article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
binary systems can be roughly divided into close ( semi - major axes @xmath18 au ) and wide ( semi - major axes @xmath19 au , @xcite ) pairs .
close binary systems have long been recognized as useful tools for studies of stellar properties .
for example , the stellar parameters such as the masses and radii of individual stars are readily determined to high confidence using eclipsing binaries @xcite .
wide binary systems have proven to be a tool for studies of star formation processes , as well as an exceptionally useful tracer of local potential and tidal fields through which they traverse .
specifically , they were used to place the constraints on the nature of halo dark matter @xcite and to explore the dynamical history of the galaxy @xcite .
a further comprehensive list of current applications of wide binaries can be found in @xcite .
close binaries , owing to their relatively short orbital periods and equally short timescales of brightness or spectrum fluctuations , are fairly easy to detect . unambiguous identification of wide binary systems , on the other hand , requires accurate astrometry on much longer timescales , as these systems have orbital periods @xmath20 years .
however , instead of requiring unambiguous identification , large samples of _ candidate _ wide binaries can be selected by simply assuming that pairs of stars with small angular separation are also gravitationally bound @xcite , or by searching for common proper motion pairs @xcite .
the angular separation method is simple to apply , but it also introduces a relatively large number of false candidates due to chance association of nearby pairs .
the contamination by random associations can be reduced by imposing constraints , such as the common proper motion , or by requiring that the stars are at similar distances .
the distances can be inferred through a variety of means , one of which is the use of an appropriate photometric parallax relation .
the photometric parallax relation provides the absolute magnitude of a star given that star s color and metallicity .
there are a number of proposed photometric parallax relations for main sequence stars in the literature that differ in the methodology used to derive them , photometric systems , and the absolute magnitude and metallicity range in which they are applicable .
not all of them are mutually consistent , and most exhibit significant intrinsic scatter of order a half a magnitude or more ( see figure 2 in ( * ? ? ?
* hereafter j08 ) ) . instead of using an existing relation to select wide binaries ,
we propose a novel method that _ simultaneously _ derives the photometric parallax relation _ and _ selects a sample of wide binary candidates .
the method relies on the fact that components of a physical binary have equal distance moduli ( @xmath21 ) and therefore @xmath22 . assuming that both stars are on the main sequence , and the _ shape _ of the adopted photometric parallax relation is correct , the difference in absolute magnitudes @xmath23 calculated from the parallax relation must equal the measured difference of apparent magnitudes , @xmath24 .
the @xmath25 equality for binaries must be valid irrespective of color , and therefore represents a test of the validity of the adopted photometric parallax relation or , alternatively , a way to estimate the parallax relation . in practice
, the distribution of @xmath26 will not be a delta - function both due to instrumental ( finite photometric precision ) and physical effects ( true vs. apparent pairs ) .
however , for _ true _ wide binaries , the distribution of @xmath26 is expected to be narrow , strongly peaked at zero , and the individual @xmath26 values are expected to be uncorrelated with color .
in contrast , the distribution of @xmath26 values for randomly associated stellar pairs ( hereafter random pairs ) should be much broader even when the correct photometric parallax relation is adopted , reflecting the different distances of components of projected binary pairs .
this dichotomy can be used to assign a probability to each candidate , of whether it is a true physical binary or a result of chance projection on the sky .
the paper is organized as follows . in section [ data ] , we give an overview of the sdss imaging data , and describe the selection , completeness and population composition of two initial , independent samples of candidate binaries . in section [ method ] we describe the photometric parallax estimation method , compare the best - fit photometric parallax relations to the j08 relation , and analyze the scatter in predicted absolute magnitudes .
the properties of wide binaries , such as the color and spatial distributions , are analyzed in section [ properties ] .
finally , the results and their implications for future surveys are discussed in section [ discussion ] .
thanks to the quality of its photometry and astrometry , as well as the large sky coverage , the sdss stands out among available optical sky surveys .
the sdss provides homogeneous and deep ( @xmath27 ) photometry in five bandpasses ( @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , @xcite ) accurate to 0.02 mag ( rms scatter ) for unresolved sources not limited by photon statistics @xcite , and with a zeropoint uncertainty of 0.02 mag @xcite . the survey sky coverage of 10,000 deg@xmath33 in the northern galactic cap and 300 deg@xmath33 in the southern galactic cap results in photometric measurements for well over 100 million stars and a similar number of galaxies @xcite
the recent data release 6 @xcite lists photometric data for 287 million unique objects observed in 9583 deg@xmath33 of sky , and can be accessed through the catalog archive server[multiblock footnote omitted ] ( cas ) casjobs[multiblock footnote omitted ] interface .
astrometric positions are accurate to better than @xmath34 per coordinate ( rms ) for sources with @xmath35 @xcite , and the morphological information from the images allows reliable star - galaxy separation to @xmath36 @xcite
. the five - band sdss photometry can be used for very detailed source classification , e.g. , separation of quasars and stars @xcite , spectral classification of stars to within one to two spectral subtypes @xcite , identification of horizontal - branch and rr lyrae stars @xcite , and low - metallicity g and k giants @xcite .
proper motion data exist for sdss sources matched to the usno - b1.0 catalog @xcite .
we take proper motion measurements from the @xcite catalog based on astrometric measurements from the sdss and palomar observatory sky surveys ( poss - i ; @xcite , poss - ii ; @xcite ) . despite the sizable random and systematic astrometric errors in the schmidt surveys , the combination of a long baseline ( 50 years for the poss - i survey ) and a recalibration of the photographic data using positions of sdss galaxies , results in median random errors for proper motions of only 3 mas yr@xmath37 for @xmath38 ( per coordinate ) , with substantially smaller systematic errors @xcite . following a recommendation by munn et al . , when using their catalog we select sdss stars with only one usno - b match within @xmath39 , and require proper motion rms fit residuals to be less than 350 mas in both coordinates .
we note that the proper motion measurements publicly available as a part of sdss data release 6 are known to have significant systematic errors ( munn et al .
, in prep . ) .
here we use a revised set of proper motion measurements which will become publicly available as a part of sdss data release 7 . for objects in the sdss catalog ,
the photometric pipeline @xcite sets a number of flags that indicate the status of each object , warn of possible problems with the image itself , and warn of possible problems in the measurement of various quantities associated with the object .
these flags can be used to remove duplicate detections ( in software ) of the same object , and to select samples of unresolved sources with good photometry . according to the sdss catalog archive server `` algorithms ''
webpage[multiblock footnote omitted ] , duplicate detections of the same objects can be removed by considering only those which have the `` status '' flag set to primary .
we consider only primary objects , and select those with good photometry by requiring that the binned1 flag is set to 1 , and psf_flux_interp , deblend_nopeak , interp_center , bad_counts_error , notchecked , noprofile , peakcenter , and edge image processing flags are set to 0 in the @xmath40 bands .
the moving unresolved sources , such as asteroids , are avoided by selecting sources with the deblended_as_moving flag set to 0 .
good photometric accuracy ( mean psf magnitude errors @xmath41 mag , see figure 1 in @xcite ) is obtained by selecting sources with @xmath42 , where @xmath43 is the @xmath30 band psf magnitude uncorrected for ism extinction .
the psf magnitudes corrected for ism extinction ( using maps by @xcite ) , and used throughout this work , are noted as @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 . to create the initial sample of resolved stellar pairs , we query ] the cas `` neighbors '' table ( lists all sdss pairs within @xmath44 ) for pairs of sources that pass the above criteria , and that have @xmath45>0\label{teff},\ ] ] where the subscript 1 is hereafter assigned to the brighter component . with this condition
we require that the component with bluer @xmath46 color is brighter in the @xmath30 band .
about @xmath47 of random pairs are rejected with this condition .
we estimate that about @xmath48 of true binary systems might be excluded by this cut ( due to uncertainties in the @xmath46 color caused by photometric errors ) , but their exclusion does not significantly influence our results .
we select @xmath49 million pairs for the initial sample of resolved stellar pairs , and plot the observed distribution of angular separation @xmath50 , @xmath51 , in figure [ ang_sep ] ( _ top _ ) . for a uniform ( random ) distribution of stars , the number of neighboring stars within an annulus @xmath52 increases linearly with @xmath50 , and therefore , the number of random pairs also increases with @xmath50 . to find the number of random pairs as a function of @xmath50
, we fit @xmath53 to the @xmath54 histogram ( in the @xmath55 region ) , and find @xmath56 arcsec@xmath37 . for large separation angles ( @xmath55 ) the two distributions closely
match , indicating that the majority of observed pairs are simply random associations , and are not physically related . at separation angles smaller than @xmath57 the frequency of observed pairs shows an excess , suggesting the presence of true , gravitationally bound systems .
however , even at small separation angles , the selected pairs include a non - negligible fraction of random pairs and require further refinement , or careful statistical accounting for random contamination . throughout this work we use samples of random pairs ( random samples , hereafter ) to account for random contamination in candidate binaries .
we define the random sample as a sample of pairs with @xmath58 taken from the initial pool of stellar pairs .
since pairs in the random sample pass the same data quality selection as candidate binaries , and since virtually all of them are chance associations ( @xmath59 ; see section [ samples ] and figure [ ang_sep ] ) , the random sample is a fair representation of the population of randomly associated stars in candidate binary samples .
the excess of pairs with @xmath60 in figure [ ang_sep ] ( _ top _ ) likely indicates a presence of true binaries , and the angular separation provides a simple , _ geometric _ criterion to select candidate binary systems .
this excess , shown as the ratio @xmath61 in figure [ ang_sep ] ( _ bottom _ ) , increases for @xmath60 , reaches a relatively flat peak of @xmath62 for @xmath63 , and sharply decreases for @xmath64 due to finite seeing and inability to resolve close pairs of sources .
this excess is related to the fraction of true binaries , @xmath65 , as @xmath66 using figure [ ang_sep ] ( _ bottom _ ) , we choose @xmath63 for our geometric selection criterion , since the fraction of true binaries is expected to reach a maximum of @xmath67 in this range .
the interpretation of the excess of close stellar pairs as gravitationally bound binary pairs implies that the components are at similar distances . if this is true , and if it is possible to constrain the distance via a photometric parallax relation , than their distribution in the color - magnitude diagram should be different than for a sample of randomly associated stars .
to test this hypothesis , we select 51,753 candidate binaries with @xmath63 .
we compare their distribution in the @xmath68 vs. @xmath69 diagram to the distribution of pairs from the random sample , as shown in figure [ counts ] .
the number of pairs in this random sample is restricted to 51,753 . were the selection a random process
, the selected candidates would have the same distribution in this diagram as the random sample , and the average candidate - to - random ratio would be @xmath70 .
however , in the region where @xmath71 and @xmath72 the two distributions are different ( average candidate - to - random ratio of @xmath73 ) , implying that @xmath74 of candidates are found at similar distances . in principle
, a selection cut using equations [ color_cut1 ] and [ color_cut2 ] could be made to increase the fraction of true binaries in the candidate sample .
we do not make such a cut a priori , but instead develop a method ( described in section [ method ] ) that robustly `` ignores '' random pairs while estimating the photometric parallax relation .
after a best - fit photometric parallax relation is obtained , the contamination can be minimized by selecting only pairs where both components are at similar distances , as described in section [ samples ] .
the @xmath30 vs. @xmath46 distributions of brighter and fainter components of candidate binaries are shown in figure [ r_vs_gi ] .
we find that the brighter components in the candidate sample are mostly disk g to m dwarfs , while the fainter components are mostly m dwarfs .
as seen from figure [ ang_sep ] ( _ top _ ) , candidate binaries with @xmath75 can not be efficiently selected using angular distance only , as nearly all pairs in this range are most likely chance associations . in this regime ,
a _ kinematic _
selection based on common proper motion should be more efficient , as random pairs have a small probability ( @xmath76 determined using monte carlo simulations ) to be common proper motion pairs ( using selection criteria listed below ) .
we therefore select a second sample of 14,148 candidate binaries by searching for common proper motion pairs with proper motion difference @xmath77 mas @xmath78 , and with absolute proper motion in the range 15 mas @xmath79 400 mas @xmath78 , where @xmath80 .
these criteria require that the directions of two proper motion vectors agree at a @xmath81 level , and that the proper motion is detected at a @xmath82 level or higher .
the common proper motion pairs with orbital motion @xmath83 over 50 years are not selected because their usno - b and sdss positions place them outside the @xmath39 search radius used by munn et al .
the angular separation of common proper motion pairs is limited to @xmath84 .
pairs of sources with @xmath85 are usually blended in the usno - b data and may not have reliable proper motion measurements ( see section [ limitations ] ) , while the maximum angular separation between sources in the cas `` neighbors '' table defines the upper limit of @xmath86 . however , for purposes of section [ limitations ] , we have created a sample of common proper motion pairs that extends to @xmath87 .
we have done so by matching sdss sources ( that pass the quality flags from section [ initial_select ] ) within a @xmath88 search radius into common proper motion pairs .
since this matching is computationally expensive , we have done this only for one sample .
the @xmath30 vs. @xmath46 distributions of brighter and fainter components of kinematically - selected candidate binaries are similar to those shown in figure [ r_vs_gi ] . before proceeding with the determination of photometric parallax relations and discussion of the properties of wide binary systems
, we summarize the completeness of geometric and kinematic samples , and estimate their expected fractions of disk and halo populations .
the samples are selected from a highly - dimensional space of measured parameters and an understanding of the selection effects is a prerequisite for determining the limitations of various derived statistical properties .
for example , the geometric sample is selected using five parameters : the @xmath46 color of the two components , @xmath89 and @xmath90 , their apparent magnitudes , @xmath91 and @xmath92 , and their angular separation on the sky , @xmath50 .
the latter three can be transformed with the aid of a photometric parallax relation into a difference of their apparent magnitudes , @xmath93 , distance @xmath94 , and the projected physical separation , @xmath8 .
we seek to constrain the photometric parallax relation by minimizing the difference @xmath95 , where @xmath96 is a two - dimensional function of @xmath89 and @xmath90 ( section [ method ] ) , and at the same time derive constraints on the two - dimensional color distribution of wide binaries , on their @xmath8 distribution , and on any variation of these distributions with position in the galaxy ( section [ properties ] ) .
not all of these constraints can be derived independently of each other , and most are subject to severe selection effects . by judiciously selecting data subsets and projections of this five - dimensional parameter space ,
these effects can be understood and controled , as described below .
to illustrate the most important selection effects , we employ the photometric parallax relation and its dependence on metallicity derived by ( * ? ? ?
* hereafter i08a ) .
the quantitative differences between their photometric parallax relation and the ones derived here have negligible impact on the conclusions derived in this section . for simplicity
, we select a sample of @xmath972.8 million stars with @xmath98 observed towards the north galactic pole ( @xmath99 ) , and study their counts as a function of distance and the @xmath46 color . due to this choice of field position ,
the distance to each star is approximately equal to its distance from the galactic plane ( for a detailed study of the dependence of stellar number density on position within the milky way , see j08 ) .
figure [ fig : dmvscolor ] illustrates several important selection effects .
first , for any @xmath46 color there is a minimum and maximum distance corresponding to the sdss saturation limit at @xmath100 and the adopted faint limit at @xmath101 ; the probed distance range extends from 100 pc to 25 kpc . within
the distance limits appropriate for a given color , the sample is essentially complete ( @xmath9798% , @xcite ) .
second , these limits are strongly dependent on color : the bluest stars saturate at a distance of about 1 kpc , while the reddest stars are too faint to be detected even at a few hundred pc .
equivalently , due to the finite dynamic range of sdss apparent magnitudes , there is no distance range where the entire color range from the blue disk turn - off edge to the red edge of luminosity function is completely covered .
at best , at distances of about 1 kpc the color completeness extends from the blue edge to the peak of luminosity function at @xmath102 .
third , when pairing stars into candidate binary systems , their color distribution at a given distance ( the requirement that the differences of apparent and absolute magnitudes are similar places the two stars from a candidate pair into a narrow horizontal strip in the distance modulus ( dm ) vs. @xmath46 diagram shown in figure [ fig : dmvscolor ] ) will be clipped : the ratio of the number of candidate binaries and the number of all single stars in the sample decreases at distances significantly different from @xmath70 kpc because of a bias against blue - red pairs .
the binary samples selected from the @xmath971 kpc distance range can be used to measure the two - dimensional color distribution of wide binaries , as well as to gauge the dependence of their @xmath8 distribution on color .
the dependence of the @xmath8 distribution on distance from the galactic plane can also be studied over a substantial distance range , but _ only under the assumption that it is independent of color . _
the imposed @xmath50 range ( @xmath103 to @xmath44 ) limits the range of probed physical separation to values proportional to distance , and ranging from 3,000 au to 30,000 au at a distance of 1 kpc .
we discuss and account for these effects in more detail in section [ spatial ] .
the counts of main - sequence stars shown in figure [ fig : dmvscolor ] include both disk and halo populations . with the available data , there are three methods that might be used for separating stars ( including candidate binary systems ) into disk and halo populations ( juri et al . , in prep . ) : 1 . a statistical method based on the stellar number density profiles ( j08 ) : beyond about 3 kpc from the plane , halo stars begin to dominate
. however , as shown in figure [ fig : dmvscolor ] , only stars bluer than @xmath104 are detected at such distances .
the stellar number density profiles suggest that the fraction of halo stars is below @xmath9720% closer than 1.5 kpc from the galactic plane ( see figure 6 in i08a ) .
2 . classification based on metallicity into low - metallicity ( @xmath105<-1 $ ] ) halo stars and higher metallicity stars . as shown by i08a , this is a robust and accurate method even when using photometric metallicity estimator , but it works only for stars with @xmath106 due to the limitations of the photometric metallicity method , and the sdss spectroscopic metallicity is available only for a small fraction of stars in the candidate samples .
kinematic selection based on proper motion measurements , and implemented via a reduced proper motion diagram ( e.g. , @xcite ; @xcite , and references therein ) . however , as discussed in detail in appendix b , this method is robust only closer than 2 - 3 kpc from the galactic plane due to a rotational velocity gradient of disk stars which diminishes kinematic differences between halo and disk stars further away from the plane . given the limitations of these methods , it is not possible to reliably separate disk and halo populations throughout the explored parameter space , and in both geometric and kinematic samples . for geometric sample ,
the third method is not applicable because sdss - poss proper motions are not reliable at small angular distances ( @xmath107 ; see section [ limitations ] ) .
the requirement @xmath106 required for the second method results in a subsample with too narrow a color range to constrain the photometric parallax relation .
nevertheless , the analysis of this subsample based on results from i08a indicates that fewer than 10% of stars in geometric sample belong to halo population ( this fraction increases with the distance from the galactic plane ; see figure 6 in i08a ) , and thus we expect that halo contamination plays only a minor role in the geometric sample .
the kinematic sample is expected to include a non - negligible fraction of halo stars due to the selection of stars with substantial proper motions .
we use the reduced proper motion diagram to estimate the fraction of halo candidate binary stars in this sample .
the reduced proper motion for an arbitrary photometric bandpass , here @xmath30 , is defined as @xmath108 where @xmath109 is proper motion in arcsec @xmath78 ( sometimes an additional offset of 5 mag is added ) . using a relationship between proper motion , distance and tangential velocity , @xmath110 and @xmath111 equation [ eq : rpm ]
can be rewritten as @xmath112 where @xmath94 is distance in parsec , @xmath113 is the absolute magnitude , and @xmath114 is the heliocentric tangential velocity ( the projection of the heliocentric velocity on the plane of the sky ) , and @xmath115 is a constant ( @xmath116 if @xmath114 is expressed in km @xmath117 ) .
therefore , for a population of stars with the same @xmath114 , the reduced proper motion is a measure of their absolute magnitude . as shown using similar data as discussed here ,
halo and disk stars form two well - defined and separated sequences in the reduced proper motion vs. color diagram ( e.g. , @xcite ; @xcite ; and references therein ) .
we discuss the impact of different metallicity and velocity distributions of halo and disk stars on their reduced proper motion distributions in more detail in appendix b. figure [ fig:1wd ] shows reduced proper motion diagrams for stars observed towards the north galactic pole , constructed for two ranges of observed proper motion : 15 - 50 mas @xmath78 and 50 - 400 mas @xmath78 .
the choice of the proper motion range , together with unavoidable apparent magnitude limits , strongly affects the probed distance range : the larger is the proper motion , the closer is the distance range over which the selection fraction is non - negligible .
we find that the two sequences closely follow the expectations based on the analysis of metallicity and velocity distributions from i08a .
the halo sequence can be efficiently separated by selecting stars with reduced proper motion larger than a boundary generated using the photometric parallax relation from i08a , evaluated for the median halo metallicity ( @xmath105=-1.5 $ ] ) and with @xmath114=180 km @xmath117 ( see equation [ eq : rpm2 ] ) .
this separation method is conceptually identical to the @xmath118 separator discussed by @xcite .
they also proposed to account for a shift of the reduced proper motion sequences with galactic latitude , an effect which we discuss in more detail in appendix b. for the reasons described there , to account for the variation of the reduced proper motion sequences away from the galactic pole , we simply offset the @xmath114 value from 180 km @xmath117 to 110 km @xmath117
( i.e. , the separator moves upwards in figure [ fig:1wd ] by 1 mag ) . while this selection removes some disk binaries ,
it is designed to exclude most of halo binaries from the sample . with the aid of reduced proper motion separator
, we separate kinematic sample into candidate halo ( 1,336 pairs ) and disk binaries ( 10,112 pairs ) .
this fraction of halo systems is consistent with the above estimate obtained for the geometric sample . to assess selection effects
, we first investigate the sample of single stars .
the top left panel in figure [ fig:2wd ] shows the fraction of all the stars shown in figure [ fig:1wd ] that have proper motion larger than 15 mas @xmath78 and @xmath119 ( the latter limit ensures the sdss - poss proper motion catalog completeness above @xmath9790% ) .
the selection efficiency is a strong function of distance , and falls from its maximum of @xmath9795% for nearby stars to below 50% at a distance of about 1 kpc .
the candidate disk stars are detected in significant numbers to @xmath973 kpc , and halo stars beyond @xmath971 kpc .
the fraction of selected stars that are classified as halo stars is below 20% closer than @xmath971.5 kpc from the galactic plane , and becomes essentially 100% beyond 3 kpc .
the kinematic difference between halo and disk stars is blurred at distances beyond 2 - 3 kpc ( see appendix b ) , and the majority of disk stars at such distances are misidentified as halo stars ( the metallicity distribution implies that disk stars do exist at distances as large as 7 kpc from the galactic plane , see figure 10 in i08a ) . to demonstrate this effect
, we use subsamples of candidate disk and halo binaries identified using the reduced proper motion diagram that have @xmath120 . for these pairs
it is possible to estimate photometric metallicity ( i08a ) and use it as an independent population classifier .
figure [ plot_ug ] shows that practically all candidate binaries with @xmath105>-1 $ ] further than @xmath972 kpc from the galactic plane are misclassified as halo stars when using reduced proper motion diagram . in summary , geometric sample is heavily dominated by disk binaries , with halo contamination all but negligible closer than about 2 kpc from the plane .
kinematic sample becomes severely incomplete ( @xmath12150% ) further than @xmath972 kpc from the plane , and has a higher fraction of halo binaries than geometric sample , at a given distance from the plane .
however , this halo contamination can be efficiently removed using the reduced proper motion diagram .
unfortunately , the number of selected halo binaries is insufficient in number ( 1,336 in kinematic and 5,556 in geometric sample ) , and spans too narrow a color range to robustly constrain the photometric parallax relation .
therefore , both samples of candidate binaries are supposed to yield similar photometric parallax relations , because both are dominated by disk stars .
in principle , both the normalization and the shape of the photometric parallax relation ( i.e. , the shape of the main sequence in the hertzsprung - russell diagram ) vary as a function of color and metallicity @xcite . since our data do not allow a reliable estimate of metallicity over the entire range of observed colors , we can only estimate the `` mean '' shape of the photometric parallax relation as a function of color , for all metallicities present in the sample .
such a mean shape is approximately an average of individual @xmath105$]-dependent relations , weighted by the sample metallicity distribution .
j08 derived such `` mean '' photometric parallax relations appropriate at the red end for the nearby , metal - rich stars , and at the blue end for distant , metal - poor stars . i08a
discuss the offset of photometric parallax relation as a function of metallicity ( see their figure 20 ) , and derived the metallicity range implied by `` mean '' photometric parallax relations from j08 .
the derived metallicity range is consistent with the spatial distribution of metallicity derived by i08a and the color - magnitude limits of the sdss survey .
we adopt the j08 polynomial @xmath122 parametrization of the photometric parallax relation @xmath123 where @xmath124 are the parameters we wish to estimate . to improve their accuracy , juri et al .
used a maximum likelihood technique to estimate the @xmath122 color from the observed @xmath125 and @xmath122 colors .
because of the brighter flux limit employed here , we use the measured @xmath46 color to derive a best estimate of the @xmath122 color via a stellar locus relation ( j08 ) : @xmath126 ) + r - i\label{interp}\ ] ] the @xmath122 color estimate obtained with this method has several times smaller noise than the measured @xmath122 color .
this is because the observed dynamic range for the @xmath46 color is much larger than of the @xmath122 color ( @xmath127 mag vs. @xmath70 mag ) , while their measurement errors are similar .
the goal of parameter estimation algorithm is to determine the photometric parallax relation , @xmath128 , that minimizes the width of the distribution of @xmath26 values for _ true _ binary systems , where @xmath129 .
the @xmath130 minimization can not be used for this purpose because random pairs , if not removed from the sample , will strongly bias the best - fit @xmath113 .
the available color , angular separation , and proper motion information are insufficient to separate the random pairs from the true binaries .
therefore , we need to design a fitting algorithm that will be least affected as possible by random pairs . we begin by studying the behavior of @xmath26 values in mock catalogs .
the first step in creating a mock catalog is the selection of 51,753 ( random ) pairs from the random sample .
note that the fraction of true binaries in the random sample is only @xmath131 ( see section [ samples ] ) .
true binaries are then `` created '' in the mock catalog by replacing the observed @xmath92 magnitudes for @xmath132 of pairs with @xmath133 where @xmath134 and @xmath135 ( equation 2 coefficients from j08 ) .
the @xmath136 is a gaussian random variate added to account for the intrinsic scatter around the photometric parallax relation .
the result of this process is a mock sample of candidates where @xmath132 of pairs are `` true '' binaries , and the rest ( @xmath137 ) is the contamination made of random pairs .
the distribution of @xmath26 values for `` true '' binaries is , by definition , a 0.1 mag wide gaussian centered on zero when @xmath134 .
figure [ delta_hist_mock ] ( _ top _ ) shows the distribution of @xmath26 values for the mock sample evaluated with the `` true '' [ @xmath138 photometric parallax relation .
the observed @xmath26 distribution can be described as a sum of a gaussian and a non - gaussian component .
the non - gaussian component is due to random pairs ( the contamination ) , while the gaussian component ( 0.1 mag wide and centered on zero ) is due to the true binaries .
when an @xmath113 relation different from the `` true '' ( or best - fit ) @xmath113 is adopted , the gaussian component becomes wider and _ the peak height of the @xmath26 distribution decreases _ , as shown in figure [ delta_hist_mock ] ( _ bottom _ ) . at the same time , the peak height of the @xmath26 distribution of the contamination changes much less since the distribution is much wider ( @xmath139 mag wide ) . therefore , _ minimizing _
the width of the @xmath26 distribution of true binaries , is equivalent to _ maximizing _ the peak height of the entire @xmath26 distribution .
we quantify this peak height as the number of candidate binaries in the most populous @xmath26 bin . to robustly explore the parameter space that defines the photometric parallax relation , and to find the best - fit coefficients @xmath140
, we implement our algorithm as a markov chain monte carlo ( mcmc ) process . the mcmc description given here and our implementation of the algorithm
are based on examples given by @xcite , @xcite , and @xcite .
the basic idea of the mcmc aproach is to take an _
n_-step intelligent random walk around the parameter space while recording the point in parameter space for each step .
each successive step is allowed to be some small distance in parameter space from the previous position .
a step is always accepted if it improves the fit , and is sometimes accepted on a random basis even if the fit is worse , where the goodness of the fit is quantified by some parameter ( usually with @xmath130 ) .
the random acceptance of a bad fit ensures that the mcmc does not become stuck in a local minimum , and allows the mcmc to fully explore the surrounding parameter space .
we start a monte carlo markov chain by setting all coefficients from equation [ photo_parallax ] to zero ( @xmath141 ) . using this initial set of coefficients
we evaluate @xmath142 for all candidate binaries assuming @xmath143 , and bin @xmath26 values in 0.1 mag wide bins .
the number of candidate binaries in the most populous bin , @xmath144 , is used to quantify the relative goodness of the fit . given @xmath145 , a new candidate step , @xmath146 , is generated , where the step size , @xmath147 , is a vector of independent gaussian random variates with initial widths , @xmath148 , set to 1 . using the candidate set of coefficients , @xmath149 ,
@xmath26 values are evaluated , binned , and the height of the @xmath26 distribution is assigned to parameter @xmath150 . following the metropolis - hastings rule @xcite the candidate step
is accepted ( @xmath151 , @xmath152 ) if @xmath153 or if @xmath154 , where @xmath155 is a random number between 0 and 1 ( @xmath156 $ ] ) .
otherwise , the candidate step is rejected .
while the metropolis - hastings rule guarantees that the chain will converge , it does not specify _ when _ the convergence is achieved .
the speed of the convergence depends on the gaussian scatter @xmath148 used to calculate the step size @xmath147 .
if the scatter is too large , a large fraction of candidate steps is rejected , causing the chain to converge very slowly .
if the scatter is too small , the chain behaves like a random walk , and the number of steps required to traverse some short distance in the parameter space scales as @xmath157 .
the choice of optimal gaussian scatter @xmath148 ( for each fitted coefficient ) , as a function of the position in the parameter space , is not trivial and it can be very complicated even if the fitted coefficients are uncorrelated . to determine the optimal @xmath148 values we follow the @xcite prescription ( see their appendix a ) .
after every 100 accepted steps we compute the coefficient covariance matrix @xmath158 from the chain itself , diagonalize it as @xmath159 , and use it to calculate a new step size @xmath160 for each coefficient separately .
we find that this transformation greatly accelerates the convergence of a chain .
due to the stochastic nature of the mcmc , the best - fit relations ( coefficients with the highest @xmath144 value in a chain ) from different chains will not necessarily be the same . to quantify the intrinsic scatter between different best - fit relations , we run fifty 10,000-element long chains , and select the best - fit coefficients from each chain for subsequent comparison ( see section [ robustness ] ) .
the proper mixing and convergence of chains is confirmed using the gelman & rubin @xmath161 statistic @xcite .
gelman & rubin suggest running the chains until @xmath162 for all fitted coefficients . with 10,000 elements in each chain ,
we obtain @xmath163 for all fitted coefficients . in the end
, we select @xmath124 with the highest @xmath144 value among all chains as our best - fit relation . the constant term @xmath164 is not constrained with our algorithm , because @xmath164 ( from @xmath165 and @xmath166 ) cancel out when evaluating @xmath26 .
instead , we constrain @xmath164 by requiring @xmath167 at @xmath168 , obtained from trigonometric parallaxes of nearby m dwarfs @xcite . to test the robustness of our algorithm , we apply it to the mock sample described in section [ algorithm ] .
the best - fit relations ( obtained from markov chains ) are compared on a @xmath169 grid in 0.01 mag steps .
we find an rms scatter of 0.05 mag between markov chains , and 0.05 mag rms scatter between the true and the best - fit relation with the highest @xmath144 value .
we repeat this test with a mock sample containing @xmath170 of true binaries .
the rms scatter between the best - fit relations decreases to 0.03 mag , and the rms scatter between the true and the best - fit relation with the highest @xmath144 value decreases to 0.01 mag .
even when only @xmath132 of sources are true binaries ( i.e. , contamination by random pairs is @xmath137 ) our algorithm recovers the `` true '' photometric parallax relation at the 0.05 mag ( rms ) level . the accuracy of the fit increases ( to 0.01 mag rms ) as the contamination decreases ( from @xmath137 to @xmath171 ) .
we apply the method described in section [ implementation ] to two samples of candidate binaries and obtain the best - fit photometric parallax relations @xmath172
@xmath173 for the geometrically- and kinematically - selected samples , respectively .
candidate halo binaries were removed from the kinematically - selected sample using reduced proper motion diagrams ( section [ separation ] ) before the equation [ mr_kin ] was derived .
the photometric parallax relations for halo stars can not be robustly constrained using geometrically- or kinematically - selected halo binaries because the color range they span is too narrow ( @xmath174 at 3 - 4 kpc , see figures [ fig : dmvscolor ] and [ fig:2wd ] ) .
we test the correctness of the shape by studying the dependence of median @xmath26 values on the @xmath46 colors of the brighter and the fainter components .
if the _ shape _ of these photometric parallax relations is correct , the distribution of @xmath26 values will be centered on zero , and the individual @xmath26 values will not correlate with color .
the medians are used because they are more robust to outliers ( random pairs in the sample ) .
we start by calculating @xmath26 values for each candidate binary sample ( using the appropriate @xmath113 relation ) , and then select candidates with @xmath175 .
this cut reduces the contamination by random pairs , as demonstrated in section [ scatter ] .
the selected candidate binaries are binned in @xmath46 colors of the brighter and the fainter component , and the median @xmath26 values are shown in figure [ medians ] .
the distributions of the median @xmath26 for each pixel are fairly narrow ( 0.07 mag ) , and centered on zero .
irrespective of color and the choice of the two best - fit photometric parallax relations , the deviations are confined to the 0.25 mag range , placing an upper limit on the errors in the mean shape of the adopted relations . in figure [ mr_ri ]
we compare the adopted photometric parallax relations to j08 `` faint '' @xmath176 and `` bright '' @xmath177 photometric parallax relations .
the rms difference between equations [ mr_geo ] and [ mr_kin ] , and equation [ mr_bright ] is @xmath178 mag , comparable to the rms difference between our equations [ mr_geo ] and [ mr_kin ] ( @xmath178 mag ) .
the maximum difference between equations [ mr_geo ] and [ mr_kin ] , and equation [ mr_bright ] is @xmath179 mag , again comparable to the maximum difference between our equations [ mr_geo ] and [ mr_kin ] ( @xmath179 mag ) .
the different color distributions of the two samples , shown in figure [ geo_kin_gi1_gi2_comp ] , together with metallicity effects , is the most likely explanation for differences between the two photometric parallax relations .
the scatter in @xmath26 values can be expressed as @xmath180 where @xmath181 is the scatter in predicted absolute magnitudes , and @xmath182 is the scatter in measured apparent magnitudes . since the photometric uncertainties of sdss are well understood , the intrinsic scatter around the @xmath2 relation is possible to measure and characterize . in figure
[ delta_hists ] we plot the observed distributions of @xmath26 values for the geometrically- and kinematically - selected binaries , and overplot the @xmath26 distribution of the random sample .
the @xmath26 values for the random sample were calculated with equations [ mr_geo ] and [ mr_kin ] , respectively .
the @xmath26 distribution of the random sample was fitted to the observed @xmath26 distribution in the @xmath183 range using the kolmogorov - smirnov test . by comparing the random and the observed @xmath26 distributions ,
we find that the two match well for @xmath183 ( the kolmogorov - smirnov test reports @xmath184 ) , indicating that candidate binaries with @xmath183 are almost certainly random pairs . on the other hand , as @xmath26 approaches zero , the two distributions become remarkably different ( @xmath185 for @xmath186 ) , indicating that these candidate binaries are dominated by true binary systems , and not by random pairs .
the @xmath26 distribution for true binaries ( figure [ delta_hists ] , dashed line ) , obtained by subtracting the random from the observed @xmath26 distribution , is clearly not gaussian .
it can be modeled as a sum of two gaussian distributions ( `` narrow '' and `` wide '' ) centered close to zero , and about 0.1 mag and 0.55 mag wide .
the centers , widths , and areas for the best - fit gaussian distributions are given in table [ gauss ] .
to determine the consistency of the observed scatter with photometric errors , we normalize the @xmath26 values for the kinematically - selected sample with expected formal errors , @xmath187 and plot the @xmath188 distribution in figure [ delta_norm ] .
the @xmath188 distribution for true binaries is not a gaussian with a width of 1 , as we would expect if the scatter in the @xmath26 distribution was only due to photometric errors in the @xmath40 bands ( note that the expected random error in @xmath113 is about 5 - 10 times larger than the random error of the @xmath46 color because @xmath189 varies from @xmath9710 at the blue edge to @xmath975 at the red edge ) .
the width of @xmath188 distribution for the geometrically - selected candidate binaries is about 3 times smaller than in the kinematically - selected sample .
we find that this is due to _ overestimated _ photometric errors in the geometrically - selected sample , as shown in figure [ photo_errors ] .
the overestimated photometric errors in the @xmath40 bands overestimate the expected formal error @xmath190 , and the overall @xmath188 distribution is too narrow .
we speculate that the small angular separation ( @xmath191 ) between the components is the cause of overestimated photometric errors ( perhaps due to sky background estimates ) .
the small angular separation of components does not affect the magnitudes of stars in the geometrically - selected sample .
if it did , the two @xmath26 distributions would be significantly different which , as shown in figure [ delta_hists ] , is not the case .
the observed non - gaussian scatter in predicted absolute magnitudes may be due to photometric parallax variation as a function of metallicity . as noted at the beginning of section [ method ]
, we can only estimate the `` mean '' shape of the photometric parallax relation .
since the intrinsic photometric parallax for a given wide binary system is different from the mean relation , @xmath96 ( the difference of predicted absolute magnitudes ) and @xmath192 ( the measured difference of apparent magnitudes ) will differ .
this discrepancy will increase for systems where the components have significantly different colors . to test the assumption
that the shape of photometric parallax relation increases the scatter in predicted absolute magnitudes , we use the mock sample constructed in section [ algorithm ] and add a color - dependent offset to apparent magnitudes @xmath193 where @xmath155 is a random number between zero and one ( the same for both components ) .
these color - dependent offsets simulate the change in the shape of the photometric parallax relation due to metallicity .
we apply the algorithm described in section [ implementation ] to the mock sample , and obtain a revised photometric parallax relation . using this relation ,
we analyze the distribution of @xmath26 values and find that it can be modeled as a sum of two gaussians centered on zero , with widths of 0.1 and 0.3 mag .
this result suggests that the non - gaussian scatter observed in candidate samples may be caused by the difference between the shapes of the mean photometric parallax relation and a true relation for a given metallicity ( and perhaps other effects , such as age ) .
this model - based conclusion is consistent with a direct comparison of relations derived here and the relations from i08a evaluated for the median halo metallicity ( @xmath105=-1.5 $ ] ) and the median disk metallicity ( @xmath105=-0.7 $ ] for distances probed by our sample ; see figure 5 in the above paper ) .
the two relations corresponding to halo and disk stars are offset by 0.6 mag due to metallicity difference .
our relations match the low - metallicity relation at the blue end and the high - metallicity relation at the red end .
therefore , in the worst case scenario of extremely blue ( @xmath194 ) and red ( @xmath195 ) disk stars , the maximum error in the difference of their absolute magnitudes is 0.6 mag . when convolved with the observed color distribution of pairs , the expected scatter is about 0.2 - 0.3 mag , consistent with the observed and simulated widths of the @xmath26 distributions .
unresolved binarity of components in candidate samples may also contribute to the non - gaussian scatter in predicted absolute magnitudes .
the multiplicity studies of g dwarfs @xcite and m dwarfs @xcite find that a significant fraction of g and m dwarf stars ( @xmath196 ) are unresolved binary systems .
if a component of a wide binary system is an unresolved binary system , its luminosity will be underestimated ( with the magnitude of the offset depending on the actual composition of the binary ) and the @xmath26 value for the wide binary system will systematically deviate from zero . in appendix [ model ]
we model the presence of unresolved binaries in wide binary systems , and find that the model can explain the observed @xmath26 scatter .
therefore , both the intrinsic variations of the photometric parallax relation and unresolved binaries can explain the observed non - gaussian scatter of @xmath26 .
the data discussed here are insufficient to disentangle these two effects .
finally , the uncertainty in predicted absolute magnitudes ( error distribution for photometric parallax method ) can be obtained by drawing random values , @xmath197 , from a non - gaussian distribution @xmath198 where @xmath199 are gaussian distributions , and the best - fit parameters are listed in table [ gauss ] .
the best - fit photometric parallax relation can be utilized to further refine the samples of candidate binaries and to address questions about their dynamical and physical properties such as * do wide binaries have the same spatial distribution as single stars ? * do wide binaries have the same color distribution as single stars ?
* are the color distributions of components in wide binary systems consistent with random pairings ? * what is the distribution of semi - major axis for wide binaries ? *
does the distribution of semi - major axis vary with the position in the galaxy ? we use the best - fit photometric parallax relations to select samples of candidate binaries with high selection efficiency ( high fraction of true binaries ) by imposing further constraints on @xmath26 values in geometric and kinematic sample . as shown in figure [ delta_hists ] , the fraction of random pairs in the candidate sample is simply @xmath200 , where @xmath201 and @xmath202 are the integrals of the random ( triangles ) and total ( thick solid line ) @xmath26 distributions .
the fraction of true binaries , or the _ selection efficiency _
, is then @xmath203 without a cut on @xmath26 , the fraction of true binaries ( the selection efficiency ) in the geometrically- and kinematically - selected samples is @xmath204 and @xmath205 , respectively .
it is reassuring to find that the @xmath206 value for the geometrically - selected sample obtained here , and the one measured in section [ geo_select ] match so well ( at a @xmath207 level ) , even though the two methods for estimating @xmath206 are independent .
the selection efficiency of @xmath205 for the kinematically - selected sample may seem low , given that only @xmath208 of random pairs pass the common proper motion criteria .
this points to a low fraction of true binaries with angular separation greater than @xmath209 .
if this fraction is about 1/400 ( @xmath210 ) , the common proper motion criteria will select 2 random pairs ( 0.5% out of a 400 ) , and only 1 true binary system .
therefore , @xmath211 of the sample ( 2 out of 3 ) will be random pairs , and @xmath204 ( 1 out of 3 ) will be true binary systems , similar to what we find for the kinematically - selected sample .
the result that only 1/400 pairs with @xmath75 are true binaries puts the fraction of random pairs in the random sample at @xmath59 .
figure [ delta_hists ] shows that the true binaries have a much smaller range of @xmath26 values than the random pairs .
therefore , a cut on @xmath26 would reduce the contamination , and increase the fraction of true binaries in a sample . by requiring @xmath175
, we construct samples where @xmath212 and @xmath213 of candidates are true binaries .
the numbers of candidate binaries in these cleaner samples are 16,575 ( geometrically - selected ) and 5,157 candidates ( kinematically - selected ) , with the expected total number of true binaries about 13,743 .
the sample efficiency for the geometric sample can be further increased to 90% by requiring @xmath214 and @xmath215 kpc , where @xmath13 is the height above the galactic plane .
compared to the existing catalogs of wide binaries by @xcite and @xcite , our samples represent a 20-fold increase in the number of candidate binaries and probe much deeper into the halo ( to @xmath216 kpc ) .
although a non - negligible fraction of candidate pairs are due to random pairings ( @xmath67 ) , the increase in the number of potential physical pairs is substantial .
we emphasize that our method only selects candidates where both components are _ main - sequence _ stars , while rejecting systems where one of the components has evolved off the main sequence .
this is due to the photometric parallax relation , as defined here , being correct for main - sequence stars only .
together with the small expected fraction of giant stars in our sample due to faint apparent magnitudes ( 1 - 2% , @xcite ; i08a ) , this bias results in practically pure main - sequence sample .
we note that the application of a photometric parallax relation that corresponds to some mean metallicity distribution introduces systematic errors in estimated @xmath113 .
we partially mitigate this problem by averaging distances determined for each binary component ( using equation [ mr_geo ] ) .
based on the behavior of photometric parallax relations and @xmath26 distribution discussed in section [ scatter ] , the systematic uncertainty in obtained distances is most likely not larger than 10 - 15% ( an understimate due to faint bias for blue stars ) .
another source of overall systematic uncertainty in distances is the normalization of equation [ mr_geo ] adopted from @xcite .
this normalization corresponds to nearby ( @xmath121100 pc ) metal - rich stars , while most stars in our sample are distances of the order 1 kpc .
the disk metallicity gradient discussed by i08a implies systematic distance overestimate of about 10 - 20% , partially cancelling the above underestimate .
these systematic uncertainties propagate as systematic uncertainties of derived semi - major axes discussed in section [ spatial ] .
the luminosity of a main - sequence star , and thus its color via photometric parallax relation , can be used as a proxy for stellar mass .
the color - color distribution of wide binaries , therefore , provides constraints on the distribution of stellar masses in wide binary systems . to find the color distribution of wide binaries , we select a volume - complete ( @xmath217 ) subsample of geometrically - selected candidate binaries with @xmath175 , and plot their distribution in the @xmath90 vs. @xmath89 color - color diagram in figure [ counts_0.7_1.0 ] ( _ top _ ) .
the sample is complete in the @xmath218 color and 4,200 au @xmath219 10,000 au semi - major axis range ( see section [ spatial ] ) . even though the @xmath175 cut increases the fraction of true binaries , about 14% of candidates ( in the @xmath217 range )
are still random pairs that contaminate the map . to remove the contamination , first we select pairs from the random sample ( see the end of section [ initial_select ] ) with @xmath175 and @xmath220 .
the @xmath175 cut on the random sample will not increase the fraction of true binaries ( @xmath206 ) above @xmath221 because @xmath206 decreases rapidly with @xmath50 ( see figure [ logz_plots ] ( _ middle left _ ) in section [ spatial ] ) , and the pairs in the random sample have @xmath222 .
the @xmath90 vs. @xmath89 distribution of this random sample is shown in figure [ counts_0.7_1.0 ] ( _ middle _ ) .
the maps are essentially probability density maps as pixels sum to 1 .
to correct for the contamination in the top map , we multiply each pixel in the random map with 0.14 ( that being the contamination in the candidate binary sample ) , and subtract two maps .
the corrected map , presented in figure [ counts_0.7_1.0 ] ( _ bottom _ ) , shows that the color - color distribution of true binary systems is fairly uniform , has a local maximum around @xmath223 , and reflects the underlying luminosity function which peaks for red stars ( c.f .
figure [ fig : dmvscolor ] ) . the map shown in figure [ counts_0.7_1.0 ] ( _ bottom _ ) describes the probability density , @xmath224 $ ] , of a wide binary system with components that have @xmath89 and @xmath90 colors falling into a given pixel .
this probability density can be expressed as a product @xmath225 = p[(g - i)_b|(g - i)_a ] \
, p[(g - i)_a ] \label{eq_wide_binary_prob}\ ] ] where @xmath226 $ ] is the conditional probability density of having one component with @xmath227 color in a wide binary system where the other component has @xmath228 , and @xmath229 $ ] is the probability density that a star with @xmath230 color is in a wide binary system .
these probability densities may also vary with galactic coordinates ( e.g. , with the height above the galactic plane ) , but we can not study such effects directly because the samples are volume - complete only in the @xmath217 range . the conditional probability density , @xmath226 $ ] , can be extracted from figure [ counts_0.7_1.0 ] ( _ bottom _ ) map by selecting pixels where either @xmath231 , or @xmath232 .
the resulting @xmath226 $ ] for several values of @xmath228 are shown in figure [ wide_binary_prob ] .
red stars ( @xmath233 ) are more likely to be associated with another red star than with a blue star , while for blue stars the companion color distribution is flat .
the best - fit analytic functions that describe the observed trends are given in table [ tbl_wide_bin_prob ] .
the probability density , @xmath229 $ ] , that a star with @xmath230 is in a wide binary system can be derived by comparing the @xmath46 color distribution of stars in wide binary systems with the @xmath46 color distribution of all the stars in the same volume . as shown in figure [ prob_gi ] ( _ top _ ) , the @xmath46 color distribution of stars in the volume - complete wide binary sample roughly follows the @xmath46 color distribution of all the stars in the same volume .
the ratio of the two distributions ( renormalized to an area of 1 ) gives the @xmath229 $ ] , and is shown in the bottom panel .
the probability for a star to be in a wide binary system ( @xmath229 $ ] ) is independent of its color . given this color , the companions of red components seem to be drawn randomly from the stellar luminosity function , while blue components have a larger blue - to - red companion ratio than expected from luminosity function .
these results are consistent with recent results by @xcite . the overall fraction of stars in wide binary systems is discussed in the next section .
if the semi - major axis distribution function , @xmath234 , is known , the number of stars in wide binary systems can be determined by integrating @xmath234 from some lower cutoff , @xmath235 , to the maximum semi - major axis , @xmath236 .
the power - law frequency distribution , @xmath237 , @xmath238 , is known in the context of wide binaries as the pik distribution ( od ; @xcite ) . when semi - major axis distribution of wide binaries follows the od ,
the frequency distribution of @xmath239 is a straight line with a slope of zero @xcite . alternatively ,
an equivalent representation of od is the cumulative distribution @xmath240\propto \log(a)$ ] . in this form ,
od is a straight line with a positive slope .
we use the cumulative representation , instead of differential , because it reduces the counting noise in sparsely populated bins ( though the errors become correlated between bins ) .
we utilize geometrically - selected candidate binaries ( see section [ limitations ] for a discussion of the kinematic sample ) , but do not limit the selection to @xmath241 , as we did in section [ geo_select ] . since @xmath242 ,
the removal of upper limit on @xmath50 allows us to probe an extended range of semi - major axes .
the downside is that random pairs dominate at large @xmath50 and a careful accounting for contamination as a function of @xmath50 is required before the the @xmath234 distribution can be constrained .
since we only know the projected separation of our pairs , we use a statistical relation to calculate the average semi - major axis , @xmath243 , as @xmath244 , where @xmath245 is the heliocentric distance @xcite .
hereafter , we drop the brackets and simply note the average semi - major axis as @xmath8 . figure [ loga_cum ] ( _ top left _ ) shows the cumulative distribution of @xmath239 for candidate wide binaries with @xmath214 selected from the @xmath246 range . the cumulative distribution does not follow a straight line , as predicted by the od , but actually increases its slope with @xmath239
we assume that this is due to an increasing fraction of random pairs at high @xmath239 , and proceed to verify this assumption .
figure [ loga_cum ] ( _ top right _ ) shows the differential distribution of angular separation for the selected sample .
for @xmath247 the observed and random distributions closely match , demonstrating that random pairs dominate at high @xmath50 ( or high @xmath239 ) . to calculate how the fraction of true binaries ( or random pairs ) changes as a function of @xmath50 , we fit @xmath53 to the observed histogram , and calculate the fraction or true binaries , @xmath206 , using equation [ eff2 ] ( see section [ geo_select ] ) .
the calculated @xmath206 values , as well as the best - fit second - degree polynomial , @xmath65 , are shown in figure [ loga_cum ] ( _ middle left _ ) . as an independent test
, the selection efficiency was calculated using equation [ eff ] ( i.e. , from the @xmath26 distribution ) for three @xmath50-selected subsamples , and the obtained values agree with @xmath65 at a level of 1% . the angular separation for which @xmath206 falls below @xmath248
is defined as @xmath249 .
the fraction of true binaries ( @xmath65 ) also changes as a function of @xmath13 , and is determined separately for different distance bins . since the candidates are restricted in @xmath13 ( @xmath250 kpc to @xmath251 kpc in this example ) and @xmath50 ( @xmath103 to @xmath249 ) , to ensure a uniform selection in the @xmath13 vs. @xmath8 space we define @xmath252 and @xmath253 as the minimum and maximum probed semi - major axis , shown as the selection box in figure [ loga_cum ] ( _ middle right _ ) .
the @xmath254 factor is to account for the fact that the candidates are restricted to high ( @xmath255 ) galactic latitudes . to correct the cumulative distribution of @xmath239
, we assign a probability @xmath65 to each candidate binary in the @xmath256 to @xmath257 range , and add the probabilities ( instead of counting candidates ) when making the cumulative @xmath239 distribution .
the corrected cumulative distribution , shown in figure [ loga_cum ] ( _ bottom left _ ) , follows a straight line up to the turnover point , @xmath12 .
we define @xmath12 as the average semi - major axis for which the straight line fit to the cumulative distribution deviates by more than @xmath258 .
in addition to @xmath12 , we also measure the slope of the cumulative distribution where it follows the straight line .
it can be shown that the slope of the cumulative distribution is equal to the constant of proportionality , @xmath259 , in pik distribution , @xmath260 .
the number of binaries can be calculated by integrating @xmath234 from @xmath235 to @xmath236 , and we obtain @xmath261 . for integration limits
we choose @xmath262 where we assume that systems with semi - major axes greater than @xmath12 are no longer bound , and @xmath263 au ( since @xmath264 , the results are not very sensitive to the choice of @xmath235 ) . the uncertainty in @xmath12 , shape of @xmath234 ( or power - law index @xmath265 ) , and number of binaries ( @xmath266 )
are estimated using monte carlo simulations .
we find that the uncertainty in @xmath12 is less than 0.1 dex , and the error on the power - law index ( @xmath265 ) is @xmath267 . the uncertainty in measuring @xmath266 is about @xmath268 .
the corrected cumulative @xmath239 distribution obtained from one of these simulations is shown in figure [ loga_cum ] ( _ bottom right _ ) .
the semi - major axis distribution of `` true '' binaries in the simulation sample is @xmath269 , and is valid between 100 au and @xmath270 au .
the turnover in the distribution happens because there are no `` true '' binaries above 10,000 au , only random pairs , similar to what we observe in real data .
this similarity is a strong warning not to over - interpret the slope of @xmath234 beyond @xmath12 . to estimate the dependence of @xmath265 ( shape of @xmath234 ) , @xmath12 , and @xmath259 on color
, we divide the @xmath271 sample into three color subsamples using @xmath272 and @xmath273 lines .
we find that @xmath234 follows od in all three subsamples ( @xmath274 ) , and that the average @xmath12 is 3.99 , with a 0.07 root - mean - square scatter .
the @xmath12 for the full @xmath275 sample is 4.02 .
these results suggest that @xmath12 and the shape of @xmath234 are _ independent of color of binaries_. the @xmath259 value , and subsequently the number of binaries , will depend on the sample s color range . for the full @xmath275 sample ,
the number of binaries is @xmath276 where @xmath277 , @xmath278 , are @xmath259 values measured for each color subsample .
therefore , the number of binaries calculated for a distance bin will change as the color range changes . assuming that the @xmath46 color distribution of binaries does not change with @xmath13 , we can use the @xmath46 color distribution for the @xmath275 sample ( solid line in figure [ prob_gi ] ( _ top _ ) ) , to correct for color incompleteness .
we also assume that the fraction of binaries outside the @xmath218 color range is small .
the correct number of binaries is then @xmath279 \,\log_{10}(a_2/a_1),\ ] ] where @xmath280 $ ] is the area underneath the solid line histogram in figure [ prob_gi ] ( _ top _ ) , between @xmath281 and @xmath282 ( @xmath46 color range for a given distance bin ) . the estimated systematic error in @xmath12 due to the choice of the @xmath283 cut is measured using @xmath284 and @xmath175 samples .
we find that @xmath12 changes by @xmath285 dex between these samples .
this result suggests that @xmath12 is not sensitive to the choice of the @xmath283 cut .
similarly , the change in @xmath12 is less than 0.03 dex if the estimate of @xmath65 is off by @xmath286 ( @xmath287 change ) from the best - fit @xmath65 . to establish whether semi - major axis distribution follows the od in other @xmath13 bins , we repeat the @xmath234 and @xmath12 measuring procedure on 8 @xmath13 bins , and show the corrected cumulative distributions with best - fit straight lines in figure [ loga_cum_fits ] . in general , the corrected cumulative distributions follow a straight line , and then start to deviate from it at @xmath12 . in the @xmath288 bin
we do not see a turnover due to a narrow range of probed projected separations ( @xmath289 limits the range to 3193 au , see figure [ theta_eff_0.1_0.4_geo ] ) , and only determine the upper limit on @xmath12 . as the average height above the galactic plane increases , the @xmath12 moves to higher values .
we investigate this correlation in more detail in figure [ logz_plots ] ( _ top left _ ) .
the data follow a straight line @xmath290)+l$ ] , where @xmath291 and @xmath292 , or approximately , @xmath293 = 12,300 \ , z[kpc]^{0.7}$ ] in the @xmath294 range .
it is possible that @xmath12 also depends on the cylindrical radius , @xmath161 , with the sun at @xmath295=8 kpc , and perhaps on the local density of stars , @xmath296 .
because the sample is dominated by stars at high galactic latitudes , it is hard to disentangle the @xmath13 dependence from the other two effects ( the @xmath161 range is small , and @xmath296 varies strongly with @xmath13 ) .
we attempt to do so using the volume - complete @xmath275 sample .
first we divide this sample into three subsamples with median galactic latitudes , @xmath297 , of @xmath298 , @xmath299 , and @xmath300 and determine @xmath12 for each subsample .
the best - fit @xmath12 varies by @xmath970.3 dex between the low - latitude and high - latitude subsample , despite the same median @xmath13 .
when the @xmath275 sample is divided into the galactic anticenter ( @xmath301 ) and the galactic center ( @xmath302 or @xmath303 ) subsamples , the best - fit @xmath12 varies by @xmath970.1 dex .
these variations suggest that the best - fit @xmath13 dependence does not fully capture the behavior of @xmath12 .
nevertheless , they are smaller ( @xmath304 dex ) than the observed variation of @xmath12 ( @xmath971 dex ) . the spatial distribution of wide binaries
can now be compared to the number density of all stars as a function of height above the galactic plane . in figure [ logz_plots ] ( _ bottom left _ ) we show that wide binaries closely follow the spatial distribution of stars , with exponential decline in the number density as a function of @xmath13 .
the fraction of binaries relative to the number of all stars , shown in the bottom right panel , changes by only a factor of 2 over a range of 3 kpc , starting from @xmath305 at @xmath306 pc and declining to @xmath208 at @xmath307 pc
. it would be informative to repeat the @xmath234 and @xmath12 analysis using kinematically - selected binaries , but unfortunately , the apparent incompleteness of sdss - poss proper motion data at @xmath85 prevents us in doing so .
as shown in figure [ theta_fit_500pc ] , the number of common proper motion pairs drops sharply below @xmath308 , probably due to blending of close sources in the poss data . because of this @xmath50 cut - off , for the same range in @xmath13 , the effective @xmath256 for the kinematic sample is three times that of the geometric sample ( where the lower limit on @xmath50 is @xmath103 ) . in the case of @xmath288 sample
observed here , the smallest probed semi - major axis ( @xmath256 ) is at 5079 au , well above the @xmath12 value of 4534 au predicted by the @xmath309 relation .
since we are outside the range where od is valid , we can not measure where the turnover in @xmath234 happens , and can not determine @xmath12 or @xmath266 .
in all the other @xmath13 bins , @xmath256 is also above the predicted @xmath12 value , and therefore outside the pik regime .
we have presented a novel approach to photometric parallax estimation based on samples of candidate wide binaries selected from the sloan digital sky survey ( sdss ) imaging catalog .
our approach uses the fact that binary system s components are at equal distances and estimates the photometric parallax relation for main - sequence stars by minimizing the difference of their distance moduli . while this method is similar to constraints on photometric parallax relation obtained from globular clusters in that it does not require absolute distance estimates , it has the advantage that it extends to redder colors than available for globular clusters observed by the sdss , and it implicitly accounts for the metallicity effects .
the derived best - fit photometric parallax relations represent metallicity - averaged relations and thus provide an independent confirmation of relations proposed by j08 in their study of the galactic structure .
an important result of this work is our estimate of the expected error distribution for absolute magnitudes determined from photometric parallax relations ( a root - mean - square scatter of @xmath970.3 mag , see section [ scatter ] ) , which is in good agreement with modeling assumptions adopted by j08 .
the mildly non - gaussian error distribution is consistent with both the impact of unresolved binary stars , and the variation of photometric parallax relation with metallicity ; we are unable to disentangle these two effects .
the best - fit photometric parallax relations enabled the selection of high - efficiency samples of disk wide binaries with @xmath310 candidates , that include about 14,000 true binary systems ( efficiency of @xmath311 ) . using the photometric measurements and angular distance of the two components , samples with efficiency exceeding 80%
can be constructed ( see section [ samples ] ) .
such samples could be used as a starting point to further increase the selection efficiency with the aid of radial velocity measurements .
spectral observations of systems where the brighter component is an f / g star , for which it is easy to estimate metallicity , could be used to calibrate both spectroscopic and photometric methods for estimating metallicity of cooler k and m dwarfs .
compared to the state - of - the - art catalogs of wide binaries by @xcite and @xcite , the samples discussed here represent a significant increase in the number of potential binaries , and probe larger distances ( to @xmath216 kpc ) . to facilitate further studies of wide binaries
, we make the catalog publicly available .
using the high - efficiency subsamples , we analyzed their dynamical and physical properties .
we find that the spatial distribution of wide binaries follows the distribution of single stars to within a factor of 2 , and that the probability for a star to be in a wide binary system is independent of its color .
however , given this color , the companions of red components seem to be drawn randomly from the stellar luminosity function , while blue components have a larger blue - to - red companion ratio than expected from luminosity function ( see section [ color ] ) .
these results are consistent with recent results by @xcite , and provide strong constraints for the scenarios describing the formation of such systems ( e.g. , @xcite and references therein ; @xcite ; @xcite ) .
we also study the semi - major axis distribution of wide binaries in the @xmath312 au range ( see section [ spatial ] ) . the observed distribution
is well described by the pik distribution , @xmath10 , for @xmath11 , where @xmath12 increases roughly linearly with the height above the galactic plane ( @xmath313 au at @xmath314 kpc ) .
alternatively , the @xmath12 correlates with the local number density of stars as @xmath315 , but we are unable to robustly identify the dominant correlation ( @xmath13 and @xmath296 are highly correlated ) .
the distribution of semi - major axes for wide binaries was also discussed by @xcite .
they used a sample of wide binaries selected using common proper motion from the rnltt catalog @xcite , and found @xmath316 , with no evidence of a turnover at @xmath317 .
their sample extended to larger angular separations than ours , and probed smaller distances . on the other hand
, @xcite used wide binaries from the same @xcite sample , and detected pik distribution , @xmath10 , for @xmath318 , consistent with the result of @xcite .
in a recent study , lpine & bongiorno searched for faint common proper motion companions of hipparcos stars and detected a turnover from pik distribution to a steeper distribution around @xmath319 au .
their sample also probed much smaller distances than ours .
we compare these results in figure [ fig : compare ] .
as evident , the variation of @xmath12 with distance from the galactic plane detected here ( approximately with distance , as shown in figure [ fig : compare ] , since stars in our sample are mostly at high galactic latitudes ) , is consistent with the above results that are based on more local samples .
in particular , this comparison of different studies suggests that the flattening of @xmath234 for small @xmath8 that `` puzzled '' chanam & gould ( see their section 4.3 ) is probably due to a combination of selection effects and the approach of the domain where pik distribution is valid in their sample .
the pik distribution suggests that the process of star formation produces multiple stars , which evolve towards binaries after ejecting one or more single stars @xcite .
the departure from the pik distribution may be evidence for disruption of wide binaries over long periods of time by passing stars , giant molecular clouds , massive compact halo objects ( machos ) , or disk and galactic tides @xcite .
however , we note that the @xmath315 correlation ( see figure [ logz_plots ] ) is outside the expected range discussed by @xcite ( @xmath320 for close strong encounters , and @xmath321 for weak encounters )
. the samples presented here can be further refined and enlarged .
first , the sdss covers only a quarter of the sky . upcoming next - generation surveys , such as the skymapper @xcite , the dark energy survey @xcite , pan - starrs @xcite and the large synoptic survey telescope ( @xcite , lsst hereafter ) , will enable the construction of such samples over most of the sky . due to fainter flux limits ( especially for the pan - starrs and lsst )
, the samples will probe a larger distance range and will reach the halo - dominated parts of the galaxy .
furthermore , due to improved photometry and seeing ( e.g. , for the lsst , by about a factor of two ) , the selection will be more robust .
we scale the 20,000 candidates discussed here , assuming @xmath322 , to the lsst depth that enables accurate photometric metallicity ( @xmath323 ; i08a ) and predict a minimum sample size of @xmath97400,000 candidate wide binary systems in 20,000 deg@xmath33 of sky .
it is likely that the sample would include more than a million systems due to the increase of the stellar counts close to the galactic plane .
another important development will come from the gaia mission @xcite , which will provide direct trigonometric distances for stars with @xmath324 . with trigonometric distances
, accurate photometric parallax relation can be used to provide strong constraints on the incidence and color distribution of unresolved multiple systems . until then , a radial velocity survey of candidate binaries assembled here could help with pruning the sample from random associations , and with better characterization of various selection effects .
this work was supported by the nsf grant ast-0707901 , and the nsf grant ast-0551161 to the lsst for design and development activity .
m. j. gratefully acknowledges support from the taplin fellowship and from the nsf grant phy-0503584 .
we are grateful to nick cowan and eric agol ( uw ) for help with the markov chain monte carlo code
. early motivation for this analysis came from a preliminary study by taka sumi .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
the following sql queries were used to select initial samples of candidate binaries through the sdss casjobs interface . when running these queries , the database context must be set to `` dr6 '' or higher . .... select -- geometric selection of candidate binaries round(p1.ra,6 ) as ra1 , round(p1.dec,6 ) as dec1 , round(p1.extinction_r,3 ) as rext1 ,
round(p1.psfmag_u,3 ) as psf_u1 , round(p1.psfmag_g,3 ) as psf_g1 , round(p1.psfmag_r,3 ) as psf_r1 , round(p1.psfmag_i,3 ) as psf_i1 , round(p1.psfmag_z,3 ) as psf_z1 , round(p1.psfmagerr_u,3 ) as psferr_u1 , round(p1.psfmagerr_g,3 ) as psferr_g1 , round(p1.psfmagerr_r,3 ) as psferr_r1 , round(p1.psfmagerr_i,3 ) as psferr_i1 , round(p1.psfmagerr_z,3 ) as psferr_z1 , p1.objid as objid1 , round(p2.ra,6 ) as ra2 , round(p2.dec,6 ) as dec2 , round(p2.extinction_r,3 ) as rext2 , round(p2.psfmag_u,3 ) as psf_u2 , round(p2.psfmag_g,3 ) as psf_g2 , round(p2.psfmag_r,3 ) as psf_r2 , round(p2.psfmag_i,3 ) as psf_i2 , round(p2.psfmag_z,3 ) as psf_z2 , round(p2.psfmagerr_u,3 ) as psferr_u2 , round(p2.psfmagerr_g,3 ) as psferr_g2 , round(p2.psfmagerr_r,3 ) as psferr_r2 , round(p2.psfmagerr_i,3 ) as psferr_i2 , round(p2.psfmagerr_z,3 ) as psferr_z2 , p2.objid as objid2 , round(nn.distance*60,3 ) as theta into mydb.binaryclose from neighbors as nn join star as p1 on p1.objid = nn.objid join star as p2 on p2.objid = nn.neighborobjid where nn.mode = 1 and nn.neighbormode = 1 and nn.type = 6 and nn.neighbortype = 6 and p1.psfmag_r between 14 and 20.5 and ( p1.flags_g & ' 229802225959076 ' ) = 0 and ( p1.flags_r & ' 229802225959076 ' ) = 0 and ( p1.flags_i & ' 229802225959076 ' ) = 0 and ( p1.flags_g & ' 268435456 ' ) > 0 and ( p1.flags_r & ' 268435456 ' ) > 0 and ( p1.flags_i & ' 268435456 ' ) > 0 and p2.psfmag_r between 14 and 20.5 and ( p2.flags_g & ' 229802225959076 ' ) = 0 and ( p2.flags_r & ' 229802225959076 ' ) = 0 and ( p2.flags_i & ' 229802225959076 ' ) = 0 and ( p2.flags_g & ' 268435456 ' ) > 0 and ( p2.flags_r & ' 268435456 ' ) > 0 and ( p2.flags_i & ' 268435456 ' ) > 0 and ( p1.psfmag_r-p1.extinction_r ) < ( p2.psfmag_r-p2.extinction_r ) and ( p1.psfmag_g-p1.extinction_g - p1.psfmag_i+p1.extinction_i ) < ( p2.psfmag_g-p2.extinction_g - p2.psfmag_i+p2.extinction_i ) and nn.distance*60 between 3 and 4 .... .... select -- kinematic selection of candidate binaries round(p1.ra,6 ) as ra1 , round(p1.dec,6 ) as dec1 , round(p1.extinction_r,3 ) as ext1 , round(p1.psfmag_u,3 ) as u1 , round(p1.psfmag_g,3 ) as g1 , round(p1.psfmag_r,3 ) as r1 , round(p1.psfmag_i,3 ) as i1 , round(p1.psfmag_z,3 ) as z1 , round(p1.psfmagerr_u,3 ) as uerr1 , round(p1.psfmagerr_g,3 ) as gerr1 , round(p1.psfmagerr_r,3 ) as rerr1 , round(p1.psfmagerr_i,3 ) as ierr1 , round(p1.psfmagerr_z,3 ) as zerr1 , ( case when ( ( p1.flags & ' 16 ' ) = 0 ) then 1 else 0 end ) as isolated1 , nn.objid as objid1 , round(p2.ra,6 ) as ra2 , round(p2.dec,6 ) as dec2 , round(p2.extinction_r,3 ) as ext2 , round(p2.psfmag_u,3 ) as u2 , round(p2.psfmag_g,3 ) as g2 , round(p2.psfmag_r,3 ) as r2 , round(p2.psfmag_i,3 ) as i2 , round(p2.psfmag_z,3 ) as z2 , round(p2.psfmagerr_u,3 ) as uerr2 , round(p2.psfmagerr_g,3 ) as gerr2 , round(p2.psfmagerr_r,3 ) as rerr2 , round(p2.psfmagerr_i,3 ) as ierr2 , round(p2.psfmagerr_z,3 ) as zerr2 , ( case when ( ( p2.flags & ' 16 ' ) = 0 ) then 1 else 0 end ) as isolated2 , nn.neighborobjid as objid2 , round(nn.distance*60,3 ) as theta , round(s1.pml,3 ) as pml1 , round(s1.pmb,3 ) as pmb1 , round(s2.pml,3 ) as pml2 , round(s2.pmb,3 ) as pmb2 into mydb.binarypm from neighbors as nn join star as p1 on p1.objid = nn.objid join star as p2 on p2.objid = nn.neighborobjid join propermotions as s1 on s1.objid = nn.objid join propermotions as s2 on s2.objid = nn.neighborobjid where nn.mode = 1 and nn.neighbormode = 1 and nn.type = 6 and nn.neighbortype = 6 and p1.psfmag_r between 14 and 19.5 and ( p1.flags_g & ' 229802225959076 ' ) = 0 and ( p1.flags_r & ' 229802225959076 ' ) = 0 and ( p1.flags_i & ' 229802225959076 ' ) = 0 and ( p1.flags_g & ' 268435456 ' ) > 0 and ( p1.flags_r & ' 268435456 ' ) > 0 and ( p1.flags_i & ' 268435456 ' ) > 0 and p2.psfmag_r between 14 and 19.5 and ( p2.flags_g & ' 229802225959076 ' ) = 0 and ( p2.flags_r & ' 229802225959076 ' ) = 0 and ( p2.flags_i & ' 229802225959076 ' ) = 0 and ( p2.flags_g & ' 268435456 ' ) > 0 and ( p2.flags_r & ' 268435456 ' ) > 0 and ( p2.flags_i & ' 268435456 ' ) > 0 and ( p1.psfmag_r-p1.extinction_r ) < ( p2.psfmag_r-p2.extinction_r ) and ( p1.psfmag_g-p1.extinction_g - p1.psfmag_i+p1.extinction_i ) < ( p2.psfmag_g-p2.extinction_g - p2.psfmag_i+p2.extinction_i ) and s1.match = 1 and s2.match = 1 and s1.sigra < 350 and s1.sigdec < 350 and s2.sigra < 350 and s2.sigdec
< 350 and sqrt(power(s1.pml - s2.pml,2 ) + power(s1.pmb - s2.pmb,2 ) ) < 5 and ( case when sqrt(power(s1.pml,2 ) + power(s1.pmb,2 ) ) > sqrt(power(s2.pml,2 ) + power(s2.pmb,2 ) ) then sqrt(power(s1.pml,2 ) + power(s1.pmb,2 ) ) else sqrt(power(s2.pml,2 ) + power(s2.pmb,2 ) ) end ) between 15 and 400 ....
recent analysis of metallicity and kinematics for halo and disk stars by i08a provides sufficient information to understand the behavior of the reduced proper motion diagram in quantitative detail ( including both the sequence separation and their widths ) , and to demonstrate that its efficiency for separating halo and disk stars deteriorates at distances beyond a few kpc from the galactic plane .
as equation [ eq : rpm2 ] shows , for a population of stars with the same @xmath114 , the reduced proper motion is a measure of their absolute magnitude . for two stars with the same color that is sensitive to the effective temperature ( such as the @xmath46 color ) , but with different metallicities and tangential velocities , the difference in their reduced proper motions is @xmath325 where @xmath326 and @xmath94 denote the two stars . in the limit that the _ shape _ of the photometric parallax relation does not depend on metallicity ,
@xmath327 does not depend on color , and is fully determined by the metallicity difference of the two stars ( or populations of stars ) .
using metallicity distributions for disk and halo stars obtained by i08a , and their expression for @xmath328)$ ] ( equation a2 ) , we find that the expected offset between @xmath113 for halo and disk stars with the same @xmath46 color varies from 0.6 mag for stars at 1 kpc from the galactic plane to 0.7 mag at 5 kpc from the plane , where the variation is due to the vertical metallicity gradient for disk stars
. the finite width of halo and disk metallicity distributions induces a spread of @xmath113 ( root - mean - square scatter computed using interquartile range ) of 0.15 mag for disk stars and 0.18 mag for halo stars .
the effect of metallicity on the separation of halo and disk sequences in the reduced proper motion diagram is smaller than the effect of different tangential velocity distributions .
assuming for simplicity that stars are observed towards a galactic pole , and that the median heliocentric tangential velocities are 30 km s@xmath37 for disk stars and 200 km s@xmath37 for halo stars , the induced separation of their reduced proper motion sequences is @xmath974.1 mag ( the expected scatter in the reduced proper motion due to finite velocity dispersion is @xmath971 - 1.5 mag ) .
together with the @xmath329 mag offset due to different metallicity distributions , the separation of @xmath975 mag between the two sequences makes the reduced proper motion diagram a promising tool for separating disk and halo stars .
however , the reduced proper motion diagram is an efficient tool only for stars within 1 - 2 kpc from the galactic plane .
the main reason for this limitation is the decrease of rotational velocity for disk stars with distance from the galactic plane , with a gradient of about @xmath330 km s@xmath37 kpc@xmath37 ( see section 3.4.2 in i08a ) .
as the difference in rotational velocity between halo and disk stars diminishes with the distance from the plane , the separation of their reduced proper motion sequences decreases , too . a mild increase in the velocity dispersion of disk stars , as well as a decrease of their median metallicity with the distance from the plane , also decrease the sequence separation , but the dominant cause is the rotational velocity gradient . to illustrate this effect
, we select a sample of @xmath9760,000 stars with @xmath331 and @xmath332 , that are observed towards the north galactic pole ( @xmath99 ) . in this color range
it is possible to separate disk and halo stars using photometric metallicity estimator from i08a , and we further select a sample of @xmath9716,000 likely disk stars with @xmath333 > -0.9 $ ] , and a sample of @xmath9734,400 likely halo stars with @xmath333 < -1.1 $ ] ( see figure 9 in i08a for justification ) . their proper motion distributions as functions of distance from the galactic plane , @xmath13 , are shown in the top left panel in figure [ fig : app1 ] . because of the gradient in the rotational velocity for disk stars , their median proper motion becomes constant at @xmath978 mas yr@xmath37 beyond @xmath334 kpc , while the median proper motion for halo stars is roughly proportional to @xmath335 , with a value of @xmath9711 mas yr@xmath37 at @xmath336 kpc .
the top right panel in figure [ fig : app1 ] shows the positions and widths of the reduced proper motion sequences for disk and halo stars as functions of @xmath13 , and the two bottom panels show the sequence cross - sections for stars with @xmath337 kpc and @xmath338 kpc . at distances beyond @xmath972 kpc from the plane ,
the reduced proper motion diagram ceases to be an efficient tool for separating halo and disk stars because the two sequences start to significantly overlap .
this increasing overlap is a result of the rotational velocity gradient for disk stars , and the finite width of halo and disk velocity distributions , and would be present even for _ infinitely accurate _ measurements ( with the proper motion errors of @xmath973 mas yr@xmath37 per coordinate , @xcite , the sequence widths of @xmath971.0 - 1.5 mag are dominated by velocity dispersions ) .
hence , beyond @xmath972 kpc from the plane , metallicity measurements are necessary to reliably separate disk and halo populations .
the above analysis is strictly valid only for fields towards the north galactic pole .
