 vertices are listed in table 3 .
the last column contains information about knots or links in @xmath230 , if any , whose complements are represented by the corresponding @xmath1-colored graphs ( whose codes are shown in the second column ) . whenever it was possible to find a prime link
whose complement is a given manifold , we indicated it by its name in the hoste - thistlethwaite table .
when no prime link was available we referred to the diagrams of figure [ links2 ] .
surprisingly , up to 12 vertices no complement of knots in the @xmath0-sphere appears , except for the complement of the trivial one ( @xmath171 in table 3 ) , whereas in this context 18 different manifolds are complements of classical links with more than one component .
so we carried on the research for the case of connected toric boundary , in order to find the complement of some non - trivial classical knot . with 14 vertices
there are exactly 13 contracted @xmath1-colored graphs without @xmath4-dipoles representing manifolds with connected toric boundary , but all of them contain good @xmath95-pairs and therefore the represented manifolds are either not prime or just solid tori . with 16 vertices
the number of graphs is 84 , and among them only two are rigid
. the codes of these two graphs are : + dabchefghgfedcbagceabhdf and dabchefghgfedcbagheacbdf , + and they both represent the complement of the trefoil knot @xmath232 .
figure [ fig : trefoil ] depicts the first one of these graphs . _
a construction producing a ( non - contracted ) @xmath1-colored graph representing the complement of a knot / link @xmath233 in @xmath230 , starting from a connected planar diagram of @xmath233 , is presented in @xcite .
however , this construction usually does not provide minimal representations , even after cancelling proper dipoles and switching good @xmath79-pairs .
for example , in the case of the complement of the trefoil ( resp . the hopf link )
a graph with @xmath234 vertices ( resp .
@xmath235 vertices ) is obtained , whereas the minimal one has @xmath236 vertices ( resp .
@xmath59 vertices , as shown in table 3 ) .
_ e. fominykh and v. tarkaev were supported in part by the laboratory of quantum topology , chelyabinsk state university ( contract no .
14.z50.31.0020 ) , the ministry of education and science of the russia ( the state task number 1.1260.2014/k ) and rfbr ( grant number 16 - 01 - 00609 ) . | a representation for compact @xmath0-manifolds with non - empty non - spherical boundary via @xmath1-colored graphs ( i.e. , @xmath1-regular graphs endowed with a proper edge - coloration with four colors ) has been recently introduced by two of the authors , and an initial classification of such manifolds has been obtained up to 8 vertices of the representing graphs .
computer experiments show that the number of graphs / manifolds grows very quickly as the number of vertices increases . as a consequence ,
we have focused on the case of orientable @xmath0-manifolds with toric boundary , which contains the important case of complements of knots and links in the @xmath0-sphere . in this paper
we obtain the complete catalogation / classification of these @xmath0-manifolds up to 12 vertices of the associated graphs , showing the diagrams of the involved knots and links . for the particular case of complements of knots ,
the research has been extended up to 16 vertices .
_ 2010 mathematics subject classification : _
57m25 , 57n10 , 57m15 . _ key words and phrases : _
3-manifolds , colored graphs , knot / link complements . |
simulations of structure formation within the standard cold dark matter cosmology predict that cluster - scale dark matter halos should typically be far from spherical , with shortest to longest axis ratios in density typically around @xmath11 , rarely greater than @xmath12 , and with no strong preference for oblate ( @xmath13 ) or prolate ( @xmath14 ) halos @xcite .
such studies have typically been limited to studying the shapes of halos on cluster - scales because the computational expense of resolving a large enough sample of galaxy - scale halos was prohibitive .
indeed , observations leading to estimates of @xmath15 for a handful of external galaxies proved to be roughly consistent with these findings ( see merrifield 2002 , fig . 3 for a summary ) .
more recently , preliminary analysis of _
galaxy_-scale dark matter halos extracted from cosmological simulations predict that they could be systematically rounder than their larger counterparts , peaking around @xmath16 and with examples as round as @xmath17 @xcite , and @xcite have demonstrated that simulations which include gas cooling effects also produce rounder halos than similar simulations which neglect such dissipation .
this result is especially intriguing given at least one estimate for the shape of the milky way s dark matter halo that suggests that it could indeed be rather spherical ( @xmath18 , * ? ? ?
* ) , placing it at the extreme of the extragalactic @xmath15 range ( but cf . * ? ? ? * who find @xmath19 0.5 ) .
however , this estimate was made using just a few dozen carbon stars thought to be associated with tidal debris from the sagittarius ( sgr ) dwarf galaxy because of their alignment with its orbit on the sky ( velocity measurements and distance estimates could not conclusively support this association in most cases ) , and hence the accuracy of this measurement remains unclear .
debris from the destruction of a satellite galaxy can provide a sensitive probe of deviations of the potential of the milky way from spherical symmetry .
such debris occupies orbits with a range of azimuthal time periods about the satellite s own , and this leads to the phase mixing of debris ahead and behind the satellite along its orbit to form tidal tails @xcite .
@xcite showed that in spherical potentials these tidal tails will gradually thicken within the orbital plane , due to the range of precession rates of turning points in debris orbits .
in non - spherical potentials , the range in precession rates of the orbital poles will lead in addition to thickening of the debris perpendicular to the ( instantaneous ) orbital plane of the satellite .
hence , if the milky way is close to spherical , debris should always remain planar and appear close to a single great circle in the sky @xcite , but if the milky way is non - spherical the debris should , over time , spread to cover a significant fraction of the sky .
such an alignment of a sample of faint , high latitude carbon stars @xcite along a single great circle on the sky previously led @xcite to conclude @xmath20 .
more recently , @xcite has cautioned that , since it can take a few orbits for debris to spread beyond its intrinsic thickness , stars released only a few orbits ago can not be sensitive probes of potential flattening , and hence that none of the sgr data sets thus far detected provide any significant constraint on @xmath21 .
in addition to helmi s ( 2004a ) concern , the thickness of a debris stream will also increase with satellite mass .
hence , if either the current mass or recent mass loss rate is poorly known these introduce additional uncertainties in trying to measure @xmath21 from the thickness of a stream . as a final concern ( and one that is much harder to correct for ) thickening can also result from interactions with other massive
and possibly invisible lumps in the halo @xcite . in this paper
we revisit the question of what constraint tidal tail stars associated with the sgr dwarf can place on the flattening of the galactic potential .
our approach differs significantly from previous work in that we use the orbit - alignment rather than thickening of the debris to measure the halo shape . for our analysis , we use m giants selected from the two micron all sky survey ( majewski et al , 2003 hereafter paper i ) . the sgr m giants are thought to have ages of 2 - 3 gyr ( paper i ) and hence , the _ thickness _ of tidal streams constituted of these stars can not be expected to provide a strong constraint on @xmath21 ( even ignoring the additional concerns about this method we raised above ) .
however , the approximate distances estimated from the apparent magnitude of the m giants permit a clear separation between stars leading or trailing sgr along its orbit ( because continuous streams of debris can be traced directly back to the sgr core ) and , moreover , yield a detailed view of the three - dimensional configuration of these tidal tails ( e.g. , paper i ) . as a consequence , rather than simply looking at the thickness of debris projected on the sky , we are able to measure the poles of the best - fit great circles ( in two dimensions ) or best - fit planes ( in three dimensions ) of the leading and trailing debris separately .
we use the difference in orbital poles between the two debris trails to quantify the precession of the orbital plane over this range of orbital phase .
the advantage of this powerful approach is that what we are measuring is sensitive to @xmath21 alone the thickness of the debris due to parent satellite mass , debris age , or scattering off of lumps in the halo potential do not have systematic effects in this signal and merely contribute to the uncertainty in our measurement . to determine the expected variation in orbital poles for different values of @xmath22 , we run a series of n - body simulations that mimic the sgr system as traced by m giants .
we describe our numerical simulations in 2 , compare them to the m giant data set in 3 and summarize our results and discuss future prospects for this approach in 4 .
our simulation technique closely follows that outlined in @xcite .
the milky way is represented by a smooth , rigid potential , and sgr by a collection of @xmath23 self - gravitating particles whose mutual interactions are calculated using a self - consistent field code @xcite . as a starting point for our study
we use the results of law , johnston , & majewski ( 2004 hereafter referred to as paper iv ; preliminary results were presented in law et al .
2004 ) , who perform simulations of satellite disruption along orbits that are consistent with sgr s current position , line - of - sight velocity and direction of motion tangential to the line - of - sight ( deduced from the orientation of the m giant plane ) .
the mass and amplitude of tangential motion ( @xmath24 ) of sgr and the potential of the milky way are systematically varied to find the closest possible fit ( hereafter referred to as the default model ) to the angular position , distance , and available line - of - sight velocity data for the 2mass m giants along the leading / trailing streams .
the default model is run in a three - component model for the galactic potential , consisting of a miyamoto - nagai ( 1975 ) disk , hernquist spheroid , and a logarithmic halo : @xmath25 @xmath26 @xmath27 where @xmath28 1 , @xmath29 @xmath30 , @xmath31 @xmath30 , @xmath32 km s@xmath33 , @xmath34 kpc , @xmath35 kpc , @xmath36 kpc , and @xmath37 kpc and @xmath38 ( i.e. the halo component is spherical in the default model ) .
the distance of the sun from the galactic center is taken to be @xmath39 7 kpc .
initially , the particles in the sgr model are distributed according to a plummer ( 1911 ) model @xmath40 where @xmath41 @xmath42 @xmath30 is the initial mass of sgr and @xmath43 0.82 kpc is its scale length .
the satellite is allowed to evolve over five pericentric passages along an orbit with a pericenter of 14 kpc , an apocenter of 58 kpc , and a radial orbital time period of 0.9 gyrs .
( the exact extent of the orbit depends on the distance scale adopted for the m giants , which is currently set by assuming a 24 kpc heliocentric distance for the sgr core . ) at the point matching sgr s current position and line - of - sight velocity , the satellite in the default model has a bound mass of @xmath44 . in paper
iv we find that the data still allow some freedom in the exact form for the galactic potential , but that within a given potential the final bound mass of sgr is constrained to within a factor of two , and , once this is known , @xmath24 can be found to within a few km s@xmath33 .
paper iv , in effect , finds the best - fitting set of parameters that describe the configuration of sgr debris _ projected onto _ the instantaneous orbital plane , whereas the current discussion focuses on the variations in the position of the orbital plane itself . in this paper , we rerun paper iv s default model with a variety of halo flattenings ranging from @xmath45 0.8 to 1.45 ( i.e. , we consider both oblate and prolate potentials ) . since we are interested in constraining the contours of the potential , we flatten these directly via parameter @xmath21 in equation ( [ haloeqn ] ) rather than generating them from a flattened density distribution .
these oblate / prolate potentials correspond to even greater oblateness / prolateness ( e.g. fig . 1 in helmi 2004a ) of the isodensity contours in the effective distance range that sgr s orbit explores .
note that even our models with spherical halos have modestly flattened potential contours due to the contribution of the galactic disk .
we recalculate the values of the galactic halo scalelength @xmath46 and @xmath24 for each value of @xmath21 considered to obtain the best possible fit to the m giant distance and line - of - sight velocity data within each model of the galactic potential ( see paper iv for a description of this fitting process ) .
the calculated values for each choice of @xmath21 are given in table 1 .
note that the orbital characteristics of the model dwarf will vary slightly for each of these models , and therefore the mass - loss history of the dwarf will depend on @xmath21 .
however , we do not revisit the question of the best - fit mass for sgr in each of these potentials because we are only concerned with the differential precession of the poles of the tidal debris , which is independent of satellite mass . throughout the paper
we deliberately compare the m giant data with our simulations as viewed from the sun , despite the fact that orbital precession is most naturally discussed in reference to the galactic center .
this is because transforming the data to a galactocentric viewpoint would introduce distortions due to uncertainties in the distances to both the m giants and to the galactic center that might be confused with orbital precession .
in contrast , our heliocentric viewpoint of the simulations is fixed by only requiring that our simulated sgr has the correct current position and velocity relative to the sun .
moreover , we prefer a heliocentric system because is the natural system for the great circle cell count analysis we perform in 3.2 .
in 3.1 we present a qualitative assessment of aitoff projections of sgr debris , and demonstrate similarly to previous discussions ( e.g. helmi 2004a ) that such qualitative analyses provide little information on halo flattening .
in contrast , we adopt two quantitative approaches to assess the degree of orbital plane precession of the sgr system .
the first method ( 3.2 ) relies on careful measurements of the mean great circle described by the @xmath47 distributions of m giants in the leading and trailing tidal tails . as applied here , this method is virtually free of observational errors or errors in derived quantities , particularly the derived distances of the m giants . in the second method ( 3.3 )
we analyze the orbital plane variation using the derived three dimensional distribution of the sgr m giants
. both techniques advance previous discussions of the qualitative appearance of sgr debris in aitoff projections to show that a slightly oblate milky way potential is preferred .
figure [ xyplot ] shows the projection of the final positions of particles in the default simulation onto sgr s orbital plane , and demonstrates the @xmath48 coordinate system of paper i ; @xmath48 is the angular distance from sgr along the plane of its orbit , defined to be zero at the core and increase in the trailing direction .
the colors represent different debris `` eras '' , i.e. orbits ( denoted as one apogalacticon to the next apogalacticon ) on which the debris was stripped from the satellite yellow for particles lost since apogalacticon about 0.4 gyr ago , magenta , cyan , and green for particles stripped from the dwarf two , three , and four orbits ago , respectively ( the orbital period of sgr in all of our model galactic potentials is about 0.75 - 1.0 gyr ) . in paper
iv we find that the positions and available velocities for sgr m giants presented in paper i are consistent with debris younger than about 1.9 gyr ( i.e. yellow , magenta or cyan ) in these simulations .
this age estimate is consistent with the expected 2 - 3 gyr lifetimes of these stars for the m giants to be a significant contributor to debris older than about 1.9 gyr the stars would have to be stripped from the satellite less than one orbital period following their birth ( see paper i for a fuller discussion of this problem ) .
figure [ aitoffplot ] plots aitoff projections of the final positions of particles in both spherical and moderately flattened halos ( @xmath38 and @xmath49 )
. visual inspection of this figure confirms the @xcite conjecture that _ qualitative _ differences in the apparent thickening or precession of the most recent debris ( i.e. yellow , magenta , and cyan points ) are fairly small over the range of potential flattenings shown , especially for younger debris .
however , it is premature to conclude that these effects do not lead to measurable variations in the disposition of the sgr debris .
we now demonstrate that _ orbital precession _ of even the most recent sgr debris can be measured and used to discriminate between halo flattening models . in their current configuration about the milky way , sgr leading arm m giants predominate in the northern galactic hemisphere while trailing debris m giants dominate the south .
paper i demonstrated ( see their fig .
6 ) clear differences in the location of the sgr m giant great circle cell count @xcite[gc3 " ] peaks when the m giant sample is limited to @xmath50 versus @xmath51 subsamples , and suggested that the differences could relate to a combination of precessional and parallax differences between leading and trailing arm debris .
we repeat the paper i gc3 analysis on our simulations ( using the yellow , magenta , and cyan debris that best represent the paper i m giant data ) to determine the degree of precessional shift as a function of @xmath21 . figure [ gc3plot ] compares the shape and positions of the gc3 peaks of the models to those of the m giant data , divided into @xmath50 ( north / mostly leading arm ) and @xmath51 ( south / mostly trailing arm ) subsamples .
other aspects of sample selection , including limiting to stars with distances between 13 and 65 kpc and ignoring the region around the magellanic clouds ( @xmath52 and @xmath53 ) were utilized to exactly match the selection critria used to generate the gc3 plots in paper i ( their fig .
the 13 - 65 kpc distance range limit serves to highlight the primary distance range for the debris corresponding to the yellow , magenta and cyan parts of the leading and trailing arms . for the observations
, this limit also removes nearby disk m giants and more distant m giants " at magnitudes where 2mass photometry becomes less reliable , whereas for both the observations and the models it limits the effects of overlapping contaminating debris on each sgr arm from extensions of the opposite sgr arms .
the shapes of gc3 peaks reflect departures from great circle symmetry _ as viewed from the sun _ , while the size ( e.g. , fwhm ) of the peak is a function of the debris width on the sky , convolved with the cell size ( here adopted as a 5@xmath54 wide cell wide cells presented here , albeit with poorer signal - to - noise because of the fewer numbers of stars that fall within the narrower cells . ] ) .
the first thing to notice is the overall consistency of the simulated and actual m giant data in both the shape and size of the gc3 peaks for _ both _ hemispheres .
the detailed matches are encouraging support for the conjecture that the simulations provide rather accurate representations of the actual sgr m giant distribution .
the second thing to note is that while there is virtually no difference in the location of the gc3 peaks for the southern ( trailing arm ) data as a function of @xmath21 , the position of the gc3 peak for northern ( leading arm ) data is , in contrast , highly and systematically sensitive to @xmath21 .
the effect is more pronounced in the north because ( a ) the leading arm extends very close to ( virtually on top of ) the sun and a given precessional shift relative to the galactic center will be foreshortened to a larger angular shift on the sky for closer debris , and ( b ) the northern sample utilized includes a larger fraction of _ older _ debris , torn off from sgr one orbit earlier ( i.e. , the cyan debris in fig . 1 ) , and these very stars , which will have experienced more precession , constitute much of the sgr debris closest to the sun in the north .
note that the simulations were not adjusted to create a best match to the gc3 poles in figure [ gc3plot ] , but rather the orbit of sgr in the simulations was set by the mean pole ( [ @xmath55 = [ 273.75,-13.46]@xmath54 ) found in the paper i single plane fit to the cartesian positions of both leading and trailing m giant debris ; the close match to the general positions of the separated north / south gc3 poles arises naturally from the evolution of the debris .
the poles in figure [ gc3plot ] show the simulations with @xmath56 and @xmath57 to be the closest match to both the absolute positions of the northern hemisphere gc3 peak as well as the relative difference in north / south gc3 peak positions in the m giant data .
more oblate and all prolate models yield northern gc3 peaks that are tens of degrees off from the observed positions .
the gc3 analysis shows that subtle variation in the @xmath47 distribution of debris ( fig .
2 ) _ can _ be measured and , in a way virtually free of observational bias , be used to constrain @xmath22 .
the gc3 analysis gives a phase - averaged view of the expected differential precession of sgr leading and trailing debris .
in contrast , the lines in figure [ poleplot ] show the pattern that the _ instantaneous _ pole of sgr s orbit traces on the sky within @xmath01.5 gyrs of sgr s current position , as viewed from the sun . the upper / lower panels show the pole evolution in potentials with oblate / prolate halo components respectively , with the solid / dotted portions of each line corresponding to portions of the orbit trailing / leading sgr .
the black triangle shows the best fit pole of the 2mass data set ( paper i ) , which the orbits were constrained to go through at sgr s current position .
the dots along each curve in figure [ poleplot ] indicate the pole position at 0.2 gyr intervals both leading and trailing sgr s current location .
the m giants discussed in paper i explore orbital phases corresponding to up to ( roughly ) + 0.6 gyrs and -0.4 gyrs from sgr s current position , ahead and behind it in time along the orbit .
( note that the debris age i.e. the time since stars were lost from sgr is much greater than this . ) hence , the poles actually measured from the m giant and simulation data can be thought of as averages of 3 points along the dotted curves and 2 points along the solid curves , weighted by the density of stars or particles along the streamers .
( note that , for fixed flattening , the pole paths of orbits with @xmath24 varying from the adopted value by as much as @xmath010 km s@xmath33 more than the maximum range permitted given constraints from the m - giant distances and velocities , as shown in paper iv were found to be virtually indistinguishable from the orbits illustrated . ) .
this figure demonstrates ( as already seen in fig .
[ gc3plot ] ) that : ( i ) in general , the separation of poles derived for the leading and trailing simulated data should increase with increasing deviations from @xmath38 ; ( ii ) pole separation should be more dramatic for the oblate cases than the prolate cases ; ( iii ) the sense of precession in prolate potentials is opposite to that in oblate potentials ; and ( iv ) some ( oblate - like ) precession should be present even in the @xmath38 case because of the presence of a disk component in our galactic potential .
we might expect the lines in figure [ poleplot ] to move monotonically in galactic longitude .
the retrograde - looping evolution of the leading ( dotted ) portion in the oblate cases is due to our perspective of these orbits , which have pericenters that lie only just outside the solar circle . indeed , some asymmetry between the evolution of the poles of leading and trailing debris is already apparent in the more dramatic shift of the gc3 peaks of the northern ( leading ) debris when compared to the southern ( trailing ) , as noted in the previous section ( fig .
[ gc3plot ] ) . however , the figure [ gc3plot ] gc3 analysis is a somewhat blunt tool for assessing the precessional shifts because it does not take advantage of all available information namely , the distances to the stars / particles . to quantify the degree of precession exhibited by sgr debris better ,
the poles of best - fit orbital planes can be determined for both the m giant and the simulated data by plane - fitting to the leading and trailing arm debris in three - dimensional space .
we constrain these planes to pass through the solar position to allow us to quantify the effect precession has on sgr tidal debris as projected on the sky ( the natural observational regime , e.g. , as used in a gc3 analysis ) .
this constraint means that the poles derived by this method should be interpreted only as tools for tracing the amount of precession in a given model of the galactic potential ; they are not the true galactocentric orbital poles and do not precisely represent the angular momentum vector of the debris .
the best - fit planes to the m giant and simulation distributions were found by minimizing the @xmath58 distribution of point distances from the plane , applying an iterative 2.5-@xmath8 rejection algorithm .
confidence limits on the corresponding poles were determined by statistical analysis of synthetic data sets generated using a bootstrapped monte carlo technique , with the confidence ellipse resulting from this analysis projected into error bars at the 68% confidence level in @xmath59 and @xmath60 .
the open symbols in figure [ planeplot ] show the results of this precession analysis applied to data points in the leading tail in the range @xmath61 - @xmath62 , while the filled symbols show the results for the trailing tail in the range @xmath63 -@xmath64 , @xmath65 , @xmath66 kpc , @xmath67 , and distances 13 kpc @xmath68 kpc .
see paper i for definitions of these criteria .
the magellanic clouds were removed from this data set using the @xmath47 cuts given in paper i. for the simulations , yellow , magenta , and cyan debris were considered and subjected to the same cuts in @xmath48 , @xmath69 , and @xmath60 as the m giants . ] see figure [ xyplot ] .
the colored triangles / squares correspond to simulated data run in prolate / oblate potentials .
( the error bars on the simulated points are comparable to the symbol sizes and are omitted for clarity . ) note there is a systematic offset in @xmath47 of order a few degrees between the results from the observed data and the trend of the simulated data .
this can be attributed to our assumption that the pole derived from the full 2mass data corresponds to sgr s present orbital pole when , in fact , it merely represents an average along tidal debris with a small range of orbital phases .
nevertheless it is clear that : ( i ) the _ sense _ of precession in the data strongly favors oblate potentials ; and ( ii ) the evolution of pole _
differences _ in the plot implies that the precession rate is most consistent with our slightly flattened halo models in which @xmath70 .
table 1 emphasizes this result by quoting the separation of the leading / trailing poles for the simulations and the m giants .
flattenings within the range @xmath71 ( @xmath72 ) lie within the 1.5-@xmath8 error bars on the data , while those with @xmath73 and @xmath74 are ruled out at the 3-@xmath8 level .
more extreme values of @xmath9 ( @xmath5 0.6 ) and @xmath10 1.25 ( @xmath7 1.6 ) are ruled out at the 7-@xmath8 and 5-@xmath8 levels respectively .
the open / filled circles in figure [ planeplot ] show the results of the same analysis performed on the older ( color coded green in figs .
[ xyplot ] and [ aitoffplot ] ) simulated data at larger separations along the orbit ( @xmath75 for the leading debris and @xmath76 for the trailing see fig .
[ xyplot ] ) for the @xmath45 0.90 simulation , which best reproduces the poles of the younger debris .
the positions of the circles are suggestive of the evolution traced by the orbital poles in figure [ poleplot ] .
note that the m giants discussed in paper i are primarily in the yellow , magenta and cyan portions of the tidal debris ( see paper iv ) and there is no clear evidence yet for m giants in the portions of the tails corresponding to the green debris ; the phase - mixing time for debris to reach these points is comparable to the stellar evolution ages of these stars ( see 3.1 ) .
the circles in figure [ planeplot ] demonstrate that an analysis such as that conducted here will provide a more powerful constraint on @xmath21 if tracers of this older debris can be found , so long as leading debris , which will be mainly in the _ southern _ hemisphere for green debris " , can clearly be separated from the _ trailing _ debris there using distance and/or velocity information .
indeed , at these larger phase - differences from sgr , the leading / trailing poles differ by more than 45 degrees even for flattenings of @xmath77 .
this suggests that the accurate determination of the centroid of an older piece of the sgr tidal stream , even at a single longitudinal point , can provide strong leverage on @xmath21 via comparison with test - particle integrations in different potentials .
we have shown that the precession of sgr s orbit can be accurately traced by stars in its debris streams even the relatively recently released m giant stars and hence be used to constrain the flattening of the galactic potential .
a difference of @xmath78 degrees is found between the best - fit orbital poles for stars in sgr s leading and trailing streams ( in the sense that the angular momentum vector _ increases _ in @xmath59 and _ decreases _ in @xmath60 from trailing to leading debris ) when those poles are measured from debris at azimuthal orbital phases within + 140/-145 degrees along its orbit as viewed from the sun .
such a low amount of precession is most consistent ( within 1.5-@xmath8 ) with simulations of the destruction of sgr run in models of the galaxy with a slightly flattened halo where @xmath21 is in the range 0.90 - 0.95 ( @xmath79 0.83 - 0.92 ) .
flattenings for the halo potential of @xmath49 ( @xmath80 ) or less and @xmath81 ( @xmath82 ) or more are ruled out at the 3-@xmath8 level , and oblate models are strongly preferred over prolate models .
note that these results depend on the assumed form of the potential of the galactic disk .
recently , helmi ( 2004b ) has suggested that prolate milky way potentials offer a means by which to solve a dilemma discussed in paper iv namely that the sgr leading arm radial velocities are one discrepant observable not well fit by oblate milky way + sgr models that provide compelling overall fits to all other available observables ( see paper iv ) . while a prolate model solves this one problem , as shown here
prolate models also introduce a serious discordance with the observed m giant precession ; indeed , prolate models are found to induce precession in the _ opposite _ direction to that observed . since orbital pole precession is almost solely sensitive to the shape of the potential , whereas a variety of other effects in addition to the halo shape e.g. , evolution in the sgr orbit and/or the strength of the galactic potential can conceiveably alter the dynamics of debris _ within _ the orbital plane , we are inclined to the simpler explanation of an oblate potential to match the precessional data while admitting the need for yet more sophisticated models to resolve the problem with the leading arm velocities using these other free parameters . future detections of older sgr debris at larger phase - differences along its orbit will provide even stronger orbital precession constraints than obtained here .
recently , @xcite announced a new detection of sgr debris as an over - density of a - colored stars in the sloan digital sky survey at sufficient angular separation from sgr ( @xmath83 ) and distance from the sun ( @xmath84 ) kpc to possibly associate it with the oldest ( green ) debris in our simulations .
unfortunately , the debris has not yet been mapped fully enough for a determination of its mean position ; when done , however , a more accurate determination of @xmath21 may be possible .
thus , sgr has still more to contribute to our understanding of the galactic potential .
srm acknowledges support from a space interferometry mission key project grant , nasa / jpl contract 1228235 , nsf grant ast-0307851 , a david and lucile packard foundation fellowship , and the f.h .
levinson fund of the peninsula community foundation .
kvj s contribution was supported through nasa grant nag5-9064.and nsf career award ast-0133617 .
.orbital pole precession of tidal debris between @xmath85 ( trailing debris ) and @xmath86 ( leading debris ) for sgr m giants and simulations with indicated values of the galactic halo flattening @xmath21 .
positive separation values indicate simulations whose leading debris orbital pole is at higher / lower l / b than the trailing debris pole , and negative values indicate simulations whose leading debris pole is at lower / higher l / b than the trailing debris pole ( i.e. in the direction opposite to that observed for sgr m giants , see fig .
the values assumed for the halo scale length ( @xmath46 ) and the velocity of the model dwarf tangential to the line of sight ( @xmath24 ) in each model are also given . [ cols="^,^,^,^ " , ] | m giants selected from the two micron all sky survey ( 2mass ) have been used to trace streams of tidal debris apparently associated with the sagittarius dwarf spheroidal galaxy ( sgr ) that entirely encircle the galaxy . while the sgr m giants are generally aligned with a single great circle on the sky , we measure a difference of 10.4 @xmath0 2.6 degrees between the mean orbital poles of the great circles that best fit debris leading and trailing sgr , which can be attributed to the precession of sgr s orbit over the range of phases explored by the data set
. simulations of the destruction of sgr in potentials containing bulge , disk and halo components best reproduce this level of precession along the same range of orbital phases if the potential contours of the halo are only slightly flattened , with the ratio between the axis length perpendicular to and in the disk in the range @xmath1 ( corresponding to isodensity contours with @xmath2 0.83 - 0.92 ) .
oblate halos are strongly preferred over prolate ( @xmath3 ) halos , and flattenings in the potential of @xmath4 ( @xmath5 0.75 ) and @xmath6 ( @xmath7 1.1 ) are ruled out at the 3-@xmath8 level .
more extreme values of @xmath9 ( @xmath5 0.6 ) and @xmath10 1.25 ( @xmath7 1.6 ) are ruled out at the 7-@xmath8 and 5-@xmath8 levels respectively .
these constraints will improve as debris with larger separation in orbital phase can be found . |
currently , one of the biggest worries of our society is the future of the climate .
common belief is that our planet is heating up at an accelerated rate , caused by the rapid increase in carbon dioxide ( co@xmath0 ) concentration in the atmosphere , henceforth called [ co@xmath0 ] .
this increased carbon dioxide finds its origin in human activity ; humans burn fossil fuels , thereby injecting large quantities of carbon into the troposphere by converting it into co@xmath0 .
the co@xmath0 contributes to the greenhouse effect of our atmosphere and it is believed that the anthropogenic co@xmath0 will heat by up the planet by up to six degrees during this century ( page 45 of ipcc 2007 report@xcite ) . here
we will analyze these ideas and come up with some remarkable conclusions .
for that , while the subject is the atmosphere , we do not have to go into much detail of atmospheric science .
there are ( nearly philosophical ) observations one can make about climate systems , even without going into technical details .
they are in the realm of signal processing and feedback theory .
the model of anthropogenic global warming ( agw ) stands or falls with the idea that temperature is strongly correlated to [ co@xmath0 ] by the so - called greenhouse effect .
serious doubt is immediately found by anybody analyzing the data .
the contribution of co@xmath0 to the greenhouse effect can easily be estimated to be about 3.612%@xcite .
the total greenhouse effect is also well known ; without our atmosphere our planet would be 32 degrees colder .
this makes the co@xmath0 greenhouse effect only 1 kelvin in a simple analysis .
we arrive at a similar value if we use statistics and do a linear regression on contemporary [ co@xmath0 ] and temperature data , the maximum of the effect we can thus expect in a linear model when doubling the concentration artificially by burning up fossil fuels@xcite .
this is far below the global warming models even if we were to use a linear model .
it is however , unlikely that the effects are linear .
the system is more likely to be sublinear .
that is because the greenhouse effect is governed by absorption of light which unavoidably follows the beer - lambert law : the absorption is highly sublinear ; twice as much co@xmath0 will not cause twice as much absorption .
the classical arrhenius greenhouse law states that the forcing is logarithmic .
yet , later models incorporating non - linear positive - feedback effects as proposed by many climate scientists do predict a super - linear behavior and come up with an estimate of between 1.1 and 6.4 degrees heating for the next century as caused by our carbon dioxide injection into the atmosphere@xcite .
the positive feedback can come from secondary effects such as an increase in water in the atmosphere , a strong greenhouse agent , or a co@xmath0-degassing of ground in the permafrost regions when these thaw .
climate scientists are basing these conclusions mainly on research of the so - called finite - elements type , dividing the system in cells that interact , the same way the weather is studied .
such systems are complicated , but by tuning the processes and parameters that are part of the simulations they manage to explain the actual climate data to an impressive accuracy , as evidenced by the quality of pictures presented in the official climate reports , see for example the ipcc 2007 report where simulation and reality are as good as indistinguishable@xcite , and , moreover , alarmingly , they conclude that the recent rise in temperature can only be attributed to co@xmath0 .
but , from a philosophical point of view , the fact that the past was explained very accurately does not guarantee the same quality for the prediction of the future .
the climate system is chaotic .
small deviations in parameters and initial conditions or assumptions made in the simulations can cause huge changes in the outcome .
this is easily explained in an example from electronics .
if we have a chaotic circuit with , for instance , critical feedback , we can go to our spice or cadence simulator and find the parameters of our components that exactly explain the behavior of our circuit .
so far so good .
the problem is that if we now go back and switch on the same circuit , we will get a different result .