@xcite found that the position of disk and halo reduced proper motion sequences , relative to their positions at the north galactic pole , varies with galactic latitude as @xmath339 where @xmath340 is the median value of @xmath114 for stars observed towards the north galactic pole .
this result is a bit unexpected because it does not contain longitudinal variation due to projection effects of the rotational motion of the local standard of rest .
we show the variation of @xmath341 , for stars with @xmath332 , as a function of galactic coordinates in figure [ fig : app2 ] .
we use photometric metallicity to separate stars into disk and halo populations .
as figure demonstrates , the longitudinal dependence is present for halo sample , but not for disk samples .
we have generated simulated behavior of @xmath341 using kinematic model from i08a , and reproduced the observed behavior to within the measurement noise .
it turns out that the vertical gradient of rotational velocity for disk stars is fully responsible for the observed strong dependence of @xmath341 on latitude , and which masks the dependence on longitude .
hence , the @xmath342 term proposed by @xcite is an indirect discovery of the vertical gradient of rotational velocity for disk stars !
these empirical models also show that a linear dependence of @xmath343 on @xmath342 is only approximately correct , and that it ignores the dependence on distance . while a more involved best - fit expression is possible ( full two - dimensional consideration of proper motion also helps to better separate disk and halo stars ) , we find that halo stars can always be efficiently rejected at @xmath344 , if the separator shown in figure [ fig:1wd ] is shifted upwards by 1 mag .
one major uncertainty when using a photometric parallax relation is the lack of information whether the observed `` star '' is a single star , or a binary ( multiple ) system .
if the observed `` star '' is a binary system , its luminosity will be underestimated , with the magnitude of the offset depending on the actual composition of the binary . to model this offset , or to correct for it
, one would ideally like to have a probability density map that gives the probability of a magnitude offset , @xmath327 , as a function of the _ observed _ binary system s color .
to construct such a map , we have generated a sample of 100,000 unresolved binary systems by randomly pairing stars drawn from the @xcite luminosity function . by independently drawing the luminosities of each component to generate unresolved binary systems
, we implicitly assume that the formation of each component is unaffected by the presence of the other . while there are other proposed mechanisms for binary formation ( @xcite , and references therein ) , we have chosen this one because it was easy to implement . for every unresolved binary system
we calculate the total @xmath30 band luminosity , and the @xmath122 and @xmath46 color of the system .
the magnitude offset , @xmath327 , caused by unresolved binarity , is obtained as the difference between the true @xmath30 band absolute magnitude , and the absolute magnitude for the pair s joint @xmath122 color calculated using equation [ mr_bright ] .
the probability density map is then simply the number of unresolved binary systems ( normalized with the total number of systems at a given color ) as a function of @xmath327 and pair s joint @xmath46 color , shown in figure [ dmr_gi ] .
it is worth noting that , with the adopted binary formation mechanism , the magnitude offset is the smallest ( @xmath345 mag ) for the bluest stars , and greatest ( @xmath346 mag ) for the reddest stars . because of this , the scatter due to unresolved binarity in the @xmath26 distribution should be more pronounced in a sample of red stars ( @xmath347 ) , than in a sample of blue stars .
the map shown in figure [ dmr_gi ] can be parametrized as a gaussian distribution @xmath348 , where @xmath349 is the median @xmath327 , and @xmath350 is the scatter ( determined from the interquartile range ) . to verify the validity of this parametrization
, we subtract @xmath327 and @xmath109 , normalize the difference with @xmath148 , find the distribution of such values , and fit a gaussian to it . as shown in figure [ dmr_chi_gauss ] , the distribution of normalized residuals is well described by a gaussian with @xmath351 .
the two peaks in the distribution are due to highly asymmetric distributions of @xmath327 values around the median @xmath327 for the bluest ( @xmath352 ) and reddest ( @xmath353 ) systems . to create a sample of wide binaries where some of the stars are unresolved binary systems , first we select pairs with @xmath58 from the initial sample of stellar pairs . following the procedure described in section [ algorithm ] , we create the `` true '' wide binaries by changing the @xmath92 magnitude using equation [ r2_mag ] , and add 0.15 mag of gaussian noise to simulate the scatter in the photometric parallax due to photometric errors .
a fraction of stars is then randomly converted to unresolved binary systems by subtracting a @xmath327 value from the @xmath30 band ( apparent ) magnitude , where the @xmath327 is drawn from a @xmath46 color - dependent @xmath348 distribution .
figure [ gauss_fit_red ] shows the @xmath26 distribution for such a mock sample , where the components are redder than @xmath354 and have a 40% probability to be unresolved binary systems .
different configurations of single stars and unresolved binaries that contribute to the observed @xmath26 distribution can be easily identified .
wide binaries where both components are single stars contribute the central narrow gaussian , with its width due to photometric errors .
if the brighter component is an unresolved binary system , its absolute magnitude is underestimated , and the result is an offset in @xmath26 in the negative direction .
a similar outcome happens if the fainter component is an unresolved binary system , but the offset is positive .
single star - unresolved binary configurations , therefore , contribute the left and the right gaussians .
if both components are unresolved binary systems , the @xmath26 will be centered on zero and will be @xmath355 wide , where @xmath356 is the width of the @xmath357 distribution .
this behavior is consistent with the @xmath26 distributions observed in figure [ delta_hists ] .
adelman - mccarthy , j. k. et al .
2008 , , 175 , 297 allen , c. , poveda , a. , & herrera , m. 2000 , , 356 , 529 allen , c. , poveda , a. , & hernndez - alcntara , a. 2007 , in iau symp .
240 , binary stars as critical tools & test in contemporary astrophysics , ed .
b. hartkopf , e. guinan , & p. harmanec , 405 andersen , j. 1991 , , 3 , 91 bahcall , j. n. & soneira , r. m. 1981 , , 246 , 122 bochanski , j. et al , 2008 , in prep chanam , j. 2007 , in iau symp .
240 , binary stars as critical tools & tests in contemporary astrophysics , ed .
b. hartkopf , e. guinan , & p. harmanec , 1 chanam , j. & gould , a. 2004 , , 601 , 289 clarke , c. j. 2007 , in iau symp .
240 , binary stars as critical tools & tests in contemporary astrophysics , ed .
b. hartkopf , e. guinan , & p. harmanec , 337 covey , k. r. , et al .
2007 , , 134 , 2398 couteau , p. 1960 , j. des observateurs , 43 , 41 croll , b. 2006 .
, , 118 , 1351 duquennoy , a. & mayor , m. 1991 , , 248 , 485 finlator , k. et al .
2000 , , 120 , 2615 fischer , d. a. & marcy , g. w. 1992 , , 396 , 178 flaugher , b. & the dark energy survey collaboration 2007 , , 209 , 22.01 ford , e. b. 2005 , , 129 , 1706 gelman , a. & rubin , d. 1992 , stat .
, 7 , 457 giersz , m. 2006 , , 371 , 484 gould , a. 1995 , , 440 , 510 gould , a. , & salim , s. 2003 , , 582 , 1001 gunn , j. e. et al .
1998 , , 116 , 3040 gunn , j. e. et al .
2006 , , 131 , 2332 hastings , w. k. 1970 , biometrika , 57 , 97 hawley , s. l. et al .
2002 , , 123 , 3409 heggie , d. c. 1975 , , 173 , 729 helmi , a. et al .
2003 , , 586 , 195 hogg , d. w. , finkbeiner , d. p. , schlegel , d. j. & gunn , j.e .
2002 , , 122 , 2129 hurley , j. r. , aarseth , s. j. , & shara , m. m. 2007 , , 665 , 707 ivezi , .
2003 , , 74 , 978 ivezi , .
2004 , astron .
nachr . , 325 , 583 ivezi , . , vivas , a. k. , lupton , r. h. , & zinn , r. 2005 , , 129 , 1096 ivezi , .
2008a , accepted to ( also astro - ph/0804.3850 ) ivezi , .
et al . 2008b , astro - ph/0805.2366 juri , m. et al .
2008 , , 673 , 864 kaiser , n. et al .
2002 , , 4836 , 154 keller , s.c .
2007 , 24 , 1 kroupa , p. , tout , c. a. & gilmore , g. 1990 , , 244 , 76 laird , j. b. , carney , b. w & latham , d. w. 1988 , , 95 , 1843 lenz , d. d. , newberg , j. , rosner , r. , richards , g. t. , & stoughton , c. 1998 , , 119 , 121 lpine , s. & bongiorno , b. 2007 , , 133 , 889 lupton , r. h. , ivezi , . ,
gunn , j. e. , knapp , g. r. , strauss , m. a. & yasuda , n. 2002 , 4836 , 350 luyten , w. j. 1979 , new luyten catalog of stars with proper motions larger than two tenths of an arcsecond ( minneapolis : univ . of minnesota press ) minkowski , a. & abell , g. o. 1963 , in stars and stellar systems , vol .
3 metropolis , n. , rosenbluth , a. w. , rosenbluth , m. n. , teller , a. h. , & teller , e. , j. chem .
phys . , 21 , 1087 monet , d. g. et al .
2003 , , 125 , 984 munn , j. a. et al .
2004 , , 127 , 3034 pik , e. j. 1924 , tartu obs .
25 , no . 6
perryman , m. a. c. et al .
2001 , , 369 , 339 pier , j. r. , munn , j. a. , hindsley , r. b. , hennesy , g. s. , kent , s. m. , lupton , r. h. & ivezi , .
2003 , , 125 , 1559 poveda , a. et al .
1994 , revista mexicana de astronoma y astrofisca , 28 , 43 poveda , a. , allen , c. & hernndez - alcntara , a. 2007 , in iau symp .
240 , binary stars as critical tools & test in contemporary astrophysics , ed .
b. hartkopf , e. guinan , & p. harmanec , 417 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 1992
, numerical recipes in c ( 2nd ed . ;
new york : cambridge univ .
press ) reid , i. n. et al .
1991 , , 103 , 661 ribas , i. 2006 , asp conf . 349 , 55 richards , g. t. et al .
2002 , , 123 , 2945 salim , s. & gould , a. 2003 , , 582 , 1011 schlegel , d. , finkbeiner , d. p. , & davis , m. 1998 , 500 ,
525 scranton , r. et al .
2002 , , 579 , 48 sesar , b. et al .
2007 , , 134 , 2236 siegel , m. h. , majewski , s. r. , reid , i. n. & thompson , i. b. 2002 , , 578 , 151 sirko , e. et al .
2004 , , 127 , 899 smith , j. a. et al .
2002 , , 123 , 2121 stoughton , c. et al .
2002 , , 123 , 485 tegmark , m. et al .
2004 , , 69 , 103501 tucker , d. et al .
2006 , astron .
, 327 , 821 weinberg , m. d , shapiro , s. l. , & wasserman , i. 1987 , , 312 , 367 west , a. a. , walkowicz , l. m. & hawley , s. l. 2005 , , 117 , 706 wilkinson , m. i. et al .
2005 , , 359 , 1306 yanny , b. et al .
2000 , , 540 , 825 yoo , j. , chanam , j. & gould , a. 2004 , , 601 , 311 | using sdss data release 6 , we construct two independent samples of candidate stellar wide binaries selected as i ) pairs of unresolved sources with angular separation in the range @xmath0 , ii ) common proper motion pairs with @xmath1 angular separation , and make them publicly available . these samples are dominated by disk stars , and we use them to constrain the shape of the main - sequence photometric parallax relation @xmath2 , and to study the properties of wide binary systems .
we estimate @xmath2 by searching for a relation that minimizes the difference between distance moduli of primary and secondary components of wide binary candidates .
we model @xmath2 by a fourth degree polynomial and determine the coefficients using markov chain monte carlo fitting , independently for each sample .
both samples yield similar relations , with the largest systematic difference of 0.25 mag for f0 to m5 stars , and a root - mean - square scatter of 0.13 mag . a similar level of agreement is obtained with photometric parallax relations recently proposed by @xcite .
the measurements show a root - mean - square scatter of @xmath3 mag around the best fit @xmath2 relation , and a mildly non - gaussian distribution .
we attribute this scatter to metallicity effects and additional unresolved multiplicity of wide binary components .
aided by the derived photometric parallax relation , we construct a series of high - quality catalogs of candidate main - sequence binary stars .
these range from a sample of @xmath4 candidates with the probability of each pair to be a physical binary ( the `` efficiency '' ) of @xmath5 , to a volume - limited sample of @xmath6 candidates with an efficiency of @xmath7 . using these catalogs
, we study the distribution of semi - major axes of wide binaries , @xmath8 , in the @xmath9 au range . we find the observations to be well described by the pik distribution , @xmath10 , for @xmath11 , where @xmath12 increases roughly linearly with the height @xmath13 above the galactic plane ( @xmath14}^{0.7}$ ] au ) . the number of wide binary systems with @xmath15 , as a fraction of the total number of stars , decreases from 0.9% at @xmath16 kpc to 0.5% at @xmath17 kpc .
the probability for a star to be in a wide binary system is independent of its color . given this color , the companions of red components seem to be drawn randomly from the stellar luminosity function , while blue components have a larger blue - to - red companion ratio than expected from luminosity function . |
following the unification scheme the central nucleus of an active galaxy ( agn ) consists of a black hole ( bh ) , an accretion disk , line - emitting clouds , a dust torus , and emanates prominent jets when classified as radio - loud .
the properties of radio - loud agn viewed at a small angle to the line - of - sight are in general agreement with the common blazar properties ( e.g. , * ? ? ?
* ; * ? ? ?
their broadband emission covers the complete electromagnetic band , from the radio up to the @xmath0-ray band , in some cases even reaching tev - energies , and is widely dominated by beamed non - thermal emission from a relativistic jet .
the blazar class subdivides into bl lac objects and flat - spectrum radio quasars ( fsrqs ) .
the difference dividing both subclasses is generally considered in the detection of strong emission lines in the case of fsrqs , while in bl lac objects the equivalent width of emission lines is depressed , or lines are absent at all .
the physical reason is thought to lie predominantly in the weak accretion disk radiation field of bl lac objects , whereas bh in the nuclei of fsrqs accrete with high rates leading to luminous accretion disk photon fields .
the present work deals with @xmath0-ray loud agn observed during an epoche of a bright accretion disk and accompagnied with the apprearance of strong emission lines .
for simplicity i will refer to them as quasars in the following , although some radio - loud agn classified conventionally as bl lac objects may occasionally fall into this category as well ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and vice versa .
gamma - ray production mechanisms , leptonic as well as hadronic ones , in agn jets often involve either the external radiation fields associated with the immediate agn environment or internal jet photons .
for example , contributions from interactions in the accretion disk or broad - line region ( blr ) radiation fields are often mandatory to explain the overall @xmath0-ray spectral energy distribution ( sed ) from fsrqs ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , while the sed of low - luminosity bl lac objects is often fitted with either synchrotron - self - compton models ( e.g. , * ? ? ?
* ) or their hadronic equivalent , the synchrotron - proton blazar models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the former scenario places the @xmath0-ray emission region rather close to the blr clouds .
this has immediate consequences : if external radiation fields in quasar environments play a non - negligible role for @xmath0-ray production , then , at the same time they are also significant for the quasi - resonant process of photon - photon pair production with its peak cross section possessing a comparable value to the thomson cross section .
it should be noted though , that the position of the high - energy emission region is still a matter of debate , with locations proposed to be also far away from the blr ( e.g. , * ? ? ?
* ; * ? ? ?
if this is the case , external compton emission ( e.g. , * ? ? ?
* ) or pair cascade radiation initiated by ultra - high energy cosmic ray interactions on external photon fields ( e.g. , * ? ? ?
* ; * ? ? ?
* ) in those sources does not provide an appreciable contribution to the observed high - energy emission .
the present work concerns specifically opacity features in the @xmath1 gev regime by photon absorption through e@xmath2-e@xmath3 pair production in radiation fields in the vicinity of the bh , but external to the jets in @xmath0-ray loud quasars .
past works on this subject ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) indicate the importance of this process for constraining agn properties like the location of the @xmath0-ray region above the disk , disk radiation fields , torus temperature , etc .
in contrast to these works , i will be focusing on the evolution of external radiation fields in quasar environments , and its consequences for the resulting opacity features in the @xmath0-ray band covered by current and near future instruments .
this is primarily motivated by the anticipated studies of the evolution of the extragalactic background light ( ebl ) through the detection of absorption features in a large sample of high - redshift sources using the large area telescope ( lat ) onboard glast .
dedicated methods have been developed here to probe the evolution of the ebl via detecting the horizon of @xmath0-rays emitted from extragalactic sources like agn and grbs while propagating through the ebl to earth ( e.g. , * ? ? ?
* ; * ? ? ?
they involve either the determination of the ratio of absorbed to unabsorbed flux versus redshift @xmath4 , or the detection of the e - folding cutoff energy @xmath5 versus redshift ( fazio - stecker relation ; * ? ? ?
* ) in a large number of sources at various redshifts in order to disentangle intrinsic blazar features from absorption caused ones during propagation in the ebl .
the common underlying reasoning for this procedure is that the observation of any redshift - dependent attenuation in @xmath0-ray agn can only be attributed to absorption in the ebl , and no other sources of redshift - dependent opacity exist in those sources . here
i will demonstrate that opacity due to photon absorption from @xmath6-pair production caused in external radiation fields within the agn system ( in the following called local absorption
, to be distinguished from `` self - absorption '' in the internal jet radiation fields ) will most likely result in optical depth values that increase with the source redshift , and coincidentally mimic redshift - dependent ebl - caused absorption .
the outline of this paper is organized as follows : sect . 2 and sect .
3 describe the considered external target photon fields ( accretion disk and blr radiation field ) for photon - photon pair production , and the corresponding optical depth calculation , respectively . in sect . 5
i will apply various models of supermassive bh growth and accretion rate evolution , that are described in sect . 4 , to the @xmath0-ray attenuation calculations .
the results with particular emphasis on the possibility of a redshift - dependence of the local opacity are presented in sect .
the paper closes with conclusions and a discussion in sect .
the most relevant target radiation fields in agn environments for photon absorption in the lat energy range , @xmath70.02 - 300 gev , are the optical / uv bands of the accretion disk photon field and the radiation field of the blr .
these will be considered in the following .
i assume the radiation fields to be located azimuthally symmetric with respect to the jet axis , and to radiate persistently during @xmath0-ray emission - ray domain , and thus allows the disk emission to be approximated as a constant photon field for the purpose of the present work . ] . the accretion disk spectrum in fsrqs is assumed to follow the cool , optically thick blackbody solution of shakura & sunyaev ( 1973 ) with a given accretion rate @xmath8 , suitable for agn that show strong emission lines ( `` strong - line agn '' ) .
the differential photon density @xmath9 into a solid angle @xmath10 , @xmath11 reads @xmath12^{-1}\ ] ] with @xmath13 , @xmath14 the distance of the emission region above the bh , @xmath15 the photon energy , and @xmath16 with @xmath17 for a schwarzschild metric , @xmath18 and @xmath19 the boltzmann constant . for the opacity calculations the disk
is considered to extend from 6 to @xmath20 .
typical accretion rates for strong quasars approach the eddington accretion rate @xmath21 , @xmath22 , respectively the eddington luminosity @xmath23 .
the accretion disk is considered as the photo - ionizing source of the blr material , and emission lines are produced through recombination .
the blr geometry is approximated as a spherical shell with radius @xmath24 filled with clouds , extending from @xmath25 to @xmath26 . in this picture the accretion disk as located at @xmath27 . in the following the shell size
is fixed to @xmath28pc to @xmath29pc , which are typical values for quasars ( e.g. , * ? ? ?
a geometrically extended shell is supported from the observations by the agn watch campaigns .
the opacity calculations require spectral as well as spatial knowledge of the emissivity of the blr radiation field .
information about the cloud sizes and their spatial distribution can be deduced in general from reverberation mapping .
although most details on blr physics and geometry have been derived by the study of radio - quiet agn , no major differences in the broad - line flux between radio - quiet and -loud sources have been found so far ( e.g. , * ? ? ?
* ; * ? ? ?
for the present calculations the number density @xmath30 and cross - section @xmath31 of the clouds are assumed to follow a power law with exponents @xmath32 and @xmath33 as derived by kaspi & netzer ( 1999 ) .
a fraction @xmath34 of the central source luminosity @xmath35 is re - processed into line radiation such that the total blr luminosity , assumed to be optically thin , is @xmath36 , and these lines are assumed to radiate isotropically .
observational support for a very narrow range of @xmath37 , i.e. the ratio of emitted to ionizing continuum photons , is provided by the observation of a linear correlation between balmer line and optical disk luminosity in agn over several orders of magnitude ( e.g. , * ? ? ?
* ; * ? ? ?
this is compatible with photoionization / recombination theory which expects the h - line brightness to be correlated with the ionizing continuum flux as the line luminosity is driven by this ionizing disk continuum . a statistical blazar study of celotti et al .
( 1997 ) implies @xmath38 , which is used in the following . changing the here fixed parameters ( e.g. @xmath37 , @xmath25 , @xmath26 , etc . )
will alter the absolute values of the @xmath0-ray attenuation , however , any redshift - dependence remains unaffected .
the calculation of the @xmath0-ray opacity follows the procedure outlined in donea & protheroe ( 2003 ) for the geometrically thick shell case , but uses a more refined blr line spectrum here .
average blr spectrum of francis et al .
( 1991 ) together with the h@xmath39 line strength reported by gaskell et al .
( 1981 ) sums up to 35 lines , with h@xmath39 and ly@xmath39 being the strongest lines . for the present work the blr spectrum is approximated as a two - component spectrum , @xmath40 , with the total luminosity of all lines at @xmath41 ( @xmath42 ) to be radiated at the strongest optical line ( h@xmath39 ) wavelength , and the total line luminosity at @xmath43 ( @xmath44 ) emitted at the strongest uv - line , ly@xmath39 .
a refined treatment of the blr spectrum is straight forward , but will not alter the results on the redshift - dependence of the local opacity , nor add qualitatively new insights to the subject of the present work .
the calculation of the optical depth @xmath45 with @xmath46 the primary @xmath0-ray photon energy , @xmath11 , @xmath47 , @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath52 takes into account the full angle - dependent cross section ( e.g. , * ? ? ?
* ; * ? ? ?
@xmath53 , @xmath54 are determined by the respective geometry of the target photon field .
the total cross section maximizes at @xmath55 , where @xmath56 ( @xmath57=photon interaction angle ) is the threshold condition of the pair production process .
the very prominent peak of the cross section near threshold reaches roughly @xmath58 , where @xmath59 is the thomson cross section .
the narrowness of the pair production cross section forces over half the interactions to occur in a small target photon energy interval @xmath60 centered on @xmath61ev for a smooth broadband spectrum @xcite .
[ fig1 ] shows the resulting opacity from different distances @xmath62 of the @xmath0-ray production region above the black hole to @xmath63 , in the accretion disk ( red curves ) and blr radiation field ( blue curves ) for typical quasar accretion rates and bh masses .
the two `` bumps '' in the blue opacity curves are the result of absorption in the h@xmath39 and ly@xmath39 lines of the blr , and smooth when a detailed multi(@xmath64)-line spectrum is used . in typical quasar environments
the strength of local absorption is strongly dependent on the location of the emission region with respect to the target photon field . if the @xmath0-ray emission region is located not well beyond the blr , which is mandatory for @xmath0-ray production that involve external photon fields , local @xmath0-ray absorption features in fsrq spectra have to be expected at @xmath65 several tens of gev .
the key step of this work is the application of cosmological black hole and quasar evolution to the expected pair production opacity of @xmath0-ray photons in agn , with direct implications for studies of the evolution of the ebl . with the availability of large agn data archives enormous advances in bh demographics were made .
the cosmic evolution of the bh mass accretion rate has recently been studied by netzer & trakhtenbrot ( 2007 ) on the basis of @xmath66 sdss type - i radio - loud and radio - quiet agn in a large redshift range ( @xmath67 ) .
significantly higher accretion rates at larger redshifts were derived with an eddington ratio of the accretion luminosity @xmath68 with @xmath69 . the intriguing similarity observed of the time history of star formation ( sf ) and bh accretion rate density ( e.g. , * ? ? ?
* ) as well as established relations between bh mass and some bulge properties of the host galaxy lead to the widely accepted picture of a joint evolution of qsos / bhs and their host galaxies ( see also * ? ? ?
* ; * ? ? ?
moreover , the observed bh mass function is consistent with that inferred from quasar luminosities for simple assumptions of the accretion efficiency ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the evolution of their luminosity functions at both , soft and hard x - rays , shows the number density of fainter agn to peak at lower redshifts than that of the brighter ones ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this evidence of downsizing directly leads to the picture of an anti - hierarchical bh growth ( i.e. high - mass bhs grow faster and low - mass bhs grow preferably at lower redshift ) , and has meanwhile been confirmed by multiple studies : an analysis based on the fundamental plane of accreting bhs @xcite ; the phenomenological approach of marconi et al .
( 2004 ) for determining the evolution of the bh mass function using observational constraints from the local bh mass function , the evolving x - ray luminosity functions and energetics from the x - ray background ; determinations of low - luminosity high - redshift quasar luminosity functions by the goods collaboration ( e.g. , * ? ? ?
* ) ; the combo-17 survey at higher luminosities @xcite , just to name a few . the combination of these observational findings with the theoretically advocated hierarchical clustering paradigm ( based on cold dark matter ) led to the awareness of feedback processes working during the process of bh mass growth in agn systems ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
while the parameter space is still large , the advocated evolution of bh growth and thus accretion rates has severe implications for the @xmath0-ray quasar population where local @xmath0-ray absorption in accretion disk and blr radiation fields is potentially important . in the following i use three exemplary models for the evolution of the cosmic bh accretion rate ( fig .
[ fig2 ] ) : a ) the netzer & trakhtenbrot ( 2007 ) analysis , complemented by the lapi et al .
( 2006 ) model for redshifts @xmath70 where only modest evolution is proposed there , b ) the anti - hierarchical bh growth picture of marconi et al .
( 2004 ) , and c ) a non - evolution scenario for comparison
. the evolution of the accretion rate of model ( b ) has been derived by using the average bh growth history as published in marconi et al .
( 2004 ) from redshift @xmath71 ( where @xmath72 source activity has been assumed as an initial condition ) to @xmath73 together with their supported accretion radiation efficiency of 0.1 .
model ( a ) and
( b ) both show strongly redshift - dependent bh accretion rates with higher rates at larger redshifts for bh masses of @xmath74m@xmath75 , typical for quasars . for the present work
the chosen models are used plainly as means and agencies to demonstrate how any evolution of accretion rates transforms into a redshift - dependence of the local optical pair production depth in strong - line quasars .
when studying the evolution of the ebl by means of a statistical analysis of signatures of @xmath0-ray attenuation with effective opacity @xmath76 from extragalactic objects , recognizing and disentangling absorption taking place within the source system ( local opacity ; @xmath77 ) and in the ebl during photon propagation to earth ( ebl - caused opacity ; @xmath78 ) is a crucial task .
if both opacities each depend on redshift in the same direction , but with the exact value of each being unknown , the redshift parameter alone will not be sufficient to extract the evolution of the ebl - caused opacity . in order to assess the probability of this to take place in quasar - like agn , in the following i will apply realistic evolving , as well as non - evolving , cosmic accretion rate curves ( see fig . [ fig2 ] ) to the calculation of the expected opacity from @xmath0-ray absorption in their accretion disk and blr radiation fields , using the procedure outlined in sect . 2 .
any evolution of accretion rates translates into an evolution of the target photon fields under consideration for photon - photon interactions , and thus into a redshift - dependence of the local opacity . for this study the position of the @xmath0-ray production
is fixed to @xmath79pc .
a location of the @xmath0-ray emission region close to the blr is particularly preferred by leptonic as well as hadronic blazar emission models that require external photon fields like accretion disk and blr radiation as an ingredient for @xmath0-ray production through particle - photon interactions .
if @xmath0-ray production is located well beyond the blr , @xmath6-pair production on external photon fields ceases to be an important process , and so will external compton scattering above pair production threshold owing to their comparable cross section values there , or photopion production on accretion disk and blr photons .
the goal of this excercise is to verify a possible _ redshift - dependence _ of the local opacity .
the absolute @xmath80 values depend on the details of the target radiation fields and location of the @xmath0-ray production ( see e.g. * ? ? ?
* ) , and thus come with uncertainties that reflect the dimension of the free parameter space .
[ fig3 ] shows the resulting opacity curves in the observer frame when using the evolutionary behaviour of accretion rates following netzer & trakhtenbrot ( 2007 ) and lapi et al .
( 2006 ) for the target photon fields for @xmath6-interactions .
the fast evolution of accretion rates at low redshifts is apparent .
the corresponding curves for a m@xmath81m@xmath75 are inflated towards higher @xmath82 by more than an order of magnitude , and for @xmath83 gev the photon - photon collisions occur predominantly in the accretion disk radiation field .
energy redshifting is the reason for the threshold and energy of maximum interaction probability , @xmath84tev , to decrease with redshift .
note that for most energies @xmath85tev @xmath0-ray absorption occurs preferentially near the increasing part of the @xmath82 function near pair production threshold .
if the distance @xmath62 of the @xmath0-ray production site from the disk rises beyond the blr , @xmath0-ray attenuation tails off as indicated in fig.[fig1 ] .
an increase in blr size leads to an opacity decrease for an unchanged blr luminosity by an amount that corresponds to the decrease in target photon density with increasing blr volume .
a direct view on the redshift - dependence of the local opacity opens up by slicing fig .
[ fig3 ] at the energies of interest .
[ fig4 ] shows the resulting @xmath86 curves at observer energy 100 gev ( all solid lines ) and 300 gev ( all dashed lines ) for all three evolution pictures for the accretion rate as shown in fig .
[ fig2 ] and for typical quasar bh masses ( m@xmath87m@xmath75 : all red curves , m@xmath81m@xmath75 : all blue curves ) .
the black curves represent the optical depths for the situation of a non - evolving high - accretion rate disk with a high - mass bh , and of a non - evolving low - accretion rate system with a lower - mass bh .
a strong redshift - dependence is apparent in almost all cases . even for the case of non - evolving accretion rates the local opacities
show redshift - dependence , if @xmath6 pair production occurs predominantly close to threshold , with an increasing slope of @xmath86 with redshift .
the reason lies basically in the prominent peak of the pair production cross section , together with cosmological energy red - shifting : the prominent peak in the cross section leads to most photon - photon collisions occuring in a rather narrow energy range @xmath88 near threshold for smooth broadband target photon spectra .
if those source spectra are located at cosmological distances , the energy red - shifting into the observer frame leads to the presented redshift - dependence of @xmath80 .
it is straight forward to then determine the evolution of the e - folding cutoff energy @xmath89 for the series of accretion rate evolution models used so far .
[ fig5 ] shows a fazio - stecker - like presentation for the local absorption and compares with the typically expected behaviour for absorption of @xmath0-rays in the evolving ebl ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
potential ebl probes of glast s lat are located at a redshift @xmath90 . in all cases , and this includes also non - evolving accretion rates , decreases the e - folding cutoff energy with redshift , similar to the fazio - stecker relation for ebl - caused absorption . if local absorption in the external radiation fields of agn leaves measurable imprints in the @xmath0-ray spectra , these will be almost unavoidably redshift - dependent , and remain to be distinguished from the @xmath0-ray opacity of the ebl .
this work is devoted to investigate whether any redshift - dependence of the opacity @xmath80 of @xmath0-rays produced in systems of strong - line blazars can be uniquely attributed to photon absorption in the ebl .
the possible existence of further sources of redshift - dependence of @xmath80 other than caused during propagation in the ebl will lead to ambiguities in estimating the ebl evolution .
strong - line quasars are in general considered as high - luminosity sources in the @xmath0-ray domain , which have a marked probability of providing sufficient photon statistics to search for absorption breaks in a spectral analysis .
their bright external photon fields ( i.e. accretion disk , blr ) are used in many high - energy emission models as a necessary ingredient for @xmath0-ray production , e.g. via inverse compton scattering . if this , however , takes place in the klein - nishina regime , and therefore above the pair production threshold , @xmath0-ray absorption due to photon - photon interactions is of comparable importance . in some hadronic blazar models
the observed @xmath0-ray output is the result of re - distributing the injected nucleon energy via pair cascades that develop in external photon fields , and thus inevitably involves photon - photon pair production there .
the focus of this study is therefore directed to @xmath6-pair production in photon environments ( here specifically accretion disk and blr radiation field ) of @xmath0-ray loud blazars where the interactions of relativistic particles in external photon fields potentially contributes significantly to the @xmath0-ray output above @xmath91 gev . *
if the @xmath0-ray emission region is located not well beyond the blr , mandatory for @xmath0-ray production that involve external photon fields , local @xmath0-ray absorption features in strong - line quasar spectra have to be expected at @xmath92 several tens of gev .
local @xmath0-ray absorption in external agn radiation fields ceases importance if the @xmath0-ray production site is sufficiently distant from the blr .
while then evolutionary studies of the ebl by means of @xmath0-ray absorption signatures are not affected , it will lead to important implications for the high - energy blazar emission models .
* following recent progress in bh demographics , bh growth and corresponding accretion rates turn out to show a redshift - dependence with higher rates at larger redshifts . correspondingly , the critical energy @xmath5 due to local absorption in quasar disk and blr radiation fields decreases with redshift , similar to fazio - stecker s relation for ebl absorption . * even for the case of no evolution of quasar disk accretion rates , @xmath5 decreases with redshift for sources of a given bh mass , at @xmath93 very similar to the fazio - stecker presentation of ebl absorption with the current knowledge of those systems .
it results from the interplay of local absorption near pair production threshold and cosmological energy red - shifting . *
any observed redshift - dependence of absorption features in blazars , that are prone to local @xmath0-ray absorption , can therefore _ not _ serve as a unique signature for absorption occurring in the ebl radiation field .
this complicates approaches for estimating the evolution of the ebl using gev - sensitive instruments , that utilize the fazio - stecker relation or similar methods , and @xmath0-ray agn whose external photon fields are considered important in @xmath0-ray production . as a result , it seems that only
naked @xmath0-ray jet sources
( i.e. agn without noticable optical / uv radiation fields close to the @xmath0-ray emission region ; `` true type-2 agn '' ) are unbiased probes for studies of the evolution of the ebl on the basis of the fazio - stecker relation and using gev instruments like glast , etc .
consequently , an obvious choice for suitable candidate sources for this task would be blazars with particular weak or absent emission lines , generally classified as bl lac objects ( although exceptions exist , see sect . 1 ) .
predictions for the expected number of gev bl lacs range from several hundred @xcite to a few thousand @xcite above the lat sensitivity .
it remains to be seen whether the near future gev instruments will be detecting a sufficient number of suitable sources at @xmath90 to allow a sensible analysis on the basis of naked @xmath0-ray jet sources only .
though the finding of this work adds fundamentally to already recognized complications ( e.g. flux and spectral source variability ) in analysis aiming to probe ebl evolution through the @xmath0-ray horizon , it also offers new options for constraining agn physics .
the relevance of @xmath0-ray absorption for cutting off the sed at the high energy end in the different blazar types , possibly as a consequence of the location of @xmath0-ray production , could be probed by means of a statistical study of @xmath94 as a function of source type or object parameters .
if simultaneously measured emission lines and/or accretion disk signatures indicate the presence of luminous photon fields external to the jet , the position of the @xmath0-ray production site could be constrained .
monitoring both , the time history of the external target photon fields in agn and the jet @xmath0-ray flux , may offer independent verification of the importance of local absorption and extern inverse compton scattering there , and allows conclusions on the @xmath0-ray production location and on some properties of the blr material @xcite . at the same time
, any non - detection of absorption features with sensitive gev instruments puts limits on the external radiation fields in those agn . if local opacity shapes part of the @xmath0-ray loud agn population , its evolution may influence luminosity function and extragalactic @xmath0-ray background contribution of agn above the 50 - 100 gev energy range . in both cases
a redshift - dependence of the local opacity of @xmath0-ray loud quasars will have far - reaching implications .
i d like to thank for valuable feedback from the glast - lat agn science working group , in particular for very useful comments from c. dermer , g. madejski and b. lott .
i also thank f. stecker for providing his model curves and for interesting discussions .
this work is supported by the national aeronautics and space administration under contract nas5 - 00147 with stanford university . | the case of @xmath0-ray absorption due to photon - photon pair production of jet photons in the external photon environment like accretion disk and broad - line region radiation field of @xmath0-ray loud active galactic nuclei ( agn ) that exhibit strong emission lines is considered .
i demonstrate that this local opacity , if detected , will almost unavoidably be redshift - dependent in the sub - tev range .
this introduces non - negligible biases , and complicates approaches for studying the evolution of the extragalactic background light with contemporary gev instruments like e.g. the gamma - ray large area space telescope ( glast ) , etc . , where the @xmath0-ray horizon is probed by means of statistical analysis of absorption features ( e.g. fazio - stecker relation , etc .
) in agn spectra at various redshifts .
it particularly applies to strong - line quasars where external photon fields are potentially involved in @xmath0-ray production . |
one of the central problems in astronomy is that of galaxy formation and evolution : when were the various parts of galaxies assembled , when were the stars formed , and how did these processes depend on environment ? at low redshift the dependence of many galaxy properties on environment is well established .
dense environments contain a larger fraction of red , passive galaxies , while low density environments contain more blue , star - forming galaxies ( e.g. , @xcite ) .
the distributions of colour and h@xmath5 equivalent width , and hence inferred star - formation histories , are found to be bimodal @xcite . surprisingly , the two sequences vary little with environment , except in terms of their relative proportions @xcite .
however , some studies find that star - forming galaxies in clusters tend to have a reduced global star formation rate with respect to field galaxies of the same morphological type @xcite .
the chemical abundances of stars and interstellar gas within galaxies provide a fundamental tool for tracing the evolution of their stellar and metal content .
these abundances depend on various physical processes , such as the star formation history , gas outflows and inflows , etc .
variation in the chemical enrichment of galaxies may potentially provide insight into how environment affects galaxy evolution .
measurements of this variation can thus assist in constraining the likely scenarios of galaxy evolution .
the correlation between galaxy metallicity and luminosity is one of the most significant observational results in chemical evolution studies ( e.g. , @xcite ; @xcite ; @xcite ) .
the analysis of large samples of star - forming galaxies shows that metallicity correlates with luminosity over @xmath6 magnitudes , with a factor of @xmath7 increase in metallicity @xcite .
the metallicity relation is tighter versus stellar mass than luminosity @xcite .
spectrophotometric observations of hii regions for a sample of nine virgo cluster spiral galaxies have shown that hi - deficient objects , near the cluster core , appear to have higher oxygen abundance than field galaxies of comparable luminosity or morphology ( @xcite ) .
oxygen abundances of spirals at the periphery of the cluster with normal hi properties are however indistinguishable from those of field galaxies ( @xcite ; see also @xcite for similar results ) . due to their lower gravitational potentials ,
dwarf galaxies ought to be more sensitive to their surroundings and should especially be less able to retain their gaseous contents . for a sample of dwarf star forming galaxies located at different environments , @xcite found that galaxies located in low density regions tend to display higher excitation - sensitive emission line ratios , with higher emission line equivalent widths and total luminosities .
dwarf galaxies appear however to follow the same metallicity - luminosity relation , i.e. , at a given luminosity , there is no systematic difference in oxygen abundance between cluster and field star - forming dwarf galaxies . @xcite
have confirmed this result for a sample of dwarf irregulars in the virgo cluster .
a subsample of virgo star - forming galaxies display a much lower baryonic gas fraction than their counterparts in the field at comparable oxygen abundances .
they argue that the observed gas - poor star forming galaxies have being stripped of their gaseous content without significant effect on their luminosities and metallicities . the large dispersion in the chemical properties of field galaxies ( see e.g. , @xcite ) , and
previous small sample sizes in dense regions make it difficult to draw definitive conclusions regarding the effect of environment on the chemical properties of galaxies .
the large data sets provided by the sloan digital sky survey ( sdss ) now allow the environmental effects on galaxy properties to be followed statistically over the full range of environments , from the sparse field to dense cluster cores , at least for bright galaxies .
additionally , the homogeneity of the sdss sample enables us to sample the entire range of galactic environments , in a uniform manner .
we can now , therefore , expand the study of galaxy chemical properties from the usual cluster versus field comparison to general environment , in this case characterised by local density , irrespective of actual cluster or group membership .
this will help strongly in constraining the physical processes responsible for determining galaxy properties . in this paper ,
we analyze the variation in oxygen abundance of the interstellar star - forming gas for a large sample of local galaxies as a function of their stellar mass and environment , in order to investigate environmental effects on the chemical content of galaxies , and how it depends on galaxy mass .
the paper is organized as follows . in section [ data ]
we describe the sample selection . in section [ results
] , we investigate the environmental dependence of the stellar mass versus gas - phase oxygen abundance relationship . the implications of our results , and our conclusions are summarized in section [ summary ] . the cosmological model with @xmath8 , @xmath9 , @xmath10 has been adopted throughout the paper .
the sdss is a project with a dedicated 2.5-m telescope designed to image @xmath11 sq .
deg . and
take spectra of @xmath12 objects @xcite .
the imaging covers five broadbands , @xmath13 , with effective wavelengths from 350 nm to 900 nm .
the spectra are obtained using a fibre - fed spectrograph with @xmath14 apertures .
this aperture size corresponds to 3kpc at @xmath15 .
the wavelength range is from 380 to 920 nm and the resolution is @xmath16 .
the majority of spectra are taken of galaxies with @xmath17 , called the main galaxy sample ( mgs ) @xcite .
the sample analysed here was selected from sdss data release four @xcite . *
the initial sample consisted of 151168 mgs galaxies in the redshift range 0.0100.085 .
the sample is volume - limited for galaxies brighter than @xmath18 , which is the density defining population . for magnitudes fainter than this limit , the sample is magnitude - limited . *
data on these galaxies were taken from the standard photo and spectroscopic data reduction pipelines @xcite , along with additional measurements from the mpa garching dr4 data @xcite and environmental measurements @xcite .
environmental density measurements were determined using an @xmath19th nearest neighbour technique .
galaxies more luminous than @xmath20 were used for the density defining population ( ddp ) : with an average space density of @xmath21 . the environmental density for each galaxy
is then given by @xmath22 where @xmath23 and @xmath24 are the projected distances to the 4th and 5th nearest ddp neighbours within @xmath25 .
for typical galaxies , @xmath26 ranges from @xmath27 , typical of void - like regions , to @xmath28 , typical of the centres of galaxy clusters .
a best - estimate @xmath26 , and minimum and maximum values , were determined taking account of edge effects and galaxies that had not been observed spectroscopically .
for example , @xmath29 was determined by using the smaller of ( i ) the distance to the nearest edge and ( ii ) the distance to the 4th/5th nearest photometrically confirmed neighbours .
see @xcite for details . from this sample , only galaxies for which emission lines are well detected were retained .
we require galaxies included in our star - forming sample to have lines of , , , , and detected at greater than @xmath30 , reducing the sample to 68603 .
galaxies with @xmath31 are excluded from the sample due to the blue wavelength cutoff of the spectrograph .
we have distinguished star - forming galaxies from non - thermal sources , such as active galactic nuclei ( agn ) , using the classical diagnostic ratios of two pairs of relatively strong emission lines , i.e. , / vs. / diagrams @xcite .
we have used the empirical demarcation line , separating star - forming galaxies from agn , provided by @xcite .
the selection of star - forming galaxies reduced the sample to 51934 galaxies . in order to avoid edge effects , galaxies with @xmath32
were excluded from the analysis .
this reduced the sample to 45107 galaxies . for fibre - fed spectroscopy , the fraction of the light of a galaxy captured within the fiber will depend on its redshift , intrinsic size , and surface brightness profile , as well as the size of the fibre and seeing during the observation . @xcite
have shown that using the spectra of the inner parts of galaxies could have a substantial systematic effects on the estimate of their global properties . to reduce systematic and random errors from aperture effects
, they recommended selecting sdss galaxies with @xmath33 . doing so , reduces the size of the sample to 37866 galaxies .
gas - phase oxygen abundances are taken from the catalogues of derived physical properties for dr4 galaxies released by the mpa garching group @xcite . briefly , gas - phase oxygen abundances are estimated based on simultaneous fits of all the most prominent emission line fluxes ( , , , , , ) with a model designed for the interpretation of integrated galaxy spectra based on a combination of population synthesis and photoionization codes @xcite .
the likelihood distribution of the metallicity of each galaxy in the sample is calculated , based on comparisons with a large library of models corresponding to different assumptions about the effective interstellar star - forming gas parameters . *
the median of this distribution is adopted as the best estimate of the galaxy metallicity .
the estimated metallicities show good agreement with those estimated using the so - called strong emission line method .
the width of the likelihood distribution provides a measure of the error .
the median @xmath34 error for our final sample of star - forming galaxies is @xmath35 dex . * * the computation of stellar masses by the mpa garching group is obtained by fitting a grid of population synthesis models , including bursts , to the spectral features d4000 and h@xmath36 absorption .
the predicted colours are then compared with broad - band photometry to estimate dust attenuation .
stellar mass - to - light ratios are determined and applied to the petrosian @xmath37-band magnitude .
the nominal @xmath34 random errors are typically @xmath38 dex . for details
see @xcite . assuming most galaxies have stellar mass estimates with relative uncertainties less than 0.2 dex , these have little impact on our results . *
* selecting only galaxies with high @xmath39 emission lines could potentially bias the distribution of galaxy properties in the selected final sample in such a way to affect the intrinsic correlations between galaxy properties that we aim to investigate .
therefore it is important to assess to what extent the final sample of star - forming galaxies covers similar regions in the parameter space as star - forming galaxies in original sample . to ensure that the final sample of star - forming galaxies is representative of the parent sample , we compare the distribution of galaxy properties in the final sample with those of a more lenient selection .
these were selected from the initial mgs sample , as those with @xmath39 larger than 3 in emission lines needed to classify the ionizing source , i.e. , , , , and . *
[ selection ] shows the distributions of stellar mass , star formation rate , abundance - sensitive diagnostic ratio @xmath40\lambda4959,\lambda5007}$}}}+{{\mbox{${\rm [ oii]\lambda3727}$}}})/{{\mbox{${\rm h\beta}$}}}$ ] , and excitation - sensitive diagnostic ratio @xmath41\lambda4959,\lambda5007}$}}}/{{\mbox{${\rm [ oii]\lambda3727}$}}}$ ] for both the final and the parent samples of star - forming galaxies .
the distributions of the properties of the star - forming galaxies in the parent sample are shown in the upper panels , and those of the final sample are shown in the bottom panels .
solid lines show the distribution of the properties of galaxies situated in rarefied field regions , i.e. , @xmath42 , and dotted lines show the distribution of the properties of those situated in dense region , i.e. , @xmath43 .
the figure shows clearly that the final sample has galaxy parameter distributions similar to those of the parent star - forming galaxy sample .
most importantly , the differences with respect to the environment are the same for the two different selections . observed abundance- and excitation - sensitive diagnostic ratios
are distributed similarly in both samples .
however the redshift cut , imposed to minimize the effects of the aperture bias , appears to slightly affect the extent of the low end tails of stellar mass and star formation rate distributions , i.e. , the low ends of both distributions for the final sample are less extended than for the parent star - forming galaxy sample .
the cut on redshift excludes low stellar mass galaxies , as those galaxies tend to be at low redshift end .
those galaxies represent however a small fraction of the sample .
the high ends of the star formation rate and the stellar mass distributions are however unaffected by the redshift cut .
the similar distributions of both star - forming galaxy samples suggest that our final sample is a fair representation of the parent sample in terms of its stellar mass , star formation activity , interstellar gas - phase properties .
the correlations we aim to investigate are therefore expected to not be severely biased by the selection procedure of the final sample of star - forming galaxies . * the distributions of global properties for our final sample , i.e. , @xmath44-band
absolute magnitudes , specific star formation rate , local galaxy density , and the concentration index , defined as the ratio of the radii enclosing 50% ( @xmath45 ) and 90% ( @xmath46 ) of the petrosian @xmath44-band galaxy light , are shown in fig .
[ prop_dist ] .
[ cmr_sel ] shows the relationship between the integrated @xmath47 colour and the stellar mass for six local galaxy density bins ranging from @xmath48 to @xmath49 ( see below ) .
the distribution of the final sample of galaxies is shown as solid contours and points .
the full parent sample of galaxies in the redshift range 0.0100.085 is shown as the dashed - dotted contours .
the bimodal nature of the colour - stellar mass relation for the complete sample is clear : red galaxies dominate in dense regions and at higher masses , while the dominance of blue galaxies is clear at lower masses and for lower local densities ( see @xcite for more details ) .
the selected galaxies are distributed along the blue sequence .
this is no surprise , as star - forming galaxies with well detected emission - lines tend to have blue colours .
the selected sample of star - forming galaxies was divided into bins of local galaxy density and stellar mass in order to determine the variation of gas - phase oxygen abundances as a function of these quantities .
figure [ ohdist ] shows the gas - phase oxygen abundance distribution of galaxies in bins of galaxy local density and stellar mass .
the median local galaxy density in the selected sample is @xmath50 ; thus , the lowest local galaxy density bin , with a median density of @xmath51 is underdense by a factor of @xmath52 , while the densest bin with a median density of @xmath53 , corresponds to the typical galaxy density found in the inner regions of galaxy clusters .
our selection function preferentially selects morphologically late - type galaxies .
however because the strongest effect with environment is the early / late - type galaxy fraction ; we split the sample into nominally early and late - type galaxies .
the ( inverse ) concentration index , is used as a proxy for galaxy morphology : with @xmath54 marking the boundary between early and late types @xcite .
@xcite have found that a concentration index around @xmath55 is a natural dividing line between red and blue galaxy populations for all stellar masses .
@xcite also show that galaxies naturally divide into two populations in the concentration
colour plane , which can be associated with late- and early - types .
solid ( dotted ) line histograms in fig .
[ ohdist ] represent gas - phase oxygen abundance distributions for galaxies with concentration index lower ( higher ) than 0.4 .
the distributions of gas - phase oxygen abundances of star - forming galaxies with concentration indices larger ( late - type population ) and lower ( early - type population ) than the dividing value are indistinguishable for all environments . in all local galaxy density bins ,
the median of the gas - phase oxygen abundance distribution increases with increasing stellar mass , i.e. , the so - called stellar mass metallicity relation is present in all environments .
the dispersion of the gas - phase oxygen abundance distribution increases at lower stellar mass , where the fraction of star - forming galaxies increases ( e.g. @xcite ) , but appears to be insensitive to environment .
most remarkable is the mass dependence of the environmental effect on galaxy gas - phase oxygen abundance . for massive galaxies , i.e. , stellar masses larger than @xmath56
, the oxygen abundance distribution median does not show any significant dependence on the environment .
however , for low mass galaxies ( @xmath57 ) , the abundance distribution median increases by 0.060.08 dex from low density environments , i.e. , @xmath58 , to highly dense regions , i.e. , @xmath59 . to determine the variation of the stellar mass versus gas - phase oxygen abundance relation as a function of the environment , we have divided the star - forming galaxy sample into local galaxy density bins .
figure [ ms_oh_env ] shows the relationship for six local density bins ranging from @xmath48 to @xmath49 . in each panel ,
the large filled circles show the median gas - phase oxygen abundance in bins of 0.15 dex in stellar mass containing more than @xmath60 galaxies , and the thick line shows a polynomial fit to these points .
the thin solid line shows the fit to the stellar mass versus oxygen abundance for galaxies in low density regions , i.e. , @xmath48 ( only visible in the higher density bins ) .
the correlation is relatively steep from @xmath61 to @xmath62 , but flattens for more massive galaxies ( as discussed by @xcite ) .
most striking of all is the similarity of the correlation between stellar mass and gas - phase oxygen abundance over a large range of local galaxy densities , from @xmath48 , typical of void - like environments , to @xmath3 , typical of the outer regions of galaxy clusters .
this is in agreement with previous findings , that star - forming galaxies in the periphery of the virgo cluster exhibit similar oxygen abundances to those of field galaxies with similar luminosities @xcite . the spread of oxygen abundances about the median is identical at all environments across the entire galaxy stellar mass range . the bottom right panel of fig .
[ ms_oh_env ] shows , however , that stellar mass versus oxygen abundance relationship at denser environments , i.e. , @xmath49 typical of the inner parts of galaxy clusters , is shallower than the relationship for galaxies in less dense regions .
gas - phase oxygen abundances show a mass - dependent shift to higher values that is larger for low stellar mass galaxies .
this seems to disagree with the results of @xcite who found that the interstellar star - forming gas of bright galaxies near the core of the virgo galaxy cluster appear to be overabundant by 0.20.3 dex in comparison with field galaxies with similar luminosities .
@xcite have confirmed the higher abundances for the same virgo galaxies studied by @xcite , although they have found counterparts for both the periphery and core cluster objects among field star - forming galaxies . they have concluded that if there is a difference in the abundance properties of the virgo and field bright spirals , this difference appears to be small , in agreement with our results .
figure [ oh_density ] shows the variation of gas - phase oxygen abundance as a function of local galaxy density for different stellar mass bins . over a factor @xmath63 in local galaxy density the median gas - phase oxygen abundance changes by only @xmath64 dex , depending on stellar mass , being greater at the low stellar mass end .
these trends are weak , relative to the stellar mass dependence : the median gas - phase oxygen abundance is @xmath65 dex higher in the massive galaxies , compared with the low mass galaxies in the same environment .
figure [ nsfr_density ] shows the relationship between the specific star formation rate and the stellar mass for galaxies in dense environments , @xmath49 .
we use the star formation rates corrected to total computed by @xcite . note that this sample only includes galaxies which are actively star forming , due to our requirement on significant emission - line detections .
however , it still meaningful to consider star - forming galaxies alone in this diagram , due to the bimodality in specific star formation rates , e.g. , as seen by @xcite considering the h@xmath5 equivalent width distribution .
as in fig .
[ ms_oh_env ] , the thick line shows a polynomial fit to the variation of the specific star formation rate median , shown as large field circles , as a function of the stellar mass .
the thin solid line shows the fit to the same correlation for galaxies situated in low density regions , i.e. , @xmath48 .
similar to the stellar mass versus oxygen abundance relation , the correlation between the specific star formation rate and the stellar mass of star - forming galaxies shows very little dependence on local galaxy density from void - like environments , i.e. , @xmath66 , to the periphery of galaxy clusters , i.e. , @xmath3 . for star - forming galaxies in dense regions ,
i.e. , @xmath49 , the median of the specific star formation rate distribution is , however , @xmath67 dex higher than for galaxies with similar stellar masses in less dense regions .
using a complete sample of 151168 galaxies in the redshift range of @xmath68 from the sdss dr4 , we draw a sample of 37866 star - forming galaxies in the the redshift range of @xmath69 to examine how oxygen abundances of the interstellar star - forming gas relate to environment .
we use the projected distance to the fourth and fifth nearest neighbours as an indicator of local galaxy density .
the stellar mass versus gas - phase oxygen abundance relationship appears to be environment - free over a large range of local galaxy densities . for galaxies located in environments with local galaxy densities typical of the inner regions of galaxy clusters ,
the interstellar star - forming gas is found to be slightly overabundant compared to what is found for their counterparts of similar stellar masses in less dense environments .
this change in oxygen abundance increases as galaxy stellar mass decreases , ranging from @xmath70 dex for galaxies with stellar mass larger than @xmath71 to @xmath72 dex for galaxies with stellar mass around @xmath73 . over that last few years
, observational evidence has been accumulating that the dominant evolution in the galaxy population properties as a function of environment is the change of the ratio between star - forming / blue galaxies and passive / red galaxies ( e.g. @xcite ; @xcite and reference therein ) .
the environmental dependence of different galaxy properties , i.e. , colour , concentration index , colour gradient , are found to be almost entirely due to the dependences of galaxy morphology and luminosity on the environment : when morphology and luminosity are fixed , galaxy physical properties are nearly independent of local galaxy density @xcite .
the scaling relations for both passive and star - forming galaxies are virtually unaffected by environment . studies of the fundamental plane have shown that the properties of early - type galaxies are nearly independent of environment ( e.g. , @xcite ) .
the scaling relations for star - forming galaxies are virtually unchanged with environment , whilst over the same density range the fraction of star - forming galaxies is changing strongly ( e.g. @xcite ) .
our finding that oxygen abundances of galaxies with a given stellar mass depend weakly on environment is consistent with ( i ) the weak dependence on environment of the mean colour , at a given luminosity , of blue galaxies that are still actively growing and evolving , compared with the dependence on luminosity , the strongest effect also being seen for faint galaxies @xcite , and ( ii ) the lack of any environmental effect on the distribution of the emission line equivalent width for star - forming galaxies @xcite .
the weak dependence of the star - forming gas oxygen abundance shown here is a further suggestion that the primary physical driver(s ) of galaxy evolution must depend primarily on galaxy intrinsic properties rather than on their environment .
in addition , whatever mechanism is responsible for the increasing fraction of passive galaxies in dense environments , it must truncate their star formation on a short timescale to avoid affecting the relations of oxygen abundance and specific star - formation versus stellar mass for actively star - forming galaxies .
yet @xcite have argued , based on the absence of dependence on environment of the correlations between different spectral indicators that probe star formation history on different timescales ( @xmath74 ) , that the decrease in star formation activity in dense environments occurs over long ( @xmath75 ) timescales .
it is not clear however how slow quenching could affect the chemical properties of galaxies , and for which kind of star formation histories galaxies would move along the observed relations in different environments .
self - consistent modelling is clearly needed in order to set tight constraints on quenching timescales from observed chemical and spectro - photometric properties of galaxies .
this is obviously well beyond the scope of this paper , and we will leave it for future investigations .
the modest change of oxygen abundances of star - forming galaxies in regions with local galaxy densities typical of the inner regions of galaxy clusters indicates however that the chemical evolution of galaxies is moderately modulated by the environment .
furthermore , the mechanism responsible for this change must be more efficient in low stellar mass galaxies , as the change of oxygen abundances is larger for those galaxies .
the relationship between stellar mass and oxygen abundance can be understood either as a sequence in astration , i.e. , more massive galaxies are able to convert a larger fraction of their gas into stars than low mass counterparts , or as a depletion sequence , i.e. , the efficiency of galaxies to retain their gas increases with galaxy mass . based on the correlation between effective yield and baryonic mass for local star - forming galaxies , @xcite have argued that the most straightforward interpretation of the correlation is the selective loss of metals from galaxies with shallow potential wells via galactic winds .
this suggests that the change of 0.020.08 dex in the mean gas - phase oxygen abundance of star - forming galaxies as a function of environment can be explained by a reduced effectiveness of galactic winds in removing metals from the potential wells of low mass halos in dense regions .
there is consequently more gas left to be turned into stars .
massive galaxies can however retain all their gas whether they are in the field or in a cluster .
in this scenario , therefore , the efficiency of star formation within low mass halos is higher in the presence of a confining intergalactic medium , because supernovae are less efficient at removing gas from a galaxy if there is hot material surrounding the galaxy .
an alternative explanation of the observed mass - dependent change in oxygen abundances of galaxies as a function of environment might be that low stellar mass galaxies have assembled their stellar contents at a faster rate in denser regions .
whatever triggers star formation in galaxy clusters , it causes it to happen rapidly , hence increasing the efficiency of recycling during star formation .
environment - driven processes , i.e. , tidal interactions , mergers , could enhance the star formation activity in low mass galaxies in cluster environment
. the observed systematic difference of specific star formation rate between field and cluster galaxies may be the signature of such an effect . | we use a @xmath0 sample of @xmath1 star - forming galaxies from the fourth data release of the sloan digital sky survey to investigate the dependence of gas - phase chemical properties on stellar mass and environment .
the local density , determined from the projected distances to the fourth and fifth nearest neighbours , is used as an environment indicator .
considering environments ranging from voids , i.e. , @xmath2 , to the periphery of galaxy clusters , i.e. , @xmath3 , we find no dependence of the relationship between galaxy stellar mass and gas - phase oxygen abundance , along with its associated scatter , on local galaxy density .
however , the star - forming gas in galaxies shows a marginal increase in the chemical enrichment level at a fixed stellar mass in denser environments .
compared with galaxies of similar stellar mass in low density environments , they are enhanced by a few per cent for massive galaxies to about 20 per cent for galaxies with stellar masses @xmath4 .
these results imply that the evolution of star - forming galaxies is driven primarily by their intrinsic properties and is largely independent of their environment over a large range of local galaxy density .
[ firstpage ] galaxies : evolution galaxies : abundances galaxies : fundamental parameters galaxies : clusters : general |
a quantum particle in a three - dimensional disordered lattice is either localized or diffusive ; this depends on the strength of the disorder @xcite .
if the particles are localized within a finite localization length @xmath2 , one observes current over length scales @xmath3 , while no conductivity over length scales @xmath4 @xcite .
if quantum transport is allowed over infinite length scales , particles are naturally assumed to diffuse over any arbitrary length scale . as was pointed out already in the original work of anderson @xcite , quantum particles with long - range hopping are diffusive in three - dimensional ( 3d ) lattices of infinite size , for any disorder strength .
this absence of localization was later explained by levitov @xcite , who showed that diffusion occurs due to resonant transitions between distant lattice sites .
the number of such resonances diverges with the lattice size , leading to particle transport over infinite length scales .
however , if the lattice size is finite , the number of resonances is also finite .
this raises the question : if the lattice size is reduced , can the system localize to length scales smaller than the lattice size or will the localization length always remain larger than the system size ? to answer this question , we consider particles with dipolar hopping in 3d disordered lattices .
dipolar hopping is long - range in 3d . for generality , we consider lattices with random on - site energies and random dilution .
we show that as the lattice size is _ reduced _ , delocalized particles may , depending on the dilution and amount of disorder , become localized to length scales much smaller than the lattice size .
this leads to a rather unusual phenomenon : particles can diffuse through a lattice of size @xmath0 , but not through a lattice of size @xmath1 .
this means that quantum currents can be switched off by shrinking the size of the system . for conventional disordered systems ,
the localization is typically characterized by computing certain properties as functions of the lattice size .
for example , one can examine the distribution of nearest spacings of the hamiltonian eigenvalues , which must undergo a change from the wigner distribution to a poisson distribution near the diffusion localization crossover . this change becomes sharper as the lattice size increases , corresponding to a phase transition in the infinite size limit . for particles with dipolar long - range hopping in 3d lattices ,
such conventional analyses are not possible as the parameters of interest are functions of the system size itself and the localization length diverges with the system size , in contrast to the case with nearest neighbour hopping . therefore , in order to examine the properties of particles with long - range hopping in finite - size lattices , we use the following approach : * first , we compute the wavepacket expansion dynamics of particles placed in an individual site of the lattice to show that , for certain system parameters , a particle placed in the middle of the lattice does not reach the lattice edge over long times greatly exceeding the heisenberg time ; * second , we compute the inverse participation ratio for each eigenstate of the hamiltonian to illustrate that , for certain system parameters , all eigenstates are localized to parts of the lattice much smaller than the lattice size ; * third , we examine the fluctuations of the wavefunctions between instances of disorder to illustrate that these fluctuations become log - normal at strong disorder ; * fourth , we map out the localization - diffusion crossover region as a function of dilution and on - site disorder using the wavepacket dynamics described in ( i ) ; * finally , we combine our results with scaling theory from refs .
@xcite in order to obtain the dependence of the localization - diffusion crossover region on the lattice size
@xcite were originally introduced to explain delocalization of particles with long - range hopping .
we consider particles with dipolar hopping in 3d disordered lattices , known to be diffusive over infinite length scales .
we start by illustrating the absence of diffusion for such particles in finite - size lattices by computing the time evolution of wavepackets placed on an individual lattice site . for generality ,
we compute the wavepacket dynamics in a diluted , disordered lattice and characterize the localization - diffusion phase diagram as a function of both dilution and disorder strength , as was done for particles with short - range hopping in ref . @xcite .
the dynamics are governed by the following hamiltonian : @xmath5 where the operator @xmath6 removes the particle from site @xmath7 , @xmath8 is the energy of the particle in site @xmath7 , and @xmath9 is the amplitude for particle tunnelling from site @xmath10 to site @xmath7 .
we introduce disorder by randomizing both the values of @xmath8 and @xmath9 , which makes eq .
( [ model ] ) relevant for both disordered lattices and amorphous systems .
we consider a single particle in a 3d cubic lattice with @xmath11 sites per dimension .
we randomize the values @xmath12 $ ] drawn from a uniform distribution and the values @xmath9 as in the site percolation model . for a given realization of disorder
, we divide the lattice sites into two subsets @xmath13 and @xmath14 , with @xmath15 sites in the @xmath13 subset and @xmath16 in the @xmath14 subset . for a given value @xmath17 ,
the lattice sites are assigned to the subsets at random .
the hopping amplitudes of the hamiltonian ( [ model ] ) are then defined as follows : @xmath18 where @xmath19 is the distance between sites @xmath7 and @xmath10 and @xmath20 determines the hopping range , with @xmath21 representing dipolar hopping .
the value of @xmath22 is chosen such that @xmath23 for nearest neighbour ( nn ) sites . with @xmath9
thus defined , the hamiltonian ( [ model ] ) describes a particle with dipolar hopping in a disordered , diluted lattice with @xmath24 sites .
the system parameters of interest are @xmath17 , @xmath25 , and @xmath11 . for a particle initially at the centre site .
@xmath26 is the diameter of a sphere containing no less than 90% of the particle density .
the particle has @xmath27 and is in a disordered lattice with @xmath28 , @xmath29 and @xmath30 ( green , blue , and magenta , respectively ) .
the horizontal dashed line shows the side - length of the cubic lattice ; the grey area highlights sizes smaller than this .
the inset shows the same observables , but over shorter time scales .
the values of @xmath26 are averaged over 1050 realizations of disorder .
the heisenberg time for these systems is @xmath31 , as can also be seen from the inset .
, title="fig : " ] + we compute the wavepacket dynamics for a particle placed at the center of the cubic lattice . as described in ref
@xcite , we use the eigenstates of the hamiltonian ( [ model ] ) to construct the time propagator and compute the wavefunction of the particle at extremely long times @xmath32 , far exceeding the heisenberg time ( typically @xmath31 for the systems considered here ) . to characterize the spread of the wavepacket in real space , we compute the diameter @xmath26 of the sphere enclosing at least 90% of the particle density as a function of time .
the values of @xmath26 are averaged over 200 to 1050 realizations of disorder , as required for convergence and as specified in the figure captions .
figure [ figurewavepackdyn ] shows the dependence of @xmath26 on time for @xmath33 and for different values of @xmath25 .
as expected , for systems with low values of @xmath25 , the particle density expands to fill the entire lattice .
however , for lattices with @xmath34 and @xmath35 , the particle density , on average , does not reach the lattice size and remains localized on a scale smaller than the lattice size . ) computed for a point with strong disorder ( @xmath36 ; @xmath37 ) and a point with weak disorder ( @xmath38 ; @xmath39 ; @xmath40 ) .
the ipr distributions for the point with strong disorder are computed for three lattice sizes with @xmath41 , @xmath42 and @xmath43 , illustrating a slow shift towards the diffusion regime recovered in the infinite size limit ( see figure [ figurescaling ] ) .
the distributions are averaged over 1300 ( @xmath41 ) and 300 ( @xmath44 and @xmath45 ) instances of disorder . ] for @xmath28 , @xmath29 , @xmath46 computed at time @xmath47 .
the different colours refer to histograms at different values of @xmath48 .
the values of @xmath2 and @xmath49 are obtained by scaling the computed histograms using the formula shown on the abscissa .
@xmath2 corresponds to the localization length .
the red curve shows @xmath50 , with @xmath51 the abscissa .
the results are obtained using 3150 realizations of disorder . ] for a particle with @xmath52 initially placed in the middle of a 3d lattice with @xmath53 after @xmath54 .
the circles indicate the computed values , averaged over 200 to 600 realizations of disorder , as required for convergence . for points in the black area , @xmath55 , indicating the absence of transport .
the white area corresponds to @xmath56 , indicating diffusion to the lattice edges in finite time .
the grey area shows the crossover regime , where some values of @xmath57 allow diffusion to the lattice edges and some not .
the dashed line in the lower panel is the @xmath53 line from figure [ figurescaling ] , determined as described in the text . ]
the results of figure [ figurewavepackdyn ] illustrate that lattices with large @xmath25 do not permit diffusion of particles with dipolar hopping over long time scales .
the slopes of the curves in figure [ figurewavepackdyn ] are zero ( within the numerical uncertainties ) , which indicates the absence of diffusion over infinite time .
however , a numerical calculation of wavepacket dynamics can not strictly prove the absence of diffusion at infinite time . in order to confirm localization and
gain a better understanding of the eigenstates of the system , we compute the inverse participation ratio ( ipr ) distribution of the entire set of the eigenstates of the hamiltonian ( [ model ] ) for two different dilutions and disorder strengths and at different system sizes . figure [ figureipr ] shows these ipr distributions for a point with weak disorder ( @xmath38 ; @xmath39 ; @xmath45 ) and strong disorder ( @xmath36 ; @xmath37 ; @xmath58 ) . for states delocalized over the entire lattice
, the ipr tends to @xmath59 . for states localized on a single lattice site ,
the ipr is one .
figure [ figureipr ] shows that for weak disorder the ipr is narrowly distributed near zero , meaning that nearly all of the eigenstates are delocalized over most of the lattice .
the ipr for strong disorder is markedly different , with all states having a much higher ipr . to identify the point with strong disorder as in a localized regime
, it is important to prove that the distributions have absolutely no states with the ipr close to @xmath60 . to do this
, we analyzed the distributions in figure [ figureipr ] for each instance of disorder separately , to compute the minimum value of the ipr , @xmath61 , which represents the _ most extended _ eigenstate .
we obtained the following results : of 1300 disorders for @xmath41 , the smallest ratio @xmath62 ; of 300 disorders for @xmath44 , @xmath63 ; of 300 disorders for @xmath45 , @xmath64 .
this means that the _ most extended _ eigenstate of many disorder realizations is confined to less than one tenth of all lattice sites , proving that all eigenstates are confined to a volume much smaller than the lattice size . for comparison ,
the same ratio for the diffusive regime distribution in figure [ figureipr ] is @xmath65 . as the lattice size increases , the ipr distributions shift toward zero .
these shifts should collapse the ipr distributions to a single line at zero in the limit of infinite size .
however , figure [ figureipr ] shows that the approach of the distributions to that in the diffusive regime is very slow , suggesting that a localized regime could be expected to exist even for systems of large size .
while the results of the previous section illustrate that all eigenstates of the strongly disordered hamiltonians are confined to a fraction of the lattice sites , the ipr values do not provide information on the shape of the wavefunctions . here , we demonstrate that the wavefunctions are exponentially localized by computing the fluctuations of the particle density between instances of disorder .
we also determine the corresponding localization length .
one of the characteristic features of anderson localization is the log - normal distribution of the particle density across disorder realizations @xcite .
for example , as shown in section 2 of ref .
@xcite , the transmission @xmath66 through a disordered region in the regime of anderson localization follows a log - normal distribution , in contrast to the diffusive case where @xmath66 follows a normal distribution . in the regime of anderson localization
, @xmath67 thus follows a normal distribution .
since @xmath67 is proportional to the extinction coefficient @xmath68 , the extinction coefficient also follows a normal distribution .
the average of @xmath68 over instances of disorder is @xmath69 ( see eqn
. 21 of ref .
@xcite ) , where @xmath70 is the system length and @xmath2 is the localization length . the typical amplitude of an anderson - localized wavefunction @xmath71 , where @xmath51 is distance . for a specific instance of disorder ,
an anderson - localized wavefunction would then be @xmath72 .
thus , at a specific value of @xmath51 , the magnitude of the wavefunction will vary log - normally between instances of disorder .
this log - normal distribution is equivalent to the exponential decay of the wavefunction in real space .
we plot in figure [ figurelognormalfluct ] the histograms of @xmath73 for the lattice with @xmath28 , @xmath29 , and @xmath46 .