( just take an operational amplifier with 100% positive feedback , it can saturate at the output at the positive as easily as the negative supply voltage , either one can be simulated ) .
an additional problem is that even the parameters themselves are not constant and seem to change without any apparent reason , for instance the el nio phenomena in the climate .
this is one of the reasons electronic engineers talk about phase margins , the zone in nyquist plots , real vs. imaginary parts of gain , that should be avoided because the circuit will become unpredictable even if it is perfectly simulatable .
in fact , recent temperature data fall way out of the prediction margins of earlier models . in view of the discussion above
, this does not come as a surprise . where extrapolation from the 2007 ipcc report predicted 2011 to be a year with an anomaly of close to one degree ( 0.95 @xmath2c is our personal estimate based on fig .
2.5 of the ipcc 2007 report ) , in reality the anomaly is closer to zero .
since 1998 , the hottest year in recent history , the planet has actually been cooling , something that was not foreseen by the predictions of 2007 where a continuing exponential increase in temperature was forecasted by the then generally accepted model .
the scientific community is now going back to their drawing boards and fine - tunes its models to new perfection and manages to simulate the new data as well .
this is a bayesian way of doing science and is significantly less reliable .
the correctness of this statement is evidenced by the fact that there now apparently exist many models that explain the data up to a certain point in time ; every correction of the model that is still consistent with earlier data proves this .
apparently , there are a manifold of models that can explain certain data quite satisfactorily ( but that diverge for future predictions ) . in view of this
, one should be reluctant in making strong claims about the correctness of the latest model . just like in the weather , where the same simulation - evaluation techniques are used ,
we can only hope to get the predictions reasonably under control after thousands of iterations between predictions and reality .
each iteration takes about the amount of time as the prediction span one week with the weather , 30 years with the climate . honestly speaking , before we get it right
, it ll take at least some hundreds of centuries if we uniquely use the approach of finite - elements calculations on supercomputers . in the meantime
, we should not see any climate models as proven indisputable facts .
a skeptic approach to any scientific model is not an illness , it is an essential ingredient in science .
theories are correct until proven wrong . ideas that stand up to scrutiny are more likely to be correct than ideas one is supposed to not question . still , undeniably , a strong correlation is found between the co@xmath0 concentrations and the temperatures as measured by gas - analysis in drillings in ice shelves , see for example the data of the pangaea project indicating that one is the function of the other for the past hundreds of thousands of years@xcite . that is a very strong point .
however , proving only statistical correlation , it is not clear from these data which one comes first .
are temperature variations the result of [ co@xmath0 ] variations , or vice versa ? while the data are consistent with the model of agw they can not serve as proof of these models .
in fact , upon closer scrutiny , the temperature always seems to be _
ahead _ of co@xmath0 variations . see figure [ fig : algore ] , where a detail of the temperature and [ co@xmath0 ] history as measured by ice - trapped gases is plotted , picturing the most blatant example of this effect . a simulation ( dashed line ) is also shown with an exponential - decay convolution of 15 kyr , quite adequately reproducing the results .
indermhle and coworkers@xcite made a statistical analysis and find a value of 900 yr for the delay and note that `` this value is roughly in agreement with findings by fischer et al . who reported a time lag of co@xmath0 to the vostok temperature of ( 600 @xmath3 400 ) yr during early deglacial changes in the last 3 transitions glacial - interglacial''@xcite .
this is inexplicable in the framework of global warming models and we honestly start having some legitimate doubts .
the apparent time lag may possibly be due to a calibration problem of the measurements , and indeed corrections have been made to the data since then , to make [ co@xmath0 ] variations and temperature variations coincide . while these corrections are the result of circular reasoning , where the magnitude is found by modeling the behavior of ice based on climate models and the climate models based on the ice behavior , these corrections are not even sufficient to remove our doubts . if the correlations are true and we continue to claim that temperature variations are the result of [ co@xmath0 ] variations , something is still not correct .
the vostok data of figure [ fig : algore ] show a sensitivity of 10 degrees for 50 ppm [ co@xmath0 ] .
contemporary [ co@xmath0 ] are of the order of 80 ppm rise from the preindustrial value .
we are thus in for a 16-degree temperature - rise . the fact that we did not reach that level means that either co@xmath0 is not climate forcing , or that there is a delay between [ co@xmath0 ] variations ( cause ) and temperature variations ( effect ) . to get a rough idea of the magnitude of this delay , in 25 years
, only 2.5% ( 0.4 of 16 degrees ) of this rise occurred .
the relaxation time is thus ( 25 years)/@xmath4 , which is about 1000 years .
these are back - of - the - envelope calculations
any real values used for the calculation could anyway be debated by anybody . yet
, the outcome will always be more or less this order of magnitude .
in other words , either the vostok plots should show a delay between [ co@xmath0 ] and @xmath5 of the order of 1000 years , or the carbon dioxide is not climate forcing .
the data , however , show a delay of @xmath6 years@xcite or zero , the latter value resulting from questionable corrections . as far as we know
, no correction was proposed to result in the + 1000 yr delay necessary to explain contemporary behavior . what is more , modern correlation figures such as given in fig .
[ fig : algore ] also include methane ch@xmath7 ( available at noaa paleoclimatology@xcite ) and , remarkably , this methane shows the same correlation with [ co@xmath0 ] and @xmath5 .
this leaves us flabbergasted .
we know that methane is also ( assumed to be ) a strong climate - forcing greenhouse agent .
the enigma is then , how did the information from the [ co@xmath0 ] variations percolate to [ ch@xmath7 ] variations ? was this information from [ co@xmath0 ] transmitted to the methane through the temperature variations ?
in other words , [ ch@xmath7 ] variations are the _ result _ of @xmath5 variations , rather than their cause
? then we may equally assume that [ co@xmath0 ] variations are the _ effect _ of @xmath5 variations rather than their _
cause_. there are several mechanisms that may explain such an inverse phase relation , such as outgassing of co@xmath0 ( and ch@xmath7 ) from the warming oceans and thawing permafrost , the correlation between [ co@xmath0 ] and [ ch@xmath7 ] then stems from a common underlying cause .
if that is the case , artificially changing the co@xmath0 in the atmosphere will not change the temperature of our planet , just like heating up a can of soda will liberate the gases contained therein into the atmosphere , while increasing the concentrations of gases above the can of soda will not raise its temperature .
this unidirectional relation between temperature and gas concentrations is what is called henry s law ; the ratio of concentrations of gas dissolved in the liquid and mixed in the air above it in equilibrium is a parameter that depends on temperature .
al - anezi and coworkers have studied this effect in more detail in a laboratory setup under various conditions of salinity and pressure , etc . for co@xmath0 in and above water
an increase in temperature will cause outgassing with a proportionality that is consistent with the correlation found by the historic correlations of global temperature and co@xmath0 in the atmosphere .
also , fischer and coworkers find the delay of [ co@xmath0 ] relative to @xmath5 , as discussed above , likely caused by this ocean outgassing effects@xcite and find that at colder times , the delay is longer , which is itself consistent with arrhenius - like behavior of thermally - activated processes , such as most in nature . in the presence of an alternative explanation ,
there is room for doubt in the agw ideas that increased [ co@xmath0 ] will cause an increased temperature .
inspired by this uncertainty in the ( anthropogenic ) global warming model , we tried to see if we can find more evidence for this failure of the cause - and - effect idea .
we looked at the recent historic climate data ( from just before the agw model prevalence ) and meticulously - measured [ co@xmath0 ] data and came to the same conclusion , as we will present here .
we started with the data of a climate report from before the global warming claims .
we deem these data more reliable since they were for sure not produced under the tutelage of a political committee ( ipcc ) .
at least we are more convinced about the neutrality of the scientists reporting these data .
moreover , the work contains all the useful data and are even available on - line ; the ideas presented here do not need recent data and thus we refrained from looking at them altogether .
the authors of the work , balling and coworkers@xcite analyzed the global warming ( without capitals because it is not the name of a model ) and concluded `` our analysis reveal a statistically significant warming of approximately 0.5 @xmath9c over the period 1751 to 1995 .
the period of most rapid warming in europe occurred between 1890 and 1950 , ... no warming was observed in the most recent half century '' .
note that at the onset of the global warming ideas , no warming was observed that can be correlated to the ( accelerated ) increase of [ co@xmath0 ] .
note also that since 1998 it has not warmed up at all , as confirmed by satellite data ( 1998 was the warmest year)@xcite , in spite of the continuing exponential increase in atmospheric co@xmath0@xcite .
the temperature seems to be unaffected by the anthropogenic co@xmath0 .
balling and coworkers then went on to analyze the increase in temperature as a function of the time of the year for the data between 1851 and 1991 .
they calculated for each of the twelve months the increase in temperature .
they found a distribution as given in figure [ fig : year ] ( open circles ) .
+ this figure based on the data of balling is again remarkable .
the first thing we note is that , while there has been an average of warming , this is not spread equally over the year .
in fact , summer months have become cooler . without knowing the underlying reason , this is remarkable , since [ co@xmath0 ]
has increased in all months .
there are seasonal fluctuations of the co@xmath0 concentrations , see the black dots which represent the monthly [ co@xmath0 ] fluctuations relative to the yearly average at the mauna loa site ( source : noaa , visited 2008@xcite ) .
these rapid fluctuations are mainly attributed to biological activity ( the northern hemisphere has more land and in colder times - in winter - more plants are converted into co@xmath0 and in warmer times - in summer - more photosynthesis takes place converting co@xmath0 into biomass , i.e. , [ co@xmath0 ] is a natural function of temperature ) .
part of the fluctuations , however , are attributed to human activity ( in winter the northern hemisphere - where more people live - is cold and humans thus burn more fuel to warm their houses , i.e. , [ co@xmath0 ] is a function of temperature ) . as a side note , these two things show us that it is very straightforward to understand how [ co@xmath0 ] can be a function of temperature , in these cases through biological activity , including that of humans , in this case resulting in a rapid inverse proportionality ( warmer @xmath10 less co@xmath0 ) .
other , long - term processes such as degassing of oceans can have opposite effects , i.e. , warmer @xmath10 more co@xmath0 .
while we bear this in mind , we will continue the reasoning of anthropogenic global warming and assume an opposite correlation , that is , temperature is a function of [ co@xmath0 ] , and analyze the oscillations .
we will show that this assumption is inconsistent with the data .
while the natural oscillations have always existed and thus do nt result in seasonal oscillations of global warming , the human - caused fluctuations should be represented in the temperature fluctuations .
what we would expect in the framework of agw is that all months have warmed up ( because of general injection of anthropogenic co@xmath0 into the atmosphere ) , but winter months a little bit more ( because of seasonal fluctuations of these injections ) . as a response to the sinusoidal [ co@xmath0 ] oscillations ,
a sinusoidal oscillation in temperature is to be expected that is i ) offset vertically by an amount to make it fully above the zero line , ii ) offset ( delayed in time ) by a time that can be up to 3 months maximum , as will be discussed here .
neither is the case .
+ + + comparing the monthly fluctuations in temperature increase with monthly fluctuations in [ co@xmath0 ] we see again that the latter lags behind , this time by about 3 months ( to be precise , fitting sine curves to the data give a difference of 2.9 months ) .
one might think that the temperature lags behind 9 months
after all , months are periodic but upon second thought , this is not possible .
this is best explained in a relaxation model .
electronic engineers model things with electronic circuits and this case of temperature and co@xmath0 is also very adequately studied by such circuits . using an equivalent electronic circuit does not mean that the processes are electronic , but that they can be modeled by such circuits , as in an analog computer .
( the appendix gives the mathematical link between a relaxation model and the equivalent electronic circuit ) . in this case
we have a model between driving force ( either [ co@xmath0 ] , as we are wo nt to believe , or temperature @xmath5 ) and the response ( respectively , @xmath5 or [ co@xmath0 ] ) .
for instance , an increase in [ co@xmath0 ] will cause an increase in @xmath5 by the greenhouse effects .
this is necessarily a simple relaxation system , where the changes of the force cause the system to be off - equilibrium until a new equilibrium is reached .
this restoring of the equilibrium comes with a certain relaxation time .
the reasons for relaxation can be various .
for instance , co@xmath0 has to diffuse to places where it can do its temperature effect .
there can even be more than a single relaxation process , and instead be a complicated multi - relaxation process comparable to multi - stage nuclear decay .
the fact is that one of the relaxation times is dominant , and we can describe the relaxation by a single relaxation time ( that is the sum of all relaxation times ) . as long as there is no resonance in the system ( something that can only be achieved with positive feedback ) it will behave as described here .
we will model our climate system with a simple electronic relaxation system consisting of a resistance and a capacitance , @xmath11 and @xmath12 respectively , see figure [ fig : relax ] .
the product of the two yields the relaxation time , @xmath13 . at the entrance of this system
we connect our oscillating driving voltage @xmath14 ( representing , for example , [ co@xmath0 ] oscillations ) , in which @xmath15 is time .
the response is measured as the charge @xmath16 in the capacitor which represents for instance the temperature variations .
this charge is also measured by the output voltage by the standard capacitor relation @xmath17 .
thus our output voltage @xmath18 represents the response ( for example temperature ) . applying a sinusoidal input signal , @xmath19 [ co@xmath0]@xmath20 ( with @xmath21 the frequency of oscillation ) we get a sinusoidal wave at the output , with the same frequency , but with a phase at the output that is not equal to the phase at the input signal , @xmath22 .
the phase difference @xmath23 is directly and uniquely determined by the relaxation time of the system @xmath24 and the oscillation frequency @xmath21 , see figure [ fig : relax ] . for very low oscillating frequencies , the system can easily relax and the phase of the output signal is equal to that of the input signal . for increased frequencies or for increased relaxation times the system has difficulty accompanying the driving force .
the amplitude at the output drops and starts lagging behind the input .
the maximum phase difference for infinite frequencies or infinite relaxation time is exactly one - quarter period . in our case
our oscillating period is one year .
one quarter period is thus 3 months and that is the maximum delay we can expect between driving force and response . for relaxation times much longer than the oscillating period of one year , that is the delay one expects .
the delay time provides information about the system . as an example
, the comparable system of solar radiation and temperature comparable in that the oscillating period is one year and both deal with the weather and climate has a delay of one month ; the solar radiation and temperature oscillate with one year period , but the warmest day is nearly everywhere one month after the day with the most daylight and the on average coldest day is one month after the day with least daylight . in figure
[ fig : relax](bottom ) we see that the relaxation time of the \{radiation @xmath10 temperature } system therefore must be about 0.1 year ( 1.2 months ) . in the plot
this is indicated with a dot .
we can get a similar estimation value of the relaxation time of the atmosphere temperature through daily oscillations . as a rough figure , the temperature drops by about 4 degrees at night in about 8 hours after the sun has set .
assume that the relaxation upon this step - like solar radiation is a simple exponential ( situation b shown in the appendix ) and would finish eventually at close to absolute zero ( say 10 kelvin ) , and starts at 290 k , 4 degrees in 8 hours , we solve the equation @xmath25 which yields 23 days , similar to the value found above from yearly oscillations .
going back to the data of [ co@xmath0 ] and temperature ( fig . [
fig : year ] ) we can now understand the behavior , that is , the phase difference .
but only if we assume the temperature to be the driving force .
for instance : for some reason the temperature has increased more in winter months , and , as a result , to the natural [ co@xmath0 ] oscillations has been added a component with a maximum in spring months .
the alternative , [ co@xmath0 ] being the driving force and a delay of 9 months ( 3 quarter periods ) is mathematically not possible .
another explanation , which we do not consider a valid alternative , the temperature might be lagging behind [ co@xmath0 ] if it has a negative gain , i.e. , [ co@xmath0 ] increments lower the temperature .
this negative sign of the gain would add another 180@xmath2 phase shift and a total apparent phase shift of 270@xmath2 would be possible .
this goes even more against agw models and we do not see an easy physical explanation how co@xmath0 might lower the temperature .
this simple analysis opposes the hypothesis that [ co@xmath0 ] is causing serious temperature rises .
as said , the model assumes that no resonance occurs that can possibly cause longer delay times .
this , in our opinion , is a valid assumption since resonance is not likely .
first of all , for this strong positive feedback effects would be needed and they are not likely .
although many climate scientists have proposed positive feedback as discussed in the introduction and they make heavy use of them in order to explain and model the needed non - linear behavior of the greenhouse effect , this goes against intuition . in a chaotic system
these feedback factors are then extremely critical .
scientists of any plumage , when making such simulations , know this ; if they change their parameters just slightly ( sometimes even in the scale of the numerical resolution of their floating point numbers ) , the outcome can be hugely different .
there is also an experimental argument against positive feedback factors , namely the conscientious satellite measurements , see for instance the work of lindzen and choi@xcite , roy spencer@xcite , or wielicki et al.@xcite .
these , in fact , prove a _ negative _ feedback in the climate system . without feedback , in standard theory ,
if the earth warms up ( by global warming in a radiation imbalance ) , the temperature rises and the outward earth radiation increases by a certain amount , until establishing a new equilibrium . in the agw model ,
a positive feedback is used of the form : if the temperature increases , the outward earth radiation is less than that predicted by standard theory or the incoming solar radiation increases because of reasons like cloud ( non)forming , thus increasing the temperature even further .
the contrary can also happen : in negative feedback , if the planet heats up by a radiation imbalance for whatever reason , new channels of earth radiation can be opened or incoming solar radiation blocked ( for instance , by increased cloud cover ) , thus reducing the temperature with respect to standard theory . as demonstrated by the scientists mentioned above ,
the earth climate is a negative - feedback auto - stabilizing system , without going into detail what these feedbacks are .
this is also in agreement with the fact that , whereas the conditions on our planet have significantly changed over the geological history ( the sun for instance has been 25% less bright than today ) , the climate has been rather stable , always restoring from climate perturbations to median values instead of saturating in extreme values ; the latter one would expect in a thermal - runaway positive - feedback climate system .
note that , if large positive feedback exists , the temperature is unstable and will change until it saturates , that is until negative feedback becomes important .
in other words , it is technically not even possible that we are in a positive - feedback situation , considering the stable temperatures .
( compare this to the positive - feedback of a shop - a - holic buying always makes him buy even more his funds are acceleratingly depleted or his credit increasingly rising , until the banks put a lid on his spending , i.e. , negative feedback ) .
we _ must _ be in a negative - feedback situation and lindzen and choi , spencer , and wielicki , et al .
, have proven this by measurements .
negative feedback was already argued to be significant when the consensus of the scientists was for a global cooling , see the work of idso@xcite .
additional arguments against positive feedback come from the fact that every day , and every year the temperature system is brought off equilibrium . at night
it cools down , in the daytime it warms up . in the winter
it cools down and in summer it warms up .
these temperature disturbances are much larger and much faster than those that may have been produced by greenhouse gases ( 20 degrees / day or 30 degrees / year vs 0.7 degrees/100 years ) . the same accounts for co@xmath0 disturbances .
the human - caused co@xmath0 is insignificant compared to the large and noisy emissions naturally occurring on this planet ( only the accumulated effect of the tiny human - originated co@xmath0 is supposed to have an effect ) .
to give an idea , segalstad and coworkers established that of the current rise in [ co@xmath0 ] levels relative to the preindustrial level , only 12 ppm is attributable to human activity while 68 ppm is attributed to natural phenomena@xcite .
these fluctuations are also visible in the extensive summary of beck@xcite and show that even in recent history the [ co@xmath0 ] levels were sometimes higher than the modern values , while as everyone knows , the human emissions have monotonously increased , showing that these huge fluctuations can only have a natural origin .
relevant for the discussion here , the fluctuations would rapidly push the climate off equilibrium if it were unstable . yet , in spite of these huge disturbances , both in temperature and co@xmath0 , the equilibrium is restored every day and every year and every century . had the earth climate been a positive - feedback system , in summer or in winter the temperature would have been in a runaway situation , unrecoverable in the following compensating half - period .
apparently the system can recover very easily and repeatedly from such huge disturbances .
the reason is that the climate is a negative - feedback system that stabilizes itself .
this is an unavoidable conclusion .
one might think that the seasonal fluctuations are too fast to be causing a runaway scenario and that before the system runs away it already recovers .
that is a misapprehension ; changes can not be too fast .
if the system is unstable , it is unstable .
if starting oscillations are much faster than the response time of the system , the effective amplitude is reduced , but in a runaway system they will be amplified up to the point of saturation .
the system can only be stable if the feedback factor at that specific frequency is not positive .
look at it like this : in the first half of the year , it is hot and the system tries to runaway . in the second half of the year it is colder and it will restore , but it has a minute memory that the temperature has already run off a little in the first half and the second half therefore does not compensate completely . in the first year
we remain with a tiny temperature offset .
once this offset is introduced , the system will runaway .
of course , it can runway in both directions .
chance will determine which one , but if the system is unstable ( positive feedback ) , the system will runaway . like the metastable system of a ball placed on top of a hill .
it can only stay there in the absence of noise or any fluctuation in general . in conclusion
, only negative feedback makes sense .
relevant to the current work , such negative feedback will make any delay longer than 1/4 period impossible .
thus , the fact that we find a delay close to a quarter period means that i ) the temperature signal is the origin for [ co@xmath0 ] signal ( or the two are uncorrelated ) and ii ) the relaxation time @xmath24 linking the two is ( much ) longer than the period ( 12 months ) of oscillation .
moreover , even if positive feedback were present , for the resonance itself to be significant , the oscillating frequency needs to be close to the resonance frequency , i.e. , 12 months .
it is highly unlikely that the natural frequency of the climate-[co@xmath0 ] system is close to the 12-months - periodic driving force .
even more so , since also the long - term ice - drilling data need to be explained somehow , where delays of several thousands of years are observed . in our analysis ,
relaxation times of several thousands of years will explain both the ice - drilling data , as well as the yearly temperature and [ co@xmath0 ] oscillations . finally , the set of data we used is rather limited .
we only used data presented by balling , et al .
, that ends at the end of the 20th century .
moreover , they only have data from the northern hemisphere .
future research should tell if the ideas presented here can stand up to scrutiny when more recent data and pan - global data are used . as a note of proof , humlum et al.@xcite ,
have recently investigated correlation between temperature and [ co@xmath0 ] variations on the time scale of decades , similarly concluding that [ co@xmath0 ] changes are delayed in relation to temperature , and can therefore not be the reason for temperature changes .
in conclusion , the idea tested here that [ co@xmath0 ] is the _ cause _ of temperature changes does not pass our signal analysis .
it goes a little too far to say that this what we present here is proof for the opposite , namely that [ co@xmath0 ] is the _ effect _ of temperature , but our analysis does not contradict this .
future will tell if such an hypothesis may be postulated with some confidence .
acknowledgements : this research was paid by no grant .
it received no funding whatsoever , apart from our salaries at the university where we work . nor are we members of any climate committees ( political or other ) or are we linked to companies or ngos , financially or otherwise .
this is an independent opinion that does not necessarily represent the opinion of our university or of our government .
in simple relaxation models the ( negative ) change of a quantity is proportional to the magnitude of the remaining quantity .
simple examples are nuclear decay , in which the change of number @xmath26 of atoms at a certain time @xmath15 is given by @xmath27 , or the velocity @xmath28 of an object under friction is given by @xmath29 .
( @xmath30 and @xmath31 positive constants ) . from experience , and by solving the differential equation , we know that such systems show exponential decay , @xmath32 and @xmath33 respectively .
now , we can take a function @xmath34 that is the driving force of another quantity @xmath35 , the response function , respectively the cause and the effect .
we can decompose the function @xmath21 into an integral of dirac - delta functions .
the response to each delta function is given by the function @xmath36 .
assuming linearity , the total response is then a convolution @xmath37 where the heaviside function @xmath38 ( @xmath39 for @xmath40 and 0 otherwise ) was used to force the causality ; the response @xmath41 can only come after the driving force .
( note that non - linearities will not change the sign of these calculations , i.e. , a delay can not become an advance . ) for instance , if the response function is an exponential decay , as mentioned above , @xmath42 substituting a delta - function at @xmath43 for the driving force @xmath21 will reproduce the exponential decay : @xmath44){\updelta}(s){\rm d}s\\ & = & g_0\exp(-{\upalpha}t)u(t)\end{aligned}\ ] ] in other words , the response to a spike , a delta function at @xmath43 is an exponential decay with an amplitude @xmath45 , and time constant @xmath46 .
the response to a heaviside ( step)function @xmath47 is then given by @xmath48/{\upalpha}\end{aligned}\ ] ] more interesting
more relevant for our work is the case of a sinusoidal driving force .
this can now easily be calculated by substituting the driving - force function @xmath21 into eq .
[ eq : conv ] : @xmath49{\rm d}s\\ \nonumber & = & \frac{f_0g_0}{{\upalpha}^2+{\upomega}^2 } \left [ { \upalpha}\sin({\upomega}t ) - { \upomega}\cos({\upomega}t)\right]\\ & = & \frac{f_0g_0}{\sqrt{{\upalpha}^2+{\upomega}^2 } } \sin\left({\upomega}t - \tan^{-1}[{\upomega}/{\upalpha}]\right)\end{aligned}\ ] ] ( for the second step in eq .
[ eq : grad ] gradshteyn and ryshik@xcite was used ) .
figure [ fig : apndx ] shows these three cases of driving forces and response functions . figure [ fig : algore ] shows a simulation with the driving function @xmath34 equal to the measured temperature and a delay of @xmath24 ( @xmath50 ) = 15 kyr , which results in a quite good representation of the [ co@xmath0]curve . an electronic circuit such as presented here has these properties of exponential response to a heaviside function and linearity and the response of equation [ eq : grad ] .
for this reason , such ( virtual ) circuits are widely used in simulations of phenomena including phenomena far away from electronics .
the interesting and relevant conclusion of eq .
[ eq : grad ] is that the maximum phase shift is 90@xmath9 and this occurs for frequencies that are much higher than the relaxation speed , @xmath51 .
h. fischer , m. wahlen , j. smith , d. mastroianni , and b. deck , 1999 : ice core records of atmospheric co2 around the last three glacial terminations , science 283 , 17121714 .
( doi : 10.1126/science.283.5408.1712 ) .
% noindent r. c. balling , r. s. vose , g .-
weber , 1998 .
analysis of long - term european records : 1751 - 1995 .
climate research 10 , 193 - 200 . also available on - line at http://www.int-res.com/articles/cr/10/c010p193.pdf ( visited sept . 2012 ) .
r. spencer , 2008 .
satellite and climate model evidence against substantial manmade climate change .
_ journal of climate _ , submitted .
available at : http://www.drroyspencer.com/research-articles/satellite-and-climate-model-evidence/ ( visited sept .
2012 ) .
b. a. wielicki , t. wong , r. p. allan , a. slingo , j. t. kiehl , b. j. soden , c. t. gordon , a. j. miller , s .- k .
yang , d. a. randall , f. robertson , j. susskind , h. jacobowitz , 2002 .
evidence for large decadal variability in the tropical mean radiative energy budget .
science 295 , 841 - 843 . | the primary ingredient of anthropogenic global warming hypothesis is the assumption that atmospheric carbon dioxide variations are the cause for temperature variations . in this paper
we discuss this assumption and analyze it on basis of bi - centenary measurements and using a relaxation model which causes phase shifts and delays . note : this paper was ( and is ) submitted to various journals and received no scientific criticism whatsoever .
it is being rejected principally on reasons of format or for having an alleged political agenda .
scientific comments can be sent to our contact addresses mentioned above .
key words : global warming , cause and effect , relaxation model |
these are lecture notes from the saalburg summer school 2004 lecture of m. shifman , william i. fine theoretical physics institute , university of minnesota , minneapolis , mn 55455 , usa .
they deal with supersymmetric gauge theories , in particular with the calculation of the gluino condensate . a more in - depth
treatment can be found in the review @xcite and the book @xcite . a concise and very clearly written introduction to supersymmetry
can be found in the book @xcite which i also recommend .
there are various older papers on the subject , see e.g. @xcite .
in contrast to the generators of the lorentz group , the generators of supersymmetry transformations ( `` supercharges '' ) are spinors and obey anticommutation relations . in this lectures we only consider @xmath0 supersymmetry in four spacetime dimensions .
then the susy generators are a left - handed weyl spinor @xmath1 , @xmath2 , and its hermitean conjugate @xmath3[s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}}$ ] , @xmath4 which satisfy the ( anti)commutation relations @xcite @xmath5[s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } } \right\ } & = \left\{{\ensuremath { { \settowidth{\textlength}{$\mathcal{q}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}},{\ensuremath { { \settowidth{\textlength}{$\mathcal{q}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\beta}}}}}\right\}=0\\ \left\{\mathcal{q}_\alpha,{\ensuremath { { \settowidth{\textlength}{$\mathcal{q}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}}\right\ } & = 2\sigma^m_{\alpha{{\ensuremath{\dot{\alpha } } } } } \mathcal{p}_m \label{susyalgebra - qqb}\\ \left[\mathcal{q}_\alpha,\mathcal{p}_m\vphantom{{\ensuremath { { \settowidth{\textlength}{$\mathcal{q}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } } \right ] & = \left[{\ensuremath { { \settowidth{\textlength}{$\mathcal{q}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\mathcal{q}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\mathcal{q}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}},\mathcal{p}_m\right]=0 \label{susyalgebra - qp } \end{aligned}\ ] ] from eq . ( [ susyalgebra - qqb ] ) , wee see that the mass dimension of @xmath6 is @xmath7=\tfrac{1}{2}\left[\mathcal{p}_m\right]=\tfrac{1}{2}$ ] .
the last line tells us that @xmath8=0 $ ] , which means that supersymmetry transformations do not change the mass of a state .
poincar and supersymmetry basically exhaust the possible symmetries of the s matrix of a sensible qft , apart from a compact internal symmetry .
the part of the internal symetry which does not commute with the generators of supersymmetry transformations is called @xmath9 symmetry , which in the @xmath0 case is at most a @xmath10 rotating the supercharges .
just as the momentum operator is realised as translations @xmath11 in minkowski space , the supersymmetry generators can be represented by differential operators @xmath12 , @xmath13[s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}}$ ] on a larger space which is ( of course ) called superspace ( it was introduced in @xcite ) .
in addition to the usual minkowski coordinates @xmath14 , this space contains two additional grassmann valued ( i.e. anticommuting ) weyl spinors @xmath15 , @xmath16 .
points in this space are thus labeled by coordinates @xmath17 .
( our conventions regarding spinors and @xmath18-matrices are collected in the appendix . )
note that the @xmath19 s have mass dimension @xmath20=\left[{\ensuremath{\bar{\theta}}}\mspace{2mu}\right]=-\tfrac{1}{2}$ ] .
if an @xmath9 symmetry is present , the @xmath19 s are rotated as well .
the supersymmetry generators are then represented by the operators @xmath21[s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha } } } } } & = -{\ensuremath{\bar{\partial}}}_{{{\ensuremath{\dot{\alpha } } } } } + { \ensuremath{\text{i}}}\theta^\alpha\sigma^m_{\alpha{{\ensuremath{\dot{\alpha}}}}}\partial_m .
\end{split } \end{aligned}\ ] ] other important operators are the supersymmetry - covariant derivatives @xmath22[s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha } } } } } & = -{\ensuremath{\bar{\partial}}}_{{{\ensuremath{\dot{\alpha } } } } } -{\ensuremath{\text{i}}}\theta^\alpha\sigma^m_{\alpha{{\ensuremath{\dot{\alpha}}}}}\partial_m \end{split } \end{aligned}\ ] ] which fulfill @xmath23[s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}}\right\}=-2{\ensuremath{\text{i}}}\sigma^m_{\alpha{{\ensuremath{\dot{\alpha}}}}}\partial_m\ , , \qquad \left\{d_\alpha , d_\beta\right\}=\left\{{\ensuremath { { \settowidth{\textlength}{$d$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}},{\ensuremath { { \settowidth{\textlength}{$d$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\beta}}}}}\right\}=0\,,\\ \left\{d_\alpha , q_\beta\right\}= \left\{d_\alpha,{\ensuremath { { \settowidth{\textlength}{$q$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\beta}}}}}\right\}= \left\{{\ensuremath { { \settowidth{\textlength}{$d$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}},q_\beta\right\}= \left\{{\ensuremath { { \settowidth{\textlength}{$d$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}},{\ensuremath { { \settowidth{\textlength}{$q$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\beta}}}}}\right\}=0 .
\end{gathered}\ ] ] superfields ( which were introduced in the same work of salam and strathdee ) are functions of the superspace coordinates which are defined by their expansion in the anticommuting coordinates .
since @xmath24 , the expansion can contain only terms with at most two @xmath19 s and two @xmath25 s : @xmath26[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } + \theta^2 m + { \ensuremath{\bar{\theta}}}^2 n + \theta\sigma^m{\ensuremath{\bar{\theta}}}a_m + \theta^2{\ensuremath{\bar{\theta}}}{\ensuremath{\bar{\psi}}}+{\ensuremath{\bar{\theta}}}^2\theta\lambda + \theta^2{\ensuremath{\bar{\theta}}}^2 d , \end{split } \end{aligned}\ ] ] where the component fields @xmath27 , @xmath28 , @xmath29 and @xmath30 are scalars , @xmath31 is a vector and @xmath32 , @xmath33 , @xmath34 and @xmath35[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } $ ] are left- or right - handed spinors , all functions of @xmath14 .
moreover , @xmath36 one may or may not impose the condition of reality , @xmath37 . the superfield @xmath38 is usually referred to as the vector superfield .
the reason will become clear shortly .
the component fields in @xmath38 form a representation of the supersymmetry algebra .
they will be involved in construction of the gauge sector . for constructing supersymmetric yang - mills theories with matter
, we will also use shorter superfields as generalisations of matter fermions .
the matter fields will be components of the ( irreducible ) chiral multiplet @xmath39 , defined as @xmath40[s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha } } } } } \phi & = 0 \ , . \label{chiralcondition } \end{aligned}\ ] ]
the chiral superfield can be regarded as a function defined not on the full superspace , but , rather , on the left - handed subspace parametrised by the coordinates @xmath41 and @xmath19 , with no explicit dependence on @xmath25 , since in this basis , @xmath42[s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ { { { \ensuremath{\dot{\alpha}}}}}=-{\ensuremath{\bar{\partial}}}_{{{\ensuremath{\dot{\alpha}}}}}$ ] , and thus eq .
( [ chiralcondition ] ) means that @xmath39 does not depend on @xmath25 explicitly .
the possibility of introducing such superfield is due to the fact that the supertransformations of @xmath43 and @xmath19 are closed , namely , @xmath44 where @xmath45 is the ( weyl spinor valued ) transformation parameter .
the component expansion of the chiral superfield is particularly simple .
the expansion in terms of the usual coordinates can be obtained by the taylor expansion in the grassmann variables , @xmath46 analogously , antichiral fiels @xmath35[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } $ ] are defined as @xmath47[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } & = 0 , \end{aligned}\ ] ] and their expansion is best expressed in terms of @xmath48 and @xmath25 : @xmath49[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } & = { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ( x_{\scriptscriptstyle r } ) + { \ensuremath{\sqrt{2}\,}}{\ensuremath{\bar{\theta}}}{\ensuremath{\bar{\chi}}}(x_{\scriptscriptstyle r } ) + \theta^2 { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ( x_{\scriptscriptstyle r } ) \\ & = { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ( x ) -{\ensuremath{\text{i}}}\theta\sigma^m{\ensuremath{\bar{\theta}}}\partial_m { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ( x ) + \tfrac{1}{4}\theta^2{\ensuremath{\bar{\theta}}}^2\box { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ( x)\\ & \quad + { \ensuremath{\sqrt{2}\,}}{\ensuremath{\bar{\theta}}}{\ensuremath{\bar{\chi}}}(x ) + \tfrac{{\ensuremath{\text{i}}}}{{\ensuremath{\sqrt{2}\,}}}{\ensuremath{\bar{\theta}}}^2\theta\sigma^m\partial_m{\ensuremath{\bar{\chi}}}(x ) + \theta^2 { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ( x ) \ , .
\end{split } \end{aligned}\ ] ] conjugates of chiral fields are antichiral , and a field that is both chiral and antichiral is constant
. products of chiral fields are again chiral , a product of a chiral and an antichiral field is neither chiral nor antichiral . if we want the scalar @xmath50 to have canonical mass dimension @xmath51=1 $ ] , we must assign to @xmath39 the dimension @xmath52=1 $ ] .
since @xmath20=-\tfrac{1}{2}$ ] , the spinor field @xmath32 also has the correct dimension .
the field @xmath38 , however , has mass dimension two .
this signifies that it is an auxiliary field , appearing in the lagrangean without derivatives , and that it can be eliminated by its equations of motion .
the physical components of the chiral superfields are scalar or spinor . to incorporate the gauge field we need to turn to eq .
( [ generalsuperfield ] ) .
the vector superfield @xmath53 , which is a generalisation of the gauge vector fields in non - supersymmetric theories , contains a vector field , a weyl fermion and its conjugate ( the gaugino , equivalent to a majorana fermion ) and a few other fields which can be eliminated by a supergauge transformation or are auxiliary .
the reality condition @xmath54 is implied .
the component expansion of the vector superfield is given in eq .
( [ generalsuperfield ] ) .
supergauge transformations acting on this field are inhomogeneous .
they involve chiral superfields as their parameters , ina way such that lower components of @xmath53 can be gauged away .
this is the famous wess zumino gauge which leaves us with the following decomposition : @xmath55 the vector field @xmath56 still has the usual gauge symmetry @xmath57 ( in the abelian case ) .
the fields @xmath33 and @xmath58 are gauge invariant in the abelian case . in the non - abelian case
they transform homogeneously . the wess
zumino gauge is the one most frequently used in dealing with supersymmetric gauge theories . on the other hand ,
if @xmath53 becomes massive by some kind of the higgs effect , the most convenient gauge is unitary , in which the lower components are no longer gauged away ; rather , they represent additional degrees of freedom , a real scalar @xmath59 and another weyl fermion @xmath32 , as we will see later .
the vector superfield contains an auxiliary field as well , the real scalar @xmath58 . just as the @xmath38 field in the chiral superfield
, it has mass dimension two ( since we want @xmath56 to have canonical dimension ) .
the auxiliary fields are needed to make the algebra close off - shell and will be useful when we consider the scalar potential of supersymmetric yang - mills theories .
finally , we need a way to write down invariant actions from superfields .
this can be achived by constructing lagrangeans which vary by a total derivative under supersymmetry transformations .
the highest ( auxiliary ) components of chiral and real superfields have exactly this property .
they can be projected out by the rules of intergration over the grassmann variables ( berezin integral ) , which for one anticommuting variable @xmath19 are can be expanded as .
this definition furthermore is unique up to a factor if we require translational invariance of the integral . ]
@xmath60 this means that @xmath61=-\left[\theta\right]$ ] .
note also that integration and differentiation give the same result for the grassmann variables , so a @xmath62-integration over a chiral field vanishes ( actually , it is a total derivative , so its spacetime integral vanishes and it does not contribute to the action ) . for superspace with two
grassmann coordinates , we define @xmath63 with coefficients chosen such that integration over the superspace coordinates gives @xmath64 so , the integral over the grassmann coordinates projects out the @xmath38-components of chiral superfields and the @xmath58-components of real superfields .
we can for example consider a theory with just a chiral superfield @xmath39 , i.e. a scalar and a spinor .
for the lagrangean we need a kinetic term and , maybe , mass or interaction terms .
the kinetic term is provided just by @xmath65[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } \phi .
\end{aligned}\ ] ] the surviving terms need to have two @xmath19 s and two @xmath25 s , so we have to collect the possible combinations from eqs .
( [ chiralsf ] ) and ( [ antichiralsf ] ) .
this gives @xmath66[s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \box a + \box { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } a\right ) -\tfrac{1}{2}\partial_m{\ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \,\partial^m a + \tfrac{{\ensuremath{\text{i}}}}{2}\left(\partial_m{\ensuremath{\bar{\chi}}}{\ensuremath{\bar{\sigma}}}^m\chi -{\ensuremath{\bar{\chi}}}{\ensuremath{\bar{\sigma}}}^m\partial_m\chi\right ) + { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } f \\ & = -\partial_m { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \partial^m a -{\ensuremath{\text{i}}}{\ensuremath{\bar{\chi}}}{\ensuremath{\bar{\sigma}}}^m\partial_m\chi + { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } f \end{split } \end{aligned}\ ] ] and we see that indeed @xmath38 appears without derivatives . in going from the first to the second line we have performed an integration by parts and dropped the total derivative terms .
mass and interaction terms appear as purely chiral combinations of fields in the so - called superpotential , @xmath67 where @xmath28 and @xmath33 are in general complex coefficients . in components , this gives @xmath68 in eq .
( [ simplesuperpot ] ) , a term linear in @xmath39 can be absorbed by a redefinition of the field .
if we now combine the two parts of the lagrangean to @xmath69 , we can eliminate @xmath38 and @xmath70[s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } $ ] by their equations of motion , @xmath71[s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \right ) } = \frac{\delta { \ensuremath{\mathscr{l}}}}{\delta { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } } = f + { \ensuremath { { \settowidth{\textlength}{$m$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } + { \ensuremath{\bar{\lambda}}}{\ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ^2 .
\end{aligned}\ ] ] this means that the part of the lagrangean containing @xmath38 and @xmath70[s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } $ ] reduces to @xmath72[s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } { \ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } + { \ensuremath{\bar{\lambda}}}{\ensuremath { { \settowidth{\textlength}{$a$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$a$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$a$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ^2\right)\left(ma+\lambda a^2\right)= -\left|\frac{\partial w}{\partial \phi } \right|^2\equiv -v(a ) , \end{aligned}\ ] ] the potential for the scalar field which is always nonnegative .
super - qed ( sqed ) is the supersymmetric generalisation of qed , so it must contain an electron , and a photon , together with their superpartners , selectron and photino .
what kinds of interactions can we expect ?
first of all , there should be the ordinary electron - electron - photon vertex from qed [ eegamma ] .
but in this vertex , we can exchange two of the fields for their superpartners , resulting in an electron - selectron - photino vertex [ ephilambda ] and a selectron - selectron - photon vertex [ phiphigamma ] .
we will see that there are also two quartic vertices , namely a four - selectron vertex [ phihochvier ] and a selectron - selectron - photon - photon vertex [ phiphigammagamma ] .
the form of these vertices is fixed by supersymmetry , as can most convieniently be seen in superfields : [ sqedvertices ] the electron and positron together form a dirac spinor @xmath73 which can be decomposed into two weyl spinors of opposite chirality , @xmath74 .
both @xmath32 and @xmath75 have electric charge @xmath76 .
the complex conjugate field @xmath77 also contains a left- and a right - handed spinor , both with electric charge @xmath78 .
so we can describe the electron by two left handed weyl spinors @xmath32 and @xmath45 with electric charges @xmath76 and @xmath78 , respectively .
the conjugate fields @xmath79 and @xmath75 have opposite charges .
we arrange @xmath32 and @xmath45 in two chiral superfields : @xmath80 they are accompanied by scalar fields @xmath81 and @xmath82 , the selectrons ( or selectron and spositron ) and by two auxiliary fields @xmath38 and @xmath83 .
the photon @xmath84 is placed in a real superfield @xmath53 together with its superpartner , the photino and the auxiliary field @xmath58 . in the wess
zumino gauge @xmath53 reads @xmath85 the gauge transformations we consider are direct generalisations of normal qed , where the fields transform as @xmath86 in sqed , the gauge parameter @xmath87 becomes a chiral superfield , and the gauge transformations become @xmath88 so the real superfield @xmath53 transforms inhomogeneously .
this is precisely the property that allows us to take @xmath53 to be in the wess
zumino gauge .
this procedure not only eliminates the lower components of @xmath53 , but also fixes all the additional gauge parameters in the superfield @xmath87 apart from the imaginary part of the scalar component which plays the rle of the usual gauge parameter of qed .
the photino and the auxilary field are gauge invariant . the kinetic term for the chiral multiplet as it stands in eq . ( [ simplekaehlerpot ] ) ,
however , is not gauge invariant .
it has to be modified to @xmath89[s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } e^{2v}q + { \settowidth{\textlength}{${\ensuremath{\widetilde{q}}}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{${\ensuremath{\widetilde{q}}}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{${\ensuremath{\widetilde{q}}}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } e^{-2v}{\ensuremath{\widetilde{q}}}\right ) .
\end{aligned}\ ] ] the exponential @xmath90 is is actually quite simple in the wess
zumino gauge , since @xmath91 . in zeroth order
it just reproduces the kinetic terms of eq .
( [ simplekaehlerpot ] ) ; the higher terms mainly amount to changing ordinary derivatives to covariant ones , but introduce one important additional term that includes the @xmath58-component of @xmath53 and is @xmath92 .
we also need kinetic terms for the gauge superfield and , possibly , a superpotential . the kinetic term for @xmath53 is usually written in terms of the field strength superfield ( or gaugino superfield ) @xmath93[s]{$d$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$d$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } ^2 d_\alpha v\ ] ] as @xmath94 where @xmath95 is the complexified coupling constant , @xmath96 with @xmath97 the vacuum angle .
we are using a complex parameter ( just as in eq .
( [ simplesuperpot ] ) ) because chiral quantities depend holomorphically on these parameters , which can be a powerful tool in many calculations . in the abelian case we deal with now
@xmath98 is a gauge invariant chiral superfield .
its component expansion is @xmath99 where @xmath100 is the photon field strength with spinor indices , @xmath101 so @xmath102 takes the component form @xmath103 + \frac{\vartheta}{64\pi^2}\varepsilon^{mnkl}f_{mn}f_{kl}. \end{aligned}\ ] ] the superpotential is strongly restricted by gauge invariance . since we have two chiral superfields of opposite charge , the only possible term ( restricting to renormalisable couplings ) is @xmath104 since the superpotential can only depend on the neutral product @xmath105 , and a term @xmath106 would already require a coefficient with mass dimension -1 .
again , @xmath28 is a complex parameter whose absolute value will be the mass of the electron and selectron .
now we have kinetic terms for matter and gauge fields as as well as a superpotential .
what else can be present ?
actually , there is one more term , which is possible only for abelian theories , the so - called fayet - iliopoulos ( or @xmath107 ) term , @xmath108 which is ( super)gauge invariant by itself ( this is clearly seen fron eq .
( 16c ) ; the @xmath62 integral of chiral and antichiral superfields vanishes ) .
it looks rather innocent , but has important consequences as we will see shortly .
the coefficient of the real superfield , @xmath107 , is real too .
the the dependence on @xmath107 is not holomorphic .
now we can collect all the terms of the lagrangean , @xmath109[s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } e^{2v}q + { \settowidth{\textlength}{${\ensuremath{\widetilde{q}}}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{${\ensuremath{\widetilde{q}}}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{${\ensuremath{\widetilde{q}}}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } e^{-2v}{\ensuremath{\widetilde{q}}}+2\xi v\right)\\ & \quad + \left\{\int{\ensuremath{\text{d}}}^2\theta\left(m{\ensuremath{\widetilde{q}}}q+\frac{1}{4\mathfrak{e}^2}w^\alpha w_\alpha\right ) + \text{h.c.}\right\}\\ & = -d_m\bar{{\ensuremath{\tilde{\varphi}}}}d^m { \ensuremath{\tilde{\varphi}}}-d_m{\ensuremath{\bar{\varphi}}}d^m \varphi - m\chi\eta -{\ensuremath { { \settowidth{\textlength}{$m$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } { \ensuremath{\bar{\chi}}}{\ensuremath{\bar{\eta}}}\\ & \quad -\frac{{\ensuremath{\text{i}}}}{2}\left(d_m{\ensuremath{\bar{\chi}}}{\ensuremath{\bar{\sigma}}}^m\chi -{\ensuremath{\bar{\chi}}}{\ensuremath{\bar{\sigma}}}^m d_m\chi\right ) -\frac{{\ensuremath{\text{i}}}}{2}\left(d_m{\ensuremath{\bar{\eta}}}\,{\ensuremath{\bar{\sigma}}}^m\eta -{\ensuremath{\bar{\eta}}}\,{\ensuremath{\bar{\sigma}}}^m d_m\eta\right)\\ & \quad -\frac{1}{e^2}\left(2{\ensuremath{\text{i}}}\lambda{{\ensuremath{\rlap{\raisebox{1pt}{$\mspace{2mu}/$}}\partial}}}{\ensuremath{\bar{\lambda}}}+f_{mn}f^{mn}\right ) + \frac{\vartheta}{8\pi}\varepsilon^{mnkl}f_{mn}f_{kl } \\ & \quad + { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } f + { \settowidth{\textlength}{${\ensuremath{\widetilde{f}}}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{${\ensuremath{\widetilde{f}}}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{${\ensuremath{\widetilde{f}}}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } { \ensuremath{\widetilde{f}}}+m\varphi{\ensuremath{\widetilde{f}}}+{\ensuremath { { \settowidth{\textlength}{$m$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } { \ensuremath{\bar{\varphi } } } { \settowidth{\textlength}{${\ensuremath{\widetilde{f}}}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{${\ensuremath{\widetilde{f}}}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{${\ensuremath{\widetilde{f}}}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } + m{\ensuremath{\tilde{\varphi}}}f + { \ensuremath { { \settowidth{\textlength}{$m$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \bar{{\ensuremath{\tilde{\varphi}}}}{\ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \\ & \quad + \frac{2}{e^2}d^2 + \xi d + d\left({\ensuremath{\bar{\varphi}}}\varphi-\bar{{\ensuremath{\tilde{\varphi}}}}{\ensuremath{\tilde{\varphi}}}\right)\ , .
\end{split } \end{aligned}\ ] ] here , @xmath110 , where @xmath111 and their conjugates and the sign depends on the charge .
the first two lines contain the kinetic terms for electrons and selectrons and mass terms for the fermions , the third line contains the gauge sector kinetic terms and the last two lines are the auxiliary field contribution . to obtain mass terms for the scalars and exhibit other interesting effects
, we have to eliminate these auxiliary fields via their equations of motion .
the equations of motion are @xmath112[s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } \bar{{\ensuremath{\tilde{\varphi}}}}\ , , \\ 0&={\ensuremath{\widetilde{f}}}+{\ensuremath
{ { \settowidth{\textlength}{$m$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$m$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$m$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } { \ensuremath{\bar{\varphi}}}\ , , \\ 0&=d + \frac{e^2}{4}\left(\xi + { \ensuremath{\bar{\varphi}}}\varphi-\bar{{\ensuremath{\tilde{\varphi}}}}{\ensuremath{\tilde{\varphi}}}\right)\ , , \end{aligned}\ ] ] so the auxiliary field lagrangean becomes the potential term for the scalar fields , @xmath113 this is called the scalar potential ; it consists of two parts : the @xmath38 term part and the @xmath58 term part .
the scalar potential allows us to analyse the vacua of the theory .
we see that the potential is a sum of squares , so it can never be negative .
that means that field configurations for which @xmath114 are vacua ( and supersymmetry is unbroken in these cases ) .
however , the existence of supersymmetric vacua depends on the parameters @xmath107 and @xmath28 .
let us consider four cases : [ [ case - mxi0 ] ] case @xmath115 : + + + + + + + + + + + + + + + + + + + + + + + + + + + the scalar potential reduces to @xmath116 which vanishes for any field configuration where @xmath117 . both fields ,
being equal , can take arbitrary complex value .
thus , the vacuum manifold , i.e. the set of minima of the scalar potential @xmath53 , is a complex line .
what is more important is that distinct vacua on this line are not physically equivalent .
this is different from the degenerate vacua situation in ordinary qft .
typically the degenerate vacua situation takes place when a global symmetry is spontaneously broken . in this case
one deals with a compact set of physically equivalent vacua ( e.g. the mexican hat potential ) . in supersymmety
non - compact continuous vacuum manifolds are typical . they are usually called flat directions .
the fields corresponding to these directions are referred to as moduli fields . actually , the requirement @xmath117 is not necessary , since only the absolute value enters the potential , and it might seem that there is a larger vacuum manifold .
however , the relative phase of @xmath81 and @xmath82 is a gauge artefact , and one can use the product @xmath118 as a gauge - independent parametrisation of the vacua ( sometimes called a composite modulus ) .
all vacua in this case have vanishing potential , so supersymmetry is unbroken , but for @xmath119 or @xmath120 , the gauge symmetry is spontaneously broken .
it is clear that the vacuum @xmath121 _ and _ @xmath122 , with unbroken gauge symmetry , is special .
[ [ case - m0xineq0 ] ] case @xmath123 : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in this case @xmath124 still holds , but @xmath125 gets modified , @xmath126 if we minimise the potential ( assuming , just exchange @xmath81 and @xmath82 in the following discussion . ] @xmath127 ) , we arrive at the condition @xmath128 this equation has solutions which preserve supersymmetry ( i.e. @xmath114 ) , but the gauge symmetry is always broken . even if we put @xmath129 , still @xmath130 .
since in this case the absolute values of the fields are fixed , all that is left of the vacuum manifold is a compact @xmath10-circle , with all vacua on this circle being equivalent . since @xmath131 need not vanish and its absolute value can be arbitrary , the full vacuum manifold is non - compact .
[ [ case - mneq0xi0 ] ] case @xmath132 : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this case is quite simple , since the potential has just one minimum at @xmath133 , so neither gauge nor supersymmetry is broken .
[ [ case - mneq0xineq0 ] ] case @xmath134 : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in the full potential , it is obvious that there will be no supersymmetric vacua , since @xmath135 vanishes only at the origin , where @xmath125 does not , and hence @xmath136 always .
so supersymmetry is broken .
now we proceed to find the minima : @xmath137 \varphi\\ 0&=\frac{\partial v}{\partial \bar{{\ensuremath{\tilde{\varphi } } } } } = 2\left[-\left({\ensuremath{\bar{\varphi}}}\varphi -\bar{{\ensuremath{\tilde{\varphi}}}}{\ensuremath{\tilde{\varphi}}}+\xi\right ) + \tfrac{1}{2}|m|^2\right ] { \ensuremath{\tilde{\varphi}}}\end{aligned}\ ] ] we distinguish four possibilities : a. [ fia]@xmath133 : this is an extremal point of the potential where @xmath138 .
whether is is a maximum or minimum depends on the magnitudes of @xmath28 and @xmath107 .
this can be checked directly , but we will see that when we consider the other cases .
the potential for the case where this is the minumum is sketched in figure [ fayetiliopoulosa ] .
b. [ fib]@xmath129 , @xmath139 : this means that @xmath140 , which is only possible if @xmath141 .
the value of the potential is @xmath142 , which is smaller than the value at the origin , @xmath143 , if the above condition holds .
c. [ fic]@xmath144 , @xmath145 : now we need @xmath146 , which is only possible if @xmath147 .
the potential is in this case @xmath148 .
so in this case as in the one before , the extremum at the origin is actually not a minimum .
the mimima occur at finite values of the fields , so gauge symmetry is broken in the vacuum .
the potential in these cases is sketched in figure [ fayetiliopoulosb ] .
d. @xmath149 , @xmath150 : in this case , there is no extremal point at all for nonzero mass .
if @xmath151 , it reduces to the case considered before where gauge symmetry is broken , but supersymmetry is not .
after sqed , we will now consider non - abelian gauge theories .
the simplest example is a pure super - yang - mills theory ( sym ) without matter , sometimes called susy gluodynamics .
this theory contains a vector field @xmath152 and a majorana spinor @xmath153 , both in the adjoint representation of the gauge group with generators @xmath154 .
they are again grouped in a vector superfield @xmath155 which we can take to be in the wess
zumino gauge .
the action looks rather simple , @xmath156 where @xmath157 is the covariant derivative and @xmath158 is the complexified coupling constant , @xmath159 moreover , @xmath160 this theory is believed to be confining and is asymptotically free , i.e. the first coefficient of the @xmath161-function ( gell - mann low function ) is negative . for a @xmath162 gauge group , @xmath163 .
function of a non - supersymmetric @xmath164 gauge theory with @xmath165 flavours , where @xmath166 .
the supersymmetric theory can be thought of as a usual theory with one gaugino , so @xmath167 .
there is an additional factor of @xmath168 for the gaugino contribution because it is in the adjoint representation of the gauge group rather than in the fundamental .
the graphs contributing to this term contain a casimir operator for the representation , and the ratio of the casimirs in the adjoint and fundamental representation is @xmath169 .
furthermore , the gaugino is a majorana fermion with only half as many degrees of freedom as the dirac one , so there is another factor of @xmath170 , resulting in @xmath171 . ]
the aspect we will focus on is the gluino condensate @xmath172 , which vanishes perturbatively , but does not vanish in the nonperturbative treatment and can be computed _ exactly _ @xcite .
we will calculate its value in chapter [ calcchap ] .
now we give the result and argue for its plausibility .
the result is @xmath173 in this expression , @xmath174 is the scale where the theory becomes strongly coupled .
it is the only dimensionful parameter , so it has to appear to the third power .
the pre - factor @xmath175 can be understood since the sum over group indices involves @xmath176 terms , but a factor of @xmath177 is included in @xmath33 , and @xmath178 . in the exponential , the parameter @xmath179 , @xmath180 labels distinct supersymmetric vacua , reflecting the @xmath175-fold vacuum degeneracy .
the above general form can be derived from holomorphicity in conjunction with renormalization group ( rg ) arguments .
the only rg invariant expression of dimension of mass which can be built from @xmath177 and the renomalisation scale @xmath181 is @xmath182 , so @xmath183 ^ 3 \\ & = c \,\mu^3 \exp\left\{-\frac{8\pi^2}{n}\frac{1}{g^2}\right\ } \end{split } \end{aligned}\ ] ] where @xmath184 is a constant and the second equality holds for @xmath162 gauge groups ( ) . on the other hand , @xmath185 is a chiral quantity , so it must depend analytically on @xmath158 , not just on @xmath177 , see @xcite , @xmath186 with a possibly different constant @xmath187 .
next we notice that physical results must be @xmath188-periodic in the vacuum angle @xmath97 .
the above result is not , so we have to change the last factor in eq .
( [ condensate1 ] ) to @xmath189 the parameter @xmath179 labels the distinct ( but degenerate ) vacua , of which there are @xmath175 this is exactly the value of the witten index @xmath190 in @xmath162 theories @xcite , which in fact counts the number of vacua . if there are no matter fields , supersymmetry is always unbroken .
we can also deduce the number of vacua by considering a global symmetry of the lagrangean ( [ symaction ] ) ( besides supersymmetry ) .
there is a global chiral @xmath191 symmetry which acts as @xmath192 this global symmetry corresponds to the current @xmath193 which is classically conserved , @xmath194 .
r80pt ( 80,90)(0,0 ) ( 37,103)(43,97)(37,97)(43,103 ) ( 40,100)(40,70 ) ( 42,85)[l]@xmath195 ( 40,70)2 ( 40,70)(20,35.4 ) ( 20,35.4)2 ( 20,35.4)(60,35.4 ) ( 60,35.4)2 ( 60,35.4)(40,70 ) ( 20,35.4)(20,5)34 ( 15,23)[r]@xmath196 ( 60,35.4)(60,5)34 ( 65,20)[l]@xmath197 however , in fact , this current is not conserved because of an anomaly originating from the triangle diagram in figure ( [ triangleanomaly ] ) .
this means that the current is no longer conserved on the quantum level .
its divergence can be calculated exactly , @xmath198 thus , the chiral @xmath191 is not a symmetry of the theory .
the anomaly explicitly breaks the chiral @xmath191 down to a discrete subgroup , namely the @xmath199 corresponding to @xmath200 , .
this is not the end of the story .
a nonvanishing @xmath185 is not invariant even under this @xmath201 , but breaks it further down to @xmath202 .
this spontaneous breaking of the discrete symmetry implies that we are left with @xmath175 distinct ( discrete ) vacua .
now we will actually calculate the gluino condensate . to do this head on in the strong coupling regime
is , however , impossible . to reach the result
, we will have to take a rather complicated detour to stay with theories that are calculable , like like in crossing an ocean by jumping from small island to small island .
one false step and we will be lost in the sea of noncalculability .
we will go through the following steps : * restrict to @xmath203 gauge group : this is mainly for technical simplicity , and since @xmath204 , the result generalises to any @xmath205 . *
introduce one flavour of matter fields ( fermions and sfermions ) : they will generically break the gauge symmetry , making the gauge fields massive .
the remaining ( massless ) fields can ge assembled in one chiral superfield . *
go to the low energy effective theory : since we are interested in the gluino condensate and hence in the vacuum structure , we can integrate out the heavy modes .
we will find that instantons generate a nonperturbative superpotential .
the resulting scalar potential is a runaway potential , i.e. the minima are at infinite field values
. we will cure this by adding a small mass term for the formerly massless fields .
* at the next step , we will relate the vacuum expectation value of the light scalar fields to the gluino condensate via the konishi anomaly .
the result will still depend on the mass introduced before .
* finally , we take the mass to infinity ( to get rid of the matter fields we introduced in the beginning ) .
luckily , we have full control over the extrapolation procedure , the gluino condensate stays finite in this limit , and eventually we arrive at the result given in eq .
( [ gluinocondensate ] ) . from now on
, we will consider only @xmath206 as gauge group .
the generators are given by the pauli matrices , @xmath207 , @xmath208 .
we also add one flavour of matter , i.e. a dirac spinor in the fundamental representation and its scalar superpartners . the fields can be assembled in two chiral superfields @xmath209 , @xmath210 transforming as doublets under @xmath206 .
the antidoublet is isomorphic to the doublet ( via @xmath211 ) . ] for notational brevity , we will write them as @xmath212 with a `` subflavour '' index @xmath213 to denote both @xmath214 and @xmath215 .
the component fields are called @xmath216 ( scalars ) , @xmath217 ( fermions ) and @xmath218 ( auxiliaries ) .
the matter lagrangean is just the kinetic term , we will add a mass term only later , @xmath219[s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } e^{2v } q + { \settowidth{\textlength}{${\ensuremath{\widetilde{q}}}$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{${\ensuremath{\widetilde{q}}}$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{${\ensuremath{\widetilde{q}}}$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } e^{2v}{\ensuremath{\widetilde{q}}}\right\}= \int{\ensuremath{\text{d}}}^4\theta\,{\ensuremath { { \settowidth{\textlength}{$q$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$q$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$q$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ f e^{2v } q_f\\ & = -{\ensuremath{\bar{\psi}}}_f{\ensuremath{\text{i}}}{{\ensuremath{\rlap{\raisebox{1pt}{$\mspace{4mu}/$}}d}}}\psi_f -d^m{\ensuremath { { \settowidth{\textlength}{$\phi$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } _ f d_m \phi_f + \tfrac{{\ensuremath{\text{i}}}}{{\ensuremath{\sqrt{2}\,}}}\left[{\ensuremath { { \settowidth{\textlength}{$\phi$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } _ f\left(\lambda\psi_f\right)-\text{h.c.}\right]\\ & \quad + d^a { \ensuremath { { \settowidth{\textlength}{$\phi$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } _ f t^a \phi_f + { \ensuremath { { \settowidth{\textlength}{$f$ } \setlength{\overlinelength}{3pt } \addtolength{\overlinelength}{0.4\textlength } \makebox[\textlength][s]{$f$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$f$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } } } _ f f_f .\\ \end{split } \end{aligned}\ ] ]
since there is no superpotential so far , we can forget about the @xmath38-terms .
the @xmath58-term , however , gives a contribution to the scalar potential after elimination of @xmath220 through its equation of motion ( there is a @xmath221 term from the kinetic lagrangean of the gauge field ) , @xmath222[s]{$\phi$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.8\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$\phi$ } } } \hspace{-.8\overlinelength}\hspace{.5\textlength } } } } _ f t^a \phi_f\right)^2 .
\end{split } \end{aligned}\ ] ] this potential vanishes for the following field values : @xmath223 where @xmath224 is a complex number . _ a priori _ , @xmath225 and @xmath226 could have different phases , but the relative phase can be gauged away . that the potential indeed vanishes can be seen from the explicit form of the generators : * @xmath227 : the generators are off - diagonal , and the product of the form @xmath228 vanishes for any @xmath229 , @xmath230 . * @xmath231 : @xmath232 is diagonal , but traceless , and hence @xmath233 so there is a non - compact vacuum manifold parameterised by @xmath224 . the definition of @xmath224 in eq .
( 39 ) seems to be gauge dependent .
in fact , the degree of gauge dependence is minimal , since @xmath234 is gauge independent .
this is clearly seen from @xmath235 hence the flat direction is a one - dimensional complex line .
for any value of @xmath236 , the gauge symmetry is broken and the gauge bosons acquire the mass @xmath237 , proportional to the as yet undetermined parameter @xmath224 .
we will later see that the theory actually wants to develop a large @xmath224 . if then @xmath238 , the theory is weakly coupled and we can actually calculate everything .
the gluons thus behave rather like @xmath239 bosons .
the gauge symmetry breaking and higgs mechanism reshuffle the degrees of freedom , since the vector bosons eat some part of the scalar fields acquiring longitudinal components .
the degrees of freedom before and after gauge symmetry breaking are listed in table [ dof - table ] .
sf & component field & b & f & sf & component field & b & f & @xmath240 + & 3 gluons & 6 & & & 3 @xmath239 bosons & 9 & & } 44 mm + & 3 gluinos & & 6 & & 3 gluinos & & 6 & + & 2 complex scalars & 4 & & & 3 gauge fermions & & 6 & + & 2 weyl fermions & & 4&&3 real scalars & 3 & & + & 2 complex scalars & 4 & & & 1 complex scalar & 2 & & + & 2 weyl fermions & & 4 & & 1 weyl fermion & & 2 & + to analyse the vacuum structure , we can integrate out the heavy fields and consider just the low - energy effective action .
all we are left with is a chiral multiplet @xmath229 defined by the formula @xmath241 its scalar component is the higgs field @xmath242 , and there are superpartners , @xmath243 for the resulting lagrangean , we get the usual kinetic term and nothing else , at the perturbative level . however , instantons induce a nonperturbative contribution to the superpotential
. this contribution is severely restriced by the non - anomalous @xmath9-symmetry of the theory , on which i will dwell momentarily .
the only allowed form is @xmath244[s]{$x$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$x$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } x + \left\{c \int{\ensuremath{\text{d}}}^2\theta\frac{\lambda^5}{x^2 } + \text{h.c.}\right\}. \end{aligned}\ ] ] the nonperturbative nature of the superpotential can be seen from its dependence on @xmath245 , as opposed to perturbative terms which could contain only powers of @xmath246 .
division by @xmath229 is to be understood as an expansion in @xmath19 , @xmath247 now i return to the issue of the @xmath9-symmetry , which is a global ( chiral ) @xmath10 symmetry of the fundamental theory with respect to rotations of the fermions and sfermions .