@xmath74 is obtained from wavepacket dynamics calculations at time @xmath47 .
the histograms for different values of @xmath48 ( distance from the site of maximum density ) are scaled to lie on the same curve .
the scaling is performed using the formula displayed on the abscissa of figure [ figurelognormalfluct ] .
the parameter @xmath2 corresponds to the localization length .
the values obtained are displayed in figure [ figurelognormalfluct ] .
figure [ figurelognormalfluct ] shows that these scaled histograms all follow a gaussian distribution , illustrating the _ log - normal _ distribution of the particle density vs disorder realizations : @xmath75}.\end{aligned}\ ] ] this log - normal behaviour shows that particles with long - range hopping undergo exponential decay in finite - size systems .
exponential decay is one key feature of anderson localization .
moreover , the data in figure [ figurelognormalfluct ] show that the localization length ( see @xmath2 in figure [ figurelognormalfluct ] ) is much smaller than the side - length of the lattice ( @xmath76 ) .
to understand the dependence of the diffusion / localization properties on the dilution and the on - site disorder , we repeat the wavepacket dynamics calculations of figure [ figurewavepackdyn ] for different values of @xmath17 and @xmath25 . the results are presented in figure [ figurephasediagram ] .
figure [ figurephasediagram ] shows two regimes : the black area corresponds to what would be observed in any experiment as the localization regime ; the white area corresponds to the diffusive regime . to obtain these results , we first fix the lattice size and determine by a large number of computations whether a particle initially in the middle of the 3d lattice spreads to fill the entire lattice or forms a localized wavepacket that , after the long time @xmath77 , stays , on average , within a sphere of size @xmath26 smaller than the lattice size .
we would like to emphasize that figure [ figurephasediagram ] is _ not _ a phase diagram as phase diagrams correspond to the thermodynamic limit and the present results are for a finite - size lattice .
similarly , what we identify is a localization - diffusion _ crossover _ , not a localization - diffusion _
transition_. the area of figure [ figurephasediagram ] coloured in black corresponds to disorders where the average size @xmath26 of the wavepacket is smaller than the lattice size ; i.e. , where , on average , there is no diffusion to the lattice edges over the long time @xmath77 .
we note that the absence of localization along the horizontal axis of figure [ figurephasediagram ] is consistent with the recent work of ref .
@xcite that shows no localization for a 3d system with dipolar hopping on a diluted lattice and no on - site disorder ( @xmath78 ) . clearly ,
the off - diagonal disorder caused by the diluted lattice is insufficient to cause localization ; diagonal disorder is required .
this also illustrates the absence of either a quantum or classical percolation threshold for a system of this size . .
the solid line corresponds to the dashed line in figure [ figurephasediagram ] , determined as discussed in the text . ]
it is important to analyze how figure [ figurephasediagram ] changes with the size of the lattice . to do this , we combine the results of wavepacket calculations in figure [ figurephasediagram ] with the scaling analysis of levitov @xcite . from ref .
@xcite , delocalization occurs through resonances defined by @xmath79 .
the number of resonances diverges with the system size and delocalization must occur when @xmath80 where @xmath81 and @xmath82 is the index of a concentric sphere with radius @xmath83 .
re - written as @xmath84 , where @xmath85 is some large constant , eq .
( [ condition ] ) produces a line in the @xmath86-phase diagram separating the diffusive and localized regimes : @xmath87 where we assumed that @xmath88 and @xmath1 and @xmath89 are unknown constants . to determine @xmath1 and @xmath89
, we first note that the probability distribution of the site energy differences is @xmath90,\ ] ] and calculate the probability of resonance between sites @xmath7 and @xmath10 as @xmath91 using eq .
( [ probability - resonance ] ) , we calculate the probability @xmath92 that the central site has _ no _ resonance in the entire system of a given size and draw the isoprobability lines for @xmath93 on the diagram from figure [ figurephasediagram ] .
the isoprobability line with @xmath94 goes through the localization - diffusion crossover diagram inside the grey crossover regime , as shown in figure [ figurephasediagram ] . using this value of @xmath92 as characterizing the diffusion - localization crossover
, we obtain the crossover lines for @xmath95 and fit eq .
( [ crossover - line ] ) to the resulting curves to obtain @xmath1 and @xmath89 .
to verify the procedure , we compute the isoprobability lines for @xmath96 and @xmath97 , and obtain a perfect agreement with the values from eq .
( [ crossover - line ] ) .
figure [ figurescaling ] shows the crossover lines for different lattice sizes .
remarkably , figure [ figurescaling ] illustrates that the localization crossover may occur at finite values of @xmath25 for systems with a macroscopically large size .
consider , for example , the case with @xmath98 , which corresponds to a 3d lattice with @xmath99 sites , which exceeds avogadro s number .
in the infinite size limit , the crossover line tends to the @xmath100-axis , which corresponds to the absence of localization and is consistent with prior work @xcite .
our results illustrate that the detection of the crossover for particles with dipolar hopping in 3d lattices is within reach of current experiments .
this can be done with ions in rf - traps , ultracold magnetic atoms , ultracold rydberg atoms , or ultracold molecules trapped in optical lattices .
an optical lattice typically contains @xmath101 sites per dimension .
localization can be observed with internal excitations of the trapped atoms or molecules @xcite .
if an optical lattice is partially populated with atoms or molecules , the empty lattice sites serve as impurities that scatter the excitations . the parameter @xmath17 in the diagram shown in figure [ figurephasediagram ]
can thus be controlled by simply varying the number of atoms or molecules loaded in the optical lattice .
the disorder in @xmath8 can be applied using an optical speckle potential , as in @xcite .
for example , figure [ figurescaling ] shows that for molecules on an optical lattice with @xmath102 and a lattice population of 30 % @xcite , the crossover can be observed by varying the optical speckle potential from below to above @xmath103 .
our results have , at least , three major applications .
first , they can be used for the design of miniaturized quantum devices relying on the transport of small excitons that move via dipolar hopping .
our results show how much disorder can be tolerated in such devices and what is the minimum length scale of such devices that allows high exciton mobility .
second , our results show how one can engineer exciton transport on small length scales by doping molecular crystals with impurities .
this is relevant for bulk heterojunction organic solar cells and similar light - harvesting devices that rely on the high mobility of excitons @xcite .
third , our results are critical for the design of experiments aimed at detection of interaction - induced many - body localization ( mbl ) .
mbl is predicted to hinder the return of a perturbed quantum system to equilibrium , thus precluding thermalization , conductivity , density equilibration or energy transport .
it is the subject of numerous theoretical studies , many of which propose experiments for the observation of the localized phase .
of particular interest is the localization phase transition in 3d dipolar systems , where excitations / spins hop between carrier particles due to dipolar interactions .
a recent startling discovery @xcite shows that the localized phase can be induced by interactions between the particles .
in order for such a phase transition to be observable , the system must be in the diffusive regime in the limit of non - interacting particles .
our work shows that this is not guaranteed and provides the conditions required to perform these experiments .
particles with long - range hopping are known to be delocalized in lattices of infinite size . here
, we have presented evidence that particles with dipolar long - range hopping may undergo exponential localization in 3d disordered lattices of finite size .
our conclusions are based on three independent but mutually consistent results : * first , we have demonstrated by rigorous quantum dynamics calculations that particles placed in an individual lattice site can not diffuse through lattices of finite size over long time scales . by itself , this is not sufficient proof of localization as this result does not rule out the possibility that particles can fill the entire lattice at infinite time . * to further confirm the presence of localization , we have calculated the inverse participation ratio for each eigenstate of the hamiltonian .
our results show that no eigenstate occupies more than 10 % of the lattice for lattices of sufficiently small finite size .
our results also show how the participation ratio changes with increasing system size towards a system with delocalized states .
one may argue that this is not sufficient proof of localization as it does not rule out the possibility of states that occupy a small number of lattice sites that are spread throughout the lattice ( for example , every tenth site ) .
however , the presence of such states would allow for significant transport through the lattice , which we do not observe in ( i ) . the results in ( i ) and ( ii ) thus provide strong evidence for localization in finite - size lattices . however , the shape of the wavefunctions is still not clear . to better understand this , *
we have computed the fluctuations of the wavefunctions between instances of disorder .
our results illustrate that the wavefunction fluctuations follow a log - normal distribution , which corresponds to exponential localization of the wavefunctions .
our results in figure [ figurephasediagram ] illustrate the crossover between what would be observed in any experiment as the localization regime and the diffusion regime .
we have demonstrated that introducing diagonal disorder can bring the system to the localization regime , even if introducing dilution alone can not @xcite .
furthermore , introducing diagonal disorder alone is sufficient to observe the localization regime .
the crossover is shown to be within reach of current experiments .
we would like to emphasize that the absence of localization with dilution alone ( x - axis of figure [ figurephasediagram ] ) shows that there is no classical or quantum percolation threshold for a system of this size .
the lack of classical localization means that any localization induced by diagonal disorder must be due to quantum interference effects .
we have combined our wavepacket calculations in figure [ figurephasediagram ] with the scaling theory from refs .
@xcite to demonstrate how the crossover changes with the system size ( figure 5 ) .
surprisingly , we find that one should expect to observe the absence of quantum transport even in systems with macroscopically finite size , providing the on - site disorder and the dilution are significant enough .
these results are of practical interest as they can be used to inform the design of minituarized quantum devices , which always have finite size .
we acknowledge useful discussions with andreas buchleitner . this work is supported by nserc of canada .
c. a. mller and d. delande , _ disorder and interference : localization phenomena _ in _ les houches 2009 - session xci : ultracold gases and quantum information _ , edited by c. miniatura , l .- c .
kwek , m. ducloy , b. gremaud , b .-
englert , l. cugliandolo , a. ekert , k. k. phua ( oxford university press , oxford 2011 ) . | non - interacting particles with long - range hopping are known to be delocalized in disordered systems of infinite size .
it is thus natural to assume that such particles can traverse any finite - size lattice . here
, we demonstrate that this is not generally true .
the delocalization mechanism is induced by resonances between distant lattice sites .
the number of such resonances diverges with the system size . for a finite - size lattice
the number of resonances is finite and may not be sufficient to result in delocalization .
we consider particles with long - range ( dipolar ) hopping in three - dimensional lattices with diagonal disorder and random dilution .
we compute the wavepacket dynamics of particles placed in an individual lattice site , the inverse participation ratios and the fluctuations of the wavefunctions with disorder .
we show that , for certain finite - size disordered lattices , particles remain localized within a finite volume much smaller than the lattice size , and that the wavefunctions exhibit the log - normal fluctuations characteristic of anderson localization .
we characterize the localization properties as functions of dilution and diagonal disorder .
we combine our results with scaling theory to obtain the size dependence of the localization diffusion crossover .
our results indicate that particles with long - range hopping undergo exponential localization in lattices of finite size , even macroscopically finite .
our phase diagrams illustrate a rather unusual phenomenon : quantum particles can diffuse through a lattice of size @xmath0 , but not through a lattice of size @xmath1 . |
conventional wireless mobile networks are designed with a cellular architecture , where base stations ( bss ) from different cells control communications for their associated mobile stations ( mss ) independently .
the resulting inter - cell interference is treated as additive noise and minimized by applying a predesigned frequency reuse pattern such that the same frequency band is reused only by non - adjacent cells . due to the rapidly growing demand for high - rate wireless multimedia applications ,
conventional cellular networks have been pushed towards their throughput limits .
consequently , many beyond-3 g wireless technologies such as wimax and 3gpp umts long term evolution ( lte ) have relaxed the constraint on the frequency reuse such that the total frequency band becomes available for reuse by all cells .
however , this factor - one frequency reuse pattern renders the overall network performance limited by the inter - cell interference ; consequently , more sophisticated interference management techniques with multi - cell cooperation become crucial . among others , one effective method to cope with the inter - cell interference in the cellular network is via joint signal processing across different bss . in this correspondence
, we study a particular type of multi - bs cooperation for the downlink transmission , where we are interested in evaluating the benefit in terms of network throughput by cooperatively optimizing the transmit beamforming vectors for different bss each with multiple antennas .
notice that the problem setup of our work is different from that for a fully cooperative multi - cell system considered in , e.g. , @xcite-@xcite , where a central processing unit is assumed with the global knowledge of all the required downlink channels and user messages to jointly design the transmitted signals for all bss .
in contrast , our work focuses on the decentralized implementation of the multi - cell cooperative downlink beamforming assuming only the local message and neighboring - channel knowledge at each bs , which is more practical than implementing the full baseband - level coordination .
it is worth noting that decentralized multi - cell cooperative downlink beamforming has been studied in @xcite to minimize the total power consumption of all bss to meet with mss individual signal - to - interference - plus - noise ratio ( sinr ) targets , based on the uplink - downlink beamforming duality . in this work ,
we provide a different design approach for rate - optimal strategies in decentralized multi - cell cooperative beamforming . for the purpose of exposition ,
in this work we consider a simplified scenario , where each ms is equipped with a single antenna , and at any given time there is only one active ms in each cell ( over a particular frequency band ) .
accordingly , we can model the multi - cell cooperative downlink transmission system as a multiple - input single - output ( miso ) gaussian interference channel ( ic ) , termed as miso - ic . from an information - theoretic viewpoint ,
the capacity region of the gaussian ic , which constitutes all the simultaneously achievable rates for all users , is still unknown in general @xcite , while significant progresses have recently been made on approaching this limit @xcite .
capacity - approaching techniques for the gaussian ic in general require certain signal - level encoding / decoding cooperations among the users , while a more pragmatic approach that leads to suboptimal achievable rates of the users is to allow only single - user encoding and decoding by treating the interference from other users as additive gaussian noise . in this work , we adopt the latter approach to study the design of cooperative transmit beamforming for the miso - ic . particularly , we focus on the design criterion to achieve different rate - tuples for the users on the pareto boundary of the achievable rate region for the miso - ic . due to the coupled signal structure ,
the achievable rate region for the miso - ic with interference treated as noise is in general a non - convex set , which renders the joint optimization of beamforming vectors to achieve different pareto - boundary rate - tuples a challenging task .
note that this problem has been studied in @xcite , where for the special two - user case , it was shown that the optimal transmit beamforming vector to achieve a pareto - boundary rate - pair for the miso - ic can be expressed as a linear combination of the zero - forcing ( zf ) and maximum - ratio transmission ( mrt ) beamformers .
the rate maximization for the ic with interference treated as noise has also been studied in the literature via various `` pricing '' algorithms ( see , e.g. , @xcite and references therein ) , while in general the price - based approach does not achieve the pareto - optimal rates for the miso - ic . in @xcite , the maximum sum - rate for the gaussian ic with interference treated as noise
is characterized in terms of degrees of freedom ( dof ) over the interference - limited regime . in this correspondence
, we develop a new _
parametrical _ characterization of the pareto boundary for the miso - ic in terms of the interference - power levels at all receivers caused by different transmitters , also known as the _ interference temperature _ ( it ) levels in the newly emerging `` cognitive radio ( cr ) '' type of applications @xcite .
we show that each pareto - boundary rate - tuple can be achieved in a decentralized manner when each of the users maximizes its own miso channel capacity subject to a certain set of it constraints at the other users receivers , which is identical to the cr miso channel transmit optimization problem studied in @xcite and thus shares the same solution structure .
we derive new closed - form solutions for the optimal transmit covariance matrices of all users to achieve an arbitrary rate - tuple on the pareto boundary of the miso - ic rate region , from which we see that the optimal transmit covariance matrices should all be _ rank - one _
( i.e. , beamforming is optimal ) .
furthermore , we derive the conditions that are necessary for any particular set of mutual it constraints across all users to guarantee a pareto - optimal rate - tuple for the miso - ic . based on these conditions , we propose a new _ decentralized _ algorithm for implementing the multi - cell cooperative downlink beamforming . for this algorithm ,
all different pairs of bss independently search for their mutually desirable it constraints ( with those for the mss associated with the other bss fixed ) , under which their respective beamforming vectors are optimized to maximize the individual transmit rates .
this algorithm improves the rates for the bss in a pairwise manner until the transmit rates for all bss converge with their mutual it levels .
_ notation _ :
@xmath0 and @xmath1 denote the identity matrix and the all - zero matrix , respectively , with appropriate dimensions . for a square matrix @xmath2 ,
@xmath3 , @xmath4 , @xmath5 , and @xmath6 denote the trace , determinant , inverse , and square - root of @xmath2 , respectively ; and @xmath7 means that @xmath2 is positive semi - definite @xcite .
@xmath8 denotes a diagonal matrix with the diagonal elements given by @xmath9 . for a matrix @xmath10 of arbitrary size , @xmath11 ,
@xmath12 , and @xmath13 denote the hermitian transpose , transpose , and rank of @xmath10 , respectively .
@xmath14 $ ] denotes the statistical expectation .
the distribution of a circularly symmetric complex gaussian ( cscg ) random vector with the mean vector @xmath15 and the covariance matrix @xmath16 is denoted by @xmath17 ; and @xmath18 stands for `` distributed as '' .
@xmath19 denotes the space of @xmath20 complex matrices .
@xmath21 denotes the euclidean norm of a complex vector ( scalar ) @xmath15 .
the @xmath22 function is with base @xmath23 by default .
we consider the downlink transmission in a cellular network consisting of @xmath24 cells , each having a multi - antenna bs to transmit independent messages to one active single - antenna ms .
it is assumed that all bss share the same narrow - band spectrum for downlink transmission .
accordingly , the system under consideration can be modeled by a @xmath24-user miso - ic .
it is assumed that the bs in the @xmath25th cell , @xmath26 , is equipped with @xmath27 transmitting antennas , @xmath28 . the discrete - time baseband signal received by the active ms in the @xmath25th cell
is then given by @xmath29 where @xmath30 denotes the transmitted signal from the @xmath25th bs ; @xmath31 denotes the direct - link channel for the @xmath25th ms , while @xmath32 denotes the cross - link channel from the @xmath33th bs to the @xmath25th ms , @xmath34 ; and @xmath35 denotes the receiver noise .
it is assumed that @xmath36 , and all @xmath35 s are independent .
we assume independent encoding across different bss and thus @xmath37 s are independent over @xmath25 .
it is further assumed that a gaussian codebook is used at each bs and @xmath38 , where @xmath39 $ ] denotes the transmit covariance matrix for the @xmath25th bs , with @xmath40 and @xmath41 .
notice that the cscg distribution has been assumed for all the transmitted signals .
furthermore , the interferences at all the receivers caused by different transmitters are treated as gaussian noises .
thus , for a given set of transmit covariance matrices of all bss , @xmath42 , the achievable rate of the @xmath25th ms is expressed as @xmath43 next , we define the achievable rate region for the miso - ic to be the set of rate - tuples for all mss that can be simultaneously achievable under a given set of transmit - power constraints for the bss , denoted by @xmath44 : @xmath45 the upper - right boundary of this region is called the _ pareto boundary _ , since it consists of rate - tuples at which it is impossible to improve a particular user s rate , without simultaneously decreasing the rate of at least one of the other users .
more precisely , the pareto optimality of a rate - tuple is defined as follows @xcite .
[ definition : pareto optimal ] a rate - tuple @xmath46 is _ pareto optimal _ if there is no other rate - tuple @xmath47 with @xmath48 and @xmath49 ( the inequality is component - wise ) . in this work ,
we consider the scenario where multiple bss in the cellular network cooperatively design their transmit covariance matrices in order to minimize the effect of the inter - cell interference on the overall network throughput .
in particular , we are interested in the design criterion to achieve different pareto - optimal rate - tuples for the corresponding miso - ic defined as above .
it is worth noting that in prior works on characterizing the pareto boundary for the miso - ic with interference treated as noise ( see , e.g. , @xcite and references therein ) , it has been assumed ( without proof ) that the optimal transmit strategy for users to achieve any rate - tuple on the pareto boundary is _ beamforming _
, i.e. , @xmath50 is a rank - one matrix for all @xmath25 s . under this assumption
, we can express @xmath50 as @xmath51 , where @xmath52 denotes the beamforming vector for the @xmath25th user .
similarly as in the general case with @xmath53 , the achievable rates and rate region of the miso - ic with transmit beamforming ( bf ) can be defined in terms of @xmath54 s .
however , it is not evident whether the bf case bears the same pareto boundary as the general case with @xmath53 for the miso - ic . in this work
, we will show that this is indeed the case ( see section [ sec : it ] ) .
accordingly , we can choose to use the rate and rate - region expressions in terms of either @xmath50 s or @xmath54 s to characterize the pareto boundary of the miso - ic , for the rest of this correspondence .
= 0.7 in the following , we review some existing approaches to characterize the pareto boundary for the miso - ic with interference treated as noise . for the purpose of illustration , in fig .
[ fig : rate region ] , we show the achievable rate region for a two - user miso gaussian ic with interference treated as noise ( prior to any time - sharing of achievable rate - pairs ) , which is observed to be non - convex . a commonly adopted method to obtain the pareto boundary for the miso - ic
is via solving a sequence of weighted sum - rate maximization ( wsrmax ) problems , each for a given set of user rate weights , @xmath55 , and given by @xmath56 where @xmath57 is the receiver sinr for the @xmath25th user defined as @xmath58 it is easy to verify that this problem is non - convex , and thus can not be solved efficiently . in addition , the above method can not guarantee the finding of all pareto - boundary points for the miso - ic ( cf .
[ fig : rate region ] ) . an alternative method to characterize the _ complete _
pareto boundary for the miso - ic is based on the concept of _ rate profile _ @xcite .
specifically , any rate - tuple on the pareto boundary of the rate region can be obtained via solving the following optimization problem with a particular rate - profile vector , @xmath59 : @xmath60 with @xmath61 denoting the target ratio between the @xmath25th user s achievable rate and the users sum - rate , @xmath62 . without loss of generality
, we assume that @xmath63 , and @xmath64 . for a given @xmath65 , let the optimal solution of problem ( [ eq : srmax ] ) be denoted by @xmath66 .
then , it follows that @xmath67 must be the corresponding pareto - optimal rate - tuple , which can be geometrically viewed as ( cf .
[ fig : rate region ] ) the intersection between a ray in the direction of @xmath65 and the pareto boundary of the rate region .
thereby , with different @xmath65 s , solving problem ( [ eq : srmax ] ) yields the complete pareto boundary for the rate region , which does not need to be convex .
next , we show that problem ( [ eq : srmax ] ) can be solved via solving a sequence of feasibility problems each for a fixed @xmath68 and expressed as @xmath69 if the above problem is feasible for a given sum - rate target , @xmath70 , it follows that @xmath71 ; otherwise , @xmath72 .
thus , by solving problem ( [ eq : feasibility problem ] ) with different @xmath70 s and applying the simple bisection method @xcite over @xmath70 , @xmath73 can be obtained for problem ( [ eq : srmax ] ) .
let @xmath74 . then , for problem ( [ eq : feasibility problem ] ) , we can replace the rate constraints by the equivalent sinr constraints given by @xmath75 similarly as shown in @xcite , the resultant feasibility problem can be transformed into a second - order cone programming ( socp ) problem , which is convex and can be solved efficiently @xcite .
in this section , instead of investigating centralized approaches , we present a new method to characterize the pareto boundary for the miso - ic in a distributed fashion , by exploring its relationship with the cr miso channel @xcite . we start with the general - rank transmit covariance matrices @xmath50 s for the miso - ic .
first , we introduce a set of auxiliary variables , @xmath76 , where @xmath77 is called the interference - power or interference - temperature ( it ) constraint from the @xmath25th bs to @xmath33th ms , @xmath34 , @xmath78 . for notational convenience ,
let @xmath79 be the vector consisting of all @xmath80 different @xmath77 s , and @xmath81 be the vector consisting of all @xmath82 different @xmath77 s and @xmath83 s , @xmath84 , for any given @xmath85 .
next , we consider a set of parallel transmit covariance optimization problems , each for one of the @xmath24 bss in the miso - ic expressed as @xmath86 where @xmath85 .
note that in the above problem for a given @xmath25 , @xmath81 is fixed .
for notational convenience , we denote the optimal value of this problem as @xmath87 .
if in the objective function of ( [ eq : cr capacity ] ) we set @xmath88 , @xmath87 becomes equal to the maximum achievable rate of an equivalent miso cr channel @xcite , where the @xmath25th bs is the so - called `` secondary '' user transmitter , and all the other @xmath89 bss , indexed by @xmath84 , are the `` primary '' user transmitters , each of which has a transmit covariance matrix , @xmath90 , and its intended `` primary '' user receiver is protected by the secondary user via the it constraint : @xmath91 . in @xcite
, it was proved that the solution for problem ( [ eq : cr capacity ] ) is rank - one , i.e. , beamforming is optimal , and in the special case of @xmath92 ( i.e. , one single primary user ) , a closed - form solution for the optimal beamforming vector was derived . in the following proposition , we provide a new closed - form solution for problem ( [ eq : cr capacity ] ) with arbitrary values of @xmath24 , from which we can easily infer that beamforming is indeed optimal .
[ proposition : optimal cr solution ] the optimal solution of problem ( [ eq : cr capacity ] ) is _ rank - one _ , i.e. , @xmath93 , and @xmath94 where @xmath95 , @xmath34 , and @xmath96 , are non - negative constants ( dual variables ) corresponding to the @xmath25th bs s it constraint for the @xmath33th ms and its own transmit - power constraint , respectively , which are obtained as the optimal solutions for the dual variables in the dual problem of problem ( [ eq : cr capacity ] ) ; and @xmath97 is given by @xmath98 where @xmath99 and @xmath100 . please see appendix [ appendix : proof optimality rank one ] .
now , we are ready to present a _
parametrical _ characterization of the pareto boundary for the miso - ic in terms of @xmath79 as follows .
[ proposition : relationship ] for any rate - tuple @xmath101 on the pareto boundary of the miso - ic rate region defined in ( [ eq : rate region ] ) , which is achievable with a set of transmit covariance matrices , @xmath42 , there is a corresponding interference - power constraint vector , @xmath102 , with @xmath103 , and @xmath85 , such that @xmath104 , and @xmath50 is the optimal solution of problem ( [ eq : cr capacity ] ) for the given @xmath25
. please see appendix [ appendix : proof relationship ] . from proposition [ proposition : relationship ]
, it follows that the pareto boundary for the miso - ic is parameterized in terms of a lower - dimensional manifold over the non - negative real vector @xmath79 , in comparison with that over the complex transmit covariance matrices , @xmath50 s , or with that over the complex beamforming vectors , @xmath54 s .
furthermore , by combining propositions [ proposition : optimal cr solution ] and [ proposition : relationship ] , it follows that _ beamforming _ is indeed optimal to achieve any rate - tuple on the pareto boundary for the miso - ic .
it is worth noting that the dimensionality reduction approach proposed in this work for characterizing the pareto boundary of the miso - ic is in spirit similar to that proposed in @xcite , where it has been shown that the transmit beamforming vectors to achieve any pareto - boundary rate - tuple of the @xmath24-user miso - ic with interference treated as noise can be expressed in the following forms : @xmath105 where @xmath106 s are complex coefficients .
note that under the assumption of independent @xmath107 s , the above beamforming structure is non - trivial only when @xmath108 .
for this case , from remark [ remark : kkt ] in appendix [ appendix : proof optimality rank one ] , it is known that for the optimal beamforming structure given in ( [ eq : optimal bf ] ) , we have @xmath109 . with this and by applying the matrix inversion lemma @xcite , it can be shown ( the detailed proof is omitted here for brevity ) that the optimal beamforming structure given by ( [ eq : optimal bf ] ) is indeed in accordance with that given by ( [ eq : old bf solution ] ) . the main difference for these two methods to characterize the pareto boundary for the miso - ic lies in their adopted parameters : the method in our work uses @xmath80 real @xmath77 s , while that in @xcite uses @xmath80 complex @xmath106 s .
note that @xmath77 corresponds to the it constraint from the @xmath25th user transmitter to the @xmath33th user receiver , whereas there is no practical meaning associated with @xmath106 .
consequently , as will be shown next , the proposed method in our work leads to a practical algorithm to implement the multi - cell cooperative downlink beamforming , via iteratively searching for mutually desirable it constraints between different pairs of bss .
in this section , we develop a new _ decentralized _ algorithm that practically implements the multi - cell cooperative downlink beamforming based on the results derived in the previous section .
it is assumed that each bs in the cellular network has the perfect knowledge of the channels from it to all mss .
furthermore , it is assumed that all bss operate according to the same protocol described as follows .
initially , a set of prescribed it constraints in @xmath79 are set over the whole network , and the @xmath25th bs is informed of its corresponding @xmath110 .
accordingly , each bs sets its own transmit beamforming vector via solving problem ( [ eq : cr capacity ] ) and sets its transmit rate equal to the optimal objective value of problem ( [ eq : cr capacity ] ) , which is achievable for its ms since the actual it levels from the other bss must be below their prescribed constraints .
then , by assuming that there is an error - free link between each pair of bss , all different pairs of bss start to communicate with each other for updating their mutual it constraints ( the details are given later in this section ) , under which each pair of bss reset their respective beamforming vectors via solving problem ( [ eq : cr capacity ] ) such that the achievable rates for their mss both get improved .
notice that each pair of updating bss keeps the it constraints for the mss associated with the other bss excluding this pair fixed ; and as a result , the transmit rates for all the other mss are not affected .
therefore , the above algorithm can be implemented in a pairwise decentralized manner across the bss , while it converges when there are no incentives for all different pairs of bss to further update their mutual it constraints .
next , we focus on the key issue on how to update the mutual it constraints for each particular pair of bss to guarantee the rate increase for both of their mss . to resolve this problem , in the following proposition
, we first provide the necessary conditions for any given @xmath102 ( component - wise ) to correspond to a pareto - optimal rate - tuple for the miso - ic , which will lead to a simple rule for updating the mutual it constraints between different pairs of bss .
note that from proposition [ proposition : relationship ] , it follows that for any pareto - optimal rate - tuple of the miso - ic , there must exist a @xmath79 such that the optimal solutions of the problems given in ( [ eq : cr capacity ] ) for all @xmath25 s are the same as those for the general formulation of miso - ic to achieve this rate - tuple .
however , for any given @xmath102 , it remains unknown whether this value of @xmath79 will correspond to a pareto - optimal rate - tuple for the miso - ic .
[ proposition : conditions ] for an arbitrarily chosen @xmath102 , if the optimal rate values of the problems in ( [ eq : cr capacity ] ) for all @xmath25 s , @xmath87 s , are pareto - optimal on the boundary of the miso - ic rate region defined in ( [ eq : rate region ] ) , then for any pair of @xmath111 , and @xmath112 , it must hold that @xmath113 , where @xmath114 is defined as @xmath115.\end{aligned}\ ] ] please see appendix [ appendix : proof conditions ] .
note that @xmath114 s for all different pairs of @xmath116 can be obtained from the ( primal and dual ) solutions of the problems given in ( [ eq : cr capacity ] ) for all @xmath25 s with the given @xmath79 ( for the details , please refer to appendix [ appendix : proof optimality rank one ] ) .
more specifically , we have @xmath117 where @xmath118 is the solution for the dual problem of problem ( [ eq : cr capacity ] ) with @xmath119 , which corresponds to the @xmath33th it constraint , and from the objective function of problem ( [ eq : cr capacity ] ) , @xmath120 where @xmath121 is the optimal solution of problem ( [ eq : cr capacity ] ) with @xmath119 . similarly , @xmath122 and
@xmath123 can be obtained from solving problem ( [ eq : cr capacity ] ) via the lagrange duality method with @xmath124 . from proposition [ proposition : conditions ] , the following observations can be easily obtained ( the proofs are omitted for brevity ) : * for any particular @xmath79 that corresponds to a pareto - optimal rate - tuple , it must hold that @xmath125 , @xmath126 , where @xmath127 corresponds to the case of using maximum transmit power with mrt beamforming for the @xmath128th bs ; * for any particular @xmath79 that corresponds to a pareto - optimal rate - tuple , it must hold that @xmath129 , @xmath130 , i.e. , the it constraints across all bss must be tight . from the above observations
, we see that if we are only interested in the values of @xmath79 that correspond to pareto - optimal rate - tuples for the miso - ic , it is sufficient for us to focus on the subset of @xmath79 within the set @xmath102 , in which @xmath125 and @xmath131 .
based on proposition [ proposition : conditions ] , we can develop a simple rule for different pairs of bss in the cooperative multi - cell system to update their mutual it constraints for improving both of their transmit rates , while keeping those of the other bss unchanged . from the proof of proposition
[ proposition : conditions ] given in appendix [ appendix : proof conditions ] , it follows that the method for any updating bs pair @xmath116 to fulfill the above requirements is via changing @xmath132 and @xmath133 according to ( [ eq : gamma update ] ) .
note that in general , the choice for @xmath134 in ( [ eq : gamma update ] ) to make @xmath135 is not unique . for notational conciseness ,
let @xmath136}$ ] ; it can then be shown that one particular choice for @xmath134 is @xmath137^t\end{aligned}\ ] ] where @xmath138 if @xmath139 and @xmath140 otherwise ; @xmath141 is a constant that controls the ratio between the rate increments for the @xmath128th and @xmath33th bss .
it can be easily verified that when @xmath142 , a larger rate increment is resulted for the @xmath128th bs than that for the @xmath33th bs , and vice versa when @xmath143 ( provided that the step - size @xmath144 in ( [ eq : gamma update ] ) is sufficiently small ) .
more specifically , the procedure for any bs pair @xmath116 , @xmath145 , @xmath146 , and @xmath147 , to update their mutual it constraints is given as follows .
first , the @xmath128th bs computes the elements @xmath148 and @xmath149 in @xmath114 according to ( [ eq : a ] ) and ( [ eq : b ] ) , respectively , with the present value of @xmath150 .
similarly , the @xmath33th bs computes @xmath151 and @xmath152 with the present value of @xmath153 .
next , the @xmath128th bs sends @xmath148 and @xmath149 to the @xmath33th bs , while the @xmath33th bs sends @xmath151 and @xmath152 to the @xmath128th bs .
then , assuming that @xmath154 and @xmath144 are preassigned values known to these two bss , they can both compute @xmath134 according to ( [ eq : d update ] ) and update @xmath132 and @xmath133 according to ( [ eq : gamma update ] ) in appendix [ appendix : proof conditions ] .
last , with the updated values @xmath155 and @xmath156 , these two bss reset their respective beamforming vectors and transmit rates via solving ( [ eq : cr capacity ] ) independently .
note that the above operation requires only local information ( scalar ) exchanges between different pairs of bss , and thus can be implemented at a very low cost in a cellular system .
one version of the decentralized algorithm for cooperative downlink beamforming in a multi - cell system is described in table [ table ] . since in each iteration of the algorithm the achievable rates for the pair of updating bss both improve and those for all other bss are unaffected ( non - decreasing ) , and the maximum achievable rates for all bss are bounded by finite pareto - optimal values , the convergence of this algorithm is ensured .
= 0.7 .algorithm for cooperative downlink beamforming .
[ cols= " < " , ] in fig .
[ fig : rate region new ] ( with the same two - user miso - ic as for fig .
[ fig : rate region ] ) , we show the pareto boundary for an example miso - ic with @xmath92 , @xmath157 , @xmath158 , and @xmath159 , which is obtained by the proposed method in this correspondence , i.e. , solving the problems given in ( [ eq : cr capacity ] ) for @xmath160 , and a given pair of values @xmath161 and @xmath162 with @xmath163 and @xmath164 , and then taking a closure operation over the resultant rate - pairs with all different values of @xmath161 and @xmath162 within their respective ranges .
we demonstrate the effectiveness of the proposed decentralized algorithm for implementing the multi - cell cooperative downlink beamforming with two initial rate - pairs , indicated by `` zf '' and `` mrt '' in fig .
[ fig : rate region new ] , which are obtained when both bss adopt the zf and the mrt beamforming vectors , respectively , with their maximum transmit powers .
it is observed that the achievable rates for both mss increase with iterations and finally converge to a pareto - optimal rate - pair .
are not only necessary ( as proved in this work ) but also _ sufficient _ for any given @xmath79 to achieve a pareto - optimal rate - tuple for the miso - ic . ] by comparing the two cases with @xmath165 and @xmath166 , it is observed that a larger value of @xmath167 results in larger rate values for the first ms in the converged rate - pairs , which is in accordance with our previous discussion .
in this correspondence , based on the concept of interference temperature ( it ) and under a cellular downlink setup , we have developed a new method to characterize the complete pareto boundary of the achievable rate region for the @xmath24-user gaussian miso - ic with interference treated as noise .
it is shown that the proposed method also leads to a new decentralized algorithm for implementing the downlink beamforming in a cooperative multi - cell system to achieve maximal rates with a prescribed fairness guarantee .
there are a number of directions along which the developed results in this work can be further investigated .
first , it would be interesting to extend the multi - cell cooperative beamforming design based on the principle of it to the scenario where each bs supports simultaneous transmissions to _ multiple _ active mss each with a single antenna or multiple antennas .
second , it remains yet to be proved whether the necessary conditions derived in this work for any particular set of it constraints across the bss to guarantee a pareto - optimal rate - tuple for the miso - ic are also _ sufficient _ , even for the special two - user case .
this proof is essential for the proposed downlink beamforming algorithm to achieve the global convergence ( pareto - optimal rates ) .
last but not least , it is pertinent to analyze the proposed decentralized algorithm that iteratively updates the mutual it constraints between different pairs of bss from a _ game - theoretical _ viewpoint .
it can be verified that problem ( [ eq : cr capacity ] ) is convex , and thus it can be solved by the standard lagrange duality method @xcite .
let @xmath95 , @xmath34 , and @xmath96 be the non - negative dual variables for problem ( [ eq : cr capacity ] ) associated with the @xmath25th bs s it constraint for the @xmath33th ms and its own transmit - power constraint , respectively .
the lagrangian function for this problem can be written as @xmath168 where @xmath169 $ ] .
the dual function of problem ( [ eq : cr capacity ] ) is given by @xmath170 accordingly , the dual problem is defined as @xmath171 where @xmath172 means component - wise non - negative . since problem ( [ eq : cr capacity ] ) is convex with strictly feasible points @xcite , the duality gap between its optimal value and that of the dual problem is zero ; thus , problem ( [ eq : cr capacity ] ) can be equivalently solved via solving its dual problem . in order to solve the dual problem , we need to obtain the dual function @xmath173 for any given @xmath174 .
this can be done by solving the maximization problem given in ( [ eq : dual function ] ) , which , according to ( [ eq : lagrangian ] ) , can be explicitly written as ( by discarding irrelevant constant terms ) : @xmath175 where @xmath176 and @xmath177 of dimension @xmath178 .
in order for problem ( [ eq : dual optimization ] ) to have a bounded objective value , it is shown as follows that @xmath179 should be a full - rank matrix .
suppose that @xmath179 is rank - deficient , such that we could define @xmath180 , where @xmath181 and @xmath182 satisfying @xmath183 and @xmath184 .
thereby , the objective function of problem ( [ eq : dual optimization ] ) reduces to @xmath185 due to the independence of @xmath186 and @xmath107 s , and thus the independence of @xmath186 and @xmath187 , it follows that @xmath188 with probability one such that ( [ eq : dual objective func ] ) goes to infinity by letting @xmath189 . since the optimal value of problem ( [ eq : cr capacity ] ) must be bounded , without loss of generality , we only need to consider the subset of @xmath190 in the set @xmath174 to make @xmath179 full - rank . [
remark : kkt ] note that from the definition of @xmath179 and the karush - kuhn - tucker ( kkt ) optimality conditions @xcite of problem ( [ eq : cr capacity ] ) , it follows that @xmath179 is full - rank only when either of the following two cases occurs : * @xmath109 : in this case , the transmit power constraint for the @xmath25th bs is tight for problem ( [ eq : cr capacity ] ) ; * @xmath191 , but there are at least @xmath27 @xmath95 s , @xmath34 , having @xmath192 : in this case , regardless of whether the transmit power constraint for the @xmath25th bs is tight , there are at least @xmath27 out of the @xmath89 it constraints of the @xmath25th bs are tight in problem ( [ eq : cr capacity ] ) .
note that this case can be true only when @xmath193 . from the above discussions
, it is known that @xmath194 exists .
thus , we can introduce a new variable @xmath195 for problem ( [ eq : dual optimization ] ) as @xmath196 and substituting it into ( [ eq : dual optimization ] ) yields @xmath197 without loss of generality , we can express @xmath195 into its eigenvalue decomposition ( evd ) as @xmath198 , where @xmath199\in\mathbb{c}^{m_k\times m_k}$ ] is unitary and @xmath200)\succeq 0 $ ] . substituting the ed of @xmath195 into ( [ eq : dual optimization new ] ) yields @xmath201 for any given @xmath202
, it can be verified that the optimal solution of @xmath203 for problem ( [ eq : dual optimization new 2 ] ) is given by @xmath204 thus , it follows that for the optimal solution of problem ( [ eq : dual optimization new ] ) , @xmath205 .
furthermore , let @xmath206 denote the index of @xmath128 for which @xmath207
. the objective function of problem ( [ eq : dual optimization new 2 ] ) reduces to @xmath208 clearly , the above function is maximized with any @xmath209 when @xmath210 from ( [ eq : bf vector ] ) and ( [ eq : power ] ) , it follows that the optimal solution for problem ( [ eq : dual optimization new ] ) is @xmath211 combining the above solution and ( [ eq : transform ] ) , it can be easily shown that the optimal solution @xmath50 for problem ( [ eq : cr capacity ] ) is as given by proposition [ proposition : optimal cr solution ] . with the obtained dual function @xmath173 for
any given @xmath212 , the dual problem ( [ eq : dual problem ] ) can be solved by searching over @xmath172 to minimize @xmath173 .
this can be done via , e.g. , the ellipsoid method @xcite , by utilizing the subgradient of @xmath173 that is obtained from ( [ eq : lagrangian ] ) as @xmath213 for @xmath214 and @xmath215 for @xmath96 , where @xmath216 is the optimal solution for problem ( [ eq : dual optimization ] ) with the given @xmath190 .
when @xmath212 converges to the optimal solution for the dual problem , the corresponding @xmath216 becomes the optimal solution for problem ( [ eq : cr capacity ] ) .
proposition [ proposition : optimal cr solution ] thus follows . since the given set of @xmath42 achieves the pareto - optimal rate - tuple @xmath101 for the miso - ic , from ( [ eq : user rates ] ) and ( [ eq : rate region ] ) it follows that for any @xmath85 @xmath217 since @xmath218 , ( [ eq : rk ] ) can be rewritten as @xmath219 note that ( [ eq : rk new ] ) is the same as the objective function of problem ( [ eq : cr capacity ] ) for the given @xmath25 .
furthermore , from ( [ eq : rate region ] ) it follows that @xmath220 . using this and the fact that @xmath221 , it follows that @xmath50 satisfies the constraints given in problem ( [ eq : cr capacity ] ) for the given @xmath25 .
therefore , @xmath50 must be a feasible solution for problem ( [ eq : cr capacity ] ) with the given @xmath25 and @xmath81 .
next , we need to prove that @xmath50 is indeed the optimal solution of problem ( [ eq : cr capacity ] ) for any given @xmath25 , and thus the corresponding achievable rate @xmath222 is equal to the optimal value of problem ( [ eq : cr capacity ] ) , which is @xmath87 .
we prove this result by contradiction .
suppose that the optimal solution for problem ( [ eq : cr capacity ] ) , denoted by @xmath223 , is not equal to @xmath50 for a given @xmath25 .
thus , we have @xmath224 furthermore , since @xmath225 , we have for any @xmath34 , @xmath226 thus , for another set of transmit covariance matrices given by @xmath227 , the corresponding achievable rate - tuple for the miso - ic , @xmath228 , satisfies that @xmath229 and @xmath230 , which contradicts the fact that @xmath101 is a pareto - optimal rate - tuple for the miso - ic .
hence , the presumption that @xmath231 for any given @xmath25 can not be true .
thus , we have @xmath232 and @xmath104 . proposition [ proposition : relationship ]
thus follows .
as given in proposition [ proposition : conditions ] , with @xmath79 , the corresponding optimal values of the problems in ( [ eq : cr capacity ] ) for all @xmath25 s , @xmath87 s , correspond to a pareto - optimal rate - tuple for the miso - ic , denoted by @xmath101 .
let @xmath42 denote the set of optimal solutions for the problems in ( [ eq : cr capacity ] ) .
we thus have @xmath233 next , we prove proposition [ proposition : conditions ] by contradiction .
suppose that there exists a pair of @xmath116 with @xmath234 , where @xmath114 is defined in ( [ eq : d ] ) .
define a new @xmath235 over @xmath79 , where all the elements in @xmath79 remain unchanged except @xmath236^t$ ] being replaced by @xmath237^t=[\gamma_{ij } , \gamma_{ji}]^t+\delta_{ij}\cdot{\mbox{\boldmath{$ d $ } } } _ { ij}\end{aligned}\ ] ] where @xmath238 is a small step - size , and @xmath134 is any vector that satisfies @xmath239 ( component - wise ) , with one possible value for such @xmath134 is given by ( [ eq : d update ] ) in the main text . with @xmath235 , the optimal solutions for the problems in ( [ eq : cr capacity ] )
remain unchanged @xmath240 , while for those with @xmath119 and @xmath124 , the optimal solutions are changed to be @xmath121 and @xmath241 , respectively .
accordingly , the new achievable rates in the miso - ic for any @xmath242 are given by @xmath243 where ( [ eq : inequality ] ) is due to ( [ eq : r k renew ] ) and the facts that @xmath244 and @xmath245 .
also , it can be shown that @xmath246 where ( [ eq : inequality 2 ] ) is due to the facts that @xmath247 and @xmath121 achieves the optimal value of problem ( [ eq : cr capacity ] ) with @xmath119 and the given @xmath248 , denoted by @xmath249 .
similarly , it can be shown that @xmath250 .
thus , from ( [ eq : gamma update ] ) and @xmath239 , it follows that with sufficiently small @xmath144 @xmath251 \geq & \left[\begin{array}{l } c_i({\mbox{\boldmath{$ \gamma $ } } } ' _ i ) \\
c_j({\mbox{\boldmath{$ \gamma $ } } } ' _ j)\end{array } \right]\\ \cong & \left[\begin{array}{l } c_i({\mbox{\boldmath{$ \gamma $ } } } _
i ) \\ c_j({\mbox{\boldmath{$ \gamma $ } } } _ j)\end{array } \right]+\delta_{ij}{\mbox{\boldmath{$ d $ } } } _ { ij}{\mbox{\boldmath{$ d $ } } } _ { ij } \\
> & \left[\begin{array}{l } r_i \\
r_j \end{array } \right].\end{aligned}\ ] ] therefore , we have found a new set of achievable rate - tuple for the miso - ic with @xmath235 , @xmath228 , which has @xmath252 , @xmath253 , and @xmath254 .
clearly , this contradicts the fact that @xmath101 is pareto - optimal for the miso - ic .
thus , the presumption that there exists a pair of @xmath116 with @xmath234 can not be true .
proposition [ proposition : conditions ] thus follows .
m. karakayali , g. j. foschini , and r. a. valenzuela , `` network coordination for spectrally efficient communications in cellular systems , '' _ ieee wireless commun .
_ , vol . 13 , no .
56 - 61 , aug . 2006 .
d. schmidt , c. shi , r. berry , m. honig , and w. utschick , `` distributed resource allocation schemes : pricing algorithms for power control and beamformer design in interference networks , '' _ ieee trans . sig . process . mag .
_ , vol . 26 .
5 , pp . 53 - 63 , sep .
2009 .
m. mohseni , r. zhang , and j. m. cioffi , `` optimized transmission for fading multiple - access and broadcast channels with multiple antennas , '' _ ieee j. sel .
areas commun .
24 , no . 8 , pp . 1627 - 1639 , aug . | this correspondence studies the downlink transmission in a multi - cell system , where multiple base stations ( bss ) each with multiple antennas cooperatively design their respective transmit beamforming vectors to optimize the overall system performance . for simplicity
, it is assumed that all mobile stations ( mss ) are equipped with a single antenna each , and there is one active ms in each cell at one time .
accordingly , the system of interests can be modeled by a multiple - input single - output ( miso ) interference channel ( ic ) , termed as miso - ic , with interference treated as noise .
we propose a new method to characterize different rate - tuples for active mss on the pareto boundary of the achievable rate region for the miso - ic , by exploring the relationship between the miso - ic and the cognitive radio ( cr ) miso channel .
we show that each pareto - boundary rate - tuple of the miso - ic can be achieved in a decentralized manner when each of the bss attains its own channel capacity subject to a certain set of interference - power constraints ( also known as interference - temperature constraints in the cr system ) at the other ms receivers .
furthermore , we show that this result leads to a new decentralized algorithm for implementing the multi - cell cooperative downlink beamforming .
beamforming , cooperative multi - cell system , interference channel , multi - antenna , pareto optimal , rate region . [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] |
a large class of phenomena in biology , chemistry , engineering and physics occurs via the transport of particles driven along periodic substrates by an external bias .
these phenomena include polymer diffusion at interfaces @xcite , motion of fluxons in superconductors containing a periodic arrangement of defects @xcite , adsorption on crystal surfaces @xcite , super - ionic conduction @xcite , motion of molecular motors along microtubules @xcite , granular flows on rough inclined substrates @xcite or in a narrow pipe @xcite .
the study of transport properties of granular systems represents an open issue in statistical mechanics of considerable interest and difficulty in view of the continuous energy dissipation caused by particle inelastic interactions @xcite . in this perspective
, we consider a simple model consisting of a granular fluid moving on a tilted rough substrate , which may favor clustering and jamming behaviors . in our formulation , a granular system is a large number of particles ( grains ) colliding with one another and losing a little energy in each collision @xcite .
if such a system is shaken to keep it in motion its dynamics resembles that of fluids as the grains move randomly .
@xcite we carry out a comparative study of the behavior of three models : the inelastic particle system ( ips ) , the elastic particle system ( eps ) , and finally the non interacting system or independent brownian particles ( ibp ) to understand how the interactions influence the collective transport .
the absence of interactions between particles makes the basic phenomenology of the ibp well understood .
it reduces , indeed , to the motion of a single particle in a force field resulting from the interplay between a constant force @xmath0 , such as gravity or electric field , a spatially periodic force , simulating the presence of a rough substrate , a viscous force accounting for the friction , and a noise term , representing thermal fluctuations @xcite .
the scenario is the following : particles diffuse with a bias in the direction of the steady force , @xmath0 , with an average velocity , @xmath1 , which is an increasing function of the applied load @xmath0 .
however , when the load is below a critical value , @xmath2 , and the noise sufficiently small , the average velocity tends to zero . in other words
the particles remain locked in the minima of the periodic potential . on the other hand , above the threshold , @xmath2
, the particles may travel from one minimum to the other along the tilt direction .
the critical value , @xmath2 , depends on roughness , friction and temperature . according to the value of the friction coefficient two different scenarios
may be observed . in the so called overdamped limit
, @xmath2 occurs only when the tilt is such that the local minima completely disappear . in the opposite limit , _
i.e. _ , in the underdamped regime , the particles may overcome many barriers even for low values of @xmath0 at which the potential still displays local minima .
in fact , the particles may cross a barrier , provided their gain in potential energy , to go from one minimum to the next , exceeds the energy dissipated along this pathway . in the underdamped limit the exact value of @xmath2 depends on the temperature .
two indicators may serve to characterize the collective transport properties of the system , the mobility , @xmath3 , and the diffusion constant , @xmath4 .
the former relates the average velocity to the tilt @xmath0 according to @xmath5 .
the latter measures the average spreading of particles : @xmath6 ^ 2 \rangle\;. \label{diffusion}\ ] ] the average is meant over the thermal noise realizations . in all the cases where @xmath7 asymptotically grows linearly in time
, we can identify the normal diffusion constant @xmath4 defined as the slope of the law @xmath8 that generally is computed numerically through simulations of the system evolution .
the study of the dependence of @xmath3 and @xmath4 on @xmath0 and other system parameter is necessary to determine the efficiency of the transport , i.e. the ratio between drift and spreading of a swarm of particles .
a detailed theory capable of describing the behavior of the ibp for arbitrary values of the damping has been developed by risken @xcite who derived an analytical expression for the mobility , @xmath3 , as a function of noise level , tilting force and substrate characteristics .
a closed formula for the diffusion coefficient exists only for the overdamped situation @xcite .
it predicts the presence of a striking enhancement of @xmath4 near a threshold , @xmath9 , separating the locked from the running phase and corresponding to the disappearance of the local minima of the potential . on the other hand , in the underdamped regime
there is no analytic expression for @xmath4 , but the enhancement of the diffusion coefficient , occurs at a lower threshold , @xmath10 @xcite due to the effect of inertia , and the phenomenon remains qualitatively similar .
the goal of the present paper is to consider the effects of interactions among particles on the transport properties in such systems .
specifically , we will address the following question : how do interactions occurring via elastic / inelastic hard - core collisions affect the mobility and the diffusion of an assembly of particles ?
this issue is particular relevant in one dimension where a repulsive hard - core interaction inhibits the particles to pass each other ; this constraint is known to influence dramatically the dynamics of a group of particles . as an example , we can mention the anomalous self - diffusion in single file systems @xcite .
we focus our dynamical approach only on exploring the effects of repulsive interactions via impulsive contact collisions .
we consider two possibilities , energy conserving collisions and dissipative inelastic collisions .
both situations are still largely unexplored in systems with washboard potentials and display non - trivial behaviors , as we shall illustrate .
for instance , the mutual repulsion between the particles induces dynamical correlations which may favor or hinder their motions : it can promote their exit from a potential well via energetic collisions , or , on the contrary , it can prevent a jump to a given site when this is too crowded .
in addition , the granular temperature , defined as the average kinetic energy per particle , in inelastic systems is in general lower than the corresponding temperature of elastic systems .
therefore , the transport , occurring via thermally activated processes across substrate barriers , is expected to be less efficient for inelastic particles @xcite
. then few questions can be addressed .
does the diffusivity @xmath4 present an enhancement analogous to that of non interacting systems ? does the threshold locked - to - running occur at larger values of the tilt ? in the following ,
we answer the questions by examining a variety of situations and analyzing the overdamped and the underdamped regimes separately .
the paper is organized as follows . in section
ii , we present the model and discuss the contact elastic / inelastic interaction with their main implication in the system dynamics . in section iii , we analyze the results from the molecular dynamics ( md ) simulations of the over - damped regime . in section iv
, we summarize the corresponding results for the under - damped regime discussing the salient differences with the over - damped case .
finally , conclusions are drawn in section v.
we consider a randomly driven granular gas , already introduced in previous works @xcite , composed of @xmath11 impenetrable hard - rods , of mass @xmath12 and size @xmath13 , moving on a line of length @xmath14 in the presence of an external washboard potential .
the overall dynamics of this gas is described by the langevin equation for each rod @xmath15}{d x_i } + f \label{kramers2}\ ] ] that embodies four types of physical phenomena : friction with the surroundings , random accelerations due to external driving , inelastic collisions among the particles and external time independent , but spatially varying , force .
we model the first two effects by means of a stochastic bath with a viscous friction @xmath16 and gaussian random force , with zero average and covariance @xmath17 satisfying a fluctuation - dissipation relation , with @xmath18 proportional to the intensity of the stochastic driving @xcite .
the damping term fulfilling einstein s relation renders the system stationary even in the absence of collisional dissipation and physically can represent the friction between the particles and the container .
the interactions among the particles , indicated formally by @xmath19 in eq .
( [ kramers2 ] ) , amount to simple impulsive forces acting on the rod @xmath20 due to the collisions with the neighboring particles @xmath21 @xcite .
thus a rod @xmath20 performs , under the influence of the bath and collisions only , independent brownian trajectories , unless it gets in contact with particle @xmath21 , i.e. @xmath22 , at which point the velocities of the colliding pair @xmath23 change instantaneously according to the inelastic rule @xmath24 @xmath25 the prime indicating post - collisional variables and @xmath26 the coefficient of restitution , with @xmath27 . the total force , in eq .
( [ kramers2 ] ) , on each particle is supplemented by terms representing the presence of a constant external bias , @xmath0 , and a periodic substrate generating a potential @xmath28 , with period @xmath29 and amplitude @xmath30 , @xmath31 a quantitative criterion to distinguish between an overdamped and an underdamped regime is provided by the dimensionless parameter @xmath32 , where @xmath33}_{min}\ ] ] is the oscillation frequency at the potential minimum .
if @xmath34 the dynamics is overdamped , and in the opposite limit is underdamped .
as discussed by borromeo and marchesoni @xcite , the present model is equivalent to consider the particles in a traveling potential , i.e. a periodic potential moving at speed @xmath35 , which is related to the tilt force by the relation @xmath36 . interestingly for speeds lower than a certain threshold the traveling wave has the capability of dragging the particles , a mechanism known as stokes drift @xcite . thus , the system can equivalently describe the physics of a one dimensional granular gas where the periodic potential @xmath28 represents a series of compartments separated by walls , an experimental set - up recently employed to study the clustering behavior of vibrated granular gases @xcite , or the roughness of an inclined plane @xcite .
we shall begin by considering the nearly overdamped regime ( @xmath37 ) , whose study is better understood because the system reaches a steady state rapidly due to the large value of the friction .
our md simulations were carried out by evolving an initial configuration , where the particles were all located in the central well without overlaps and their velocities were extracted from a maxwell - boltzmann distribution of temperature @xmath18 .
each run involved @xmath38 particles of size @xmath39 , mass @xmath40 , with three different values of the coefficient of restitution , @xmath41 , @xmath42 , and @xmath43 , respectively . the substrate potential was characterized by wells of width @xmath44 and height @xmath45 , in units of @xmath46 .
finally , the heat - bath temperature has been chosen to be @xmath47 and @xmath48 . the time is measured in units of @xmath49 .
the friction coefficient is @xmath50 .
the cyclic system studied has been taken to be of length @xmath51 so that it is virtually equivalent to a system with open boundary conditions . in this condition of high dilution ,
the global system density @xmath52 is extremely low , however it is not meaningful parameter , rather it is the initial density profile , characterized by the number of particles in a well @xmath53 , that has a strong influence on the system behaviour .
indeed , at the beginning of the evolution , the system needs to be packed enough to reach a not negligible collision rate . only in the later stages of the simulations ,
a crossover is observed toward the behaviour of rarefied gases . at first sight ,
the md simulations of the interacting system display a mobility quite similar to that of the non - interacting system : the curves @xmath54 versus @xmath0 for elastic and inelastic hard - rod systems are shown in fig .
[ fig : mob ] . in the same figure
, we also plot the corresponding quantity relative to the ibp as calculated numerically by means of the continued fraction method @xcite . as a function of the rescaled external force @xmath55 ( @xmath56 ) for @xmath57 ( circles ) and @xmath58 ( squares ) for a system with @xmath38 particles , temperature @xmath59 , friction @xmath50 , and potential amplitude @xmath60 , @xmath39 , @xmath61 .
the system evolution is simulated for @xmath62 time units .
the full line indicates the corresponding curve for the non interacting particles obtained by the continued fraction method @xcite.,width=321 ] however , a closer inspection reveals that differences do exist between the ibp and the interacting systems with different inelasticity @xmath26 .
figure [ fig : rapmob ] shows the ratio of the mobilities of interacting systems to the mobility of the ibp .
this ratio , for small values of the load , @xmath63 , can be significantly different from @xmath64 , and more specifically the mobility of the eps exceeds the mobility of the ibp , whereas the mobility of the ips is lower .
the reason for these differences can be found in the fact that contact interactions may change drastically , with respect to the ibp , the time that particles spend in a given potential minimum .
( @xmath56 ) for @xmath65 ( circles ) , @xmath66 ( squares ) and @xmath41 ( triangles ) .
the remaining parameters are the same as in fig .
[ fig : mob].,width=321 ] in more detail , these behaviors can be explained by recalling that two competing mechanisms contribute to the system mobility : the excluded volume and the inelasticity .
the first leads to an effective reduction of the barrier height and favors the escape from the wells , thus basically increasing the mobility .
the second , instead , tends to decrease the average kinetic energy , rendering longer , on average , the time spent by the particles in a given well . for low values of the forcing field @xmath0
, the mutual repulsion dominates and thus we observe a larger mobility of the eps with respect to the ibp at the same values of @xmath0 and @xmath18 .
however , as we switch the inelasticity on , the mobility of the ips can become lower than that corresponding to the ibp . for @xmath67
, @xmath3 turns out to be the same for all systems because in this regime only the forcing field @xmath0 matters .
in fact , the absence of minima in the total potential reduces the influence of the mutual interactions so the particles follow coherently the strong effect of the drift .
the role of the excluded volume can be highlighted by monitoring the trajectories of some specific tagged particles whose dynamics can be very representative of the evolution of the whole system .
more specifically , we consider the two extremal particles , @xmath64 and @xmath11 ( having labeled the rods along the drift direction from @xmath64 to @xmath11 ) and the central particle @xmath68 .
the first particle is , on average , braked by the collisions of the preceding particles , whereas the @xmath11-th particle is pushed ahead by the pressure of those behind .
their overall behavior is expected to be qualitatively rather different from that of the central particle . in figure
[ fig : raptraj ] , we display , for @xmath69 and @xmath70 , the evolution of the mean displacement of these three tracer particles from their initial position .
each displacement is normalized with respect to the displacement of the center of mass of the whole system from its initial position : @xmath71 with @xmath72 and @xmath11 .
all the averages are performed over @xmath73 independent runs .
we consider , first , the eps for three values of the tilt @xmath0 , first column of fig .
[ fig : raptraj ] . for all values of @xmath0 , the central particle moves with velocity very close to the center of mass velocity @xmath74 as shown by the fact that @xmath75 stays almost pinned around the value @xmath64 .
the velocities of the extremal particles , instead , are very different from @xmath74 only during an initial transient when the system remains compact .
indeed , it is clear that , in a system not yet too diluted , the motion of the first particle is frequently hindered by the others , while the motion of the last is favored .
asymptotically , the differences in the motion of rods @xmath64 , @xmath76 and @xmath11 become less evident because the interactions become less effective as the packet spreads over and over .
notice , also , the asymmetry between the first and the last particle .
the major change observed when the inelasticity is turned on ( @xmath77 ) is that the velocity of the last particle displays a pronounced deviation from @xmath74 .
these different behaviors are also evident by an inspection of the shapes of the instantaneous coarse grained distributions of particle positions ( fig .
[ fignx ] ) , @xmath78 computed , in the simulations , by binning the number of particles to widths of the size of a single potential well . with @xmath79 and @xmath80 .
the circles correspond to the elastic case ( @xmath65 ) with @xmath39 , while the squares and triangles to @xmath66 and @xmath41 , respectively .
the diamonds correspond , instead , to the elastic particles of size @xmath81 .
the distributions extend only over few hundred wells , while the system size is much larger @xmath82 .
inset : the trend of the skewness as a function of @xmath83 in the same cases .
the other parameters are the same as in fig .
[ fig : mob].,width=321 ] in the eps , the asymmetry of the packet is determined only by the excluded volume effects and , when the diameter of the rods increases from @xmath13 to @xmath84 , both asymmetry and drift velocity become more pronounced .
the dissipation , favoring clustering , makes the packets of the ips more compact and , in addition , changes the direction of the asymmetry , as seen for @xmath66 and @xmath41 , in fig .
[ fignx ] .
this effect can be quantified by means of the skewness of the particle distribution around their center of mass , @xmath85 where @xmath86 is the standard deviation and @xmath87 the center of mass position .
positive values of @xmath88 entail distributions with an asymmetric shape extending out toward more positive tails .
negative values are associated with distributions extending out toward more negative tails @xcite . in the inset of fig .
[ fignx ] , we plot the parameter @xmath89 as a function of time for three systems with different inelasticities .
let us consider , now , the quantity @xmath90 in the ibp , @xmath4 is constant and corresponds to the diffusion coefficient .
we have found numerically that such a behavior persists both in the eps and ips for all values of @xmath26 we explored , as clearly indicated by the linear growth of @xmath7 in the inset of fig .
[ fig : diff - sv ] . , ( @xmath91 ) as a function of @xmath55 .
each point is the result of an average over @xmath92 independent runs .
the circles correspond to the elastic case ( @xmath65 ) , while the squares and triangles to @xmath66 and @xmath41 , respectively .
the full line represents the theoretical result obtained by reimann et al .
inset : standard diffusion of the system rescaled spreading , @xmath93 , [ see eq .
( [ diffusion ] ) ] with respect to the dimensionless time @xmath83 in the same cases for @xmath79 .
the other parameters are the same as in fig .
[ fig : mob].,width=321 ] the interactions change quantitatively the dependence of @xmath4 on @xmath0 , in fact , as @xmath0 varies , the diffusion coefficient displays a maximum for @xmath94 ( see fig . [
fig : diff - sv ] ) , with a behavior similar to that found by reimann et al . in the ibp @xcite . as @xmath26 decreases , however , we observe some differences . in particular ,
the larger diffusivity of the eps with respect to the ibp is basically explained by the excluded volume repulsion . at first
, the diffusivity lowers as @xmath26 decreases , but a further decrease of @xmath26 determines a new enhancement of @xmath4 ( fig . [
fig : diff - sv ] ) .
such a feature is a consequence both of the larger broadening and of the larger asymmetry of the packet occurring as the particle size grows ( see fig .
[ fignx ] ) ; as a matter of fact , there is a huge increase of the diffusion since , for @xmath79 , we found @xmath95 ( these data are not shown ) .
when the damping force , in eq .
( [ kramers2 ] ) , is smaller or comparable with respect to the inertial term @xmath96 the situation changes drastically . at low friction
, the particles can travel across several wells before being trapped in a minimum .
again , as a guide , we can use the results obtained by risken in his thorough study on the ibp in washboard potentials , providing a full theoretical treatment of the mobility as a function of @xmath97 and temperature .
risken s theory shows the existence of two dynamical states : _ locked _ and _ running_. at @xmath98 , a characterization of the dynamical behavior of the non interacting system is straightforward .
indeed , for @xmath99 only running states exist .
when @xmath100 , with @xmath101 particles may be either locked or running depending on the initial velocity and position .
this fact determines a hysteresis loop in a @xmath3 vs @xmath0 diagram .
when the temperature is finite , the particles can switch from one state to the other under the influence of the thermal bath . as a result
, the hysteresis is suppressed and the locked - running transition occurs smoothly as a function of the tilt @xmath0 , being the @xmath102 curve a sigmoid in @xmath0 . only as @xmath103 the sigmoid becomes a step function whose discontinuity is located at @xmath104 .
md simulations of the underdamped regime at different values of @xmath0 allowed us to determine the curve of @xmath3 as a function of the tilt , shown in fig .
[ fig : mob - ud ] for the eps ( @xmath105 ) and the ips with @xmath41 at temperatures @xmath106 and @xmath107 . the mobility , @xmath3 , appears to be roughly similar to that of the overdamped regime , but
the critical @xmath0 separating locked and unlocked situations lies at values lower than @xmath9 and depends both on @xmath97 and @xmath18 . in the eps ,
the mobility is rather close to the ibp value , while for the ips the mobility is reduced , as shown explicitly in the case @xmath41 in fig .
[ fig : mob - ud ] . as a function of the external rescaled force @xmath108 for different temperatures and coefficients of restitution .
the open and closed symbols refer to @xmath107 and @xmath106 , respectively .
the squares indicate the elastic case and the triangles the case with @xmath41 .
the other parameters of the system are @xmath109 , @xmath38 , @xmath60 , @xmath110 , @xmath61 and @xmath111 .
the full line indicates the corresponding curve for the non interacting particles , obtained via the continued fraction method by risken.,width=321 ] while , the mobility of particles interacting inelastically displays , in the underdamped regime , no peculiar behavior with respect to the ibp and eps , their collective diffusion presents some anomalies . at low temperatures , the dynamics
strongly depends on the initial conditions , the evolution of the packet is very sensitive to the tilt and exhibits a rather inhomogeneous and irregular structure .
this can be readily visualized by looking at the time behavior of the coarse grained particle density @xmath112 [ eq .
] at temperature @xmath113 shown in figures [ fig : n_xt_08](a ) and [ fig : n_xt_08](b ) for @xmath114 and @xmath115 , respectively .
it indicates that , for @xmath114 , particles remain partially trapped in the well where they have been initially deposited and only a fraction of them escape acquiring a drift .
for comparison we also show the corresponding behavior of @xmath112 for the eps , where the diffusion is standard .
interestingly , the inelastic distribution @xmath112 corresponds to a lower mobility , but to a larger spread of the particles .
moreover , one can see that new clusters , indicated by the spikes in fig .
[ fig : n_xt_08](a ) , spontaneously form and persist for long periods before being dissolved .
such clustering phenomena are favored by the moderate value of the tilt .
the situation , instead , looks different at @xmath115 [ see fig . [
fig : n_xt_08](b ) ] , where the initial cluster `` evaporates '' earlier , and the formation of new clusters is prevented by the effects of the drift @xmath0 .
one observes a clear difference with respect to the behavior of the overdamped ips illustrated in fig .
[ fignx ] showing a more compact structure of @xmath112 which does not lose `` debris ''
. plotted as a function of the dimensionless quantity @xmath116 , for @xmath114 ( panel a ) and @xmath115 ( panel b ) , for the ips with @xmath41 and eps ( broken line ) .
the remaining parameters are the same as in fig .
[ fig : mob - ud ] .
we notice that the system size is @xmath82 much larger than the well width , thus the peaks are not commensurate with the system size.,width=321 ] such an early stage is sufficient to determine a late collective transport characterized by an anomalous spreading .
the closely packed initial configuration has deep repercussions on the late spreading , @xmath7 .
this quantity , despite a very long simulation , does not appear to reach a linear or any other simple functional asymptotic dependence on time ( fig .
[ fig : sqrm ] ) .
the situation described above corresponds to an early very steep growth of @xmath7 , growth that becomes slower after a characteristic time @xmath117 whose duration decreases with the tilt amplitude .
as shown in fig .
[ fig : sqrm ] , the vertical position of the knee strongly depends on the number of particles . by no means the @xmath7 shows the expected linear behavior of standard diffusion .
such a feature is present only in the inelastic systems .
however , when the temperature is raised , even the ips recovers a linear behavior as seen in fig .
[ fig : sqrm](b ) .
the estimated diffusion coefficient is shown in fig .
[ fig : diff - ud ] . , , for the ips with @xmath41 .
panel ( a ) reports three runs with @xmath114 , @xmath107 and @xmath38 ( circles ) two runs with @xmath118 ( squares ) and two runs with @xmath119 ( triangles ) .
the full line is the average of the same quantity over @xmath92 independent runs involving @xmath38 particles .
panel ( b ) shows @xmath93 vs @xmath83 at temperature @xmath106 for different forces : @xmath120 ( circles ) , @xmath121 ( squares ) and @xmath122 ( triangles ) .
the curves represent an average over five runs .
the remaining parameters are the same as in fig.[fig : mob - ud].,width=302 ] ( @xmath91 ) , for different coefficients of restitution , as a function of @xmath108 . the @xmath4 values represent averages over ten trajectories .
the squares correspond to the elastic case ( @xmath65 ) , while the triangles to @xmath41 .
the temperature is @xmath123 , the remaining parameters are the same as in fig . [
fig : mob - ud ] .
the dashed line indicates the result relative to the ibp at a lower temperature @xmath113 , showing that the effect of inelastic collisions can not be accounted for by a mere reduction of the effective temperature of the system .
the inset , finally , displays the ratio between the kinetic and the heat bath temperatures.,width=321 ] the peak of @xmath4 relative to the ips is higher than the corresponding peak of the eps .
the larger value of the ips peak can be associated with the shape of the phase - space distribution , @xmath124 , which is defined as the probability of finding a particle at distance @xmath125 from the center of mass and with velocity @xmath126 with respect to center of mass velocity @xmath74 .
+ the enhancement of the diffusion in the ibp is determined by the locked - running bistability , i.e. by the existence of two peaks at velocities @xmath127 and @xmath128 in the velocity distribution . at the transition both populations are present and the center of mass velocity does not represent the most probable velocity in the system . hence , it is clear that in this situation the variance of the velocity distribution can be very large . on the other hand , when one of the two populations becomes dominant the variance tends to decrease .
the spatial part of the distribution @xmath124 does not play any role in the ibp as the @xmath129-dependence of the distribution is symmetric with respect to the center of mass coordinate .
the scenario changes in the interacting cases , since in one dimension , the excluded volume plays a fundamental role in the dynamics .
as we have seen in the inelastic case , at low temperatures the interaction was such as to lead to a fragmentation of the system and to a non linear behavior of the diffusion .
as @xmath18 increases the fragmentation decreases and @xmath7 display a linear behavior with respect to @xmath130 . in order to explain the different values of @xmath4 near the peak between the ips and the eps , we have collected from the simulations the double histograms @xmath124 .
the comparison is reported in fig .
[ fig : sei ] for drift values @xmath0 near the peak . analyzing the different behaviors of @xmath124 in the various cases
may help clarify the role of the interactions . in the eps for loads @xmath121 and @xmath131
, @xmath124 displays the same velocity bistability as the ibp and no spatial asymmetry and indeed the diffusion in these two systems results very similar as shown in fig .
[ fig : diff - ud ] . in the ips , instead , for the same loads we observe a strong spatial asymmetry which enhances the diffusion
. in the ips , the interactions besides determining the peak of the diffusion
also determine its width .
from our simulations , as shown in fig .
[ fig : diff - ud ] , we see that the peak occurs at lower values of @xmath0 with respect to the ibp and eps and the width is much broader .
this means that it is not possible to reproduce the behavior of the ips by an appropriate choice of an effective temperature accounting for the inelasticity . a smaller temperature , in fact , would give a higher peak , but would not change the value of @xmath0 at which it occurs .
moreover , the width of the peak decreases with @xmath18 and fails to reproduce the observed broadening in the ips , which is due to the inter - particle interactions . in the inset of fig .
[ fig : diff - ud ] , we display the ratio between the kinetic temperature @xmath132 and the heat bath temperature @xmath18 .
notice that in both cases the peak occurs in correspondence of the peak of the curve @xmath133 of the elastic system and the ratio approaches , as expected , the value @xmath64 at low and high bias @xmath0 .
in summary , in this work we have reported a numerical study on the dynamics and the transport of an array of identical particles in a inclined washboard potential .
we investigated , at several tilting @xmath0 , the role of hard - core repulsive interactions among the particles on both the mobility and the spreading of the system with respect to its center of mass , indicators that are customarily used to characterize collective transport is such systems .
we compared the behavior of three models involving respectively : independent brownian particles ( ibp ) , elastic particles ( eps ) and inelastic particles ( ips ) .
we have found that , in the high friction regime , the transport behavior of the system with interacting particles is qualitatively similar to that of the non interacting system ( ibp ) .
basically , the interactions do not strongly affect the mobility which remains close to the ibp value , while they modify the diffusivity that appears to be larger .
this conclusion is intuitive , if we consider that @xmath3 is related the center mass velocity of the system , a quantity rather insensitive to the presence of interparticle interactions . to achieve a a finer information about the structure of the collective motion
, we monitored the evolution of the particle spatial distribution .
such a distribution strongly depends on the choice of the geometrical parameters and of inelasticity .
this analysis shows that an initial localized packet of particle spreads differently in the inelastic system from the elastic one .
the scenario in the underdamped regime is more complex .
the eps systems , in fact , are similar to ibp either analyzing the mobility , or the diffusion of the particles .
the inelastic interactions practically do not affect the mobility , whereas determine a marked change in the behavior of the diffusion at low temperatures and forces .
the spreading , @xmath7 ( eq . ( [ diffusion ] ) ) , presents a clear two stage behavior : it increases nonlinearly in a transient regime which is then followed by a nonlinear growth , as shown in fig .