@xmath9-symmetry does not commute with supersymmetry , as it requires a rotation of the grassmann parameters @xmath19 and @xmath248 .
assume that @xmath249 .
( naturally @xmath250 . )
since the kinetic term of the gauge fields is proportional to @xmath251 the invariance of the lagrangean then requires that @xmath252 which entails , in turn that @xmath253 .
the @xmath10 charges of the matter fermions are determined by anomaly cancellation : we want the @xmath9 current to be conserved not only classically , but , rather , exactly conserved .
the exactly conserved @xmath9 current is @xmath254 the emergence of the factor 2 in front of the matter fermions is explained by the fact that for @xmath206 the contributions to the anomaly of the gaugino is twice larger that that of the matter fermions , see footnote before eq .
therefore , from the law of the @xmath9 rotation of @xmath33 above we deduce that @xmath255 . what remains to be established is the @xmath9 charge of the sfermion field @xmath216 .
the easiest way is to examine the vertex @xmath256 in the lagrangean .
its invariance requires that @xmath257 , which implies that @xmath258 .
finally , let us note that if @xmath259 then @xmath260 . combining this with the above results we conclude that the matter superfields are transformed as @xmath261 while @xmath262 .
the @xmath9-charges of the fields are collected in table [ r - table ] .
the @xmath9-charge @xmath29 means that a field @xmath39 transforms as @xmath263 .
r149pt [ cols="^,^,^,^,^,^",options="header " , ] the non - anomalous @xmath9-symmetry must be conserved in the low - energy effective action which emerges after integration over the heavy fields .
note that @xmath264 ; in other words , the @xmath9 charge of @xmath229 is unity .
since the kinetic terms are @xmath265[s]{$x$ } \hspace{-.5\textlength}\hspace{-\overlinelength}\hspace{.625\overlinelength } \overline { \makebox[\overlinelength][s ] { \vphantom{$x$ } } } \hspace{-.625\overlinelength}\hspace{.5\textlength } } x$ ] , there is no restriction from @xmath9-symmetry on the kinetic term .
the superpotential is quite constrained , however . indeed ,
if we impose analyticity ( @xmath266 ) , we find @xmath267 so the superpotential must be @xmath268 . to guarantee the appropriate mass dimension , we then have to insert the fifth power of @xmath174 , the only dimensionful parameter of the theory .
note that exactly @xmath245 emerges from the one - instanton measure .
now , the superpotential is completely fixed up to a coefficient .
this coefficient , in principle , could be zero .
therefore , strictly speaking , we do not yet know whether a superpotential is generated nonperturbatively , but if it is , it must be of the form given in eq .
( [ nonperturbativesuperpotential ] ) . we will now check that instantons do generate this superpotential .
instantons ( at weak coupling ) are classical solutions to the gauge field equations of motion with non - trivial topology in field space and finite action @xmath269 .
their contributions are suppressed by @xmath270 , and so perturbative effects ( corresponding to @xmath271 ) are dominant .
but in the case of the superpotential , there are no perturbative contributions ( this is the famous non - renormalisation theorem for superpotentials ) , so the instanton effects decide the outcome .
our model is constructed in such a way that at @xmath272 the gauge symmetry of the model is completely broken and _ all _ gauge bosons become massive ( in fact , very heavy if @xmath224 is large ) .
why is it that the requirement of the complete breaking of the gauge symmetry is so important ?
if all gauge bosons are heavy ( much heavier than @xmath174 ) the theory is at weak coupling .
this means that all calculations , including instanton calculations , are under complete control and are reliable .
if , on the other hand , a non - abelian subgroup remains unbroken , the analysis is dragged , with necessity , into a strong coupling domain , where theoretical control is lost .
technically this loss manifests itself in the fact that , with unbroken non - abelian gauge group , integrals over the instanton size become divergent , and there is no way one can make instanton calculation well - defined . after this remark
let us pass to determination of the coefficient @xmath184 in the model at hand .
usually , instanton analyses are quite cumbersome .
supersymmetry makes it rather trivial , however .
the instanton suppression factor @xmath273 quoted above is just the classical instanton action . to carry out the actual calculation we must integrate over fluctuations of all fields over the instanton solution . to this end one expands the field fluctuations in modes .
there are a few zero modes , and an infinite number of non - zero modes .
handling non - zero modes is the most labour - intensive and time - consuming part of the problem . in supersymmetric theories one can prove , however , that all nonzero modes cancel each other , so the instanton analysis reduces to that of zero modes @xcite .
pictorially , the instanton is represented by a circle with an @xmath274 in it in figure [ f : instantons ] . in the model under consideration
the instanton has four gaugino zero modes and two zero modes corresponding to two subflavors of the matter fermions ( see fig .
a ) . in order to determine the number of the zero modes one
does not have to actually find them .
the anomaly relation ( [ e : triangleanomaly ] ) , combined with @xmath275 fixes the number of the gaugino zero modes , while a similar anomaly relation for the matter fermion axial current determines the number of the matter fermion zero modes .
all fermion lines are directed outward , so the instanton breaks the chiral symmetry .
it makes all anomalies explicit .
the chiral currents of both gauginos and matter fermions are anomalous .
as i have already mentioned , however , the combination @xmath276 is not . at first sight
, it seems that the fact that there are six fermion zero modes in the instanton background forbids generation of the effective lagrangean ( 42 ) ( see also eq .
indeed , this effective lagrangean requires two and only two fermion zero modes , since the only term surviving after integration over @xmath277 is the last one in eq .
let us not rush to hasty conclusions , however .
the fundamental lagrangean of our model contains the vertex @xmath278 .
it allows one to convolute a pair of the gaugino zero modes with those of the matter fields exploiting the non - vanishing vacuum value of @xmath279 , fig .
b. next , we use the same vertex again to trade two remaining gluino zero modes into @xmath280 ( fig .
c ) . thus ,
the one - instanton contribution does generate a @xmath281-term , and supersymmetry ensures that the @xmath282 term needed for the scalar potential is generated as well @xcite , although the emergence of this term is harder to see ( an off - shell background field would be required ) .
this contribution also has the @xmath245-factor needed for dimensional reasons .
higher instanton effects vanish .
for instance , two instantons would produce @xmath283 in the superpotential , which , as we know on general grounds , is impossible . now that we have established generation of the superpotential in eq .
( [ nonperturbativesuperpotential ] ) , we can use the expansion in ( [ oneoverxsquaredexpansion ] ) and perform the @xmath277-integration .
note that the numerical constant @xmath184 is fully determined by normalizations of the zero modes , which are very well known .
for further technical details i refer to @xcite . eliminating the auxiliary field
, we obtain the scalar potential @xmath284 which goes to zero at infinity and does not have a minimum at finite field values .
such a situation is usually referred to as a _
runaway vacuum_. we can , however , cure this disaster by adding a small quark mass to the superpotential of the fundamental theory ( small for now , we will later take the limit @xmath285 , so the ( s)quarks will drop out again ) . with the mass term switched on the superpotential
will take the form @xmath286 now @xmath287 ( and hence @xmath288 ) vanishes at finite values of @xmath229 , namely , at @xmath289 these are the minima of the scalar potential
. as was expected , there are two vacua rather than @xmath290 .
the point is that @xmath291 is gauge invariant , see the definition before eq .
( 41 ) . ] .
the vacuum expectation value of @xmath291 is stabilized far away from the origin .
this justifies _ a posteriori _ our starting assumption that the gauge bosons are very heavy and can be integrated out .
thus , we found @xmath292 .
remember , however , that our task was to determine the gluino condensate , @xmath293 .
since the gluon and gluino fields are integrated out , it might seem that information on @xmath293 is lost .
this is not the case we can relate the vacuum expectation value of @xmath291 to the gluino condensate by virtue of the konishi anomaly .
the konishi anomaly is a supersymmetric generalisation of the chiral anomaly , i.e. the nonconservation of the axial current which is proportional to @xmath294 .
this product is contained in the highest component of the superfield @xmath295 , and the konishi relation is @xmath296 where @xmath297 is the field strength superfield and the @xmath239 appearing in the last term is the superpotential
. spelled out explicitly , the last term is @xmath298 .
however , the left hand side of eq .
( [ konishi ] ) is a total superspace derivative , so its expectation value in any supersymmetric vacuum must vanish .
this , in turn , gives us the relation we desire , namely @xmath299 or in particular its lowest component , @xmath300 inserting the expression for @xmath291 from eq .
( [ hsquaredvev ] ) , we get @xmath301 the above result was computed at small @xmath28 . however , in the very end we want to take the limit @xmath285 .
how do we know that eq .
( [ lambdalambdakonishi ] ) is still valid ?
here we can exploit the fact that @xmath303 is a chiral quantity which must depend _ analytically _ on the ( complex ) mass parameter @xmath28 , so its functional dependence is determined by its singularity structure .
furthermore , we can consider the @xmath9-symmetry of the theory .
invariance of @xmath304 under the @xmath9 transformation requires the @xmath9-charge of @xmath28 to be @xmath305 .
since the @xmath9-charge of @xmath306 is @xmath307 , the analytic @xmath28 dependence of @xmath303 implies that @xmath308 with @xmath309 , which means that eq .
( [ lambdalambdakonishi ] ) is exact at all values of @xmath28 . still , the limit @xmath285 of eq .
( [ lambdalambdakonishi ] ) seems not to make sense , since it would imply @xmath310 as well .
however , the scale @xmath311 is the scale of a theory with one flavour . in the @xmath285 limit
, the theory reduces again to pure sym theory , where the relevant scale @xmath312 may and does differ .
actually , there is an exact result relating the scales , @xmath313 which follows from the matching of the gauge coupling running with appropriate @xmath161 functions .
we see that the right - hand side of this expression is precisely the right - hand side in eq .
( [ lambdalambdakonishi ] ) .
so we can now safely take the limit and arrive at @xmath314 this is exactly the expression advertised in eq .
( [ gluinocondensate ] ) restricted to @xmath206 , since the exponential just reduces to @xmath315 .
i am very grateful to misha shifman for many comments and corrections of the manuscript .
i also would like to thank the organisers of the saalburg summer school for two very interesting weeks in the forest .
in this notes we are always using two component ( weyl ) spinors in van - der - waerden notation , i.e. with dotted and undotted greek letters as indices , indicating the @xmath318 or @xmath319 representation of @xmath320 . these indices are raised and lowered with the @xmath321-symbol , @xmath322 in products of spinors , the convention is that undotted indices are contracted from upper left to lower right , and dotted ones from lower right to upper left : @xmath323 since the spinors anticommute and raising and lowering of a contracted index also gives a factor of @xmath78 , we have @xmath324 und also @xmath325
. the rle of the @xmath326 matrices is played by the @xmath18 matrices , @xmath327 with @xmath328 the pauli matrices .
we also need the `` commutator '' of the @xmath18 s , @xmath329 which are antisymmetric in @xmath330 .
if one of the spinor indices is raised or lowered , they are also symmetric in @xmath331 or @xmath332 , respectively .
the transition to usual four - component dirac spinors @xmath73 is most easily done in the chiral ( or weyl ) basis for the @xmath326-matrices , @xmath339 in this basis , a dirac spinor @xmath73 directly reduces to two weyl ones , @xmath340 and kinetic and mass terms become @xmath341 some care has to be taken when raising ore lowering indices on spinorial derivatives ( there is an additional minus sign ) .
they act in the following way ( note the order of the indices on the @xmath321 s ) : @xmath344 vector indices @xmath28 can be converted to a pair of spinor ones @xmath346 by contraction with the @xmath18 matrices , specifically for vectors @xmath347 @xmath348 and for antisymmetric tensors @xmath349 , @xmath350 where the dual tensor appears as the imaginary part .
99 m. a. shifman and a. i. vainshtein,``instantons versus supersymmetry : fifteen years later , '' in m. shifman _
itep lectures on particle physics and field theory _
( world scientific , singapore , 1999 ) , vol .
, hep - th/9902018 .
m. a. shifman and a. i. vainshtein , `` on holomorphic dependence and infrared effects in supersymmetric gauge theories , '' nucl .
b * 359 * ( 1991 ) 571 .
m. a. shifman and a. i. vainshtein , `` on gluino condensation in supersymmetric gauge theories .
su(n ) and o(n ) groups , '' nucl .
b * 296 * ( 1988 ) 445 ] .
i. affleck , m. dine and n. seiberg,``dynamical supersymmetry breaking in supersymmetric qcd , '' nucl .
b * 241 * , 493 ( 1984 ) .
e. witten,``constraints on supersymmetry breaking , '' nucl .
b * 202 * , 253 ( 1982 ) .
y. a. golfand and e. p. likhtman , `` extension of the algebra of poincare group generators and violation of p - invariance , '' pisma zh .
* 13 * ( 1971 ) 452 [ jetp lett . *
13 * , 323 ( 1971 ) ; reprinted in _
supersymmetry _ , ed .
s. ferrara , ( north - holland / world scientific , 1987 ) vol . 1
r. haag , j. t. opuszanski and m. sohnius , `` all possible generators of supersymmetries of the s matrix , '' nucl .
b * 88 * , 257 ( 1975 ) [ reprinted in _
supersymmetry _ , ed .
s. ferrara , ( north - holland / world scientific , 1987 ) vol . 1
, p. 51 ] .
a. salam and j. strathdee , `` supergauge transformations , '' nucl .
b * 76 * , 477 ( 1974 ) .
f. berezin , _ methods of second quantization _ , ( academic press , new york , 1966 ) . | this article contains lecture notes of m. shifman from the saalburg summer school 2004 .
the topic is supersymmetric yang - mills theory , in particular the gluino condensate in pure susy gluodynamics .
desy-04 - 248 + from super - yang - mills to qcd + + _ deutsches elektronensynchrotron desy + _ notkestr .
85 + _ 22603 hamburg + _ germany _ _ _ _ |
the conductivity of a conductor - insulator composite material is characterized by a strong dependence on the volume fraction @xmath4 of the conducting phase present in the system , and is generally understood as a percolation phenomenon arising from the electrical connectivity between neighboring or adjacent conducting particles .
specifically , percolation theory considers the conducting particles as either electrically connected , with some finite inter - particle conductance , or disconnected.@xcite the introduction of this sharp cut - off implies then that below a specific fraction @xmath5 ( the percolation threshold ) the conductivity is zero because there is no sample - spanning network of connected particles , while for @xmath6 such network is formed and the conductivity increases as @xmath7 , where @xmath8 is a critical exponent .
it is easy to see that the notion of a sharp cut - off applies well to composites made of large ( of the order of one micron or more ) conducting particles , for which two particles can be considered electrically connected only if they essentially touch each other , it is less clear for the case in which the conducting particles have sizes limited to a few nanometers.@xcite in that situation even if the particles do not physically touch each other , their mean separation in the composite is such that electrons can still flow via tunneling processes from one particle to the other .
the resulting tunneling conductance decays exponentially with the inter - particle distance over a characteristic tunneling length @xmath2 which is of the order of a fraction to a few nanometers depending on the material properties .
since the tunneling decay does not imply any sharp cut - off , the basic assumption of percolation theory in describing composites made of nanometric conducting fillers is not justified a priori .
in fact , it turns out that nanocomposites in the dielectric regime , for which the conducting particles do not touch each other , are better explained by inter - particle tunneling , with no imposed sharp connectivity criterion , than by the classical percolation theory.@xcite the resulting conductivities of these systems follow therefore hopping - like ( or tunneling - like ) behaviors,@xcite with no critical percolation thresholds as a function of @xmath4 . the possibility of having percolation - like or tunneling - like regimes depending on the size @xmath1 of the conducting particles compared to the tunneling length @xmath2 , as suggested above and in ref .
[ ] , does not appear to have been further elaborated in the literature , despite of its relevance to the understanding of transport properties in conductor - insulator composites . in the present paper
we address this issue by considering a global tunneling network ( gtn ) model of conductor - insulator composites,@xcite where each conducting particle is connected to all others via tunneling processes .
we show that this model permits to treat the percolation and the tunneling regimes on equal footing , and that one can switch from one regime to the other depending not only on @xmath0 but , most notably , also on the specific distribution of the conducting phase in the composite .
we illustrate this behavior by considering two idealized realizations of conductor - insulator composites : a lattice model , where conducting spherical particles of diameter @xmath1 occupy randomly a fraction of the sites of a regular lattice , and a continuum model , where the conducting particles are dispersed with an equilibrium distribution in a continuous insulating medium . by using both the effective medium approximation ( ema ) and monte carlo ( mc ) calculations , we find that the conductivity of the continuum model has a tunneling - like behavior independent of the value of @xmath0 , while the lattice model displays a percolation behavior of the conductivity only for sufficiently large @xmath0 values , which gradually crosses over to a tunneling - like regime as @xmath0 decreases .
for @xmath9 we show that the conductivity of the lattice model is basically indistinguishable from that of the continuum case , establishing therefore a crossover point from percolation to tunneling behaviors .
this paper is organized as follows . in sec .
[ ema ] we introduce the gtn model , we formulate the ema for both lattice and continuum models and we calculate the resulting ema conductances . in sec .
[ mc ] we present our mc calculations for both the lattice and continuum models , and in sec .
[ crossover ] we discuss the crossover between percolation and tunneling regimes . discussions and conclusions are given in sec .
the gtn model for conductor - insulator composites is defined by considering @xmath10 identical conducting particles contained within a volume @xmath11 .
this defines the volume fraction @xmath12 of the conducting phase , where @xmath13 is the particle density and @xmath14 is the volume of a single particle . for the purpose of the present work we limit the analysis to the relatively simple case of spherical particles of identical diameter @xmath1 , so that @xmath15
next , we assume that the tunneling conductance between any two particles centered at @xmath16 and @xmath17 is given by : @xmath18,\ ] ] where @xmath19 is a constant contact " conductance which in the following we shall set equal to unity , @xmath2 is the characteristic tunneling length , and @xmath20 is the distance between two sphere centers . for simplicity
we further assume that selective tunneling mechanisms arising from excitation energies can be safely neglected ( as in the case for nearest neighbor hopping@xcite ) .
the set of all tunneling conductances of eq
. defines a resistor network whose conductivity depends on @xmath2 and @xmath1 , as well as on the volume fraction @xmath4 and on the specific distribution of the particle centers .
as shown below , it turns out that all these dependencies are well captured by the ( single bond ) ema applied to the tunneling resistor network.@xcite this is given by the solution of the following equation : @xmath21\bar{g}}\right\rangle=0,\ ] ] where @xmath22 is the effective bond conductance and @xmath23 and @xmath24 run over the positions of the @xmath10 particles . in the above expression ,
the symbol @xmath25 denotes the statistical average over all realizations of the @xmath10-particle system .
furthermore , @xmath26 and @xmath27 are the coordination numbers of the two ends of a resistor @xmath28 , _
i.e. _ , given a particle at @xmath23 ( @xmath24 ) , @xmath26 ( @xmath27 ) is the number of particles that are electrically connected to @xmath23 ( @xmath24 ) .
since it is assumed that all particles are connected to each other regardless of their relative distances , @xmath29 . in this way , eq
. reduces to @xmath30\bar{g}}\right\rangle=0.\ ] ] by noticing that @xmath31 and by using eq .
, we can write eq .
more conveniently as @xmath32 + 1}\right\rangle=2,\ ] ] where for large @xmath10 the dimensionless conductance @xmath33\bar{g}/g_0\simeq n\bar{g}/2g_0 $ ] coincides with the conductance between two random nodes of the network with all bond conductances equal to @xmath34.@xcite by multiplying each term of the summation over @xmath35 by @xmath36 , where the integration is extended over the entire volume @xmath11 , the left - hand side of eq
. becomes @xmath37 + 1}\right\rangle\nonumber \\ & = \int\ !
d{\bf r}\left\langle\frac{1}{n}\sum_{i\neq j}\delta({\bf r}-{\bf r}_{ij})\right\rangle \frac{1}{g^*\exp[2(r - d)/\xi]+1}.\end{aligned}\ ] ] the above expression can be rewritten in terms of the pair distribution function @xmath38 by noticing that , by definition:@xcite @xmath39 so that eq .
reduces to @xmath40 + 1}=2,\ ] ] where @xmath41 is the radial distribution function ( rdf ) and the upper limit of the integration has been set to infinity because of the exponential decay of the integrand .
note that eq .
[ ema5 ] is similar to the result obtained in ref . .
an alternative and useful version of eq . can be obtained by introducing the function @xmath42 + 1}=\frac{1}{\exp[2(r - r^*)/\xi]+1},\ ] ] with @xmath43 and integrating eq . by parts : @xmath44 where @xmath45 is the cumulative coordination number function ( _ i.e. _ , the number of spheres whose centers are within a distance @xmath46 from the center of a given sphere ) . for the lattice ( dotted lines ) and
the continuum ( solid lines ) models . for the lattice model
the vertical dotted lines indicate the positions of the delta - peaks centered at @xmath47 for @xmath48 .
the @xmath49 curves for different @xmath4 values of the continuum model have been obtained by using the results of ref . .
( b ) cumulative coordination number @xmath50 , eq .
, for the lattice ( dotted lines ) and the continuum ( solid lines ) models for different @xmath4 values . ]
as it is clear from the ema equations and , all the informations on the spatial distribution of the conducting particles is contained in the rdf @xmath49 , whose dependencies on @xmath46 and @xmath51 govern the behavior of the ema conductance @xmath52 . here ,
we consider two possible realizations of a conductor - insulator composite which corresponds to two rather extreme forms of @xmath49 . in the first case
, we consider a simple cubic lattice with a lattice constant equal to the sphere diameter @xmath1 , where only a fraction @xmath53 of the lattice sites is occupied , randomly , by the spherical conducting particles , while the remaining fraction @xmath54 is occupied by insulating spheres of equal diameter ( as in the scher and zallen model of ref . ) .
the @xmath23-th conducting particle occupies therefore the position @xmath55 with probability @xmath56 , where @xmath57 is the direct lattice vector running over all @xmath58 sites of the cube . in this way ,
the pair distribution function defined in eq .
reduces to @xmath59 which leads to an rdf of the form : @xmath60 where @xmath61 is the number of the @xmath62-th nearest neighbors being at distance @xmath47 from a reference particle set at the origin .
the characteristic feature of this fractionally occupied lattice model is therefore that its rdf , shown in fig .
[ fig1](a ) by dotted lines , is given by delta - peaks whose positions do not change with @xmath53 . as we shall see in the following , this feature is directly related to the appearance of a percolation behavior of the ema conductance @xmath52 for large @xmath0 values .
the second model of particle distribution is given by an equilibrated dispersion of impenetrable spheres in the continuum .
typical examples of the resulting @xmath49 for this case , obtained by using the formula provided in ref . , are shown in fig .
[ fig1](a ) by solid lines and for different values of the volume fraction @xmath4 . as opposed to the lattice case ,
the rdf of an equilibrium dispersion of hard spheres is continuous and nonzero in the whole @xmath63 range , which we anticipate here to be the pre - requisite for an hopping behavior of the composite conductivity . in fig .
[ fig1](b ) we compare the coordination number function @xmath50 , eq .
, for the lattice ( dotted lines ) and continuum ( solid lines ) models .
contrary to the continuum model in which @xmath50 increases smoothly with @xmath46 , the lattice @xmath50 has a step - like increase . at large @xmath46
however the main contribution of the lattice @xmath50 goes as @xmath64 , meaning that the distribution of spheres in a sparsely occupied lattice is basically that of point particles in the continuum .
let us start by considering the solution of the ema equation for the fractionally occupied cubic lattice case . by using eq .
we find that eq .
reduces to : @xmath65 + 1}=2.\ ] ] in the limiting case of very large particle sizes such that @xmath66 , only the first ( @xmath67 ) nearest neighbors with @xmath68 and @xmath69 contribute to the summation in eq . , which reduces to @xmath70 .
hence , in this limit , the resulting conductivity @xmath71 has a percolation behavior and vanishes at the critical fraction @xmath72 . for finite @xmath0 values
, the next nearest neighbors also contribute to the network conductivity .
for example , by assuming that @xmath0 is large enough to retain only the terms up to the second nearest neighbors ( @xmath73 and @xmath74 ) we get that eq
. becomes @xmath75 + 1}=2,\ ] ] whose solution still behaves as @xmath76 for @xmath77 but remains finite , albeit exponentially small , at lower @xmath53 values .
this is because the tunneling to the @xmath78-nd nearest neighbors vanishes only at @xmath79 .
considering the whole set of neighbors when @xmath53 decreases , one finds that the solution of eq .
becomes characterized by a monotonous decrease of @xmath52 punctuated by sharp ( for small @xmath0 values ) drops at @xmath80 with @xmath81 .
furthermore , in the vicinity of each @xmath82 , the conductance follows the ema power law behavior @xmath83 for @xmath84 .
this feature is illustrated in fig .
[ fig2](a ) where we show numerical solutions of eq . for different values of @xmath0 . for @xmath85 , the first percolation transition at @xmath72 ( @xmath86 ) causes a drop of the conductivity of about @xmath87 orders of magnitude ( not shown in the figure ) compared to the conductivity at the close packing fraction @xmath88 that corresponds to the lattice sites that are all occupied by the conducting spheres .
such drop of the conductivity is well beyond the maximum range of conductivity values measured in real composites,@xcite and so , for all practical purposes , the @xmath85 case behaves as @xmath89 with @xmath90 being the ema critical exponent .
for @xmath91 the first transition at @xmath92 leads to a drop of @xmath52 of only @xmath93 orders of magnitude , and the transitions at lower values of @xmath4 are clearly visible in the figure . for lower values of @xmath0 , the drops of @xmath52
are further reduced and the transitions are much smoother , due of course to the fact that for these @xmath0 values the probability of tunneling to neighbors that are farther apart is enhanced . for @xmath94
the variation of @xmath52 is so smooth in the whole range of @xmath4 that the underlying lattice structure can be considered as completely irrelevant .
in contrast with the lattice case , the ema conductance @xmath52 resulting from an equilibrium distribution of conducting spheres in the continuum does not display any percolation behavior even for large @xmath0 values .
this is shown in fig .
[ fig2](b ) where solutions of eq .
, with @xmath49 as given in ref .
, are plotted for the same @xmath0 values of fig .
[ fig1](a ) .
the lack of percolation behavior in this case is due to the fact that the corresponding rdf [ solid lines in fig .
[ fig1](a ) ] is always nonzero for @xmath63 and it does not vary much even for @xmath4 values close to the packing fraction of the simple cubic lattice ( @xmath95 ) . to see how the corresponding tunneling behavior arises ,
let us consider the ema in the form of eq . .
since , as shown in fig .
[ fig1](b ) , the coordination number function @xmath50 is a smooth increasing function of @xmath96 and , given that @xmath97 is peaked at @xmath98 with spread @xmath99,@xcite we can approximate eq . by @xmath100
so that , by applying eq . , the ema conductance becomes @xmath101\ ] ] where @xmath102 is such that eq .
is satisfied . in passing ,
we note that eqs . and represent the ema equivalent of the critical path approximation ( cpa ) of refs .
.@xcite for sufficiently large @xmath102 ( _ i.e. _ , small @xmath52 ) the coordination number goes as @xmath103 and from eq .
we obtain that @xmath104,\ ] ] which coincides , if the coefficient @xmath105 in the exponent is replaced by @xmath106 , with the low - density hopping behavior as obtained from the cpa.@xcite in the same low - density limit , eq .
( with a slightly different coefficient in the exponent ) has been derived also in ref . by using an ema - based procedure similar to the one presented here .
the hopping - like ( or tunneling - like ) dependence of @xmath52 for the continuum model is further illustrated in fig . [ fig2](c ) where the results of fig
. [ fig2](b ) have been re - plotted as a function of @xmath107 ( dotted lines ) . in fig .
[ fig2](c ) we show for comparison also the @xmath52 versus @xmath107 curves ( solid lines ) of the lattice model results of fig . [ fig2](a ) , which tend , with the decrease of @xmath0 , to the tunneling behavior of the continuum model .
an asymptotic equivalence between the two models is indeed expected because , as noted above , the underlying microstructures become irrelevant for small @xmath0 .
mathematically , this can be seen from eq . since @xmath97 averages @xmath50 within a distance @xmath99 around @xmath102 , and thus the differences between the lattice and the continuum models are blurred for sufficiently large @xmath2 .
note also that for the lattice model @xmath108 for large @xmath102 , implying that @xmath52 tends asymptotically to the results given by eq .
, as shown by the solid lines in fig .
[ fig2](c ) .
let us turn now to a comparison of our ema results of the previous section with full mc calculations of the conductivity for both the lattice and continuum models .
for the continuum we have generated equilibrium distributions of impenetrable spheres inside a cubic cell with periodic boundary conditions following the procedure that we outlined previously in refs . .
for the lattice model , we simply populated a given fraction of the cubic lattice sites with conducting spheres with diameter equal to the cubic lattice constant . to calculate the conductivity resulting from the lattice and continuum models , we ascribed to each pair of particles the tunneling conductance of eq . , and performed a numerical decimation of the resulting resistor network.@xcite to reduce computational times of the decimation procedure we introduced an artificial maximum distance between the particles beyond which the resulting bond conductance can be safely neglected.@xcite
the results of the calculated conductivity are plotted in fig .
[ fig3 ] for the same parameter values that were used for fig .
each symbol is the outcome of @xmath109 realizations of systems of @xmath110 spheres .
the logarithm average of the results was considered since , due to the exponential dependence of eq . , the distribution of the computed conductivities was approximately of the log - normal form .
it is interesting to notice that there is an overall quasi - quantitative agreement between the ema and the mc results , for both lattice and continuum models , meaning that the ema formulation of sec .
[ ema ] captures well the physics of the problem . for example , the first percolation threshold at @xmath72 ( @xmath111 ) obtained from the ema on the cubic lattice is very close to the critical value @xmath112 ( @xmath113 ) for the site percolation problem on the simple cubic lattice.@xcite an important expected difference is however found in the region @xmath114 where the mc conductivity for the lattice case should follow a @xmath89 dependence with a critical exponent @xmath115 instead of the ema exponent of @xmath90.@xcite this is indeed verified in fig .
[ fig4 ] where the mc conductivities for @xmath91 ( a ) and @xmath116 ( b ) are both fitted with @xmath117 , which is slightly lower than @xmath115 due to finite size effects ( for the same reason we find @xmath118 instead of @xmath113 ) .
similar critical behaviors are expected in the vicinity of all consecutive percolation thresholds @xmath119 with @xmath120 , but due to the limited number of particle densities considered in our mc calculations and the close proximity of successive @xmath82 values we have been able to fit only the @xmath91 case in the vicinity of the second percolation threshold .
we have found @xmath121 , which agrees within errors with the exponent at @xmath92 , and @xmath122 ( @xmath123 ) which is close to the expected value @xmath124 ( @xmath125).@xcite these same values of @xmath8 and @xmath126 appear to reproduce quite well the percolation behavior around the second percolation threshold also for the @xmath116 case [ fig . [ fig4](b ) ] . for ( a ) @xmath91 and ( b ) @xmath116 .
the solid lines are the best fits with the percolation formula @xmath127 . ]
another important agreement between the ema and the mc results is the tendency of the lattice model conductivity to become tunneling - like as @xmath0 decreases .
this is illustrated in fig .
[ fig3](c ) where the mc conductivities of fig . [ fig3](a ) and fig . [ fig3](b )
are plotted as functions of @xmath107 .
the asymptotic regime for small @xmath128 of the continuum model ( open symbols ) follow eq . with the coefficient @xmath105 replaced by @xmath106 , as obtained from the critical path approximation applied to a dilute system of impenetrable spheres.@xcite in complete analogy with the ema results , the conductivities of the lattice model [ filled symbols in fig .
[ fig3](c ) ] steadily tend to this tunneling regime as @xmath0 decreases , and for @xmath94 the mc results of the two models are practically indistinguishable for all values of @xmath4 .
both ema and mc results point toward a substantial equivalence between the cubic lattice and the continuum models for sufficiently small values of @xmath0 . given that the two models considered represent two extreme limits of how the conducting particles may be ideally arranged in an insulator - conductor composite , such equivalence is important for the understanding of the transport properties in real composites , whose microstructure is neither completely lattice - like nor exactly an equilibrium distribution in the continuum .
we find it therefore useful to define a measure for the deviation between the lattice and continuum models in order to follow quantitatively how these two extremes approach each other as @xmath0 decreases . to this end , we introduce the following quantity : @xmath129 where @xmath130 and @xmath131 are the conductivities for the lattice and the continuum models , respectively , and the @xmath132 is the logarithm with base @xmath133 . hence , according to eq .
, @xmath134 if for a certain @xmath4 value @xmath130 differs from @xmath131 by @xmath135 orders of magnitude . in fig .
[ fig5 ] we plot the maximum value @xmath136 and the mean value @xmath137 of @xmath138 as calculated over the entire range of @xmath4 considered and for several values of @xmath0 .
the open symbols are the @xmath136 and @xmath137 values as extracted from the ema calculations while the filled symbols refer to the mc results . as clearly seen in the figure , when @xmath0 decreases both ema and mc data display a steady decrease of @xmath136 and @xmath137 . for @xmath139 ,
@xmath136 becomes less than unity , which means that @xmath130 and @xmath131 differ at most by less than one order of magnitude in the whole range of @xmath4 values .
we also note that for @xmath139 the mc results display a somewhat stronger decrease than the ema results .
this feature does not seems to be due to finite size effects in the mc calculations , and we attribute it to a real deviation from the ema results .
we can then identify the value of @xmath3 as a crossover between percolation and hopping regimes below which the conductivities of the lattice and continuum models differ significantly less than one order of magnitude . and mean value @xmath137 extracted from the whole @xmath4 dependence of eq . and for several values of @xmath0 .
the open symbols are the ema results , while the filled symbols refer to the mc calculations . ]
as repeatedly stressed above , the lattice and continuum models are rather idealized representations of the true microstructure of real conductor - insulator composites .
however they also define two extreme boundaries which delimit somehow all the possible configurations that can be found in isotropic and homogeneous composites . for this reason , the results of fig .
[ fig5 ] can be considered as upper boundaries of more realistic composites , and our @xmath3 result probably underestimates the crossover point between percolation and hopping for composites whose microstructure deviates from an ideal lattice model .
the results presented in the previous sections consistently show that the gtn model of conductor - insulator composites , where each particle is connected via tunneling to all others , is capable of explaining the appearance of both percolation and hopping regimes depending on few characteristics of the composite .
in particular , we have identified the ratio @xmath0 and the composite microstructure as the relevant variables that control the switch from one regime to the other . for large values of @xmath0
, percolation arises in composites whose microstructure can approach a regular lattice that is fractionally occupied by conducting particle , while a hopping regime of the conductivity is always obtained for equilibrium distributions of conducting particles in the continuum . as @xmath0 decreases , the conductivity of the lattice - like composites gradually loses its percolating character and approaches the hopping regime which characterizes the continuum limit . for @xmath9
the composite conductivity displays hopping behavior independent of the specific microstructure and is practically indistinguishable from that arising from an equilibrium distribution of particles in the continuum . given that the values of the tunneling characteristic length @xmath2 do not exceed a few nanometers , our analysis predicts that composites whose conducting fillers have nanometric sizes should always display a hopping - like ( _ i .
e. _ , percolation - less ) behavior independent of the particular distribution of the conducting particles in the composite . instead
, composites with conducting filler sizes larger than a fraction of microns should display percolation or hopping behaviors depending on the whether the microstructure is more lattice - like or more continuum - like , respectively . in view of the above
, it appears then reasonable that dilute filler polymer - based composites will display hopping - like behavior also for filler diameters substantially larger than @xmath2 ( _ i .
e. _ , in the order of some hundreds of nanometers ) as we have verified recently in ref . .
at the same time , our theory also explains the conductivity behavior of composites made of mixtures of hard conducting and insulating particles .
in particular , we note that in dense ensembles of conducting spheres the arrangement in continuum systems is similar to that of lattices.@xcite in this respect , the results of ref . on
the conductivity of co - sputtered ni - sio@xmath140 cermets represent a nice example of tunneling - driven percolating behavior on a lattice - like microstructure .
indeed , the data of ref .
display multiple percolation thresholds as the concentration of conducting ni grains is reduced , very much like the behavior shown in fig .
[ fig4 ] for the mc conductivity of a fractionally occupied lattice model .
the difference of about @xmath141 orders of magnitude between the conductivity at the largest ni concentration and that at the first percolation threshold can be reproduced by our model by setting @xmath142 which , by using the measured mean ni grain size @xmath143 nm,@xcite leads to @xmath144 nm , which is of the expected order of magnitude ( see e.g. refs . ) and compares well with the @xmath2 values extracted from other composites with spherical fillers.@xcite it should be noted however that the lattice model applies only partially to the ni - sio@xmath140 data because the nonuniversal value of the conductivity exponent @xmath8 observed in the vicinity of the second percolation threshold in ref .
can not be reproduced by our lattice mc results ( see fig . [ fig4 ] ) . as to be discussed elsewhere , a nonuniversal value of @xmath8 ( in the sense that we have specified in refs .
) could be obtained within the gtn model by allowing a finite dispersion in the distances of the second nearest neighbors . before concluding it is worth to point out some limitations of the theory presented above .
first , for simplicity , we have considered composites whose conducting fillers are given by monodispersed spheres . even maintaining that for some classes of composites the shape of the fillers
can be approximated by a sphere , the collection of such spheres in a real composite has however some degree of polydispersivity , which can be large enough to make the fractionally occupied periodic lattice model an inappropriate description of the possible configuration of the microstructure . however , as stressed above , the lattice model should be regarded as an extreme deviation from a continuum dispersion of particles ( monodispersed or not ) , so that the crossover studied in sec .
[ crossover ] represents an upper limit also for the case of polydispersed spherical particles .
the second limitation is associated with the fact the we have limited ourselves to the case of isotropic particle fillers , and we have not attempted to analyze how percolation and hopping behaviors arise for anisotropic filler shapes such as rodlike or platelike particles .
an analysis of these cases would be interesting also in relation to the appearance of nematic phases for large volume fractions of anisotropic fillers and to lattice - like arrangements driven by attractive forces in dispersions of rodlike particles .
v. ambegaokar , b. i. halperin , and j. s. langer , phys .
b * 4 * , 2612 ( 1971 ) ; m. pollak , j. non cryst .
solids * 11 * , 1 ( 1972 ) : b. i. shklovskii and a. l. efros , sov .
jetp * 33 * , 468 ( 1971 ) ; * 34 * , 1084 ( 1972 ) .
we remind that the cpa amounts to keep only the subset of tunnel conductances @xmath28 having @xmath149 , where @xmath150 is the largest among the inter - particle distances such that the so - defined subnetwork forms a conducting percolating cluster .
it then follows that @xmath150 is such that @xmath151 is satisfied , where @xmath152 is the critical coordination number .
finally the cpa is obtained by assigning the value @xmath153 $ ] to all the tunnel conductances of the entire network . in the present ema version of the cpa ,
@xmath154 is replaced by @xmath52 , eq . , and the criterion @xmath151
is replaced by eq . . | here , we show that the conductivity of conductor - insulator composites in which electrons can tunnel from each conducting particle to all others may display both percolation and tunneling ( _ i.e. _ , hopping ) regimes depending on few characteristics of the composite . specifically , we find that the relevant parameters that give rise to one regime or the other are @xmath0 ( where @xmath1 is the size of the conducting particles and @xmath2 is the tunneling length ) and the specific composite microstructure . for large values of @xmath0 , percolation arises when the composite microstructure can be modeled as a regular lattice that is fractionally occupied by conducting particle , while the tunneling regime is always obtained for equilibrium distributions of conducting particles in a continuum insulating matrix .
as @xmath0 decreases the percolating behavior of the conductivity of lattice - like composites gradually crosses over to the tunneling - like regime characterizing particle dispersions in the continuum . for @xmath0 values
lower than @xmath3 the conductivity has tunneling - like behavior independent of the specific microstructure of the composite . |
lasers are widely used in many fields of applied and fundamental research . however , for some applications , a serious drawback is their residual frequency instability .
free - running lasers without any measures of active or passive frequency stabilization exhibit frequency fluctuations ( laser linewidths ) much larger than the schawlow - townes limit @xcite and are subject to continuous frequency drifts .
for example , free- running diode lasers show a laser linewidth of about 100 mhz and frequency drifts of up to many 10 mhz / h , due to temperature and air pressure fluctuations in the laboratory , diode current instabilities etc . @xcite .
whereas the linewidth of a diode laser can be narrowed to typically less than 1 mhz using the passive optical feedback from a grating @xcite , the unwanted frequency drifts on longer time scales are more difficult to eliminate . a common technique is to actively stabilize the laser onto an external high - finesse fabry - perot interferometer ( fpi ) or , alternatively , onto suitable atomic or molecular transitions . for the latter ,
transition lines of various substances , e.g. , alkaline atoms or molecules like iodine or tellurium , have been widely in use .
the thorough investigation of the iodine spectrum by gerstenkorn et al .
@xcite together with the ease with which an iodine vapor cell can be handled make iodine a particularly attractive element for this purpose @xcite .
on the other hand there are also a number of drawbacks : the main spectrum of iodine ranges from 500 to 900 nm , and within this range the number of strong lines is limited so that for many frequencies of interest no excitation lines for the laser frequency stabilization are available . in this case
it is principally always possible to use an external fpi to stabilize the laser .
however , to set up an external fpi may entail considerable experimental efforts and , moreover , can not completely eliminate the unwanted residual frequency drifts .
therefore , if possible , it is usually advantageous to use atomic or molecular transitions as a reference for the active laser frequency stabilization since the atomic excitation frequencies do not change in time .
apart from iodine , there is thus a high demand to search for other atoms or molecules suitable for this purpose . in this paper , we investigate a hitherto unexplored element in use for active laser frequency stabilization , namely the atom praseodymium . because of its five outer valence electrons , the energy spectrum of praseodymium exhibits an extremely rich level structure starting already @xmath3 above the ground state . since the spectrum resembles more that of a solid than that of an atom , it is sometimes also called ` quasi - continuous ' .
the praseodymium spectrum has been well investigated by several groups over the last few decades @xcite .
so far , more than 1200 energy levels and transition lines are known in the spectral range from 320 nm to 3.5 @xmath4 m of which one third has been classified so far .
this multitude of excitation lines , including a rich hyperfine structure ( hfs ) for each resonance , in combination with the comprehensive knowledge of the spectrum make praseodymium an attractive element for the frequency stabilization of laser sources . in what follows ,
we report the first active frequency stabilization of a laser onto a praseodymium resonance .
several praseodymium transitions between 1105 nm and 1123 nm were systematically investigated for this purpose , corresponding to the tuning range of the diode laser used in the experiment .
the transitions were excited in a hollow cathode lamp ( hcl ) and recorded via laser - induced fluorescence ( lif ) .
active frequency stabilization of the diode laser was demonstrated using standard laser locking techniques . in this way ,
the frequency drift of the diode laser of more than 30 mhz / h was eliminated and its frequency continuously stabilized to within 1.4(1 ) mhz .
in addition , one new strong praseodymium transition close to the 4@xmath5 sub - harmonic of the d@xmath1 transition of mg@xmath2 at was found , making it possible to actively lock the diode laser at this wavelength .
after frequency quadruplication this will allow the continuous excitation of the mg@xmath2 d@xmath1 transition at .
the paper is organized as follows : in sects .
[ pr ] and [ setuplif ] we present the element praseodymium and introduce the experimental setup used to obtain the praseodymium lif spectra . in sects .
[ lif ] and [ lock ] we display several experimental lif spectra recorded in the range 1105 to 1123 nm and discuss the possibility to actively frequency stabilize a laser onto one of these lines . in sect .
[ mg ] we present a new praseodymium transition , close to the 4@xmath5 sub - harmonic of the d@xmath1 line of mg@xmath2 .
frequency quadruplication of an infrared laser , locked to this transition , would make it possible to continuously excite the d@xmath1 transition in mg@xmath2 . in sect .
[ conclusion ] we finally draw conclusions .
the chemical element praseodymium ( pr ) belongs to the lanthanide series ( rare earths group ) and is placed between the elements cerium ( ce ) and neodymium ( nd ) .
one of the prominent pr electron configurations with three valence electrons inside the @xmath6 shell is [ xe]@xmath7 .
the lowest term under a total of 17 terms of this configuration is a @xmath8 , where @xmath9 corresponts to the electronic ground state of pr .
hereby , the designation @xmath9 describes the coupling of the three equivalent electrons in the open @xmath6 shell in ls - notation ( ls - coupling nomenclature : @xmath10 denotes the parity of the level and @xmath11 in this case ) .
the naturally occurring pr is composed of only one stable isotope pr with nuclear spin @xmath12 .
the hyperfine interaction causes a hyperfine level splitting determined by the magnetic dipole coupling constant @xmath13 and the electric quadrupole coupling constant @xmath14 ( magnetic dipole moment @xmath15 = 4.2754(5 ) @xmath16 @xcite , electric quadrupole moment @xmath17 = -0.0024 b @xcite ) . for a hyperfine sublevel characterized by the quantum number @xmath18 the shift @xmath19 with respect to the corresponding fine structure level energy
can be written in first order perturbation theory in the form @xmath20 with @xmath21 where @xmath22 , @xmath23 , and @xmath18 are the nuclear spin , the total electron and total atom angular momentum quantum numbers , respectively .
eq.([hfs ] ) is usually called _
casimir s formula_. since the pr nucleus is almost spherical , the electric quadrupole term can be largely neglected in eq.([hfs ] ) .
the hyperfine splitting leads to a manifold of hfs sublevels .
the number of sublevels is given by : @xmath24 if at least one of the @xmath23-values of the combining fine structure levels is greater than 5/2 , then the hyperfine structure shows a flag pattern with six main ( diagonal ) hyperfine components . besides these main components two additional groups of weaker off - diagonal hyperfine components can be observed where the number of off - diagonal elements depends on @xmath25 . if @xmath26 one can find 10 off - diagonal components
whereas if @xmath27 there are only 9 off - diagonal components can be observed . since the frequency spacing between neighboring hyperfine components is typically on the order of 2 ghz and with doppler - widths of approximately 400 mhz at 1115 nm the full spectrum of the hfs can be usually resolved . to obtain the spectra the spectroscopic method of lif
is employed using a specially designed hcl .
the active laser frequency stabilization onto a particular pr transition can be roughly divided into two parts .
the first part deals with the generation of a strong lif signal in the hcl with high signal - to - noise ( s / n ) ratio .
the second part refers to the laser frequency stabilization itself , i.e. , the stabilization of the laser onto a particular lif resonance .
both parts will be discussed in more detail in the following three sections . increases the population of the upper level @xmath28 .
therefore the fluorescence intensity of the light stemming from the spontaneous decay of this level is modified .
this modification is recorded and used for the active laser frequency stabilization . the fluorescence light is usually composed of a set of different fluorescence channels with wavelengths @xmath29 , @xmath30 , @xmath31 , , which give information on the excited state .
each energy level is characterized by four parameters , namely @xmath32 , @xmath23 , @xmath13 , and @xmath14 .
this set of parameters represents an atomic fingerprint of an energy level and can be calculated from the observed lif spectrum .
this information in connection with the set of measured lif wavelengths allows to classify the transition.,scaledwidth=40.0% ]
lif spectroscopy in combination with an hcl is a sensitive and technique of recording the spectra of various gaseous substances .
the broad - band excitation in the hollow cathode discharge leads to a highly - excited plasma with a local thermodynamic equilibrium of the population distribution among the atomic energy levels .
therefore , lif in an hcl makes it also possible to investigate the level schemes of high - lying energy states , in the case of pr of up to above the ground state . the basic principle of lif spectroscopy is depicted in fig .
[ fig : prinzip_lif ] .
the laser beam with wavelength @xmath33 is led through the plasma column of the hollow cathode discharge and is slowly scanned over the spectral range of interest .
the selective absorption of the laser beam increases the population of the upper level @xmath28 when hitting a resonance .
this change in the population distribution can be detected by measuring the rise of fluorescence light intensity at the wavelengths @xmath34 , emitted by the spontaneous decay from the upper level @xmath28 to corresponding lower levels ( lif ) .
a lif signal will only be recorded if the laser frequency is in resonance with a dipole allowed transition between certain hyperfine sublevels @xmath18 of the lower level @xmath35 and @xmath36 of the upper level @xmath28 . a tunable monochromator ( spectral bandwidth @xmath37 = 0.2 nm ) which registers the fluorescence light makes it possible to separate all decay channels of the fluorescence spectrum . at the same time , the wavelength band - pass filter of the monochromator avoids an overload of the photomultiplier tube caused by the bright light emitted by the hollow cathode plasma over the whole spectral region .
there are two conceptionally different kinds of lif experiments , depending on the operation of the monochromator .
for the first type of experiments , the laser frequency is tuned to an optical resonance , so that the monochromator can be used to analyze the lif emission spectrum . since every energy level decays spontaneously into particular final states depending on the selection rules , the monitored fluorescence light can give information on the position , parity and quantum number @xmath23 of the upper state .
in the other type of experiment , the monochromator is tuned to a strong lif line and the laser frequency is then scanned over a frequency region of interest to obtain the hyperfine structure of the excitation line . in this way , the entire hfs of a spectral line can be studied .
both lif methods are displayed in fig .
[ fig : prinzip_lif_2arten ] .
the last mentioned method has been applied mostly in this work . the experimental setup for the generation of the lif spectra is presented in fig .
[ fig : setup_lif ] . a commercial diode laser from toptica photonics ( dl 100 ) at 1115
nm center wavelength is used for the excitation of pr atoms produced by sputtering inside the hcl .
the grating - stabilized extended cavity diode laser ( ecdl ) has a linewidth of less than 1 mhz which is sufficiently narrow to record the doppler - broadened pr resonances .
the ecdl has a maximum output power of 83 mw at 208 ma and is tunable over a range of almost 20 nm , from 1105 to 1123 nm ; without mode hopping the laser can be tuned in single mode over 20 ghz .
the internal collimator lens of the ecdl is used to focus the laser beam into the hcl to yield optimum illumination of the reaction volume . in order to prevent unwanted optical feedback from the hcl and other optical elements an optical isolator with isolation @xmath38 60 db
is installed in front of the ecdl .
a beam splitter placed behind the optical isolator feeds parts of the laser light into a wavemeter and an fpi .
both devices are used to determine the laser frequency .
the wavemeter ( uncertainty 50 mhz ) is used for the absolute frequency calibration of the lif spectra whereas the fpi ( free spectral range 2.109(12 ) ghz ) serves as an additional frequency marker in order to linearize the recorded lif spectra .
the non - commercial hcl especially designed for this experiment consists of several parts ( see fig . [
fig : hcl ] ) .
the central element of the symmetrical hcl is a copper tube into which an axially hollowed cathode of 20 mm length is screwed , filled on the inner walls with pr .
in addition , axially hollowed anodes made from aluminum are fixed at both ends of the copper tube .
the inner diameter of the hollow cathode , best for the investigation of neutral pr , has been found to be 3 mm .
the copper tube together with the cathode and the two anodes is placed inside a glass cylinder , sealed at both ends with brewster windows to allow maximum transmission of the laser beam through the hcl .
the hcl was operated with a buffer gas ( argon ) at a pressure of 0.5 mbar . in order to obtain a constant discharge current of about 60 ma , a voltage of 300
v was applied . under these conditions
one can assume that the fraction of atomic pr in the plasma amounts to more than 90 percent , the rest of the sputtered material being mainly singly ionized pr@xmath2 . in order to achieve a more stable discharge ,
the hcl can be cooled with liquid nitrogen . the nitrogen cooling results in an internal temperature of the hcl of about 100 to 200 @xmath39c ( boltzmann temperature ) which greatly reduces the doppler - widths of the spectra down to typically 350 to 400 mhz at 1115 nm .
passing laser radiation through the gas discharge yields a new population distribution among the two energy levels of pr participating in a resonance .
the fluorescence light that leaves the hcl in case of resonant laser excitation is collected by a pin - hole mirror ( placed on the side of the incoming laser beam ) and is directed to the entrance slit of the monochromator by a set of mirrors and lenses .
the monochromator with a tuning range from 300 to 900 nm and a resolution of @xmath40 can be used to filter single lif lines out of the total fluorescence spectrum .
a photomultiplier with high sensitivity for the visible spectrum is mounted at the exit slit of the monochromator allowing the detection of extremely low fluorescence signal changes .
in addition , a mechanical chopper ( 280 hz ) placed inside the laser beam enables a phase - sensitive detection and amplification of the lif signals by use of a lock - in amplifier ( lia ) .
even very low - intensity lif signals in an extremely noisy environment can thus be observed and recorded . the lif signal together with the marker signal from the fpi
are simultaneously recorded and stored in a computer .
the acquired lif spectra can then be processed by an especially designed evaluation program @xcite .
this program serves to determine the hfs parameters of the measured lines by fitting the experimentally recorded spectra by a theoretical curve composed of a specified number of voigt profiles using the method of least- squares .
fitting parameters of the theoretical curve are the hyperfine constants @xmath13 and @xmath14 of the involved lower and upper energy levels , the linewidths of the gaussian and lorentzian parts of the voigt profiles , and the relative intensities of all hyperfine components .
the total angular momentum quantum numbers @xmath23 and @xmath41 of the lower and upper levels can be chosen separately before the algorithm of the fit is carried out .
the choice of specific @xmath23 values defines the number of hyperfine components and is therefore a prerequisite for a good theoretical reconstruction of the recorded hfs patterns . as the positions of a maximum of 16 ( if @xmath42 ) or 15 ( if @xmath27 ) hfs components obey the casimir relations ( see eq.([hfs ] ) ) , only one additional parameter ( the center of gravity position of the line ) is needed . the final fit and the experimental spectrum are plotted together .
an additional curve called deviation curve is shown at the bottom of each spectrum .
this curve displays the error between the best fitted theoretical and the experimental curves and illustrates the quality of the fit .
[ fig : bestfit ] shows an example of a typical lif spectrum together with the adjusted theoretical curve in the best - fit situation .
= 1119.841 nm of neutral pr recorded in an hcl via lif .
@xmath33 is the vacuum wavelength of the laser excitation defining the center position of the hyperfine structure of the transition .
after laser excitation of a transition between the hyperfine structure sublevels of the lower level @xmath43 and the upper level @xmath44 _ _ 20643.071__@xmath45 , according to the selection rules of electric dipole radiation , the excited sublevels of the upper level @xmath28 decay into the energy level @xmath46 by emission of the fluorescence light at @xmath47 = 616.7 nm ( given in air ) .
the voigt - width of the hyperfine components is 405 mhz .
the figure shows the experimental excitation spectrum and the corresponding theoretical fit on a relative frequency scale .
the theoretical profile was calculated by a especially designed evaluation program @xcite based on a least - squares - fit method .
the plotted line below the spectrum is called the deviation curve and displays the difference between the theoretical fit and the experimental curve .
it illustrates the quality of the fit .
this pr spectral line was newly classified during this work.,scaledwidth=48.0% ] + +
in the first stage of our investigations of neutral pr we recorded a number of lines in the range between 1105 and 1123 nm , accessible with the tunable ecdl . since up to now all lif experiments with pr have been performed using ti : sapphire or single - mode dye lasers , the pr spectra recorded so far have been limited to laser excitation in the visible range of the spectrum ; apart from experiments employing fourier spectroscopy @xcite no direct electric dipole transitions in the infrared range of the spectrum have thus yet been observed . the spectral line density and the amount of known lines in the infrared region between 1105 and 1123 nm is almost ten times lower compared to the amount of lines classified in the visible . since lif spectroscopy is based on an optical coincidence measurement , where simultaneously the laser excitation and the detection window of the monochromator must coincide with a dipole allowed excitation line and a radiating emission line , respectively , it is a hopeless undertaking to find a resonance line in this wavelength region just by trial and error .
( the wavelength range between 1105 and 1123 nm corresponds to @xmath48 4 thz ; if the linewidth of a hfs component is @xmath48 400 mhz , the possibility of hitting the transition will be 1 : 10 000 . in combination with the monochromator tunable in the range 300 to 900 nm and a bandwidth of @xmath37 @xmath48 0.2
nm the overall possibility of recording a lif signal is in the best case 1 : 30 000 000 . ) to facilitate the task , we made use of a special database in which all fluorescence lines as well as all energy levels of pr recorded so far are collected @xcite .
the database contains not only the almost 600 levels of neutral pr given in ginibre s published and unpublished work @xcite but also about 600 further levels , most of them discovered in hamburg @xcite .
the constantly updated database gives information on the wavelengths of all recorded excitation lines and their combining energy levels , including @xmath23 quantum numbers , hyperfine coupling constants @xmath13 and @xmath14 , and parities @xmath32 .
these characteristic data are supplemented by a set of fluorescence wavelengths which were observed in earlier lif experiments @xcite .
apart from the identification of already measured lines and transitions the database allows in particular the prediction of new spectral lines and energy levels as well as the classification of hitherto unidentified hyperfine spectra . according to this database
, 66 lines were measured between 1105 and 1123 nm , which corresponds to the tuning range of the diode laser .
only a few of them had been classified so far .
for all other lines only the transition wavelengths and no further information about the combining energy levels were listed .
since it was not our intention to check all 66 measured lines , we picked out a few of them , in particular combining energy levels .
energy levels are usually much more populated than the higher energy levels , so that the corresponding lif resonances are expected to be much stronger .
this property becomes all the more important , as the laser power of the ecdl was restricted to 60 mw inside the hcl . in fig .
[ fig : spektrum_s12 ] three typical hfs spectra with different s / n ratios around 1119 nm are shown , together with the corresponding excitation and emission schemes .
the wavelengths of the fluorescence lines are given in air .
spectrum ( a ) is a confirmation of a classification given in @xcite .
spectra ( b ) and ( c ) were observed on the base of new levels taken from @xcite .
the particular hfs patterns originate from the different types of electronic transitions , defined by @xmath49 , as well as from the hyperfine splitting of the combining levels : well separated main hyperfine components with a characteristic flag pattern can be explained by clearly different hyperfine splitting of the lower and upper energy levels .
this is the case for example for the spectra shown in fig .
[ fig : spektrum_s12 ] ( a ) and ( b ) . on the contrary , the spectrum in fig .
[ fig : spektrum_s12 ] ( c ) , apart from five small off - diagonal components , displays only one broad asymmetric structure , resulting from the overlap of all six main components because of almost equal splitting of the lower and upper energy levels .
from all three spectra shown in fig .
[ fig : spektrum_s12 ] the spectrum in fig .
[ fig : spektrum_s12 ] ( b ) offers the best properties for an active frequency stabilization of a laser due to the good s / n ratio and the clearly separated hfs components .
+ in the case that no resonance lines are known in a particular wavelength region , the already mentioned database for the lines and energy levels of pr together with a software developed at the helmut - schmidt- universitt hamburg , germany , makes it possible to search for new transitions .
( note that l. windholz s group from the tu graz , austria , developed a similar program .
it is free to the public and can be used for any element . )
the list of theoretically proposed transitions contains hereby not only the information on the wavelength needed for the laser excitation and their combined energy levels but also the known fluorescence lines of the levels .
the latter can be used for the identification of the laser - induced transition .
the question remains , however , whether the theoretically predicted transitions can be actually confirmed in the experiment .
since the amount of known resonance lines in the infrared wavelength range is relatively low , the program turned out to be especially helpful .
for example , the program predicted ten transitions between @xmath50 = 1118.519 nm and @xmath51 = 1118.563 nm ( @xmath52 ) close to the 4@xmath5 sub - harmonic of the d@xmath1 transition of mg@xmath2 at which have never been excited or recorded before .
however , in most of these resonances the transitions start from relatively energy levels , @xmath53 to @xmath54 above the ground state , for which the chance of recording a strong lif signal is low . with the available laser power of 60 mw inside the hcl only two of the ten predicted pr lines
were confirmed in the experiment , namely those with the lowest initial states ( @xmath55 and @xmath56 ) .
they were detected by their strongest fluorescence decay channels at 567.75 and 573.58 nm , respectively .
the latter pr line is shown in fig .
[ fig : hfs_lock ] .
both lines were recorded and classified for the first time .
the two lines show a flag pattern with clearly separated hfs components and linewidths of about 375 mhz . with a s / n of about 20 and 70 , respectively , it is relatively simple to stabilize a laser onto any of the six main hfs components of each of the two lines . moreover , the presented s / n ratios could be improved by using a higher laser power .
however , for the diode laser at 1115 nm , higher intensities can only be achieved with special amplifier units , e.g. , tapered or fiber amplifiers .
a larger laser intensity would most probably allow the detection of the other eight theoretically predicted pr lines .
line of singly ionized magnesium ( mg@xmath2 ) with isotope shifts for the three naturally occurring isotopes where the level isotope shifts of the @xmath57 states are set to zero @xcite.,scaledwidth=47.0% ]
the active frequency stabilization of the infrared diode laser onto a particular pr hfs component was performed using a so - called lock - in - regulator ( lir ) from toptica photonics ( see fig . [
fig : setup_lif_lock ] ) .
the lir module is a phase- sensitive lock - in detection unit with an integrated proportional , integral , and differential ( pid ) regulator .
based on a modulation of the laser frequency by dithering the piezo of the ecdl grating the first derivative of the lif resonance is generated .
this error signal is used to accomplish via the pid regulator the active frequency stabilization of the laser or , outside the lock , to investigate the frequency fluctuations and drifts of the laser . for the latter ,
the linear part of the error signal around the zero crossing is used as a frequency discriminator ( see fig .
[ fig : drift ] ( a ) and ( b ) ) .
[ fig : drift ] ( c ) shows the drift behavior of the unlocked laser . for the measurements
the strongest hfs component of the pr transition + @xmath58 at was used , having a linewidth of 375 mhz ( see fig .
[ fig : hfs_lock ] ) . within five hours of measurement a drift of more than 90 mhz was observed .
after active frequency stabilization of the laser this frequency drift was strongly reduced .
[ fig : lock_pr_lock ] shows the frequency stability of the laser ( averaging time @xmath59 s ) after actively locking the laser onto the mentioned hfs component . as can be seen in the figure the laser frequency
was stabilized to within 1.4(1 ) mhz for an unlimited amount of time .
the ultimate frequency stability of the laser is determined by the s / n ratio of the lif signal .
since the lif signal is relatively low and furthermore obscured by the bright fluorescence background of the hcl , it is necessary to work with a lock - in amplifier ( lia ) ( see sect . [ setuplif ] ) .
the lif signal , after amplification by the lia , is fed into the lir to produce the error signal for the active frequency lock ( see fig .
[ fig : setup_lif_lock ] ) . since the laser frequency is also modulated for the lock - in detection unit of the lir , a too large time constant of the lia actually improves the s / n of the lif signal but at the same time also averages over the lir modulation and thus suppresses the lir error signal . a good compromise between a reasonable s / n of the lif signal and an acceptable error signal of the lir is therefore mandatory . for this purpose the modulation frequency of the lir was reduced to the smallest selectable value ( 8.2 hz ) .
the bandwidth of the locking loop , given approximately by half of the modulation frequency of the lir , therefore corresponded to about 4 hz .
this allows mainly the compensation of the low - frequency instabilities of the laser , i.e. , the laser drift .
in the previous section we presented the active frequency stabilization of a diode laser onto a particular hyperfine transition of pr at . the selected pr line served however merely as an example to demonstrate the feasibility of locking a laser onto a pr resonance using lif spectroscopy . in principle , the diode laser could have been stabilized onto any other suitable pr line in the range from 1105 to 1123 nm . by frequency quadrupling the stabilized infrared diode laser , in the range 276 to 281 nm can be produced .
this technique has already been employed to convert laser radiation at 922 nm to narrow - band uv - light at 231 nm @xcite .
since diode lasers are not yet available at 280 or 560 nm , this approach makes it possible to transfer all well - known advantages of grating - stabilized ecdls to these wavelength regions .
this is an attractive alternative to ( frequency - doubled ) dye lasers or tunable fiber lasers @xcite otherwise employed in these wavelength regions .
[ tab : mg_lines ] cccc isotop & @xmath60 ( nm ) & @xmath61 ( nm ) & @xmath62 ( nm ) + @xmath63 & 279.6355 & 559.2710 & 1118.5420 + @xmath64 & 279.6349 & 559.2698 & 1118.5396 + @xmath65 & 279.6347 & 559.2694 & 1118.5388 + as it turns out , this technique can also be used to produce laser light at 279.635 nm , making it possible to excite the d@xmath1 transition of singly ionized magnesium mg@xmath2 . in recent years
there has been a growing interest in mg@xmath2 since due to its internal level structure this ion is a favorite candidate for experiments in spectroscopy , quantum optics or in quantum information science @xcite .
the precise wavelengths of the d@xmath1 lines of the three stable isotopes together with the corre- sponding line isotope shifts are shown in fig .
[ fig : mg_level_scheme ] .
the corresponding fundamental , second , and fourth sub - harmonics for the three isotopes are listed in table [ tab : mg_lines ] .
= 1118.5548 nm together with the 4@xmath5 sub - harmonics of the d@xmath1 transitions @xmath66 of the three isotopes @xmath67mg@xmath2 , @xmath68mg@xmath2 , and @xmath69mg@xmath2 .
the relative positions and the adapted linewidths are drawn to scale .
of the six components of the hyperfine structure , the strongest hyperfine component at has a deviation of less than 600(100 ) mhz with respect to @xmath66 of the three isotopes and is therefore the closest and most suitable line for locking the frequency of the infrared diode laser in order to perform experiments with trapped mg@xmath2
ions.,scaledwidth=48.0% ] the previously discussed strongest hfs com- ponent of the pr transition at is less than 600(100 ) mhz away from the 4@xmath5 sub - harmonics of the mg@xmath2 d@xmath1 line @xmath66 at 1118.540 nm = 4 @xmath70 @xmath71 ( see fig . [
fig : hfs_lock ] ) .
this resonance can therefore be conveniently used to lock the laser in the infrared and - after frequency quadruplication - to obtain light in the uv in order to excite the d@xmath1 transition .
the remaining frequency gap of 600(100 ) mhz in the infrared can be easily bridged by an acousto - optical modulator ( aom ) .
the second nearest hyperfine component @xmath72 of the transition is already almost 2 ghz away from @xmath66 , usually too large to be bridged by an aom .