[ fig : sqrm ] . at higher temperatures
, however , the diffusion of the inelastic particles exhibits a single regime , with a linear growth typical of a standard diffusion .
our study indicates , moreover , that a group of particles interacting via inelastic hard - core collisions spread over much more than elastic particles .
the mobility @xmath3 seems to be a global indicator that poorly encodes the information about the details of the interactions . on the contrary ,
the diffusion coefficient @xmath4 is a much more representative observable being more sensitive to the presence of elastic or inelastic interactions .
we believe that in the overdamped regime , it should be possible to apply theoretical methods already employed in the study of dense molecular fluid in narrow channels @xcite , whereas the treatment of the underdamped case remains more problematic .
the aim of the present numerical work is , perhaps , to stimulate further investigation and new theoretical proposals . | we study numerically the motion of a one dimensional array of brownian particles in a washboard potential , driven by an external stochastic force and interacting via short range repulsive forces .
in particular , we investigate the role of instantaneous elastic and inelastic collisions on the system dynamics and transport .
the system displays a locked regime , where particles may move only via activated processes and a running regime where particles drift along the direction of the applied field . by tuning the value of the friction parameter controlling the brownian motion we explore both the overdamped dynamics and the underdamped dynamics . in the two regimes we considered the mobility and the diffusivity of the system as functions of the tilt and other relevant control parameters such as , coefficient of restitution , particle size and total number of particles .
we find that , while in the overdamped regime , the results for the interacting systems present similarities with the known non - interacting case , in the underdamped regime , the inelastic collisions determine a rich variety of behaviors among which is an unexpected enhancement of the inelastic diffusion . |
the mapping of optical information into an atomic ensemble , a phenomenon usually named light storage ( ls ) , plays a fundamental role for both classical and quantum information processing @xcite . to date several observations of ls have been reported both in thermal and in cold atomic systems @xcite .
long storage times as well as the possibility of controlling and manipulating the stored information @xcite , is of essential importance for the realization of any practical quantum protocol . in most of the experimental observations of ls ,
ground state hyperfine levels of alkali atoms are employed to store the optical information into zeeman coherences .
these coherences may be conveniently manipulated by external magnetic fields @xcite .
however , they are also very sensitive to stray magnetic fields gradients , which can strongly reduce the coherence time of the system @xcite . collapses and revivals due to larmor precession of zeeman coherence associated with different hyperfine ground states of rubidium atoms
have already been observed , both using poissonian classical light and single photon quantum states @xcite .
recently , we have demonstrated the coherent larmor precession of orbital angular momentum of light stored into zeeman coherences of cesium atoms @xcite .
it is also worth mentioning the recently demonstrated long - lived gradient echo memory , which is a different technique for ls and allows the control of the retrieved pulse sequence @xcite . in this work
we investigate the coherent evolution of a light grating stored onto the ground state zeeman coherence in the presence of an external dc magnetic field .
the presence of the magnetic field give rise to a periodic evolution of the atomic state ( larmor precession ) that determines an oscillatory response of the retrieved pulse efficiency as a function of the time separation between the writing and reading of the coherence grating .
nevertheless , the coherent response of the sample is damped as a consequence of the dephasing of the individual atomic contributions due to residual magnetic field inhomogeneities .
interestingly enough , the damping time of the total atomic response strongly depends on the direction of the applied magnetic field .
we have observed a five - fold increase in the coherence decay time , with respect to the damping time at zero applied field , for a particular direction of the applied magnetic field .
precise theoretical modelling of our observations is prevented by the lack of a priori knowledge of the residual magnetic field in the sample .
nevertheless , the essential features of the experimental results are qualitatively reproduced by a simplified theoretical model , based on optical bloch equations , describing the atomic state evolution in the presence of a homogeneous magnetic field together with an ad hoc inhomogeneous field .
this model accounts for the oscillations ( revivals ) of the retrieved pulse amplitude as well as for the variation of the storage decay time with magnetic field direction .
the key role played by the relative orientations of the mean magnetic field and its inhomogeneity is also illustrated through the classical calculation of the precession of an ensemble of magnetic dipoles in the presence of an inhomogeneous field .
the experimental apparatus is similar to the one previously described in @xcite and employs cold cesium atoms obtained from a magneto - optical trap ( mot ) .
the atoms are initially prepared into the @xmath0 hyperfine ground state by optical pumping induced by the mot trapping beams , with the repumping beam switched off for about 1 ms .
the mot quadrupole magnetic field is switched off during the optical pumping process and the subsequent light storage sequence .
three pairs of orthogonal helmholtz coils are used to compensate for stray magnetic fields . to write and read the coherence grating we use light from an external cavity diode laser ,
locked to the cesium d2 transition @xmath1 . immediately after the optical pumping interval ,
two laser beams are incident on the atomic cloud making a small angle of @xmath2 mrad as indicated in fig .
these two beams , named from now on as the grating writing beams , @xmath3 and @xmath4 , have opposite circular polarizations and induce a zeeman coherence grating between pairs of zeeman sublevels , as indicated in the fig.1 ( a ) .
the writing beams have their intensity controlled by a pair of acousto - optic modulators ( aom ) and are applied during a time long enough ( @xmath5s ) so to allow the system to reach a steady state regime . during the coherence - grating writing interval ,
an external dc magnetic field is applied , either along the propagation direction of the @xmath3 beam ( chosen as quantization axis in fig .
1 ) or orthogonal to this axis , by two additional independent pairs of helmholtz coils shown in fig.1 . after the switching off of the writing beams
, the induced zeeman coherence grating evolves in the presence of the external magnetic field undergoing larmor precession . the instantaneous zeeman coherence grating can be used to retrieve the optical field ( the diffracted beam @xmath6 ) by switching on a reading beam , @xmath7 , counterpropagating to the writing beam @xmath3 and having a circular polarization opposite to that of this beam , as shown in fig .
for this beam geometry , the retrieved light pulse is phase conjugated to the writing beam @xmath4 and couterpropagating with respect to this beam .
the reading beam also passes through an independent pair of aoms that allows for the variation of the time interval between writing and reading processes as indicated in the time sequence shown in fig .
we have experimentally verified that the retrieved beam has a polarization which is orthogonal to the polarization of the reading beam as indicated in fig .
1-(b ) . the retrieved signal intensity for different storage times is shown in fig.2-(a ) corresponding to the polarization configuration specified in fig .
1-(a),(b ) for the case where no external dc magnetic field is applied . for these data , the intensities of the writing beams
@xmath3 and @xmath4 are equal to 17.6 mw/@xmath8 and 1.1 mw/@xmath9 respectively , while the intensity of the reading beam is equal to 7.2 mw/@xmath9 .
as can be observed , the retrieved signal peak intensity has an approximately exponential decrease with a free decay time of the order of 4 @xmath10s @xcite .
this measured coherence decay time is mainly determined by the non compensated residual magnetic field gradients @xcite .
the mot temperature was estimated to be in the range of mk and the measured density of atoms in the f=3 ground state was of the order of @xmath11 @xmath12 . from these values , we estimated the collision rate to be of order of @xmath13 khz @xcite , therefore having a negligible effect in the decoherence process . also , at the estimated temperature , the time needed for an atom to move one grating period is much larger than the measured coherence time , therefore ruling out the effect of atomic motion .
the dynamics of the retrieved signal changes completely when a dc magnetic field of magnitude @xmath14=0.7 g is applied perpendicularly to the propagation axis , i.e. , along the @xmath15 direction .
the observed result is presented in fig .
2-(b ) , which clearly shows a periodic oscillation of the retrieval efficiency with a period of approximately @xmath16=4.4 @xmath10s . on the other hand , as shown in fig .
2-(c ) , if the dc magnetic field is applied along the @xmath17 axis , we also observe a periodic modulation of the retrieved signal , but now with smaller visibility and a period given approximately by @xmath16/2=2.2 @xmath10s .
these values are consistent with larmor frequency predicted for the hyperfine level @xmath18 , which is given by @xmath19 , with @xmath20 the land factor , @xmath21 the bohr magneton and @xmath22 the magnitude of the magnetic field @xcite . from the applied magnetic field in our setup
we estimate a larmor period of @xmath23=4.1 @xmath10s , which agrees reasonably well with the observed value . and
@xmath4 ) configuration and partial zeeman level scheme of the hyperfine transition @xmath24 , showing the coupling of the writing beams .
( b ) spatial configuration and coupling of the incident reading ( @xmath7 ) and the retrieved diffracted ( @xmath6 ) beams .
the beams @xmath3 and @xmath4 make a small angle @xmath25 and are circularly polarized with opposite handedness , while the beam @xmath7 is counterpropagating to the beam @xmath3 and have a circular polarization opposite to this beam .
the diffracted beam is detected in a direction opposite to the beam @xmath4 .
( c ) switching time sequence for the writing and reading fields . ]
= 0.7 g dc magnetic field is applied along the @xmath15 direction ; ( c ) a @xmath26=0.7 g dc magnetic field is applied along the @xmath17 direction . each graphic represents a superposition of the retrieved pulses associated with different reading time . ] a simple picture can be given for a qualitative understanding of the observed behavior by considering the evolution of the coherent superposition of zeeman sublevels , prepared by the writing beams , in the presence of the applied magnetic field .
considering the larger intensity of beam @xmath3 with respect to @xmath4 , one should expect orientation of the atomic system with the largest population in the zeeman sublevel with the highest @xmath27 value .
thus , in the former case where the magnetic field is perpendicular to the quantization direction , the magnetic hamiltonian will couple the zeeman states differing by @xmath28 and the probability amplitude associated to each component of the initial superposition state will oscillate at the larmor frequency , therefore repeating itself after the larmor period .
it is worth noticing that after half the larmor period the evolution changes the probability amplitude associated with a sublevel @xmath27 into that of @xmath29 and vice - versa , therefore also repeating the value of the zeeman coherence .
however , at this instant the reading beam will be coupled to a different transition , which can lead to a decrease of the diffracted signal .
indeed , in the limit of strong saturation of the reading beam , the signal is reduced by the ratio between the two clebsch - gordan coefficients of the transitions coupled by the reading in the two different situation @xcite . differently , when the dc magnetic field is applied along the quantization direction , each component of the initial superposition will simply acquire a phase factor proportional to @xmath30 and the zeeman coherences associated with sublevels differing by @xmath31 will therefore oscillate with a period @xmath32 .
this simple consideration accounts qualitatively well for the series of collapses and revivals observed experimentally .
in addition , we have performed similar measurements for the case where the writing beams have orthogonal linear polarization with the reading beam being linearly polarized parallel to the polarization of the writing beam @xmath4 . in this case
, alignment is produced in the atomic system leading to a symmetric distribution of the zeeman population . in such case , for an applied transverse magnetic field , after half the larmor period all the zeeman coherences and populations repeat themselves and therefore the retrieved signal oscillates with half the larmor period .
this behavior was experimentally observed .
finally , we also have measured the evolution of the zeeman coherence grating for different values of the external magnetic field . in fig .
3 we plot the inverse of the larmor period as a function of the transverse magnetic field amplitude . as expected , a linear dependence is observed . the most remarkable feature of the measurements presented in fig .
2 is the observed increase of the coherence decay time when the dc magnetic field is applied . as mentioned before , the measured coherence decay time is mainly determined by stray magnetic field gradients that are not precisely known in our apparatus .
the experimental data shown in fig.2(b ) and fig.2(c ) clearly indicate that a larger increase in the coherence time is obtained when a dc magnetic field is applied along a specific direction ( @xmath15 axis ) , which is orthogonal to the propagation direction of @xmath3 and @xmath7 beams , although a smaller increase is also observed when the magnetic field is applied along the propagation direction of these beams ( @xmath17 axis ) .
in this section we provide a simplified model of the atomic evolution in the presence of the writing and reading light pulses , which gives rise to the emission of the retrieved pulse .
although the model does not intend to precisely fit the experimental observations , the main features of the observed signals are qualitatively reproduced .
the model gives a clear indication of the relevant physical parameters determining the atomic evolution .
the model is based on the numerical solution of the bloch equations for an ensemble of atoms evolving in a dc magnetic field .
the calculation is a direct extension of the one presented in @xcite .
the reader is referred to this publication for details .
we briefly remind here the main lines of the calculation .
two atomic hyperfine levels @xmath33 and @xmath34 are considered with well defined total angular momentum @xmath35 and @xmath36 respectively , connected by an optical transition .
all zeeman sublevels are taken into account .
the excited state @xmath34 has a total decay rate @xmath37 that determines the relaxation of the excited state population and of the optical coherence . unlike in @xcite , however , we do not include in the model any relaxation mechanism for the ground state coherence .
the light field acting upon the atoms is resonant with the atomic transition .
the bloch equations are numerically integrated for successive time intervals , corresponding to the light storage sequence with square pulses .
each interval corresponds to the duration of an applied light pulse . during each interval
, the optical field has a well defined amplitude , polarization and phase . in a given time interval
, the optical field may correspond to a unique ( @xmath7 ) light pulse or be the total field resulting from two light beams ( @xmath3 + @xmath4 ) .
our simplified model does not include light propagation effects across the atomic sample .
the atoms are subject to a single optical field , which can however have one or two polarization components . in consequence ,
no spatial grating is represented in the model . numerically integrating the bloch equations
, we calculate the atomic polarization during the reading time - interval .
we consider the beams @xmath3 and @xmath4 with orthogonal linear polarizations .
beams @xmath4 and @xmath7 have the same polarization .
notice that the choice of linear polarizations is different from the experimental situation previously described .
the reasons for such choice are explained next .
we are interested in the transient atomic response corresponding to the emission of the phase conjugate " retrieved pulse . in the experiment such pulse is characterized by its propagation direction opposite to that of field @xmath4 . since our calculation does not include propagation , phase matching effects determining the emission of light in well defined directions are not accounted for and can not be used to characterize the phase conjugate " retrieval . in consequence , only polarization can be used in the calculation for the identification of the phase conjugate " retrieved pulse .
the corresponding light intensity is proportional to the square modulus of the optical coherence of the atomic sample with polarization opposite to that of fields @xmath7 and @xmath4 .
the fact that in the numerical simulation , only the light polarization can be used to separate the phase conjugate " retrieved light pulse , imposes some constraints in the choice of the light polarizations . in particular , we have to be careful to choose the common polarizations of fields @xmath7 and @xmath4 such that it will not acquire a component along the detected " polarization ( the same as @xmath3 ) as a consequence of faraday rotation under the dc magnetic field .
such effect , that would contaminate the retrieved pulse with a background independent of the previously stored light pulses , is not relevant in the experiment where only the phase conjugate response is phase matched in the direction opposite to that of field @xmath4 .
we have considered in the numerical simulation a linear polarization along axis @xmath15 for field @xmath3 , and linear polarization along @xmath38 for fields @xmath4 and @xmath7 .
the phase conjugate " retrieved pulse is due to the atomic polarization during the reading interval oriented along @xmath15 .
two orientations of the mean magnetic field were considered : @xmath15 and @xmath38 . in cold atoms , where atomic motion and collisions can be neglected , the decay of the coherent response of the atomic sample comes from dephasing of the contributions of atoms in different parts of the sample .
such dephasing is due to the presence of an inhomogeneous magnetic field along the sample that causes different evolutions of the atomic ground state @xcite . in order to simulate the inhomogeneous response of the atomic sample
, we have calculated the evolution of homogeneous atomic sub - samples subject to randomly distributed magnetic fields .
the total response is given by the sum of the optical coherence in each sub - sample .
we have considered in our calculation different orientations of the mean value @xmath39 of the vector magnetic field . however , in order to simplify the calculation and illustrate the importance of the inhomogeneity orientation , we have only considered magnetic field departures from the mean value with a constant orientation ( along @xmath15 ) .
a gaussian distribution of the magnetic field variation was assumed with variance @xmath40 .
figures [ campo0]-[campoy ] show the result of the numerical simulations of the retrieved phase conjugate " light pulse as a function time .
although the simulation reproduces the complete light storage sequence , only the storage and retrieval intervals are represented .
the origin of the time axis corresponds to the simultaneous turn off of the two writing fields @xmath3 and @xmath4 . in each figure , several retrieved pulses are presented corresponding to different durations of the dark ( storage ) interval . for simplicity , the calculations were carried on a model transition @xmath41 .
other choices of the angular momenta give qualitatively similar results .
the amplitudes of the applied fields are determined by the corresponding reduced rabi frequency for the transitions .
the simulations presented in figs .
[ campo0]-[campoy ] correspond to reduced rabi frequencies of @xmath42 , @xmath43 and @xmath44 for fields @xmath3 , @xmath4 and @xmath7 , respectively .
the duration of the writing period ( @xmath3 and @xmath4 both present ) was taken as @xmath45 figure [ campo0 ] show the retrieved pulses obtained for zero mean magnetic field in the presence of a gaussian distributed inhomogeneous magnetic field oriented along @xmath15 with variance @xmath46 .
the magnetic field inhomogeneity is responsible for the observed decay of the retrieved pulse amplitude with a characteristic decay time of the order of a few microseconds .
when , in addition to the inhomogeneous magnetic field distribution , a dc magnetic field with @xmath47 is applied along axis @xmath15 an oscillation is present at twice the corresponding larmor frequency ( fig .
[ campox ] ) .
the envelope decay time of the retrieved pulses amplitude oscillation is similar to that in fig .
[ campo0 ] .
notice that the shape of the individual pulses varies depending on whether the retrieval process occurs during the rising or the falling of the magnetically induced oscillation .
broad pulses are observed on the rising side of the oscillation and narrow ones on the falling side .
a similar effect is observed in the experiment ( see fig .
2 ) . in the presence of the magnetic field ,
individual retrieved pulses also present oscillations that are rapidly damped with a decay time dependent on the @xmath7 pulse intensity .
a dramatically different behavior is seen in fig .
[ campoy ] where the mean magnetic field is oriented along axis @xmath38 while the field inhomogeneity is given by field variations along @xmath15 . in this case
, the oscillation of the peak of the retrieved pulse is clearly visible while the envelope decay is almost unnoticeable on the time scale of the figure .
notice the presence of two frequency components in the oscillation : a larger amplitude oscillation at the larmor frequency and a smaller oscillation at twice that frequency .
figure [ campoy ] dramatically illustrates the importance of the relative orientation of the mean magnetic field and the field inhomogeneity for the effective decay rate of the retrieved pulse amplitude .
when the mean field orientation is perpendicular to the inhomogeneous field , the dephasing of the contributions to the signal from atoms at different locations occurs at a much slower rate .
this effect has a purely classical origin as can be illustrated by considering the ( classical ) motion of an ensemble of magnetic dipoles precessing in a homogeneous magnetic field .
we present in fig .
[ classico ] the evolution of the total magnetic dipole , initially oriented along @xmath38 evolving in a magnetic field with inhomogeneity in the @xmath15 component . if the mean magnetic field is zero or oriented along @xmath15 a relatively short decay of the total dipole is observed . for a mean magnetic field oriented along @xmath17 the evolution of
the total dipole corresponds to a slowly damped precession .
this difference in dephasing time can be simply understood by noting that while an inhomogeneous field oriented along the mean field will produce a first order correction to the total field , a magnetic field variation in a direction perpendicular to the mean field only causes a second order total field modification .
the same argument can be used for the atomic evolution considered previously .
in spite of the simplicity of the theoretical model presented above and the use of a different choice of optical polarizations , several features of the experimental observations are qualitatively reproduced .
an almost exponential decay of the retrieved pulse amplitude is observed in the absence of a mean magnetic field illustrating the inhomogeneous nature of the coherence decay . in this case
all retrieved pulses have the same shape with a characteristic decay time dependent on the reading pulse intensity .
when a dc magnetic field is present , the retrieved pulse amplitude oscillates at the larmor frequency and/or its second harmonic .
this is consistent with the expected evolution of the sample orientation or alignment , respectively . in the general case ,
both frequencies are present in the evolution of the atomic coherence .
the individual retrieved pulses have different shapes depending on the turn - on time of the reading pulse with respect to the magnetic oscillation cycle .
most importantly , the similarities between experimental results and the calculated evolution illustrates the decisive role of the magnetic field inhomogeneity on the coherence signal decay . when the dc field is applied perpendicular to orientation of the local variations of the magnetic field a significant enhancement in the coherence decay can be observed .
such feature may be used for the characterization of unknown field inhomogeneities .
the above result can be seen as a simple means of coherence control in a sample where the decoherence is a consequence of spatial inhomogeneity of the atomic sample .
coherence control has attracted considerable attention in recent times due to its crucial role in the preservation of quantum information . in the case where decoherence results from the coupling of the observed system to a large environment , and is thus homogeneous and irreversible in nature , coherence control
poses a really challenging problem .
significant progress has been recently achieved in this field both with passive and active control protocols relying on the shaping of the interaction of the small system with its environment @xcite . in the present case , however , the decoherence time for the state of individual atoms is long and the shorter observed decay of the coherent emission is the result of the dephasing of the individual atomic contributions evolving in slightly different magnetic environments .
we have shown here that the resulting effective dephasing time can be manipulated by the addition of a constant magnetic field properly oriented .
however , one has to keep in mind that inhomogeneous dephasing is not intrinsically irreversible .
the initial information stored in the sample is , in principle , available in the atomic sample where individual atoms undergo unitary evolution .
the stored information can be recovered , by inducing the rephasing of the atomic contributions through photon echoes @xcite . moreover
, the inhomogeneous nature of the atomic dephasing allows the simultaneous recording in the sample of different light pulse fourier components . as first suggested by mossberg @xcite , this allows the storage and recording of light pulses with complex temporal profiles . while early experiments used the pre - existing sample inhomogeneity , an interesting recent improvement consists in the at - will manipulation of the sample inhomogeneity by the application of a user controlled external magnetic field gradient . in such case ,
storage as well as manipulation ( delay and reversal ) of complex pulse temporal information is possible @xcite .
finally , it is worth emphasizing the significant progress recently achieved in the direction of suppressing the inhomogeneous dephasing altogether by eliminating the influence of the magnetic field through the use of magnetically insensitive two - photon transitions ( clock transitions ) , where storage time of the order of millisecond have been observed @xcite . in summary ,
we have experimentally observed the larmor precession of a coherence grating stored into the zeeman sublevels of the cesium @xmath0 ground state in the presence of an applied magnetic field .
collapses and revivals of the retrieved signal with distinct larmor oscillation frequency were observed for different distributions of zeeman populations and coherences associated with different polarization configuration of the writing beams . an appreciable increase in the coherence decay time
was observed for a specific direction of the applied magnetic field .
we have also presented a simple theoretical model based on the numerical solution of the bloch equation including the zeeman degeneracy which qualitatively agrees with the main experimental observations .
moreover , we have shown that the increasing in the coherence time has a even simpler and purely classical explanation .
finally , we think that the possibility of controlling the coherence time in a larger ensemble of atoms can be of considerable interest for several applications and specially for quantum information .
we gratefully acknowledge marcos aurelio for his technical assistance during the experiment .
this work was supported by the brazilian agencies cnpq / pronex , cnpq / inst .
milnio and finep . | we investigate the evolution of a zeeman coherence grating induced in a cold atomic cesium sample in the presence of an external magnetic field .
the gratings are created in a three - beam light storage configuration using two quasi - collinear writing laser pulses and reading with a counterpropagating pulse after a variable time delay .
the phase conjugated pulse arising from the atomic sample is monitored .
collapses and revivals of the retrieved pulse are observed for different polarizations of the laser beams and for different directions of the applied magnetic field . while magnetic field inhomogeneities are responsible for the decay of the coherent atomic response , a five - fold increase in the coherence decay time , with respect to no applied magnetic field ,
is obtained for an appropriate choice of the direction of the applied magnetic field .
a simplified theoretical model illustrates the role of the magnetic field mean and its inhomogeneity on the collective atomic response . |
research in spintronics and carbon nanotubes sparked renewed interests in the effects of spin - orbit coupling on quantum transport @xcite .
one of the remarkable features of non - interacting disordered electron systems with spin - orbit coupling is the existence of a metallic phase in two dimensions ( 2d ) @xcite .
more exactly the 2d metallic phase appears in systems with symplectic symmetry , i.e. , in systems with time reversal symmetry but without spin rotation symmetry .
such systems are an exception to the prediction of abrahams _ et al . _ that there is no metallic phase in 2d disordered electron systems @xcite .
recently some 2d metals have been discovered experimentally @xcite .
although the mechanism inducing the 2d metal might be different from spin - orbit coupling , we think that detailed understanding of the 2d symplectic systems gives some insights on general transport properties of 2d metals .
a scaling theory for the symplectic systems @xcite can be described by the @xmath0 function for the conductance @xmath1 ( in units of @xmath2 ) @xmath3 where @xmath4 is the linear size of a system .
( the conductance is actually a distributed quantity .
see refs .
@xcite for discussions on the scaling hypothesis of the conductance . ) because of the so - called anti - localization effect , the @xmath0 function shows non - monotonic behavior in the symplectic systems .
the asymptotic behavior in the strongly metallic region in 2d is conjectured to be @xmath5 this indicates that there is a metallic phase in 2d . in the scaling theory for the symplectic systems in 2d
, however , there is a long standing problem for more than two decades @xcite : the logarithmic divergence of the conductance in the 2d metallic phase in the limit @xmath6 indicated by eq .
( [ eq : betaasymptotic ] ) .
since the conductivity and the conductance are the same quantity in 2d , this predicts a perfect conductivity , @xmath7 , in spite of the system being disordered .
this is in contrast to 3d disordered metals , in which the conductivity converges to a finite value .
according to the renormalization group theory of continuous phase transitions , the scaling hypothesis is valid when the correlation length is much larger than all other microscopic lengths .
the scaling argument which deduced ( [ eq : betaasymptotic ] ) , however , is based on the weak localization theory in which disorder is supposed to be weak . in such a weakly disordered metallic phase this condition on the correlation length may not be satisfied . therefore , a numerical check of the scaling hypothesis in the 2d metallic region is desirable . in this paper
, we report a numerical check of the scaling hypothesis in the 2d metallic region .
we also report numerical calculation of the landauer conductance in 2d , which is a direct check of the prediction of a perfect conductivity .
last , numerical simulations of transport through a sierpinski carpet are reported .
among the models with symplectic symmetry , we employ the su(2 ) model @xcite , @xmath8 the random potential @xmath9 is distributed with box distribution in the range @xmath10 $ ] .
random spin - orbit coupling is included in the nearest neighbor hoping term . we parameterize the hopping matrix as @xmath11 and we distribute @xmath12 and @xmath13 with uniform probability in the range @xmath14 , and @xmath15 according to the probability density @xmath16 in the range @xmath17 $ ] . in the actual simulations
@xmath18 s in one direction are set to the unit matrix with the aid of the local su(2 ) gauge transformation .
the su(2 ) model has an advantage that corrections to scaling arising from irrelevant variables are smaller than other models @xcite , so it is quite a useful model when we study universal aspects of anderson localization and the anderson transition .
we analyze the quasi-1d localization length @xmath19 on a quasi-1d strip with width @xmath4 @xcite .
periodic boundary conditions are imposed in the transverse direction .
the scaling hypothesis implies that the dimensionless quantity @xmath20 obeys @xmath21 where @xmath22 is the 2d correlation ( localization ) length which depends on @xmath23 and the fermi energy @xmath24 .
the subscript @xmath25 indicates the metallic and insulating phases .
we have calculated @xmath26 for sizes @xmath27 $ ] with an accuracy from @xmath28 to @xmath29 .
the ensemble transfer matrix method has been used @xcite .
in addition to the data at @xmath30 analyzed in ref .
@xcite , we have also accumulated data at @xmath31 in @xmath32 $ ] . from the numerical data the correlation length @xmath22 at each pair of @xmath33 and the scaling function @xmath34 are estimated using numerical fit described in ref .
the quality of the fit is assessed by the goodness of fit probability @xmath35 . to eliminate the ambiguity of the absolute scale of @xmath22 , we set @xmath36 at @xmath37 . figure [ fig : metalsps1 ] demonstrates the single parameter scaling in the metallic region .
all data fall on a common scaling curve when @xmath26 is plotted as a function of @xmath38 , indicating the validity of the scaling hypothesis . in a strongly metallic region ( @xmath39 ) , the data are well fitted to @xmath40 as shown in fig . [
fig : metalsps2 ] . in ref .
@xcite , the @xmath0 function for @xmath26 from the metallic to the insulating limits was estimated .
the estimate has not changed significantly when we add data for @xmath31 .
is @xmath41 . ]
is @xmath42 . ]
in last section and in ref .
@xcite , the scaling hypothesis has been checked numerically .
this supports the scaling theory for the 2d symplectic systems .
furthermore , the logarithmic increase of @xmath26 in the strongly metallic region gives the impression that the conductance may also increase logarithmically . in this section ,
we study the size dependence of the mean conductance directly .
we now attach two perfect leads to the square sample with linear size @xmath4 and calculate the two terminal landauer conductance @xmath1 ( in units of @xmath2 ) with the transfer matrix method @xcite .
a fixed boundary condition is imposed in the direction transverse to the current .
the hopping matrices in the transverse direction are set to the unit matrix , and the fermi energy @xmath30 .
we have accumulated 10000 samples for @xmath43 $ ] and 5000 samples for @xmath44 . as shown in fig .
[ fig:2dmeang ] , the mean conductance @xmath45 in the strongly metallic region does show the logarithmic increase @xmath46 in agreement with the prediction of the scaling theory @xcite .
the prefactor of the logarithmic term , which is an important parameter for the scaling in the strongly metallic region , is also consistent with it .
note that a perfect conductivity does not mean perfect transmission .
the transmission probability per channel rather goes to zero as @xmath4 increases , so most of incoming electrons are reflected .
it is left for future to understand physics behind this possible perfect conductivity in 2d . vs. system size @xmath4 in the 2d metallic region .
the solid lines are a guide to the eye only . for a reference , @xmath47 is indicated . ]
recently we have obtained numerical results which suggest that an anderson transition occurs even below 2d in the presence of spin - orbit coupling @xcite .
this is based on numerical simulations of electrons on a sierpinski carpet sc(@xmath48 ) , where @xmath49 is the generation number . in the limit @xmath50 , sc(@xmath48 )
becomes a true fractal whose spectral dimension is @xmath51 @xcite . here we study the size dependence of the landauer conductance through the sierpinski carpet in a delocalized phase .
we have attached two perfect leads to sc(@xmath48 ) and have calculated the conductance with the recursive green s function method @xcite .
a fixed boundary condition is imposed in the transverse direction .
the hopping matrices in the transverse direction are set to the unit matrix .
we set the fermi energy @xmath52 and the disorder @xmath53 where relatively large mean conductance is obtained .
the number of samples is 10000 for @xmath54 and 500 for @xmath55 .
figure [ fig : sc51meang ] shows the mean conductance @xmath56 as a function of the linear size @xmath57 .
the mean conductance increases with @xmath4 , possibly indicating a delocalized phase .
it also indicates that the increase is slower than that in 2d .
it is an open problem whether the conductance in the possible delocalized phase of the sierpinski carpet diverges . | electron transport phenomena in disordered electron systems with spin - orbit coupling in two dimensions and below are studied numerically .
the scaling hypothesis is checked by analyzing the scaling of the quasi-1d localization length .
a logarithmic increase of the mean conductance is also confirmed .
these support the theoretical prediction that the two dimensional metal in systems with spin - orbit coupling has a perfect conductivity .
transport through a sierpinski carpet is also reported . , , quantum transport , spin - orbit coupling , symplectic class , scaling theory , perfect conductivity |
until very recently , studies of @xmath1@xmath51 galaxies have focused primarily on luminous , vigorously star - forming objects such as submillimetre sources or @xmath6@xmath2@xmath0 lbgs that are forming stars at rates of 10s or 100s @xmath4/yr .
such studies have largely neglected the less glamorous but far more numerous faint , sub-@xmath0 galaxies that are forming stars at much lower rates .
although individually faint , these sub-@xmath0 objects are very numerous and so collectively they generate more than half the total uv luminosity density of the universe at high redshift .
they are , thus , extremely important contributors to the story of star formation and metal enrichment in the early universe .
moreover , the presence of a break at @xmath0 in the galaxy lf suggests that galaxies below @xmath0 differ substantially from those above it . because of this , our understanding of the mechanisms that drive galaxy evolution at high redshift may profit from comparing the properties of the better - studied luminous ( @xmath6 @xmath7 @xmath0 ) objects with those of the neglected sub-@xmath0 galaxies . this paper summarizes the results of our recent work aimed at understanding these hitherto neglected , individually modest but collectively very important objects .
they are the long faint tail of the high-@xmath1 galaxy population .
the shape of the galaxy lf bears the imprint of galaxy formation and evolution processes .
we study the global statistics of sub-@xmath8 galaxies at @xmath1@xmath22 , 3 , 4 , and 5 using two deep , large - area , multi - field imaging surveys . at @xmath1@xmath22 , 3 , and 4 we use the keck deep fields ( kdf ) of sawicki & thompson ( 2005 , 2006a , b ) .
the kdf use the _ very same _
@xmath9 filter set and color - color selection criteria as are used in the well - known work of @xcite .
however , in contrast to the steidel et al .
work , the kdf reach @xmath10=27 , which is 1.5 mag deeper than the steidel et al.surveys and significantly below @xmath8 ; even at @xmath1@xmath24 we reach 2 mag fainter than @xmath11 .
the kdf have an area of 169 arcmin@xmath12 divided into three spatially independent regions on the sky that allow us to monitor the effects of cosmic variance .
the @xmath13 filters and color - color selection commonality with the work of steidel et al .
lets us confidently select very faint , sub-@xmath8 galaxies at @xmath1@xmath22 , 3 , and 4 by relying on their extensive spectroscopic characterization of sample selection .
it also allows us to confidently combine our data with theirs , thereby for the first time consistently spanning such a large range in galaxy luminosity at high redshift .
our @xmath1@xmath25 work is based on very deep @xmath14 imaging of two subaru suprime - cam fields presented in @xcite .
these data are wider and deeper than earlier work we presented in @xcite : they now span a total of 1,300 arcmin@xmath12 divided over two fields with @xmath15 = 26.5 and 25.5 .
the large area covered by these data is essential for determining the abundance of the rare , luminous objects , while at the same time the depth of the data gives us the necessary ability to probe below @xmath0 in the lf .
we use @xmath14 lyman break selection to select @xmath1@xmath25 galaxies and our ongoing spectroscopic program has already confirmed a number of @xmath1@xmath25 galaxies , validating our @xmath1@xmath25 color - color selection technique @xcite .
we find that the luminosity function of high-@xmath1 galaxies undergoes evolution that is differential with luminosity . as fig .
1(a ) shows , the bright end of the lf remains virtually unchanged , but the faint end evolves from @xmath1@xmath25 to at least @xmath1@xmath23 .
our tests show that it is highly unlikely that this observed differential evolution is due to some systematic effect such as selection bias , our modeling of the survey volume , etc.(for details see * ? ? ?
the effect is also quite significant the statistical probability that the @xmath1@xmath24@xmath163 evolution is _ not _ differential with luminosity is only 1.5% , and the result is even more robust for @xmath1@xmath25@xmath163 .
we conclude that the _ differential , luminosity - dependent evolution _ is very likely real .
if so , it must reflect some real , intrinsic evolutionary differences between faint and luminous lbgs .
understanding what causes these differences will give us insights into what drives galaxy evolution at high redshift .
the importance of the lf s faint end extends into measurements of the cosmic luminosity density and the cosmic star formation rate that is often derived from it . for reasonable faint - end slopes of the lf
the bulk of the luminosity density resides in galaxies that are fainter than @xmath0 and it is only with deep , multi - field , large - area surveys such as ours that it is possible to reliably measure the contribution of this dominant population .
figure 1(b ) shows the evolution of the uv luminosity density using data from galex @xcite , keck sawicki & thompson ( 2006b ) , and subaru @xcite .
the _ total _
uv luminosity density , i.e. , luminosity density due to galaxies of _ all _ luminosities , rises from early epochs , @xmath1@xmath175 , experiences either a plateau or a broad peak at @xmath1@xmath231 , and then a decline to @xmath1=0 .
this behavior of the total luminosity density is dominated not by luminous galaxies , but by sub-@xmath8 objects .
indeed , it is galaxies within the rather narrow luminosity range @xmath6=(0.11)@xmath8 that dominate at high redshift , @xmath1@xmath225 , and it is the rapid evolution of this population that drives the evolution in the _ total _ uv luminosity density . thus far , this faint population has been largely neglected in high-@xmath1 follow - up studies which have focused on @xmath6 @xmath7 @xmath8 lbgs .
in contrast to @xmath1@xmath172 , at lower redshifts , @xmath1@xmath181 , the total uv luminosity density is dominated by _ very _ faint galaxies with @[email protected]@xmath8 .
the @xmath6=(0.11)@xmath8 galaxies that dominated at @xmath1@xmath172 are still important but no longer dominant .
this switch in the luminosity of the galaxies that dominate the total luminosity density is reminiscent of galaxy downsizing , although the effect we see so far is a downsizing in uv luminosity rather than galaxy mass .
as we have discussed above , sub-@xmath0 lbgs dominate over their more luminous cousins , but evolve differently from them as a population .
we now wish to know what makes them _ physically _ different .
we attempt to address this issue through broadband spectral energy distribution ( sed ) studies . unlike in the lf work
, here we do not need large samples of objects but instead must have deep multiwavelength photometry that spans from the uv to beyond the balmer break in the rest frame .
we need to go particularly deep because we wish to study galaxies significantly fainter than @xmath0 at @xmath1@xmath52 .
no spectroscopic samples of such very faint lbgs currently exist and so we must use photometric redshifts .
our first foray into this area is at @xmath1@xmath22 and uses the hst wfpc2+nicmos @xmath20 data of the northern hubble deep field ( hdf ) .
these data are ideal for us as they are very deep , include the @xmath21-band imaging essential for selecting @xmath1@xmath22 lbgs , and span the age - sensitive balmer break at the target epoch . after smoothing the multiband images to a common psf and obtaining sextractor @xcite photometry
, we constructed a @xmath21-band drop - out sample using the lbg color - color selection criteria defined by steidel et al , ( 1996 ) for the hdf .
we then applied a photometric redshift cut , 1.8@xmath18@[email protected] , where the photometric redshifts are determined as part of the sed - fitting procedure ( see below ) .
the lower limit of this photo-@xmath1 cut eliminates potential low-@xmath1 interlopers and the higher limit excludes high-@xmath1 lbgs for which the @xmath20 filter set does not span the crucial , age - sensitive balmer break .
this procedure ( two - color selection followed by a photo-@xmath1 cut ) is very similar to the procedure ( two - color selection followed by spectroscopy ) used by @xcite in their work .
our resulting sample of @xmath270 objects has a mean redshift @xmath23=2.3 and in other ways also closely mimics the bx samples of @xcite and @xcite .
however , it contains much fainter galaxies , reaching down to @xmath24=27 .
magnitude is defined via averaging the @xmath25 and @xmath26 fluxes . ]
our sed fitting procedure follows the now well - established approach first developed for lbgs by @xcite and subsequently used by many others ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we compare the observed @xmath27 is essential for _ selecting _ our lbg sample , we omit it from the sed fitting to avoid contaminating our sed analysis with the stochastic effects of intergalactic hydrogen absorption . ]
galaxy photometry with predictions of @xcite spectral synthesis models attenuated by @xcite dust .
we fix metallicity and star formation history and fit for burst age , star formation rate , stellar mass , dust reddening , and redshift .
while for brighter samples redshifts are usually constrained from spectroscopy , here we must keep redshift as a free parameter .
however , our tests as well as previous experience with this approach @xcite show that lack of spectroscopic redshifts does not significantly affect the results .
[ sedfit.fig ] figure 2a shows that the bulk of the galaxies in our hdf sample follows a relation between stellar mass and uv luminosity and this relation extends into the brighter galaxies of @xcite .
there appears to be good correlations between stellar mass and uv luminosity ( fig .
2a ) and stellar mass and sfr ( fig . 2b ) .
outliers in both samples turn out to be galaxies with constant - sfr best - fit ages apparently older than the age of the universe at @[email protected] ( triangles ) and galaxies with very young ages @xmath29@xmath1930gyr ( squares ) . however , most objects show a fairly tight correlation which can be described by the relation @xmath30 $ ] ( solid line in fig
the inclusion of the outliers in the fit does not significantly affect this result . the @xmath0 knee in the @[email protected] lf ( @xmath31=@xmath32 at @[email protected] : sawicki & thompson , 2006a ) corresponds to a stellar mass @xmath33@xmath210@xmath3@xmath4 ( fig 2a ) .
this @xmath1@xmath22 characteristic mass is remarkably similar to the 3@xmath3410@xmath3@xmath4 characteristic mass that at low redshift marks the transition between the two distinct galaxy `` bimodality '' families @xcite .
this similarity of characteristic masses suggests that the processes ( see , e.g. , * ? ? ?
* ) responsible for the observed galaxy bimodality at low @xmath1 may have also been operating at @xmath1@xmath22 . moreover , the near - unity slope of the logarithmic relation between stellar mass and uv luminosity ( fig .
2a ) suggests that the bulk of the detectable stellar mass in @[email protected] lbgs resides in the thus - far poorly studied galaxies with @xmath0@xmath5@[email protected]@xmath0 , just as does the bulk of the total uv luminosity density .
the outliers with extremely young fit ages ( square symbols ) are apparently either undermassive for their sfrs or overluminous for their masses . in the former case they may move up onto the mass - sfr relation as they age ; in the latter they could be experiencing strong star - forming bursts that dramatically elevate their normally low luminosities .
outliers with very old best - fit ages ( triangles ) , are also consistent with variable star formation histories : constant star formation histories tend to yield the largest ages ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and thus the old ages here may simply be telling us that star formation in these galaxies was more intense in the past .
all these scenarios advocate that at least some @xmath1@xmath22 lbgs , both faint and luminous , have variable star formation histories consistent with one of the scenarios proposed by @xcite to explain the evolving lbg luminosity function .
the thus - far largely unexplored sub-@xmath0 lbgs dominate the uv luminosity density of the universe at @xmath1@xmath52 .
moreover , their stellar masses appear to correlate with uv luminosities down to very faint magnitudes at least at @xmath1@xmath22 and so , just as they dominate the uv luminosity density , these objects may also contain a substantial fraction of the formed stellar mass by that redshift . at the same time
, however , the differentially - evolving luminosity function suggests that these sub-@xmath0 lbgs differ substantially from their more luminous cousins .
it is not yet clear what mechanism is behind this differential evolution , but the bursty nature of at least some of the @xmath1@xmath22 galaxies means that changes in the timescales of star - burst episodes may play a role .
we will address this and other issues as we extend our sed - fitting and other differential studies to higher redshifts . | we study the properties of very faint , sub-@xmath0 lyman break galaxies at @xmath1@xmath225 thus far a largely neglected but numerically and energetically very important population . we find that the lbg luminosity function undergoes luminosity - dependent evolution : the number of luminous galaxies remains constant while the number of faint ones grows with time .
the total uv luminosity density increases with cosmic time from at least @xmath1@xmath25 until reaching a peak or a plateau around @xmath1@xmath22 behaviour that is governed by the sub-@xmath0 galaxies in the lf s `` faint tail '' . using broadband
sed fitting we find a nearly - linear relationship between sfr and galaxy stellar mass at @xmath1@xmath22 .
a typical @xmath0 lbg at @xmath1@xmath22 shows a stellar mass of @xmath210@xmath3@xmath4 , remarkably similar to the bimodality mass at low redshift .
this similarity suggests that the mechanisms responsible for the galaxy bimodality at low-@xmath1 may have also been at play at @xmath1@xmath22 . |
extragalactic x - ray sources with luminosities in excess of 10@xmath2 erg s@xmath3 have now been widely detected in the nearby universe ( fabbiano 1989 ; miller & colbert 2004 ; roberts 2007 ) .
these are designated as ultraluminous x - ray sources ( ulxs ) if they are not coincident with the nucleus of their host galaxy ( i.e. an agn ) nor correspond to a foreground or background object , e.g. a qso .
they are too luminous to be powered by a sub - eddington accretion flow in a stellar - mass black hole binary ( bhb ) unless substantial beaming of their emission is invoked ( king et al .
2001 ) , so either require super - eddington mass accretion onto a stellar mass black hole , or sub - eddington accretion onto an intermediate mass black hole ( imbh ; see colbert & mushotzky 1999 ) .
unambiguously distinguishing between these possibilities requires radial velocity measurements of the underlying binary motion to directly determine the mass ( see e.g. charles & coe 2006 ) .
until these become available , we can only use more indirect arguments to identify the nature of these objects , primarily by comparing the x- ray spectral and timing properties of ulxs with those of stellar - mass bhbs ( see e.g. the review by roberts 2007 ) .
the x - ray spectra of ulxs can be fairly well described by a two component ( disc plus power - law ) model , as commonly used for bhbs .
taking the disc temperatures at face value implies masses of order 10@xmath410@xmath5m@xmath1 ( kaaret et al .
2003 ; miller et al . 2003
; miller et al . 2004a ; cropper et al . 2004 ; dewangan et al .
2004 ; roberts et al .
2005 ) , favouring sub - eddington accretion flows onto an imbh ( colbert & mushotzky 1999 ) .
however , the disc component generally does not dominate these spectra , and bhbs show that in these circumstances the derived disc radii vary dramatically , so can not be used to estimate mass ( kubota & done 2004 ) .
kajava & poutanen ( 2009 ) show that this cool disc component varies as @xmath6 , implying that the radius of the thermally emitting material increases with increasing luminosity , in sharp contrast to the @xmath7 expected from a disc with constant inner radius ( see also feng & kaaret 2009 ) .
the observed behaviour may instead be produced from a super - eddington source powering a strong wind , where the photosphere of the wind expands with increasing luminosity ( shakura & sunyaev 1973 ; king & pounds 2003 ; begelman , king & pringle 2006 ; poutanen et al .
2007 ) .
while the disc spectrum is ambiguous , the power - law is somewhat clearer in distinguishing between imbh and stellar mass systems ( or rather in distinguishing between sub and super - eddington accretion ) .
sources with the highest quality data show clear signs that the high energy ` power - law ' tail rolls over above 6 kev , so it is better fit by low temperature thermal comptonisation ( stobbart , roberts , & wilms 2006 ; miyawaki et al .
2009 ) . a rollover at such low temperatures is not seen in any of the sub - eddington bhbs
, so this suggests that ulxs are super - eddington and in a new , _ ultraluminous _ state ( roberts 2007 ; gladstone et al .
2009 , hereafter g09 ; vierdayanti et al .
similarly , low temperature thermal comptonisation is also required in the brightest narrow - line seyfert 1s ( nls1s , e.g. re j1034 + 396 ) and the brightest bhb in our own galaxy , grs 1915 + 105 ( zdziarski et al . 2001 ; middleton et al . 2009 ; ueda et al . 2009 ) .
super - eddington accretion then implies a much smaller mass black hole in ulxs , of the order 10 - 100 m@xmath1 .
another diagnostic of the nature of the accretion flow is its rapid variability as this is strongly correlated with spectral state .
sub - eddington bhbs show power density spectra ( pds ) which can be very roughly described by band limited noise , with a low frequency break at @xmath8 from a slope of @xmath9 to @xmath10 , followed by a high frequency break , @xmath11 , to @xmath12 .
this broad - band noise is often accompanied by strong low frequency quasi - periodic oscillations ( lfqpos ) which correlate with the low frequency break ( @xmath13 : wijnands & van der klis 1999 ) .
the lfqpo increases in frequency , coherence and power as the source flux increases and the spectrum softens from a faint low - hard state through the hard - intermediate state up to the soft - intermediate state ( belloni , mndez & homan 2005 ; van der klis 2004 ; remillard & mcclintock 2006 ) . during this transition ,
the broad - band power spectrum narrows in @xmath14 , with @xmath8 increasing while the power at high frequencies remains fairly constant ( czerny et al .
2008 ; done , gierliski & kubota 2007 ) .
l whilst the characteristic breaks are difficult to constrain in the low signal - to - noise ulx data ( see heil , vaughan & roberts 2009 ) , qpo features are much easier to measure .
three ulxs are known to demonstrate strong qpos in their pds , m82 x-1 ( strohmayer & mushotzky 2003 ) , ngc 5408 x-1 ( strohmayer et al .
2007 , hereafter s07 ; strohmayer & mushotzky 2009 , hereafter sm09 ) and , more recently , a second source in m82 ( feng , rao , & kaaret 2010 ) . identifying these with the type c lfqpo seen in sub - eddington bhbs
implies that these are all imbhs ( casella et al . 2008 ; s07 ; sm09 ) .
however , it also implies that the power spectra should show a low frequency break , and that the energy spectra should be consistent with a low - hard or very high state as these are the states in which the lfqpo is seen .
we reanalyse the data from ngc 5408 x-1 , looking at both spectra and variability , and show that neither the energy spectra nor power spectra support the identification of the qpo with the standard lfqpo seen in sub - eddington bhbs .
instead we show that the spectra , power spectra , and variability as a function of energy are all better matched by a super - eddington flow , implying a stellar mass compact object , in which case the qpo may be analogous to the milli - hz qpos seen in grs 1915 + 105 .
l [ cols="<,^,^,^,^,^,^,^,^,^,^ " , ] we can further resolve the variability by looking at long and short timescales .
we recalculate the rms using a binning of 50 s , to show only the power on long timescales ( red points ) .
this shows a very similar pattern of variability to that derived from the total lightcurve , with much more variability at hard energies than at soft .
the blue points show the energy dependence of the rapid variability , formed from the difference ( in quadrature ) between @xmath15 from the 10 s binning and that from the 50 s binning .
thus this effectively isolates variability between frequencies of 0.01 - 0.05 hz , but again the pattern of variability with energy is very similar .
both observations show a very similar pattern , but the rms variability at 1 - 2 kev is even larger for obs 2 than obs 1 .
the black points in fig 6 show the ratio of the total rms spectrum in obs 2 compared to obs 1 , while the red and blue points show the ratio of rms for the long and short timescale variability respectively . apart from on short timescales ,
these show significantly more variability above 1 kev in obs 2 , as expected from the power spectral analysis ( fig 4 ) .
thus , variability on all timescales and in both observations shows a strong increase as a function of energy .
this is consistent with a model in which there is a relatively stable component at soft energies diluting the variability as might be expected by a thermal disc ( see churazov et al .
2001 ) or optically thick photospheric component ( g09 , zdziarski , misra & gierliski 2010 ) .
this can be modelled by changing the normalisation of the hard energy spectral component by the percentage rms and taking the ratio of this to the best - fit model .
the result on 10s binning is shown as the solid black line in fig 5 , which closely resembles the shape of the rms .
we note that the shape of this rms is inconsistent with models in which the spectrum is dominated by a single component which changes only in normalisation .
this would give an rms which is constant with energy .
this rules out the reflection dominated model of caballero - garcia & fabian ( 2010 ) unless the variability is strongly fine - tuned .
the rms shape requires that the unseen illuminating power - law spectrum pivots around 0.3 kev ( see the spectral and rms analysis of a reflection dominated model for re j1034 + 396 in middleton et al .
2009 ) .
the spectral and rms analysis presented here both support a two - component model , where the soft component can be modelled by low temperature thermal emission that is mostly constant on timescales of up to a day , while the hard component , best modelled by a cool , optically thick corona , is steep and varies quite dramatically on timescales down to at least a hundred seconds . whilst this superficially resembles the properties of the steep power - law state ( mcclintock & remillard 2006 ) seen at fairly high luminosities ,
both the spectra and variability are inconsistent with such an interpretation .
the low temperature of the thermal component means that if this were from an accretion disc then the black hole mass is of order @xmath16 .
a similarly large mass is derived if the qpos are identified with the bhb type c lfqpo . however , the ulx luminosity of @xmath17 ergs s@xmath3 is then only @xmath18 .
bhbs have hard tails ( @xmath19 ) at such low mass accretion rates , rather than the soft tailseen in ngc 5408 x-1 .
similarly , the bhb power spectra at these low mass accretion rates show a low frequency break tightly correlated with the qpo frequency .
these breaks are not present at the predicted frequency in ngc 5408 x-1 .
soft , variable tails are only seen in bhbs in the steep power - law state , at @xmath20 ( remillard & mcclintock 2006 ) , giving a mass of @xmath21 if ngc 5408 x-1 is in an analogous state .
however , the large fractional variability of 0.3 - 0.4 is only seen above 10 kev in bhbs in this state , where the tail dominates over the constant disc component ( see the energy - dependent rms spectra in gierliski & zdziarski 2005 ) . scaling this energy for the difference in mass from @xmath21 to @xmath22
gives a prediction that the tail should start to dominate above 4.5 - 6 kev , at significantly higher energies than seen in ngc 5408 x-1 , where the variable tail dominates above 2 kev .
thus , ngc 5408 x-1 does not look convincingly similar to any of the sub - eddington bhb states .
instead , the large implied emission radius of the soft thermal component can also be incorporated into the super - eddington flow models by identifying this not with the disc itself but with much more extended emission from the photosphere of a wind ( shakura & sunyaev 1973 ; begelman et al .
2006 ; poutanen et al .
similarly , the extreme variability associated with the tail can be matched by occasional states of the pathologically variable bhb , grs 1915 + 105 , where it is probably associated with instabilities in a super - eddington flow ( belloni et al . 1997 ) .
fig 7 ( left hand panel ) shows the power spectrum of the tail in ngc 5408 x-1 ( 1 - 8 kev ) compared to that of _ rxte _ data from @xmath23 and @xmath24 variability states of grs 1915 + 105 ( centre panel , obsids 20405 - 01 - 33 - 00 : belloni et al .
these are calculated using standard 1 data ( 0.125 s resolution ) which has no energy information , but the source spectrum in the _ rxte _ bandpass means that the counts are concentrated in the 3 - 10 kev energy range . as well as showing similarly extreme amounts of broad band noise ,
these also show small milli - hertz qpo features at similarly low frequencies as seen in ngc 5408 x-1 ( 0.01 hz for 33 ) .
these ultra - low frequency qpos can dominate the power spectrum of grs 1915 + 105 in other states where there is much less broad band variability ( obsid : 10408 - 01 - 05 - 00 shown in magenta in the same panel ) , but this is unlike ngc 5408 x-1 , where the qpos are weak and the broadband variability is strong .
nonetheless , it is clear that stellar remnant black holes can produce ultra - lfqpos as well as the standard type c lfqpos .
however , while the variability properties of ngc 5408 x-1 and some states of grs 1915 + 105 are very similar , the energy spectra are subtly different .
the @xmath23 and @xmath24 state observations of grs 1915 + 105 have spectra which can be well modelled by a disc plus low temperature , optically thick comptonisation ( zdziarski et al .
2001 , 2005 ; done , wardziski & gierliski 2004 ; ueda et al .
2009 , 2010 ) , but _
unlike _ ngc 5408 x-1 , the comptonisation component dominates the bolometric luminosity , rather than the thermal component ( see fig .
2 , compared to middleton & done 2010 ) .
thus , even the most super - eddington bhb is not really a good match to the spectrum of ngc 5408 x-1 .
the only known systems which show a similar combination of steep spectral tail and extreme variability are a small subset of narrow line seyfert 1 galaxies ( leighly 1999 ; oneill et al .
these are likely to be super - eddington sources , but there are large uncertainties in deriving mass and mass accretion rate for these sources ( grupe & mathur 2004 ; marconi et al .
fig 7 ( right hand panel ) shows the power spectrum from _ xmm - newton _ data ( obsid 0110890101 ) of the most extremely variable nls1 , iras 13224 - 3809 ( boller et al .
1997 ; gallo et al . 2004
; ponti et al .
since any soft thermal component in the agn should be at much lower energies , we use the full bandpass from 0.3 - 10 kev .
clearly the peak in @xmath25 reaches similar values , though the frequencies are very different . since ngc
5408 x-1 does not match with any of the sub - eddington bhb states , nor to any of the multiple states of the mildly super - eddington bhb grs 1915 + 105 , it seems most likely that it represents a strongly super - eddington source .
such flows power a strong wind , where the wind photosphere gives the large radius , luminous , soft thermal component at low energies ( shakura & sunyaev 1973 ; begelman et al . 2006 ; poutanen et al .
this extended emission will be constant on short timescales .
we also see a much more variable tail , so the optically thick wind can not cover all lines of sight to the inner disc . a reasonable geometry for this is an equatorial disc wind ( abolmasov , karpov & kotani 2009 ; ohsuga et al . 2007 , 2009 ) . the low temperature , optically thick compton emission could be produced by a corona over the inner disc , a more extreme version of the marginally optically thick corona seen in the steep power - law state ( e.g. done & kubota 2006 ; g09 )
. however , it could also be the spectrum of the inner disc itself as at such high mass accretion rate , the high disc temperature means that there is very little true continuum opacity ( beloborodov 1998 , done & davis 2008 ) .
this means that the energy does not thermalise , and can not even be well approximated by a colour temperature corrected blackbody ( see e.g. watarai et al .
we favour this latter origin as the probably super - eddington agn re j1034 + 396 shows clear evidence for a _ separate _ , variable corona , as well as a more constant low temperature comptonised component ( the energy shift from the much larger black hole mass means that any wind photosphere emission is at too low an energy to be observed ) .
this predicts there should also be an additional weak tail in the ulx , which is so far unobserved .
while the small size scale of the inner disc means that it could vary intrinsically , we note that other ulxs with similar spectra do not show similarly extreme variability ( heil et al .
2009 , spectra from g09 ) . instead
, a solution which can explain the range of variability behaviour in these wind - dominated ulxs could be if the extreme variability is extrinsic , from clumps in the wind .
numerical simulations show that this wind is turbulent ( ohsuga 2007 ; ohsuga et al .
2009 ) , giving strong variability of the comptonisation component from the inner disc on sightlines which intercept the major wind streamlines , but the soft thermal emission will remain mostly constant due to its much larger size scale . we show a simple cartoon of this model in fig 8 ( see also leighly 2004 ) . at extreme super - eddington mass accretion rates , a large fraction of the accretion energy
is ejected in the wind , making this energetically dominant ( poutanen et al . 2007 ) .
the photosphere of the wind ( orange ) is the source of soft thermal emission , while the inner disc ( green ) is the source of the low temperature , optically thick comptonised emission .
we propose that this has only low level intrinsic variability in strongly super - eddington sources , perhaps because the disc structure is stabilised by advection ( abramowicz et al .
thus these sources have very low variability when viewed along a line of sight which does not intercept the wind , while there is strong variability from obscuration of the comptonised emission when viewed through the turbulent wind .
the intrinsically variable , but rather weak corona ( blue ) gives rise to a weak high energy tail , so far not observed in ulxs .
scaling this model up to nls1s could perhaps similarly explain the extreme variability from sources such as iras 13324 - 3809 and 1h0707 - 495 from absorption in strong winds ( gallo 2004 , see also miller et al .
2010 for how this can also match the spectrum ) , the existence of which are supported by uv emission lines ( leighly 2004 ) . the much lower variability seen from the low temperature comptonised component in re j1034 + 396
is then associated with a more face - on viewing angle .
where timing characteristics scale with mass , they can be used to determine the unknown mass of the black hole in ulx .
however , the wide variety of properties seen in bhb means that this only works if the correct spectral state is identified for scaling .
the key issue in ulxs is whether they are analogous to any of the classic , sub - eddington , bhb states .
if so , the large luminosities of these systems implies an imbh . instead , if ulxs are super - eddington , then there are no well defined templates from bhb except for the unique source grs 1915 + 105 .
the suggested identification of the qpo in ngc 5408 x-1 is with the type c lfqpo seen in bhbs .
however , bhbs show a tight correlation between this qpo frequency and that of the low frequency break in the broad band power spectrum ( wijnands & van der klis 1999 ) .
we use extensive numerical simulations to constrain the position of this break in the two observations of ngc 5408 x-1 , and find that they are inconsistent with the predicted frequencies from a type c lfqpo .
the energy spectra of both observations of ngc 5408 x-1 are best described by a soft thermal component ( disc or photosphere ) together with a low temperature , optically thick , thermal compton component , as proposed for other high quality ulx spectra ( g09 ) .
the rms spectra show that the variability increases strongly with energy , consistent with a constant soft component which dilutes a separate , strongly variable component at higher energies .
the amount of variability in the high energy component is extreme , and we speculate that this is extrinsic , from stochastic obscuration by the turbulent wind .
the only objects with comparable variability and spectra are grs 1915 + 105 and some extreme nls1s , both of which are most probably super - eddington accretors .
this also suggests an alternative identification for the qpo in ngc 5408 x-1 as being more closely related to the ultra low frequency qpo seen occasionally in grs 1915 + 105 . both spectra and timing are then consistent with identifying ngc 5408 x-1 as a super - eddington source .
this places a natural upper limit on the central bh mass of @xmath26 100m@xmath1 , consistent with a bh formed from stellar collapse in a low metallicity environment ( e.g. belczynski et al .
we thank the referee , juri poutanen , for useful comments and insights .
mm thanks stfc for their support in the form of a post - doctoral position funded by a standard grant .
this work is based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and nasa . | the ultraluminous x - ray source ( ulx ) , ngc 5408 x-1 , is one of only 3 such objects to show a quasi - periodic oscillation ( qpo ) in its power spectrum .
previous analysis of this signal identified it with the well - studied type c low - frequency qpo ( lfqpo ) seen in black hole binaries ( bhbs ) , implying an intermediate mass black hole ( imbh ) .
however , in bhbs this qpo has a centroid frequency which scales tightly with the position of the low - frequency break in the broad band power spectrum .
we use this relation to predict the frequency of the power spectral break in ngc 5408 x-1 , and show that this is inconsistent with the break frequencies in both available , archival _ xmm - newton _ observations .
thus the broad band power spectral shape does not support this identification of the qpo .
the energy spectra also do not support an imbh interpretation .
they can be fit by a two - component model , best described by soft thermal emission at low energies , together with low - temperature , optically thick comptonisation producing a tail which dominates above 2 kev .
the parameters of the tail are unlike those seen in any of the sub - eddington bhb spectral states .
the energy dependent variability supports this deconvolution , as it is consistent with the soft thermal component below 2 kev diluting extreme variability of the high energy tail .
the only objects with similar spectra which have similar amounts of variability are the bhb , grs 1915 + 105 , and some extreme nls1s .
this suggests that ngc 5408 x-1 is in a similar super - eddington state , placing a natural limit on the mass of @xmath0 100m@xmath1 .
its qpo could then be similar to the ultra - lfqpo seen occasionally in grs 1915 + 105 , consistent with a large stellar mass black hole .
we suggest a model geometry which may explain the spectra and variability of highly super - eddington sources .
= -0.5 cm [ firstpage ] accretion : accretion discs black hole physics
x - rays : binaries x - rays : individual : ngc 5408 x-1 |
most biological processes within the cell are carried out by proteins that physically _ interact _ to form stoichiometrically stable _
complexes_. even in the relatively simple model organism _ saccharomyces cerevisiae _ ( budding yeast ) , these complexes are comprised of many subunits that work in a coherent fashion .
these complexes interact with individual proteins or other complexes to form functional modules and pathways that drive the cellular machinery .
therefore , a faithful reconstruction of the entire set of complexes ( the ` complexosome ' ) from the physical interactions among proteins ( the ` interactome ' ) is essential to not only understand complex formations , but also the higher level cellular organization .
protein complexes constitute _ modular _ functional units within the network of physical interactions , the ppi network@xcite . from a biological perspective , this modularity is a result division of labor and of evolution to provide robustness against mutation and chemical attacks@xcite . from a topological perspective , this modularity is a result of proteins within complexes being densely connected to each other than to the rest of the network@xcite . since the advent of high - throughput " screening in molecular biology in the late 1990s and early 2000s
, several techniques have been introduced to infer physical interactions among proteins within organisms in a large - scale ( genome - wide " ) manner .
these have helped to catalogue significant amount of protein interactions in organisms such as yeast thereby fueling computational techniques to systematically mine and analyze such large - scale interaction data . in yeast ,
the yeast two - hybrid ( y2h)@xcite , protein complementation assay ( pca)@xcite and tandem affinity purification followed by mass spectrometry ( tap - ms)@xcite are some of the widely adopted experimental systems that have helped to identify a considerable fraction of physical interactions among proteins .
however , even at the current ` state - of - the - art ' , these high - throughput techniques have been shown to produce considerable proportion of spurious ( false positive ) interactions@xcite .
therefore , once the interactions are identified their qualities need to be first assessed to generate a reliable set of interactions that is deemed suitable for further mining and analysis .
this process includes assigning each interaction a confidence score that typically accounts for the biological variability and technical limitations of the experimental conditions , and therefore reflects the reliability of the inferred interaction @xcite .
the interactions with confidence scores below a certain threshold are discarded to build a reliable cleaned - up " ppi network .
this ppi network is then mined to identify groups of proteins potentially forming complexes .
the whole process can be summarized in the following steps : 1 .
integrating high - throughput datasets from multiple experiments and assessing the reliabilities of interactions ; 2 . constructing a reliable ppi network
; 3 . identifying modular subnetworks from the ppi network to generate a candidate list of complexes ; 4 . evaluating the identified complexes against _ bona fide _ complexes , and validating and assigning roles to novel complexes .
the identification of complexes from high - throughput interaction datasets has attracted considerable attention from both biologists as well as computational research communities , and over the years , several computational techniques have been developed to systematically identify complexes .
quite naturally , a number of surveys have come out from time - to - time evaluating and comparing these techniques for their performance on available ppi datasets .
one of the earliest comprehensive evaluations was by brohee and van helden ( 2006)@xcite .
this was followed by vlashblom et al .
( 2009)@xcite and li et al . ( 2010)@xcite . while brohee and vlasblom et al . evaluated and compared some early methods on ppi datasets available at that time ( till 2006 ) , li et al . covered some of the more recent methods developed until 2009 .
the purpose of our work is to provide an up - to - date survey , classification ( taxonomy ) and evaluation of some the representative works done till date ( 2011/2012 ) .
we build upon the existing surveys so as to not repeat entirely the evaluations and conclusions already drawn , yet we provide our own classifications and evaluations of more recent techniques across up - to - date ppi datasets .
we also compare across unscored ( raw ) and scored ppi datasets , which is missing in these existing surveys .
we also highlight and comment on some of the newer challenges and open problems in complex prediction , which can guide future directions for research in this area .
we begin by mentioning some definitions and terminologies widely adopted across the reviewed works .
a ppi network is modeled as an undirected graph @xmath0 , where @xmath1 is the set of proteins and @xmath2 is the set of interactions among protein pairs . for any protein @xmath3
, @xmath4 is the set of direct neighbors of @xmath5 , while @xmath6 is the degree of @xmath5 .
the interaction density of @xmath7 is defined as @xmath8 .
this is a real number between 0 and 1 , and typically quantifies the richness of interactions " within @xmath7 : 0 for a network without any interactions and 1 for a fully connected network .
the clustering coefficient @xmath9 measures the cliquishness " of the neighborhood of @xmath5 : @xmath10 , where @xmath11 is the set of edges in the neighborhood of @xmath5 .
if the interactions of the network are reliability scored ( weighted ) , these definitions can be extended to their corresponding weighted versions : @xmath12 , @xmath13 , and @xmath14 , where @xmath15 is a scoring function on the interactions in @xmath16 .
there are several interesting variants proposed for weighted clustering coefficient @xmath17 ; for a survey see@xcite .
although at a very generic level most existing methods make the key assumption that complexes are embedded among densely - interacting groups of proteins within ppi networks , these methods vary considerably either in the algorithmic methodologies or the kind of biological insights employed to detect complexes .
accordingly , we classified some popular complex detection methods into two broad categories ( a soft classification ) : ( i ) methods based solely on graph clustering ; ( ii ) methods based on graph clustering and some additional biological insights .
these biological insights may be in the form of functional , structural , organizational or evolutionary information known about complexes or their constituent proteins from experimental or other biological studies .
we present this classification in two snapshots .
the first snapshot , shown in figure [ figure : taxonomy_chronology ] , gives a _ chronology - based _ bin - and - stack " classification , while the second snapshot , shown in figure [ figure : taxonomy_complex_detection_methods ] gives a _ methodology - based _
tree " classification of the methods . in the chronology - based classification , we _ binned _ methods based on the years in which they were developed , and stacked them based on the kind of biological insights used ( see figure [ figure : taxonomy_chronology ] ) .
the biological insights are grouped as : core - attachment structure , evolutionary information , functional coherence , and mutually exclusive and co - operative interactions .
it is interesting to note from this classification that , over the years , as researchers tried to improve the basic graph clustering ideas , they also incorporated a variety of biological information into their methods . in the methodology - based classification
, we distributed the methods to different branches of a _ classification tree _
based on the kind of computational strategy used ( see figure [ figure : taxonomy_complex_detection_methods ] ) . at the first level from the root , we grouped these methods into those based solely on graph clustering , and those employing additional biological insights . at subsequent levels , we further divided these methods based on the kind of algorithmic strategies used , into : ( i ) methods employing merging or growing of clusters ; ( ii ) methods employing repeated partitioning of networks ; and ( iii ) methods employing network alignment .
the methods employing merging or growing clusters go bottom - up " , that is , typically start with small seeds " ( for example , triangles or cliques ) , and repeatedly add or remove proteins or merge clusters based on some similarity measures to arrive at the final set of complexes . on the other hand , the methods based on network partitioning go top - down " , that is , repeatedly partition or break the network into multiple subnetworks based on certain divisive criteria .
the methods based on network alignment use multiple networks ( typically from different species ) to arrive at isomorphic regions that likely correspond to complexes , the inituition being that proteins belonging to real complexes should generally be conserved through the evolution process to act as an integrated functional unit@xcite .
most methods that cluster the ppi network into multiple dense subnetworks make use of solely the topology of the network .
mcode , proposed by bader and hogue ( 2003)@xcite , is one of the first computational methods ( and therefore , seminal ) developed for complex detection from ppi networks .
the mcode algorithm operates in mainly in two stages , vertex weighting and complex prediction , and an optional third stage for post - processing . in the first stage , each vertex @xmath5 in the network @xmath0 is weighted based on its neighborbood density . instead of directly using clustering coefficient
, mcode uses core - clustering coefficient which measures the density of the highest @xmath18-core in the neighborhood of @xmath5 .
this amplifies the weighting of densely connected regions in @xmath7 . in the second stage ,
the vertex @xmath5 with the highest weight is used to seed a complex .
mcode then recursively moves outwards from the seed vertex , including vertices into the complex whose weight is a given percentage ( vertex weight parameter - vwp ) away from the seed vertex .
a vertex once added to a complex is not checked subsequently .
the process stops when there are no more vertices to be added to the complex , and is repeated using the next unseeded vertex . at the end of this process multiple non - overlapping complexes are generated .
the optional third stage performs a post - processing on the complexes generated from the second stage .
complexes without 2-cores are filtered out , and new vertices in the neighborhood with weights higher than a given ` fluff ' parameter are added to existing complexes .
the resultant complexes are scored and ranked based on their densities .
the time complexity of the algorithm is @xmath19 , where @xmath20 is the vertex size of the average vertex neighbourhood in the network @xmath7 .
the markov clustering ( mcl ) algorithm , proposed by stijn van dongen ( 2000)@xcite , is a general graph clustering algorithm that simulates random walks ( called _ flow _ ) to extract out relatively dense regions within networks . in biological applications ,
it was first applied to cluster protein families and ortholog groups@xcite before it proved to be effective in detecting complexes from protein interaction networks@xcite .
mcl manipulates the adjacency matrix of networks with two operators called _ expansion _ and _ inflation _ to control the random walks ( flow ) .
expansion models the spreading out of the flow , while inflation models the contraction of the flow , making it thicker in dense regions and thinner in sparse regions .
these parameters boost the probabilities of intra - cluster walks and demote those of inter - cluster walks .
mathematically , expansion coincides with normal matrix multiplication , while inflation is a hadamard power followed by a diagonal scaling .
therefore , mcl is highly efficient and scalable .
the iterative expansion and inflation separates the network into multiple non - overlapping regions .
cmc was proposed by liu et al .
( 2009)@xcite to detect complexes from ppi networks based on repeated merging of maximal cliques
. some earlier algorithms like cfinder@xcite and local clique merging algorithm ( lcma)@xcite also adopted clique merging to find dense neighborhoods , but the distinct advantage of cmc over these algorithms is its ability to work on weighted networks and to find relatively low density regions ( in subsequent improved versions of cmc ) .
cmc begins by enumerating all maximal cliques in the ppi network using the cliques algorithm proposed by tomita et al.@xcite .
although enumerating all maximal cliques is np - hard , this does not pose a problem in ppi networks because these networks are usually sparse .
cmc then assigns a score to each clique @xmath21 based on its weighted density , which considers the reliabilities ( weights ) of the interactions within the clique : @xmath22 cmc then ranks these cliques in decreasing order of their scores and iteratively merges or removes highly overlapping cliques based on their inter - connectivity scores .
the inter - connectivity score of two cliques @xmath23 and @xmath24 is based on the non - overlapping regions of the two cliques and is defined as : @xmath25 cmc determines whether two cliques @xmath23 and @xmath24 sufficiently overlap : @xmath26 .
if so , @xmath24 is either removed or merged with @xmath23 based on the inter_score : if the @xmath27 , then @xmath23 and @xmath24 are merged , else @xmath24 is removed . finally , all the resultant merged clusters are output as the predicted complexes .
nepusz et al .