[ fig : hfs_lock ] shows the complete hyperfine structure of the transition together with @xmath66 of the three stable isotopes @xmath67mg@xmath2 , @xmath68mg@xmath2 , and @xmath69mg@xmath2 . the relative positions and the linewidths are drawn to scale .
other pr hfs resonances investigated in this wavelength region exhibited either worse s / n ratios or were at least 5 ghz away from the desired @xmath66 .
the obtained frequency stability of the laser of 1.4(1 ) mhz at - after frequency quadruplication - corresponds to a frequency stability of @xmath73 6 mhz in the uv .
this is by far sufficient for the resonant excitation of the d@xmath1 line of mg@xmath2 which has a natural linewidth of mhz . by appropriately detuning the frequency - stabilized diode laser it thus becomes possible to excite or laser cool individual mg@xmath2 isotopes for an unlimited amount of time .
in conclusion , we discussed the atom praseodymium ( pr ) as a versatile and hitherto unexplored reference for the active frequency stabilization of a laser . with five
valence electrons pr displays an extremely rich level structure , starting already @xmath3 above the ground state .
so far , more than 1200 energy levels and transition lines between 320 nm and 3.5 @xmath4 m are known , including a rich hyperfine structure ( hfs ) for each resonance .
this multitude of excitation lines together with the detailed knowledge of the spectrum makes pr a very attractive reference for active laser frequency stabilization . to demonstrate the feasibility to use this element for the active frequency stabilization of lasers we actively locked a diode laser onto a particular pr line .
the corresponding excitation signals were recorded in a hollow cathode lamp ( hcl ) via laser - induced fluorescence ( lif ) spectroscopy .
using standard locking techniques we were able to eliminate the frequency drifts of the unlocked diode laser of more than 30 mhz / h and stabilize its frequency permanently to within 1.4(1 ) mhz for averaging times @xmath0 s. the strongest hfs component of a newly found pr line at allows - after frequency quadruplication - the production of frequency stable uv light at which can be used to excite the d@xmath1 transition in mg@xmath2 . in future
it might be possible to simplify the experimental design by replacing the hcl by a common rf - lamp filled with pr .
the authors would like to especially thank prof .
hermann harde from the helmut - schmidt - universitt der bundeswehr hamburg for making measurements in his laboratory possible .
the friendly assistance from dr .
hans - otto behrens , gnther helmrich , and marion kollmeier during the measurements in hamburg was also appreciated .
thanks also go to thomas heine , thorsten schmitt , and dr .
harald ellmann from toptica photonics for their valuable technical support for the diode laser system .
furthermore , s.o .
thanks the universitt bayern e.v . for financial support .
a. l. schawlow , c. h. townes , phys . rev . * 112 * , 1940 ( 1958 ) . c. e. wieman , l. hollberg , rev .
62 * , 1 ( 1991 ) .
et al . _ , opt . commun . * 117 * , 541 ( 1995 ) .
s. gerstenkorn , p. luc , _ atlas du spectre dabsorption de la molcule diode _ , laboratoire aim - cotton , cnrs ii , orsay ( france ) , 14000 - 15600 @xmath74 ( 1978 ) , 15600 - 17600 @xmath74 ( 1977 ) , 17500 - 20000 @xmath74 ( 1977 ) .
s. gerstenkorn , j. verges , j. chevillard , _ atlas du spectre dabsorption de la molcule diode _ , laboratoire aim - cotton , cnrs ii , orsay ( france ) , 11000 - 14000 @xmath74 ( 1982 ) .
et al . _ , opt . commun . * 192 * , 263 ( 2001 ) .
s. picard _
_ , ieee trans .
52 * , 236 ( 2003 ) .
. commun . * 235 * , 377 ( 2004 ) .
m. n. skvortsov _
et al . _ , quantum electron . *
34 * , 1101 ( 2004 ) .
r. m. macfarlane , d. p. burum , r. m. shelby , phys .
* 49 * , 636 ( 1982 )
. k. d. bklen _ et al .
_ , z. phys .
a * 274 * , 195 ( 1975 ) .
a. ginibre , phys .
* 23 * , 260 ( 1981 ) .
a. ginibre , thse universit de paris - sud , centre dorsay ( 1988 ) .
g. h. guthhrlein , helmut - schmidt - universitt der bundeswehr hamburg ( germany ) , unpublished energy levels and transition lines of praseodymium .
l. windholz , technische universitt graz ( austria ) , unpublished energy levels and transition lines of praseodymium .
r. zalubas , b. r. borchardt , j. opt .
63 * , 102 ( 1973 ) .
w. c. martin , r. zalubas , l. hagan , `` atomic energy levels - the rare earth elements '' , national bureau of standards , u.s.g.p.o .
( washington dc 1978 ) . w. j. childs , l. s. goodman , phys .
a * 24 * , 1342 ( 1981 ) .
t. kuwamoto _
et al . _ , j. phys .
japan * 65 * , 3180 ( 1996 ) .
a. krzykowski _
. commun . * 140 * , 216 ( 1997 ) .
b. furmann , thesis , tu pozna ( 1998 ) .
a. krzykowski , thesis , tu pozna ( 1998 ) .
d. el bakkali , thesis , helmut - schmidt - universitt der bundeswehr hamburg ( germany ) , fakultt fr elektrotechnik ( 2006 ) .
b. furmann _
et al . _ ,
. scr . * 74 * , 658 ( 2006 ) .
evaluation program _
fitter _ for hyperfine structure measurements , helmut - schmidt - universitt der bundeswehr hamburg ( germany ) , laboratory for experimental physics , unpublished .
a. ginibre - private communication to g. h. guthhrlein .
c. schwedes _ et al .
b * 76 * , 143 ( 2003 ) . a. friedenauer _
et al . _ ,
b * 84 * , 371 ( 2006 ) . j. t. hffges _ et al . _ , j. mod . opt . * 44 * , 1999 ( 1997 ) .
v. batteiger _
et al . _ , phys .
a * 80 * , 022503 ( 2009 ) .
t. schtz _
et al . _ , j. mod .
opt . * 54 * , 2317 ( 2007 ) .
a. friedenauer _
et al . _ , nat
. phys . * 4 * , 757 ( 2008 ) .
j. p. home _
et al . _ ,
science * 325 * , 1227 ( 2009 ) .
h. schmitz _ et al .
lett . * 103 * , 090504 ( 2009 ) . | we present a new possibility for the active frequency stabilization of a laser using transitions in neutral praseodymium . because of its five outer electrons , this element shows a high density of energy levels leading to an extremely line - rich excitation spectrum with more than known spectral lines ranging from the uv to the infrared .
we demonstrate the active frequency stabilization of a diode laser on several praseodymium lines between 1105 and 1123 nm .
the excitation signals were recorded in a hollow cathode lamp and observed via laser - induced fluorescence .
these signals are strong enough to lock the diode laser onto most of the lines by using standard laser locking techniques . in this way , the frequency drifts of the unlocked laser of more than 30 mhz / h were eliminated and the laser frequency stabilized to within 1.4(1 ) mhz for averaging times @xmath0 s. frequency quadrupling the stabilized diode laser can produce in the range from 276 to 281 nm . in particular , using a strong hyperfine component of the praseodymium excitation line at makes it possible - after frequency quadruplication - to produce laser radiation at , which can be used to excite the d@xmath1 line in mg@xmath2 . |
entanglement @xcite and nonlocality @xcite are considered to be significant resources that quantum mechanics offers.they have a ubiquitous role in information processing tasks and foundational principles , whenever one has to certify quantum advantage over conventional classical procedures .
entanglement is a physical phenomenon in which many particles interact in a manner that the description of each particle separately does not suffice to describe the composite system .
a resource of this kind can be used to demonstrate one of the strongest form of non classical feature i.e non - locality where the statistics generated from each subsystem can not be reproduced by any local realistic theory analogous to classical physics @xcite .
bell nonlocal correlation along with entanglement are found to be key resources for many information processing tasks such as teleportation @xcite , dense coding @xcite , randomness certification @xcite , key distribution @xcite , dimension witness@xcite , bayesian game theoretic applications @xcite .
+ however , the question of identifying an entangled state remains one of the most involved problems in quantum information . commonly phrased as the , it has been shown to be np hard@xcite . in lower dimensions viz .
( @xmath1 and @xmath2 ) there is an elegant necessary and sufficient criterion criterion to identify entangled states .
negative partial transpose of a quantum state is considered to be a signature of entanglement whereas states having positive partial transpose(ppt ) are separable @xcite .
the solution in higher dimensions lacks a bi - directional logic to certify a state to be entangled , more so with the presence of ppt entangled states @xcite .
nevertheless , an extremely useful operational criteria to detect entanglement is provided through entanglement witnesses(ew)@xcite.an outcome of the well - known hahn - banach theorem in functional analysis , entanglement witnesses @xmath3 are hermitian operators having at least one negative eigenvalue which satisfy the inequalities ( i ) @xmath4 separable states @xmath5 and ( ii ) @xmath6 for at at least one entangled state @xmath7 .
the geometric form of the theorem states that points lying outside a convex and closed set can be efficiently separated from the set by a hyperplane @xcite .
the completeness of this existence guarantees that whenever a state is entangled there is a ew to detect it @xcite . on the virtue of being hermitian
, entanglement witnesses have proved their efficacy in experimental detection of entanglement @xcite .
the notion of this separability axiom has been extended to identify useful resources for teleportation using teleportation witnesses @xcite .
an elegant procedure to capture non - locality is through a bell - chsh witness @xcite .
this bell - chsh witness is a translation of an ew to detect states which violate the bell - chsh inequality . on a different note , for two qubits , all quantum states @xmath8 do not violate the bell - chsh inequality iff @xmath9 , where @xmath10 is defined as the sum of the two largest eigenvalues of the matrix @xmath11 , @xmath12 being the correlation matrix in the hilbert - schmidt representation of @xmath8 .
thus , @xmath13 is a signature of the non - locality of the state@xcite .
+ the prominent role of non - local states are highlighted through non - local games and randomness certification .
randomness @xcite plays a key role in many information theoretic tasks .
it has already been shown to be an important resource for quantum key distribution and cryptography .
so an interesting question could be whether one can classify the states which are helpful to certify randomness , which has been answered affirmatively in recent times .
now the question arises whether this class of resources can be expanded in a scenario where prior to the task all the subsystems are subjected to a global unitary operation . in this paper
we deal with this question by characterizing the class of local states which can never be _ bell chsh - nonlocal _ under any possible global unitary operation termed as absolutely _ bell chsh - local _ states .
as one can see that this class of states can never be useful for randomness certification task .
we have further shown that for systems with hilbert space @xmath14 these states form a convex and compact set .
this implies the existence of a hermitian operator which can detect such non - absolutely _ bell chsh - local _ states , a potential resource in the modified scenario .
+ in the following section ( sec.[motivate ] ) we first outline the need for an operator to identify non - absolutely _ bell chsh - local _ states and its importance for a number of information theoretic tasks .
parallely , the question of existence of a set containing absolutely _ bell chsh - local _ states has been addressed in relation to problems of an analogous type . in sec.[def ] we introduce the relevant notations and definitions to prepare a mathematical formulation . in sec.[proof ] we present the proof of the existence and a definite scheme of constructing such operators and illustrate with examples of absolutely _ bell chsh - local _ states in sec.[examples ] .
finally we conclude in sec.[discuss ] .
( b ) . now the unitary - evolved system is used to play a non - local game ( c ) . ] pertaining to separability of quantum states , questions have been raised on the characterization of absolutely separable@xcite and absolutely ppt states@xcite .
precisely , a quantum state which is entangled(respectively ppt ) in some basis might not be entangled(resp .
ppt ) in some other basis .
this depends on the factorizability of the underlying hilbert space .
thus , the characterization of states which remain separable(resp .
ppt ) under any factorization of the basis is pertinent .
literature already contains results in this direction @xcite . precisely for two qubits ,
a state is absolutely separable iff its eigenvalues(arranged in descending order ) satisfy @xmath15 @xcite . in a different perspective ,
violation of bell - chsh inequalities exhibits non - local manifestations of a quantum state . a state which violates the bell - chsh inequality
is considered non - local . however , the violation of such an inequality invariably depends on the factorization of the underlying hilbert space and in this regard the study on states which do not violate the bell - chsh inequality under any factorization assumes significance .
this study forms the main context of our present work.while intuition permits one to state that states which are absolutely separable will be eligible candidates under this classification , it is interesting to probe the existence of entangled states which come under this category .
our results underscore the existence of states outside the absolute separable class , which do not violate bell - chsh inequality under any global unitary operation .
global unitary operations play the anchor role here as they transform states to different basis .
+ in standard bell - scenario the subsystems are allowed to share classical variables prior to the game and perform local operations only . for our purpose
we consider a modified bell - scenario ( depicted in fig.[belldiag ] ) in which two parties are allowed to perform a global unitary @xmath16 prior to the collection of statistics from the joint - system .
one can easily note that by performing a cnot operation on the initial state @xmath17 ( which is bell - chsh local ) , it can be transformed to @xmath18 which has a maximum bell - chsh violation of @xmath19 .
this clearly shows that the set of resources for this modified bell - chsh scenario can be expanded to a plausible extent .
+ now the question is how far can this extension is possible .
contrary to the common intuition there exists states which do not violate bell - chsh inequality under any global unitary operation
. then one can depict the set of resources in this modified bell - scenario as the green and yellow regions in a schematic diagram as shown in fig.[sets ] . to deal with this not - so - obvious scenario
we prepare the mathematical framework hereafter beginning with a description relevant notations and definitions in the following section .
we begin with some notations and definitions needed for our analysis.@xmath20 denotes the set of bounded linear operators acting on x. the density matrices that we consider here , are operators acting on two qubits , i.e. , @xmath21 .
@xmath22 denotes the set of all density matrices .
we denote by @xmath23 , the set of all states which do not violate the bell - chsh inequality @xcite , i.e. , @xmath24 .
we denote by @xmath25 as the set containing states which do not violate the bell - chsh inequality under any global unitary operation ( @xmath26 ) i.e. @xmath27 .
one can easily see that @xmath28 forms a non - empty subset of @xmath23 , as @xmath29 .
a schematic diagram of the sets has been shown in fig.[sets ] . .
the set @xmath30 is characterized by the existence of the bell - chsh witness @xcite .however , we give a formal characterization below : + @xmath23 is a convex and compact subset of @xmath22 .
first note that the statements below are equivalent : + ( i ) @xmath31 + ( ii ) @xmath32 bell - chsh operator @xmath33 + ( iii ) @xmath32 bell - chsh witness @xmath34 + in view of the above , we can rewrite @xmath30 as , @xmath35 .
now consider a function @xmath36 , defined as @xmath37 where , @xmath38 is a fixed bell - chsh witness .
let @xmath39 .
@xmath40 will have a maximum value @xmath41 ( say ) .
therefore , one may write @xmath42 $ ] .
@xmath43 is a continuous function as @xmath44 is a continuous function .
this in turn implies @xmath45 is a closed set .
continuing as above , one may define @xmath46 for a fixed @xmath47 .
@xmath46 will be closed @xmath48 . since , arbitrary intersection of closed sets is closed , @xmath49 is closed .
it is easy to see that @xmath50 .
hence , @xmath30 is closed .
if we now take two arbitrary @xmath51 , then @xmath52 \geq 0 $ ] for any @xmath53 , @xmath54 $ ] .
this follows from the fact that @xmath55 \geq 0 , i=\lbrace 1,2 \rbrace$ ] for any @xmath56.thus @xmath30 is convex . since @xmath57 is compact , @xmath30 being a closed subset of @xmath57 , is thus compact .
hence the lemma .
this theorem facilitates the characterization of the set @xmath25 as stated in the theorem below : + @xmath28 is a convex and compact subset of @xmath23 .
we only show that @xmath25 is convex as the compactness follows from a retrace of the steps presented in @xcite .
+ take two arbitrary @xmath58 .
one may rewrite @xmath59 \geq 0 , \forall b_{chsh}^{w } , \forall u \rbrace $ ] .
therefore , for any @xmath26 , @xmath60u^{\dagger } = \lambda \sigma_{1}^{\prime } + ( 1-\lambda)\sigma_{2}^{\prime } \in \mathbf{al } $ ] .
this follows , since @xmath30 is convex .
[ @xmath61 . +
as noted earlier , one may see the compactness with a re - run of the steps in @xcite .
hence , the theorem .
the above characterization enables to formally define an operator(@xmath62 ) which detects states that violate bell - chsh inequality under global unitary .
@xmath63 consider @xmath64 .
there exists a unitary operator @xmath65 such that @xmath66 violates bell - chsh inequality .
consider a bell - chsh witness @xmath3@xcite that detects @xmath67 , i.e. , @xmath68 .
using the cyclic property of the trace , one obtains @xmath69 .
we thus claim that @xmath70 is our desired operator .
to see that it satisfies inequality ( [ ineq1 ] ) , we consider its action on a state @xmath71 from @xmath28 .
we have @xmath72 . as
@xmath73 @xmath74,and @xmath75 is a bell - chsh witness @xmath76 .
this implies that @xmath77 has a non - negative expectation value on all states @xmath78 .
while the above theorem has provided a tool to identify states which can augment non - local resources under global unitary , it has also highlighted the existence of a set which contains states from which no non - local resource ( in terms of the bell - chsh inequality ) can be generated .
it is therefore important to look for certain states which can belong to the absolutely bell - chsh local set .
it is evident that , any separable state obviously belongs to @xmath23 . from the definition of absolutely separable states ,
i.e. @xmath79 , @xmath80 being any global unitary operation , it is clear that after the operation of the global unitary the state remains in @xmath23 , i.e. all the absolutely separable states are absolutely bell - chsh local states .
consider the bell - diagonal states , i.e. @xmath81 , where @xmath82 are the usual bell states . if now one imposes the dual conditions @xmath83 , it is easy to see that this state is not only separable but also belongs to @xmath84 and hence @xmath85 .
werner states @xmath86 , where @xmath87 being singlet state , in @xmath88 are absolutely separable for @xmath89 , as a result it is also absolutely local here .
it can now be asked whether there exist states which are not separable but belong to @xmath28 .
it is well known that @xmath90 is entangled but does not violate bell - chsh for @xmath91 $ ] . in case of werner states there
exist no global unitary operations @xmath80 such that @xmath92 will violate bell - chsh inequality for @xmath93 , as the first part of the modified state @xmath94 will always have bell - chsh value less than or equal to @xmath95 and the second part will have zero for any global unitary .
it is evident that the werner states with visibility factor @xmath96 belong to @xmath28 but are entangled .
consider another state ; @xmath97 , where @xmath98 , @xmath99 , which is also @xmath28 for @xmath100 but entangled for @xmath101 .
+ one may note here that , since unitary similarity is an equivalence relation , it partitions the absolutely bell - chsh local set into distinct equivalence classes . for e.g if we define an equivalence class for a definite @xmath102 as , @xmath103= \lbrace \tau : \tau = u \sigma_{al } u^{\dagger } \rbrace\ ] ] then , all states @xmath104 $ ] are absolutely bell - chsh local . + hence if one considers the state having weights @xmath105 diagonal in the computational basis @xmath106 is a separable state but not absolutely separable for @xmath107 and it belongs to @xmath108 $ ] .
it is evident from above that the state is in @xmath28 for @xmath109 .
in standard bell - scenario the free resources are local operation and shared randomness . here
we have considered an modified scenario where prior to the non - local game the subsystems are allowed to undergo a global unitary evolution .
contrary to common intuition all quantum states can not be made to violate bell - chsh inequality even in this modified scenario . in this work we have shown that for two qubit systems these useless states which we call absolutely _ bell chsh - local _ , form a convex and compact set implying the existence of a hermitian operator which can detect non - absolutely _ bell chsh - local _ states , a potential resource in the modified bell - scenario .
we also present a characterization of absolutely _ bell chsh - local
_ states for a number of generic class of states .
this analysis of bell non - locality presents a new paradigm for asking a number of important questions .
firstly , one could seek for a generic characterization of absolutely _ bell chsh - local _ states even for two qubits .
secondly , the question remains whether one could demonstrate the existence of non - absolutely _ bell - local _ witness operators for higher dimensional systems in different bell - scenarios@xcite and subsequently characterizing the set of absolutely _ bell - local _
states for such systems .
* acknowledgment : * we would like to gratefully acknowledge fruitful discussions with prof .
guruprasad kar .
we also thank tamal guha and mir alimuddin for useful discussions .
am acknowledges support from the csir project 09/093(0148)/2012-emr - i . c. h. bennett , g. brassard , c. crpeau , r. jozsa , a. peres , and w. k. wootters , `` teleporting an unknown quantum state via dual classical and einstein - podolsky - rosen channels '' , http://dx.doi.org/10.1103/physrevlett.70.1895[phys . rev
. lett . * 70 * , 1895 ( 1993 ) ] .
s. pironio , a. acn , s. massar , a. boyer de la giroday , d. n. matsukevich , p. maunz , s. olmschenk , d. hayes , l. luo , t. a. manning and c. monroe , `` random numbers certified by bell s theorem '' , http://www.nature.com/nature/journal/v464/n7291/full/nature09008.html[nature * 464 * , 1021 - 1024 ] .
r. colbeck and r. renner , free randomness can be amplified " , http://www.nature.com/nphys/journal/v8/n6/full/nphys2300.html[nat .
phys.*8 * , 450 ( 2012 ) ] ; a. chaturvedi and m. banik , measurement - device
independent randomness from local entangled states " , http://iopscience.iop.org/article/10.1209/ 0295 - 5075/112/30003/meta;jsessionid = d6c96abb3e61 c42c542a9553e8a4f4dc.c3.iopscience.cld.iop.org[epl * 112 * , 30003 ( 2015 ) ] .
j. barrett , l. hardy , and a. kent , no signaling and quantum key distribution " , http://dx.doi.org/10.1103/physrevlett.95.010503[phys .
lett . * 95 * , 010503 ( 2005 ) ] ; a. acn , n. gisin , and l. masanes , from bells theorem to secure quantum key distribution " , http://dx.doi.org/10.1103/physrevlett.97.120405[phys .
lett . * 97 * , 120405 ( 2006 ) ] ; n. brunner , s. pironio , a. acn , n. gisin , a. a. methot , and v. scarani , testing the dimension of hilbert spaces " , http://dx.doi.org/10.1103/physrevlett.100.210503[phys . rev
. lett . * 100 * , 210503 ( 2008 ) ] ; r. gallego , n. brunner , c. hadley , and a. acn , device independent tests of classical and quantum dimensions " , http://dx.doi.org/10.1103/physrevlett.105.230501[phys .
lett . * 105 * , 230501 ( 2010 ) ] ; s. das , m. banik , a. rai , md r. gazi , and s.kunkri ,
hardy s nonlocality argument as a witness for postquantum correlations " , https://journals.aps.org/pra/abstract/10.1103/physreva.87.012112[phys .
a * 87 * , 012112 ( 2013 ) ] ; a. mukherjee , a. roy , s. s. bhattacharya , s. das , md .
r. gazi , and m. banik ,
hardy s test as a device - independent dimension witness " , https://journals.aps.org/pra/abstract/10.1103/physreva.92.022302[phys .
a * 92 * , 022302 ( 2015 ) ] ; n. brunner and n. linden , connection between bell nonlocality and bayesian game theory `` , http://www.nature.com/ncomms/2013/130703/ncomms3057/full/ncomms3057.html#references[nature communications * 4 * , 2057 ( 2013 ) ] .
et al . _ ' ' nonlocality and conflicting interest games `` , http://journals.aps.org/prl/abstract/10.1103/physrevlett.114.020401[phys . rev .
lett . * 114 * , 020401 ( 2015 ) ] .
a. roy , a. mukherjee , t. guha , s. ghosh , s. s. bhattacharya , m. banik , `` nonlocal correlations : fair and unfair strategies in bayesian game '' , http://arxiv.org/abs/1601.02349[arxiv : 1601.02349 ] .
l.gurvits , _ proceedings of the thirty - fifth annual acm symposium on theory of computing _ , eds . l. l. larmore and m. x. goemans , 10 ( 2003 ) .
a. peres,''separability criterion for density matrices``,http://dx.doi.org/10.1103/physrevlett.77.1413[phys .
77 , 1413 ( 1996 ) ] .
m. horodecki , p. horodecki , r. horodecki,''separability of mixed states : necessary and sufficient conditions `` , doi:10.1016/s0375 - 9601(96)00706 - 2[phys .
lett . a 223 , 1 ( 1996 ) ] .
p. horodecki,''separability criterion and inseparable mixed states with positive partial transposition `` , doi:10.1016/s0375 - 9601(97)00416 - 7[phys .
a 232 , 333 ( 1997 ) ] ; m. horodecki , p. horodecki , r .
horodecki,''mixed - state entanglement and distillation : is there a `` bound '' entanglement in nature?``,http://dx.doi.org/10.1103/physrevlett.80.5239[phys .
80 , 5239 ( 1998 ) ] . b. m. terhal,''bell inequalities and the separability criterion " , doi:10.1016/s0375 - 9601(00)00401 - 1[phys .
lett . a 271 , 319 ( 2000 ) ] .
m. barbieri , f. de martini , g. di nepi , p. mataloni , g. m. dariano , c. macchiavello,``detection of entanglement with polarized photons : experimental realization of an entanglement witness'',http://dx.doi.org/10.1103/physrevlett.91.227901 [ phys .
91 , 227901 ( 2003 ) ] .
w. wieczorek , c. schmid , n. kiesel , r. pohlner , o. guhne , h. weinfurter,``experimental observation of an entire family of four - photon entangled states'',http://dx.doi.org/10.1103/physrevlett.101.010503[phys .
lett . 101 , 010503 ( 2008 ) ] .
n. ganguly , s. adhikari , a. s. majumdar , j. chatterjee,``entanglement witness operator for quantum teleportation'',http://dx.doi.org/10.1103/physrevlett.107.270501[phys .
107 , 270501 ( 2011 ) ] .
s. adhikari , n. ganguly , a. s. majumdar,``construction of optimal teleportation witness operators from entanglement witnesses'',http://dx.doi.org/10.1103/physreva.86.032315[phys .
a 86 , 032315 ( 2012 ) ] .
d.collins , n. gisin , n. linden , s. massar , s. popescu,``bell inequalities for arbitrarily high - dimensional systems'',http://dx.doi.org/10.1103/physrevlett.88.040404[phys .
88 , 040404 ( 2002 ) ] . | the action of global unitary operations can have intriguing effects on the non - local manifestations of a quantum state .
states which admit a local hidden variable model can violate bell s inequality when transformed by a global unitary .
this phenomena assumes significance when one exercises to augment non - local resources from these seemingly useless states(i.e .
, in terms of non - local tasks )
. however , equally intriguing is the existence of states from which no non - local resource can be generated with a global unitary.the present work confirms the existence of such a set , pertaining to states living in @xmath0 .
the set exhibits counter intuitive features by containing within it some entangled states which remain local on the action of any unitary . furthermore through an analytic characterization of the set we lay down a generic prescription through which one can operationally identify states which can generate non - local resources under the action of a unitary on the composite system . |
one of the fundamental questions in quantum statistical inference problem is to establish the ultimate precision bound for a given quantum statistical model allowed by the laws of statistics and quantum theory . mainly due to the non - commutativity of operators and nontrivial optimization over all possible measurements ,
this question still remains open in full generality .
this is very much in contrast to the classical case where the precision bounds are obtained in terms of information quantities for various statistical inference problems .
the problem of point estimation of quantum parametric models is of fundamental importance among various quantum statistical inference problems .
this problem was initiated by helstrom in the 1960s and he devised a method to translate the well - known strategies developed in classical statistics into the quantum case @xcite . a quantum version of fisher information was successfully introduced and the corresponding precision bound , a quantum version of cramr - rao ( cr ) bound , was derived .
it turned out , however , that the obtained bound is not generally achievable except for trivial cases .
a clear distinction regarding the quantum parameter estimation problem arises when exploring possible estimation strategies since there is no measurement degrees of freedom in the classical estimation problem .
consider @xmath0 identical copies of a given quantum state and we are allowed to perform any kinds of quantum measurements according to quantum theory .
a natural question is then to ask how much can one improve estimation errors by measurements jointly performed on the @xmath0 copies when compared to the case by those individually performed on each quantum state . the former class of measurements is called _ collective _ or _ joint _ and the latter is referred to as _
separable _ in literature .
it is clear that the class of collective measurements includes separable ones and one expects that collective measurements should be more powerful than separable ones in general .
since one can not do better than the best collective measurement , the ultimate precision bound is the one that is asymptotically achieved by a sequence of the best collective measurements as the number of copies tends to infinity .
this fundamental question has been addressed by several authors before @xcite .
it was holevo who developed parameter estimation theory of quantum states by departing from a direct analogy to classical statistics .
he proposed a bound , known as the holevo bound , in the 1970s aiming to derive the fundamental precision limit for quantum parameter estimation problem , see ch .
6 of his book @xcite . at that time
, it was not entirely clear whether or not this bound is a really tight one , i.e. , the asymptotic achievability by some sequence of measurements . over the last decade
, there have been several important progress on asymptotic analysis of quantum parameter estimation theory revealing that the holevo bound is indeed the best asymptotically achievable bound under certain conditions @xcite .
these results confirm that the holevo bound plays a pivotal role in the asymptotic theory of quantum parameter estimation problem . despite the fact that we now have the fundamental precision bound
, the holevo bound has a major drawback : it is not an explicit form in terms of a given model , but rather it is written as an optimization of a certain nontrivial function . therefore , unlike the classical case , where the fisher information can be directly calculated from a given statistical model , the structure of this bound is not transparent in terms of the model under consideration .
having said the above introductory remarks , we wish to gain a deeper insight into the structure of the holevo bound reflecting statistical properties of a given model . to make progress along this line of thoughts ,
we take the simplest quantum parametric model , a general qubit model , and analyze its holevo bound in detail . since explicit formulas for the holevo bound for mixed - state models with one and three parameters and pure - state models are known in literature , the case of two - parameter qubit model is the only one left to be solved .
the main contribution of this paper is to derive an explicit expression for the holevo bound for any two - parameter qubit model of mixed - states without referring to a specific parametrization of the model .
remarkably , the obtained formula depends solely on a given weight matrix and three previously known bounds : the symmetric logarithmic derivative ( sld ) cr bound , the right logarithmic derivative ( rld ) cr bound , and the bound for d - invariant models .
this result immediately provides necessary and sufficient conditions for the two important cases .
one is when the holevo bound coincides with the rld cr bound and the other is when it does with the sld cr bound .
we also show that a general model other than these two special cases exhibits an unexpected property , that is , the structure of the holevo bound changes smoothly when the weight matrix varies .
we note that similar transition has been obtained by others @xcite for a specific parametrization of two - parameter qubit model . here
we emphasize that our result is most general and is expressed in terms only of the weight matrix and two quantum fisher information .
the main result of this paper is summarized in the following theorem ( the detail of these quantities will be given later ) : consider a two - parameter qubit model of mixed states , which changes smoothly about variation of the parameter .
denote the sld and rld fisher information matrix by @xmath1 and @xmath2 , respectively , and define the sld and rld cr bounds by @xmath3&={\mathrm{tr}\left\{w g_\theta^{-1}\right\}},\\ c_\theta^r[w]&={\mathrm{tr}\left\{w \re\tilde{g}_\theta^{-1}\right\ } } + { \mathrm{trabs}\left\{w\im\tilde{g}_\theta^{-1}\right\}},\end{aligned}\ ] ] respectively . here , @xmath4 is a given @xmath5 positive definite matrix and is called a weight matrix ( @xmath6 is defined after eq . . ) .
introduce another quantity by c_^z[w]:=\{w z _ } + \{wz _ } , where @xmath7 is a @xmath8 hermite matrix defined by z_:=_i , j\{1,2},z_^ij:= ( _ l_^j l_^i ) .
here @xmath9 are a linear combination of the sld operators ; @xmath10 with @xmath11 denoting the @xmath12 component of the inverse of sld fisher information matrix and @xmath13 sld operators . with these definitions ,
we obtain the following result : [ thm1 ] the holevo bound for any two - parameter qubit model under the regularity conditions is [ eq1 ] c_^h[w]= c_^r[w ] c_^r[w ] + c_^r[w]+s _ , where the function @xmath14 $ ] , which is nonnegative , is defined by [ hsbound ] s_:= .
note that the condition @xmath15 implies @xmath16 so that @xmath14 $ ] is well defined .
( see the discussion after eq . .
) the above main result [ thm1 ] sheds several new insights on the quantum parameter estimation problems .
first note that the form of the holevo bound changes according to the choice of weight matrices .
this kind of transition phenomenon has never occurred in the classical case .
second surprise is to observe the appearance of the rld cr bound in the generic two - parameter estimation problem .
as we will provide in the next section , the rld cr bound has been shown to be important for a special class of statistical models , known as a d - invariant model . here
, we explicitly show that it also plays a major role for non d - invariant models .
last , in many of previous studies parameter estimation problems , the precision bound is either expressed in terms of the sld or rld fisher information , but the second case of the above expression depends _ both _ on the sld and rld fisher information . to see this explicitly ,
we can rewrite it as @xmath17+s_\theta[w]= { \mathrm{tr}\left\{wg_\theta^{-1}\right\ } } \\ + \frac14\ \frac{\left({\mathrm{trabs}\left\{w\im\tilde{g}_\theta^{-1}\right\}}\right)^2}{{\mathrm{tr}\left\{w(g_\theta^{-1}-\re\tilde{g}_\theta^{-1})\right\}}}. \end{gathered}\ ] ] all these findings will be discussed in details together with examples .
the rest of this paper continues as follows .
section [ sec2 ] provides definitions and some of known results for parameter estimation theory within the asymptotically unbiased setting . in sec .
[ sec3 ] , a useful tool based on the bloch vector is introduced and then the above main theorem is proved .
discussions on the main theorem are presented in sec .
section [ sec5 ] gives several examples to illustrate findings of this paper .
concluding remarks are listed in sec .
most of the proofs for lemmas are deferred to appendix [ sec : appb ] .
supplemental materials are given in appendix [ sec : appc ] .
in this section , we establish definitions and notations used in this paper .
we then list several known results regarding the holevo bound to make the paper self - contained .
consider a @xmath18-dimensional hilbert space @xmath19 ( @xmath20 ) and a @xmath21-parameter family of quantum states @xmath22 on it : : = \{_| = ( ^1,^2, ,^k)^k } , where @xmath23 is an open subset of @xmath21-dimensional euclidean space .
the family of states @xmath24 is called a _ quantum statistical model _ or simply a _ model_. the model discussed throughout the paper is assumed to satisfy certain regularity conditions for the mathematical reasons @xcite . for our purpose ,
the relevant regularity conditions are : i ) the state @xmath22 is faithful , i.e. , @xmath22 is strictly positive .
ii ) it is differentiable with respect to these parameters @xmath25 sufficiently many times .
iii ) the partial derivatives of the state @xmath26 are all linearly independent . in the rest of this paper ,
the regularity conditions above are taken for granted unless otherwise stated .
for a given quantum state @xmath27 , we define the sld and rld inner products by @xmath28 respectively , for any ( bounded ) linear operators @xmath29 on @xmath19 . here , @xmath30 denotes the hermite conjugation . given a @xmath21-parameter model @xmath31 , the sld operators @xmath32 and rld operators @xmath33 are formally defined by the solutions to the operator equations : @xmath34 the sld fisher information matrix is defined by @xcite @xmath35_{i , j\in\{1,\dots , k\ } } \\
\nonumber g_{\theta , ij}&:={\langlel_{\theta , i},l_{\theta , j}\rangle_{\rho_{\theta}}}={\mathrm{tr}\left(\rho_{\theta}\frac12 \big(l_{\theta , i}l_{\theta , j}+l_{\theta , j}l_{\theta , i } \big)\right ) } , \end{aligned}\ ] ] and the rld fisher information is @xcite @xmath36_{i , j\in\{1,2,\dots , k\ } } \\ \label{rldfisher } \tilde{g}_{\theta , ij}&:={\langle\tilde{l}_{\theta , i},\tilde{l}_{\theta , j}\rangle_{\rho_{\theta}}^{+}}={\mathrm{tr}\left(\rho_{\theta}\tilde{l}_{\theta , j}\tilde{l}_{\theta , i}^*\right)}. \end{aligned}\ ] ] define the following linear combinations of the sld and rld operators : l_^i:= _
j=1^k(g_^-1)^jil_,j _ ^i:=_j=1^k(_^-1)^ji_,j . from these definitions ,
the following orthogonality conditions hold .
[ orthcond ] _
^i , l_,j__=^i_j , _ ^i,_,j__^+=^i_j .
these operators with upper indices are referred to as the _ sld and rld dual operators _ , respectively .
consider @xmath0th i.i.d .
extension of a given state and we define @xmath0th extended model by [ nmodel ] ^(n):=\{_^n | ^k } .
the main objective of quantum statisticians is to perform a measurement on the @xmath0 tensor state @xmath37 and then to make an estimate for the value of the parameter @xmath25 based on the measurement outcomes . here
measurements are described mathematically by a positive operator - valued measure ( povm ) and is denoted as @xmath38 .
an estimator , which is a purely classical data processing , is a ( measurable ) function taking values on @xmath23 and is denoted as @xmath39 .
they are @xmath40 where @xmath41 is the identity operator on @xmath19 and we assume that povms consist of discrete measurement outcomes . for continuous povms
, we replace the summation by an integration .
a pair @xmath42 is called a _ quantum estimator _ or simply an _ estimator _ when is clear from the context and is denoted by @xmath43 @xcite .
the performance of a particular estimator can be compared to others based on a given figure of merit and then one can seek the best " estimator accordingly . as there is no
universally accepted figure of merit , one should carefully adopt a reasonable one depending upon a given situation .
for example , a specific prior distribution for the parameter @xmath25 is known , the bayesian criterion would be meaningful to find the best bayesian estimator .
if one wishes to avoid bad performance of estimators , the min - max criterion provides an optimal one that suppresses such cases . in this paper , we are interested in analyzing estimation errors at specific point @xmath44 , that is , the pointwise estimation setting . for a given model and an estimator @xmath45
, we define a _ bias _ at a point @xmath25 as @xmath46&:=\sum_{x\in\cx_n}(\hat\theta_{n}(x)-\theta)p^{(n)}_\theta(x)=\eof{\hat\theta_{n}}-\theta,\\ \mathrm{with}&\quad p^{(n)}_\theta(x ) : = { \mathrm{tr}\left(\rhon\pi^{(n)}_x\right)},\end{aligned}\ ] ] where @xmath47 denotes the expectation value of a random variable @xmath48 with respect to the probability distribution @xmath49 .
note that the bias @xmath50 $ ] is a @xmath21-dimensional real vector .
an important class of estimators when estimating the specific point of the model is the locally unbiased estimatior .
this is to restrict estimators such that the bias vanishes at the true point @xmath25 up to the first order in the taylor expansion .
mathematically , an estimator @xmath43 is called _ locally unbiased _ at @xmath25 if [ lucondition ] b^(n)_=0 = ^j_i , hold at @xmath44 .
it is known that the quantum cr bounds hold for any locally unbiased estimator @xcite . upon analyzing performance of estimators within the asymptotic regime ,
we should impose some conditions that restrict the class of estimators . in statistics ,
a sequence of an estimator is said ( weakly ) _ consistent _ , if it converges to the true value in probability for every value @xmath44 , i.e. , > 0_n\{|_n-|>}=0 , holds for all @xmath44 . in this expression , @xmath51 denotes the standard euclidean norm and the right hand side means that error probability can be made arbitrary small . as a good estimator must converge to the true value as @xmath0 goes to infinity , it is reasonable to look for the class of consistent estimators in quantum parameter estimation as well . in classical statistics
, this condition of consistency alone turns out to be weak in order to exclude artificial estimators .
there are several approaches to handle these problems in the classical case @xcite . rather than going into mathematical discussions ,
we simply look for the following class of estimators to avoid such a situation .
a sequence of estimators @xmath52 is called _ asymptotically unbiased _ if it satisfies [ aucondition ] _
n b^(n)_=0 , _
n = ^j_i , for all @xmath53 and for all @xmath44 . that is to require the locally unbiased condition in the @xmath54 limit . to quantify estimation errors of a given estimator
, we consider the _ mean - square error _ ( mse ) matrix defined by @xmath55&:=\left [ v^{ij}_{\theta , n}[\hat\pi^{(n ) } ] \right]_{i , j\in\{1,2,\dots , k\}},\\ v^{ij}_{\theta , n}[\hat\pi^{(n)}]&:=\sum_{x\in\cx_n } ( \hat{\theta}_{n}^i(x)-\theta^i ) ( \hat{\theta}_{n}^j(x)-\theta^j)p^{(n)}_\theta(x).\end{aligned}\ ] ] by definition , the mse matrix is a @xmath56 real symmetric matrix and it is straightforward to show that it is nonnegative . as stated in the introduction , we wish to find the best precision bound allowed by the laws of quantum theory and statistics , which is achievable in the @xmath54 limit . in the classical case
, one can directly minimize the mse matrix as a matrix inequality over the class of asymptotically unbiased estimators and to find the lowest mse error achievable as @xmath54 .
this line of approach does not work in the quantum case .
one way to tackle this question is to deal with a weighted trace of the mse matrix , which is a scalar quantity , and it is defined by _
^(n)[^(n)|w]:=\{w v^(n)_}. here the matrix @xmath4 is called a _ weight matrix _ and can be chosen arbitrary as long as it is real and strictly positive .
since the weight matrix is one of the important ingredient for our discussion , let us denote the set of all possible weight matrices by [ weightset ] : = \{w^kk| w>0}. by changing the weight matrix , one can explore trade - off relations in estimating different parameter components @xmath57 .
we note that the weight matrix can depend on the value of the estimation parameter @xmath25 as well .
for example , it can be chosen as the sld fisher information matrix . defining these terminologies
, we now state the problem : for a specific point of a given i.i.d .
model @xmath58 , what is the best sequence of estimators @xmath59 and the minimum value of the weighted trace of mse ? to put it differently , one wishes to find the optimal sequence of estimators that minimum of the first order coefficient @xmath60 $ ] in the large @xmath0 expansion : _
^(n)[^(n)|w ] + o(n^-2 ) , i.e. , the fastest decaying rate for the mse .
mathematically , we define the cr type bound for the mse by the following optimization problem : c_:=_\{^(n ) } \ { _ nn_^(n)[^(n)|w ] } , where the infimum is taken over all possible sequences of estimators @xmath52 that is asymptotically unbiased ( a.u . ) .
note that this bound depends both on the weight matrix @xmath4 and the model @xmath22 at @xmath25 .
the symbol @xmath25 appearing in the bound @xmath60 $ ] represents the model @xmath22 at @xmath25 . unlike the bayesian or the min - max settings mentioned before
, we are interested in understanding statistical properties of a given parametric model
. this would be important in particular study of quantum states from geometrical point of view @xcite .
to define the holevo bound , we need some definitions first . for a given quantum statistical model on @xmath19 , denote a @xmath21 array of hermite operators on @xmath19 by @xmath61 , @xmath62 , i.e. , @xmath63 , for all @xmath64 , and define the set @xmath65 by @xmath66 the holevo function @xcite in the quantum estimation theory is defined by [ hfunction ] h_:=\{wz_}+\{wz _ } , where the @xmath56 hermite matrix @xmath67 $ ] is [ zmatrix ] z_:= _ i , j\{1, ,k } , and trabs@xmath68 denotes the sum of the absolute values of @xmath69 with @xmath70 for some invertible matrix @xmath71 .
we note the following relation also holds for any anti - symmetric operator @xmath68 : \{wx}=_i|_i|=\{|w^1/2xw^1/2 | } , where @xmath72 denotes the absolute value of a linear operator @xmath68 .
the holevo bound is defined through the following optimization : [ hbound ] c_^h[w]:=__h_. the derivation of the above optimization is well summarized in hayashi and matsumoto @xcite .
holevo showed that this quantity is a bound for the mse for estimating a single copy of the given state under the locally unbiased condition @xcite : _
^(1)[|w]c_^h[w ] , holds for any locally unbiased estimator @xmath73 .
the nontrivial property of the holevo bound is the additivity @xcite : c_^h[w,]=n^-1c_^h[w , _ ] , where the notation @xmath74 $ ] represents the holevo bound about @xmath0th extended model .
the following theorem establishes that the holevo bound is the solution to the problem of our interest : [ thm2 ] for a given model satisfying the regularity conditions , @xmath60=c_\theta^h[w]$ ] holds for all weight matrices .
there exist several different approaches upon proving the above theorem .
hayashi and matsumoto @xcite proved the case for a full qubit model first .
gu and kahn @xcite introduced a different tool based on ( strong ) quantum local asymptotic normality to prove the qubit case .
this was further generalized to full models on any finite dimensional hilbert space @xcite . however , all these proofs depend on a specific parametrization of quantum states .
more general proof has been recently established by yamagata , fujiwara , and gill @xcite .
this theorem implies that if we choose an optimal sequence of estimators , the mse behaves as \{w v^(n ) _ } + o(n^-2 ) , for sufficiently large @xmath0 .
that is the holevo bound is the fastest decaying rate for the mse .
although the holevo bound stands as an important cornerstone to set the fundamental precision bound , the definition contains a nontrivial optimization .
the main motivation of our work , as stated in the introduction , is to perform this optimization explicitly for any given model for qubit case .
the result shows several nontrivial aspects of parameter estimation in quantum domain . before going to present our result ,
we summarize several known results . in this subsection
, we consider two special cases where analytical forms of the holevo bound are known . for a given @xmath21-parameter model on the hilbert space @xmath19 ,
let us denote sld and rld fisher information matrices by @xmath1 and @xmath2 , respectively , eqs .
( [ sldfisher ] , [ rldfisher ] ) .
define the sld and rld cr bounds by @xmath75&:={\mathrm{tr}\left\{wg_\theta^{-1}\right\}},\\ \label{rldbound } c_\theta^r[w]&:={\mathrm{tr}\left\{w\re\tilde{g}_\theta^{-1}\right\}}+{\mathrm{trabs}\left\{w{\im}\tilde{g}_\theta^{-1}\right\}},\end{aligned}\ ] ] respectively . throughout the paper ,
we use the notation @xmath76 ( @xmath77 ) representing the real ( imaginary ) part of the inverse matrix of the rld fisher information matrix .
the well - known fact is that the sld and rld cr bounds can not be better than the holevo bound : [ lem2 ] for a given model satisfying the regularity conditions , the holevo bound is more informative than the sld and the rld cr bound , i.e. , @xmath78\ge c_\theta^s[w]$ ] and @xmath78\ge c_\theta^r[w]$ ] hold for an arbitrary weight matrix @xmath4 .
proof can be found in the original work by holevo that is summarized in his book @xcite .
more compact proof was stated by nagaoka @xcite .
see also hayashi and matsumoto @xcite .
when the number of parameters is one , the problem can be reduced significantly . in this case , there can not be any imaginary part for the matrix and thus the minimization is reduced to minimizing the mse itself .
[ thm3 ] for any one - parameter model , the holevo bound coincides with the sld cr bound , i.e. , [ 1bound ] c_^h= , holds for all @xmath44 where @xmath79 is the sld fisher information at @xmath25 .
note that there is no weight matrix since we are dealing with a scalar mse for the one - parameter case .
importantly , there is no gain from collective povms for one - parameter models .
existence of a povm whose mse is equal to this bound is discussed independently by several authors @xcite .
consider an arbitrary @xmath21-parameter model @xmath24 and let @xmath13 ( @xmath64 ) be the sld operators at @xmath25 .
the linear span of sld operators with real coefficients is called the _ sld tangent space _ of the model at @xmath25 : t_():=_\{l_,1,l_,2, ,l_,k } .
any elements of the sld tangent space , @xmath80 , satisfy @xmath81 and it is not difficult to see that the space @xmath82 is essentially a real vector space with the dimension @xmath21 .
holevo introduced a super - operator @xmath83 , called a _ commutation operator _ , as follows .
given a state @xmath84 on @xmath19 , let @xmath85 be the set of linear operators on @xmath19 , then @xmath83 is a map from @xmath85 to itself defined through the following equation : y,(x)_=[y , x]_,x , y . here
, @xmath86 is the sld inner product and @xmath87_\rho={\mathrm{tr}\left(\rho[x^*,\,y]\right)}/(2{\mathrm{i}})$ ] is a sesqui - linear form .
( here , the definition is different from the original one by a factor . )
when considering a parametric model , we denote @xmath88 for simplicity .
we say that a model is _ d - invariant _ at @xmath25 if the sld tangent space at @xmath25 is invariant under the action of the commutation operator .
mathematically , this definition is expressed as ( x)t _ ( ) , xt _ ( ) . when a model is d - invariant for all @xmath44 , we say the model @xmath24 is globally d - invariant .
lemma [ lem4 ] in appendix [ sec : appc-1 ] characterizes equivalent conditions for d - invariant models . from the definitions of two inner products and the commutation operator , the relationship @xmath89 holds for all linear operators @xmath29 on @xmath19 .
for a given model , another important relation ( _ i_x)=,l_,i__=,_,i__^+ , holds for @xmath90 . combining them
gives @xmath91 , and hence we obtain [ app-0 ] l_,i=(i+)(_,i ) .
two more useful relations are @xmath92 which can be checked directly from the definitions .
it is well known that the holevo bound gets simplified significantly if the model is d - invariant @xcite .
importantly , d - invariant model enjoys the following proposition , which is due to holevo : [ propdinv ] let @xmath9 ( @xmath93 ) be the sld ( rld ) dual operator and @xmath1 ( @xmath2 ) be the sld ( rld ) fisher information matrix , respectively . define a @xmath56 hermite matrix by @xmath94_{i , j\in\{1,\dots , k\}}$ ] .
when the model is d - invariant at @xmath25 , @xmath95 holds at @xmath25 and further the holevo bound is expressed as w c_^h[w]=c_^r[w ] = h_. this statement can be proven in several different manners @xcite .
in passing , we note that the expression @xmath96 $ ] in the above proposition is also expressed as @xmath96={\mathrm{tr}\left\{w \re z_\theta\right\ } } + { \mathrm{trabs}\left\{w\im z_\theta\right\}}$ ] in terms of the matrix @xmath7 .
when the model is not d - invariant , @xmath96 $ ] does not seem to play any important role .
this is because the quantity [ grldbound ] c_^z[w]:=\{w z _ } + \{wz _ } , is always greater or equal to the holevo bound , i.e. , @xmath97\ge c_\theta^h[w ] $ ] for all weight matrices . nevertheless ,
as will be shown in this paper , this is an important quantity and we call it as the _ d - invariant bound _ in our discussion
. we note that this quantity was also named as the _ generalized rld cr bound _ by fujiwara and nagaoka @xcite in the following sense .
when a model fails to satisfy some of the regularity conditions , the rld operators do not exist always . even in this case , when the model is d - invariant , then the above bound is well defined and provides the achievable bound for a certain class of models , known as the coherent model @xcite .
another remark regarding this proposition is that the converse statement also holds .
[ thmdinv ] for any @xmath21-parameter model @xmath24 on any dimensional hilbert space under the regularity conditions , the following equivalence holds : @xmath98=c_\theta^r[w].\\ \leftrightarrow&\ \forall w\in\cw\ c_\theta^h[w]=c_\theta^z[w ] .
\end{aligned}\ ] ] this equivalence for the d - invariant model might have been known for some experts , but it was not stated explicitly in literature to our knowledge @xcite .
sketch of proof is given in appendix [ sec : appc-2 ] for the sake of reader s convenience .
we remark that the holevo bound for a general model , which is not d - invariant , exhibits a gap among @xmath97 $ ] , @xmath99 $ ] , and @xmath100 $ ] .
the following relation holds in general : c_^z[w]c_^h[w]\{c_^s[w],c_^r[w ] } , for all weight matrices @xmath4 . from this general inequality ,
it is clear that the condition of d - invariance shrinks the gap between @xmath97 $ ] and @xmath99 $ ] to zero .
the holevo bound then coincides with the rld and d - invariant bounds .
in this section we consider a model for quantum two - level system , a qubit model . for mixed - state models , where parametric states are rank-2 for all @xmath44 , possible numbers of the parameters are from one to three . as stated in sec .
[ sec2 - 3 ] , the holevo bound for one - parameter qubit model is solved and is given by theorem [ thm3 ] . when the number of parameters is equal to three , on the other hand , it is easy to show that the model becomes d - invariant as follows . since three sld operators are linearly independent , they expand the set of all linear operators @xmath68 satisfying the condition @xmath81 .
in other words , the sld tangent space is same as this space and hence the sld tangent space is always d - invariant . in this case , the holevo bound is given as eq . .
therefore , the two - parameter case needs to be solved explicitly . in the following , we consider two - parameter qubit models of mixed states only .
further , the regularity conditions mentioned before are assumed throughout our discussion . upon performing this optimization to derive an explicit formula for the holevo bound , it is convenient to utilize the bloch - vector formalism .
a similar technique has been used by watanabe , where all operators are expanded in terms of a basis of lie algebras @xcite . in the next subsection
, we present necessary machinery and then solve the two - parameter case . in this subsection , we present a formalism in which sld operators are represented by three - dimensional real vector .
this is motivated by the well - known one - to - one mapping between a given qubit state and three - dimensional real vector .
thus , any qubit model can be represented by a family of three - dimensional real vectors as [ qbmodel ] _
= \{_=(s_^1,s_^2,s_^3 ) | } , with @xmath101 the interior of the bloch ball . to simplify notations ,
we define the standard inner product and the outer product for three - dimensional complex vectors by @xmath102_{i , j\in\{1,2,3\}},\ ] ] respectively , where @xmath103 denotes the complex conjugation of @xmath104 .
the outer product is a @xmath105 matrix whose action onto a vector @xmath106 is @xmath107 .
we first observe that the one - to - one correspondence between the sld operator and a four - dimensional vector when is expanded in terms of the basis @xmath108 with @xmath109 usual pauli spin matrices for spin-1/2 particles .
since the sld operators belong to the sld tangent space , the relation @xmath110 holds , i.e. , they are orthogonal to the identity operator with respect to the sld inner product .
this leads to the following constraint : ( l_,i)=-| v_,i , where @xmath111 with @xmath112 is a three - dimensional real vector .
thus , we have a one - to - one mapping from the sld operator @xmath13 to the three - dimensional real vector @xmath113 .
the vector @xmath113 shall be referred to as the _ sld bloch vector _ in this paper .
it is straightforward to solve the operator equation , which defines the sld operators , and the sld bloch vector is v_,i= _ i+ , where @xmath114 denotes the length of the bloch vector and @xmath115 is the @xmath116th partial derivative . to proceed further
, we find it convenient to introduce a @xmath117 matrix : [ qop1 ] q_:=+ , with @xmath118 the identity matrix acting on the three - dimensional vector space @xmath119 .
it follows from the definition that @xmath120 is a real and positive matrix with eigenvalues @xmath121 and its inverse is [ qop2 ] q_^-1=-|| .
the sld bloch vector is then expressed as v_,i = q__i ( q^-1_v_,i= _ i ) .
the @xmath12 component of the sld fisher information is [ sldqubit ] g_,ij = v_,i|q_^-1v_,j= _ i |q__j .
let @xmath122 be the @xmath12 component of the inverse sld fisher matrix and we define the sld bloch dual vector @xmath123 by [ sldcot ] v_^i=_jg^ji_v_,j , then , the following orthogonality condition holds : v_^i|q_^-1v_,j=^i_j , which corresponds to eq . .
the inverse of sld fisher information matrix is also expressed as [ sldqinv ] g^ij_=v_^i|q_^-1v_^j . the same line of arguments holds for rld operators and rld fisher information .
the only difference is here is that the rld bloch vector becomes complex in general .
define a @xmath117 complex matrix : [ qtop1 ] _ : = ( -f _ ) , where @xmath124 with @xmath125 the completely antisymmetric tensor .
the action of @xmath126 is to give the exterior product of two vectors , i.e. , @xmath127 for @xmath128 . from this definition , @xmath129 is also strictly positive and its inverse is given by [ qtop2 ] _ ^-1=-||+f_=q_^-1+f_. the rld bloch vector is _ , i=__i ( ^-1__,i= _ i ) , and the rld fisher information matrix is [ rldqubit ] _ ,
ij=_,i|_^-1_,j= _ i |__j .
define the rld bloch dual vector by [ rldcot ] _
^i=_j^ji__,j , then we have _ ^i|_^-1_,j=^i_j .
other useful relations are listed below without detail calculations .
first , there is a one - to - one correspondence between sld and rld bloch vector .
this is given by v_,i=(+f _ ) _ , i ( _ , i=(q_^-1-f _ ) v_,i ) .
second , the vector @xmath130 is orthogonal to the bloch vector @xmath131 . defining [ gamma ] _ ,
i:=|v_,i,_,i:=|_,i , this is expressed as _
, i=_,i .
third , the sld fisher information and real part of the rld fisher information are related by [ sldqrld ] g_,ij-_,ij=_,i_,j . in other words , the matrix @xmath132 is rank one . in this subsection
, we consider an arbitrary two - parameter qubit model , that is the parameter to be estimated is @xmath133 . in order to derive an explicit expression for the holevo bound , we first rewrite the holevo bound in terms of the bloch vectors .
a linear operator which satisfies @xmath134 can be expressed as x^i=-|x^ii+x^i .
let @xmath135 be the orthogonal space to the @xmath116th derivative of the bloch vector and an element of the set appearing in the definition takes the form of @xmath136 with @xmath137 thus , the set of operators @xmath138 can be mapped to the set of vectors @xmath139 . using this form of bloch vector representation ,
the @xmath12 component of the @xmath140 $ ] matrix and the holevo function read @xmath141&= { \langle\v x^i|\tilde{q}_\theta^{-1}\v x^j\rangle } , \\
\nonumber h_\theta[\vec{x}|w]&=\sum_{i , j=1}^2 \left [ w_{ij}{\langle\v x^i|q_\theta^{-1}\v x^j\rangle } + \sqrt{\det w}\big| { \langle\v x^i|f_\theta\v x^j\rangle } \big| \right],\end{aligned}\ ] ] for a given @xmath5 weight matrix @xmath142_{i , j\in\{1,2\}}$ ] .
we note that the holevo function is a quadratic function of the six - dimensional vector @xmath143 .
the minimization of this function under the constraints can be handled with a standard procedure .
the only point needs to be taken is that the function is not differentiable for all points . since the number of free variables for the optimization is @xmath144(the number of independent constraints)@xmath145 , we take the following substitution : x^i = v_^i+^i v_^ , where @xmath146 is a vector orthogonal to both @xmath147 and @xmath148 and @xmath149 is a free variable without any constraint . with this substitution , the holevo function
is significantly simplified as follows .
[ lem - holevo2 ] for a two - parameter qubit model , the holevo bound takes the following minimization form without any constraint : c_^h[w]=_^2h _ , where the function @xmath150 $ ] is defined by @xmath151={\mathrm{tr}\left\{wg_\theta^{-1}\right\ } } + { \langle\v\ell_{\theta}^\bot|{q}_{\theta}^{-1}\v\ell_{\theta}^\bot\rangle } ( \vec\xi|w\vec\xi)\\ + 2\sqrt{\det w}\left| { \langle\v \ell_\theta^1|f_\theta\v \ell_\theta^2\rangle}+(1-s_\theta^2)(\vec{\gamma}_\theta|\vec{\xi } ) \right| . \end{gathered}\ ] ] in this expression , we introduce the standard inner product for two - dimensional real vector space by @xmath152 and @xmath153 is given by eq . .
the derivation for this lemma is given in appendix [ sec : appb-0 ] .
in the following , we carry out the above optimization to derive the main result of this paper , theorem [ thm1 ] .
we first list several definitions and lemmas . for a given two - parameter qubit mixed - state model ,
the sld cr , rld cr , and d - invariant bounds are defined by @xmath3&={\mathrm{tr}\left\{w g_\theta^{-1}\right\}},\\ \nonumber c_\theta^r[w]&={\mathrm{tr}\left\{w \re\tilde{g}_\theta^{-1}\right\ } } + { \mathrm{trabs}\left\{w\im\tilde{g}_\theta^{-1}\right\}},\\ \nonumber c_\theta^z[w]&={\mathrm{tr}\left\{w \re z_\theta\right\ } } + { \mathrm{trabs}\left\{w\im z_\theta\right\ } } , \end{aligned}\ ] ] respectively , where @xmath1 is the sld fisher information matrix , @xmath2 is the rld fisher information matrix , and @xmath154_{i , j\in\{1,2\}}$ ] with @xmath155 as before .
[ lem7 ] for any two - parameter qubit mixed - state model , the following relations hold : 1 . @xmath156 .
2 . @xmath157 .
3 . @xmath158-c_\theta^r[w ] \right)$ ] .
[ lem8 ] for any two - parameter qubit model @xmath24 , the following conditions are equivalent . 1 .
@xmath24 is d - invariant at @xmath25 .
2 . @xmath159 at @xmath25 .
3 . @xmath160 at @xmath25 .
furthermore , we have the following equivalent characterization for global d - invariance . 1 .
@xmath24 is globally d - invariant .
@xmath159 for all @xmath161 .
@xmath162 is independent of @xmath25 .
three remarks regarding the above lemmas are listed : first , imaginary parts of the inverse of the rld fisher information matrix and the @xmath7 matrix are always identical for two - parameter qubit mixed - state models , i.e. , @xmath163 , see proof in the appendix .
second , if a model expressed as the bloch vector contains the origin @xmath164 , the model is always d - invariant at this point .
this is because the condition @xmath165 is met at @xmath166 .
last , a globally d - invariant model is possible if and only if the state is generated by some unitary transformation .
this is because the condition [ lem8]-6 in lemma [ lem8 ] is equivalent to preservation of the length of the bloch vector .
finally , we need the following lemma for the optimization : [ lem9 ] for a given positive @xmath8 matrix @xmath167 , a real vector @xmath168 , and a real number @xmath169 , the minimum of the function f()=(|a)+2|(|)+c| , is given by _ ^2f()=
2|c|-(|a^-1 ) |c|(|a^-1 ) + |c|<(|a^-1 ) , where the minimum is attained by _ * = -(c)a^-1
|c|(|a^-1 ) + -a^-1 |c|<(|a^-1 ) , where @xmath170 is the sign of @xmath169 .
proofs for the above three lemmas are given in appendix [ sec : appb-2]-[sec : appb-4 ] .
we now prove theorem [ thm1 ] . from the expression of the holevo function
, we can apply lemma [ lem9 ] by identifing @xmath171 we need to evaluate @xmath172 and @xmath173 and they are calculated as follows .
@xmath174-c_\theta^r[w ] , \end{aligned}\ ] ] where lemma [ lem7]-1 and [ lem7]-3 are used to get the last line .
lemma [ lem7]-2 immediately gives 2|c|=\{wz_}=c_^z[w]-c_^s[w ] .
we obtain if |c|(|a^-1 ) c_^r[w]12 ( c_^z[w]+c_^s[w ] ) is satisfied , the holevo bound is @xmath175&=\nonumber \ds c_\theta^s[w]+ ( c_\theta^z[w]-c_\theta^s[w])- ( c_\theta^z[w]-c_\theta^r[w])\\ & = c_\theta^r[w ] . \end{aligned}\ ] ] if @xmath176 @xmath177 @xmath99<({c_\theta^z[w]+c_\theta^s[w]})/{2}$ ] is satisfied , on the other hand , the holevo bound takes the following form : @xmath178 & = c_\theta^s[w]+\frac{\left[(c_\theta^z[w]-c_\theta^s[w])/2\right]^2}{c_\theta^z[w]-c_\theta^r[w ] } \nonumber\\ & = c_\theta^r[w]\nonumber\\ + & \frac{1}{c_\theta^z[w]-c_\theta^r[w ] } \left(\frac{c_\theta^z[w]+c_\theta^s[w]}{2}- c_\theta^r[w ] \right)^2 \nonumber\\ & = c_\theta^r[w]+s_\theta[w ] , \end{aligned}\ ] ] where the function @xmath14 $ ] is defined in eq . .
this proves the theorem .
@xmath179 we remark that from lemma [ lem7 ] and the positivity of @xmath4 we always have the relation [ zger ] c_^z[w]c_^r[w ] , and the equality if and only if @xmath180
@xmath24 is d - invariant at @xmath25 by lemma [ lem8]-3 .
note if @xmath24 is d - invariant at @xmath25 ( @xmath160 ) , the condition @xmath99<({c_\theta^z[w]+c_\theta^s[w]})/{2}$ ] @xmath177 @xmath181-c_\theta^r[w ] $ ] can not be satisfied .
thus , the obtained holevo is well defined for all @xmath25 and for arbitrary weight matrix @xmath4 .
the optimal set of hermite operators attaining the holevo bound can be given by lemma [ lem9 ] as follows .
define an hermite matrix by l_^:= -|v_^i+v_^ , and the function @xmath182 by @xmath183\ge\frac12(c_\theta^z[w]+c_\theta^s[w])\\[1ex ] \frac{w^{1/2}\im z_\theta^{12}}{c_\theta^z[w]-c_\theta^r[w ] } \quad \mathrm{otherwise } \end{cases}. \end{gathered}\ ] ] then , we have @xmath184 $ ] as @xmath185 before moving to discussion and consequence of the main result , we present the following two - alternative expressions . define a function , h(x):= 2|x|-1 + x^2 , which is continuous and the first derivative is also continuous for all @xmath186 . then , the holevo bound is written in a unified manner : @xmath187=c_\theta^s[w]+(c_\theta^z[w]-c_\theta^r[w ] ) \\
\times h\left(\frac{(c_\theta^z[w]-c_\theta^s[w])/2}{c_\theta^z[w]-c_\theta^r[w]}\right ) .