( 2012)@xcite proposed clusterone , a method for detecting overlapping protein complexes from weighted ppi networks , based on seeding and greedy growth , similar to mcode@xcite .
clusterone uses a cohesiveness measure to determine how likely a group of proteins form a complex , and is based on the weight of the interactions within the group and with the rest of the network . to begin with
, clusterone identifies seed proteins and greedily grows them into groups with high cohesiveness . when the greedy growth for a group can not progress any more , a next seed protein is selected to repeat the procedure until no more seed proteins remain . in the second step , clusterone identifies highly overlapping cohesive groups and merges them into potential complex candidates .
clusterone allows identification of overlapping complexes if each of the merged groups represent individual complexes that share proteins .
nepusz et al.s comparisons with methods like mcode , mcl and cmc showed that the complexes from clusterone achieved comparable accuracies when matched against known ` gold standard ' complexes and mcl achieved the closest performance to clusterone with the exception that mcl produced only non - overlapping clusters - a distinct advantange of clusterone .
apart from these discussed methods , three other methods worth mentioning here are lcma ( 2005)@xcite , pcp ( 2007)@xcite and haco ( 2009)@xcite . the lcma algorithm first locates cliques within local neighborhoods using vertex degrees and then merges them based on overlaps to produce complexes . protein complex prediction ( pcp ) uses fs weight scoring to remove unreliable interactions and add indirect interactions , and then merges cliques to produce the final list of complexes . on the other hand , haco uses hierarchical agglomerative clustering to produce the intial set of ( non - overlapping ) clusters .
proteins are then assigned to multiple clusters based on their interactions to the clusters to produce the final list of overlapping clusters .
a few other recently proposed ( 2010 - 2011 ) methods include those by zhang et al.@xcite , ma et al.@xcite , wang et al.@xcite and chin et al.@xcite .
these use the property of bridgeness " of cross - edges among clusters along with the internal connectivities to detect complexes .
gavin and colleagues ( 2006)@xcite performed large - scale analysis of yeast complexes and found that the proteins with complexes were divided into two distinct groups , cores " and attachments " .
the cores formed central functional units of complexes , while the attachment proteins aided these cores in performing their functions .
several computational methods were proposed to reconstruct complexes from ppi newtorks by capitalizing on this structural organization .
wu min et al .
( 2009)@xcite proposed the coach method which reconstructs complexes in two stages - it identifies dense core regions , and subsequently includes proteins as attachments to these cores .
figure [ figure : coach_xli2010 ] summarizes how coach identifies core and attachment proteins to build complexes .
leung et al .
( 2009)@xcite proposed the core method to identify protein cores within the ppi network .
they defined the probability of two proteins @xmath28 and @xmath29 ( of degrees @xmath30 and @xmath31 , respectively ) to belong to the same core using two main factors : whether the two proteins interact or not and the number of common neighbors @xmath32 between them .
the probability that @xmath28 and @xmath29 have @xmath33 interactions and @xmath34 common neighbors is calculated under the null hypothesis that @xmath30 edges connecting @xmath28 and @xmath31 edges connecting @xmath29 are randomly assigned in the ppi network according to a uniform distribution .
this probability is used to arrive at a @xmath35-value for whether @xmath28 and @xmath29 belong to the same core .
subsequently , core merges sets of core proteins of sizes two , three , etc . until further increase in size is not possible , to produce the final set of cores .
core then scores and ranks the predicted cores based on the number of internal and external interactions in them . the attachments are added to these cores in a manner similar to coach to produce the final set of complexes .
srihari et al .
proposed mcl - ca ( 2009)@xcite and the improved ( weighted ) version mcl - caw ( 2010)@xcite which identify complexes by refining clusters produced by the mcl algorithm@xcite by incorporating core - attachment structure .
essentially , mcl - caw categorizes proteins within mcl clusters into core " and attachment " proteins based on their connectivities , and then selects only these categorized proteins into complexes while discarding the remaining noisy " proteins .
this enables mcl - caw to trim " the raw mcl clusters .
further , unlike core and coach , the refinement in mcl - caw capitalizes on reliability scores assigned to the interactions .
consequently , mcl - caw reconstructs significantly higher number of ` gold standard ' complexes ( @xmath3630% higher ) and with better accuracies compared to plain mcl .
chin et al .
( 2010)@xcite proposed the hunter algorithm which begins by generating a module seed @xmath37 for each node @xmath5 in the ppi network .
@xmath37 is then pruned by removing vertices having low weight edges to other members of @xmath37 .
then the maximal @xmath38-connected subnetwork of @xmath37 is selected as the initial core @xmath39 .
this core is then expanded into a module by adding new vertices that share many neighbors with @xmath39 .
if two modules overlap beyond a certain threshold , these modules are merged .
the resultant collection of modules form the final set of predicted complexes .
proteins within complexes are generally enriched with same or similar functions@xcite .
if the functional information for proteins from an organism are available , then this information can be combined with topological information from ppi networks for the reconstruction of complexes from the organism .
one possible way to incorporate functional information is to score the interactions based on the functional similarity between the interacting pairs of proteins .
alternately , functional annotations ( for example , from gene ontology@xcite ) can be used to aid decisions where including or excluding a protein into complexes purely based on topological information might be difficult . king et al .
( 2004)@xcite proposed the rnsc algorithm that combines topological and gene ontology information to detect complexes .
the algorithm operates in two steps - it begins by clustering the ppi network and then filters the clusters based on cluster properties and functional homogeneity .
the network @xmath0 is first randomly partitioned into multiple subnetworks , which is essentially a partitioning of the node set @xmath1 .
the algorithm then iteratively moves nodes from one cluster to another in a randomized fashion till an integer - valued cost function is optimized .
a common problem among such clustering algorithms is the tendency to settle in poor local minima . to avoid this
, the rnsc algorithm adopts diversification moves , which shuffle the clustering by occasionally dispersing the contents of a cluster at random .
once the clustering process is completed , clusters of small sizes or densities ( the lower bound on cluster sizes and densities are experimentally determined ) are discarded .
finally , a @xmath35-value is calculated using functional annotations ( from go ) for each cluster that measures the functional homogeneity of the clusters .
all clusters above a certain @xmath35-value are discarded to produce the final list of predicted complexes . based on experiments , king et al .
recommend cluster density cut - off of 0.70 and @xmath35-value cut - off of @xmath40 .
( 2007)@xcite proposed the decaff algorithm which essentially is an extention of the lcma algorithm@xcite proposed earlier by the same group .
decaff identifies dense subgraphs in a neighborhood graph using a hub - removal algorithm .
local cliques are identified in these dense subgraphs and merged based on overlaps to produce clusters . each cluster
is assigned a functional reliability score , which is the average of the reliabilities of the edges within the cluster .
all clusters with low reliabilities are discarded to produce the final set of predicted complexes .
the pcp algorithm@xcite described earlier can also be categorized into this set of methods because pcp uses a weighting scheme based on functional similarity ( though the similarity is inferred from topology ) to assign reliability scores to interactions , and then uses a clique merging strategy to detect complexes .
the increasing availability of ppi data from multiple species like yeast , fly , worm and some mammals has made it feasible to use insights from cross - species analysis for detection of ( conserved ) complexes .
the assumption is that proteins belonging to real complexes should generally be conserved through the evolution process to act as an integrated functional unit@xcite .
sharan et al . proposed methods ( 2005 - 2007)@xcite for detection of conserved complexes across species based on the evolution of ppi networks . in these methods ,
an orthology network ( network alignment graph ) is constructed from the ppi networks of different species , which essentially represents the orthologous proteins and their conserved interactions across the species . for a protein pair @xmath41 in network @xmath42 of species @xmath43 and @xmath44 in @xmath45 of species @xmath46 , the orthology network @xmath47 contains the pair @xmath48 if @xmath49 is orthologous to @xmath50 , and @xmath51 is orthologous to @xmath52 .
the edge @xmath53 is weighted by the sequence similarities between the pairs @xmath41 , and @xmath54 .
any subgraph in this orthology network @xmath47 is therefore a conserved subnetwork of @xmath42 and @xmath45 .
such candidate subgraphs are then evaluated for parts of conserved complexes .
based on this idea , a tool _ _
qnet__@xcite was developed which returns conserved complexes from different species when queried using known complexes from yeast . the overlapping binding interfaces in a protein may prevent multiple interactions involving these interfaces from occurring simultaneously@xcite . in other words , the set of interactions in which a protein participates may be either co - operative or mutually exclusive
. the information about the co - occurrence or exclusiveness of interactions can therefore be useful for predicting complexes with higher accuracy .
this information can be gathered from the interacting domains of protein pairs or the three - dimensional structures of the interacting surfaces .
ozawa et al . (
2010)@xcite proposed a refinement method over mcode and mcl to filter predicted complexes based on exclusive and co - operative interactions .
they used domain - domain interactions to identify conflicting pairs of protein interactions in order to include or exclude proteins within candidate complexes . based on their results , the accuracies of predicted complexes from mcode and mcl
improved by two - fold . on the other hand , jung et al .
( 2010)@xcite used structural interface data to construct a simultaneous ppi network ( spin ) containing only co - operative interactions and excluding competition from mutually exclusive interactions .
mcode and lcma algorithms tested on this spin displayed a sizeable improvement in correctly predicted complexes .
even though incorporating information about co - operative and exclusive interactions shows promising improvement in complex detection algorithms , there are still several practical problems related to this approach . gathering more data on conflicting interactions , especially based on three - dimensional structures of interfaces , needs to be addressed before this approach can be more easily adopted . in a recent foresightful survey by przytycka et al.@xcite , the application of network _ dynamics _ ( temporal information ) into current computational analysis is discussed at good lengths , especially with respect to detection of complexes and pathways from protein interaction networks .
the authors suggest that if sufficient information about the ` timing activities ' of proteins can be obtained , the dynamical nature of the underlying organizational principles in interaction networks can be better understood .
this shift from static to dynamic network analysis is vital to understanding several cellular processes , some of which may have been wrongly understood due to ignoring dynamic information .
considering the wide variety of proposed methods for complex detection , one can gauge the seriousness in the research effort towards computational identification and categorization of complexes . several surveys and experiments@xcite have focused on the comparative analyses of these proposed methods for complex detection .
each new work on complex detection also comes with detailed comparative analyses of the new method with some earlier methods .
however , due to the differences in ppi and benchmark datasets , evaluation criteria , thresholds and parameters used , and the subset of methods considered for these comparative assessments , different works arrive at different results on the relative performance of methods .
but , typically the following broadly accepted criteria are used across the works .
if a reasonably large ` gold standard ' set of complexes is available ( as in the case of yeast ) , the performance of a method can be gauged on how accurately its predicted complexes reconstruct or recover the ` gold standard ' .
two commonly adopted measures for this are precision and recall@xcite .
precision measures how many among the predicted complexes match some ` gold standard ' complex , in turn measuring the proportion of realible predictions ( accuracy ) from the method .
recall measures how many of the ` gold standard ' complexes are reconstructed by the method , in turn measuring the coverage or sensitivity of the method .
some methods tend to produce too many ( sometimes arbitrary ) predictions resulting in high recall but very low precision , and therefore too many false positives to consider the method even reasonably reliable . to handle this , a combination of precision and recall , usually through a harmonic mean called f - measure , is used to evaluate how balanced "
is the method . on the other hand , if a ` gold standard ' set is not available ( as in the case mammals , currently ) , self - evaluatory " measures like cluster cohesiveness and separability is used@xcite .
the cohesiveness of a predicted complex ( cluster ) usually measures topological characteristics of the cluster , for example , its interaction density or size , while separability measures how separated is the cluster from others@xcite .
a combination of cohesiveness and separability reveals how modular is the clustering and therefore how likely the individual clusters represent real complexes .
another typically independent way to evaluate the predictions is to measure the functional or co - localization coherence of the clusters subjected , however , to availability of appropriate annotation data@xcite .
this captures how functionally coherent are the proteins within a predicted complex and whether they are co - localized within the cell .
the usual annotations required for these calculations are functions and sub - cellular localizations of the proteins .
this evaluation is particularly useful for alternative validation of the predictions .
now , we present a summary of some representative surveys and comparative assessments and their conclusions .
one of the first comprehensive assessments was performed by brohee and van helden ( 2006)@xcite .
they performed a detailed empirical comparison between mcode@xcite , mcl@xcite , rnsc@xcite and super - paramagnetic clustering ( spc)@xcite .
these algorithms were tested on ppi datasets from high - throughput experiments , and the resultant complexes were evaluated against benchmark complexes from mips@xcite .
additionally , the ppi datasets were introduced with artificial noise ( random edge addition and deletion ) to test the robustness of these algorithms .
they concluded that mcl and rnsc outperformed mcode and spc in terms of precision ( the proportion of correctly predicted complexes ) and recall ( the proportion of correctly derived benchmarks ) .
rnsc was robust to variation in its input parameter settings , while the performance of the other three varied widely for parameter changes .
mcl was remarkably robust even upon introducing 80%-100% random noise .
overall , the experiments confirmed the general superiority of mcl over the other three algorithms .
vlasblom et al .
( 2009)@xcite compared mcl with another clustering algorithm , affinity propagation ( ap)@xcite on unweighted as well as weighted ppi networks .
the initial unweighted network was built from a set of 408 hand - curated complexes from wodak lab@xcite followed by random addition and removal of edges to mimic real ppi networks .
the weighted network was obtained from the collins et al.s work@xcite , generated from gavin and krogan datasets@xcite .
they concluded that mcl performed considerably better than ap in terms of accuracy and separation of predicted clusters , and robustness to random noise .
in particular , mcl was able to achieve about 90% accuracy and 80% separation compared to only 70% accuracy and 50% separation of ap on unweighted ppi networks with introduced random noise .
mcl was able to discover benchmark complexes even at high ( 40% ) noise levels . more recently ( 2010 ) , li et al.@xcite performed a detailed comparative evaluation of several algorithms : mcode@xcite , mcl@xcite , core@xcite , coach@xcite , rnsc@xcite and decaff@xcite .
these algorithms were tested on ppi datasets from dip@xcite and krogan et al.@xcite .
the dip network consisted of 17203 interactions among 4930 proteins , while the krogan dataset consisted of 14077 interactions among 3581 proteins .
they used a total of 428 benchmark complexes from mips@xcite , aloy et al.@xcite and sgd@xcite .
a cluster @xmath55 from a method was considered a correct match to a benchmark complex @xmath56 using the bader score@xcite @xmath57 , where @xmath58 denotes the number of proteins in @xmath55 , and @xmath59 denotes the number of proteins in @xmath56 . based on this criteria , the precision , recall and f - measure values were calculated .
the comparisons of these algorithms is shown in figure [ figure : comparison_methods_dip_krogan_xli2010 ] ( adapted from@xcite ) .
the methods are arranged in chronological order , and it is interesting to note that over the years , the f - measures have improved .
li et al . concluded that mcl , rnsc , core , coach and decaff attained the best recall values .
mcode was able to achieve the highest precision , but it produced very few clusters resulting in very low recall . in our assessment , we experimentally evaluated some key complex detection methods on both unscored ( raw ) and scored ppi networks .
to build our unscored network , we combined the physical interaction from two tap - ms experiments , gavin et al .
( 2006)@xcite and krogan et al .
( 2006)@xcite , which we call the gavin+krogan ( g+k ) network .
we then gathered the scored version of this network , the consolidated network from collins et al .
( 2007)@xcite .
this network comprises of interactions from gavin et al . and krogan et al .
scored using the purification enrichment scheme@xcite .
some of the properties of these networks are shown in table 1 . [ cols="^,^,^,^ " , ] let @xmath60 and @xmath61 be the sets of benchmark and predicted complexes , respectively .
we use the jaccard coefficient @xmath62 to quantify the overlap between a benchmark complex @xmath63 and a predicted complex @xmath24 : @xmath64 we consider @xmath63 to be covered by @xmath24 , if @xmath65 _ overlap threshold _ @xmath66 . in our experiments , we set the threshold @xmath67 , which requires @xmath68 .
for example , if @xmath69 , the overlap between @xmath63 and @xmath24 should be at least 6 .
we use previously reported@xcite definitions of _ recall _ ( coverage ) and _ precision _ ( sensitivity ) of the set of predicted complexes : @xmath70 here , @xmath71 gives the number of _ derived benchmarks_. @xmath72 here , @xmath73 gives the number of _ matched predictions_. we calculated the _ f - measure _ as the harmonic mean of precision and recall , @xmath74 we considered the following methods for our evaluation : * on the unscored network , mcl ( 2002 , 2004)@xcite , mcl - ca ( 2009)@xcite , mcl - caw ( 2010)@xcite , core ( 2009)@xcite , coach ( 2009)@xcite , cmc ( 2009)@xcite and haco ( 2009)@xcite ; * on the scored network , mcl ( 2002 , 2004)@xcite , mclo ( 2007)@xcite , mcl - caw ( 2010)@xcite , cmc ( 2009)@xcite and haco ( 2009)@xcite .
we do not show comparisons for older methods like mcode ( 2003)@xcite and rnsc ( 2004)@xcite because these have been evaluated extensively in several earlier surveys@xcite , instead we included mcl as a benchmark in all our comparisons since mcl has been repeatedly shown to perform better than these older methods@xcite .
further , not all methods are devised to make use of interaction confidence scores , and therefore we selected only the ones capable of doing so for the evaluations on the scored network .
table 3 shows the precision and recall values for methods evaluated on the unscored gavin+krogan network across the three benchmark sets .
the table shows that core , haco and mcl - caw performed significantly better in terms of recall compared to the rest of the methods . in particular ,
mcl - caw performed considerably better than plain mcl indicating that incorporating core - attachment structure into raw mcl clusters helped to improve the accuracies of the predicted complexes .
this indicated that incorporating some kind of biological knowledge helped to identify complexes more accurately .
next , table 4 shows these values for the methods evaluated on the scored consolidated network .
this table shows that all methods were able to reconstruct significantly higher number of complexes upon scoring as compared to on the unscored network .
this clearly indicated that noise in raw datasets ( negatively ) impacted the performance of methods , and reliability scoring aided in alleviating this impact and thereby improving the performance of methods .
this demonstrated the effectiveness of current reliability scoring schemes in cleaning raw interaction datasets for focused studies such as complex detection .
c || l || c | c | c | c | c | c | c + + & & + & & mcl & mcl - ca & mcl - caw & coach & core & cmc & haco + & # predicted & 242 & 219 & 130 & 447 & 386 & 113 & 278 + & # matched & 55 & 49 & 69 & 62 & 83 & 60 & 78 + * wodak & precision & 0.226 & 0.224 & 0.531 & 0.139 & 0.215 & 0.531 & 0.281 + * ( # 182 ) & # derived & 62 & 49 & 75 & 49 & 83 & 60 & 85 + & recall & 0.338 & 0.269 & 0.412 & 0.269 & 0.456 & 0.330 & 0.467 + * * & # matched & 35 & 42 & 42 & 45 & 59 & 41 & 45 + * mips & precision & 0.143 & 0.192 & 0.323 & 0.101 & 0.153 & 0.363 & 0.162 + * ( # 177 ) & # derived & 40 & 42 & 53 & 38 & 59 & 41 & 57 + & recall & 0.226 & 0.237 & 0.300 & 0.215 & 0.333 & 0.232 & 0.322 + * * & # matched & 43 & 41 & 47 & 54 & 59 & 43 & 59 + * aloy & precision & 0.179 & 0.187 & 0.362 & 0.121 & 0.153 & 0.381 & 0.212 + * ( # 76 ) & # derived & 42 & 41 & 52 & 37 & 59 & 43 & 59 + & recall & 0.556 & 0.539 & 0.684 & 0.487 & 0.776 & 0.566 & 0.776 + * *
c || l || c | c | c | c | c + + & & + & & mcl & mclo & mcl - caw & cmc & haco + & # predicted & 116 & 119 & 130 & 77 & 101 + & # matched & 70 & 80 & 83 & 67 & 57 + * wodak & precision & 0.603 & 0.672 & 0.638 & 0.870 & 0.564 + * ( # 145 ) & # derived & 79 & 80 & 90 & 67 & 64 + & recall & 0.545 & 0.552 & 0.621 & 0.462 & 0.441 + * * & # matched & 48 & 65 & 53 & 56 & 40 + * mips & precision & 0.414 & 0.546 & 0.408 & 0.727 & 0.396 + * ( # 157 ) & # derived & 63 & 65 & 67 & 56 & 57 + & recall & 0.401 & 0.414 & 0.427 & 0.357 & 0.363 + * * & # matched & 54 & 56 & 57 & 45 & 44 + * aloy & precision & 0.466 & 0.471 & 0.438 & 0.584 & 0.436 + * ( # 76 ) & # derived & 55 & 56 & 55 & 45 & 45 + & recall & 0.724 & 0.737 & 0.724 & 0.592 & 0.592 + * * we next plugged - in " these evaluation results as well as results obtained from some earlier surveys@xcite into our bin - and - stack " classification , as shown in figure [ figure : taxonomy_chronology_f1 ] . for each method
, we show the f - values _ before / after _ scoring , that is , on the unscored g+k network and the scored consolidated network .
such a representation in our classification revealed two interesting insights , 1 . incorporating biological information in addition to ppi topology improved performance of the methods ( the f measures have increased in the vertical layers compared to the lowest layer ) ; 2 .
reliability scoring significantly improved performance of the methods , as shown by the before - after values .
this representation also shows how complex detection methods have evolved over the years to improve performance , and therefore our taxonomy can be useful to guide future directions for further improvement .
the review and evaluation of computational methods in the above sections reveal several critical challenges facing accurate identification of complexes from high - throughput interaction datasets .
we saw that most methods are considerably impacted by noise in raw datasets .
further , most methods are able to reconstruct only a fraction of the known complexes ( achieve at the most @xmath75 recall ) even upon scoring .
this points towards some inherent limitations within the methods itself . on this basis
, we broadly classify the challenges facing current methods into two categories , ( i ) challenges originating from biological datasets ; ( ii ) challenges originating from existing computational techniques .
even though over the last few years , several independent high - throughput experiments have helped to catalogue enormous amount of interactions from yeast , they show surprising lack of correlation with each other , and lack of coverage - bias towards high abundance proteins and against proteins from certain cellular compartments ( like cell wall and plasma membrane)@xcite . also , each dataset still contains a substantial number of false positives ( noise ) that can compromise the utility of these datasets for more focused studies like complex detection , as seen from our evaluation results . in order to reduce the impact of such discrepancies , a number of data integration and reliability scoring schemes have been devised@xcite . to overcome these challenges to some extent
, in our evaluation , we combined multiple datasets ( from gavin et al.@xcite and krogan et al.@xcite ) to account for the lack of interaction coverage , and also adopted scoring prior to predicting complexes . in spite of these precautionary steps , we notice that most methods are able to reconstruct only a fraction of the known complexes , and we still have a long way to go towards identification of meaningful novel complexes through computational means . as noted earlier ,
even though there have been numerous methods developed for complex detection , most of them suffer from low recall ( at most 65% recall even on the scored network ; table 4 ) .
even a union " of these methods achieves at most @xmath76-@xmath77 recall on average across a variety of ppi datasets@xcite .
one of the crucial reasons for this limitation is that every method , in one way or another , relies on the key assumption that complexes are embedded among dense " regions of the network .
however , recent experiments have indicated that relying too much on this assumption in the wake of insufficient credible interaction data causes these methods to miss many complexes that are of low densities , that is , sparse " in the network@xcite .
therefore the need is to find alternative ways to model complexes than mere dense subnetworks , and also to compensate for the sparsity of topological information by augmenting other kinds of biological information . in the attempt to detect sparse complexes , the recent work by srihari et al .
( 2012)@xcite is insightful and can guide futher directions towards tackling this challenge .
srihari et al .
( 2012)@xcite noticed that most existing computational methods based on ppi networks rely overly on the assumption that complexes are embedded among
dense " regions of the network , and therefore miss most of the sparse " complexes that have very low interaction densities or lie disconnected in the network .
these complexes are missed by current methods due to the lack of crucial topological information ( missing interactions and/or proteins ) - for example , even in the well - studied organism yeast , only @xmath78 of the interactome has been validated and catalogued@xcite . in order to overcome these topological gaps " , the authors proposed careful augmenting of functional interactions to ppi networks .
functional interactions represent logical or conceptual associations among proteins , and therefore encode " a variety of biological similarities or affinities among proteins beyond just physical interactivity , thereby compensating for the lack of topological information in ppi networks . in order to do this augmentation
systematically they proposed the sparc algorithm .
sparc selects low quality clusters predicted from the physical ( ppi ) network using existing methods , and then selectively enhances these clusters to reconstruct ( sparse ) complexes .
the key idea is that many of these low quality clusters are in fact fractions ( or pieces " ) of sparse complexes embedded in the ppi network , but due to missing interactions they lie scattered " in the network and do not represent whole complexes in their current forms .
if these clusters can be carefully enhanced by augmenting functional interactions , then several of the sparse complexes can be reconstructed accurately .
this enhancement involves increasing their interaction densities and pulling - in " together their disconnected components .
however , during the selection of the initial set of low quality clusters , many may just represent noisy predictions ( false positives ) . in order to determine the clusters that are more likely to represent complexes
, sparc makes use of a novel component - edge ( ce ) score .
the ce - score is a topological measure combining connectivity and relative density of the clusters , and is shown to more accurately correlate with the topological characteristics of real complexes compared to traditionally accepted measures like edge density .
the ce - score is calculated for every low quality cluster predicted from the ppi network .
sparc then augments functional interactions to the clusters and checks if their ce - scores improve beyond a certain threshold . if a cluster shows the required improvement , it is output as a potential candidate representing a sparse complex .
srihari et al . showed through extensive experiments on clusters produced from methods like mcl@xcite , mcl - caw@xcite , cmc@xcite and haco@xcite that sparc was capable of improving the overall recall of these methods by upto @xmath79 on average across a variety of networks .
specifically , sparc helped to reconstruct @xmath80 more complexes among the ones that were sparse .
small complexes ( complexes of two or three proteins ) also pose severe challenges in identification , particularly if ppi network topology is the only available information .
in fact most complex detection methods based solely on the ppi network only attempt to identify complexes with at least 4 or more proteins in the network@xcite .
the attempt to predict small complexes ( pairs or triplets ) from the network based only on connectivity typically produces a significant fraction of false positives .
further , the smaller the size of a complex , the more prone it is to sparsity - missing even a few interactions can result in very low densities or render the complexes disconnected .
due to these challenges , additional information apart from ppi network topology is required for detection of such complexes .
in a recent work ( 2012 ) , srihari et al.@xcite incorporated temporal information to identify small complexes .
the authors focused on identifying small complexes that are assembled during the yeast cell cycle in a temporal manner .
they noticed that several high - degree proteins such as kinases interacted with different subsets of proteins during different phases of the yeast cell cycle to assemble multiple phase - specific complexes .
however , due to the lack of temporal information to disambiguate ( segregate ) the different phase - specific complexes , existing topology - based methods produced large clusters of complexes fused together .
srihari et al .
decomposed such large clusters by incorporating temporal information on the yeast cell cycle in which each protein showed peak expression , and thereby segregated the individual constituent complexes .
many of the segregated complexes were small and represented complexes of kinases and their temporal substrates .
srihari et al . also noticed that by incorporating such temporal information the dynamics " of protein complexes could be better understood .
they observed an interesting relationship between the staticness " ( constitutive expression ) of a protein and its participation in multiple complexes@xcite - cells tend to maintain generic proteins as ` static ' to enable their reusability " across multiple temporal complexes .
membrane protein complexes are formed by physical interactions among membrane proteins .
membrane proteins are attached to or associated with the membranes of the cell or its organelles .
membrane proteins constitute approximately @xmath81 of the proteomes of organisms , yet they are one of the least studied subsets of proteins .
the study of membrane proteins and their complexes is crucial in understanding diseases and aiding new drug discoveries@xcite .
membrane protein complexes are notoriously difficult to study using traditional high - throughput techniques like y2h and tap - ms@xcite .
this is due in part to the hydrophobic ( insoluble ) nature of membrane proteins , as well as the ready dissociation of subunit interactions , either between transmembrane subunits or between transmembrane and cytoplasmic subunits@xcite . in order to counter the disadvantages of conventional techniques ,
new biochemical techniques have been developed recently to facilitate the characterization of interactions among membrane proteins . among these
is the split - ubiquitin membrane yeast two - hybrid ( myth ) system@xcite . with the development of the high - throughput myth system ,
a fair number of interactions among membrane proteins have been recently catalogued in species such as _ _ arabidopsis thaliana__@xcite and yeast _ _ saccharomyces cerevisiae__@xcite .
the identification of membrane complexes requires understanding their assembly - how the individual proteins come together to form complexes , and how these complexes are eventually degraded .
this is because membrane proteins are not stable entities as their soluble counterparts .
studies reveal that this assembly occurs in an orderly fashion , that is , membrane complexes are formed by an ordered assembly of intermediaries , and in order to prevent unwanted intermediaries , this assembly is highly aided by chaperones@xcite .
many membrane complexes are formed by transient interactions involving exchange of proteins in and out of existing complexes via membranes - a phenomenon called ` dynamic exchange'@xcite .
the need therefore now is to develop sophisticated algorithms that take into account these aspects specific to membrane complexes to mine them effectively from membrane sub - interactomes .
protein complexes are the fundamental functional units responsible for many biological mechanisms within the cell .
their identification is therefore necessary to understand the cellular organization and machinery .
the advent of high - throughput techniques for inferring protein interactions in a large - scale fashion has fueled development of computational techniques to systematically mine for potential complexes from the network of interactions . in this work
, we reviewed , classified and evaluated some of the key computational methods developed till date for the detection of protein complexes from ppi networks .
we presented two insightful taxonomies of existing methods - ` bin - and - stack ' and ` tree ' . from these taxonomies
we note that scoring of raw interaction datasets ( followed by filtering of false positives ) and integrating key biological insights with topology can significantly improve complex prediction .
even though more than 20 methods have been developed over the years , complex detection still requires careful attention in handling errors and noise in experimental datasets , and reconstructing complexes with high accuracies . on this front , we identified some of the crucial limitations and challenges facing current experimental and computational techniques .
interaction datasets coming from different experimental sources show surprising lack of correlation and also contain sizeable fraction of spurious ( false positive ) interactions .
this severely impacts the accuracy and coverage of complex detection methods .
further , computational methods also overly rely on the assumption that complexes are embedded among densely connected groups of proteins , an assumption that is not fully valid in the wake of insufficient credible interactions . finally , the interactions among membrane proteins have not been catalogued adequately making it difficult to identify an important group of complexes necessary for understanding diseases - membrane complexes .
we hope that our review and assessment of computational methods as well as the challenges highlighted here will provide valuable insights to drive future research for further advancing the ` state - of - the - art ' in computational prediction , characterization and analysis of protein complexes from organisms .
this work was supported in part by the national university of singapore under arf grants r-252 - 000 - 361 - 112 and r-252 - 000 - 461 - 112 .
ss is now associated with the institute of molecular bioscience , university of queensland , and would like to thank the institute for providing the resources to complete this work .
uetz , p. , giot , l. , cagney , g. , traci , a. , judson r. , knight , s.r . ,
lokshon , d. , narayan , v. , srinivasan , m. , pochart , p. , qureshi - emili , a. , li , y. , godwin , b. , conover , d. , kalbfleish , t. , vijayadamodar , g. , yang , m. , johnston , m. , fields , s. , rothberg , j.m . , a comprehensive analysis of protein - protein interactions in saccharomyces cerevisiae , _ nature _ * 403(6770)*:623627 , 2000 .
gavin , a.c . ,
bosche , m. , krause , r. , grandi , p. , marzioch , m. , bauer , a. , schultz , j. , rick , j.m . ,
michon , a - m . ,
cruciat , c - m . ,
remor , c. , hofert , c. , schelder , m . ,
brajenovic , m . ,
ruffner , m . ,
merino , a . , klein , k . , hudak , m. , dickson , d . , rudi , t . ,
gnau , v . ,
bauch , a .
, bastuck , s . ,
huhse , b . ,
leutwein , c . , heurtier , m - a . , copley , r - r . ,
edelmann , a . ,
querfurth , e . ,
rybin , v . ,
drewes , g . ,
raida , m. , bouwmeester , t. , bork , p . ,
seraphin , b . , kuster , b . , neubauer , g .
, superti - furga , g . ,
functional organization of the yeast proteome by systematic analysis of protein complexes , _ nature _ * 415(6868)*:141147 , 2002 .
ho , y. , gruhler , a. , heilbut , a. , bader , g. , moore , l. , adams , s.l . , millar , a. , taylor , p. , bennett , k. , boutilier , k. , yang , l. , wolting , c. , donaldson , i. , schandorff , s. , shewnarane , j. , vo , m. , taggart , j. , goudreault , m. , muskat , b. , alfarano , c. , dewar , d. , lin , z. , michalickova , k. , willems , a.r .
, sassi , h. , nielsen , p.a . ,
rasmussen , k.j . ,
andersen , j.r . ,
johansen , l.e . ,
hansen , l.h .
, jespersen , h. , podtelejnikov , a. , nielsen , e. , crawford , j. , poulsen , v. , sorensen , b.d . , matthiesen , j. , hendrickson , r.c . , gleeson , f. , pawson , t. , moran , m.f . ,
durocher , d. , mann , m. , hogue , c.w . , figeys , d. , tyers , m. , systematic identification of protein complexes in saccharomyces cerevisiae by mass spectrometry , _ nature _ * 415(6868)*:180183 , 2002 .
gavin , a.c .
, aloy , p. , grandi , p. , krause , r. , boesche , m. , marzioch , m. , rau , c. , jensen , l.j . ,
bastuck , s. , dumpelfeld , b. , edelmann , a. , heurtier , m.a . , hoffman , v. , hoefert , c. , klein , k. , hudak , m. , michon , a.m. , schelder , m. , schirle , m. , remor , m. , rudi , t. , hooper , s. , bauer , a. , bouwmeester , t. , casari , g. , drewes , g. , neubauer , g. , rick , j.m . ,
kuster , b. , bork , p. , russell , r.b .
, superti - furga , g. , proteome survey reveals modularity of the yeast cell machinery , _ nature _ * 440(7084)*:631636 , 2006 .
krogan , n.j . ,
cagney , g. , yu , h. , zhong , g. , guo , x. , ignatchenko , a. , li , j. , pu , s. , datta , n. , tikuisis , a.p . , punna , t. , peregrin - alvarez , j.m . , shales , m. , zhang , x. , davey , m. , robinson , m.d .
, paccanaro , a. , bray , j.e . ,
sheung , a. , beattie , b. , richards , d.p . ,
canadien , v. , lalev , a. , mena , f. , wong , p. , starostine , a. , canete , m.m . ,
vlasblom , j. , wu , s. , orsi , c. , collins , s.r . ,
chandran , s. , haw , r. , rilstone , j.j . ,
gandi , k. , thompson , n.j . , musso , g. , onge , p. , ghanny , s. , lam , m.h . ,
butland , g. , altaf - ul , a.m. , kanaya , s. , shilatifard , a. , oshea , e. , weissman , j.s .
, ingles , c.j . ,
hughes , t.r . , parkinson , j. , gerstein , m. , wodak , s.j . ,
emili , a. , greenblatt , j.f . , global landscape of protein complexes in the yeast saccharomyces cerevisiae , _ nature _ * 440(7084)*:637643 , 2006 .
von mering , c. , krause , r. , snel , b. , cornell , m. , oliver , s.g . , fields , s. , bork , p. , comparative assessment of large - scale datasets of protein - protein interactions , _ nature _ * 417(6887)*:399403 , 2002 .
cusick , m.e .
, yu , h. , smolyar , a. , venkatesan , k. , carvunis , a - r . , simonis , n. , rual , j - f . , borick , h. , braun , p. , dreze , m. , vandenhaute , j. , galli , m. , yazaki , j. , hill , d.e . ,
ecker , j.r . ,
roth , f.p . , vidal , m. , literature - curated protein interaction datasets , _ nature methods _ * 6(1)*:3946 , 2009 .
collins , s.r .
, kemmeren p. , zhao , x.c .
, greenbalt , j.f . , spencer f. , holstege , f. , weissman , j.s . , krogan , n.j .
, toward a comprehensive atlas of the physical interactome of saccharomyces cerevisiae , _ molecular cellular proteomics _ * 6(3)*:439450 , 2007 .
liu , g. , li , j. , wong , l. , assessing and predicting protein interactions using both local and global network topological metrics , _ genome informatics series : proceedings of the 19th international conference on genome informatics _ * 21*:138149 , 2008 .
kalna , g. , higham , d. , a clustering coefficient for weighted networks , with application to gene expression data , _ j. ai communications - network analysis in natural sciences and engineering _ * 20(4)*:263271 , 2007 .
li x.l . , tan , s.h . ,
foo , c.s .
, ng , s.k . , interaction graph mining for protein complexes using local clique merging , _ genome informatics series : proceedings of the 16th international conference on genome informatics _ * 16(2)*:260269 , 2005 .
wang , h. , kakaradov b. , collins s.r .
, karotki , l. , fiedler , d. , shales m. , shokat , k.m . , walter , t. , krogan n.j . , koller , d. , a complex - based reconstruction of the saccharomyces cerevisiae interactome , _ molecular cellular proteomics _ * 8*:13611377 , 2009 .
wang , y. , gao , l. , chen , z. , an edge based core - attachment method to detect protein complexes in ppi networks , _ proceedings of the ieee international conference on systems biology ( isb ) _
pp.7277 , 2011 .
chin , c.h . ,
chen , s.h .
, ho , c.w . , ko , m.t . , lin , c.y
. , a hub - attachment based method to detect functional modules from confidence - scored protein interactions and expression profiles , _ bmc bioinformatics _ * 11(suppl 1)*:s25 , 2010 .
srihari , s. , ning , k. , leong , h.w . , refining markov clustering for complex detection by incorporating core - attachment structure , _ genome informatics series : proceedings of the 20th international conference on genome informatics _ * 23(1)*:159168 , 2009 .
srihari , s. , ning , k. , leong , h.w .
, mcl - caw : a refinement of mcl for detecting yeast complexes from weighted ppi networks by incorporating core - attachment structure , _ bmc bioinformatics _ * 11*:504 , 2010 .
ashburner , m. , ball c.a . , blake j.a .
, botstein , d. , butler , h. , cherry , m. , davis ap , dolinski k , dwight ss , eppig jt , harris ma , hill dp , issel - tarver l , kasarskis a , lewis s , matese jc , richardson je , ringwald m , rubin gm , sherlock g. , gene ontology : a tool for the unification of biology , _ nature genetics _ * 25(1)*:2529 , 2000 .
li , x.l . ,
, ng , s.k . , discovering protein complexes in dense reliable neighborhoods of protein interaction networks , _ computational systems bioinformatics series : proceedings of the computational systems bioinformatics conference _ * 6*:157168 , 2007 .
sharan , r. , ideker , t. , kelley , b.p . , shamir , r. , karp , r.m .
, identification of protein complexes by comparative analysis of yeast and bacterial protein interaction data , _ j. computational biology _ * 12(6)*:835846 , 2005 .
ozawa , y. , saito , r. , fujimori , s. , kashima , h. , ishizaka , m. , yanagawa , h. , miyamoto - sato , e. , tomita , m. , protein complex prediction via verifying and reconstructing the topology of domain - domain interactions , _ bmc bioinformatics _
* 11*:350 , 2010 .
xenarious , i. , salwinski , l. , duan , j.x . ,
higney , p. , kim , s - l .
, eisenberg , d. , dip , the database of interacting proteins a research tool for studying cellular networks of protein interactions , _ nucleic acids research _ * 30(1)*:303305 , 2002 .
aloy , p. , bottcher , b. , ceulemans , h. , leutwein , c. mellwig , c. , fischer , s. , gavin , a.c . ,
bork , p. , superti - furga , g. , serrano , l. , russell , r. , structure - based assembly of protein complexes of yeast , _ science _ * 303(5666)*:20262029 , 2004 .
cherry , j.m . ,
adler , c. , chervitz s.a . ,
. , jia , y. , juvik , g. , roe , t. , schroeder , m. , weng , s. , borstein , d. , sgd : saccharomyces genome database , _ nucleic acids research _
* 26(1)*:7379 , 1998 .
srihari , s. , leong , h.w . , employing functional interactions for the characterization and detection of sparse complexes from yeast ppi networks , to appear in _
international j. bioinformatics research and applications _ , 2012 .
lalonde , s. , ehrhardt , d.w . ,
logue , d. , chen , j. , rhee , s.y . ,
frommer , w.b . , molecular and cellular approaches for the detection of protein - protein interactions latest techniques and current limitations , _ plant j. _ * 53(4)*:610 - 635 , 2008 .
miller , j.p . , russel.s .
ben - hur , a. , desmarais , c. , stagljar , i. , novel , w.s . ,
fields , s. , large - scale identification of yeast integral membrane protein interactions , _ proc .
sci . _ * 102(34)*:12123 - 12128 , 2005 .
kittanakom , s. , chuk , m. , wong , v. , snyder , j. , edmonds , d. , lydakis , a. , zhang , z. , auerbach , d. , stagljar , i. , analysis of membrane protein complexes using the split - ubiquitin membrane yeast two - hybrid ( myth ) system , _ methods mol .
* 548*:247271 , 2009 . | complexes of physically interacting proteins are one of the fundamental functional units responsible for driving key biological mechanisms within the cell .
their identification is therefore necessary not only to understand complex formation but also the higher level organization of the cell . with the advent of high - throughput " techniques in molecular biology , significant amount of physical interaction data has been catalogued from organisms such as yeast , which has in turn fueled computational approaches to systematically mine complexes from the network of physical interactions among proteins ( ppi network ) . in this survey , we review , classify and evaluate some of the key computational methods developed till date for the identification of protein complexes from ppi networks .
we present two insightful taxonomies that reflect how these methods have evolved over the years towards improving automated complex prediction .
we also discuss some open challenges facing accurate reconstruction of complexes , the crucial ones being presence of high proportion of errors and noise in current high - throughput datasets and some key aspects overlooked by current complex detection methods .
we hope this review will not only help to condense the history of computational complex detection for easy reference , but also provide valuable insights to drive further research in this area . |
recently , ultracold atomic gases with strong p - wave interaction have attracted broad interest both experimentally expk40 , expk40b , expk40mola , expk40dipo , expk40molb , expli6a , expli6b , expli6c , expli6molsca , expli6r87 , explik , expli6mol , expkrb and theoretically pyou,2dpsfa , psfi,2dpsfb , psf1certa , psf1certb , psferta , psfa , p2dsfb , psff,2dpsfc , psfh , p2dsfa , psfc , 1dpsfb , psfe , pssfa , psfertc , gpra , 1dpsfa , psfb , psfertb , psf1certc , pfrvortex , psfola , pbosonert , psfbosefermi , twofermionerta , maqidajpsj1 , maqidajpsj2,zinnerepjd , maqidaprl1,pra2010a , maqidapra1,maqidapra2,maqidapra3,xxxa , stoofprl , p3bert , ppseudopotentialerta , p2bfr , ppseudopotentialertb , ppseudopotentialc , ppseudopotentiald , ppseudopotentiale , prlblume , nf1,nf2,nf3,nf4 .
the p - wave magnetic feshbach resonances , which can generate tunable p - wave ineratomic interactions , have been observed in the cold gases of @xmath11k @xcite , @xmath12li expli6a , expli6b , @xmath12li-@xmath13rb mixture expli6r87 , @xmath12li-@xmath14k mixture @xcite , and @xmath15k-@xmath13rb mixture @xcite .
the p - wave feshbach molecules have also been created and studied in the gases of @xmath16k @xcite and @xmath12li expli6a , expli6c , expli6mol , expli6molsca .
these experimental achievements stimulate theoretical researches on the quantum superfluid in ultracold atomic gases with strong p - wave interactions pyou,2dpsfa , psfi,2dpsfb , psf1certa , psf1certb , psferta , psfa , p2dsfb , psff,2dpsfc , psfh , p2dsfa , psfc , 1dpsfb , psfe , pssfa , psfertc , gpra , 1dpsfa , psfb , psfertb , psf1certc , pfrvortex , psfola , pbosonert , psfbosefermi , maqidajpsj1,maqidajpsj2 , zinnerepjd , maqidaprl1,pra2010a , maqidapra1,maqidapra2,maqidapra3,xxxa , stoofprl , as well as the relevant few - body problems @xcite .
until now , most theories of ultracold atomic gases with strong p - wave interactions psf1certa , psf1certb , psferta , psfertc , psfertb , psf1certc , pbosonert , stoofprl , twofermionerta , p3bert , ppseudopotentialerta , p2bfr , ppseudopotentialertb are based on the low energy expansion of the p - wave scattering amplitude @xmath17 given by the effective - range theory @xcite @xmath18here @xmath19 is the relative momentum of the two atoms ; @xmath20 is the scattering volume and @xmath4 is the effective - range .
the effective - range theory for the scattering amplitude is used in both theories of p - wave atomic superfluids psf1certa , psf1certb , psferta , psfertc , psfertb , psf1certc , pbosonert , stoofprl and related few - body problems @xcite . in particular , the separable two - body potential @xcite @xmath21 used in the many - body hamiltonian for the p - wave atomic superfluid is derived directly from low energy expansion ( 1a ) of the scattering amplitude . however , the effective - range theory is correct only for the _ short range _ potentials ( _ e.g. _ , yukawa potential ) @xcite which decays faster than any power function @xmath22 in the large interatomic distance limit @xmath23 . here
@xmath24 is the relative coordinate between two atoms . for a realistic interaction between two cold atoms , which is described by a _
long - range _ potential dominated by the van der waals term @xmath25 in the limit @xmath26 . here
@xmath27 is the single - atom mass , @xmath28 is the plank constant and @xmath29 is the van der waals length . due to the long - range van der waals potential ,
the effective - range theory and the low - energy expansion ( [ 1a ] ) of the p - wave scattering amplitude is not applicable any longer @xcite . in the presence of a p - wave magnetic feshbach resonance ( pmfr ) in the ultracold gases of polarized fermionic atoms , the p - wave scattering amplitude of the atomic collision
is contributed by both the background potential in the open channel and the bound state in the closed channel .
the long - range nature of the background potential makes the final scattering amplitude to be inconsistent with the effective - range theory .
therefore , it is essential to investigate the condition under which the effective - range theory ( [ 1a ] ) can be used as an approximation of the exact p - wave scattering amplitude under a pmfr .
if the effective - range theory provides a good approximation of the scattering amplitude , then the previous theories on p - wave superfluid would be applicable ; nevertheless , if the exact scattering amplitude is found to be significantly different with the one in eq .
( [ 1a ] ) , then the previous theories should be modified . especially , the separable two - body potential can not be used any more .
the low - energy p - wave scattering amplitude near pmfr has been investigated in refs .
@xcite . however , these studies are based on simplified models of the atomic interaction , e.g. , zero background potential p2bfr or a separable background potential that decays exponentially in the momentum space @xcite .
the long - range van der waals potential is not taken into account in either case .
due to these simplifications , the scattering amplitudes given in refs .
@xcite automatically have the form of eq .
( [ 1a ] ) , and can not be used to judge the applicability of the effective - range theory . in this paper ,
based on the realistic long - range inter - atomic potential , we provide an explicit calculation for the low - energy p - wave scattering amplitude of two spin polarized fermonic atoms near a pmfr , and then discuss the condition under which the effective - range theory can be used as a good approximation .
we get sufficient conditions for the effective - range theory , and show that for the ultracold gases of of @xmath9li and @xmath10k with the fermi temperature of the order 1@xmath30k , the effective - range theory can be used as a good approximation in the resonance regime where the scattering volume is enhanced the main results of this paper are summarized as follows . in this work
we first calculate the exact expression of the low energy p - wave scattering amplitude with a pmfr .
we prove that the scattering amplitude can be expressed as @xmath31with @xmath32 the strength of the magnetic field applied along the @xmath33 axis . here
@xmath34 is the @xmath33-component of the angular momentum ; @xmath35 @xmath36 and @xmath8 are @xmath2-dependent scattering parameters .
we obtain the general expressions for them .
it is pointed out that , the denominator of @xmath37 can not be expressed as a laurent series with @xmath2-independent coefficients because @xmath38 @xmath36 and @xmath39 are not analytical functions of @xmath2 .
equation ( [ 1c ] ) shows that , in the presence of a pmfr the inconsistency of the scattering amplitude with the effective - range theory is displayed in a more complicated manner . the low - energy p - wave scattering amplitude in eq .
( [ 1c ] ) is different from the one obtained from the standard effective - range theory in the following two senses : \1 . the scattering parameters @xmath40 depend on the incident momentum @xmath41 .
the term @xmath42 $ ] can not be included in the effective - range theory . after obtaining the p - wave scattering amplitude under a pmfr
, we discuss the applicability of the effective rang theory as an approximation of the scattering amplitude ( [ 1c ] ) .
we find that , in the bec side of the pmfr where @xmath43 and @xmath8 have the same sign , sufficient conditions for the validity of effective - range theory are @xmath44 . in the bcs side of the resonance , the sufficient conditions become @xmath45 . here
@xmath46 , @xmath47 and @xmath48 are defined in eqs .
( [ r1new ] ) , ( [ r2 ] ) and ( [ r3 ] ) . if these conditions are satisfied , the scattering amplitude of our system can be approximated as @xmath49 which has the same form as eq .
( [ 1a ] ) derived by the effective - range theory .
qualitatively speaking , the above sufficient conditions means that we can use the effective - range theory if the fermonic momentum of the cold gas is low enough , the magnetic field is tuned close enough to the resonance point and the background scattering potential in the open channel is far away from the zero - energy shape resonance point . for the realistic cold gases of fermi atoms ,
if the background scattering is far away from the shape resonance , the effective range theory can usually be used in the total region where the @xmath50-wave interaction is negligible .
the paper is organized as follows . in sec .
ii , we calculate the p - wave scattering amplitude near a pmfr , and obtain the low energy expansion in ( [ 1c ] ) .
the parameters @xmath51 are expressed in terms of the background scattering parameters and the magnetic field . in sec .
iii we discuss the range of applicability of the effective - range theory .
we show that , under some simple sufficient conditions @xmath52 or @xmath53 , the scattering amplitude given by the effective - range theory is a good approximation for the exact one obtained in sec .
we further show that these conditions are well satisfied in the cold gases of @xmath14k and @xmath12li when the scattering volume is enhanced by a pmfr , and then the previous results based on the effective - range theory are applicable for these systems . in sec .
iv there are some conclusion and discussions .
we describe some detail of our calculations in the appendixes .
in this section we calculate the p - wave scattering amplitude in presence of a pmfr induced by a magnetic field along the @xmath33 direction .
we begin with the two - channel hamiltonian for the relative motion of two atoms ( fig .
1 ) : @xmath54where @xmath55 is the kinetic energy of relative motion , @xmath56 is the background scattering potential in the open channel , @xmath57 is the coupling between the open and closed channel , and @xmath58 is the interaction potential in the close channel , that has a @xmath32-dependent positive threshold @xmath59 . in this paper , for simplicity , we assume that the background potential @xmath60 is independent on the direction of @xmath24 and invariant under the so(3 ) rotation .
we further assume that , in the closed channel there are only three bare p - wave bound states @xmath61 which are near resonance with the threshold of the open channel . here
@xmath62 is the angular momentum along the @xmath33 axis .
the self energy @xmath63 of @xmath64 with @xmath62 depends on the atomic magnetic dipole . for the atoms with small magnetic dipole ,
@xmath9li , the values of @xmath65 and @xmath66 are close with each other p2bfr , while for the the atoms with large magnetic dipole , _
@xmath67k , the difference between @xmath68 and @xmath69 is quite large @xcite .
the p - wave scattering amplitude in the open channel can be defined with the standard scattering theory @xcite . to this end
we firstly introduce the two - component stationary scattering state@xmath70where @xmath71 is the eigenstate of the atomic relative momentum with eigen value @xmath72 , and @xmath73 is the mller operator defined as@xmath74 in the large interatomic distance limit @xmath26 , the asymptotic behavior of the state @xmath75 can be expressed as@xmath76with @xmath77 and @xmath78 the scattering amplitude which can be further expanded in terms of different partial waves:@xmath79here @xmath80 is the spherical harmonic function . for a scattering potential with so(3 ) symmetry ,
the partial wave scattering amplitude only depends on the quantum number of the angular momentum @xmath81 . in our case , the so(3 )
symmetry is broken by the interaction between the atomic magnetic dipole and the magnetic field .
then we have a @xmath82-dependent scattering amplitude @xmath83 . in the case of low - energy scattering between two spin polarized fermonic atoms
, one can neglect all the high partial wave scattering amplitudes @xmath84 with @xmath85 , and only consider the p - wave amplitudes @xmath86 , which can be further expressed in terms of the p - wave phase shifts @xmath87:@xmath88 during the scattering process , the bare bound state @xmath89 is coupled with the p - wave background scattering states in the open channel and significantly change the p - wave scattering amplitude @xmath37 .
this effect can be directly treated via the feshbach resonance theory , _
e.g. _ , the methods in ref .
@xcite and ref .
@xcite . after a straightforward calculation in appendix a , we find that the final phase shift @xmath87 is the sum of the background phase shift @xmath90 for the background potential @xmath91 and a correction @xmath92 given by the closed channel : @xmath93 here @xmath92 satisfies@xmath94 with @xmath95 and @xmath96 given by@xmath97 in the above we have used the background green s function @xmath98 @xmath99 and the background scattering state @xmath100 defined as@xmath101 in the following subsections , we evaluate the low - energy expression of the scattering amplitude @xmath37 by expanding the term @xmath102 in eq .
( [ fdelta ] ) in the limit @xmath103 . as shown in eq .
( [ delta1a ] ) , the phase shift @xmath104 is the sum of @xmath105 and @xmath92
. the low energy behavior of background phase shift @xmath106 is already known to be @xcite @xmath107 therefore , if we can further obtain the low - energy expansion of term @xmath108 , then the expressions of @xmath109 and @xmath37 can be calculated straightforwardly . in this subsection
we investigate the expression of @xmath111 in the limit @xmath103 . to this end , we need to expand both the numerator and the denominator of ( [ delta1 ] ) in the low - energy limit . in this paper
we assume the background scattering volume in the open channel is finite .
it can be proved that ( appendix c ) , in this case the function @xmath112 can be expanded as@xmath113with @xmath114 . on the other hand , due to the long - range nature of the van der waals potential ,
the partial wave scattering state @xmath115 is not an analytical function of the incident momentum @xmath2 in the neighborhood of @xmath116 @xcite . to investigate the low - energy behavior of @xmath95 and then the denominator of ( [ delta1 ] ) , we separate the non - analytical part of @xmath117 by introducing the background jost function @xmath118 @xcite defined as@xmath119 here @xmath120 is the canonical solution of the radial equation@xmath121 with boundary condition@xmath122 where @xmath123 is the first - order regular riccati - bessel function @xcite . according to the standard scattering theory @xcite
, @xmath124 is an analytical function of @xmath2 , and can be expanded as a talyor series of @xmath2 : @xmath125 it is pointed out that , all the odd order terms of the above taylor series are exactly zero @xcite .
thus the non - analytical part of @xmath95 is included in the term with the jost function @xmath118 .
substituting eqs .
( [ ftuter ] ) , ( [ jost1 ] ) and ( [ gk2 ] ) into eq .
( [ delta1 ] ) , we find that in the low - energy limit the factor @xmath126 takes the form @xmath127here we have the @xmath2-dependent parameters:@xmath128 ^{-1}. \nonumber \\ & & \label{rdet}\end{aligned}\]]with the parameters @xmath129 and @xmath130 defined by@xmath131 _
{ k=0 } \nonumber \\
w_{m_{z}}^{\prime } & = & \frac{1}{120}\frac{d^{5}}{dk^{5}}\left [ |\langle \phi _ { \mathrm{res}}^{(m_{z})}|w|\psi _ { k1m_{z}}^{(\mathrm{bg}+)}\rangle & & \end{aligned}\ ] ] in the above subsection we get the expansion ( [ kcot2 ] ) of the factor @xmath108 .
substituting eq .
( [ kcot2 ] ) , ( [ delta1bg ] ) and ( [ delta1a ] ) into ( [ fdelta ] ) , we finally get the low - energy behavior of the p - wave scattering amplitude @xmath37 in presence of a pmfr : @xmath132where the @xmath2-dependent scattering parameters are given by@xmath133here the parameters @xmath134 and @xmath135 are defined as@xmath136 so far we have obtained the low - energy expression of the p - wave scattering amplitude @xmath37 in the case of pmfr . with the help of the scattering theory , we obtain the general expressions ( [ f1]-[v2effa ] ) for the scattering amplitude @xmath37 as well as the scattering parameters @xmath137 , which are formulated in terms of the background scattering parameters @xmath138 and the magnetic field .
it is pointed out that , although due to the long - range nature of the van der waals potential we can not express the denominator of @xmath37 as an laurent series with @xmath2-independent coefficients , we successfully include all the @xmath2-dependence of the parameters @xmath40 into the jost function @xmath139 .
( [ v0eff ] ) clearly show the effect of pmfa that the scattering volume @xmath140 diverges under the magnetic field @xmath141 . in the realistic cold atom gases , for the creation of an observable effects with @xmath50-wave interaction ,
the scattering volume @xmath142 should be large enough .
particularly , @xmath143 should be much larger than the van der waals length so that in the bcs region , the transition temperature @xmath144 $ ] @xcite of superfluid is realizable and in the bec region the binding energy of the @xmath50-wave feshbach molecule be roubst with respect to the detail of the atom - atom interaction potential . in the end of this section
, we consider the dependence of the effective range @xmath145 on the magnetic field @xmath32 .
according to eq .
( [ v2effa ] ) , @xmath145 depends on @xmath32 through the ratio @xmath135 between @xmath146 and @xmath140 , and the quantatity @xmath147 . in the cold gases of @xmath9li and @xmath10k , the background scattering volumes @xmath146 are of the order of @xmath148 . according to our above discussion , they are too small for the creation of @xmath50-wave superfluids @xcite . therefore in these systems
the strong enough @xmath50-wave interactions can only be obtained in the resonance region with @xmath149 or @xmath150 which implies @xmath151 . on the other hand , according to eq .
( [ rdet ] ) , the dependence of @xmath152 on @xmath32 is significant when the magnetic field is far away enough from from the resonant point @xmath129 so that the factor @xmath153 is comparable or larger than @xmath154 .
the values of @xmath155 and @xmath130 are not available for @xmath9li and @xmath10k .
nevertheless , the binding energies of the @xmath50-wave feshbach molecules are measured to be linear functions of the magentic field @xcite in the region with large enough @xmath50-wave scattering volumes ( @xmath156 ) .
this observation shows that in these regions the term @xmath157 is negligible and the effective range @xmath145 can be approximated as a @xmath32-independent constant @xmath158 .
in the above section , we have obtained the expression ( [ f1 ] ) of the p - wave scattering amplitude @xmath37 in the region near the point of pmfr .
it is apparent that , this expression is different from the one ( 1a ) given by the effective - range theory in the following two senses : \1 . in the standard effective - range theory , the scattering volume @xmath159 and effective - range @xmath4 are independent on the incident momentum @xmath2 .
nevertheless , in the expression ( [ f1 ] ) the scattering parameters @xmath160 depend on @xmath2 through the jost function @xmath118 .
the term @xmath42 $ ] can not be included in the effective - range theory .
it is apparent that , if under some condition the scattering amplitude ( f1 ) can be approximated [email protected] .
, both the @xmath2-dependence of @xmath162 and the term @xmath163 $ ] can be neglected , the behavior of the system would be approximately described by the effective - range theory . in this section
, we investigate the conditions for the approximation ( f1er ) or ( [ 1 g ] ) , or the applicability of effective - range theory .
we will consider the importance of the term @xmath164 $ ] and the @xmath2-dependence of the jost function respectively .
in this subsection we search the sufficient condition for the ignorance of the @xmath2-dependence of the scattering parameters @xmath165 .
as we have discussed above , the @xmath2-dependence of the scattering parameters comes from the mode square of the jost function @xmath118 . in the ultracold gases of the fermionic atoms , the maximum value of the relative momentum of two atoms is on the order of the fermi momentum @xmath166 .
therefore , the importance of the @xmath2-dependence of the parameters parameters @xmath167 can be described by the factor @xmath168 obviously , when @xmath169 , we can replace @xmath170 with @xmath171 and neglect the @xmath172 dependence of the parameters @xmath160 .
( blue dash - dotted line ) , @xmath173 ( open circle ) , @xmath174 ( black solid line ) and @xmath175 ( red dashed line ) in eq .
( [ jost]).,width=283 ] to investigate the behavior of the the ratio @xmath176 , we first calculate the jost function @xmath118 . by means of the quantum defect theory gao3 , we can obtain the expression of @xmath170 ( see appendix d ) : @xmath177^{-1 } \nonumber \\\end{aligned}\ ] ] where @xmath178 is a @xmath2-independent coefficient and @xmath179 @xmath180 with @xmath181 defined in @xcite . in fig .
2 we plot the functions @xmath182 in the low - energy case . the parameter @xmath183 is denoted as @xmath183 is related to the background scattering parameters @xcite . expanding the p - wave phase shift in eq.(7 ) of ref . @xcite , we can express @xmath184 in terms of @xmath183 : @xmath185^{2}}\beta _ { 6}^{3 } ; \label{v0bg } \\
\mathcal{s}^{(\mathrm{bg)}}&=&-\frac{35(1+k_{l=1}^0)^{2}\pi } { % 324(k_{l=1}^{0})^2\gamma \lbrack 3/4]^{4}}\beta _ { 6}^{2}. \label{v1bg}\end{aligned}\ ] ] the above expression shows that , when @xmath186 ^ 2)\sim0.1 $ ] , we have @xmath187 and the background scattering potential is far away from the shape resonance ; when @xmath188 is much smaller than @xmath189 , the background potential is in the shape resonance region which gives @xmath190 .
now we consider the features of the ratio @xmath176 , which is determined by the parameter @xmath188 .
2 shows that in the low energy case with @xmath191 , the function @xmath174 is almost a @xmath2-independent constant and much larger than the other three d - functions .
therefore , if the parameter @xmath188 is large or the background scattering potential in the open channel is far away from the shape resonance , then according to eq .
( [ jost ] ) , the jost function @xmath192 is dominated by the term with @xmath193 . in this case the variation of @xmath170 with respect to @xmath2 is negligible and we have @xmath169 . on the other hand ,
if @xmath188 is close to zero and the background scattering potential is close to the shape resonance , then @xmath192 becomes a rapid changing function of @xmath2 and the ratio @xmath176 would be significant .
the above argument is quantitatively verified by fig .
2 , where the ratio @xmath46 is plotted as functions of @xmath2 with respect different values of @xmath183 or @xmath194 .
it is clearly shown that , if the fermonic momentum @xmath195 , then the ratio @xmath46 and the @xmath2-dependence of the scattering parameters can be neglected when the background potential is far enough from the shape resonance so that @xmath196 .
if @xmath197 , this restriction can be further relaxed to @xmath198 .
defined in eq .
( [ r2 ] ) as a function of the fermi momentum @xmath199 with @xmath200 ( @xmath201 , blue solid line ) , @xmath202 ( @xmath203 , green empty square ) , @xmath204 ( @xmath205 , black empty circle ) , @xmath206 ( @xmath207 , blue dash - dotted line ) , @xmath208 ( @xmath209 , green filled square ) , @xmath210 ( @xmath211 , red dashed line).,width=283 ] now we discuss the importance of the term @xmath164 $ ] .
since the purpose of this paper is to obtain the sufficient condition for the effective - range theory , or the approximation ( [ f1er ] ) , for simplicity , in this subsection we assume the condition @xmath169 obtained in the above section is already met , and the @xmath2-dependence of the coefficients @xmath212 can be neglected . in the bec side of the pmfr where @xmath213 , since @xmath214 , the parameters @xmath8 and @xmath43 have the same sign .
in that case , if the absolute value of @xmath215 is much smaller than the one of @xmath216 , then it would be also much smaller than @xmath217 , and can be neglected .
it is obvious that , in the limit @xmath103 , the term @xmath215 would be much smaller than @xmath218 .
therefore the importance of the term @xmath219 is actually determined by the ratio @xmath47 between the two terms for the upper limit of the relative momentum @xmath199 : @xmath220}{1/[% \mathcal{v}^{\mathrm{eff}}(0;b;m_{z})k_{f}^{2}]}\right\vert .\end{aligned}\ ] ] when @xmath47 is much smaller than unity , we can neglect the term @xmath221 . if @xmath47 is comparable or larger than unity , the term @xmath222 would be necessary for the theory . the straightforward calculation with eqs .
( [ v0eff ] ) , ( [ v1eff ] ) , ( v0bg ) and ( [ v1bg ] ) yields @xmath223 in the practical cold atom systems we have @xmath224 .
therefore , in the resonance region with @xmath225 , the ratio @xmath47 in eq .
( [ r1new ] ) is much smaller than unity , and then the term @xmath226 is negligible . in the bec side of the resonance with @xmath227 , the parameters @xmath228 and @xmath8 have different signs . in that case , there is a speical momentum@xmath229which makes the terms @xmath230 $ ] and @xmath231 $ ] cancel with each other or@xmath232 therefore , if the atomic relative momentum @xmath2 is far away from @xmath233 , the absolute value of @xmath234 would be quite different with the one of @xmath235 . in that case
we can still neglect the term @xmath222 under the condition @xmath236 or @xmath237 .
if the atomic relative momentum @xmath2 is in the neighborhood of @xmath238 and the terms @xmath239 is canceled with @xmath235 , the scattering amplitude ( [ f1 ] ) can be expressed as@xmath240 in that case , if the absolute value of @xmath241 $ ] is much smaller than @xmath238 , we an also neglect the term with @xmath36 , even in the neighborhood of @xmath238 .
we define a parameter @xmath48 as @xmath242then the term with @xmath36 can be neglected when @xmath243 . a further calculation with eqs .
( [ kstar ] ) , ( [ v0eff ] ) , ( [ v1eff ] ) , ( [ v0bg ] ) and ( [ v1bg ] ) implies that @xmath244 as shown above , the condition @xmath245 is obtained for the momentum region @xmath246 . since the realistic momentum of the atomic relative motion takes the value between zero and @xmath199 , in the cases with @xmath247 , we can disregard the restriction of the ratio @xmath48 , and use effective - range theory under the condition @xmath248 in both the bec and the bcs sides of the resonance . in summary , the general sufficient conditions for the effective - range theory in the bcs side of the resonance can be summarized as @xmath249while the ones for the bec side are @xmath250from the definition of the ratios @xmath251 and @xmath48 , we notice that in the realistatic cold gases of fermi atoms , the crucial factors for the usage of effective range theory is the background @xmath50-wave scattering volume @xmath252 and the @xmath32-dependence of the factor @xmath253 . if the the background @xmath50-wave scattering is far away from the shape resonance so that @xmath254 then according to our previous discussions and eqs .
( [ r1new ] ) , the conditions @xmath44 can be satisfied in the region @xmath255 where the @xmath50-wave interaction is strong enough for the creation of @xmath50-wave superfluids . in that region
we also have @xmath256 .
if @xmath257 can be further approximated as a @xmath32-independent constant which is of the order @xmath258 , then the condition @xmath245 can also be satisfied , and the effective range theory can be used a good approximation for the realistic scattering amplitude . in the following subsection
we show that the pmfrs in the cold gases of @xmath10k and @xmath9li are just of this case . in the above subsections we obtained the sufficient conditions ( [ sc1 ] ) , ( [ sc2 ] ) of the effective - range theory for the p - wave scattering amplitudes of polarized fermonic atoms near a pmfr . in this subsection , with the help of the conditions , we perform a discussion on the usage of effective - range theory in the ultracold gases of @xmath10k and @xmath9li . for the ultracold gas with @xmath10k atoms in the state @xmath259 , we have @xmath260(a.u . ) @xcite and @xmath261 @xcite .
these parameters leads @xmath262 and @xmath263 .
if the fermi temperature @xmath264k , then we have @xmath265 .
the straightforward calculation shows that @xmath266 .
therefore @xmath2-dependence of the scattering parameters can be safely neglected .
the p - wave feshbach resonance for the states with @xmath267 occurs at @xmath268 g with width @xmath269 g and effective - range @xmath270 .
the resonance for the states with @xmath271 occurs at @xmath272 g with width @xmath273 g and effective - range @xmath274 expk40dipo . according to these data we have @xmath275 when @xmath276 .
then effect from the ratio @xmath48 is also negligible .
therefore the sufficient condition for the usage of effective - range simply becomes @xmath277 .
further calculation shows that @xmath278 when @xmath279 .
then the effective - range approximation ( [ f1er ] ) is applicable for @xmath10k atoms in the state @xmath259 in the whole region of pmfr with @xmath280 .
the condition for the effective - range approximation is broken only in the small region @xmath281 g ( @xmath271 ) or @xmath282 g ( @xmath267 ) where we have @xmath283 or @xmath284 .
now we consider the gas with @xmath9li atoms in the ground hyperfine state @xmath285(a.u . )
li6parameter , xxxgao and @xmath286 .
these parameters leads @xmath287 and @xmath288 .
if the fermi temperature @xmath289k , we have @xmath290 which implies @xmath291 . then similar as above , the effective range approximation ( [ f1er ] ) is also applicable for @xmath9li atoms in the whole region of pmfr with @xmath292 .
in this work we obtain the explicit expression of the p - wave scattering amplitude of two ultracold spin polarized fermonic atoms near the p - wave feshbach resonance .
we show that due to the long rang nature of the van der waals potential , the scattering amplitude is explicitly described by eq .
( [ 1c ] ) in the low - energy case . with the help of the quantum defect theory ,
we formulate all the scattering parameters @xmath293 in terms of the background parameters and the inter - channel coupling .
based on this result , we discussed the applicability of the effective - range theory , or the eq .
( [ 1d ] ) as an approximation of the exact scattering amplitude .
we show that , in the bec side of the resonance , the sufficient conditions of the effective - range theory can be quantitatively described as @xmath294 while in the bcs side the conditions become @xmath295 , where @xmath296 are defined in eqs .
( r1new ) , ( [ r2 ] ) , and ( [ r3 ] ) .
the applicability of the effective - range theory for the ultracold gases of @xmath10k and @xmath9li are examined with our results .
the effective - range theory is shown to be a good approximation in both of the two cases in the resonance regime where the absolute value of the scattering volume is equal to or larger then the background one .
in this appendix we post the derivation of the p - wave phase shift @xmath297 in eqs .
( [ delta1a ] ) and ( [ delta1 ] ) .
our calculation is based on the method in ref .
we begin from the scattering state @xmath75 in eq .
( [ phik ] ) . according to the scattering theory @xcite , the open channel component @xmath298 and closed channel component @xmath299 of @xmath75
satisfy the two channel lippmman - schwinger equation @xmath300 with the free green s functions@xmath301we further define the background green s function @xmath302 and the closed channel green s function @xmath303 as@xmath304then we have the relationships @xmath305substituting eqs .
( [ g01 ] ) and ( [ g02 ] ) into the last terms of the right hand side ( r.h.s . ) of eqs .
( [ lp1a ] ) and ( [ lp1b ] ) , and using the lippmman schwinger equation ( [ psibg ] ) for the background scattering state @xmath306 , we get the equation which relates @xmath75 with the background scattering state @xmath306 @xcite : @xmath307 to calculate the p - wave phase shifts @xmath87 , we operate the projection operator @xmath308 for the manifold @xmath309on both of the two sides of eqs . ( [ m ] ) and ( [ n ] ) .
then we have@xmath310here we have used@xmath311which are guaranteed by the rotational symmetry along the @xmath33-axis of the system .
we also made the approximation @xmath312 that is , we only take into account the contribution from the near - resonance bound state @xmath313 in the closed channel . substituting eqs .
( [ nn ] ) into eq .
( [ mm ] ) , we get@xmath314with @xmath315replacing the @xmath316 in the r.h.s . of eq .
( [ p ] ) with the r.h.s . of ( [ o ] ) , we get@xmath317 substituting eq .
( [ q ] ) into eq .
( [ o ] ) , and using the asymptotic expression ( [ asypsi ] ) of the scattering state and the definition ( parf ) of the partial wave scattering amplitude , we can obtain the p - wave scattering amplitude @xmath318 here @xmath95 is defined in eq .
( psik1mz ) .
@xmath319 is defined as @xmath320with@xmath321the background state with in - going boundary condition . in the above calculation we also used the asymptotic behavior of the background green s function:@xmath322with @xmath77 . with straightforward calculation
, we can further rewrite the scattering amplitude @xmath37 in eq .
( [ fcomplex ] ) as @xmath323with@xmath324 in the derivation of eq .
( [ finalf ] ) we have used the relationship ( appendix b)@xmath325and @xmath326here @xmath327 refers to the principle value of the integral . considering the relationship ( [ fdelta ] ) between the scattering amplitude @xmath37 and the phase shift @xmath87 , it is easy to prove that the phase shift @xmath87 corresponding to the scattering amplitude ( [ finalf ] ) is the one given in eqs .
( [ delta1a ] ) and ( [ delta1 ] ) .
it is pointed out that , this result can also be proved with the method in ref .
( @xcite ) .
in this appendix we prove the formula ( [ nppsi ] ) in appendix a. we begin from the relationship @xcite between the three dimensional scattering states with ingoing and outgoing boundary conditions : @xmath328considering the definitions ( [ psik1mz ] ) and ( [ psik1mzn ] ) of @xmath329 , we can obtain @xmath330we further define the one dimensional functions @xmath331 as taylor @xmath332using the relationships @xmath333we get@xmath334 on the other hand , we know that @xmath335 and @xmath336 satisfy the same differential equation @xmath337with the same boundary condition @xmath338 .
then @xmath339 is proportional to @xmath340 . to calculate the ratio between @xmath341
, we consider their asymptotic behaviors in the limit @xmath26 : @xmath342 ; \nonumber \\ \\
f_{k1}^{(\mathrm{bg}-)}(r ) & = & \hat{j}_{1}(kr)+kf_{k1}^{(\mathrm{bg}% ) } ( k)^{\ast } [ \hat{n}_{1}(kr)-i\hat{j}_{1}(kr ) ] \nonumber \\\end{aligned}\]]with @xmath343defined in ( [ j1 ] ) and @xmath344 the irregular first order riccati - bessel function . comparing the coefficients of @xmath345 , we have@xmath346here we have used@xmath347substituting eq . ( [ c10 ] ) into eq .
( [ c3 ] ) , we get eq .
( [ nppsi ] ) .
in this appendix we prove the eq .
( [ gk2 ] ) .
we firstly rewrite the factor @xmath96 as @xmath348 here we have used the identity@xmath349the first term in the r.h.s . of ( [ expgk ] ) is independent on @xmath2 .
it contributes the constant term @xmath350 in eq .
( [ gk2 ] ) . on the other hand , the second term in eq .
( [ expgk ] ) can be re - written as @xmath351 in the limit @xmath116 , we have@xmath352 .
on the other hand , as we have shown in sec .
iii.b , the factor @xmath353 tends to a non - zero constant in the low energy limit if the background scattering volume in the open channel is finite . using the relationship ( [ jost1 ] ) between @xmath118 , @xmath354 and @xmath355 , and the low - energy behavior ( [ ftuter ] ) of @xmath356 , it is easy to prove that the factor @xmath357 is proportional to @xmath358 in the limit @xmath359 . therefore , the above integration in eq .
( [ c44 ] ) converges to a finite constant in the limit @xmath360 . then the expansion in eq .
( [ gk2 ] ) is proved and we have @xmath361
in this appendix we calculate the jost function @xmath118 of the background scattering state . to this end
, we introduce a function @xmath362 where @xmath363 is defined in eq .
( [ jost1 ] ) .
it is apparent that @xmath364 is a solution of the radial equation @xmath365with a @xmath2-independent boundary condition @xmath366 following the spirit of quantum defect theory @xcite , we assume the scattering potential @xmath367 can be approximated as the van der waals potential @xmath25when @xmath368 is larger than a critical distance @xmath369 which is much smaller than @xmath258 . in the region with @xmath370 , @xmath367 is assumed to be so large that the atomic kinetic energy @xmath371 is negligible in comparing with @xmath91 , and then @xmath372 is independent on @xmath2 . in the region
@xmath373 , @xmath372 is the superposition of the two independent solutions @xmath374 and @xmath375 of eq .
( [ radical ] ) @xcite ( in ref .
@xcite , @xmath374 and @xmath375 are denoted as @xmath376 and @xmath377):@xmath378 in the short distance region with @xmath379 , @xmath380 and @xmath381 are almost independent on @xmath2 @xcite .
the wave function @xmath372 in the two regions are connected at point @xmath382 , where @xmath372 , @xmath383 and @xmath375 are approximately independent on @xmath2 .
then we know that @xmath384 and @xmath385 are independent on @xmath2 and we have @xmath386with @xmath387 . the above equation yields@xmath388 in the region @xmath26 , the asymptotic behaviors of @xmath389 and @xmath390 can be expressed in terms of @xmath181 ( @xmath391 ) defined in @xcite @xmath392 ; \nonumber \\ & & \label{ff1 } \\
\kappa _ { \epsilon_k1}^{(0)}(r ) & \rightarrow & \sqrt{\frac{2}{\pi k}}\left [ z_{gf}\sin \left ( kr-\frac{\pi } { 2}\right ) -z_{gg}\cos \left ( kr-\frac{\pi } { % 2}\right ) \right ] .
\nonumber \\ & & \label{g1}\end{aligned}\]]on the other hand we know that in the same limit we have @xcite @xmath393
. \nonumber \label{sc } \\ & & \end{aligned}\]]together with eqs .
( [ ss ] ) , ( [ ff1 ] ) , ( [ g1 ] ) as well as ( [ sc ] ) ,
we obtain the expression of ( [ jost ] ) of @xmath162 and the background phase shift : @xmath394where @xmath179 and @xmath395 is the one in sec .
the result ( [ tgdeltabg ] ) is also given in @xcite .
e. wille , f. m. spiegelhalder , g. kerner , d. naik , a. trenkwalder , g. hendl , f. schreck , r. grimm , t. g. tiecke , j. t. m. walraven , s. j. j. m. f. kokkelmans , e. tiesinga , and p. s. julienne , phys .
. lett . * 100 * , 053201 ( 2008 ) . | most of the current theories on the p - wave superfluid in cold atomic gases are based on the effective - range theory for the two - body scattering , where the low energy p - wave scattering amplitude @xmath0 is given by @xmath1 $ ] , where @xmath2 is the incident momentum , and @xmath3 and @xmath4 are the @xmath2-independent scattering volume and effective - range , respectively . however , due to the long - range nature of the van der waals interaction between two colliding ultracold atoms , the p - wave scattering amplitude of the two atoms is not described by the effective - range theory @xcite . in this paper
we provide an explicit calculation for the p - wave scattering of two ultracold atoms near the p - wave magnetic feshbach resonance ( pmfr ) .
we show that the low energy p - wave scattering amplitude in the presence of pmfr takes the form @xmath5 $ ] where @xmath6 @xmath7 and @xmath8 are @xmath2-dependent parameters .
based on this result , we show sufficient conditions for the effective range theory to be a good approximation of the exact scattering amplitude . using these conditions
we show that the effective - range theory is a good approximation for the p - wave scattering in the ultracold gases of @xmath9li and @xmath10k when the scattering volume is enhanced by the resonance . |
in this paper , we have used the adelaide approach to calculating quark model parton distributions to calculate the effect of csv on the parton distributions of the proton . for the valence quarks ,
we verify in a model independent fashion the anomalously large csv effects observed in the minority valence quark distribution by several authors @xcite , and we relate these distributions to the difference between the @xmath5 and @xmath6 valence distributions . since there is little data to constrain the size of the csv distributions , proposed tests of csv @xcite will provide a sensitive test of the physics contained in qcd - inspired models of the nucleon . additionally , we have made estimates of the size of csv effects in the nucleon sea .
first , we assume a strong violation of charge symmetry , which would alter the value of sum rules which assume charge symmetry .
for the gottfried sum rule , there is no csv contributions unless there is a simultaneous violation of flavor symmetry in either the sea or valence quarks . in either case
, we estimate that the csv contributions to the gottfried sum rule are much smaller than are suggested by experiment .
we have also estimated the weak csv contribution to the nucleon sea , using methods similar to the valence quark csv calculations .
again , we find the charge symmetry violating distributions are extremely small , and highly unlikely to make significant contributions to any observable .
the authors wish to acknowledge support from nsf research contract no .
phy-9722706 , and to d. murdock for assistance in producing the figures , and a.w .
thomas for useful discussions of alternate methods for arriving at an analytic expression for the csv distributions .
99 p. amaudruz _
et al _ ( nmc collaboration ) , phys .
lett . , 2712 ( 1991 ) ; phys .
* b295 * , 159 ( 1992 ) .
a. baldit _
( na51 collaboration ) , phys .
* b332 * , 244 ( 1994 ) .
p.e . reimer _
( e866 collaboration ) , _ proceedings of the sixth conference on the intersections between particle and nuclear physics _ , edited by t.w .
donnelly , ( american inst . of physics ,
1997 ) , p. 643 ; e. hawker _
et al _ , phys .
80 , 3715(1998 ) .
b. adeva et al .
( smc collaboration ) , _ phys .
lett._*b302 * ( 1993 ) 533 ; p.
l. anthony et al .
lett._*71 * ( 1993 ) 959 .
a.i . signal and a.w .
thomas , phys . lett . *
b191 * , 206(1987 ) , + c.j .
benesh and g.a .
miller , phys .
* d36 * , 1344(1987 ) .
al . , cteq collaboration , phys . rev . * d55 * , 1280 ( 1997 ) ; + a.d.martin,w.j.stirling and r.g.roberts , _ phys .
lett . _ * b252 * , 653 ( 1990 ) .
e. n. rodionov , a. w. thomas and j. t. londergan , mod .
letts . a * 9 * , 1799 ( 1994 ) .
e. sather , phys .
b * 274 * , 433 ( 1992 ) .
benesh and t. goldman , phys .
* c55 * , 441 ( 1997 ) . r.m . davidson and m. burkardt ,
* b403 * , 134 ( 1997 ) .
w.g . seligman _
et al . _ , phys .
lett . * 79 * , 1213 ( 1997 ) .
londergan and a.w .
thomas , _ progress in particle and nuclear physics _
, in press .
jaffe , in _ relativistic dynamics and quark nuclear physics _ , proceedings of the los alamos summer school , 1985 , edited by m.b .
johnson and a. picklesimer , ( wiley , new york , 1986 ) . f.e.close and a.w.thomas , _ phys .
* b212 * , 227 ( 1988 ) .
a result similar to ours may be obtained by first integrating out the delta function and then neglecting the transverse momentum in the z - component of the quark current .
one then obtains an expression identical to eq .
4 , except that the @xmath3 derivative now acts only on @xmath106 .
the corresponding change in the minority valence quark distribution is reduced by about @xmath107 compared to our results .
thomas , private communication .
s. kumano , phys .
rep.*303 * , 183 ( 1998 ) .
r. p. bickerstaff and a. w. thomas , _
j. phys._*g15 * , 1523 ( 1989 ) .
k. gottfried , phys .
lett . * 18 * , 1174 ( 1967 ) .
ma , phys .
b274 * , 111 ( 1992 ) .
the charge symmetry violating contribution to the evolution equations has been calculated and found to be small .
h. spiesberger , phys . rev . *
d52 * , 4936(1995 ) .. see ref .
@xcite and m. glck , e. reya and a. vogt , z. phys .
* c67 * , 433 ( 1995 ) .
donoghue and e. golowich , phys . rev . * d15*. 3421 ( 1977 ) .
t. goldman , k.r .
maltman , and g.j .
stephenson jr . , phys . lett . * b228 * , 396 ( 1989 ) . * fig .
1 - diquark mass contribution to charge symmetry violating quark distributions for minority valence quarks .
curves show @xmath108 defined from eqs .
[ twelve ] and [ analyt ] .
curves were calculated using the cteq4 lq distribution at @xmath109 gev@xmath61 . also shown
are model calculations from refs .
@xcite and @xcite . * fig .
2 - charge symmetry violating quark distributions for majority ( @xmath110 ) and minority ( @xmath111 ) valence quarks .
solid curves include effects of both @xmath112 mass difference and diquark mass contribution ; dot - dashed curve shows minority quark csv neglecting the @xmath112 mass difference .
curves were calculated using the cteq lq distribution at @xmath109 gev@xmath61 . * fig .
3 - photons dressing the quark - gluon vertex during the production of sea quarks .
the probability of producing a sea quark pair will depend on the charge of the created quarks(fig .
3b ) and on the quark that initially produced the gluon(fig .
3a ) . * fig .
4 - charge symmetry violating sea quark distributions .
the solid , dotted , dashed , and dot - dashed curves are @xmath113 , @xmath114 , @xmath115 , and @xmath116 , respectively .
.charge asymmetric contributions to the mass of multi - quark composite systems . @xmath117
mev and @xmath118 mev are determined by fitting the neutron - proton and iso - vector diquark mass splittings . [
cols="^,^",options="header " , ] | recent calculations of charge symmetry violation ( csv ) in the valence quark distributions of the nucleon have revealed that the dominant symmetry breaking contribution comes from the mass associated with the spectator quark system .
assuming that the change in the spectator mass can be treated perturbatively , we derive a model independent expression for the shift in the parton distributions of the nucleon .
this result is used to derive a relation between the charge and flavor asymmetric contributions to the valence quark distributions in the proton , and to calculate csv contribution to the nucleon sea . the csv contribution to the gottfried sum rule
is also estimated , and found to be small .
-0.35 in 0.00 in 0.25 in 6.50 in 8.5 in [ [ section ] ] recent measurements of the flavor @xcite and spin @xcite dependence of quark distributions in the nucleon have led to a revival of interest in the soft qcd physics that determines the shape and normalization of parton distributions @xcite , and has led to a re - examination of some of the fundamental assumptions embodied in the parametrizations @xcite used to describe data from high energy experiments . in particular , the violation of charge symmetry in the valence quark distributions has been calculated by a number of authors @xcite in the context of quark models , and surprisingly large violations , as large as 5 - 10% , have been found at large @xmath0 @xcite . although charge symmetry is assumed in all phenomenological parton distributions , there is not a great deal of direct experimental evidence which justifies this assumption . the strongest upper limit on parton charge symmetry violation(csv ) can be obtained by comparing the @xmath1 structure function for charged lepton deep inelastic scattering with the @xmath1 structure function measured in neutrino - induced charged current reactions .
the ccfr group has recently carried out such a test using the most recent available data @xcite .
they compared their neutrino cross sections on iron with the @xmath1 structure functions extracted from muon - deuterium measurements of the nmc group @xcite . in the region @xmath2
, these recent experiments can place upper limits of about 6% on parton csv . for larger @xmath3 the upper limits are substantially larger , while for @xmath4 the present data appears to indicate a substantial violation of charge symmetry .
a number of experiments have been suggested which would measure directly charge symmetry violation in parton distributions @xcite . in this paper
, we combine the approaches of refs . @xcite and @xcite and
examine the violation of charge symmetry in the valence and sea quark distributions of the nucleon in a manner independent of the choice of any particular quark model . in the following section
, we present the formalism used to produce parton distributions from quark model wavefunctions , and examine the variation of those distributions with small changes in the mass of the system of spectator quarks . in the next section
, we use these results to examine the changes in the valence quark distribution due to su(4 ) violation via color magnetism , and charge symmetry violation via quark masses and electro - magnetism . combining these results
, we obtain a wavefunction independent prediction relating the dominant contribution to the csv valence distribution to the well known difference between the @xmath5 and @xmath6 valence distributions . in the penultimate section ,
we estimate the charge symmetry violating sea quark distributions , assuming that the dominant csv effect comes from the mass and electro - magnetic differences of the participating quarks , rather than from dependence of the quark wavefunctions on charge symmetry , as is the case for the valence quark csv contributions @xcite .
additionally , we make simple estimates of the contribution of csv to the gottfried sum rule .
finally , we compare our results with those of other authors and discuss their implications .
[ [ section-1 ] ] for the purposes of this paper , we shall adopt the adelaide method for calculating quark distributions from model wavefunctions @xcite .
the starting point for the method is the reduction of quark distributions to the form @xcite @xmath7 where @xmath3 is the bjorken scaling variable , @xmath8 is the mass of the target , @xmath9 is the light cone wavefunction for the struck quark in momentum space , and @xmath10 is the energy of the struck quark .
an analogous expression gives the antiquark distributions @xmath11 the essence of the adelaide approach lies in the kinematic assumption that the three quark system can be divided into the struck quark plus a diquark spectator system , and that the distribution of masses of the spectator system is sufficiently sharply peaked that the diquark can be thought of as an on - shell system with a definite mass .
this completely specifies the kinematics of the problem and allows the quark energy appearing in the delta functions of eqs .
[ one ] and [ two ] to be replaced by the difference between the target mass and the on - shell diquark energy .
since the direct calculation of these distributions is not our interest here , we will not reduce this expression further , but simply note that the expression for the quark distribution can be reduced to a single integral over the momentum space wavefunction of the struck quark @xcite , and is guaranteed to have the correct support as a function of @xmath0 regardless of the properties of the wavefunctions used to evaluate the integral .
[ [ section-2 ] ] with the exception of ref .
@xcite , the approach to study the violation of parton distribution symmetries has been to calculate entire distributions in the framework of a given quark model and then to compare the results from small variations in the model parameters . in refs . @xcite and @xcite , the dependence of the valence quark distribution on the model wavefunction and diquark
mass parameter has been studied in the context of examining charge and flavor symmetry in the valence sector .
the dependence of the valence distributions on the details of the quark wavefunction was found to be small in comparison to the effects generated by changes in the diquark mass parameter .
similar conclusions were reached in ref .
@xcite , where charge symmetry breaking effects were studied using a different approach to connect quark model wavefunctions to valence distributions . based on these results
, we shall assume that the dominant contribution to the breaking of parton distribution symmetries is generated not by the changes in struck quark wavefunctions , but by the changes in the kinematic constraints brought about by shifts in the masses of the spectator quark systems and the nucleon itself .
this allows a quark model independent prediction for the change in parton distributions when the symmetry is broken . to see this , begin with eq .
[ one ] for the valence quark distribution , and make a small variation in the mass of the spectator diquark , @xmath12 , where @xmath13 is the diquark mass in the symmetric limit , and @xmath14 is the shift in the diquark mass due to symmetry breaking . expanding eq .
[ one ] to first order in @xmath15 , we obtain @xmath16 where quantities with the superscript @xmath17 are taken in the symmetric limit , and @xmath18 is the diquark energy . changing the derivative of the delta function to a derivative in @xmath3 , using kinematics to solve for @xmath19 , and neglecting the transverse momentum of the diquark@xcite , one finally obtains @xmath20 . \label{dqanal}\ ] ] as we shall demonstrate in the following , this expression has two important advantages over explicit model realizations : first , as in ref .
@xcite , @xmath21 is determined in terms of @xmath22,so the change in the distribution can be extracted from measured distributions rather than models .
second , the size of the change is controlled by the diquark mass shift , which allows us to relate different symmetry breaking effects to one another .
[ [ section-3 ] ] in models where confinement is implemented via interactions which are independent of both spin and flavor , the wavefunctions of the up and down valence quarks are identical , leading via eq .
[ one ] to the prediction @xmath23 where @xmath24 and @xmath25 are respectively the up and down valence quark distributions , and the factor of two merely expresses the relative normalization of the distributions .
these relations , which relate distributions with different flavors within the same hadron , are somewhat inappropriately termed flavor symmetries @xcite .
such symmetries are a consequence of dynamical assumptions about the nature of confinement in qcd . in this instance , the quark model su(4 ) spin - isospin symmetry gets broken by the color hyperfine interaction , leading to the well - known dominance of @xmath26 over @xmath27 at large @xmath3 .
this symmetry breaking has been modeled by close and thomas @xcite , who considered the spin - flavor correlations in the su(4 ) nucleon wavefunction and the mass splitting of spin 1 and spin 0 diquark pairs brought about by the color hyperfine interaction . in this picture ,
the valence distribution depends not on the flavor of the struck quark , but rather on the spin of the spectator diquark system .
explicitly , @xmath28 where the superscript refers to the spin of the diquark spectator . in the su(4 ) symmetric limit , @xmath29 and eq . [ five ] is recovered .
color hyperfine effects are included via a hamiltonian of the form @xmath30 with @xmath31 the spin of quark @xmath32 , @xmath33 the corresponding color generator , and @xmath34 mev is normalized by the @xmath35 splitting .
the diquark masses are shifted according to @xmath36 so that , from eq . [ dqanal ] , @xmath37 and @xmath38 interestingly
, this pattern of symmetry breaking predicts that the valence distributions measured in lepto - production from neutron targets are precisely the su(4 ) symmetric distributions .
now we turn our attention to the case of charge symmetry violation , which relates distributions of opposite isospin in targets of opposite isospin . in the charge symmetric
limit , @xmath39 where for example @xmath40 denotes the distribution of @xmath5 quarks in the proton , @xmath41 is the distribution of @xmath6 quarks in the neutron , and we have analogous definitions for the minority quark distributions . at the quark level ,
charge symmetry is violated by quark mass and electro - magnetic effects and is generally expected at the level of one per cent .
since both of these effects are iso - vector , there is no csv contribution to the mass of the isoscalar diquark , and we may write , for the majority quark distributions , @xmath42 where @xmath43 is the change in the quark distribution generated by the shift in the @xmath44 diquark masses via eq .
[ dqanal ] . since the only difference between the csv case and the flavor / su(4 ) symmetry violating distribution is the value of @xmath14 , the two distributions are the same up to a normalization , so that @xmath45 from the bag model , a robust estimate of the isovector diquark mass splitting is -4 mev @xcite , yielding @xmath46 . using the cteq4 lq distributions @xcite to parametrize the valence distributions ,
the charge symmetry violating contribution to the minority quark valence distribution due to the shift in the diquark mass is shown in fig .
1 . we plot the quantity @xmath47 , where the csv terms are calculated from eq .
[ analyt ] . also shown
are result of a bag model calculations using two different approaches to extract the quark distributions from model wavefunctions @xcite . the reasonable agreement between the analytic result of eq .
[ analyt ] and the two different model calculations of csv gives us confidence that the model calculations give rather robust predictions of valence quark charge symmetry violation . to complete the calculation of the csv distributions we must include the effect of the proton - neutron mass difference
this proceeds in essentially the same manner outlined in the last section , with the slight subtlety that the variation with respect to the target mass should be taken keeping the product @xmath48 fixed , as the mass is implicit in the definition of @xmath3 ( see ref .
@xcite for an explanation of this point ) .
the resulting addition to the csv distribution yields @xmath49 where @xmath50 mev is the proton - neutron mass difference.the resulting csv distributions ( multiplied by @xmath3 ) are shown in fig . 2 for the cteq4 lq distributions , along with the csv contribution to the minority quark distribution ( @xmath51 ) from fig .
1 ( the dash - dot curve in fig . 2 ) . as a consequence of the singularities in the cteq parametrizations ,
the nucleon mass correction diverges non - integrably at @xmath52 , cf .
eq . [ dqanal ] .
this reflects the inadequacy of the diquark spectator approximation in this region , which is dominated by states having one or more pairs of sea quarks . at larger @xmath3 , where the calculation is more reliable ,
the nucleon mass correction partially cancels out the change due to the diquark mass shift , a result which can be expected on the basis of eq .
[ three ] . in accordance with the results of refs .
@xcite , the magnitude of the csv contribution to the minority and majority quark distributions are roughly comparable , so that the smaller minority quark distribution is more sensitive to charge symmetry violation than the majority distribution .
[ [ section-4 ] ] there has been much interest in the question of flavor symmetry violation in the nucleon sea , following the precise measurement of the gottfried sum rule @xcite by the nmc group @xcite .
the gottfried sum rule is obtained by integrating the @xmath1 structure functions for lepton - induced deep inelastic scattering(dis ) on protons and neutrons , @xmath53 \nonumber \\ & = & { 1\over 3 } + { 8\over 9 } \left ( p_{\bar{u}/p } - p_{\bar{u}/n } \right ) + { 2\over 9 } \left ( p_{\bar{d}/p } - p_{\bar{d}/n } \right ) ~~ , \label{gottf}\end{aligned}\ ] ] where we define the sea quark probabilities as @xmath54 in deriving eq .
[ gottf ] we have used the normalization of the valence quark distributions ; if we assume parton charge symmetry we obtain the `` standard '' form for the gottfried sum rule , @xmath55 by measurements over the range @xmath56 for muon dis on proton and deuteron targets , and extrapolation over the unmeasured region , the nmc group obtained @xmath57 .
if one assumes charge symmetry , this result implies an asymmetry in the probabilities for finding up and down sea quarks in the proton , @xmath58 however , as was pointed out by ma @xcite , we could alternatively assume flavor symmetry but not charge symmetry for the sea quark distributions , in which case the nmc result would imply @xmath59 experimental upper limits on sea quark charge symmetry can not at present rule out parton csv contributions .
so it is possible that the gottfried sum rule result arises either from csv effects , or from a combination of flavor symmetry and charge symmetry violation . in this section
we will estimate the magnitude of charge symmetry violation for sea quarks , and the contribution this might make to the gottfried sum rule .
the effect of flavor symmetry or charge symmetry violation on the sea quark distributions is complicated by the fact that there are no sum rules constraining the normalization of the sea quark distributions .
indeed , in the standard parametrizations @xcite at high momentum scales the total number of sea quarks is infinite , a fact that has been used to suggest that fractionally small csv in the sea may lead to sizeable contributions to the gottfried sum rule @xcite .
since this contribution , regardless of its source , is independent of the momentum scale at which it is measured @xcite , it is amenable to modeling at a relatively low scale , say @xmath60 gev@xmath61 , where it has been argued that the number of sea quarks in the nucleon is finite @xcite .
[ [ section-5 ] ] the csv contributions to the sea separate , somewhat artificially , into two classes : strong csv effects which change the number of sea quarks in the nucleon ( and potentially contribute to the gottfried sum rule ) , and weak csv effects which change the shape of the sea distributions without altering their normalization .
we shall begin by considering the first of these classes , since the extent to which the gottfried sum rule violation is a reflection of csv rather than flavor symmetry violation may provide an important constraint on the second class of csv contribution .
the simplest perturbative estimate of the size of the csv contribution is obtained by modeling the nucleon as a valence state coupled to a small number of states with an additional quark - antiquark pair @xmath62 with @xmath63 a normalization , @xmath64 the amplitude for finding the `` extra '' quark anti - quark pair of flavor @xmath65 , and @xmath66 is a three quark state with the same third component of isospin as the original three quark state . assuming flavor symmetry , the amplitude for creating an extra quark pair of flavor @xmath67 is given , in perturbation theory , by @xmath68 where @xmath69 is the mass of the 4 quark , anti - quark state , and @xmath70 is a typical hadronic mass scale determined by the strong coupling constant and the details of the wavefunction.if charge symmetry is also good , this amplitude is independent of the flavor of the quark - antiquark pair , so that the probability for creating a quark anti - quark pair of flavor @xmath67 is @xmath71 which , apart from pauli effects which violate the flavor symmetry we ve assumed , is essentially the result of donoghue and golowich @xcite . to incorporate charge symmetry violation into this picture requires only an estimate of the mass shift due to quark mass and electro - magnetic effects .
we parametrize this difference via a term that counts the number of up and down quark masses in the composite system , and through the inclusion of an electro - magnetic contribution to the energy which we assume is proportional to the sum of the pairwise products of the charges of the constituents . the quark mass difference and electro - magnetic energy
are then fit to the mass shifts of the diquark and nucleon , and the results are extrapolated to systems with more quarks .
this procedure clearly neglects a great deal of the physics of these systems , but should be sufficient to provide an estimate of the energy scales involved.the resulting mass shifts between charge conjugate systems can then be written @xmath72 where @xmath73 is the number of quarks plus antiquarks of flavor @xmath74 in the system in question , @xmath75 is the change in the sum of the products of charges under charge conjugation , and @xmath76 is the average electro - magnetic energy of each quark pair . the mass shifts that result from this procedure are shown in table 1 .
expanding to first order in the symmetry violating terms yields relations between the number of sea quarks of charge conjugate flavors in the proton and neutron : @xmath77 where @xmath78 mev is the constituent quark mass in the symmetric limit .
remarkably , since the energy required to make a pair of a particular flavor is the same for both the proton and neutron in this parametrization , there is no contribution to the gottfried sum rule in the flavor symmetric limit . in a more sophisticated treatment of quark mass and electro - magnetic effects
we expect that the csv contribution will be non - zero , but since a contribution to the gottfried sum rule needs top be odd under the combined operation of flavor and charge symmetry , which requires additional csv and flavor violating effects , the csv contribution to the gottfried sum rule will be suppressed below the natural scales for csv effects arising from quark mass differences and electromagnetic terms ( proportional to @xmath79 and @xmath67 , respectively ) .
another source for csv in the sea is the possibility of charge dependence of the quark gluon coupling constant . by dressing the quark - gluon vertex with a photon as in figs .
3a and 3b , the effective quark - gluon coupling for @xmath5 quarks is different than that for @xmath6 quarks @xcite , so that the probability of creating a @xmath80 pair becomes dependent on both the flavor of the pair created and the flavor of the quark that emitted the gluon . at low momentum scales , the effective gluon coupling to a quark of flavor @xmath74
is given by @xmath81 where @xmath82 is the strong coupling constant , @xmath83,@xmath84 is the quark charge , and @xmath85 with @xmath67 the electro - magnetic coupling . rewriting this ,
we get @xmath86 where now @xmath87 is the charge asymmetric piece of the strong coupling and @xmath88 . assuming that the sea quarks at low momentum scales are generated by the valence quarks
, we have @xmath89 from which we obtain , to first order in the csv terms , @xmath90 since the valence quarks provide a source of flavor asymmetry , there is a contribution to the gottfried sum rule . using the cteq distributions to calculate @xmath91 and @xmath92 mev , we find the csv contribution to the gottfried sum rule to be @xmath93 , which is far too small to explain the observed deviation .
similarly , there is a small excess of strange quarks in the proton over those in the neutron , according to @xmath94 having eliminated the possibility of a large csv contribution to the gottfried sum rule , we return briefly to the possibility that the largest part of the sum rule comes from violation of flavor symmetry , and that the violation of flavor symmetry leads to a csv contribution to the sum as well . combining eq . [ conjg ] with a flavor asymmetric sea yields @xmath95 using the cteq parton distributions .
our estimated csv contribution is still far too small to affect significantly the extraction of the flavor asymmetry .
we conclude that the prospects for a large contribution to the gottfried sum rule from sea quark csv are extremely slim , at least within the assumptions in our calculation of charge symmetry violating effects .
[ [ section-6 ] ] since we have been unable to produce a mechanism for generating a csv distribution whose integral is significantly different from zero , we are led to investigate the possibility of weak charge symmetry breaking , where the shapes of the sea distributions in the neutron and proton are different , but their normalizations remain the same . by analogy with the csv effects in the valence distributions
, we shall proceed by postulating that the 4 spectator partons can be assumed to belong to a tetra - quark , and that the mass of the recoiling tetra - quark can be assumed to be approximately constant with a value roughly equal to the sum of the constituent quark masses . since there is a greater variety of states available to the four parton spectator system ,
this assumption is much less certain than the analogous assumption in the valence case , but it will allow for an exploration of the magnitudes involved in the problem .
the basic argument proceeds as in the case of the valence distributions , by calculating the shift in the tetra - quark mass produced by the symmetry violating interactions and expanding to get @xmath96,\ ] ] where @xmath97 is the charge symmetric tetra - quark mass , and @xmath98 is the anti - quark distribution in the symmetric limit . for the light quark sea ,
we have used a tetra - quark mass of @xmath99 mev , roughly four times the constituent quark mass .
if , as many models assert , pionic effects dominate the sea at these low scales , this assumption will underestimate the size of the csv contribution to the light quark sea . for the systems containing strange quarks , we have assumed a constituent quark mass of 500 mev . in addition to this correction , there will be a correction due the change in the nucleon mass introduced by csv , which is calculated in the same fashion as the analogous contribution for the valence quarks.the values of the csv induced mass shifts for strange and nonstrange tetra - quarks are listed in table 1 , and the resulting csv sea distributions are shown in fig .
4 . just as in the case of the valence distributions , the large contribution to sea quark asymmetries at very small @xmath3 is a consequence of the singular form of the parton distributions in this region , and indicates the breakdown of the tetra - quark picture we ve used here . at larger @xmath3 , the relative size of @xmath100 over @xmath101 reflects the fact that the tetra - quark spectator has @xmath102 when a @xmath103 is struck , so that there is no tetra - quark mass shift to partially cancel the nucleon mass shift , as is the case when @xmath104 is struck . for the strange csv distributions ,
a similar cancellation occurs for the spectator tetra - quark when an @xmath105 quark is struck , but since this cancellation depends sensitively on the parameters chosen in table 1 , we conclude this result may be highly model dependent and therefore untrustworthy . in general , we do not find surprisingly large contributions for any of the sea quark distributions . |
if one puts aside philosophical questions created by the perceived counterintuitive features of the quantum mechanics that are frequently discussed in the literature on the quantum measurement problem ( an entry point to this literature is provided , e.g. , by the collections of papers @xcite or monographs @xcite ) it is easy to see that the physics of quantum measurements is fairly well understood by now at least on a qualitative level .
the process of quantum measurement is dynamic interaction between a microscopic quantum system and a macroscopic detector that establishes correlations between the states of these systems .
although it is impossible to give a universal definition of `` macroscopic '' in this context , a reasonable working definition is that the macroscopic detector is a system with negligible quantum fluctuations .
such systems are quite abundant in nature and can be described quantitatively within the framework of quantum mechanics .
since the description of quantum measurement process as an interaction between microscopic and macroscopic systems is quite broad , it is of interest to see whether there are any universal quantitative features of this process that are independent of specific physical realization of the detectors and the measured system .
the purpose of this work is to show that one such universal characteristics of quantum measurements can be obtained from the linear response theory .
this result was first pointed out in @xcite , and this work gives its more concise and accurate presentation .
an example of the area where a universal description of the quantum measurement process can be particularly useful is mesoscopic quantum dynamics of solid - state , in particular josephson - junction , qubits .
recent widespread interest in the development of solid - state qubits for quantum information processing brought with it the discussion of a large number of different detectors for measurement of quantum dynamics of these qubits .
these detectors include quantum point contacts @xcite , normal - metal @xcite and superconducting @xcite set transistors operated in different regimes , resonant tunneling structures @xcite , generic mesoscopic conductors @xcite , squids @xcite .
with such a variety of detectors ( the list of references above is by no means complete ) , a possibility of giving a quantitative description of some detector characteristics independently of its physical realization is obviously very helpful for understanding of the process of quantum measurement of mesoscopic qubits .
as was discussed in the introduction , the most essential feature of the model of the quantum measurement process considered in this work is its universality .
the model applies to measurement with an arbitrary detector which has to satisfy only some basic conditions .
schematics of this model is shown in fig . 1 and includes the measured system and the detector with the hamiltonians @xmath0 and @xmath1 , respectively .
the detector and the system are coupled through an interaction hamiltonian @xmath2 which almost without any loss of generality can be written as the product of some system `` coordinate '' @xmath3 and the detector `` force '' @xmath4 : @xmath5 the operator @xmath3 acts as the observable measured by the detector . in principle
, one could imagine a situation when a detector couples to some system through a combination of the several product terms like ( [ 1 ] ) , but the detector sensitive to several different dynamic variables of the system would be quite unusual .
( 5.,.9 ) ( .75,.05)=3.8 in the general structure of the hamiltonian of the system : @xmath6 is similar to the hamiltonians studied in the discussions of the dissipative quantum dynamics based on the `` system+reservoir '' models ( see , e.g. ,
@xcite ) if the detector is identified with the reservoir .
this analogy is not accidental , and many features of the dynamics of the measurement process are similar to the dissipative quantum dynamics of open systems .
for instance , as will also be discussed below , coupling to the detector leads to the dephasing of the measured system , similar to the reservoir - induced dephasing .
there are , however , important differencies between the two situations .
the main difference is that in the case of the dissipative quantum dynamics induced by a reservoir , changes in the reservoir caused by the system of interest are assumed to be unobservable , while the detector as a matter of principle should have an output observable @xmath7 that provides some information about its state . to represent the complete measurement process
, @xmath7 should behave as a classical variable , i.e. magnitude of quantum fluctuations in it should be much smaller than its classical component .
this condition implies that there is another difference between the detector and the dissipative reservoir
. the detector should be able to convert weak quantum input signal @xmath8 into the classical output signal @xmath7 that is large on the scale of quantum fluctuations .
such a transformation requires amplification and can not be achieved in an equilibrium system , i.e. , the detector , in contrast to the reservoir has to be driven out of the equilibrium . in the case of the dissipative reservoirs , it is well known that the linear response theory is a powerful tool that produces a general quantitative statement , fluctuation - dissipation theorem , about the reservoir - induced relaxation .
the theorem is independent of the specific microscopic model of the system .
the analogy between the measurement and the environment - induced relaxation leads naturally to the question of whether some general relations characterizing quantum measurements can be derived from the linear response theory . as we will see below ,
the answer to this question is yes , and the linear response theory applied to the measurement problem enables one to derive generic restriction on the energy sensitivity of a general linear quantum detector .
the obtained relation characterizes how close the detector is to being quantum - limited , i.e. , how close the detector - induced back - action dephasing in the measured system is to the fundamental limit set by the quantum mechanics .
while such a restriction on the energy sensitivity was known before in several specific detector models , linear response theory establishes it for an arbitrary detector weakly coupled to the measured system . for the linear response theory to be applicable to the measurement set - up shown in fig . 1 , and to provide some meaningful information , the detector has to satisfy certain conditions .
two most important are : * the detector is weakly coupled to the measured system so that its response to variations in this system is linear , * the detector is in the stationary state . since the detector can not be in equilibrium , the second item on this list generalizes the equilibrium condition of the standard linear - response theory . both of these conditions are not very restrictive , and are satisfied in many experimental situations .
linearity of the response is not sufficient to specify the detector properties completely , since the linear response coefficient can have an arbitrary frequency dependence , i.e. one needs to assume some spectral characteristics of the detector . here
we consider the simplest case when the response time of the detector is much shorter that the characteristic time of the dynamics of the measured system .
this means that the response coefficient @xmath9 is constant in the relevant frequency range .
although this assumption is less fundamental that the previous two , and in principle can be avoided , it simplifies the discussion considerably .
quantitatively , linearity of the detector response means that dynamics of the detector operators can be treated in the lowest order of the perturbation theory in the detector - system coupling @xmath2 . expanding the evolution operator @xmath10 with the total hamiltonian ( [ h ] )
upto the first order in @xmath2 we see that the detector output @xmath7 can be written as : @xmath11 x(\tau ) \ , .
\label{2}\ ] ] the first term @xmath12 in this expression can be interpreted as the output noise of the detector that exists even in the absence of the input signal , while the second term represents the detector s linear response to measured system . in zeroth order in coupling ,
the detector and the system are independent , so the averages over their states in eq .
( [ 2 ] ) can be taken separately .
tracing out the detector variables in ( [ 2 ] ) we obtain the regular expression for the linear response , extended to the situation when the `` driving force '' @xmath8 is a quantum operator . with the adopted assumption of instantaneous detector response : @xmath13 \rangle_d = \lambda \delta(\tau -0 ) \ , \label{3}\ ] ] eq .
( [ 2 ] ) reduces to : @xmath14 the infinitesimal shift in the argument of the @xmath15-function in ( [ 3 ] ) represents small but finite response time of the detector , and is needed to resolve the ambiguity in eq .
( [ 2 ] ) . in equation ( [ 3 ] ) , we used the assumption ( introduced above ) that the detector is in the stationary state and the @xmath16@xmath4 correlator depends therefore only on the difference of the time arguments , and @xmath17 denotes the average of the stationary detector density matrix .
since @xmath18 represents the output noise , we took @xmath19 . similarly ,
if the detector produces some non - vanishing average force @xmath20 , it is natural to include it in the system hamiltonian @xmath0 , and one can therefore always assume that @xmath21 .
dynamics of the coupling force @xmath4 generated by the detector is characterized then by the correlation function @xmath22 and physically can be viewed as `` back - action '' noise that dephases the measured system . following the assumption of the instantaneous response (
[ 3 ] ) we also take both the output noise and the back - action noise to be @xmath15-correlated : @xmath23 here we introduced spectral densities of the output ( @xmath24 ) and back - action ( @xmath25 ) noise that are constant in the relevant low frequency range . besides the infinitely small response time of the detector , eqs . (
[ 4 ] ) also imply that dynamics of both @xmath18 and @xmath26 is essentially classical , since the frequency - dependent part of the spectrum that corresponds to zero - point fluctuations gives negligible contribution to the correlators ( [ 4 ] ) .
qualitatively , measurement dynamics defined by the total hamiltonian ( [ h ] ) and assumptions ( [ 3 ] ) and ( [ 4 ] ) about the detector characteristics , includes two related processes .
one is the dephasing and relaxation / excitation of the measured system by the back - action force @xmath26 .
this part of the measurement is very similar to the dissipative quantum dynamics of a system weakly coupled to a reservoir .
the second process is acquisition of information about the state of the measured system reflected in the evolution of the detector output @xmath7 . on the quantitative level ,
measurement dynamics depends on the hamiltonian @xmath0 of the measured system and the operator structure of the measured observable @xmath3 .
probably the simplest situation which can be used to illustrate the two qualitative aspects of the measurement dynamics is the case of the stationary system , @xmath27 , with an observable @xmath3 that has several discrete eigenstates @xmath28 : @xmath29 with eigenvalues @xmath30 .
since under the assumption ( [ 4 ] ) the back - action force @xmath26 can be treated as a classical random force generated by the detector , one can write the time - evolution equation directly for the density matrix @xmath31 of the measured system . from the hamiltonian ( [ h ] ) , this equation in the basis of @xmath3-eigenstates is : @xmath32 solution of this equation averaged over different realizations of @xmath26 using eq .
( [ 4 ] ) gives : @xmath33 from this equation one can see that when the system is prepared in the initial state which is a superposition of the eigenstates of @xmath3 , i.e. @xmath34 has non - vanishing off - diagonal elements , interaction with the detector dephases the system and suppresses these off - diagonal elements .
the suppression does not happen only if the states @xmath35 and @xmath36 are degenerate with respect to @xmath3 : @xmath37 , so that the detector does not distinguish them .
equation ( [ 5 ] ) shows also that the rate @xmath38 of such a back - action dephasing is determined by the spectral density of the back - action force @xmath26 .
the dephasing ( [ 5 ] ) leads to diagonalization of the density matrix in the basis of the measured observable and coincides with the simplest instance of the environment - induced dephasing of an open quantum system . from the measurement perspective
, it describes the dynamic side of the `` wave - function collapse '' , since when the measurement is completed , the measured system should find itself in one of the eigenstates of the observable being measured , i.e. @xmath3 . the density matrix diagonal in the @xmath3 representation corresponds precisely to such a situation .
this interpretation of the back - action dephasing is strengthened by the relation between the dephasing rate @xmath38 and the rate of information acquisition - see , e.g. , @xcite .
indeed , from this perspective , in the example of the preceding paragraph , the detector has to distinguish different eigenstates of the observable @xmath3 and to provide information in which eigenstate the system finds itself .
according to the linear response relation ( [ l ] ) , the difference between two eigenvalues of @xmath3 , @xmath39 , translates into the difference @xmath40 of the dc values of the detector output @xmath41 .
the characteristic time @xmath42 on which this difference in @xmath41 can be distinguished in the presence of the output noise is determined by the low - frequency spectral density @xmath24 of this noise . averaging the @xmath15-correlated noise over the time interval @xmath43 leaves characteristic noise amplitude @xmath44 .
the characteristic time @xmath42 of acquisition of the information about the dc value of the detector output , and therefore , about the eigenstate of the measured observable , is determined by the condition that the noise amplitude is reduced at least to half the distance @xmath45 between the signal level .
this condition gives @xmath46 ^ 2\ , , \label{6}\ ] ] and shows that the relation between the back - action dephasing rate and the `` measurement time '' @xmath42 is independent of the eigenvalue difference @xmath39 , and depends only on the linear - response parameters of the detector : @xmath47 now the final step is to show that the linear response theory relates parameters @xmath9 , @xmath24 , @xmath25 , in such a way that the dephasing rate is fundamentally linked to the measurement time : @xmath48 to see this , we start by taking fourier transform of eq .
( [ 3 ] ) : @xmath49 expanding expression for the correlator @xmath50 in the basis of the energy eigenstates @xmath51 of the detector we obtain : @xmath52 where @xmath53 is the density of the detector energy states , and since the detector was assumed to be in the stationary state , its density matrix @xmath54 in the energy basis is diagonal . making use of the expression ( [ 11 ] ) and similar expressions for noise spectral densities , we can establish an inequality relating them . to obtain it
, we note that eq .
( [ 11 ] ) can be viewed from a somewhat artificial , but useful perspective as a scalar product of two functions of energy @xmath55 , latexmath:[$\langle \varepsilon -\hbar \omega probability @xmath54 are non - negative , the scalar product defined by eq .
( [ 11 ] ) satisfies all the usual requirements of the scalar product . using the standard notation for this product
, we can write ( [ 11 ] ) simply as @xmath57 spectral densities of the back - action and the output noise can be also expressed in terms of this scalar product .
the total spectral densities @xmath24 , @xmath25 are defined as @xmath58 with similar equations for @xmath59 . writing the correlators @xmath60 and @xmath61 in the same way as @xmath50 ( [ 11 ] ) in the basis of the detector energy eigenstates one sees immediately that they again can be expressed very simply in the language of the introduced scalar product : @xmath62 schwarz inequality for this scalar product gives , when applied to eqs .
( [ 12 ] ) and ( [ 13 ] ) : @xmath63 finally , combining eq .
( [ 9 ] ) , inequality ( [ 14 ] ) , similar inequality for components of the correlators at frequency @xmath64 , and our assumption that the noise spectral densities @xmath65 and @xmath66 are constant at low frequencies , we see that @xmath67 and therefore @xmath68^{1/2 } \ , , \label{15}\ ] ] where all spectral densities and the response coefficient are now taken at low frequencies .
this inequality is the main result of application of the linear response theory to the measurement problem .
one consequence of the inequality ( [ 15 ] ) is the proof of the relation ( [ 8 ] ) between the back - action dephasing rate and the measurement time .
it follows directly from eq .
( [ 15 ] ) that when the detector is `` symmetric '' , i.e. @xmath69 , so that there are no classical correlations between the back - action noise and the output noise , inequality ( [ 15 ] ) transforms eq .
( [ 7 ] ) directly into the ( [ 8 ] ) .
( the case of an asymmetric detector is discussed below . ) in this derivation of ( [ 8 ] ) , this inequality appears to be the consequence of some relation between the correlators appearing in the linear response theory .
from the perspective of the measurement problem it can be interpreted in a much broader sense .
inequality ( [ 8 ] ) shows that in general the dephasing of the measured system by a detector can be arbitrary strong without providing any information on the system .
however , acquisition of information about the state of the system creates some minimum dephasing that suppresses coherence between the eigenstates of the measured observable .
the fact that such a minimum exists is dictated by the basic principle of quantum mechanics which requires the successful measurement of an observable to localize the system in one of the eigenstates of this observable .
if the detector is such that it is causing only this minimum back - action dephasing dictated by quantum mechanics , it is typically referred to as `` quantum - limited '' or `` ideal '' detector .
interplay between information acquisition by the detector and back - action dephasing of the measured system has a simple form of inequality ( [ 8 ] ) only in the situation of measurement of a static system .
when the system hamiltonian @xmath0 is non - vanishing , this interplay is affected by the dynamics of the system and in general manifests itself less directly .
one studied example of this @xcite is the measurement of a _ two - state system_. coherent quantum oscillations in this system are transformed by measurement into classical oscillations of the detector output . in this case , the trade - off between the back - action dephasing and information acquisition is reflected in the height of the oscillations peak @xmath70 in the output spectrum relative to the output noise @xmath24 .
inequality ( [ 8 ] ) limits then the peak height : @xmath71 , with eaquality reached by measurement with a symmetric quantum limited detector .
another system which has been studied in this context is _ harmonic oscillator _
similarly to the two - state system , dynamics of the measurement process is reflected in this case in the frequency spectrum of the detector output , which contains the detector output noise and a peak at the oscillator frequency .
the trade - off between the detector back - action and information gain becomes relevant if one asks a question what is the minimum contribution of the detector noise to the spectrum .
qualitatively , there are two ways in which the detector noise contributes to the spectrum .
the detector output noise gives direct contribution to the spectrum , while the back - action noise contributes to the output spectrum indirectly , by inducing additional oscillations at the detector input that are transformed to the output by the detector response .
minimization of the total noise contribution ( reduced to the detector input and normalized to the zero - point spectrum of the oscillator in units of @xmath72 ) shows that the minimum total noise is given by the `` energy sensitivity '' @xmath73 : @xmath74^{1/2 } \ , .
\label{16}\ ] ] energy sensitivity has the dimension of action , and relation ( [ 15 ] ) between the parameters of the linear response theory translates into the following limitation on @xmath73 : @xmath75 inequality ( [ 17 ] ) can be expressed qualitatively as a statement that the detector , when measuring a harmonic oscillator , adds at least half an excitation quantum of this oscillator to the measured signal .
this inequality was obtained for the first time in a different physical context of linear amplification of electromagnetic radiation , where this interpretation has direct meaning
see , e.g. , the papers on quantum noise in linear amplifies reprinted in @xcite , also @xcite and references therein .
the magnitude of the signal in this context is described naturally by the number of photons in a mode and amplification is characterized by the dimensionless photon gain : the ration of these numbers in the output and input modes .
relation between such a linear amplifier and the linear detector considered in this work is established by the fact that in the limit of large gain , the input signal of the amplifier which in general can be quantum ( i.e. , contain only few photons ) is transformed into the output signal which contains number of photons much larger than one and is in this sense classical .
this means that the amplifier in this regime acts essentially as a detector , since after amplification of the signal to the classical level , it can be dealt with without any restrictions .
the last point to be made here concerns the case of an asymmetric detector .
such a detector is defined by the nonvanishing classical correlations between the output and back - action noises described by the real part of the spectral density @xmath76 .
comparison of eq .
( [ 7 ] ) for the trade - off between the measurement time and the back - action dephasing , and eq . (
[ 17 ] ) for the energy sensitivity shows that they behave differently in the case of non - vanishing @xmath77 .
while the energy sensitivity ( [ 17 ] ) can reach its quantum limit even in asymmetric detectors with @xmath78 , the @xmath79 product ( [ 7 ] ) in this situation can only be larger than its quantum limit @xmath80 .
the origin of this discrepancy is that in asymmetric detectors the output noise contains some information about the back - action noise which should be utilized to reach the optimum quantum - limited performance of the detector . in the simple treatment that lead to inequality ( [ 8 ] ) , this
information is lost , and as a result the dephasing is increased beyond the minimum required by the quantum mechanics . in the case of measurement of the harmonic oscillator , the noise minimization procedure implicitly makes use of this information , since the precise frequency where the noise is minimum depends on the correlation strength @xmath77 . in the case of information / dephasing trade - off , the information provided by the correlations contained in non - vanishing @xmath77 can also be used to approach the quantum limit of dephasing . conceptually ( and probably also practically ) the simplest way to use the information contained in the output noise of the detector to reduce the back - action dephasing is to apply the output noise with an appropriate transfer coefficient to the measured system together with the back - action noise .
although physically this procedure might require a rather complicated set - up that actually applies part of the detector output to the measured system , formally this is achieved by simply redefining the back - action force of the detector : @xmath81 it is straightforward to see that the rate @xmath82 of back - action dephasing created by the redefined force @xmath83 in measurement of a static system can indeed be equal to the quantum minimum of @xmath84 even for asymmetric detectors with @xmath85 , if the relation ( [ 15 ] ) between the linear response properties of this detector is satisfied as equality .
the same conclusion can be reached for measurement of the two - state systems @xcite .
this provides one more example of the fact that many physical characteristics of the measurement process can be understood conveniently as dynamics of information .
99 , ed . by j.a .
wheeler and w.h .
zurek ( princeton univ . press , 1983 ) . ,
ed . by p. meystre and m.o .
scully ( plenum , ny , 1983 ) . , ed . by a.i .
miller ( plenum , ny , 1990 ) .
v.b . braginsky and f.ya .
khalili , _ quantum measurement _ , ( cambridge , 1992 ) .
mensky , _ quantum measurement and decoherence : models and phenomenology _ ( kluwer , dordrecht , 2000 ) .
averin , in : _ `` exploring the quantum / classical frontier : recent advances in macroscopic quantum phenomena '' _ , ed . by j.r .
friedman and s. han , ( nova publishes , hauppauge , ny , 2002 ) , p. 441 ; cond - mat/0004364
. m. field , c.g .
smith , m. pepper , d.a .
ritchie , j.e.f .
frost , g.a.c .
jones , and d.g .
hasko , phys .
* 70 * , 1311 ( 1993 ) .
m. kataoka , c.j.b .
ford , g. faini , d. mailly , m.y .
simmons , d.r .
mace , c .- t .
liang , and d. a. ritchie , phys .
83 * , 160 ( 1999 ) .
d. sprinzak , e. buks , m. heiblum , and h. shtrikman , phys .
lett . * 84 * , 5820 ( 2000 ) .
gurvitz , phys .
b * 56 * , 15215 ( 1997 ) .
korotkov , phys .
b * 60 * , 5737 ( 1999 ) .
makhlin , g. schn , and a. shnirman , phys .
lett . * 85 * , 4578 ( 2000 ) .
averin , in : _ `` macroscopic quantum coherence and quantum computing '' _ ed . by d.v .
averin , b. ruggiero , and p. silvestrini , ( kluwer , 2001 ) , p. 399 ; cond - mat/0010052 .
m.h . devoret and r.j .
schoelkopf , nature * 406 * , 1039 ( 2000 ) .
g. johansson , a. kck , and g. wendin , phys .
lett . * 88 * , 046802 ( 2002 ) .
zorin , phys .
lett . * 76 * , 4408 ( 1996 ) .
a.a . clerk , s.m .
girvin , a.k .
nguyen , and a.d .
stone , phys .
lett . * 89 * , 176804 ( 2002 ) .
averin , fortschrit .
der physik * 48 * , 1055 ( 2000 ) .
s. pilgram and m. bttiker phys .
lett . * 89 * , 200401 ( 2002 ) . s .-
li , y. yu , y. zhang , w. qiu , s. han , and z. wang , phys .
lett . * 89 * , 098301 ( 2002 ) .
u. weiss , _ quantum dissipative systems _ , ( world scientific , 1999 ) .
korotkov , cond - mat/0209629 .
a.n . korotkov and d.v .
averin , phys .
b * 64 * , 165310 ( 2001 ) . l.n .
bulaevskii , m. hruska , g. ortiz , cond - mat/0212049 .
danilov , k.k .
likharev , and a.b .
zorin , ieee trans . magn . * 19 * , 572 ( 1983 ) .
caves , phys .
d * 26 * , 1817 ( 1982 ) .
a.a . clerk , s.m .
girvin , and a.d .
stone , cond - mat/0211001 . | linear response theory describes quantum measurement with an arbitrary detector weakly coupled to a measured system .
this description produces generic quantitative relation characterizing the detector that is analogous to the fluctuation - dissipation theorem for equilibrium systems .
the detector characteristic obtained in this way shows how efficient the trade - off is between the back - action dephasing and information acquisition by the detector . |
superconducting josephson junction devices and circuits are proving to be promising systems for quantum information processing @xcite . due to the great controllability of the qubits and microwaves in the superconducting circuit systems , single - qubit @xcite ,
two - qubit @xcite , and three - qubit @xcite operations have been experimentally demonstrated .
the next natural step is to develop robust , high - fidelity and scalable gates for larger scale quantum computation .
an essential prerequisite for quantum information processing and quantum computation is precise coherent control of quantum systems or quantum bits ( qubits ) . here , we focus our discussion on high - fidelity quantum control for superconducting flux qubit systems .
the flux qubits have an advantage over other types of superconducting qubits in the larger anharmonic energy level structure , i.e. the difference between adjacent transition frequencies is larger @xcite and thus less leakage to higher energy - level states . several schemes to implement local qubit operations and controllable couplings using microwaves for multi - flux - qubit systems have been proposed or / and realized .
these schemes may be categorized into two groups : ( i ) directly coupling between qubits @xcite and ( ii ) indirectly coupling through an intermediate coupler @xcite .
for example , a scheme to control the effective qubit - qubit interaction for directly couping flux qubits with fixed qubit transition frequencies and large detuning between the neighboring qubits using time - dependent magnetic fluxes ( or microwaves ) was proposed in ref . .
while this approach is advantageous due to the resonant nature of the coupling , its disadvantage is that at least one of the qubits must be biased away from the coherence optimal point @xcite ; this makes the qubit susceptible to low - frequency flux noise and results in a shorter coherence time . for the case of indirect coupling through an intermediate coupler , the coupling subcircuit or coupling nonlinear element can also be driven with microwaves @xcite .
these indirectly coupling schemes have the advantage of enabling the qubits to be operated at their optimal coherence points @xcite .
however , in addition to the added circuit complexity , the effective couplings between two qubits in the indirectly coupling schemes are generally smaller so the two - qubit gating time is about @xmath4 ns , at least one or two order(s ) of magnitude longer than that of directly coupling schemes .
schemes for microwave controllable coupling , which allow qubits to be operated at optimal bias points and without intermediate coupler subcircuit , exist @xcite .
but the effective qubit - qubit couplings in these schemes are usually even smaller than those of the indirectly coupling schemes .
smaller two - qubit coupling or equivalently longer two - qubit gating time will usually make the qubits suffer more decoherence effect during the gate operation , thus deteriorating the gate fidelity .
the fidelities of single - qubit and two - qubit quantum gate operations in a multi - qubit register for the present directly / indirectly coupling schemes via microwaves in the unitary case ( without considering decoherence effect ) even under the two - level qubit approximation are not perfect @xcite .
the microwave pulses for single - qubit gate operations are generally obtained and applied under the approximation that the qubit with large detuning with its neighboring qubits is effectively decoupled with its neighbors even though there exist fixed two - qubit couplings between them . but
this decoupling is only valid in the first order of a small parameter that is the ratio of the two - qubit coupling to the detuning . for two - qubit operations ,
the microwave pulses are commonly constructed and performed under the rotating wave approximation , the adiabatic approximation for the nonlinear coupler , or / and the approximation of neglecting other small residual two - qubit interactions .
the qubit - decoupled approximation becomes excellent when the effective qubit - qubit coupling is very small ; however , a large qubit - qubit interaction is favorable for two - qubit operations .
therefore , reaching a delicate balance between these two situations results in , for instance , single - qubit gate errors greater than @xmath5 for practical experimental parameters even in the ideal , unitary case of simple two - level approximation for each flux qubit .
quantum optimal control theory is a powerful tool that provides a variational framework for finding optimal control field profiles or sequences by maximizing a desired physical objective ( or minimizing a physical cost function ) within certain constraints @xcite . in this paper , we apply the quantum optimal control theory to find the control pulse sequences of externally applied ac magnetic fluxes ( fields ) to implement fast and high - fidelity one - qubit and two - qubit gates on superconducting flux qubits .
there have been optimal control studies of quantum gates for superconducting qubits , but all focusing on the gate operations of cooper - pair - box charge qubit systems @xcite .
our investigation here is , to our knowledge , the first optimal control study for the flux qubit systems .
the flux qubits we consider have fixed direct qubit - qubit couplings and fixed transition frequencies but with large detuning between neighboring qubits , which ensures the qubits are effectively decoupled ( to the first order ) in the absence of time - dependent control fields or signals .
our optimal control scheme requires no additional bias or control lines beyond those used for the manipulation of individual qubits . in the absence of the time - dependent magnetic fluxes ,
the flux qubits in our scheme are biased at the coherence optimal points @xcite to reduce dephasing due to flux noise , i.e.,to have a longer coherence time .
furthermore , our optimal control scheme takes into account two - qubit interaction , qubit detuning and other time - dependent magnetic field induced residual single - qubit and two - qubit interactions when performing single - qubit and two - qubit gates .
thus the gate fidelities ( errors ) in the unitary ( closed - system ) case can be as high ( low ) as one wishes , limited only by the accuracy of the approximated two - level qubit hamiltonian used and by the machine precision of the computation . besides , the two - qubit cnot gate operations of our scheme is about several folds to two orders of magnitude faster than the directly or / and indirectly microwave controllable coupling schemes @xcite . by considering leakage to higher - energy - level states , the gate errors for single - qubit gates using the optimal control pulse sequences obtained are in the order of @xmath6 and are in the order of @xmath7 for two - qubit cnot gates
to take the effect of qubit decoherence into account , we model the qubit dynamics by a quantum master equation with experimentally available relaxation and dephasing rates .
the gate errors in the presence of decoheence by considering the master equation are still in the order of @xmath7 for single - qubit gates and in the order of @xmath8 for two - qubit cnot gates .
these gate errors are still below the error threshold @xmath9 ( @xmath5 in @xcite ; @xmath10 if surface code error correction is used @xcite ) required for fault - tolerant quantum computation .
the paper is organized as follows .
we first describe in sec .
[ sec : hamiltonian ] the hamiltonian of two inductively coupled flux qubits with each individual qubit controlled by a time - dependent magnetic flux ( field ) .
the reduced hamiltonian at the optimal bias point expressed in terms of the two - level qubit basis states is then obtained . in sec .
[ sec : qoct ] , a brief description about the quantum optimal control theory for performing state - independent quantum gate operations is presented .
the control field pulse sequences obtained by the optimal control theory and their corresponding gate errors and state evolutions are presented in sec .
[ sec : ressults ] .
it is also shown that our optimal control scheme is notably robust against leakage to states outside the computational basis state space .
the effect of qubit decoherence on gate errors is also discussed . a short conclusion with discussions of how to implement the optimal control pulses experimentally is given in sec . [ sec : conclusion ] .
( color online ) schematic illustration of two inductively coupled superconducting flux qubits .
the mutual inductance is @xmath11 and there are three junctions in each qubit loop .
the external static and time - dependent magnetic fluxes through the @xmath12th qubit are denoted as @xmath13 and @xmath14 , respectively .
, scaledwidth=50.0% ] the system we consider here is two flux qubits right next to each other and coupled by their mutual inductance @xmath11 @xcite as schematically illustrated in fig . [ fig : coupled_flux_qubits ] .
each qubit loop contains three josephson junctions @xcite , and one of them has an area @xmath15 times smaller than that of the two identical junctions .
the larger josephson junctions in the @xmath12th qubit loop have the josephson energy @xmath16 and the corresponding critical current @xmath17 , where @xmath18 is the flux quantum with the planck constant @xmath19 and the elementary charge @xmath20 .
the capacitances in the @xmath12th qubit loop satisfy the conditions @xmath21 and @xmath22 , where @xmath23 are the capacitance of the larger josephson junctions .
the coupled flux qubit system considered here is exactly the same as that discussed in ref .
fluxoid quantization around each qubit loop imposes a constraint on the phase drop across the three junctions @xcite : @xmath24 where @xmath25 is the gauge - invariant phase of each junction , @xmath26 and @xmath27 are , respectively , the static ( dc ) and time - dependent magnetic fluxes applied through the @xmath12th qubit . here
the self - inductance is considered negligible as compared to the josephson inductance .
the requirement for our coupled - qubit system is to have suitable mutual inductance for qubits placed next to each other as illustrated in fig .
[ fig : coupled_flux_qubits ] and independent flux lines to enable control over each individual qubit with external dc bias magnetic fluxes and time - dependent control magnetic fluxes .
the qubit - loop size for three - junction flux qubits could range experimentally from submicrometer to micrometer .
as long as the design of the qubit loops fulfills the above requirement , the actual size of each qubit loop or the distance between the qubits can have some flexibility .
the total hamiltonian of the two coupled three - junction flux qubits reads @xcite : @xmath28\right\ } , \label{eq : htot}\end{aligned}\ ] ] with the redefined phases @xmath29 , @xmath30 , the reduced dc bias magnetic flux @xmath31 , and the reduced time - dependent control magnetic flux @xmath32 . the effective masses are @xmath33 and @xmath34 , which correspond to the effective momenta @xmath35 and @xmath36 .
the persistence current in the @xmath12th qubit loop is @xmath37 .
\label{eq : i}\end{aligned}\ ] ] here the self - interaction terms of the external time - dependent control magnetic fields @xmath38^{2}$ ] is neglected since they vary only the global phase of the state vector and do not affect the control of the coupled flux qubits . to simulate the dynamics of the coupled flux - qubit system with hamiltonian eq .
( [ eq : htot ] ) , the parameter values of the mutual conductance @xmath11 as well as the josephson energies of the junctions @xmath16 , the capacitance of the larger josephson junctions @xmath39 and the area ratio @xmath15 of the smaller junction to one of the larger junctions in each qubit loop are required .
these parameter values can be obtained through the experimental measurement and characterization of the system .
the actual values of these parameters we choose for our simulations are given in sec .
[ sec : ressults ] .
they are all experimentally available or realistic values
. we will keep the time - dependent control amplitudes small such that the reduced time - dependent magnetic flux satisfies @xmath40 .
small time - dependent control amplitude allows the approximation of @xmath41\sim2\pi f_{c}^{\left(l\right)}(t)$ ] and @xmath42\sim 1 $ ] when expanding the last term of the hamiltonian of eq .
( [ eq : htot ] ) . as a result
, the time - dependent hamiltonian becomes linear in @xmath43 or @xmath44 , which can be readily incorporated with the krotov quantum optimal control method that we will employ later .
this weak - amplitude approximation of the control fields also keeps the qubit not deviating much from the dc bias point that is set to be the optimal coherence point in our case .
moreover , this weak - amplitude approximation helps reduce unwanted possible excitations to the higher - energy - level states outside the computational state space when we make two - level ( qubit ) approximation discussed later . making this weak - amplitude approximation
, we obtain from eq .
( [ eq : htot ] ) the hamiltonian similar to that in ref . @xcite as @xmath45 the first term @xmath46 is the single - qubit hamiltonian and reads : @xmath47.\label{eq : hamiltonian03}\end{aligned}\ ] ] the hamiltonian @xmath48 plays the role of a driving hamiltonian representing the interaction between the @xmath12th qubit and its time - dependent magnetic field .
it takes the form @xmath49 . \label{driving_ham}\ ] ] the last three terms in eq .
( [ eq : hamiltonian02 ] ) come from the inductive coupling between the two flux qubits .
the hamiltonian @xmath50 describes the qubit - qubit interaction controlled by one of the time - dependent magnetic flux [ @xmath51 or @xmath52 and is written as @xmath53 where @xmath54 \label{loop_current}\end{aligned}\ ] ] is the loop current of the @xmath12th qubit when no time - dependent magnetic fluxes are applied [ cf . , eq .
( [ eq : i ] ) ] .
the qubit - qubit interaction @xmath55 controlled by two simultaneously applied time - dependent magnetic fluxes through , respectively , the two qubits is @xmath56 we still keep this term , although it is much smaller than the term controlled by a single time - dependent magnetic flux .
the hamiltonian @xmath57 describes an always - on interaction between the two flux qubits without time - dependent magnetic fluxes being applied , where @xmath58 is defined in eq .
( [ loop_current ] ) .
the hamiltonian , eq .
( [ eq : hamiltonian02 ] ) , in the two - qubit computational state basis @xmath59 , where @xmath60 and @xmath61 are the lowest two energy - level states of @xmath46 of eq .
( [ eq : hamiltonian03 ] ) , becomes @xmath62 \nonumber \\ & & -\sum_{l\neq m=1}^{2}\left[\chi_{1}^{\left(lm\right)}(t)\sigma_{z}^{\left(l\right)}+\chi_{2}^{\left(lm\right)}(t)\sigma_{x}^{\left(l\right)}\right]\nonumber \\ & & -\sum_{l\neq m=1}^{2}\left[\xi_{11}^{\left(lm\right)}(t)\sigma_{z}^{\left(l\right)}\sigma_{z}^{\left(m\right)}+\xi_{22}^{\left(lm\right)}(t)\sigma_{x}^{\left(l\right)}\sigma_{x}^{\left(m\right)}\right .
\nonumber \\ & & \qquad \qquad \left .
+ \xi_{12}^{\left(lm\right)}(t)\sigma_{z}^{\left(l\right)}\sigma_{x}^{\left(m\right)}+\xi_{21}^{\left(lm\right)}(t)\sigma_{x}^{\left(l\right)}\sigma_{z}^{\left(m\right)}\right]\nonumber \\ & & + \theta_{11}(t)\sigma_{z}^{\left(1\right)}\sigma_{z}^{\left(2\right)}+\theta_{22}(t)\sigma_{x}^{\left(1\right)}\sigma_{x}^{\left(2\right)}+\theta_{12}(t)\sigma_{z}^{\left(1\right)}\sigma_{x}^{\left(2\right)}+\theta_{21}(t)\sigma_{x}^{\left(1\right)}\sigma_{z}^{\left(2\right)}\nonumber \\ & & + \lambda_{11}\sigma_{z}^{\left(1\right)}\sigma_{z}^{\left(2\right)}+\lambda_{22}\sigma_{x}^{\left(1\right)}\sigma_{x}^{\left(2\right)}+\lambda_{12}\sigma_{z}^{\left(1\right)}\sigma_{x}^{\left(2\right)}+\lambda_{21}\sigma_{x}^{\left(1\right)}\sigma_{z}^{\left(2\right)},\label{eq : hamiltonian04}\end{aligned}\ ] ] where the single - qubit transition frequencies @xmath63 and the driving hamiltonian amplitudes @xmath64 and @xmath65 are @xmath66 @xmath67 @xmath68 respectively , the other time - dependent controllable interaction strengths , @xmath69 , @xmath70 , @xmath71 and @xmath72 , coming from qubit - qubit inductive interaction are @xmath73 @xmath74 @xmath75 @xmath76 respectively , and the static fixed qubit - qubit interaction strengths @xmath77 are @xmath78 the relevant parameters in eqs .
( [ eq : param01])-([eq : param02 ] ) are @xmath79 which is the mutual inductive energy with respect to the critical current @xmath80 of the larger josephson junctions in each qubit loop , @xmath81 @xmath82
@xmath83 , \label{omega1}\end{aligned}\ ] ] @xmath84 and @xmath85 , \label{delta}\end{aligned}\ ] ] where the operator @xmath86 is defined as @xmath87 the hamiltonian eq .
( [ eq : hamiltonian04 ] ) with parameters defined in eqs .
( [ transition_freq])-([upsilon ] ) is valid for weak time - dependent magnetic fluxes ( fields ) and arbitrary static bias magnetic fluxes .
if we take the time dependence of the control magnetic flux to be sinusoidal , i.e. , @xmath88 , and making the relevant rotating wave approximation , we can revert to the hamiltonian of ref .
different from ref .
@xcite , our control scheme , however , does not require qubits being biased away from the optimal coherence points .
in contrast , we set the reduced static or dc bias magnetic fluxes to be @xmath89 so that the qubits are at the optimal coherence points and are thus insensitive to low - frequency flux noise in the first order . at the optimal coherence points of @xmath89 , hamiltonian @xmath46 of eq .
( [ eq : hamiltonian03 ] ) is invariant under the parity transformation of @xmath90 , i.e. , @xmath91 .
thus its eigenstates have definite parities and the lowest two energy - level eigenstates have the opposite parities @xcite . since the hamiltonian of eq .
( [ driving_ham ] ) and the loop currents of eq .
( [ loop_current ] ) are odd functions of @xmath90 and the operators @xmath92 from eq .
( [ upsilon ] ) are even functions of @xmath90 at @xmath89 , the parameters @xmath64 , @xmath93 and @xmath94 defined in eqs .
( [ kappa_1 ] ) , ( [ lambda1 ] ) and ( [ omega2 ] ) , respectively , vanish due to the parity symmetry consideration . as a consequence ,
most parameters in eqs .
( [ eq : param01])-([eq : param02 ] ) vanish , and eq ( [ eq : hamiltonian04 ] ) at the optimal bias points simplifies to @xmath95+\lambda_{22}\sigma_{x}^{\left(1\right)}\sigma_{x}^{\left(2\right)}\nonumber \\ & & -\sum_{l\neq m=1}^{2}\left[\chi_{1}^{\left(lm\right)}(t)\sigma_{z}^{\left(l\right)}+\xi_{12}^{\left(lm\right)}(t)\sigma_{z}^{\left(l\right)}\sigma_{x}^{\left(m\right)}\right]+\theta_{11}(t)\sigma_{z}^{\left(1\right)}\sigma_{z}^{\left(2\right)}.\label{eq : hamiltonian05}\end{aligned}\ ] ] we will use eq .
( [ eq : hamiltonian05 ] ) combining with quantum optimal control theory to find control magnetic field pulses @xmath43 for the implementation of high - fidelity one- and two - qubit gates .
quantum optimal control theory enables us to realize accurate state - independent quantum gates by selecting optimal pulse shapes ( arbitrarily shaped pulses and duration ; or continuous dynamical modulation ) for the external control within experimental capabilities @xcite . to perform state - independent optimal control , the equation of motion for the time evolution operator ( or propagator ) @xmath96
is needed : @xmath97u\left(t\right ) , \label{eq : sch1}\ ] ] where @xmath98 is the system hamiltonian and @xmath99 is a time - dependent control field .
we choose the trace distance between the desired target gate operation @xmath100 and the actual ( could be nonunitary ) propagator @xmath101 at the final operation time @xmath102 to characterize the gate error : @xmath103^{\dagger}\left[o - u(t)\right]\right\ } \label{error1 } , % \\ % & = & \frac{1}{2n } % \left\langle \left\langle % u\left(t\right)-o\left|u\left(t\right)-o\right\rangle % \right\rangle \right .. % \label{error2}\end{aligned}\ ] ] where @xmath104 is the dimension of the matrix @xmath96 . if the actual time evolution operator @xmath101 is equal to the target gate operation @xmath100 , then @xmath105 with @xmath106 denoting the gate fidelity .
sometimes , one is also interested in calculating the errors of the operations ( operators or matrices ) in a subspace of a unitary matrix .
when projection is made to the subspace , the projected matrix in the subspace is in general no longer unitary .
for instance , when the evolution operator of the multi - level coupled flux - qubit system @xmath101 for the multi - level hamiltonian , eq .
( [ eq : htot ] ) , is projected into the approximated two - level qubit computational state subspace , the projected time evolution operator is no longer unitary . in this case , eq .
( [ error1 ] ) with slight modification [ see eq .
( [ errorp ] ) ] is still an appropriate measure for calculating the gate error . in realistic control problems ,
it is desirable that the optimal control sequence can provide highest quality ( fidelity ) with minimum energy consumption .
therefore , we define the cost function for our optimal control problem as @xmath107^{2}dt , \label{eq : costfun02}\ ] ] where @xmath108 is defined in eq .
( [ error1 ] ) , @xmath109 , a positive shape function , and @xmath110 , a weight , can be adjusted and chosen empirically @xcite . here
, the reference field @xmath111 is chosen to be the control sequence in the previous iteration , i.e. , @xmath112 such that the control field energy constraint in the cost function @xmath113 of eq .
( [ eq : costfun02 ] ) has the physical interpretation that the change of the control pulse energy in each iteration is limited @xcite . when the iterative procedure approaches the optimal solution , the change in the control field is minimal or vanishing .
therefore , this choice of the reference field @xmath111 ensures that the iterative method is found to reduce the total objective @xmath113 of eq .
( [ eq : costfun02 ] ) by reducing the gate error @xmath108 rather than the total control pulse energy .
with eqs .
( [ eq : sch1 ] ) and ( [ eq : costfun02 ] ) and the initial propagator @xmath114 , where @xmath115 is the identity operator , one can then investigate the state - independent quantum gate optimal control problem .
we will use the krotov iterative method for the quantum gate optimal control @xcite .
the krotov method has several appealing advantages @xcite over the standard gradient optimization methods : ( a ) monotonic increase of the objective with iteration number , ( b ) no requirement for a line search , and ( c ) macrosteps at each iteration .
the optimal algorithm following the krotov method @xcite can be found in refs . .
we use the reduced hamiltonian , eq .
( [ eq : hamiltonian05 ] ) , with two lowest energy - level states for each qubit to obtain control sequences for single - qubit @xmath0 and @xmath1 gate operations and two - qubit cnot gate operations by optimizing the cost function @xmath113 , eq .
( [ eq : costfun02 ] ) , using the krotov iterative method .
simulations on the two - qubit hamiltonian , eq .
( [ eq : htot ] ) , without making the weak - amplitude approximation and with the lowest five energy states for each qubit , using the optimal control field sequences found via eq .
( [ eq : hamiltonian05 ] ) will also be performed for error comparison .
this allows us to test how well the optimal control pulses obtained by the weak - amplitude approximation perform , and to determine how severe the leakage to the higher - energy - level states is .
the target gate operation @xmath100 is defined in the two - qubit computational state basis . the evolution operator or propagator @xmath101 results from eq .
( [ eq : htot ] ) using the optimal control pulse sequences @xmath116 and @xmath117 obtained from eq .
( [ eq : hamiltonian05 ] ) is in a larger multi - level - state space .
we thus define a projection operator @xmath118 to project @xmath101 onto the subspace subtended by the computational basis states , and compare the target operator @xmath100 with the resultant propagator @xmath119 which is in general no longer unitary .
the error in this case is defined from eq .
( [ error1 ] ) as @xmath120^{\dagger}[o - pu(t)]\right\ } , % \eta_p=(1/32){\rm tr}\left\{[o - pu(t)]^{\dagger}[o - pu(t)]\right\}. % \left\langle \left\langle % pu\left(t\right)-o\left|\right.pu\left(t\right)-o\right\rangle % \right\rangle .