\end{gathered}\ ] ] this expression needs a special care when @xmath97-c_\theta^r[w]\to 0 $ ] .
this case should be understood as the limit @xmath188 . the other expression shown in eq
. follows from the first line of eq . by noting @xmath97-c_\theta^r[w]= \mathrm{tr}\{w(g_\theta^{-1}-\re\tilde{g}_\theta^{-1})\}$ ] and
@xmath97-c_\theta^s[w]=\mathrm{trabs}\{w\im \tilde{g}_\theta^{-1}\}$ ] .
in this section , we shall discuss the consequences of theorem [ thm1 ] .
this brings several important findings of our paper .
first is two conditions that characterize special classes of qubit models .
second is a transition in the structure of the holevo bound depending on the choice of the weight matrix .
the general formula for the holevo bound for any two - parameter model is rather unexpected in the following sense .
first of all , it is expressed solely in terms of the three known bounds and a given weight matrix .
second , a straightforward optimization for a nontrivial function reads to the exactly same expression as the rld cr bound when the condition @xmath99\ge(c_\theta^z[w]+c_\theta^s[w])/2 $ ] is satisfied .
as noted before , this condition explicitly depends on the choice of the weight matrix @xmath4 . at first sight , this seems to be in contradiction with the general theorem [ thmdinv ] stating that the rld cr bound can be attained if and only if the model is d - invariant .
therefore , we must examine that the holevo bound is identical to the rld cr bound if and only if the model is d - invariant based on theorem [ thm1 ] .
the following proposition confirms that this is indeed so .
we note that this statement is a special case of theorem [ thmdinv ] . here , its proof becomes extremely simple with the obtained formula .
[ thm10 ] for any two - parameter qubit model , the holevo bound at @xmath25 becomes same as the rld cr bound for all positive weight matrices if and only if the model is d - invariant at @xmath25 .
that is , [ thmeq10 ] w c_^h[w]= c_^r[w ] .
when the model is d - invariant , the condition @xmath99<({c_\theta^z[w]+c_\theta^s[w]})/{2}$ ] can not be satisfied .
( see the remark after eq . . ) therefore , the holevo bound is always identical to the rld cr bound in this case .
next we show the left condition implies the right in eq . .
if @xmath189= c_\theta^r[w]$ ] holds for all @xmath4 , the relation @xmath190 must be true for all @xmath4 .
but , if @xmath191 , this is impossible . to show it ,
let us suppose @xmath192 , then we can change @xmath193 to get [ proof10 ] | v_^1|f_v_^2| b^-1 1 , w .
it is easy to show that @xmath194 as a function positive matrix @xmath195 satisfies @xmath196 as changing positive matrix @xmath195 .
that is @xmath197 can be made arbitrary small by choosing @xmath195 and hence the condition can not hold .
this gives @xmath198 and hence this is equivalent to the d - invariant condition from lemma [ lem8 ] . @xmath179
using theorem [ thm1 ] , we now state one more important condition characterizing the model where the holevo bound coincides with the sld cr bound .
[ thm11 ] for any two - parameter qubit model , the holevo bound coincides with the sld cr bound for all positive weight matrices if and only if the imaginary part of @xmath7 vanishes .
further , this condition is equivalent to the existence of a weight matrix @xmath199 such that the holevo bound and the sld cr bound are same with this particular choice @xmath199 @xcite .
mathematically , we have @xmath200= c_\theta^s[w]\\ \label{thmeq12 } \leftrightarrow&\ \im z_\theta=0\\ \label{thmeq13 } \leftrightarrow&\ \exists w_0 \in\cw\ c_\theta^h[w_0]= c_\theta^s[w_0 ] .
\end{aligned}\ ] ] @xmath201 : the condition @xmath202 implies @xmath203 for all @xmath204 . in this case , the condition @xmath99<({c_\theta^z[w]+c_\theta^s[w]})/{2}$ ] is always satisfied , and hence the holevo bound coincides with the sld cr bound for all choices of the weight matrix .
this follows from the second line of the expression .
@xmath201 : the condition implies an existence of a weight matrix satisfying @xmath189=c_\theta^s[w]$ ] .
@xmath201 : next let us assume @xmath205= c_\theta^s[w_0]$ ] for some weight matrix @xmath206 .
the expression immediately implies @xmath207 which gives @xmath208 .
therefore , three conditions are all equivalent .
we have three remarks regarding this proposition .
first , in terms of the sld bloch vectors , the necessary and sufficient condition is also written as @xmath209\right)}=0\\ & \leftrightarrow{\langle\stheta|\v \ell_{\theta,1}\times\v \ell_{\theta,2}\rangle}=0\\ & \leftrightarrow\ { \langle\stheta| \del_{1}\stheta\times\del_2\stheta\rangle}=0,\end{aligned}\ ] ] which is easy to check by calculating the bloch vector of a given model .
second , we note that given a symmetric matrix @xmath167 , @xmath210 for all positive weight matrix @xmath4 implies @xmath211 as a matrix inequality . when the holevo bound is same as the sld cr bound , we see that the mse matrix @xmath212 $ ] for any asymptotically unbiased estimators satisfy the sld cr inequality : _ nn v^(n)_g_^-1 , as a matrix inequality
. moreover , there exists a sequence of estimators that attains this matrix equality .
this is rather counter intuitive since two sld operators @xmath213 and @xmath214 do not commute in general .
therefore , the condition @xmath215 seems to grasp asymptotic commutativity of two sld operators in some sense .
indeed , this condition can be written as @xmath216\right)}=0 $ ] , i.e. , commutativity of @xmath213 and @xmath214 on the trace of the state @xmath22 . when this holds , the quantum parameter estimation problem becomes similar to the classical case asymptotically . in the rest of the paper
, we call a model _ asymptotically classical _ when this condition is satisfied
. a similar terminology , `` quasi classical model , '' was used by matsumoto in the discussion of parameter estimation of pure states @xcite . here , we emphasize that classicality arises only in the asymptotic limit and hence , this terminology is more appropriate .
we also note that the equivalence between @xmath201 was stated in the footnote of the paper @xcite based on the unpublished work of hayashi and matsumoto . here
our proof is shown to be simple owing to the general formula obtained in this paper .
last , a great reduction occurs in the structure of the fundamental precision bound for this class of models .
we note that achievability of the sld cr bound for specific models have been reported in literature in discussion on quantum metrology @xcite . here
, we provide a simple characterization , the necessary and sufficient condition , of such special models in the unified manner .
having established the above two propositions , we can conclude that a generic two - parameter qubit model other than d - invariant or asymptotically classical ones exhibits the nontrivial structure for the holevo bound in the following sense : the structure changes smoothly as the weight matrix @xmath4 varies . for a certain choice of @xmath4 , it coincides with the rld cr bound and it becomes different expression for other choices .
put it differently , consider any model that is not asymptotically classical , then we can always find a certain region of the weight matrix set @xmath217 such that the holevo bound is same as the rld cr bound . this point is examined in detail in the next subsection and examples are provided in the next section for illustration .
let us consider a two - parameter qubit model that is neither d - invariant nor asymptotically classical . in this case
, the set of all possible weight matrices @xmath218 is divided into three subsets .
the first two sets are @xmath219 and @xmath220 in which @xmath99-({c_\theta^z[w]+c_\theta^s[w]})/{2}$ ] is positive and negative , respectively . the last is the boundary @xmath221 that consists of a family of weight matrices satisfying the equation : b_:=c_^r[w]-12(c_^z[w]+c_^s[w])=0 . according to this division , the holevo bound takes the form @xmath189=c_\theta^r[w]$ ] for @xmath222 and it is expressed as @xmath189=c^r_\theta[w]+s_\theta[w]$ ] for @xmath223 .
whereas , @xmath189=c_\theta^r[w]=s_\theta[w]$ ] holds for the boundary @xmath224 . in the following we characterize these sets explicitly . to do this
, we first note that the degree of freedom for the weight matrix in our problem is three due to the condition of @xmath4 being real symmetric .
second , we can show that a scalar multiplication of the weight matrix does not change anything except for the multiplication of over all expression of the holevo bound .
thus , we can parametrize the @xmath5 weight matrix by two real parameters . for our purpose ,
we employ the following representation up to an arbitrary multiplication factor : w = u _ ( cc1 & w w_2 + w w_2 & w_2 ^ 2 ) u_^ * , where @xmath225 and @xmath226 are imposed from the positivity condition and the real orthogonal matrix @xmath227 is defined in terms of eq . by u_= ( cc_,1 & -_,2 + _ , 2 & _ , 1 ) .
the assumption of the model under consideration then yields @xmath228&=| { \langle\v\ell_\theta^1|f_\theta\v\ell_\theta^2\rangle}|\sqrt{\det w } -{\mathrm{tr}\left\{w(g_\theta^{-1}-\re\tilde{g}_\theta^{-1})\right\ } } \\
\nonumber = & \big [ |{\langle\v\ell_\theta^1|f_\theta\v\ell_\theta^2\rangle}|\sqrt{1-w^2}-{\mathrm{tr}\left\{g_\theta^{-1}-\re\tilde{g}_\theta^{-1}\right\}}\,w_2 \big]w_2 . \end{aligned}\ ] ] therefore , by defining a constant solely calculated from the given model @xmath229 , we obtain the sets @xmath230 as follows .
@xmath231 where the common conditions @xmath232 , and @xmath233 also need to be imposed to satisfy the positivity of the weight matrix .
so far we only consider models which consist of mixed states .
it is known that collective measurements do not improve the mse for pure - state models @xcite .
in other words , the holevo bound is same as the bound achieved by separable measurements as far as pure - state models are concerned . in this subsection
, we examine the pure - state limit for our general result .
when a mixed - state model is asymptotically classical , the holevo bound is identical to the sld cr bound .
this should be true in the pure - state limit , and this agrees with the result of matsumoto @xcite . when a model is d - invariant , on the other hand , it is shown that the rld cr bound can be achieved .
this also holds in the pure - state limit and we examine the pure - state limit for a generic mixed - state model below .
we first note that we can not take the limit for sld and rld operators directly @xcite .
this is because there are the terms @xmath234 appearing in the denominators .
however , the sld and rld dual operators are well defined even in the pure - state limit . by direct calculation
, we can show that the sld and rld dual vectors ( [ sldcot ] , [ rldcot ] ) are written as @xmath235 for any two - parameter qubit model .
thus , as long as @xmath236 , they converge in the pure - state limit .
this condition is also expressed as @xmath237 , and hence is equivalent to @xmath238 in the pure - state limit .
, i.e. , @xmath24 is not asymptotically classical .
second , the same warning applies to the sld and rld fisher information @xcite .
however , the inverse of the sld fisher information matrix is well defined even for the pure - state limit .
this is because the @xmath12 component of the inverse of sld fisher information matrix is @xmath239 and this has the well - defined limit .
the same reasoning can be applied to the inverse of the rld fisher information matrix in the pure - state limit .
last , let us examine the general formula .
it is straightforward to show that the function vanishes in the pure - state limit . in other words ,
the general formula given in theorem [ thm1 ] becomes @xmath78=c^r_\theta[w]$ ] for all weight matrices .
therefore , the holevo bound becomes the rld cr bound in the pure - state limit .
this is , of course , expected because any pure - state model can be expressed as a unitary model , which is d - invariant from lemma [ lem8 ] .
in this section , we consider several examples for two - parameter qubit models to illustrate our result .
the first one is a d - invariant model whose holevo bound is identical to the rld cr bound .
the second one is a asymptotically classical model which gives the sld cr bound . as the last example
, we analyze a generic model in particular the behavior of the holevo bound depending on the weight matrix . within the setting of pointwise estimation
, we note that it is sufficient to specify the following three vectors to define a model locally ; [ localmodel ] \ { , _ 1,_2 } .
equivalently , the set @xmath240 or @xmath241 can be used to define the model at @xmath25 uniquely .
one - to - one correspondences among these three specifications of the model are easily established by the bloch vector representation discussed in sec .
[ sec3 - 1 ] . the d - invariant condition is now expressed in terms of the vectors of the set as |_1=|_2=0 .
the asymptotically classical condition is equivalent to |_1_2=0 .
all models other than satisfying the above two conditions are generic ones .
consider a simple unitary model where the parametric state is generated by a two - parameter unitary @xmath242 ( @xmath243 ) : _
= u()_0u()^*. it is easy to show that the norm of the bloch vector is independent of @xmath25 , and hence , this model is globally d - invariant . the holevo bound is @xmath189=c_\theta^r[w]$ ] for all weight matrices @xmath4 .
we note that this kind of d - invariant models possesses symmetry and has been studied by many authors , see for example ch .
4 of holevo s book @xcite and hayashi @xcite and references therein . consider the following model : _ = \{=f^1()_1+f^2()_2 | } , where @xmath244 are unit vectors , which are not necessarily orthogonal to each other , and @xmath245 are scalar ( differentialble ) functions of @xmath246 . the parameter region @xmath23 is specified by an arbitrary open subset of the set ; @xmath247 .
we can show that this model is asymptotically classical because of @xmath248 , and the holevo bound is @xmath189={\mathrm{tr}\left\{wg_\theta^{-1}\right\}}\forall w>0 $ ] . for the asymptotically classical case
, we can easily compare this result with the bound achieved by an optimal estimator comprised of separable povms . in this case , the nagaoka bound is known to be achievable that is calculated as @xcite c_^n[w]=\{wg_^-1}+2 .
therefore , the gain due to use of collective povms is @xmath249 . as a special case ,
let us set @xmath250 to be orthogonal normal vectors and @xmath251 .
then , the inverse of the sld fisher information matrix reads g_^-1= ( cc1-(^1)^2 & -^1 ^ 2 + -^1 ^ 2 & 1-(^2)^2 ) , and the gain mentioned above is @xmath252 .
thus , the role of collective povms becomes important as the state becomes more mixed . in this subsection
, we analyze a generic model other than the previous two examples . in this case
the structure of the holevo bound changes when the weight matrix @xmath4 varies , that is it takes the same form as the rld cr bound for certain choices of @xmath4 whereas it becomes different form for other cases .
this is explicitly demonstrated in the section [ sec4 - 2 ] and here we consider a simple yet nontrivial example to gain a deeper insight into this a phenomenon of this transition .
consider a two - parameter qubit model : [ gicmodel ] _
= \{=(^1,^2,_0)^t | } , where @xmath253 is a fixed parameter which satisfies @xmath254 .
the parameter region is specified by a subset of @xmath255 .
this model is neither d - invariant nor asymptotically classical when @xmath256 and @xmath257 are satisfied .
this model becomes asymptotically classical for @xmath258 as discussed before .
we can also regard this model as a sub - model embedded in the three - parameter qubit model : @xmath259 . the inverse of sld and rld fisher information matrices at @xmath25 are @xmath260 with @xmath261^{1/2}$ ] , respectively , and the three bounds appearing in theorem [ thm1 ] read @xmath262&={\mathrm{tr}\left\{w\right\}}- \frac{1}{1-\theta_0 ^ 2 } ( \vec{\theta}|w\vec{\theta}),\\
\nonumber c_\theta^r[w]&= \frac{1-s_\theta^2}{1-\theta_0 ^ 2 } { \mathrm{tr}\left\{w\right\ } } + 2\sqrt{\frac{1-s_\theta^2}{1-\theta_0 ^ 2}}\ , |\theta_0| \sqrt{\det w},\\ \nonumber c_\theta^z[w]&= { \mathrm{tr}\left\{w\right\}}- \frac{1}{1-\theta_0 ^ 2 } ( \vec{\theta}|w\vec{\theta})+2\sqrt{\frac{1-s_\theta^2}{1-\theta_0 ^ 2}}\ , |\theta_0| \sqrt{\det w } , \end{aligned}\ ] ] where @xmath263 is introduced for convenience . in the following
let us analyze the structure of the holevo bound using the following parametrization of a weight matrix : @xmath264 where we normalize the trace of @xmath4 to be one and @xmath265 , @xmath266 .
this parametrization is different from the one analyzed in sec .
[ sec4 - 2 ] , yet is convenient for the purpose of visualization .
it is easy to see that the effect of the matrix @xmath267 is to mix two parameters @xmath268 and @xmath269 by rotating them about an angle @xmath270 . since @xmath271 for this particular parametrization
, we see that the rld cr bound is independent of the weight parameter @xmath270 .
the other bounds depend on two parameters @xmath272 .
we are interested in how the holevo bound @xmath273 $ ] changes when we vary the weight parameters @xmath274 . upon plotting , we fix the direction of the estimation parameter @xmath275 as @xmath276 with @xmath277 .
we plot the holevo bound for two sets of the state parameters ; ( a ) @xmath278 and ( b ) @xmath279 for illustration .
figure [ fig1a ] shows the holevo bound for the state parameter ( a ) as a function of the weight - matrix parameter @xmath274 . in this plot
, the gray areas indicate the case for @xmath78=c^r_\theta[w]+s_\theta[w]$ ] , whereas the white - meshed region indicates the case for @xmath78=c^r_\theta[w]$ ] .
we also show the other state parameter setting ( b ) in fig .
[ fig1b ] with the same convention . in both figures
we observe smooth transitions between two - different expressions discussed in sec .
[ sec4 - 2 ] . from these figures
, we see that the holevo bound coincides with the rld cr bound for relatively large choice of the weight matrix in fig .
[ fig1a ] , where as the opposite observation holds for fig .
[ fig1b ] . to gain a deeper insight ,
let us calculate the value of quantities @xmath280 ( @xmath281 ) .
then , we get @xmath282 for the case ( a ) and @xmath283 for the case ( b ) . since vanishing of these quantities is equivalent to the d - invariance of a model ( lemma [ lem8 ] ) , we naively expect that the smaller values of them imply that a model behaves more d - invariant - like . indeed , the examples presented here agree with our intuition , yet more detail analysis is needed to make any conclusion .
we briefly discuss another important observation of this paper . a generic model , other than special cases discussed before , exhibits a transition in the structure of the holevo bound for a fixed weight matrix when we change the estimation parameter @xmath25 .
a rough sketch of this argument is that a change in the weight matrix is amount to that in the parameter and vice versa .
this is a well - known fact in the pointwise estimation setting @xcite . below we
briefly show such an example .
the model is same as the generic model analyzed the previous subsection .
as noted before , a change in the weight - matrix parameter @xmath284 is equivalent to rotate the parameter @xmath285 by the angle @xmath270 . depending upon the choice of the weight matrix
, we can also observe a similar transition when we change the parameter @xmath25 in the set @xmath23 .
we set the weight matrix to be a diagonal matrix @xmath286 and @xmath287 .
figure [ fig2 ] plots the holevo bound as a function of the state parameter @xmath246 . the gray area indicates the region where @xmath288=c^r_\theta[w_0]+s_\theta[w_0]$ ] holds , whereas the white - meshed region indicates the case for @xmath288=c^r_\theta[w_0]$ ] .
this shows that the holevo bound coincides with the rld cr bound for a certain subset of the parameter space . and @xmath287 as a function of the state parameter @xmath289 .
the gray area indicates the case for @xmath78=c^r_\theta[w]+s_\theta[w]$ ] , whereas the white - meshed region indicates the case for @xmath78=c^r_\theta[w]$].,width=207 ]
in this paper , we have derived a closed formula for the fundamental precision bound , the holevo bound , for any two - parameter qubit estimation problem in the pointwise estimation setting .
this bound is known to be asymptotically achievable by the optimal sequence of estimators consisting of jointly performed measurements under some regularity conditions . since there exist explicit formulas for the holevo bound for the pure - state qubit model ,
qubit mixed - state models with one and three parameters , our result completes a list of the fundamental precision bounds in terms of quantum versions of fisher information , which is calculated from a given quantum parametric model , as far as qubit models are concerned .
the obtained formula shows several new insights into the property of the holevo bound for quantum parameter estimation problems . in the following
we shall list concluding remarks together with outlook for future works .
first , the necessary and sufficient conditions for the asymptotic achievability of the sld and rld cr bounds are derived when estimating any two - parameter family of qubit states . in particular , the notation of asymptotically classical models is proposed , in which all sld operators commute with each other on the trace of a given parametric state . in this case , the weight matrix can be eliminated and the problem becomes similar to the classical statistics in the asymptotic limit , where the sld fisher information plays the same role as the fisher information .
we note that the notion of asymptotic classicality can be extended to any models on any finite - dimensional system and the same statement obtained for the qubit case holds : the holevo bound coincides with the sld cr bound if and only if the model is asymptotically classical .
the detail of this result shall be given in the subsequent paper .
second , the rld cr bound is shown to be achieved for a certain choice of the weight matrix even though the model is not d - invariant .
this result emphasizes the importance of the rld fisher information for general qubit parameter estimation problems .
since the imaginary parts of the inverse of the rld fisher information matrix and the @xmath7 matrix are always identical for two - parameter qubit mixed - state models , we can not immediately conclude that this is the general statement or not .
it might well be that the only real part of the inverse of rld fisher information matrix plays an important role in higher dimensional systems .
this deserves further studies and shall be analyzed as an extension of the present work .
third , our result also provides the ( unique ) minimizer to the optimization problem appearing in the definition of the holevo bound .
this set of observables , which are locally unbiased in the sense of eq .
, can be used to construct an optimal sequence of povms that attains the holevo bound asymptotically .
this line of approach has been proposed by hayashi , who reported a theorem without proof to realize the construction of povms on the tensor product states @xcite .
his approach is different from other approaches given in refs .
@xcite and it may need more refined arguments to complete his theorem .
last , smooth transitions in the structure of the holevo bound is shown to occur in general .
since this happens in the simplest quantum system , we expect similar phenomena occur in higher dimensional systems as well . however , we do not know whether or not the number of different forms is always two as demonstrated here .
the techniques used in this paper can be applied to two - parameter estimation problems in higher dimensional systems , and we shall make progress in due course .
the author is indebted to prof .
h. nagaoka for invaluable discussions and suggestions to improve the manuscript .
in particular , theorem [ thmdinv ] and lemma [ lem4 ] were motivated by his seminar @xcite .
he also thanks b .- g .
englert and c. miniatura for their kind hospitality at the centre for quantum technologies , singapore , where a part of this work was done .
proof follows from a straightforward calculation .
we substitute @xmath290 into the second line of eqs . .
the first term reads @xmath291 where the relation @xmath292 for @xmath281 and eq .
are used . note @xmath293 for all @xmath294 ,
then the second term is calculated as @xmath295 where @xmath296 and @xmath297 are used . combining the above expressions
, we get the expression of this lemma .
@xmath179 \1 .
@xmath298 : + for convenience , let us define @xmath299 which are related by @xmath300 .
the standard vector analysis shows @xmath301 from the expression , the components of the sld fisher information matrix is expressed as @xmath302 .
then , the determinant of the sld fisher information matrix is calculated as @xmath303 and hence we obtain @xmath304 .
next , we use the relation for the rld fisher information ; @xmath305 $ ] to calculate the determinant @xmath306 as @xmath307 this proves the relation ; @xmath308 .
@xmath179 \2 .
@xmath309 : + from the definition for the matrix @xmath7 , the imaginary part is expressed as z_=v_^1|f_v_^2 ( cc0 & 1 + -1 & 0 ) , and the straightforward calculation yields \{wz_}=2| v_^1|f_v_^2| . the imaginary part of the rld fisher information matrix is _
=( cc0 & -1 + 1 & 0 ) , and the imaginary part of the inverse is \{_^-1}= ( cc0 & 1 + -1 & 0 ) , where we use @xmath308 .
it is easy to show @xmath310 and thus we obtain the important relationship ; @xmath311 .
this proves the claim . @xmath179
@xmath312-c_\theta^r[w ] \right)/{(1-s_\theta^2)}$ ] : + this can be shown by the following calculations : @xmath313-c_\theta^r[w ] \right ) .
\ \square\end{aligned}\ ] ] since the imaginary parts of the inverse of the rld fisher information matrix and the matrix @xmath7 are always identical for two - parameter qubit models , equivalence between 1 and 2 is the direct consequence of lemma [ lem4 ] .
we next note the following relation : @xmath314 which follows from eq . and lemma [ lem7]-1 .
thus @xmath315 if and only if @xmath160 . to show the statement about global d - invariance
, we only need to show @xmath160 for all @xmath44 if and only if @xmath114 is independent of @xmath25 .
@xmath177 @xmath317 holds , the integration of this condition gives @xmath162 is independent of @xmath246 .
conversely , @xmath162 does not depend on @xmath25 , we obtain @xmath318 ( @xmath281 ) for all @xmath44 .
@xmath179 when @xmath319 , the function to be minimized is @xmath320 .
since @xmath167 is positive - definite , the minimum is @xmath321 and is attained by @xmath322 . for the other case ( @xmath323 ) , introducing the new variables through @xmath324 we can express the function @xmath325 as f()=(_1 ^ 2+_2 ^ 2)+2|_1+c| , where @xmath326 is a positive constant .
the minimum of this simple quadratic function is obtained by analyzing the case @xmath327 and @xmath328 separately .
the result is _ ^2f()=
-+2cc + --2cc- , and the unique minimizer @xmath329 is _ * = -c + -c |c| < + c- .
this solution can be translated into the original variables @xmath330 to prove the lemma . @xmath179
in this appendix , we summarize the concept of canonical projection and invariant subspace for an inner product space on real numbers . these results can be generalized to more general settings to be applied to the quantum estimation problem .
consider an @xmath0-dimensional vector space @xmath331 on @xmath332 and an inner product on it @xmath333 .
let @xmath334 ( @xmath335 ) be a set of linearly independent vectors and define the subspace @xmath336 spanned by these vectors .
we define a real positive semi - definite matrix @xmath337_{i , j\in\{1,2,\dots , k\}}$ ] and its inverse by @xmath338_{i , j}$ ] .
a set of vectors @xmath339 forms the dual basis of @xmath340 . given a vector @xmath341 , the canonical projection of @xmath342 onto the subspace
@xmath4 is a map @xmath343 such that _ w(v)=_i=1^k , w^iw_i w. this canonical projection is unique and it preserves the inner product as _
w(v),w=,w ww , vv .
consider any element @xmath341 and the condition @xmath344 is equivalently expressed as follows .
@xmath345 where @xmath346 is the ( unique ) orthogonal complement of @xmath4 . next ,
consider a linear map @xmath167 from @xmath331 to itself .
the subspace is said invariant under the map @xmath167 , if the image of @xmath4 is a subset of @xmath4 , i.e. , @xmath347 holds for @xmath348 . using the above equivalence ,
this can be written as @xmath349 [ lem4 ] let @xmath9 ( @xmath93 ) be the sld ( rld ) dual operator and @xmath1 ( @xmath2 ) be the sld ( rld ) fisher information matrix , respectively .
define a @xmath56 hermite matrix by @xmath94_{i , j\in\{1,\dots , k\}}$ ] , then , the following conditions are equivalent . 1 .
@xmath24 is d - invariant at @xmath25 .
2 . @xmath350 , @xmath351 .
3 . @xmath95 4 .
@xmath352 , @xmath351 .
@xmath353 , @xmath354 with respect to @xmath355 @xmath201 @xmath354 with respect to @xmath356 .
we prove this lemma by the chain ; @xmath359 .
suppose that a given model is d - invariant , this is equivalent to say that the action of the commutation operator on the sld dual operators is expressed as ( l_^i)=_jc^jil_,j , with some real coefficients @xmath360 .
these coefficients are expressed as c^ji=_^i,(l_^j ) _ _ , which directly from the orthogonality condition . using the relation ,
the right hand side is also expressed as @xmath361 , and if the model is d - invariant at @xmath25 , [ app3 ] ( l_^i)=_j ( z_)^jil_,j . hence we show @xmath362 . next , if the condition holds , the sld inner product between @xmath363 and the rld dual operator @xmath364 gives _
^i,(l_^j)__=_k c^kj _ ^i , l_^k__=c^ji .
the left hand side is also calculated from eq . as @xmath365 .
therefore , we show [ app4 ] c^ji = z_^ij=-(_^ij - g_^ij ) _ ^-1=z _ , holds , if the condition holds , that is , @xmath366 .
consider an arbitrary linear operator @xmath68 and assume the condition . in this case , the rld inner product between @xmath68 and @xmath93 is calculated as @xmath367 where eq .
, and the several equations presented in sec . [ sec2 - 3 - 2 ] are used .
since @xmath368 is arbitrary , it implies [ app5 ] l_^i=_^ii\{1,2, ,k}. therefore , we show @xmath369 .
next , let us assume the condition and we show that this implies the condition @xmath370 , that is , @xmath353 , @xmath354 with respect to @xmath355 @xmath201 @xmath354 with respect to @xmath356 .
this is because @xmath371 the remaining to be shown is that the condition , @xmath372 implies the d - invariance of the model .
consider a set of hermite operators @xmath373 and suppose the above condition .
since @xmath374 is equivalent to @xmath375 with @xmath376 the canonical projection on @xmath82 , we can rewrite it as [ app-7 ] x^i , i , j ( ^i - l_^i , l_,j _ _
^i - l_^i , l_,j__^+ ) .
the use of eq .
leads @xmath377 @xmath201 @xmath378 .
then , the equivalent condition can be applied to conclude that the subspace @xmath82 is invariant under the action of linear operator @xmath379 , that is , the model is d - invariant . @xmath179
\i ) proof for the rld cr bound case : + the sufficiency ( d - invariant model @xmath380= c_\theta^r[w]$ ] for all @xmath381 . ) follows from proposition [ propdinv ] .
if the holevo bound is identical to the rld cr bound for all weight matrices , then all the rld dual operators must be hermite , i.e. , @xmath382 for all @xmath64 . to see this , we note that @xmath383 implies @xmath384 . by rewriting @xmath385
, we see that @xmath386 ( @xmath64 ) are orthogonal to the complex span of the rld operators , @xmath387 .
it is straightforward to show that @xmath67=z_\theta[\vec{x}-\vec{\tilde{l}}_\theta]+z_\theta[\vec{\tilde{l}}_\theta]$ ] holds for @xmath388 , where @xmath389 is the collection of the rld dual operators . since the matrix @xmath390 $ ] is positive semi - definite , and hence ,
we obtain _ z__^-1 , and the equality if and only if @xmath391 holds .
thus , all @xmath93 need to be in the set @xmath65 .
next , we use the relation between the sld and rld operators , @xmath392 ( see eq . ) , to get ( i+)(_^i)=_j ^jil_,j . then , the conditions @xmath382 for all @xmath116 imply @xmath393 since @xmath2 is positive definite , so as @xmath394 . with these relations ,
the action of the commutation operator on the sld operators are calculated as ( l_,i)=_j , k [ ( \{^-1})^-1]_ji \{^kj}l_,k .
this proves that @xmath395 .
therefore , if @xmath189= c_\theta^r[w]$ ] for all @xmath396 , then the model is d - invariant .
+ ii ) proof for the d - invariant bound : + this equivalence is a direct consequence of the proposition [ propdinv ] ( @xmath95 ) and the property of the canonical projection given in appendix [ sec : app-3 ] . first , let us note [ dequiv1 ] w c_^h[w]=c_^z[w ] _ z_z_. this is because @xmath397 is an element of the set @xmath65 and @xmath398= z_\theta$ ] . if the condition [ lem4]-5 holds , it is easy to show @xmath399 . therefore , we obtain the matrix inequality : @xmath400&=[{\langlex^i , x^j\rangle_{\rho_{\theta}}^{+}}]_{i , j}\\ \nonumber & = [ { \langlex^i - l_\theta^i , x^j - l_\theta^j\rangle_{\rho_{\theta}}^{+}}]_{i , j}+[{\langlel_\theta^i , l_\theta^j\rangle_{\rho_{\theta}}^{+}}]_{i , j } \\ & \ge [ { \langlel_\theta^i , l_\theta^j\rangle_{\rho_{\theta}}^{+}}]_{i , j}=z_\theta , \end{aligned}\ ] ] holds for all @xmath401 because of the semi - definite positivity of the matrix @xmath402_{i , j}$ ] .
conversely , let us assume @xmath403\ge z_\theta$ ] is true .
this is possible if and only if @xmath404 , otherwise we can always find some @xmath401 such that the matrix inequality @xmath67\ge z_\theta$ ] is violated .
thus , lemma [ lem4]-5 and the equivalence prove this theorem .
@xmath179 m. hayashi , in _ selected papers on probability and statistics american mathematical society translations series 2 _ , vol . * 277 * , 95 - 123 , ( amer . math . soc .
( it was originally published in japanese in bulletin of mathematical society of japan , sugaku , vol . *
55 * , no . 4 , 368 - 391 ( 2003 ) . | the main contribution of this paper is to derive an explicit expression for the fundamental precision bound , the holevo bound , for estimating any two - parameter family of qubit mixed - states in terms of quantum versions of fisher information .
the obtained formula depends solely on the symmetric logarithmic derivative ( sld ) , the right logarithmic derivative ( rld ) fisher information , and a given weight matrix .
this result immediately provides necessary and sufficient conditions for the following two important classes of quantum statistical models ; the holevo bound coincides with the sld cramr - rao bound and it does with the rld cramr - rao bound .
one of the important results of this paper is that a general model other than these two special cases exhibits an unexpected property : the structure of the holevo bound changes smoothly when the weight matrix varies . in particular
, it always coincides with the rld cramr - rao bound for a certain choice of the weight matrix .
several examples illustrate these findings . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.