\label{errorp}\ ] ] where @xmath104 is the dimension of the @xmath100 matrix or @xmath121 matrix .
we choose the parameters in the two - qubit hamiltonian with experimentally available or realistic values @xcite .
the josephson energy of the larger junctions in the qubit loops is chosen to be @xmath122 ghz and the charging energy @xmath123 is chosen to be @xmath124 .
the capacitance of the smaller josephson junction is @xmath125 with the ratio @xmath126 . to fit the value of the coupling strength @xmath127 reported in the experiment of ref .
, we take the mutual inductance @xmath128 ph and obtain @xmath129
. then other parameters are determined by the calculations through the hamiltonian , eq .
( [ eq : htot ] ) .
we then obtain the qubit transition energies @xmath130 and @xmath131 , and the single - qubit driving amplitudes @xmath132 and @xmath133 .
the time - dependent coefficients of the last three coupling terms in eq .
( [ eq : hamiltonian05 ] ) are also obtained to be @xmath134 @xmath135 and @xmath136 .
we note here that the dynamics and also the optimization results that will be presented later depend on the values of the microscopic parameters .
these parameter values can be obtained through the experimental measurement and characterization of the system but always come with some error bars .
thus if the actual system is , for example , slightly away from the assumed control point ( a set of values for the system parameters ) chosen for the quantum optimal control theory , then the optimal control pulse sequences sent to the experiment may result in gate fidelities not close to optimal .
this issue of imprecise knowledge of the system parameters seems to hinder the practical use and experimental applicability of the quantum optimal control theory .
fortunately , there may be ways around , for example , using the adaptive model - free closed - loop feedback control method .
we postpone the discussions regarding this until sec .
[ sec : conclusion ] . in the following , we first present our optimal control results for single - qubit and two - qubit gates in the unitary case together with brief descriptions of corresponding conventional approaches for comparison .
the results considering the effect of leakage to higher - energy - level states and the effect of qubit decoherence are presented subsequently .
( color online ) gate errors @xmath108 as a function of the number of iterations .
the gate errors calculated using hamiltonian eq .
( [ eq : hamiltonian06 ] ) decrease as the number of iterations increases .
the stopping iteration criteria of error is set to be @xmath137 .
the single - qubit gates of the first qubit , the single - qubit gates of the second qubit and the two - qubit cnot gate are achieved at @xmath138 ns , @xmath139 ns and @xmath140 ns , respectively .
, scaledwidth=50.0% ] _ conventional approach - _ in most of the studies for the directly coupling schemes , the hamiltonian of two inductively coupled flux qubits at the optimal bias point contains only the first line of eq .
( [ eq : hamiltonian05 ] ) : @xmath141+\lambda_{22}\sigma_{x}^{\left(1\right)}\sigma_{x}^{\left(2\right)}. \label{eq : hamiltonian06}\end{aligned}\ ] ] this is because the coefficients of the last three time - dependent terms of eq .
( [ eq : hamiltonian05 ] ) are several orders of magnitude smaller than the qubit transition frequencies @xmath142 , the single - qubit driving amplitudes @xmath143 and the fixed static qubit - qubit interaction strength @xmath144 .
thus neglecting these coefficients does not introduce appreciable error .
furthermore , since the strength @xmath144 of the off - diagonal coupling @xmath145 in the presence of large qubit detuning @xmath146 affects the qubit dynamics only in second order in the parameter @xmath147 , the constant coupling @xmath144 term is often neglected at the optimal bias point provided that @xmath147 is rather small . in this case
, the qubits are regarded to be effectively decoupled .
thus the single - qubit gate , say , @xmath148 gate ( rotation of angle @xmath149 around @xmath150-axis on qubit 1 ) , can be achieved by simply driving the magnetic flux through the first qubit with microwave resonance with the transition frequency of the first qubit , i.e. , @xmath151 , while keeping the time - dependent flux through the second qubit off , i.e. , @xmath152 .
when the driving field strength @xmath153 is much smaller than the transition frequencies @xmath63 and qubit detuning @xmath146 in which the rotating - wave approximation can be made , eq . ( [ eq : hamiltonian06 ] ) in the frame rotating with the driving frequency @xmath154 becomes @xmath155 performing single - qubit @xmath156-gate on qubit 1 will also demand qubit 2 to return to its original state up to a global phase at the end of the @xmath156 operation , i.e. , an identity gate on qubit 2 . in this rotating frame
, qubit 2 takes @xmath157 to complete an identity gate up to a global phase with @xmath158 being an integer number . taking this time to be the time to complete a @xmath149-pulse on qubit 1 yields the single - qubit driving strength or the rabi frequency to be @xmath159 . by requiring the reduced time - dependent magnetic flux @xmath160 such that the weak - amplitude approximation to obtain eq .
( [ eq : hamiltonian02 ] ) is valid , one obtains the minimum integer number @xmath158 to be @xmath161 .
this then leads to @xmath162 and the operation time of @xmath156 gate @xmath163 ns . plugging these numbers into eq .
( [ eq : hamiltonian05 ] ) and eq .
( [ eq : htot ] ) to simulate an @xmath156 gate gives an error of about @xmath164 .
this result indicates that the dominant source of error is mainly due to the constant qubit - qubit interaction term of @xmath144 and does not come from the leakage to higher - energy - level states @xcite .
indeed , for the experimental parameters used here , the ratio of @xmath165 , and taking the second order correction gives an estimated error also in the order of @xmath5 .
this is about the best one can do in a unitary case if neglecting the constant qubit - qubit interaction @xmath144 term . how to include the qubit - qubit interaction term and perform a much more accurate single - qubit gate is not intuitively obvious .
_ optimal control approach - _ here , we apply the krotov iterative method to obtain the optimal control sequences for high - fidelity ( low - error ) single - qubit and two - qubit gates taking all the static and time - dependent single - qubit and two - qubit terms or interactions into account .
usually , stronger strengths of the control fields result in shorter gate operation times . in our case ,
the gate operation times are , however , chosen such that the reduced time - dependent magnetic fluxes satisfying the weak - amplitude approximation of @xmath40 .
we find that the high - fidelity single - qubit gates on qubit 1 can be achieved at a gating time of @xmath138 ns , and the gating time is @xmath139 ns for qubit 2 .
the two - qubit gates will be discussed in the next subsection . as shown in fig .
[ fig : error ] , the errors @xmath108 of @xmath166 , @xmath148 , @xmath167 and @xmath168 gates calculated using the reduced hamiltonian , eq .
( [ eq : hamiltonian05 ] ) , decrease with the number of iterations .
we note here that for simplicity , we let the positive shape function @xmath169 in eq .
( [ eq : costfun02 ] ) be a constant .
in addition , the time - dependent control field @xmath99 in eq .
( [ eq : costfun02 ] ) in our case becomes the reduced time - dependent magnetic fluxes @xmath170 that are dimensionless and required to be smaller than @xmath5 , and thus in our calculations we choose @xmath171 ( ghz)@xmath172 in order to make sure that the iterations converge to the optimal results monotonically . in principle , the gate error for a closed , unitary system can , via the optimal control theory , be as low as one wishes , limited only by the accuracy of the two - level - approximation hamiltonian , eq .
( [ eq : hamiltonian05 ] ) , and by the machine precision of the computation . considering the accuracy of the two - level - approximation hamiltonian used ,
we set the stopping criteria of error to be @xmath173 , and thus the iterations are terminated when error @xmath137 . a typical set of the control sequences of
the reduced time - dependent magnetic fluxes @xmath174 and @xmath175 for the high - fidelity @xmath156 gate is shown in figs .
[ fig : x1 ] and [ fig : x1 ] . _ conventional approach - _ many schemes using microwaves have been put forward @xcite to realize an effective controllable qubit - qubit interaction for hamiltonian ( [ eq : hamiltonian06 ] ) or ( [ eq : hamiltonian05 ] ) . however , in most of the schemes , two neighboring qubits are considered effectively decoupled ( as described earlier ) when no time - dependent microwave field is applied .
this approximation limits the maximum fidelity ( or minimum error ) these schemes can actually achieve .
one of the common goals for gate control is to implement a two - qubit cnot gate , an important universal gate for quantum computation . in all of the schemes , an entangled gate natural to their couplings
is implemented , and thus to realize a cnot gate , an additional sequence of single - qubit rotations is needed .
for example , if an iswap gate is conveniently implemented in a particular scheme , the cnot gate can be realized by the following sequence cnot@xmath176[r_z(-\pi/2)\otimes r_z(\pi/2)][{\rm iswap}][r_x(\pi/2)\otimes i][{\rm iswap}][i \otimes r_z(\pi/2)]$ ] .
typical two - qubit gate operation times in these schemes are in the order of @xmath177 to @xmath178 ns . however , decomposing a cnot gate or a general gate operation into several single - qubit and some entangled two - qubit operations in series makes its operation time normally longer than that of completing the operation in one single run of pulses .
moreover , the overall gate error may also become larger as the gate errors of the decomposed gates will add up or accumulate .
_ optimal control approach - _ in contrast , the optimal control method has great advantages of enabling the implementation of a cnot gate or other general quantum gates in one single run of pulse or in one single pulse sequence by simply setting the target operation to be the cnot gate or the general quantum gate one wishes to implement . requiring the reduced time - dependent magnetic flux @xmath40
, we then set the cnot gate operation time @xmath179 ns and use the reduced hamiltonian ( [ eq : hamiltonian05 ] ) to find the optimal control sequence .
this cnot gating time is about several folds to two orders of magnitude shorter than that of the directly or / and indirectly microwave controllable coupling schemes @xcite .
this is because single - qubit and two - qubit operations can be performed in parallel simultaneously ( rather than in series ) in one single run of the optimal control pulses .
one can see from fig .
[ fig : error ] that the cnot gate error drops quickly with the number of iterations . here , qubit 1 is the control qubit and qubit 2 is the target qubit .
the optimal control sequences obtained when the iteration terminates at the point when error @xmath137 are shown in figs .
[ fig : cnot ] and [ fig : cnot ] . to investigate the gate error due to the leakage to higher energy - level states outside the computational basis - state space
, we apply the control sequences of the single - qubit and two - qubit gates obtained from the optimization of the reduced two - level - approximation hamiltonian ( [ eq : hamiltonian05 ] ) to the multi - level hamiltonian ( [ eq : htot ] ) and calculate the errors @xmath180 of eq .
( [ errorp ] ) in the projected two - qubit computational state subspace .
the single - qubit gate errors are found to be @xmath181 , @xmath182 , @xmath183 , and @xmath184 for the @xmath166 , @xmath148 , @xmath167 and @xmath168 gates , respectively .
these results indicate that the excitations to higher energy - level states increase slightly the gate errors ; however , the gate errors in the order of @xmath6 are still much smaller than @xmath5 of those conventional approaches that apply directly a microwave @xmath149 pulse by neglecting the qubit - qubit interaction .
we also simulate the @xmath156 gate numerically with the multi - level hamiltonian ( [ eq : htot ] ) using the optimal control sequences obtained from the reduced hamiltonian ( [ eq : hamiltonian05 ] ) for initial four different qubit computational basis states @xmath185 , @xmath186 , @xmath187 and @xmath188 .
the time evolutions of the state probabilities of the @xmath156 gate are shown in figs .
[ fig : x1]-[fig : x1 ] . as expected , after the @xmath156 operation , @xmath189 , @xmath190 , @xmath191 , and @xmath192 .
it seems that there is an approximate symmetry under time - reversal between figs .
[ fig : x1 ] and [ fig : x1 ] , and between [ fig : x1 ] and [ fig : x1 ] .
this is mainly due to the fact that the control pulse sequences for the @xmath156 gate shown in figs .
[ fig : x1 ] and [ fig : x1 ] are seemingly symmetric with respect to the midpoint of the operation time .
thus the time evolutions of the state probabilities , when the first qubit making the transition between @xmath193 and @xmath194 with the second qubit returning back to the same state , appear to have such a time - reversal symmetry . in our case ,
the pulse sequences are obtained when the gate error is smaller than the error threshold of @xmath137 for a certain initial guess of the pulse sequence .
the truly global optimal pulses should , however , correspond to the minimum value of the cost function obtained by trying all possible initial guesses of the pulse shapes .
thus other initial guesses can generate slightly different @xmath156-gate pulse sequences satisfying still the same error threshold requirement but with less symmetry .
in fact , the pulse sequence of @xmath175 for the @xmath195 gate we obtain ( not shown here ) does not have an apparently symmetric shape with respect to the midpoint of the operation time as compared to @xmath174 of the @xmath156 gate in figs .
[ fig : x1 ] . as a result ,
the time evolutions of the state probability between @xmath196 ( latexmath:[$|ee\rangle \rightarrow show prominent time reversal symmetry as their corresponding counterparts for the @xmath156 gate .
there is , however , indeed an approximate symmetry in time evolutions of the state probabilities in figs .
[ fig : x1]-[fig : x1 ] .
this approximate symmetry could be understood in terms of hamiltonian eq .
( [ eq : hamiltonian05 ] ) , or more easily in terms of hamiltonian eq .
( [ eq : hamiltonian06 ] ) by neglecting the tiny contributions of the last three time - dependent terms of eq .
( [ eq : hamiltonian05 ] ) .
if one puts aside the fixed static qubit - qubit interaction @xmath144 term , then the two qubits in hamiltonian eq .
( [ eq : hamiltonian06 ] ) are decoupled and have the symmetry of exchanging @xmath198 and @xmath199 in the presence of @xmath200 term . as a result ,
the time evolutions of the state probabilities @xmath201 , @xmath202 , @xmath203 and @xmath204 in figs .
[ fig : x1]-[fig : x1 ] , respectively , behave approximately the same .
similarly , the approximately same time evolutions for the state probabilities @xmath203 , @xmath204 , @xmath201 and @xmath202 in figs
. [ fig : x1]-[fig : x1 ] , respectively , are also observed . applying the optimal control sequences of two - qubit cnot gate shown in figs .
[ fig : cnot ] and to multi - level hamiltonian eq .
( [ eq : htot ] ) with initial four different qubit computational basis states @xmath185 , @xmath186 , @xmath187 and @xmath188 , we plot the time evolutions of the state probabilities of the cnot gate in figs .
[ fig : cnot]-[fig : cnot ] . when qubit 1 is initially in the state @xmath205 , the two qubits return to their original states ; when qubit 1 is initially in the state @xmath206 , an effective not operation is performed on qubit 2 ( the target qubit ) at the end of the cnot gate operation .
this is exactly what we see in figs .
[ fig : cnot]-[fig : cnot ] .
the error of the cnot gate when including the higher energy - level states is @xmath207 . by comparing this error to the error of the single - qubit gates of @xmath208 , leakage to the higher - energy level states outside
the computational basis state space is more appreciable for the cnot gate control pulses .
this may be due to the fact that the magnitude of the reduced time - dependent magnetic flux @xmath175 of fig .
[ fig : cnot ] for the cnot gate is larger than @xmath174 and @xmath175 for the single - qubit gates [ see , e.g. , @xmath174 and @xmath175 of figs . [
fig : x1 ] and [ fig : x1 ] for the @xmath156 gate ] .
nevertheless , the cnot gate error of the two coupled flux qubits is , to our knowledge , still much lower than the currently available schemes when the leakage to higher energy - level states is considered .
we have also performed an optimal cnot gate operation with qubit 1 being the target qubit and qubit 2 being the control qubit .
this cnot gate can also be accomplished at @xmath140 ns with error @xmath209 .
we show below that our optimal control scheme is robust against qubit decoherence .
we use the born - markov master equation for the reduced density matrix of the flux qubit system in the lindblad form to model the effect of decoherence on the qubit dynamics @xcite : @xmath210+\sum_{i=1}^{2}\left(\gamma_{1,i}\mathcal{d}\left[\sigma_{i}^{-}\right]\rho(t)+\gamma_{\varphi , i}\mathcal{d}\left[\sigma_{i}^{z}\right]\rho(t)\right ) , \label{eq : mastereq1}\\ & \equiv & \mathcal{l}\rho(t ) , \label{eq : mastereq}\end{aligned}\ ] ] where the hamiltonian @xmath98 is defined in eq .
( [ eq : hamiltonian05 ] ) , the superoperator @xmath211\rho = c\rho c^{\dagger}-\frac{1}{2}c^{\dagger}c\rho-\frac{1}{2}\rho c^{\dagger}c,\ ] ] and @xmath212 and @xmath213 are the relaxation rate and dephasing rate of qubit @xmath214 , respectively . generally speaking , if one starts from a microscopic model in the weak system - environment coupling limit , the argument in the superoperator @xmath215 in the born - markov approximation should rather be operators generating transitions between the energy eigenstates of the coupled - qubit hamiltonian ( [ eq : hamiltonian05 ] ) @xcite .
however , for small inter - qubit coupling considered here , these operators in the rotating - wave approximation can be approximately represented by individual qubit transition operators , @xmath216 and @xmath217 .
in addition , at finite temperatures one should have a dissipative term describing the influx of energy from the thermal environment into the system .
but since typical flux - qubit experiments are conducted at a low temperature around 50 mk @xcite which is much smaller than the transition frequencies of the qubits considered here ( i.e. , @xmath218 ) . as a result , the thermal mean occupation number of the environment modes at the energy of about @xmath219 approaches zero , and the environment may be regarded as an effective zero - temperature bath .
these justify the use of the master equation ( [ eq : mastereq1 ] ) to phenomenologically model the decoherence dynamics of the coupled flux - qubit system .
the master equation is symbolically expressed in a concise form with a liouville s superoperator @xmath220 in eq .
( [ eq : mastereq ] ) .
the equation of motion of the propagator can be found , by substituting @xmath221 into eq .
( [ eq : mastereq ] ) , to be @xmath222 with the relaxation time @xmath223 obtained from the rabi oscillation and the decoherence time @xmath224 from ramsey interference measurements in an recent experiment @xcite that probed the noise spectrum of a superconducting flux qubit , one is able to deduce the realistic values of @xmath212 and @xmath213 using the relations of @xmath225 , @xmath226 and @xmath227 . define the gate error @xmath228 by replacing @xmath229 with @xmath230 in eq .
( [ error1 ] ) for the case considering qubit decoherence .
one can perform the optimal control calculation by minimizing the cost function eq .
( [ eq : costfun02 ] ) with @xmath231 using the equation of motion ( [ eq : g ] ) . the single - qubit gate errors considering the effect of decoherence
are found to be in the order of @xmath7 , and the errors are in the order of @xmath8 for two - qubit cnot gates .
these gate errors are still below the error threshold @xmath9 ( @xmath5 in @xcite ; @xmath10 if surface code error correction is used @xcite ) required for fault - tolerant quantum computation .
we have also tested numerically that using a finite - temperature master equation with arguments in the superoperator @xmath215 being the operators generating transitions between the energy eigenstates but with the decay rates and the dephasing rates assuming to be the same as individual single qubits leads to the similar error values with correction only to the digits in @xmath6 or @xmath232 .
.[tab : gate_error]summary of the quantum gate errors [ cols="^,^,^,^,^",options="header " , ] we have applied the quantum optimal control method to implement fast and high - fidelity single - qubit and two - qubit gates for two inductively coupled superconducting flux qubits with fixed static off - diagonal qubit - qubit coupling and fixed qubit transition frequencies .
table [ tab : gate_error ] summarizes the gate errors calculated with realistic experimental parameters for ideal unitary case , and for the cases considering the effect of leakage state and the effect of qubit decoherence .
our optimal control scheme has the following great advantages .
( 1 ) our scheme that shares the same advantage of other directly coupling scheme requires no additional coupler subcircuit and control lines and thus is simple in experimental design .
the control lines needed are only for the manipulation of individual qubits ( e.g. , the time - dependent magnetic flux or field through each qubit ) .
( 2 ) quantum gates constructed via our scheme are all with very high fidelity ( very low error ) as our optimal control scheme takes into account the fixed qubit detuning and fixed two - qubit interaction as well as all other time - dependent magnetic - field - induced single - qubit interactions and two - qubit couplings when performing single - qubit and two - qubit gates . ( 3 ) our scheme can cope with noise and decoherence very well as the qubits are biased at the optimal coherence point to reduce the influence of low - frequency flux noise , and the gate operation time ( @xmath3 ns ) is much shorter than the corresponding qubit decoherence time ( a few @xmath233s or larger ) @xcite so that the decoherence effect on such a fast gate is significantly diminished .
( 4 ) a cnot gate or other general quantum gates can be implemented in a single run of pulse sequence rather than being decomposed into several single - qubit and some entangled two - qubit operations in series by composite pulse sequences .
a natural question then arises : how to generate these optimal control pulses experimentally ?
the pulse sequences shown in fig .
[ fig : x1 ] and , and in fig .
[ fig : cnot ] and look experimentally challenging , but not impossible .
commercial devices ( e.g. , tektronix awg70001a ) for generating arbitrary wave forms with 10 bits of vertical resolution at a sample rate of 50 gsa / s , a bit rate of 12.5 gb / s and a rise / fall time smaller than 27 ps are now available .
such a device should enable generation of complex signals in a time scale of sub - nanoseconds to nanoseconds .
this high - end arbitrary wave form generator combining with ultrafast josephson electronics should be or about to be able to fulfill the necessary specifications for implementing the optimal control pulse sequences we obtain for the quantum gate operations presented here .
another challenge for the experimental implementations and applications of the quantum optimal control theory is one s imprecise knowledge of the quantum system s parameters .
typically , the quantum gates constructed by the quantum optimal control theory are computed to very high precision assuming the parameters in the model hamiltonian are exactly known . however , for real systems controlled in the experiments , the parameters and also the hamiltonians usually are not known exactly .
this poses a challenge to implement such high - fidelity gates successfully in the laboratory .
recently , a hybrid open / closed - loop optimal control method called adaptation by hybrid optimal control ( ad - hoc ) has been proposed @xcite to overcome not only the problem of inaccurate knowledge of the system parameters but also shortcomings of the assumed physical model and errors on the controls themselves .
the basic idea is to use the open - loop quantum optimal control theory to find the optimal control pulses with the best available model and parameters of the system .
then the pulses are sent to the experiment and their performance are efficiently measured by using , for example , the method of randomized benchmarking @xcite .
the closed - loop pulse calibration of ad - hoc , similar to adaptive model - free feedback control ( also referred to as closed - loop laboratory control or learning control ) @xcite , uses the physical system itself as a feedback to calibrate control pulses and optimize their performance . by using , e.g. , the robust and efficient nelder - mead algorithm @xcite ,
the control pulses are updated experimentally and the procedure is iterated until a target performance is reached or convergence stops . after the calibration is finished , the optimized pulses can then be used . even though the precise experimental parameters are never identified , ad - hoc can reduce the initial gate errors of the numerically obtained optimal control pulses implemented in an experiment by at least an order of magnitude @xcite .
thus the quantum optimal control theory is practical and applicable experimentally through , e.g. , the ad - hoc protocol @xcite . a recent experiment demonstrating single - qubit operations ( with fidelity of 0.99 ) and two - qubit entanglement for the electron spins of two proximal nitrogen - vacancy centers in diamond using optimal control has been reported @xcite .
it is thus believed that high - fidelity quantum gates with optimal control pulses obtained via our scheme will be realized experimentally in the near future .
we are very grateful to prof .
yu - xi liu for sending us his note for the derivation of the coupled flux - qubit hamiltonian of ref . @xcite .
we acknowledge support from the national science council in taiwan under grant no .
100 - 2112-m-002 - 003-my3 , from the national taiwan university under grants no . 103r891400 , no .
103r891402 and 102r3253 , and from the focus group program of the national center for theoretical sciences , taiwan .
99 j. q. you and f. nori , phys . today * 58 * , 11 , 42 ( 2005 ) ; j. q. you and f. nori , nature * 474 * , 589 ( 2011 ) , and references therein
. j. clarke and f. k. wilhelm , nature * 453 * , 1031 ( 2008 ) , and references therein .
y. nakamura , yu .
a. pashkin , and j. s. tsai , nature * 398 * , 786 ( 1999 ) .
d. vion , a. aassime , a. cottet , p. joyez , h. pothier , c. urbina , d. esteve , and m. h. devoret , science * 296 * , 886 ( 2002 ) .
i. chiorescu , y. nakamura , c. j. p. m. harmans , and j. e. mooij , science * 299 * , 1869 ( 2003 ) .
a. wallraff , d. i. schuster , a. blais , l. frunzio , j. majer , m. h. devoret , s. m. girvin , and r. j. schoelkopf , phys .
lett . * 95 * , 060501 ( 2005 ) . j. m. martinis , k. b. cooper , r. mcdermott , m. steffen , m. ansmann , k. d. osborn , k. cicak , s. oh , d. p. pappas , r. w. simmonds , and c. c. yu , phys . rev .
lett . * 95 * , 210503 ( 2005 ) . c. rigetti , j. m. gambetta , s. poletto , b. l. t. plourde , j. m. chow , a. d. crcoles , j. a. smolin , s. t. merkel , j. r. rozen , g. a. keefe , m. b. rothwell , m. b. ketchen , and m. steffen , phys .
b * 86 * , 100506 ( 2012 ) .
j. m. chow , l. dicarlo , j. m. gambetta , f. motzoi , l. frunzio , s. m. girvin , and r. j. schoelkopf , phys .
a * 82 * , 040305(r ) ( 2010 ) .
h. paik , d. i. schuster , l. s. bishop , g. kirchmair , g. catelani , a. p. sears , b. r. johnson , m. j. reagor , l. frunzio , l. i. glazman , s. m. girvin , m. h. devoret , and r. j. schoelkopf , phys .
lett . * 107 * , 240501 ( 2011 ) .
j. majer , j. m. chow , j. m. gambetta , j. koch , b. r. johnson , j. a. schreier , l. frunzio , d. i. schuster , a. a. houck , a. wallraff , a. blais , m. h. devoret , s. m. girvin , and r. j. schoelkopf , nature * 449 * , 443 ( 2007 ) .
l. dicarlo , j. m. chow , j. m. gambetta , l. s. bishop , b. r. johnson , d. i. schuster , j. majer , a. blais , l. frunzio , s. m. girvin , and r. j. schoelkopf , nature * 460 * , 240 ( 2009 ) .
j. m. chow , a. d. crcoles , j. m. gambetta , c. rigetti , b. r. johnson , j. a. smolin , j. r. rozen , g. a. keefe , m. b. rothwell , m. b. ketchen , and m. steffen , phys .
lett . * 107 * , 080502 ( 2011 ) .
a. dewes , f. r. ong , v. schmitt , r. lauro , n. boulant , p. bertet , d. vion , and d. esteve , phys .
lett . * 108 * , 057002 ( 2012 ) . j. m. chow , j. m. gambetta , a. d. crcoles , s. t. merkel , j. a. smolin , c. rigetti , s. poletto , g. a. keefe , m. b. rothwell , j. r. rozen , m. b. ketchen , and m. steffen , phys . rev
* 109 * , 060501 ( 2012 ) .
l. dicarlo , m. d. reed , l. sun , b. r. johnson , j. m. chow , j. m. gambetta , l. frunzio , s. m. girvin , m. h. devoret , and r. j. schoelkopf , nature * 467 * , 574 ( 2010 ) ; m. neeley , r. c. bialczak , m. lenander , e. lucero , m. mariantoni , a. d. oconnell , d. sank , h. wang , m. weides , j. wenner , y. yin , t. yamamoto , a. n. cleland , and j. m. martinis , nature * 467 * , 570 ( 2010 ) .
a. p. peirce , m. a. dahleh , and h. rabitz , phys .
a * 37 * , 4950 ( 1988 ) .
d. j. tannor , v. a. kazakov and v. orlov , in _ time - dependent quantum molecular dynamics _ , edited by j. broeckhove and l. lathouwers ,
nato advanced studies institute , series b : physics ( plenum press , new york , 1992 ) , vol .
347 - 360 .
j. p. palao and r. kosloff , phys .
lett . * 89 * , 188301 ( 2002 ) ; j. p. palao and r. kosloff , phys .
a * 68 * , 062308 ( 2003 ) .
n. khaneja , t. reiss , c. kehlet , t. schulte - herbrggen , and s. j. glaser , j. magn
. reson . *
172 * , 296 ( 2005 ) ; d .- b .
tsai , p .- w .
chen , and h .- s .
goan , phys .
a * 79 * , 060306(r ) ( 2009 ) .
a. sprl , t. schulte - herbrggen , s. j. glaser , v. bergholm , m. j. storcz , j. ferber , and f. k. wilhelm , phys .
a * 75 * , 012302 ( 2007 ) .
r. eitan , m. mundt , and d. j. tannor , phys .
a * 83 * , 053426 ( 2011 ) .
m. wenin and w. ptz , phys .
a * 78 * , 012358 ( 2008 ) ; m. wenin and w. ptz , phys .
b * 78 * , 165118 ( 2008 ) ; m. wenin , r. roloff , and w. ptz , j. appl . phys . *
105 * , 084504 ( 2009 ) .
r. roloff and w. ptz , phys .
b * 79 * , 224516 ( 2009 ) ; m. wenin and w. ptz , phys . rev .
a * 74 * , 022319 ( 2006 ) .
h. jirari , europhys . lett . *
87 * , 40003 ( 2009 ) .
p. aliferis and j. preskill , phys .
a * 79 * , 012332 ( 2009 ) .
d. s. wang , a. g. fowler , and l. c. l. hollenberg , phys .
a * 83 * , 020302(r ) ( 2011 ) .
a. g. fowler , a. c. whiteside , and l. c. l. hollenberg , phys .
lett . * 108 * , 180501 ( 2012 ) , y .- x .
liu , j. q. you , l. f. wei , c. p. sun , and f. nori , phys .
lett . * 95 * , 087001 ( 2005 ) .
f. deppe , m. mariantoni , e. p. menzel , a. marx , s. saito , k. kakuyanagi , h. tanaka , t. meno , k. semba , h. takayanagi , e. solano , and r. gross , nat . phys . * 4 * , 686 ( 2008 ) .
h. j. carmichael , _ statistical methods in quantum optics 1 _ ( springer , berline , 1999 ) . c. w. gardiner and p. zoller , _ quantum noise _
, 2nd edition .
( springer - verlag , berlin , 2000 ) .
breuer and f. petruccione , _ the theory of open quantum systems _
( oxford university press , oxford , 2002 ) .
d. j. egger , and f. k. wilhelm , arxiv:1402.7193 .
e. magesan , j. m. gambetta , b. r. johnson , c. a. ryan , j. m. chow , s. t. merkel , m. p. da silva , g. a. keefe , m. b. rothwell , t. a. ohki , m. b. ketchen , and m. steffen , phys .
109 * , 080505 ( 2012 ) ; e. magesan , j. m. gambetta , and j. emerson , phys . rev .
* 106 * , 180504 ( 2011 ) . j. m. chow , j. m. gambetta , l. tornberg , j. koch , l. s. bishop , a. a. houck , b. r. johnson , l. frunzio , s. m. girvin , and r. j. schoelkopf , phys .
lett . * 102 * , 090502 ( 2009 ) ; | we apply the quantum optimal control theory based on the krotov method to implement single - qubit @xmath0 and @xmath1 gates and two - qubit cnot gates for inductively coupled superconducting flux qubits with fixed qubit transition frequencies and fixed off - diagonal qubit - qubit coupling . our scheme that shares the same advantage of other directly coupling schemes requires no additional coupler subcircuit and control lines .
the control lines needed are only for the manipulation of individual qubits ( e.g. , a time - dependent magnetic flux or field applied on each qubit ) .
the qubits are operated at the optimal coherence points and the gate operation times ( single - qubit gates @xmath2 ns ; cnot gates @xmath3 ns ) are much shorter than the corresponding qubit decoherence time . a cnot gate or other general quantum gates
can be implemented in a single run of pulse sequence rather than being decomposed into several single - qubit and some entangled two - qubit operations in series by composite pulse sequences .
quantum gates constructed via our scheme are all with very high fidelity ( very low error ) as our optimal control scheme takes into account the fixed qubit detuning and fixed two - qubit interaction as well as all other time - dependent magnetic - field - induced single - qubit interactions and two - qubit couplings .
the effect of leakage to higher energy - level states and the effect of qubit decoherence on the quantum gate operations are also discussed . |
soon after the theoretical prediction of universal conductance fluctuations ( ucf ) in disordered electronic samples,@xmath0 it was shown that their distribution function is mainly gaussian.@xcite these considerations referred to a weakly disordered _ open _
sample connected to a reservoir by broad external contacts . in this case , after diffusing through the sample of size @xmath1 during time @xmath2 electrons are inelastically scattered in the reservoir ( @xmath3 is the diffusion coefficient ) . the resulting uncertainty in the level position , i.e. level broadening @xmath4 , is of the order of the thouless energy @xmath5 , and @xmath6 , where @xmath7 is the mean level spacing .
this picture remains almost unchanged in the presence of relatively strong inelastic processes within the sample@xcite which lead to dephasing ( the loss of phase - coherence ) at the time scale @xmath8 : the conductance distribution remains almost gaussian although its variance is no longer universal but decreases @xmath9 where @xmath10 is a spatial dimensionality .
in contrast , it has been shown more recently that the conductance distribution is pronouncedly non - gaussian in almost closed systems like chaotic cavities with single channel leads @xcite , quantum dots in the regime of coulomb blockade@xcite where the distribution of the heights of the coulomb blockade peaks has been measured experimentally,@xcite or isolated conducting rings threaded by an aharonov
bohm flux.@xcite weak transmission through the contacts means that the electrons typically spend more time in the system than @xmath11 so the broadening of energy levels @xmath12 .
this inequality corresponds to the ergodic regime which allows the use of non - perturbative techniques including random matrix theory @xcite and the zero dimensional supersymmetric @xmath13
model.@xcite there are two distinct types of behavior within the ergodic regime , depending on whether the level broadening @xmath4 is smaller or larger than the mean level spacing @xmath7 . in the absence of level overlapping , @xmath14
, the conductance distribution is clearly non - gaussian .
the nature of the distribution changes considerably when dephasing is substantial , @xmath15 , so that the energy levels overlap .
this case allows also a perturbative treatment@xcite in the framework of the standard diagrammatic expansion .
a crucial question here is how to describe the level broadening in quantum dots with point - like external contacts .
the first approach , following ref .
, was to attach an additional voltage probe in which electrons lose their phase coherence before being re - injected into the dot.@xcite another approach was to include an imaginary potential in the hamiltonian thus allowing for the possibility that electrons may leave the dot.@xcite however , brouwer and beenakker @xcite have recently pointed out that these approaches do not describe inelastic processes _ within _ the dot .
while the voltage probe accounts for spatially localized dephasing only , the imaginary potential in the hamiltonian does not conserve the particle number . in order to describe dephasing processes within the dot
, they have introduced@xcite in the framework of random matrix theory a new variant of the voltage probe model which produces spatially uniform dephasing .
thus it was found that the conductance distribution became gaussian for strong dephasing , @xmath16 .
note that experiments measuring conductance fluctuations in quantum dots , for example as a function of shape,@xcite also find a gaussian distribution .
in this paper we present a microscopic diagrammatic calculation of the conductance distribution function within a dirty quantum dot with two point contacts in the presence of strong dephasing processes ( @xmath16 ) within the dot .
we construct a particle - conserving model by clearly separating two different mechanisms of level broadening : dephasing processes which contribute only to the conductance fluctuations and particle leakage which contributes also to the mean value of conductance .
it enables us to resolve a contradiction between the results of the imaginary potential model @xcite that suffered from the violation of particle conservation and the voltage - probe model.@xcite by calculating the moments of the conductance distribution , we will show that for @xmath17 it is almost gaussian with the variance @xmath18 this expression is valid both in the ergodic zero - mode regime , @xmath19 , and in the diffusive regime , @xmath20 . here
@xmath21 is the mean conductance , and @xmath22 where the small parameter @xmath23 is effectively a diffusion propagator at distance @xmath24 in the presence of strong dephasing , and @xmath25 is the elastic mean free path . we will show that this parameter is given by @xmath26 here @xmath27 is the dephasing length . note that although the voltage - probe model@xcite also predicts a gaussian conductance distribution , the width of the distribution given by eq .
( [ var ] ) is different from that obtained in ref . .
in contrast to a cavity with broad leads,@xcite ( which corresponds to a multi - channel case so that the fluctuations remain universal ) the variance is non - universal even in the diffusive regime , @xmath28 . the factor @xmath29 which governs the distribution width resembles the standard weak - localization parameter .
this could be seen from the first line of eq .
( [ zr ] ) where @xmath30 is a classical dimensionless conductance of a sample with broad leads
. nevertheless , the origin of this parameter is quite different : the variance is not very sensitive to the presence of a magnetic field and thus has nothing to do with weak localization effects .
when time - reversal invariance is absent , the variance ( [ var ] ) changes only by the standard dyson s factor of @xmath31 , typically for mesoscopic effects , where @xmath32 for the orthogonal ensemble ( in the presence of potential scattering only ) , and @xmath33 for the unitary ensemble ( in the presence of a finite magnetic field or weak scattering by magnetic impurities that breaks time - reversal symmetry ) .
it is worth noting that the close - to - gaussian nature of the distribution is established in a way somewhat less straightforward than for ucf in samples with broad leads.@xcite had it been possible to restrict considerations to the lowest order of perturbation for the variance and higher moments , we would have obtained the exponential distribution as in the imaginary potential model , @xcite albeit with a shifted position of the mean .
however , the leading contribution to the moments is given by the next ( two - loop ) order of perturbation while contributions from higher orders are negligible .
in considerations of conductance fluctuations in quantum dots with single - channel leads , the lowest order of perturbation does not make a leading contribution due to the existence of two different perturbative parameters related to qualitatively different contributions to the level broadening .
although both of them arise also in considering any mesoscopic effect for a diffusive sample with broad contacts , only one of them is relevant in that case as we explain below .
we adopt the impurity diagram technique @xcite to perform ensemble averaging over disorder with allowance for dephasing processes . at the first step ,
we assume a conventional model of free electrons in a random gaussian potential @xmath34 , with @xmath35 and @xmath36 where @xmath37 is the one - electron density of states and @xmath38 is the mean elastic scattering time .
a standard perturbative expansion @xcite is a loop - expansion with the loops being built up of diffusons and cooperons ( if time reversal symmetry is preserved ) .
these are impurity ladders ( fig .
1 ) with a pole at small transferred momenta ( diffusons ) or small total momenta ( cooperons ) . for an isolated @xmath10-cubic sample of size @xmath1 with @xmath39 such a diffusion propagator ( _ diffuson _ ) at @xmath40 , i.e. in the absence of inelastic processes ,
is given by @xmath41 where @xmath42 is the diffusion constant , @xmath43 , and @xmath44 with @xmath45 being non - negative integers .
in the presence of time reversal symmetry , the cooperon propagator has the same form . for weak disorder ,
the diffusion propagator is proportional to the fourier transform of the density - density correlation function , @xmath46 which describes the probability of finding at the point @xmath47 at time @xmath48 a diffusive particle injected at the origin at @xmath49 .
particle conservation implies that @xmath50 at any time @xmath51 . on performing the fourier transform
, one then obtains that @xmath52 .
this exact identity remains valid also at @xmath53 when one includes a weak interaction leading to dephasing processes in the sample .
therefore , the diffuson associated with @xmath54 remains the same as in eq .
( [ dp ] ) : the only possible effect of dephasing could be in replacing @xmath3 with @xmath55 ( for weak disorder , @xmath56 so that this effect may also be neglected ) .
= 0.9 [ fig:1 ] fig .
[ fig:1 ] . a diffuson ladder .
a cooperon ladder is obtained by inverting the direction of one of the arrows . mathematically , particle conservation in the diffuson associated with the density - density correlation function ( which we call _ the intrinsic diffuson _ ) is due to the ward identity which provides an exact cancellation of the one - particle self - energy corrections due to inelastic processes by vertex corrections ( described graphically as interaction lines between the two sides of the diffusion ladders ) .
however , if one defines the diffuson in terms of the sum of graphs describing diffusive propagation in the particle - hole channel _ without energy exchange _ between the particle and hole line , it is no longer proportional to the density - density correlation function and it is not constrained by particle conservation . @xcite the absence of the usual cancellation of self - energy corrections by vertex corrections leads to the appearance of the cutoff @xmath57 in the diffusion pole .
@xcite this is precisely the kind of diffuson arising in typical lowest - order conductance - fluctuations diagrams@xmath58 and describing correlations between _ different members of the ensemble _ which makes energy exchange between the opposite sides of the diffusion ladder impossible . we will call this kind of diffuson _
inter - sample_. note that all the cooperon propagators , either intrinsic or inter - sample , have exactly the same cutoff , @xmath57 , as the inter - sample diffusons.@xcite therefore , one can allow for dephasing by merely including @xmath57 as an infrared cutoff in the denominators of _ inter - sample _ diffusons , i.e. substituting into eq .
( [ dp ] ) @xmath57 for @xmath59 at zero frequency . at @xmath53 , averaging over energies leads@xcite to the saturation of the denominators of _ inter - sample _ diffusons .
a consistent description of both thermal smearing and dephasing requires the use of the temperature technique@xcite and the development of a microscopic theory of dephasing in the partially open dot which would provide a dependence of @xmath60 on @xmath61 and microscopic characteristics of the dot .
such a theory is well established for relatively large diffusive samples with broad leads@xcite or for closed quantum dots.@xcite however , our aim here is just to describe how dephasing and thermal smearing influence mesoscopic fluctuations of conductances of the dot . to this end , all we need is to distinguish between the two kinds of diffusons : _ inter - sample _ diffusons which are saturated at @xmath62 , and _ intrinsic _ diffusons which diverge at @xmath62 . since the divergence of this sort could not be treated within the diagrammatic approach , we introduce a weak level broadening @xmath63 due to escape from the dot which provides a cutoff for the _ intrinsic _ diffusons . the cutoff in the intrinsic diffuson
violates the particle conservation law but since we restrict our considerations to the case @xmath64 , it is dephasing ( or thermal smearing ) within the dot rather than particle escape which governs the conductance distribution .
even with this cutoff , @xmath65 , i.e. there is a sharp distinction between the _ intrinsic _ diffusons , @xmath66 , and the _ inter - sample _ diffusons , @xmath67 . for simplicity , we shall use a zero - temperature diagrammatic technique .
if @xmath68 , one should substitute @xmath61 for @xmath69 and the thermal - smearing length @xmath70 for the dephasing length @xmath71 in all final results , including that for the variance , eq .
( [ var ] ) .
the existence of the two kinds of diffusons was practically irrelevant for the description of mesoscopic conductance fluctuations in samples with broad leads.@xmath58 firstly , there is strong leakage through the leads , @xmath72 , and one could restrict considerations to the case of weak dephasing , @xmath73 , when both intrinsic and inter - sample diffusons have the same cutoff .
moreover , even for larger samples or higher temperatures when @xmath74 , the intrinsic diffusons do not appear either in @xmath75 or in var@xmath76 but as a third - order perturbation correction which is irrelevant in a weakly disordered metal .
in contrast to this , in the problem under consideration the intrinsic diffusons determine the value of @xmath75 and enter var@xmath76 in the leading order of perturbation .
traditionally conductance fluctuations in a system with broad , spatially homogeneous contacts are considered@xmath58 by means of the kubo formula @xcite .
alternative considerations based on the landauer - bttiker formula @xcite are more convenient in the presence of tunnel barriers @xcite or for a lead geometry which involves spatially inhomogeneous currents . in this paper
we will use this formula in the form first derived by fisher and lee:@xcite @xmath77where the transmission coefficient @xmath78 is the probability of transmission from the channel labeled by @xmath79 in the left ( right ) lead to the channel labeled by @xmath80 in the right ( left ) lead . in the case of single - channel leads ( of width @xmath81 ) the point - to - point conductance , @xmath76 ,
is given by eq .
( [ lb ] ) without the summation sign , the transmission coefficients in this equation being related to green s functions by @xcite @xmath82\ , , \label{tab}%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\]]where @xmath83 ( @xmath84 ) is a retarded ( advanced ) green s function , @xmath85 , @xmath86 are the positions of the point contacts , and @xmath87 is the transmission probability through the contacts themselves . in the entire energy interval of interest the mean density of states
@xmath37 is a constant and the @xmath88 are energy independent so we will subsequently drop the @xmath89 label . thus we imply that energy dependence will also be irrelevant for the conductance fluctuations : as usual , this is valid when the energy difference between two conductances , @xmath90 , is much smaller than the level broadening which we assume to be the case .
we consider the region of parameters defined by @xmath91here the level broadening due to escape from the dot is represented by @xmath92 where @xmath93 is the level broadening due to escape from the dot other than through the point contacts , for example due to leakage through the dot matrix .
our perturbative approach is formally applicable both to the non - ergodic diffusive regime , @xmath94 , and to the ergodic zero - mode regime , @xmath95 .
the inequality @xmath96 is a crucial distinction from the imaginary potential model @xcite where the opposite inequality holds .
this allows us to separate dephasing processes inside the dot from those due to weak particle leakage .
the latter have been introduced only for convergence of the perturbative approach which is not valid for @xmath97 .
therefore , a weak violation of particle conservation is inevitable in the perturbative approach as will be explicitly shown later .
however , this weak particle non - conservation is not related to the level broadening due to dephasing inside the dot and , as we shall see , may be totally neglected in the final results . on the contrary , in the imaginary potential model both the level broadening and the particle non - conservation
have exactly the same source and can not be separated which makes this model unsuitable for considerations of the influence of inelastic processes inside the dot on the conductance distribution .
the nonperturbative region @xmath97 was the main area of interest for previous zero mode calculations .
@xcite however , the exact zero - mode calculations within the nonlinear @xmath13 model can not be extended ( at least , in a straightforward way ) to include relaxation processes inside the dot by simply introducing @xmath60 .
the reason is that all intra - sample diffusons _ must _ and all intrinsic diffusons _ must not _ contain a cutoff proportional to @xmath57 .
these two kinds of diffusons which are both elementary excitations in the nonlinear @xmath13 model must be distinguished automatically .
this can be easily done in the perturbative approach to the @xmath13 model @xcite by the introduction of an extra set of matrix indices to numerate different conductance loops .
while this additional dependence does not lead to any complication in perturbative or renormalization group calculations , it would make direct zero mode calculations rather difficult if possible at all .
obviously , one circumvents this complication when dephasing is included phenomenologically , as in the voltage probe model where one can proceed with zero - mode calculations .
@xcite however , there is a price to pay : as two kinds of diffusons are not distinguished , this model contains @xmath60 proportional corrections to the mean conductance@xcite even in the absence of time - reverse invariance when such corrections can not appear in all orders of perturbation as we will show below .
the present model overlaps with the voltage probe model in the ergodic zero - mode regime ( with strong dephasing ) : @xmath98 . even in this case , the overlapping is restricted to the region @xmath99 as the requirement of convergence of perturbation series does not allow us to take the limit @xmath100 .
nevertheless , both the models address clearly the same physical situation where dephasing occurs inside the dot , and the results of the present considerations are in a broad agreement with the voltage probe model .
moreover , we shall argue later that the perturbative approach might be valid even outside the formal limits of applicability . note finally that since our considerations always refer to strong dephasing with a noticeable leakage ( @xmath101 ) , the results are applicable both for statistics of the heights of the coulomb blockade peaks ( for @xmath102 ) and for the conductance distribution of a single - channel cavity ( for @xmath103 ) .
conductance cumulants @xmath104 are given by @xmath105 where @xmath106 should be expressed in terms of green s functions , eq .
( [ tab ] ) , and an extra prefactor of 2 is due to spin .
fisher and lee @xcite have shown that the unitarity of the @xmath107 matrix implies @xmath108 even in the absence of time reversal invariance ( in @xcite the proof of this equality for a two lead system has been explicitly based on particle conservation ) . in the imaginary potential model of refs . the particle number is not conserved and , if time reversal symmetry is broken , @xmath109 as we will show explicitly later .
that is why the conductance distribution in the absence of time reversal symmetry obtained@xcite from the microscopic formula ( [ lb ] ) turns out to be different from that obtained@xcite from the reduced formula @xmath110 .
in the model under consideration , the particle number is _ almost _ conserved : for the sake of convergence we have introduced @xmath111 which describes particle leakage ; however , in the region of interest , @xmath112 , the leakage effect is small and we will show that in the leading order approximation @xmath113 . on average though , @xmath114 so that the calculation of the mean conductance is straightforward . the mean conductance @xmath21 is clearly proportional to the fourier transform at zero frequency of the density - density correlation function , eq .
( [ cf ] ) . for the parametric region of eq .
( [ ineq ] ) and for point contacts separated by a large distance , @xmath115 , @xmath21 is dominated by the one - diffuson contribution shown in fig . 2 .
= 0.6 [ fig:2 ] fig .
[ fig:2 ] .
the diagram for the mean conductance . obviously , this is the _ intrinsic _ diffuson , obeying particle conservation . at each end of the diffuson
there are two - sided ` petal ' shapes given analytically by @xmath116 the filled circles at the petals correspond to the constants @xmath117 where @xmath118 are the transmission probabilities through the contacts ( here and elsewhere we put @xmath119 to @xmath120 in all intermediate expressions ) .
the petals correspond to motion at ballistic scales since the average green s functions @xmath121 ( drawn as edges of the petal ) decay like @xmath122 . at the diffusive scale , @xmath123 ,
the petals may be approximated by @xmath124 where the constant @xmath125 is given by@xcite @xmath126 where @xmath127 .
then one finds @xmath128 2 where @xmath129 with @xmath130 . as @xmath131 in the parametric regime ( [ ineq ] ) , @xmath132 is dominated by the zero - mode contribution ( @xmath133 ) only .
therefore , @xmath134 combining eqs .
( [ meang ] ) and ( [ esc ] ) , one finds @xmath135 as expected in the zero - mode approximation , @xmath21 is independent of the separation of the point contacts , the dimensionality , and the degree of disorder .
the perturbation expansion for @xmath21 is strictly valid only in the regime where @xmath136 . without introducing @xmath93 , we would have @xmath137 so that @xmath138 . however , substituting this value of @xmath139 into eq.([meang ] ) reproduces the result of the voltage probe model for the classical mean conductance @xcite which corresponds to the standard single - channel conductance .
this means that all perturbative corrections to @xmath21 either vanish or cancel each other . in the absence of time reversal symmetry
, higher order corrections could result from the expansion in the diffuson loops only .
in the usual case of broad contacts , one should sum over all @xmath140 in eq.([dr ] ) which leads to the parameter , @xmath141 at @xmath142 .
it is well known that all the ` main logarithms ' , @xmath143 , are mutually cancelled due to the renormalizability in all orders of perturbation .
@xcite the first nonvanishing contribution arises in the second order and is proportional to @xmath144 in correspondence with the renormalization group results .
@xcite however , in the zero mode regime considered here there is complete cancellation at least up to the fourth order similar to that occurring in the same regime in calculating the energy - level correlation function@xcite .
this is easy to verify as the calculation is much simpler than the standard one in the diffusive regime .
one could assume that the cancellation of the main logarithms in all higher orders in the diffusive regime corresponds to a similar cancellation of the zero mode contributions .
since in the latter case there is no renormalization group backing , such a cancellation remains only a plausible hypothesis .
nevertheless , it is certain that in the absence of time - reversal invariance no correction due to dephasing may appear as they are forbidden by particle conservation .
when time - reversal symmetry is not broken , there exist weak localization ( cooperon ) corrections in the lowest order which are proportional to the standard @xmath145 in the non - ergodic diffusive regime , @xmath146 , or to @xmath147 in the ergodic zero - mode regime , @xmath148 . in any case , in the region of eq .
( [ ineq ] ) these corrections are negligible , and the main result is given by eq .
( [ g ] ) .
the lowest order diagrammatic contribution to the variance is given in fig .
3 , for clarity both in the standard representation @xmath0 and in that with diffusion modes separated from ballistic ones @xcite as in fig . 1 .
although the diagram in fig .
3a looks precisely like one of the leading diagrams@xmath58 for the ucf , the result of the calculation is absolutely different due to the fact that the vertices of both conductance loops correspond to the point contacts .
this is most clear in the representation of fig .
3b where local petals are connected together by two inter - sample diffusons .
the petals reduce to the same constant as for the mean conductance , eq .
( [ chi2 ] ) .
thus , one finds by analogy with eq .
( [ meang ] ) that this diagram s contribution to the variance is @xmath149 = 0.95 [ fig:3 ] fig .
[ fig:3 ] .
the lowest - order contributions to the variance .
wavy lines represent diffuson ladders which correspond to the inter - sample diffusons .
the relation between a wavy line and a ladder is the same as for intrinsic diffusons , fig .
[ fig:1 ] . where the constant @xmath150 is the overall factor in the expression for @xmath151 in terms of green s function , eq .
( [ tab ] ) , this factor being explicitly defined in eq .
( [ meang ] ) . substituting the values of all constants and taking into account the contribution of the equivalent cooperon diagram ( obtained from that in fig .
3a by inverting the direction of arrows in one of the loops ) results in the following expression @xmath152 ^ 2 \ , .
\label{vg1}\ ] ] here @xmath153 is given by eq .
( [ dr ] ) with @xmath154 .
its value depends on whether one considers the non - ergodic diffusive regime , @xmath146 , or the ergodic zero - mode regime , @xmath148 . in the former case , the summation in eq .
( [ dr ] ) may be approximated by an integration with cutoff at @xmath155 , while in the latter case only the @xmath133 term makes a relevant contribution to the sum .
this leads to the expression for @xmath153 given in the introduction , eq .
( [ zr ] ) .
the contribution ( [ vg1 ] ) to the variance differs from @xmath156 by the substitution of @xmath153 for @xmath157 . as @xmath158 in the region ( [ ineq ] ) , this difference ( which is absent in the imaginary potential model@xcite ) is crucial for the description of dephasing within the dot .
it is important that the diffuson and cooperon diagrams originate from different terms in eq .
( [ lb2 ] ) .
indeed , a diffuson ladder connects @xmath159 and @xmath84 lines which have the opposite directions , while a cooperon ladder connects @xmath159 and @xmath84 with the same direction .
therefore , in the diffuson diagram in fig .
3a @xmath159 s have the same direction in both loops .
then it follows from the expression for @xmath106 in terms of green s functions , eq .
( [ tab ] ) , that this diagram originates from @xmath160 . as the direction of arrows
must be inverted in one of the loops in order to obtain an equivalent cooperon diagram , in this diagram @xmath159 s have the opposite directions so that it originates from @xmath161 . considering @xmath162 where ,
in this lowest diagrammatic order , the first contribution is given by the diffuson diagram while the second is given by the cooperon diagram , one can see that @xmath163 for @xmath33 ( in the absence of time reversal symmetry ) when the cooperon contribution vanishes .
for @xmath32 the two contributions in eq .
( [ l = r ] ) cancel each other . in the imaginary potential model@xcite
the lowest order diagrams make the leading contributions to the variance .
therefore , in this case the reduced formula @xmath164 is not equivalent to the microscopically derived@xcite formula ( [ lb ] ) .
since the @xmath165 equality is based on particle conservation , its breakdown in the imaginary potential model is another clear indication of the absence of particle conservation . in the present model where the particle number is ` almost ' conserved
, the breakdown of this equality in the lowest order indicates that the higher order diagrams must be more relevant .
this is indeed the case since the condition @xmath166 suggests that the dominant contribution to the variance in the region ( [ ineq ] ) is given by diagrams which contain the lowest possible number of intrinsic diffusons ( @xmath139 ) and have their conductance loops connected _ by only one _ inter - sample diffuson ( @xmath167 ) .
two such diagrams are shown in fig . 4 in the two equivalent representations described above
. directions of arrows are chosen in such a way that all propagators in both the diagrams are diffusons . inverting the direction of arrows in one of the conductance loops in diagrams ( a ) and ( c )
, one converts inter - sample diffusons into cooperons , leaving intrinsic diffusons intact .
in the diffuson diagram ( 4a ) , @xmath159 s have opposite directions in different conductance loops so that it originates from @xmath161 . on the contrary
, @xmath159 s have the same directions in both loops in the diagram ( 4c ) so that this diagram originates from @xmath160 .
thus , in contrast to the lowest order diagrams in fig . [ fig:3 ] , even for @xmath33 when time - reversal symmetry is broken , both @xmath161 and @xmath160 contribute to the conductance .
= 0.95 [ fig:4 ] fig .
[ fig:4 ] . the leading diagrammatic contribution to the variance . in the diagrams
_ b _ and _ d _ , the inter - sample diffusons are represented by wavy lines and the intrinsic diffusons by double lines .
the calculation of these diagrams is described in the appendix . for simplicity
, we did not draw single - impurity lines in diagrams in fig .
4c and 4d which `` dress '' in a standard way @xcite the inner square in fig .
such a dressing results in an exact cancellation of one of the inter - sample diffusons ( wavy lines ) by the square which makes the contribution of the diagram 4d equal to that of the diagram 4b .
in addition to these diagrams , one can construct only one more contribution to the variance which contains two diffuson loops .
this contribution describes a weak - localization correction to a diffuson and contains three intra - sample ladders thus being small even compared to the contribution of eq .
( [ vg1 ] ) .
adding together the contributions of all relevant diffuson and cooperon diagrams in this order , we find @xmath168 ^ 2 \!\ !
\zeta_{\phi}(\ell ) \ , . \label{vg2}\ ] ] comparing eqs .
( [ vg2 ] ) and ( [ vg1 ] ) , one has @xmath169 in the parametric region ( [ ineq ] ) corresponding to strong dephasing since @xmath170 is small while @xmath139 is close to 1 .
therefore , in this regime the variance is dominated by the non - universal contribution of eq .
( [ vg2 ] ) . substituting the expression for @xmath21 into eq .
( [ vg2 ] ) , we obtain the result of eq .
( [ var ] ) given in the introduction .
let us stress that the exact equality of the diffuson diagrams ( [ fig:4]a ) originating from @xmath171 and ( [ fig:4]c ) originating from @xmath172 ensures that the two terms in eq .
( [ l = r ] ) cancel each other even in the case @xmath33 when there are no cooperon contributions .
the same is valid for the leading contribution to the higher order moments of @xmath173 .
therefore , keeping only the leading contribution and neglecting sub - dominant contributions to @xmath174 which are due to a small particle leakage , one finds that @xmath165 as expected in any particle conserving model .
in order to determine the distribution , we need to find the leading contributions to the @xmath175th cumulant .
the lowest order diagrams are a generalisation of those for the variance @xcite , fig .
they have two @xmath175-petal daisy vertices , the petals being connected by @xmath175 inter - sample diffusion propagators .
each diagram gives a contribution of @xmath176^n$ ] where a factor of @xmath177 arises upon counting all possible ways of ordering the @xmath175 loops .
diagrams which generalize the leading order contribution to the variance for @xmath178 are shown in fig .
[ fig:5 ] .
= 0.9 [ fig:5 ] fig .
[ fig:5 ] . some of the diagrams making a leading contribution to the third moment of conductance .
diagrams in the second row are equivalent to their counterparts in the first row . 2 in general , such diagrams are obtained by taking @xmath175 conductance loops , each loop having an intra - sample diffuson ( as for the mean , fig .
2 ) , and connecting these loops with the least possible number of inter - sample diffusons . actually , the number of inter - sample diffusons may vary , as every inner polygon ( like the inner square in diagram [ fig:4]_d _ for the variance ) cancels one of them ( see appendix ) .
it is the number of loops consisting of inter - sample diffusons which should be minimal . for the third cumulant ( fig . 5 .
) , each diagram has two such loops .
the lowest order diagram , two three - petal daisies connected by three inter - sample diffusons , also has two diffuson loops . however , there is no cancellation there and this diagram is proportional to @xmath179 .
each diagram in fig .
5 is proportional to @xmath180 .
as we are interested in the case of strong dephasing ( @xmath181 ) and weak leakage @xmath182 , this contribution is dominant .
note that fig .
5 does not include all the diagrams of this order , and a numerical factor attached to the leading contribution to @xmath104 is not easy to calculate for @xmath178 and hardly possible for arbitrary @xmath175 .
however , this is not necessary for what follows .
it is sufficient to know that the leading contribution to the @xmath175th cumulant is proportional to @xmath183 .
thus , keeping both the lowest order and the leading contribution , we find @xmath184^n + \ , c_n \,\,\left [ \zeta_{\phi}(\ell ) \right]^{n-1}\right\ } , \label{pointn}\ ] ] where @xmath175-dependence of @xmath185 is much slower than factorial ( although it might include a trivial factor like @xmath186 or so ) . for weak particle escape and strong dephasing , eq . ( [ ineq ] ) , the second term in eq . ( [ pointn ] ) dominates .
comparison of this dominant contribution to an appropriate power of the variance gives @xmath187^{{n-2}}\ll1\,,\ ] ] so that the distribution is mainly gaussian with non - universal variance @xmath188 given by eqs .
( [ var ] ) and ( [ vg2 ] ) .
note that this is similar to the behavior of the higher order cumulants @xcite in the case of ucf in a sample with broad contacts , where such a ratio is proportional to @xmath189 .
very high cumulants , with @xmath190 , grow much faster which leads to lognormal tails of the distribution@xcite similar to those @xcite for samples with broad contacts . at weak disorder
the lognormal tails are practically irrelevant as they start at @xmath191 where @xmath192 , while typical deviations @xmath193 . in the presence of some particle leakage from the dot , @xmath194 ,
the exponential tails are more important , these tails being governed by first term in eq .
( [ pointn ] ) which dominates for larger @xmath175 .
it turns out that the exponential tails dominate the gaussian ones for @xmath195 = 0.9 [ fig:6 ] fig .
[ fig:6 ] .
solid curves show the conductance distribution for @xmath196 with dephasing rates @xmath197 and @xmath198 , calculated using the second term in eq .
( [ pointn ] ) with an arbitrarily chosen coefficient @xmath199 . for comparison
the dashed curves are gaussian distributions with the same mean and variance .
this estimation results from a comparison of the exponential distribution obtained using the first term in eq .
( [ pointn ] ) with the gaussian distribution with the variance ( [ vg2 ] ) obtained by neglecting the higher cumulants . for weak particle leakage ,
the exponential tails start at deviations which are much stronger than typical ones .
with increasing leakage , these tails become important at the same values of @xmath139 at which the variance is mainly contributed by escape processes , eq .
( [ ratio ] ) .
inequalities ( [ ratio ] ) and ( [ expt ] ) show that the crossover to the regime dominated by particle escape happens not at @xmath200 but at considerably weaker particle escape : either at @xmath201 in the ergodic regime ( @xmath202 ) , or at @xmath203 in the diffusive regime ( @xmath204 ) .
note that in eqs .
( [ ratio ] ) and ( [ expt ] ) we have disregarded the difference between @xmath205 and @xmath170 which is exact in the zero - mode regime and logarithmically accurate in the diffusive regime at @xmath142 . in the diffusive regime at @xmath206 , this difference is significant as may be seen from eq .
( [ zr ] ) . taking this into account will introduce an extra small factor of @xmath207 into the r.h.s . of the estimations ( [ ratio ] ) and ( [ expt ] ) thus suppressing a relative importance of escape processes . in the case of main interest ,
when particle leakage may be neglected , the distribution still deviates from gaussian since the cumulants , given by the second term in eq .
( [ pointn ] ) , do not vanish although they are small .
the distribution can be restored from these cumulants in a standard way ( see , e.g. , ref . ) .
exact values of the coefficient @xmath185 do not matter as the most relevant contribution to the deviation from a gaussian shape is given by the third cumulant , similarly to the case of the conductance fluctuations in samples with broad leads.@xcite the resulting distributions are shown for two values of @xmath208 in fig .
it is seen that deviations from the gaussian shape are not very strong . in conclusion
, we have found that the conductance distribution for disordered quantum dots with single - channel leads becomes mainly gaussian for strong dephasing , @xmath209 , in broad agreement with the voltage probe model of ref . .
let us stress again that since we have assumed non - zero @xmath93 in this paper , there is no exact region of overlap between our results and those of the voltage probe model . therefore
, a direct comparison of the results is not possible .
however there remain some noticeable differences , first of all in the dependence of the variance of the gaussian distribution on @xmath60 .
if we stretch our results to the limit of applicability putting @xmath210 and making the contacts transparent ( @xmath211 ) , we will have var@xmath212 while the result of ref . for transparent contacts
is var@xmath213 .
note that although in our approach resummation in all powers of @xmath139 might be necessary for @xmath214 , it could not change the dependence of the result on @xmath60 .
similarly , in contrast to another result of ref . , the dephasing rate @xmath57 could not appear in the mean conductance in the unitary case , @xmath215 .
indeed , in this case cooperons vanish , and particle conservation strictly prohibits the appearance of @xmath57 in any power of expansion in diffusons . however , this disagreement is minor , and is clearly due to differences between the models .
let us stress that in contrast to the imaginary potential model , our present microscopic considerations confirm the prediction of the voltage probe model on the gaussian character of the conductance distribution with a non - universal variance depending on the dephasing rate . the most distinctive feature of the perturbative approach used here was taking into account two different sources of level broadening , one due to the presence of contacts entering the classical part of the mean conductance , and the other due to dephasing , appearing only in weak localization corrections and fluctuations .
it remains an open question as to whether a similar introduction of two different sources of level broadening could be used within a non - perturbative approach such as the supersymmetric nonlinear @xmath13 model , thus enabling a calculation for @xmath216 .
we are grateful to c. w. j. beenakker , m. leadbeater , and r. a. smith for useful discussions .
work in birmingham has been supported by epsrc grants gr / j35238 and gr / k95505 .
the calculation of the diagrams in fig . [ fig:4 ] is straightforward in the representation ( _ b _ ) and ( _ d _ ) .
the contribution of diagram ( _ b _ ) is given by @xmath217 here the overall factor of @xmath218 takes into account that diagram ( 4_b _ ) originates from @xmath219 , and there are four ways to connect @xmath220 and @xmath221 with an inter - sample diffuson which is clear from the equivalent representation of this diagram , fig .
[ fig:4]_a_. the constant @xmath222 is contributed by a triangle consisting of three green s functions .
this is a constant since , similarly to the calculation of the mean conductance , eq .
( [ meang ] ) , the triangle may be approximated at the diffusive scale , @xmath123 , by @xmath223 .
the presence of the @xmath224 functions ensures that the contribution of this diagram , eq .
( [ 4b ] ) , reduces to a mere product of the local contributions , @xmath225 and @xmath222 , and the diffusons , where the constant @xmath226 is given by @xmath227 at the same scale , the spatial arguments in @xmath228 coincide within an accuracy of order @xmath25 .
therefore , @xmath229 by the definition of eq .
( [ dr ] ) with @xmath23 given by eq .
( [ zr ] ) . substituting all the constants into eq .
( [ 4b ] ) , we obtain the contribution of diagram ( _ b _ ) as follows @xmath230 ^ 2 \!\ !
\zeta_{\phi}(\ell ) = { \zeta_{\phi}(\ell ) } \,\left <
g\right>^2 \ , .
\label{diagb}\ ] ] the contribution of diagram ( 4_d _ ) is given by @xmath231 here @xmath232 is contributed by the `` dressed '' inner square which is the sum of one `` bare '' square and two squares containing one single - impurity line each , this line connecting opposite sides of the square.@xcite the factor of 4 arises because diagram ( 4_d _ ) originates from @xmath160 and there are only two ways to connect the conductance loops with two inter - sample diffusons which is clear from the equivalent representation , fig .
[ fig:4]_c_. the calculation is more transparent in the ergodic zero - mode regime , @xmath233 , when not only the intrinsic diffuson , eq .
( [ zr ] ) , but also the inter - sample diffuson , eq . ( [ esc ] ) , and the dressed square , @xmath232 , reduce to a constant . since the contribution of a single - impurity line is given by the correlator ( [ rp ] ) , one finds @xmath232 as follows @xmath234 the first term is calculated in a similar manner to that in eq .
( [ chi2 ] ) which gives @xmath235 .
the integral in the second term coincides with that in eq .
( [ chi3 ] ) so that one obtains @xmath236 the difference between the one - particle relaxation time , @xmath237 , and the mean elastic scattering time , @xmath38 , is crucial in this calculation ; the same difference is a formal cause of the saturation of the inter - sample diffuson , eq .
( [ dr ] ) , for @xmath133 at @xmath238 . on substituting the zero - mode values of @xmath232 , eq .
( [ chi4 ] ) , and @xmath239 ) , one reduces the integral in eq .
( [ 4d ] ) to @xmath240 so that the dressed square , @xmath232 , exactly cancels one of the inter - sample diffusons .
then , on substituting eq .
( [ int1 ] ) and all the constants into eq .
( [ 4d ] ) , one finds that in the zero - mode case ( when @xmath241 ) , this contribution equals that of diagram ( 4_b _ ) , eq .
( [ diagb ] ) . in eq .
( [ chi4 ] ) we have neglected the @xmath242 contribution to the one - particle relaxation time , @xmath243 . taken into account , this would lead to an additional contribution of diagram ( _ d _ ) proportional to @xmath244 rather than @xmath245 as in eq .
( [ diagb ] ) . in this contribution
, the inner square cancels one of the intrinsic diffusons .
overall , such a contribution of diagram ( _ d _ ) would be exactly equal to the contribution of the diagram similar to that in fig .
( [ fig:4]_b _ ) but made from one intrinsic and two inter - sample diffusons .
we neglect these contributions which are small even compared to those of the diagrams in fig .
( [ fig:3 ] ) .
we do not present here the calculation of @xmath246 in a general case , without the restriction to the zero - mode regime .
the calculation is standard@xcite and results in @xmath247 so that @xmath246 is proportional to the diffusion operator . then using the diffusion equation
one finds that @xmath248 which generalizes eq .
( [ int1 ] ) and shows that the contribution of diagram ( [ fig:4]_d _ ) in this case is also exactly equal to that of diagram ( [ fig:4]_b _ ) , eq .
( [ diagb ] ) .
finally , the overall coefficient for the contribution of all the leading diagrams equals @xmath249 , since in the orthogonal case ( @xmath32 ) an equal contribution is made by two equivalent cooperon diagrams .
this leads to the result of eq .
( [ vg2 ] ) .
a. m. chang , h. u. baranger , l. n. pfeiffer , k. w. west , and t. y. chang , phys .
rev . lett .
* 76 * , 1695 ( 1996 ) .
j. a. folk , s. r. patel , s. f. godijn , a. g. huibers , s. m. cronenwett , c. m. marcus , k. campman , and a. c. gossard , phys .
* 76 * , 1699 ( 1996 ) .
average green s functions @xmath250 include @xmath251 . in calculating petals or any boxes containing filled circles ( i.e. representing electron motion at the ballistic scale near the leads ) , one can safely substitute @xmath38 for @xmath237 . however , the difference between @xmath38 and @xmath237 is crucial in calculating ` dressed ' inner boxes without external vertices like the square in diagram 4@xmath10 , see appendix .
[ fig:3 ] fig .
[ fig:3 ] .
the lowest - order contributions to the variance .
wavy lines represent diffuson ladders which correspond to the inter - sample diffusons .
the relation between a wavy line and a ladder is the same as for intrinsic diffusons , fig .
[ fig:1 ] .
[ fig:4 ] fig . [ fig:4 ] .
the leading diagrammatic contribution to the variance . in the diagrams
_ b _ and _ d _ , the inter - sample diffusons are represented by wavy lines and the intrinsic diffusons by double lines .
[ fig:6 ] fig .
[ fig:6 ] .
solid curves show the conductance distribution for @xmath196 with dephasing rates @xmath197 and @xmath198 , calculated using the second term in eq .
( [ pointn ] ) with an arbitrarily chosen coefficient @xmath199 . for comparison
the dashed curves are gaussian distributions with the same mean and variance . | we consider the distribution of conductance fluctuations in disordered quantum dots with single channel leads . using a perturbative diagrammatic approach , valid for continuous level spectra
, we describe dephasing due to processes within the dot by considering two different contributions to the level broadening , thus satisfying particle number conservation . instead of a completely non - gaussian distribution , which occurs for zero dephasing , we find for strong dephasing that the distribution is mainly gaussian with non - universal variance and non - gaussian tails . 2 |
the study of the problem of motion in general relativity ( gr ) has been a central theme in testing its validity .
for instance , the post - newtonian program , computing gr corrections to the newtonian gravitational potential between massive bodies , is crucial to understanding gravity in our solar system . in
the past decade or more , the post - newtonian analysis of weak field gravity has also been developed to very high order in perturbation theory because of the need to model gravitational waves ( gws ) from inspiraling compact binaries , which is expected to be a major source for detectors like advanced ligo . at the same time
, the discovery of cosmic acceleration and its associated cosmological constant problem , has prompted many attempts to modify how gravity operates at large ( astrophysical ) length scales .
one such example is the family of scalar field theories known as galileons @xcite , building on interesting properties of a limit of the dvali - gabadadze - porrati model @xcite .
these scalar fields couple to the stress - energy of ordinary matter , and therefore alter large scale dynamics , but due to their self interactions they exhibit the vainshtein screening effect @xcite , such that masses close to the central source of gravity ( lying within a so - called vainshtein radius @xmath0 ) do not feel their presence , and thus allowing galileons to evade solar system tests of gr to date .
furthermore , models of this type have become even more interesting since it has been discovered that they may be extended in multiple ways to yield other new field theories @xcite with related attractive properties , and that they arise as a limit of ghost - free massive gravity @xcite in this paper , we wish to lay the groundwork for understanding analytically , and in some detail , the impact of such vainshtein screened scalar fields on the problem of motion . to achieve concrete results we will consider the cubic galileon theory about flat spacetime and couple the galileon field to the trace of the stress - energy of matter .
we will place a large mass @xmath8 at the origin of our coordinate system .
the central goal of this paper is to solve the retarded green s function of the linearized equations of motion of the galileon fluctuating around the background field sourced by @xmath8 . in a paper in preparation @xcite
, we use the static limit of our green s function here to investigate the conservative portion of the dynamics ; to compute the effective potential between well separated test masses orbiting around @xmath8 . in the current paper
we address the dissipative aspect of the dynamics : does motion of matter lying well within the vainshtein radius of @xmath8 produce galileon radiation that can carry energy - momentum away to infinity ?
this question arises as an issue of principle because the galileon is a massless scalar field , and one would expect the motion of sources of massless fields to create radiation .
yet , it is not clear how much radiation would actually be produced , because it could perhaps be suppressed by the vainshtein mechanism . beyond issues of principle , we are also motivated by the possibility that one can constrain modified gravity theories by demanding that the power loss from binary pulsar systems , such as the hulse - taylor binary psr b1913 + 16 , had better not deviate too far from the predictions of gr , since observations have confirmed the latter to high precision .
furthermore , we look forward to the prospect that , as already alluded to , within the next decade or so gw detectors may be able to directly listen in on the spacetime ripples generated by such compact binary systems .
once this is possible , we may hope to use these gw signals to search for or put further constraints on the existence of the vainshtein mechanism .
this requires that we develop a quantitative prediction of the galileon waves themselves , beyond just an estimate of the power loss through scalar emission .
even though the two comparable point mass galileon problem possibly requires different techniques to solve due to the importance of the nonlinearities of the field equations , the results of our current paper , which assume the existence of a very large central mass , may perhaps be seen as an approximation to the situation where the inspiraling binary consists of rather unequal masses , say @xmath9 .
yet another possible source of gws is the oscillations of neutron stars themselves ; as such , we will also consider a toy problem of surface waves on a spherical body stimulating galileon waves .
the outline of the paper is as follows . in section
, we set up our problem in a quantitative manner . in section
we summarize the results for the galileon retarded green s function about the background field of the central mass @xmath8 ; and following that in section we step through its derivation . in section
we describe how the linearized equations of motion of the galileon field propagating about the background sourced by @xmath8 is equivalent to a minimally coupled massless scalar in some curved spacetime .
then in section we move on to use the retarded galileon green s function to study two examples of radiative processes : galileon radiation produced by surface waves on an otherwise spherical central body , and that generated by @xmath10 point masses gravitationally bound to @xmath8 . in appendix , we work out both the curved and minkowski spacetime minimally coupled massless scalar analog of the linard - wiechert potentials in electromagnetism ; this spacetime calculation complements the frequency space one in section .
_ notation _ a few words on notation .
the speed of light is set to unity .
the galileon field is @xmath11 .
the background spacetime is ( 3 + 1)-dimensional minkowski , with metric in cartesian coordinates given by @xmath12\end{aligned}\ ] ] so that , @xmath13 in spherical coordinates @xmath14 , with @xmath15 $ ] and @xmath16 , the metric reads instead @xmath17 , \qquad x^\text{a } = ( \theta,\phi ) \ , \end{aligned}\ ] ] where @xmath18 ( with @xmath19 ) is the infinitesimal solid angle .
a hat on a variable representing spatial location , e.g. @xmath20 , denotes the unit vector , @xmath21 .
in particular , @xmath20 only depends on the spherical coordinate angles @xmath22 $ ] .
finally , the planck mass is defined in terms of newton s gravitational constant @xmath23 as @xmath24 .
we would like to understand galileon radiative processes taking place in the background galileon field @xmath25 $ ] generated by a massive central body of mass @xmath8 , which we would take to be static ( time independent ) and spherically symmetric . to model this mass @xmath8
we shall treat it as a point particle at rest , located at the spatial origin @xmath26 of the coordinate system .
the galileon radiation we are investigating is generated by matter ( described by stress - energy tensor @xmath27 ) lying well within the vainshtein radius @xmath0 of the central mass .
( the vainshtein radius @xmath0 , as we will see very shortly , is the radius below which the galileon field @xmath28 generated by the mass @xmath8 is increasingly governed by non - linear self interactions ; well outside vainshtein , the theory is linear and there @xmath8 generates a @xmath29 coulomb potential . ) this matter distribution @xmath27 is meant to be viewed as a perturbation relative to the mass @xmath8 , but can otherwise be arbitrary .
for instance @xmath27 may describe slight deviation of the mass @xmath8 from an exact spherical configuration ( its multipole moments ) , and/or a deviation from time independence ; it could also be the energy - momentum of @xmath10 light compact bodies , with masses @xmath30 , orbiting around @xmath8 .
we will take for the galileon theory the simplest cubic @xmath11 lagrangian ; in actuality , several other terms are allowed in 4 spacetime dimensions .
the galileon , being a scalar , couples to the trace of the stress - energy tensor of matter .
the total action for our setup is therefore @xmath31 where , in cartesian coordinates , @xmath32 , \\
\label{action_perturbations } \delta s & \equiv \int { \text{d}}^4 x \frac{\pi}{{m_{\text{pl } } } } \delta t , \qquad \delta t
\equiv \delta t^\mu_{\phantom{\mu}\mu}.\end{aligned}\ ] ] because we are interested in radiation that can propagate to infinity , we note that far away from the source @xmath8 , contributions to the galileon stress - energy tensor @xmath33 due to the cubic - in-@xmath11 terms in @xmath34 fall away more rapidly than their quadratic counterparts ( provided , of course , that @xmath11 and its gradients falls off with increasing @xmath35 ) .
expressed in cartesian coordinates , the asymptotic galileon energy - momentum tensor is thus that of the non - interacting massless scalar in flat spacetime , @xmath36 = \partial_\mu \pi \partial_\nu \pi - \frac{1}{2 } \eta_{\mu\nu } ( \partial\pi)^2 \ .\end{aligned}\ ] ] the background galileon field sourced by the mass @xmath8 is the exact solution to @xmath37 .
since we have a static spherically symmetric source , @xmath28 must only depend on @xmath35 , and so the euler - lagrange equation from varying @xmath38 becomes @xmath39\end{aligned}\ ] ] integrating both sides of over a sphere of radius @xmath35 centered at @xmath26 then yields @xmath40 ( even though we have modeled the central body as a point mass , is valid outside any isolated static spherically symmetric matter distribution as long as we replace @xmath8 with the integral @xmath41 r'^2 { \text{d}}r'$ ] , where @xmath42 is the stress - energy tensor of the body . )
is a quadratic equation in @xmath43 $ ] . defining the vainshtein radius @xmath44 we obtain @xmath45}{\lambda^3 } = \frac{r}{8 } \left ( 1 - \sqrt { 1 + \left(\frac{r_v}{r}\right)^3 } \right).\end{aligned}\ ] ] there is a second solution for @xmath43 $ ] in which the negative sign in front of the square root is replaced with a plus sign .
this solution is proportional to @xmath35 and hence blows up as @xmath46 .
since the stress - tensor depends on @xmath43 $ ] see eq . and note that including the contributions to @xmath33 from the cubic - in-@xmath11 terms in @xmath34 would only exacerbate the problem we may discard this second solution on the grounds that the energy - momentum of @xmath28 measured by an asymptotic observer can not be infinite .
it is possible to integrate eq .
exactly in terms of hypergeometric functions , @xmath47}{\lambda^3 } = \frac{r^2}{16 } \left ( 1 - 4 \left(\frac{r_v}{r}\right)^{\frac{3}{2 } } \,_2 f_1\left [ -\frac{1}{2 } , \frac{1}{6 } ; \frac{7}{6 } ; -\frac{r^3}{r_v^3 } \right ] \right ) - \frac{\gamma\left [ -\frac{2}{3 } \right ]
\gamma\left [ \frac{7}{6 } \right]}{8 \sqrt{\pi } } r_v^2 \ , \end{aligned}\ ] ] where we have chosen the asymptotic boundary condition to be @xmath48 = 0 $ ] . for later use ,
note the following limits of equations and .
when @xmath49 , @xmath50 & = -\frac{m}{{m_{\text{pl}}}\pi r_v } \left ( \frac{\gamma[-\frac{2}{3 } ] \gamma[\frac{7}{6}]}{2\sqrt{\pi } } + \sqrt{\frac{r}{r_v } } + \mathcal{o}\left [ \left(\frac{r}{r_v}\right)^2 \right ] \right),\end{aligned}\ ] ] and @xmath51 & = - \frac{\lambda^3 r}{8 } \left ( \left(\frac{r_v}{r}\right)^{3/2 } - 1 + \mathcal{o}\left [ \left(\frac{r}{r_v}\right)^{\frac{3}{2 } } \right ] \right ) , \\
\label{backgroundpisolution_smallradius_iiofii } { \overline{\pi}}''[r ] & = \frac{\lambda^3}{16 } \left ( \left(\frac{r_v}{r}\right)^{3/2 } + 2 + \mathcal{o}\left [ \left(\frac{r}{r_v}\right)^{\frac{3}{2 } } \right ] \right).\end{aligned}\ ] ] on the other hand , when @xmath52 , @xmath53 & = \frac{m}{4 \pi { m_{\text{pl}}}r } \left ( 1 + \mathcal{o}\left
[ \left(\frac{r_v}{r}\right)^3 \right ] \right),\end{aligned}\ ] ] and @xmath54 & = - \frac{\lambda^3 r}{16 } \left ( \left(\frac{r_v}{r}\right)^3 + \mathcal{o}\left [ \left(\frac{r_v}{r}\right)^6 \right ] \right ) , \\
\label{backgroundpisolution_largeradius_iiofii } { \overline{\pi}}''[r ] & = \frac{\lambda^3}{8 } \left ( \left(\frac{r_v}{r}\right)^3 + \mathcal{o}\left [ \left(\frac{r_v}{r}\right)^6 \right ] \right).\end{aligned}\ ] ] notice that the first term of the small radius limit @xmath43 \sim 1/\sqrt{r}$ ] in can be obtained by dropping the linear term @xmath55 $ ] in eq . , whereas the first term in of the large radius limit @xmath43 \sim 1/r^2 $ ] , which is the usual coulomb force law , comes about from dropping the non - linear @xmath56 piece in eq . .
this is the quantitative statement that , close to the matter source ( @xmath49 ) , the dynamics of galileons are primarily governed by their nonlinear self interactions .
the theory is linear when the observer is well outside the vainshtein radius . more physically , to see the vainshtein effect at work , imagine a test point mass @xmath57 , at spatial location @xmath58 , orbiting the central body .
its action takes the same form as @xmath59 in eq .
, except we evaluate the galileon field about the background @xmath28 generated by @xmath8 , @xmath60\right].\end{aligned}\ ] ] ( strictly speaking , lorentz invariance says @xmath61 needs to be replaced with @xmath62 ^ 2}$ ] ; however , we are working in the non - relativistic regime , where the square root is very close to unity . ) by employing the definition of the vainshtein radius in eq .
, followed by integrating the first term on the right - hand - sides of equations and without worrying too much about the overall numerical factors , @xmath63| } \left(\frac{|\vec{z}[t]|}{r_v}\right)^{\frac{3}{2 } } , & |\vec{z}[t]| \ll r_v \\ & \sim \int
{ \text{d}}t \frac{g_\text{n } m m}{|\vec{z}[t]| } \ , & |\vec{z}[t]| \gg r_v \end{aligned}\ ] ] thus , the galileon potential experienced by a test point mass orbiting close to the central mass @xmath8 is the newtonian gravitational potential @xmath64 multiplied by a suppression factor of @xmath65 .
only when the test mass travels well outside vainshtein does the suppression factor drop out and the galileon potential become comparable in strength to that of regular gravity . with the exact background solution @xmath43 $ ] in hand
, we now substitute @xmath66 = { \overline{\pi}}[r ] + \varphi[x]\end{aligned}\ ] ] in , and drop terms cubic - in-@xmath67 . the resulting linearized equation of motion of @xmath67 about the background @xmath28 reads @xmath68 = \frac{\delta t[x]}{{m_{\text{pl } } } } \ , \end{aligned}\ ] ] where the differential operator @xmath69 is @xmath70 \equiv \left ( e_1 \partial_t^2 - e_2 \partial_r^2 - \frac{2}{r } e_3 \partial_r - \frac{1}{r^2 } e_3 \vec{l}^2 \right ) \varphi[x]\end{aligned}\ ] ] and @xmath71 is the angular part of the laplacian in euclidean 3-space , usually called the negative of the `` angular momentum squared '' operator , given by @xmath72 here @xmath73 & \equiv 1 - \frac{8 \overline{\pi}'}{\lambda^3 r } - \frac{4 \overline{\pi}''}{\lambda^3 } = \frac{1}{4 } \left ( 3\sqrt{\frac{(2 r^3 + r_v^3)^2}{r^3(r^3+r_v^3 ) } } -2 \right ) , \\ \label{e2 } e_2[r ] & \equiv 1 - \frac{8 \overline{\pi}'}{\lambda^3 r } = \sqrt{1 + \left(\frac{r_v}{r}\right)^3 } , \\
\label{e3 } e_3[r ] & \equiv 1 - \frac{4 \overline{\pi}'}{\lambda^3 r } - \frac{4 \overline{\pi}''}{\lambda^3 } = \frac{1}{4}\sqrt{\frac{(4 r^3 + r_v^3)^2}{r^3(r^3+r_v^3)}}.\end{aligned}\ ] ] if we assume @xmath74 does not implicitly depend on @xmath11 , then @xmath67 is entirely sourced by @xmath74 , the ( for now arbitrary ) matter perturbations .
the key to solving @xmath67 in terms of @xmath74 is the retarded green s function defined by the equation , @xmath75 = \mathcal{w}_{x ' } g[x , x ' ] = \frac{1}{rr ' } \delta[t - t ' ] \delta[r - r']\delta[\phi-\phi']\delta[\cos\theta-\cos\theta ' ] , \end{aligned}\ ] ] where the @xmath76s are the dirac delta functions .
the solution to the linearized galileon equation about the static spherically symmetric background @xmath28 is now ( in cartesian coordinates ) @xmath77 = \int { \text{d}}^4 x ' g[x , x ' ] \frac{\delta t[x']}{{m_{\text{pl}}}}.\end{aligned}\ ] ] we write @xmath78 to represent a collective label for @xmath14 and @xmath79 for @xmath80 , so that the requirement that the signal does not precede the turning on of the source requires @xmath81 = 0 $ ] for @xmath82 .
note that it is not obvious that the appropriate spacetime dependence multiplying the @xmath76-functions in the green s function equation is @xmath83 , and therefore we will justify this in appendix below . when solving eq .
it is important to remember the following boundary condition . since the green s function is the field generated by a unit point mass , if we let @xmath74 describe a static point mass sitting at the origin , @xmath84
\equiv \delta m \delta^{(3)}[\vec{x } ' ] , \qquad \delta m / m \ll 1,\end{aligned}\ ] ] this merely amounts to shifting the mass of the central body by @xmath85 .
then we already know what to expect from the linear solution represented by the integral in eq .
. it should be the linear - in-@xmath86 piece of the full @xmath25 $ ] solution in eq . upon the replacement @xmath85 .
remember that the mass dependence in the full solution of eq .
is contained entirely in @xmath0 via eq . .
perturbing @xmath85 in eq . up to linear order in @xmath86 yields , @xmath87 = { \overline{\pi}}[r;m ] + \delta { \overline{\pi}}[r ] , \end{aligned}\ ] ] with @xmath88 = \frac{\delta m}{{m_{\text{pl } } } } \frac{1}{2\pi r_v } \left ( \frac{\gamma\left[\frac{1}{3}\right ] \gamma\left[\frac{1}{6}\right]}{6\sqrt{\pi } } - \sqrt{\frac{r}{r_v } } \,_2f_1\left[\frac{1}{6},\frac{1}{2};\frac{7}{6};-\frac{r^3}{r_v^3}\right ] \right ) .\end{aligned}\ ] ] hence , when solving @xmath81 $ ] below , we must obtain from eq .
, @xmath89 = \frac{\delta m}{{m_{\text{pl } } } } \int_{-\infty}^\infty { \text{d}}t g[x , x ' ] = \delta { \overline{\pi}}[r ] , \qquad x \equiv ( t,\vec{x } ) ; \ x ' \equiv ( t',\vec{0}).\end{aligned}\ ] ] ( the second equality follows from the time translation symmetry of the problem at hand . ) for later use let us note that the small and large radius limits are , respectively , @xmath90 \to \frac{1}{2\pi r_v } \left ( \frac{\gamma\left[\frac{1}{3}\right ] \gamma\left[\frac{1}{6}\right]}{6\sqrt{\pi } } - \sqrt{\frac{r}{r_v } } \right ) , \qquad r \to 0\end{aligned}\ ] ] and ( using eq .
below ) @xmath91
\to \frac{1}{4\pi r } , \qquad r\to\infty .\end{aligned}\ ] ]
in this section we solve , in different limits of physical interest , the galileon retarded green s function about the spherically symmetric background @xmath25 $ ] in equation .
the first subsection summarizes all the results in a coherent manner , and in subsequent subsections we step through the derivation systematically . our solution of the retarded galileon green s function @xmath81 $ ] obeying equation is composed of an integral over all angular frequencies @xmath92 and an infinite mode sum over all harmonics @xmath93 : @xmath94 & = \int_{-\infty}^{+\infty } \frac{{\text{d}}\omega}{2\pi } e^{-i\omega(t - t ' ) } \omega \sum_{\ell = 0}^\infty \widetilde{g}_\ell\big[\omega r , \omega r'\big ] \sum_{m=-\ell}^{+\ell } y_\ell^m[\theta,\phi ] \overline{y_\ell^m}[\theta',\phi'].\end{aligned}\ ] ] we will make frequent use of the dimensionless variables @xmath95 because we will shortly show that the radial green s function @xmath96 depends on @xmath35 , @xmath97 and @xmath0 solely through @xmath98 , @xmath99 and @xmath100 respectively . in eq .
, the @xmath101s are the usual spherical harmonics spanning a complete set of functions defined on a sphere of unit radius embedded in 3 spatial dimensions ( the over bar means complex conjugation ) ; they obey the eigenvalue equation @xmath102 because the background @xmath28 is static , the green s function reflects the time translation symmetry of the setup at hand .
moreover , spherical symmetry tells us the radial green s functions @xmath96 do not depend on the azimuthal number @xmath103 .
the separation of variables method of mode expansion employed in eq .
reduces the problem of solving the linear partial differential equation for @xmath81 $ ] in eq . to a linear second order ordinary differential equation ( ode ) for the radial green s function @xmath96 . inserting the ansatz in eq . into eq . , and
using a fourier representation of @xmath104 $ ] and the completeness relation for the spherical harmonics , one may read off the ode for @xmath96 in frequency space by equating the coefficient of @xmath105 y_\ell^m[\theta,\phi ] \overline{y_\ell^m}[\theta',\phi']$ ] on both sides of the green s function equation .
we then arrive at @xmath106 = \frac{\delta[\xi-\xi']}{\xi\xi ' } .\end{aligned}\ ] ] we have carried out a change of variables according to the rules in eq . ; for the @xmath107 this amounts to simply replacing every @xmath35 variable with its corresponding @xmath98 variable .
no longer depends explicitly on the radii nor on the angular frequency @xmath92 means that the solution for the radial green s function @xmath96 , can not depend on the lengths @xmath108 or frequency @xmath92 explicitly .
the radial green s function @xmath96 has a discontinuous first derivative at @xmath109 because its second derivatives at @xmath110 needs to yield @xmath111/(\xi\xi')$ ] .
therefore we need to distinguish between the two regions @xmath112 or @xmath113 .
we therefore let @xmath114 and @xmath115 represent the larger and smaller of the two radii @xmath35 and @xmath97 , and define @xmath116 we now proceed to summarize the results for @xmath96 .
_ radiative limit _ of central importance in this paper , is the situation where the emitter lies deep inside the vainshtein radius while the observer sits far outside ( @xmath117 ) . taking the limit where one of the radii is much smaller than @xmath0 and the other much larger ,
more specifically @xmath118 and @xmath119 , we obtain @xmath120 & = & \left\ { \begin{array}{lr } h_\ell^{(1)}[\xi _ > ] c_0^\text{(rad)}[\xi_v ] \sqrt[4]{\xi _ < } j_{-\frac{1}{4 } } \left [ \sqrt{3}\xi_</2 \right ] & \ell=0 \\ & \\ h_\ell^{(1)}[\xi _ > ] c_\ell^\text{(rad)}[\xi_v ] \sqrt[4]{\xi _ < } j_{\frac{1}{4}(2\ell+1)}\left [ \sqrt{3}\xi_</2 \right ] & \ell>0 \\ \end{array } \right .
\ , \end{aligned}\ ] ] where @xmath121 is the bessel function and the @xmath122 is the spherical hankel function of the first kind . in the high frequency regime @xmath123 ,
the coefficients @xmath124 are @xmath125 where @xmath126 } \right ) { \text{d}}\vartheta \approx 0.253 \ , \end{aligned}\ ] ] and @xmath127 \equiv -\frac{1}{4 } \frac{3(1 + 2\vartheta^3 ) - 2\sqrt{\vartheta^3(1+\vartheta^3)}}{1+\vartheta^3 } \ .\ ] ] and is plotted in fig .
the suppression factor of @xmath128 in indicates that high frequency galileon signals are indeed vainshtein screened , at least for small @xmath129s . }
$ ] ( solid line ) see eq .
as a function of the ratio @xmath130 .
the asymptotics are : @xmath131 } = \sqrt{3}/2 $ ] ( long - dashed line ) and @xmath132 } = 1 $ ] ( short - dashed line ) .
the turning point , which is a global minimum , is at @xmath133 } = \sqrt{2/3}$].,width=307 ] in the low frequency regime , @xmath134 , the coefficients @xmath124 are @xmath135{3 } \pi}{\gamma[\frac{1}{4 } ] } & \ell=0 \\ & \\
i \sqrt{\pi } \xi_v^{\frac{\ell-1}{2 } } \frac{\gamma\left[-\frac{2 \ell } { 3}-\frac{1}{3}\right ] \gamma\left[\frac{5}{6}-\frac{\ell } { 3}\right]}{2^{2\ell+1 } 3^{\frac{\ell}{4}+\frac{1}{8 } } \gamma \left[\frac{\ell}{2}+\frac{3}{4}\right ] \gamma [ -\ell ] } \left(1 + \frac{\cos\left[\frac{1}{6 } ( 2\ell+1)\pi\right]}{\sin[\pi\ell ] } \right ) & \ell>0 \\ \end{array } \right .
\ , \end{aligned}\ ] ] it is worth pointing out that although @xmath136 $ ] appears in the denominator , and this expression contains @xmath137-functions whose arguments appear they could be negative integers ; these @xmath138 terms are all in fact non - singular .
( this remark also applies to the related equation eq . below . )
we list the first ten @xmath124s here , to 3 significant figures : [ cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ] notice from equation that the @xmath139 modes do not contain @xmath0 ; this is a direct consequence of the fact that the leading order terms of the static green s function in this same @xmath140 limit ( eq . )
do not contain any @xmath0 for @xmath139 .
in fact , we may further compare the cubic galileon radial green s function @xmath96 with its non - interacting massless cousin @xmath141 j_\ell[\xi_<]$ ] ( see eq . and below ) .
let us consider the non - relativistic limit @xmath142 , where the reciprocal of the characteristic frequency of the motion is much smaller than the characteristic distance of the source to the central mass @xmath8 . in this limit
, we may replace the bessel functions with their small argument limits , and find @xmath143 , \qquad \qquad
\widetilde{g}_1 = \widetilde{g}^\text{(flat)}_1 = \frac{i}{3 } \xi _ < h_\ell^{(1)}[\xi_>]\end{aligned}\ ] ] and for @xmath144 , @xmath145 & \to \kappa_\ell \xi_v^\ell \left(\frac{r_<}{r_v } \right)^{\frac{\ell+1}{2 } } h_\ell^{(1)}[\xi _ > ] \\ \widetilde{g}^\text{(flat)}_\ell[r _
> \gg r_v \gg r _ < ] & \to \kappa'_\ell \xi_v^\ell \left(\frac{r_<}{r_v } \right)^\ell h_\ell^{(1)}[\xi _ > ] .
\nonumber\end{aligned}\ ] ] here , @xmath146 and @xmath147 are constants that depend solely on @xmath129 .
this teaches us that , while high frequency galileon power loss is vainshtein screened , low frequency signals generated from deep within the vainshtein radius of @xmath8 are comparable to or even vainshtein enhanced relative to the non - interacting massless scalar
. the results of the radiative processes described in section will reflect these observations . _ the wkb high frequency limit _
when @xmath148 $ ] , we may apply the wkb approximation .
let us first define @xmath149 } { \text{d}}\vartheta.\ ] ] where it is worth noting that @xmath131 } = \sqrt{3}/2 $ ] and @xmath132 } = 1 $ ] .
the leading order wkb solution is @xmath150 & = \left ( u[\xi/\xi_v ] u[\xi'/\xi_v ] \xi(\xi^3+\xi_v^3 ) \xi'(\xi'^3+\xi_v^3 ) \right)^{-\frac{1}{4 } } \\ & \qquad \qquad \times \left ( \frac{i}{2 } \exp\left [ i \xi_v \left ( \phi _ > - \phi _ <
\right ) \right ] + \frac{c_\ell^{++}}{2 } \exp\left [ i \xi_v \left ( \phi _ > + \phi _ < \right ) \right ] \right ) \ , \nonumber\end{aligned}\ ] ] where @xmath151 when @xmath152 \ll |\xi_v| \ll |\xi| , |\xi'|$ ] , the radial green s functions become @xmath153 & = \frac{1}{\xi \xi ' } \left ( \frac{i}{2 } \exp\left [ i \left ( \xi _ > - \xi _ <
\right ) \right ] + \frac{c_\ell^{++}}{2 } \exp\left [ i \left ( \xi _ > + \xi _ < - 2 \xi_v \mathcal{i}_\infty \right ) \right ]
\right ) \ , \end{aligned}\ ] ] whereas when @xmath154 they become @xmath155 & = \frac{2}{\sqrt{3 } \xi_v^{3/2}\sqrt[4]{\xi\xi ' } } \left ( \frac{i}{2 } \exp\left [ i\frac{\sqrt{3}}{2 } \left ( \xi _ > - \xi _ <
\right ) \right ] + \frac{c_\ell^{++}}{2 } \exp\left [ i \frac{\sqrt{3}}{2 } \left ( \xi _ > + \xi _ <
\right ) \right ] \right ) .\end{aligned}\ ] ] in this limit , observe that the @xmath114 and @xmath156 dependent portion of the combination @xmath157 in the mode expansion eq .
is @xmath158{r _ > } } \exp\left[i \frac{\sqrt{3}}{2 } \omega \left ( r _ > - \frac{2}{\sqrt{3 } } ( t - t ' ) \right)\right],\end{aligned}\ ] ] and if we imagine a source located much closer to @xmath8 than the observer is ( @xmath159 ) the green s function tells us the observer will receive purely outgoing radial waves .
it is not entirely clear from the outset that this would be the case , particularly viewed from the curved spacetime picture ( which we describe in section below ) , because one may think that the radiation from the source could backscatter off the spacetime geometry and return to the observer , thereby mimicking an ingoing radial wave .
( this scenario may in fact occur in the low frequency limit , where the longer wavelength of the galileon waves may grow more sensitive to the curvature of the background effective geometry . )
the phase @xmath160 also indicates these outgoing waves , if they are propagating only in the radial direction , are superluminal because @xmath161 .
of particular importance is the wkb radiative limit , @xmath140 .
here the radial green s function is @xmath162 & = & \left\ { \begin{array}{lr } \frac{\sqrt{2}i}{\xi _ > \sqrt[4]{3
\xi _ < } \xi_v^{3/4 } } e^{i\left ( \xi _ > - \xi_v \mathcal{i}_\infty - \frac{\pi}{8}\right ) } \cos\left [ \frac{\sqrt{3}}{2 } \xi _ < - \frac{\pi}{8 } \right ] & \ell=0 \\ & \\ \frac{\sqrt{2}i}{\xi _ > \sqrt[4]{3 \xi _
< } \xi_v^{3/4 } } e^{i\left ( \xi _ > - \xi_v \mathcal{i}_\infty + \pi\frac{5 - 2\ell}{8}\right ) } \cos\left [ \frac{\sqrt{3}}{2 } \xi _ < + \pi\frac{5 - 2\ell}{8 } \right ] & \ell>0 \\ \end{array } \right . \ .\end{aligned}\ ] ] _ large mode number _ it is useful to note that the asymptotic expansion of the bessel function @xmath163 $ ] exhibits an exponential suppression at large order @xmath164 , for @xmath165 .
therefore , the presence of @xmath166 $ ] in eq .
means , at least for the radiation problem ( @xmath140 ) the main object of this paper we may neglect mode numbers much larger than @xmath167 .
_ the static limit _ we may also obtain the zero frequency ( static ) limit of the green s function , defined as @xmath168 = \int_{-\infty}^{+\infty } g\left[x \equiv ( t,\vec{x } ) , x ' \equiv ( t',\vec{x}')\right ] { \text{d}}t\end{aligned}\ ] ] ( it does not actually matter , because of time translation symmetry of the situation , whether we integrate with respect to @xmath169 or @xmath170 in eq . . ) as the name suggests , the static green s function does not depend on time .
we may interpret @xmath171 $ ] as the galileon potential between two static point sources , both of unit mass , in the background @xmath25 $ ] .
putting the mode expansion in eq . into the integral over all time in eq .
yields the mode expansion for the static green s function @xmath172 = \lim_{\omega\to 0 } \omega \sum_{\ell = 0}^\infty \widetilde{g}_\ell[\xi,\xi ' ] \sum_{m=-\ell}^{+\ell } y_\ell^m[\theta,\phi ] \overline{y_\ell^m}[\theta',\phi ' ] .\end{aligned}\ ] ] the exact result can be expressed in terms of hypergeometric functions as @xmath173 & = \frac{1}{2\pi r_v } \left ( \frac{\gamma\left[\frac{1}{3}\right ] \gamma\left[\frac{1}{6}\right]}{6\sqrt{\pi } } - \sqrt{\frac{r_>}{r_v } } \,_2f_1\left[\frac{1}{6},\frac{1}{2};\frac{7}{6};-\frac{r_>^3}{r_v^3}\right ] \right ) \nonumber\\ & + \frac{1}{r_v } \sum_{\ell = 1}^\infty \sum_{m = -\ell}^\ell \frac { y_\ell^m[\theta,\phi ] \overline{y_\ell^m}[\theta',\phi ' ] } { 2\ell+1 } \left ( \frac{r_<}{r_v } \right)^{\frac{\ell+1}{2 } } \ _ 2f_1\left [ \frac{1}{6}-\frac{\ell}{6 } , \frac{1}{2}+\frac{\ell}{2 } ; \frac{7}{6}+\frac{\ell}{3 } ; -\frac{r_<^3}{r_v^3 } \right ] \nonumber\\ & \times \bigg ( 2 \left ( \frac{r_v}{r _ > } \right)^{\frac{\ell}{2 } } \ _ 2f_1\left [ \frac{\ell}{6}+\frac{1}{3 } , -\frac{\ell}{2 } ; \frac{5}{6}-\frac{\ell}{3 } ; -\frac{r_>^3}{r_v^3 } \right ] \\ & \qquad \qquad \qquad \qquad + \frac{\ell ! \gamma\left [ -\frac{1}{6}(2\ell+1 ) \right]}{\sqrt{\pi } \gamma\left[\frac{1}{3}(2\ell+1)\right ] } \left ( \frac{r_>}{r_v } \right)^{\frac{\ell+1}{2 } } \ _
2f_1\left [ \frac{1}{6}-\frac{\ell}{6 } , \frac{1}{2}+\frac{\ell}{2 } ; \frac{7}{6}+\frac{\ell}{3 } ; -\frac{r_>^3}{r_v^3 } \right ] \bigg ) \ . \nonumber\end{aligned}\ ] ] via the identity in eq .
below , the @xmath174 mode ( the first line on the right hand side of eq . )
is equivalent to @xmath175 .\end{aligned}\ ] ] also , since @xmath176 & = \frac{\gamma[\alpha+\beta-\gamma]\gamma[\gamma]}{\gamma[\alpha]\gamma[\beta ] } ( 1-z)^{\gamma-\alpha-\beta } \,_2 f_1[\gamma-\alpha,\gamma-\beta;\gamma-\alpha-\beta+1;1-z ] \nonumber\\ & \qquad \qquad + \frac{\gamma[\gamma-\alpha-\beta]\gamma[\gamma]}{\gamma[\gamma-\alpha]\gamma[\gamma-\beta ] } \,_2 f_1[\alpha,\beta;\alpha+\beta-\gamma+1;1-z ] , \end{aligned}\ ] ] and using that @xmath177 = 0 $ ] if @xmath103 is a positive integer or zero , one of the static mode functions can be written as , for @xmath129 even , @xmath178 & = \sqrt{\pi } \left ( \frac{r_v}{r } \right)^{\frac{\ell}{2 } } \frac{\gamma\left [ \frac{5 - 2\ell}{6 } \right]}{\gamma\left [ -\frac{\ell-1}{2 } \right ] \gamma\left [ \frac{\ell+5}{6 } \right ] } \,_2f_1\left [ -\frac{\ell}{2 } , \frac{\ell+2}{6 } ; \frac{1}{2 } ; 1 + \frac{r^3}{r_v^3 } \right],\end{aligned}\ ] ] and , for @xmath129 odd , @xmath178 & = -2 \sqrt{1+\frac{r^3}{r_v^3 } } \sqrt{\pi } \left ( \frac{r_v}{r } \right)^{\frac{\ell}{2 } } \frac{\gamma\left [ \frac{5 - 2\ell}{6 } \right]}{\gamma\left [ -\frac{\ell}{2 } \right ] \gamma\left [ \frac{\ell+2}{6 } \right ] } \,_2f_1\left [ -\frac{\ell-1}{2 } , \frac{\ell+5}{6 } ; \frac{3}{2 } ; 1 + \frac{r^3}{r_v^3 } \right ] .\end{aligned}\ ] ] ( note that @xmath179 $ ] and @xmath180 $ ] are , respectively , @xmath181th ( even @xmath129 ) and @xmath182th ( odd @xmath129 ) order polynomials in the variable @xmath183 . ) taking the limit @xmath184 hands us the green s function to the laplacian in euclidean 3-space @xmath185 & = \frac{1}{4\pi |\vec{x } - \vec{x}'|},\end{aligned}\ ] ] plus corrections that begin at relative order @xmath186 and @xmath187 .
this is to be expected , since far outside the vainshtein radius , the central mass becomes irrelevant and we ought to recover the theory of a massless scalar in flat spacetime .
next , taking the limit @xmath188 leads us to @xmath189 & = \frac{1}{2\pi r_v } \left ( \frac{\sqrt{rr'/r_v^2}}{\left\vert \sqrt{r / r_v}\widehat{x } - \sqrt{r'/r_v}\widehat{x } ' \right\vert } - \sqrt{\frac{r}{r_v } } - \sqrt{\frac{r'}{r_v } } + \frac{\gamma\left[\frac{1}{3}\right ] \gamma\left[\frac{1}{6}\right]}{6\sqrt{\pi } } \right ) , \end{aligned}\ ] ] a result obtained independently in @xcite , and which will play a key role in our field theory based analysis of the conservative portion of the galileon two body problem taking place in the background field @xmath28 of @xmath8 . here
@xmath22 $ ] and @xmath190 $ ] are the unit radial vectors of the observer and source , respectively , and vertical bars denote the euclidean length . we have expressed every occurrence of the two radii in eq . as a small ratio , @xmath130 or @xmath191 , to highlight the vainshtein mechanism at work . in section below , where we shall view the galileon propagating on the background @xmath28 as a minimally coupled massless scalar propagating in a particular curved spacetime
, we shall re - derive eq . in an alternate manner . for @xmath140
, we obtain @xmath192 = \frac{1}{4\pi r _ > } + \frac{1}{r _ > } \sqrt{\frac{r_<}{r_v } } & \sum_{\ell = 1}^\infty \sum_{m=-\ell}^\ell \frac{y_\ell^m\left [ \theta,\phi \right ] \overline{y_\ell^m } \left [ \theta',\phi ' \right]}{2\ell+1 } \left(\frac{\sqrt { r_v r _
< } } { 2 r_>}\right)^\ell \\ & \times \frac{\gamma \left[-\frac{2 \ell}{3}-\frac{1}{3}\right ] \gamma \left[\frac{5}{6}-\frac{\ell}{3}\right]}{\sqrt{\pi } \gamma[-\ell ] } \left(\frac{\cos \left[\frac{1}{6 } ( 2\ell+1)\pi\right]}{\sin[\pi\ell ] } + 1 \right ) \ , \nonumber\end{aligned}\ ] ] plus corrections that begin at order @xmath193 and @xmath194 relative to these displayed terms . notice for @xmath139 , @xmath0 drops out of these leading order terms .
this tells us for the monopole and dipole terms , when the wavelength of galileon signals are much longer than the vainshtein radius @xmath0 , the vainshtein mechanism becomes less effective .
_ well inside vainshtein _ let both the observer and emitter lie well inside the vainshtein radius , @xmath188
. in the low frequency limit , @xmath195 , the radial green s function becomes , @xmath196 & = \frac{\pi ( \xi \xi')^{\frac{1}{4}}}{2 \xi_v^{3/2 } } j_{-\frac{1}{4 } } \left[\sqrt{3}\xi_</2\right ] \left ( \frac{4 \sqrt[4]{3}\sqrt{\pi } \gamma[\frac{1}{3 } ] \gamma[\frac{7}{6}]}{\gamma[\frac{1}{4}]^2 } \sqrt{\xi_v } j_{-\frac{1}{4 } } \left[\sqrt{3}\xi_>/2\right ] - \sqrt{2 } j_{\frac{1}{4 } } \left[\sqrt{3}\xi_>/2\right ] \right ) \ , \end{aligned}\ ] ] for @xmath174 , and @xmath197 & = i \frac{\pi ( \xi \xi')^{\frac{1}{4}}}{2 \xi_v^{3/2 } } j_{\frac{1}{4}(2\ell+1 ) } \left[\sqrt{3}\xi_</2\right ] \bigg ( h^{(1)}_{\frac{1}{4}(2\ell+1 ) } \left[\sqrt{3}\xi_>/2\right ] \nonumber\\ & \qquad \qquad + \left ( \frac{2}{i(-)^\ell-1 } - \frac{i}{\xi_v^{\ell + \frac{1}{2 } } } \frac{4^{\ell + 1 } \gamma\left[-\frac{\ell}{3}-\frac{1}{6}\right ] \gamma\left[\frac{\ell}{2}+\frac{5}{4}\right]^2 \ell!}{3^{\frac{\ell}{2}+\frac{5}{4 } } \pi^{3/2 } \gamma\left[\frac{2 ( \ell + 2)}{3}\right ] } \bigg ) j_{\frac{1}{4}(2\ell+1 ) } \left[\sqrt{3}\xi_>/2\right ] \right ) , \end{aligned}\ ] ] for @xmath138 . in the wkb limit ,
i.e. @xmath198 $ ] , the radial green s function instead reads @xmath199 & = & \left\ { \begin{array}{lr } i \frac{\pi ( \xi \xi')^{\frac{1}{4}}}{2 \xi_v^{3/2 } } e^{i\frac{\pi}{4 } } h^{(1)}_{\frac{1}{4 } } \left[\sqrt{3}\xi_>/2\right ] j_{-\frac{1}{4 } } \left[\sqrt{3}\xi_</2\right ] & \ell=0 \\ & \\
i \frac{\pi ( \xi \xi')^{\frac{1}{4}}}{2 \xi_v^{3/2 } } h^{(1)}_{\frac{1}{4}(2\ell+1 ) } \left[\sqrt{3}\xi_>/2\right ] j_{\frac{1}{4}(2\ell+1 ) } \left[\sqrt{3}\xi_</2\right ] & \ell>0 \\
\end{array } \right .
\ .\end{aligned}\ ] ] _ well outside vainshtein _ when both observer and emitter lie well outside the vainshtein radius , @xmath184 , we recover the theory of a minimally coupled massless scalar in flat spacetime , with the radial green s function @xmath200 = i h_\ell^{(1)}[\xi _ > ] \left ( j_\ell[\xi _ < ] + c^{(hh)}_\ell h_\ell^{(1)}[\xi _ < ] \right),\end{aligned}\ ] ] where in the low frequency limit @xmath134 , @xmath201 & \ell=0 \\ & \\ i \xi_v^{2 \ell + 1 } \frac{\gamma \left[-\frac{2 \ell } { 3}-\frac{1}{3}\right ] \gamma \left[-\frac{\ell } { 3}-\frac{1}{6}\right ] \left(\csc \left[\frac{1}{6 } ( 2 \pi \ell + \pi ) \right]+\cot \left[\frac{1}{6 } ( 2 \pi \ell + \pi ) \right ] \csc [ \pi \ell ] \right)}{2^{2 \ell + 3 } 3 \sqrt{\pi } ( ( 2 \ell -1)!!)^2 \gamma [ -\ell ] } & \ell>0 \\ \end{array } \right . \ , \end{aligned}\ ] ] and in the high frequency limit @xmath198 $ ] , @xmath202 in this section we derive the results presented in the preceding section . in appendix
, we review the relevant facts about solving green s functions for linear second order odes , and also justify the @xmath83 measure on the right hand side of eq . .
the algorithm for obtaining @xmath96 is as follows .
_ general solution of radial green s function _ we need to first solve for the two linearly independent homogeneous solutions @xmath203 $ ] and @xmath204 $ ] to the ode in eq . , namely @xmath205 = 0 .\end{aligned}\ ] ] in the notation of eq . ,
@xmath206 and @xmath207 .
next we normalize the solutions @xmath208 such that they satisfy @xmath209 \left ( r_\ell^{(1)}[\xi ] ( r_\ell^{(2)})'[\xi ] - ( r_\ell^{(1)})'[\xi ] r_\ell^{(2)}[\xi ] \right ) = \frac{1}{\xi^2 } .\end{aligned}\ ] ] then the general solution to our radial green s function is @xmath210 = \mathcal{c}_\ell r_\ell^{(1)}[\xi _ > ] r_\ell^{(2)}[\xi _ < ] - ( 1-\mathcal{c}_\ell ) r_\ell^{(1)}[\xi _ < ] r_\ell^{(2)}[\xi _ > ] + c_\ell^{11 } r_\ell^{(1)}[\xi ] r_\ell^{(1)}[\xi ' ] + c_\ell^{22 } r_\ell^{(2)}[\xi ] r_\ell^{(2)}[\xi'],\end{aligned}\ ] ] where the constants @xmath211 , @xmath212 and @xmath213 do not depend on @xmath98 nor @xmath99 , but depend on @xmath100 .
the @xmath214 and @xmath215 terms have discontinuous first derivatives and hence contribute to the coefficient of the @xmath76-functions on the right hand side of the green s function equation in eq . . for @xmath216 , retarded boundary conditions and the demand for non - singular solutions will fix these constants uniquely
that is , we shall require that , whenever the observer is very far away from @xmath8 , @xmath52 , and the source is closer to the central mass than the observer , @xmath217 , then the observer ought to receive purely outgoing galileon waves .
furthermore , on physical grounds , we will admit only solutions that are nonsingular when either the observer or the source is situated close to the central body . ) due to @xmath8 that blows up as @xmath218 ; but here we are requiring that , as long as the observer is not sitting on top of the secondary source , i.e. the @xmath74 , she should not measure galileon forces due to @xmath74 that grow without bound . ] for @xmath174 , in addition to regularity and the retarded condition , we shall also need to invoke gauss law applied to the curved spacetime helmholtz equation ( see eq . ) to fix these constants uniquely . in the following subsections
, we will first solve for @xmath96 in the zero frequency ( static ) and high frequency ( wkb ) limits .
we will shortly also derive the @xmath208s in terms of bessel and hankel functions in the limits @xmath184 and @xmath188 .
this means we can fix the form of @xmath96 in the limits @xmath188 , @xmath184 and @xmath140 up the @xmath100-dependent constants @xmath211 , @xmath212 and @xmath213 . we will then proceed to fix these constants at least within the low and high frequency limits , @xmath134 and @xmath198 $ ] , respectively by ensuring they agree with the static and wkb @xmath96 results in the same limits .
_ well outside vainshtein _ that we have just shown that the radial wave equation in eq .
reduces to that in minkowski spacetime when @xmath52 implies that we may use the known solution there to read off the @xmath208s .
the solution in flat ( and , importantly , empty ) minkowski spacetime is textbook material , and we may represent it as @xmath219}{4\pi |\vec{x}-\vec{x}'| } = \int \frac{{\text{d}}\omega}{2\pi } e^{-i\omega(t - t ' ) } \frac{e^{i\omega |\vec{x}-\vec{x}'|}}{4\pi |\vec{x}-\vec{x}'| } .\end{aligned}\ ] ] then , using @xmath220 \overline{y_\ell^m}[\theta',\phi ' ] j_\ell[\xi _ < ] h^{(1)}_\ell[\xi_>],\end{aligned}\ ] ] we can deduce that the @xmath221 $ ] in the limit @xmath52 must be a linear combination of @xmath222 and @xmath122 . the @xmath223 implement retarded boundary conditions , since the spherical hankel function of the first kind may be understood as @xmath224 = - i ( -z)^\ell \left ( \frac{1}{z } \frac{{\text{d}}}{{\text{d}}z } \right)^\ell \frac{e^{i z}}{z } \ , \ ] ] note that @xmath225 $ ] only contains a factor of @xmath226 $ ] and does not contain @xmath227 $ ] .
thus , using the asymptotic expansion of the hankel function for large argument , we see that @xmath228 \to ( -i)^{\ell+1}\frac{\exp\left[-i\omega(t - t'-r)\right]}{\omega r } \left ( 1 + \mathcal{o}\left [ \left ( \omega r \right)^{-1 } \right ] \right),\end{aligned}\ ] ] describes radially purely outgoing waves at unit speed propagating to infinity .
a similar discussion shows that @xmath229 implements advanced boundary conditions , and because @xmath222 can be expressed as a linear combination of @xmath230 , it describes a superposition of ingoing and outgoing waves . in the ansatz of eq .
, we see that we have to choose @xmath231 = h_\ell^{(1)}[\xi_>]$ ] and set @xmath232 and @xmath233 to ensure retarded boundary conditions .
( we are able to deduce from eq . that @xmath122 and @xmath222 are already appropriately normalized to obey the wronskian condition in eq . for @xmath184 .
) this means we have determined @xmath234 $ ] to take the form in eq . .
_ well inside vainshtein _ let us now understand the forms of @xmath221 $ ] evaluated close to the central source ( @xmath49 ) .
we exploit equations and , keeping only the most dominant terms in @xmath107 ( equations through ) , to reformulate eq . as @xmath235 = 0 \ .\end{aligned}\ ] ] one may rescale the solutions @xmath236 \equiv \xi^{1/4 } \mathcal{r}^{(1,2)}_\ell[\sqrt{3}\xi/2]$ ] and find that @xmath237 $ ] satisfies bessel s equation @xmath238 = 0 \ , \end{aligned}\ ] ] where @xmath239 . noting that the wronskian between @xmath121 and @xmath240 is @xmath241 = j_\nu[z ] ( h^{(1)}_\nu)'[z ] - ( j_\nu)'[z ] h^{(1)}_\nu[z ] = \frac{2i}{\pi z},\end{aligned}\ ] ] and that @xmath242 , z^\frac{1}{4 } h^{(1)}_\nu[z ] \right ] = \sqrt{z } { \sf wr}_{(z)}\left [ j_\nu[z ] , h^{(1)}_\nu[z ] \right],\end{aligned}\ ] ] we conclude that the two linearly independent solutions normalized to obey eq . are @xmath243 & \equiv \sqrt{\frac{i\pi}{2 \xi_v^{3/2 } } } \xi^{\frac{1}{4 } } h^{(1)}_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi/2 \right ] , \\
r_\ell^{(2)}[\xi ] & \equiv \sqrt{\frac{i\pi}{2 \xi_v^{3/2 } } } \xi^{\frac{1}{4 } } j_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3 } \xi/2 \right].\end{aligned}\ ] ] with these homogeneous solutions , the general solution for the radial green s function , deep within the vainshtein radius , @xmath244 , is @xmath245 & = \frac{i\pi}{2 \xi_v^{3/2 } } \sqrt[4]{\xi \xi ' } \bigg ( \mathcal{c}_\ell h^{(1)}_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi_>/2 \right ] j_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi_</2 \right ] - \left(1-\mathcal{c}_\ell\right ) h^{(1)}_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi_</2 \right ] j_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi_>/2 \right ] \nonumber\\ & \qquad + c_\ell^\text{(jj ) } j_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi/2 \right ] j_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi'/2 \right ] + c_\ell^\text{(hh ) } h^{(1)}_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi/2 \right ] h^{(1)}_{\frac{1}{4}\left ( 2\ell+1 \right)}\left [ \sqrt{3}\xi'/2 \right ] \bigg ) .
\end{aligned}\ ] ] next , we recall the small argument limits ( @xmath246 ) of the bessel and hankel functions @xmath247 & \to \frac{(z/2)^{\frac{1}{4}\left ( 2\ell+1 \right)}}{\gamma\left[\frac{1}{4}\left ( 2\ell+5 \right)\right ] } \left ( 1 + \mathcal{o}\left [ z^2 \right ] \right ) ,
\\ \label{hankelh1_smallargument } h^{(1)}_{\frac{1}{4}\left ( 2\ell+1 \right)}[z ] & \to -\frac{i}{\pi } \gamma\left[\frac{1}{4}\left ( 2\ell+1 \right)\right ] \left(\frac{2}{z}\right)^{\frac{1}{4}\left ( 2\ell+1 \right ) } \left ( 1 + \mathcal{o}\left [ z^2 \right ] \right ) \nonumber\\ & \qquad + \left(\frac{z}{2}\right)^{\frac{1}{4}(2\ell+1 ) } \frac{(1+i\cot\left [ \frac{\pi}{4}(2\ell+1 ) \right])}{\gamma[\frac{1}{4}(2\ell+5 ) ] } \left ( 1 + \mathcal{o}\left [ z^2 \right ] \right ) .\end{aligned}\ ] ] the @xmath248 piece of the small argument behavior of @xmath249 \equiv j_\nu[z ] + i n_\nu[z]$ ] can be traced to @xmath250 $ ] .
this implies that if we want a nonsingular solution as @xmath251 , we must set @xmath232 and @xmath252 for @xmath216 . for @xmath174 , however , both @xmath253 $ ] and @xmath254 $ ] are nonsingular in the small radius limit ; the former goes to zero and the latter to a constant .
we must therefore write @xmath255 $ ] as a linear combination of these two functions . _ the radiative limit _ we may now fix the form of the radiative limit , @xmath140 , of @xmath96 .
this is crucial for studying the galileon radiation seen by an asymptotic observer at @xmath52 generated by a source moving deep within the vainshtein radius of the central mass ( @xmath256 ) .
our previous discussion leads us to the forms @xmath257 & = & \left\ { \begin{array}{lr } h_0^{(1)}[\xi _ > ] \left ( c_0^\text{(j)}\cdot \sqrt[4]{\xi _ < } j_{\frac{1}{4}}\left
[ \sqrt{3}\xi_</2 \right ] + c_0^\text{(h)}\cdot \sqrt[4]{\xi _ < } h^{(1)}_{\frac{1}{4}}\left [ \sqrt{3}\xi_</2 \right ] \right ) & \ell=0 \\ & \\ h_\ell^{(1)}[\xi _ > ] c_\ell^\text{(j)}\cdot \sqrt[4]{\xi _ < } j_{\frac{1}{4}(2\ell+1)}\left [ \sqrt{3}\xi_</2 \right ] & \ell>0 \\ \end{array } \right .
\ .\end{aligned}\ ] ] the static limit of the green s function , as defined in eq . , requires a slightly different treatment from the time dependent case because it amounts to setting @xmath258 in frequency space , making the variables @xmath98 , @xmath99 and @xmath100 in eq .
ill defined .
we therefore work instead with the original radial variables @xmath259 and @xmath0 , in terms of which becomes , in the @xmath260 limit @xmath261 = 0.\end{aligned}\ ] ] ( here @xmath262 $ ] depend on @xmath263 , not @xmath264 . ) we also express eq . as @xmath265 \equiv \lim_{\omega\to 0 } \omega \widetilde{g}_\ell[\xi,\xi ' ] & = \mathcal{c}_\ell r_\ell^{(1|s)}[r _ > ] r_\ell^{(2|s)}[r _ < ] - ( 1-\mathcal{c}_\ell ) r_\ell^{(1|s)}[r _ < ] r_\ell^{(2|s)}[r _ > ] \nonumber\\ & \qquad \qquad + c_\ell^{11 } r_\ell^{(1|s)}[r ] r_\ell^{(1|s)}[r ' ] + c_\ell^{22 } r_\ell^{(2|s)}[r ] r_\ell^{(2|s)}[r ' ] , \end{aligned}\ ] ] and eq . as @xmath266 \left ( r_\ell^{(1|s)}[r ] ( r_\ell^{(2|s)})'[r ] - ( r_\ell^{(1|s)})'[r ] r_\ell^{(2|s)}[r ] \right ) = \frac{1}{r^2}.\end{aligned}\ ] ]
as we shall see , for @xmath216 , the radial static green s function will be fixed once we demand that the solutions are regular for all radii @xmath259 . for @xmath174 ,
regularity is irrelevant ; instead , @xmath267 will be determined by ensuring that eq . is obtained and , for @xmath268 , that the green s function goes to zero . multiplying both sides of eq .
by @xmath269 ( and dropping the labels ) yields @xmath270 + \frac{4r^3+r_v^3}{4 } \left ( \frac{2}{r } r'[r ] -\frac{\ell(\ell+1)}{r^2 } r[r ] \right ) = 0\end{aligned}\ ] ] and this equation may be readily solved in mathematica @xcite .
the general homogeneous solutions to the static radial mode equation eq .
, normalized to satisfy the condition in eq .
are @xmath271 & \equiv \sqrt{\frac{2}{(2\ell+1)r_v } } \left ( \frac{r_v}{r } \right)^{\frac{\ell}{2 } } \ _
2f_1\left [ \frac{\ell}{6}+\frac{1}{3 } , -\frac{\ell}{2 } ; \frac{5}{6}-\frac{\ell}{3 } ; -\frac{r^3}{r_v^3 } \right ] , \\ r_\ell^{(2|s)}[r ] & \equiv \sqrt{\frac{2}{(2\ell+1)r_v } } \left ( \frac{r}{r_v } \right)^{\frac{\ell+1}{2 } } \ _ 2f_1\left [ \frac{1}{6}-\frac{\ell}{6 } , \frac{1}{2}+\frac{\ell}{2 } ; \frac{7}{6}+\frac{\ell}{3 } ; -\frac{r^3}{r_v^3 } \right ] .\end{aligned}\ ] ] since @xmath272 = 1 $ ] , we see that the @xmath273 r_\ell^{(2|s)}[r_>]$ ] term tends to @xmath274 and the @xmath212 term in eq
. tends to @xmath275 , as @xmath276 .
these two terms grow without bound , and therefore we must choose @xmath277 and @xmath278 .
next , we use the identity @xmath279 & = \frac{\gamma[\gamma]\gamma[\beta-\alpha]}{\gamma[\beta]\gamma[\gamma-\alpha ] } ( -z)^{-\alpha }
\ _ 2f_1\left[\alpha,\alpha+1-\gamma;\alpha+1-\beta;\frac{1}{z}\right ] \\ & \qquad \qquad + \frac{\gamma[\gamma]\gamma[\alpha-\beta]}{\gamma[\alpha]\gamma[\gamma-\beta ] } ( -z)^{-\beta } \ _ 2f_1\left[\beta,\beta+1-\gamma;\beta+1-\alpha;\frac{1}{z}\right ] \nonumber\end{aligned}\ ] ] to recast the product @xmath280 r_\ell^{(2|s)}[r_<]$ ] in eq . in terms of @xmath281 $ ] .
the two potentially divergent terms for @xmath138 are the ones proportional to @xmath282 \gamma\left [ \frac{2 ( \ell+2)}{3}\right ] + \frac{\sqrt{\pi}\ell!}{\sin\left[\frac{1}{6}\pi ( 2 \ell+1 ) \right ] } \right ) \left ( 1 + \mathcal{o}\left [ r_v^2/r_\lessgtr^3 \right ] \right ) \\
\label{staticdivergences_ii } & \left(\frac{r_>}{r_<}\right)^\ell \frac{r_v}{r _ < } \left ( 2 c_\ell^{22 } \gamma\left[\frac{\ell}{3}+\frac{1}{6}\right ] \gamma\left [ \frac{2 ( \ell+2)}{3}\right ] + \frac{\sqrt{\pi}\ell!}{\sin\left[\frac{1}{6}\pi ( 2 \ell+1 ) \right ] } \right ) \left ( 1 + \mathcal{o}\left [ r_v^2/r_\lessgtr^3 \right ] \right ) .\end{aligned}\ ] ] we may therefore collect the results @xmath283}{2\sqrt{\pi } \gamma[\frac{1}{3}(2\ell+1 ) ] } , \end{aligned}\ ] ] where we have used the @xmath137-function identity @xmath284 \gamma[1-z ] = \pi \mathrm{cosec}[\pi z]$ ] .
notice that none of the regularity constraints apply for @xmath174 .
in fact , @xmath285 $ ] is a constant .
@xmath286 $ ] varies as @xmath287 for small @xmath130 , and using eq .
, we obtain the equivalent expression @xmath288 = \sqrt{\frac{2}{r_v } } \left ( \frac{\gamma[\frac{1}{3 } ] \gamma[\frac{7}{6}]}{\sqrt{\pi } } - \frac{r_v}{2r } \,_2f_1\left [ \frac{1}{3 } , \frac{1}{2 } ; \frac{4}{3 } ; -\frac{r_v^3}{r^3}\right ] \right),\end{aligned}\ ] ] which varies as @xmath289 for large @xmath130 .
this implies that @xmath286 $ ] is regular for both large and small @xmath130 . to determine @xmath290 , @xmath291 and @xmath292 here
, we recall the discussion towards the end of section , that @xmath293/(4\pi)$ ] must correspond to the coefficient of the @xmath294 piece of @xmath295 $ ] in eq . ; by spherical symmetry , the @xmath216 do not contribute to the solution generated by a point mass at the origin .
this implies @xmath296 and @xmath297\gamma[1/6]/(6 \sqrt{\pi})$ ] . when @xmath298 , by using the identity in eq . on @xmath299/(4\pi)$ ] ,
and setting the resulting @xmath300s to unity , we find that the only constant term ( independent of both @xmath35 and @xmath97 ) reads @xmath301/(9 \cdot 2^{2/3 } \sqrt{3\pi } r_v)$ ] . because we have already chosen the asymptotic boundary condition ( see equations and ) that @xmath302 $ ] generated by a point mass located at some finite @xmath97 should approach zero , this implies @xmath303 .
we see , at this point , that @xmath304 $ ] only depends on @xmath114 and not on @xmath115 .
we may also arrive at the same result for @xmath299/(4\pi)$ ] , without invoking the background solution @xmath28 , if we refer to the curved spacetime picture described in section .
in particular , the static limit of eq . , gotten by setting @xmath305 , translates to @xmath306 = 1 - \mathcal{c}_0 = 1 .\end{aligned}\ ] ] this immediately implies @xmath307 .
taking the @xmath298 limit tells us that , as @xmath268 , we are left with @xmath308 \gamma[\frac{7}{6}]}{\sqrt{\pi } } \right ) + c_0^{22}\left ( \frac{2 \gamma^2[\frac{1}{3 } ] \gamma^2[\frac{7}{6}]}{\pi r_v } - \frac{\gamma[\frac{1}{3 } ] \gamma[\frac{7}{6}]}{\sqrt{\pi } r _ < } \right).\end{aligned}\ ] ] since we require @xmath302 = 0 $ ] for any finite radial location of the point mass , we must have @xmath309 \gamma[\frac{7}{6}]/\sqrt{\pi}$ ] and @xmath303 .
this completes the derivation of eq . . with the exact solution to the static green s function in hand
, we may now take the limits @xmath188 , @xmath184 , and @xmath140 . in the small radii limit ,
@xmath188 , we set the @xmath300s in eq .
to unity , and drop the subleading @xmath310 term relative to the dominant @xmath311 term , to obtain @xmath312 & = -\frac{1}{2\pi r_v } \left ( \sqrt{\frac{r}{r_v } } + \sqrt{\frac{r'}{r_v } } \right ) + \frac{2\sqrt{rr'}}{r_v^{3/2 } } \sum_{\ell = 0}^\infty \sum_{m = -\ell}^\ell \frac { y_\ell^m[\theta,\phi ] \overline{y_\ell^m}[\theta',\phi ' ] } { 2\ell+1 } \frac{1}{\sqrt{r _ > } } \left ( \frac{r_<}{r _ > } \right)^{\frac{\ell}{2 } } , \end{aligned}\ ] ] the infinite mode sum in eq .
may be collapsed into a closed expression by first summing over the azimuthal modes and then invoking the generating function of the legendre polynomials .
the result of the mode sum in eq .
is eq . .
in the same vein , the large radii @xmath184 static green s function may be summed into a closed form by similar means .
in fact , our solution is consistent with our earlier observation that galileon dynamics reduce to that of a minimally coupled massless scalar field in minkowski ; for since we know its static limit is eq .
, our galileon static green s function ought to reduce to the same in this asymptotic limit to lowest order in @xmath313 . in detail ,
if one begins from eq .
, applies the identity in eq . to the @xmath314s and then proceeds to set the transformed @xmath314s to unity because their arguments will go as @xmath315 , which is very small at large radii
one finds a subleading term proportional to @xmath316 and a dominant term proportional to @xmath317 . keeping only the dominant term and again converting the sum over spherical harmonics into one over legendre polynomials , followed by applying the latter s generating function
, we reach eq . .
as for the case @xmath140 , the result in eq . can be obtained by starting with the exact solution in eq . , but only applying the identity in eq .
to @xmath318 $ ] , followed by setting the @xmath314s to unity .
next we consider the high frequency limit , @xmath319 $ ] .
we first rescale the mode functions via @xmath320 \equiv \frac{\mathcal{r}_\ell^{(1,2)}[\xi]}{\left ( \xi ( \xi^3+\xi_v^3 ) \right)^{1/4}},\end{aligned}\ ] ] so that the wronskian condition in eq . becomes @xmath321 ( \mathcal{r}_\ell^{(2)})'[\xi ] - ( \mathcal{r}_\ell^{(1)})'[\xi ] \mathcal{r}_\ell^{(2)}[\xi ] = 1 $ ] , and reads @xmath322 + \left ( k[\xi,\xi_v ] + u[\xi/\xi_v ] \right ) \mathcal{r}_\ell^{(1,2)}[\xi ] \
.\end{aligned}\ ] ] here @xmath323 has already been defined in eq . , and we define @xmath324 \equiv \frac{16 \xi^6 \ell ( \ell+1)+4 \xi^3 \xi_v^3 ( 5 \ell(\ell+1)+6)+\xi_v^6 ( 4 \ell(\ell+1)-3)}{16 \xi^2 ( \xi^3+\xi_v^3)^2}. \end{aligned}\ ] ] for large @xmath319 $ ] , we see that the denominator of @xmath325 scales as 8 powers of the large quantity @xmath326 ; while its numerator contains six powers of @xmath327 times terms of order unity , order @xmath129 and order @xmath328 .
this means that the largest possible scaling of @xmath325 is that it goes as @xmath329 . for @xmath174 ,
@xmath325 scales as @xmath330 .
since @xmath323 is of order unity , in the high frequency limit we may therefore discard @xmath325 relative to @xmath323 . suppressing the irrelevant indices , we now seek to solve @xmath331 + u[\xi/\xi_v ] \mathcal{r}[\xi ] \ .\end{aligned}\ ] ] ( here and below , we are introducing a fictitious parameter @xmath332 that will be set to unity once the solutions to @xmath333 are obtained . ) observe that , viewed as a function of @xmath98 , the @xmath323 is very flat by assumption , because @xmath334 = u'[\xi/\xi_v]/\xi_v \ll 1 $ ] . this calls for the wkb method of solution , in which one uses the derivatives of @xmath323 with respect to @xmath98 as an expansion parameter .
we therefore pose the ansatz @xmath335 = \frac{e^{(i/\epsilon)\mathcal{s}[\xi]}}{\sqrt[4]{-u[\xi/\xi_v ] } } \sum_{\ell=0}^\infty \epsilon^\ell \tau_{(\ell)}[\xi].\end{aligned}\ ] ] it is important to note from eq . that @xmath336}$ ] has no real zeros , though it has a singularity at @xmath337 and a global minimum at @xmath338 } = \sqrt{2/3}$ ] .
both in eq . and the ansatz of eq .
, @xmath332 will turn out to count derivatives , so that @xmath339 implies an integral . inserting eq . into eq . and
setting the coefficient of each distinct power @xmath340 to zero , the @xmath341 term yields a relationship between @xmath323 and @xmath342 , which we may integrate to obtain two solutions @xmath343 = \pm \int^\xi { \text{d}}\xi '' \sqrt{-u[\xi''/\xi_v ] } \ .\end{aligned}\ ] ] the @xmath344 term gives a differential relationship between @xmath345{-u[\xi/\xi_v]}$ ] , its first derivative with respect to @xmath98 , and @xmath342 and @xmath346 . through eq .
, this gives @xmath347 by setting to zero the coefficients of @xmath340 , for @xmath144 , we find a recursion relation obeyed by @xmath348 , @xmath349{-u[\xi''/\xi_v ] } } \frac{{\text{d}}^2}{{\text{d}}\xi''^2 } \left ( \frac{\tau_{(\ell-1)}[\xi'']}{\sqrt[4]{-u[\xi''/\xi_v ] } } \right),\end{aligned}\ ] ] where @xmath350 ( or @xmath351 ) is chosen if we chose the @xmath351 ( or @xmath350 ) sign in eq . .
as advertised earlier , we see that every higher order in @xmath332 contains an additional derivative with respect to @xmath98 ; and the @xmath339 in the phase of eq . is the integral in eq . . for our purposes
, we shall work only to lowest order in the wkb approximation , just involving @xmath352 and @xmath353 ; the solutions @xmath208 normalized to obey the wronskian condition in eq .
are @xmath354 = \frac{\exp\left[\pm
i \xi_v \int_0^{r / r_v } { \text{d}}\vartheta \sqrt{-u[\vartheta]}\right]}{\sqrt[4]{-u[r / r_v ] } \left ( \xi ( \xi^3+\xi_v^3 ) \right)^{1/4}}.\end{aligned}\ ] ] let us pause to understand the large ( @xmath355 ) and small ( @xmath356 ) radius limits . examining fig .
reminds us that @xmath357}$ ] is basically flat for large @xmath355 .
together with the limit @xmath358 } = 1 $ ] , we may infer , for some @xmath359 and @xmath98 such that @xmath360 , @xmath361 } { \text{d}}\vartheta \approx \xi_v\int_0^{\xi_0/\xi_v } \sqrt{-u[\vartheta ] } { \text{d}}\vartheta + \xi-\xi_0 = \xi - \xi_v\int_0^{\xi_0/\xi_v } \left(1-\sqrt{-u[\vartheta]}\right ) { \text{d}}\vartheta .\end{aligned}\ ] ] we will justify below that we may now further approximate this integral by extending the upper limit of integration @xmath362 to infinity , @xmath363 } { \text{d}}\vartheta \approx \xi - \xi_v \mathcal{i}_\infty , \quad \xi/\xi_v \gg \ell\end{aligned}\ ] ] where @xmath364 was defined in eq . .
similarly , for @xmath356 , by the flatness of the potential @xmath365}$ ] near @xmath366 , we have @xmath367 } { \text{d}}\vartheta \approx \frac{\sqrt{3}}{2}\xi , \qquad \xi/\xi_v \ll 1 .\end{aligned}\ ] ] we thus have @xmath368 & \approx & \left\ { \begin{array}{lr } \frac{\exp\left[\pm i \frac{\sqrt{3}}{2}\xi \right]}{\sqrt{\sqrt{3}/2 } \left ( \xi \xi_v^3 \right)^{1/4 } } , & r / r_v \ll 1 \\ &
\\ \frac{\exp\left[\pm i \left ( \xi - \xi_v \mathcal{i}_\infty \right ) \right]}{\xi } & r / r_v \gg 1 \\ \end{array } \right . \ .\end{aligned}\ ] ] the @xmath188 , @xmath188 and @xmath140 limits reported in equations , , and follow from equations and once @xmath369 is computed .
it is important to observe that , in this high frequency limit we are working in , @xmath203 $ ] ( the @xmath351 sign solution ) in eq .
is proportional to @xmath370 $ ] in the @xmath49 regime and to @xmath225 $ ] in the @xmath52 regime . to validate this assertion
, we merely need to compare the expressions in eq . against the high frequency limit of the hankel functions .
( likewise , @xmath204 $ ] is proportional to @xmath371 $ ] and @xmath372 $ ] in the limits @xmath52 and @xmath49 respectively . )
retarded boundary conditions mean , therefore , that in eq . we need to set @xmath232 and @xmath373 for all @xmath374 . at this point , our wkb radial green s function solution takes the form in eq . . for @xmath216 ,
@xmath369 may be fixed by regularity , demanding that the limit @xmath251 yields a radial green s function that is proportional to the high frequency behavior of @xmath375 $ ] . from the form in eq and the large argument limit of the bessel function
this translates to the consistency condition @xmath376 , \end{aligned}\ ] ] where @xmath377 is a constant . by converting the cosine into exponentials and equating the coefficients of @xmath378 $ ] on both sides ,
this lets us solve for both @xmath369 and @xmath377 . for @xmath174
, the radial green s function can now be proportional to a linear combination of the high frequency limits of @xmath379 $ ] and @xmath380 $ ] , because as already discussed , both are non - singular in the zero radius limit .
let us consider setting @xmath381 , eliminating the @xmath382 term . if we now also take @xmath383 , and if we remind ourselves of eq . and the large argument limit of @xmath240 , we see that the wkb @xmath384 term must match onto the high frequency limit of the @xmath385{\xi\xi ' } h^{(1)}_{1/4}[\sqrt{3}\xi/2 ] h^{(1)}_{1/4}[\sqrt{3}\xi'/2]$ ] term in eq . , i.e. @xmath386 when @xmath198 $ ] . as we will discuss in section below , the @xmath174 radial green s function , which obeys an inhomogeneous helmholtz equation in curved spacetime ,
must obey the gauss law in eq .
: @xmath387 = 1,\end{aligned}\ ] ] where we have taken the small radius limit of the effective metric in eq . .
because the @xmath253 $ ] vanishes when evaluated at @xmath388 , this condition applied to eq .
allows us to solve for @xmath389 which in turn yields the @xmath384 in eq . .
in the preceding discussion , we invoked the flatness of @xmath365}$ ] at both large and small @xmath390 .
let us pause to quantify this flatness by examining the following expressions @xmath391 \equiv \xi_v\int_0^\vartheta { \text{d}}\vartheta ' \left ( \frac{\sqrt{3}}{2 } - \sqrt{-u[\vartheta ' ] } \right)\end{aligned}\ ] ] and @xmath392 \equiv \xi_v\int_\vartheta^\infty { \text{d}}\vartheta ' \left ( 1 - \sqrt{-u[\vartheta ' ] } \right).\end{aligned}\ ] ] the reason for the @xmath100 in front of the integrals is that , because these expressions occur in phases ( e.g. @xmath393 $ ] ) , it is not sufficient for the integrals themselves to be much less than unity .
we need the entire expression to be small , so that very little oscillation occurs .
computing the power series of @xmath394 $ ] and @xmath395 $ ] about @xmath396 allows us to understand precisely how good an approximation the expressions in eq .
for @xmath356 we obtain @xmath397 = \xi_v \frac{(\xi/\xi_v)^{5/2}}{5\sqrt{3 } } + \mathcal{o}\left [ \xi_v(\xi/\xi_v)^4 \right]\end{aligned}\ ] ] and , for @xmath355 , @xmath398 = \xi_v\frac{(\xi_v/\xi)^2}{8 } + \mathcal{o}\left [ \xi_v(\xi_v/\xi)^5 \right].\end{aligned}\ ] ] therefore , replacing @xmath399 with @xmath400 in eq .
results in an error in the phase that is of order @xmath401 , as long as we assume @xmath402 ( i.e. the observer is far enough outside vainshtein ) .
likewise , the approximation in @xmath403 for @xmath404 , replacing @xmath405 with @xmath406 , makes an error in the phase that is of order @xmath407 as long as @xmath408 . because we were able to solve for the relevant mode functions ( up to their overall @xmath100-dependent normalization constants ) in the regions very close to and very far away from the the central mass
, we were able to determine the form of the radial green s function in the limits @xmath184 ( eq . ) , @xmath188 ( eq . ) , and @xmath140 ( eq . ) .
we may calculate the @xmath100-dependent constants appearing in these expressions at least in the low frequency @xmath134 and high frequency @xmath198 $ ] limits by matching them onto the static and wkb results we have obtained in the previous two sections . in the low frequency limit
, we may simultaneously take the limit @xmath134 and replace the bessel and hankel functions , and their spherical versions , with their small argument limits .
the resulting expressions in equations , , and can then be compared against the respective expressions in equations , ( the small @xmath130 , @xmath191 limit of ) , and . similarly , in the high frequency limit
, we may simultaneously take the limit @xmath409 and the resulting expressions in , , and can then be compared with the respective expressions in equations , and .
notice that , upon these comparisons , for the radial green s function evaluated deep inside vainshtein , eq .
, @xmath410 in the high frequency limit , while it goes to zero in the low frequency limit . finally , to arrive at equations , , and
, it is useful to write @xmath411 , and to use @xmath412 = \frac{j_\nu[z ] \cos[\pi\nu ] - j_{-\nu}[z]}{\sin[\pi\nu]}.\end{aligned}\ ] ] one may wonder why the result in eq .
is not simply the usual answer in flat minkowski spacetime , i.e. @xmath413 , since if both observer and emitter are very far from the central mass @xmath8 the propagation of signals would not be expected to feel the presence of the central body @xmath8 .
the physical reason is that signals with wavelengths much longer than that of vainshtein radius @xmath134 indeed can not resolve @xmath0 very well @xmath414 is proportional to some integer power of the small quantity @xmath100 but once the wavelengths become much shorter than @xmath0 galileon signals can resolve the vainshtein scale and @xmath414 is not small but becomes a mere phase . centered at the origin of the spatial coordinate system .
the radial retarded green s function takes the form @xmath415 ( j_\ell[\xi _ < ] - \chi h_\ell^{(1)}[\xi_<])$ ] .
perfect absorber here means the scalar field observed on the surface of the sphere is identically zero , thereby imposing @xmath416/h_\ell^{(1)}[\omega r_0]$ ] . at low frequencies , @xmath417 , @xmath418 . at high frequencies , @xmath419 , @xmath420)$ ] . ]
we have now completed the derivation of the results in section .
before we move on to investigate the radiation seen by an asymptotic observer generated by the motion of matter close to the central source @xmath8 , we would like to discuss an alternative perspective for the galileon propagating on the background @xmath25 $ ] .
we shall also discuss the existence of a gauss law for the helmholtz equation obeyed by the @xmath174 mode of the radial green s function . by a direct calculation ,
it is possible to view the green s function equation as that for a massless scalar wave equation in a curved spacetime .
specifically recalling equations through dividing both sides of by @xmath421 , we arrive at @xmath422 = \box_{x ' } g[x , x ' ] = \frac{\delta^{(4)}[x - x']}{|\mathfrak{g}\mathfrak{g}'|^{1/4 } } = \sqrt{\mathfrak{h}[r / r_v ] \mathfrak{h}[r'/r_v ] } \delta[t - t ' ] \frac{\delta[r - r']}{rr ' } \delta^{(2)}[\widehat{x}-\widehat{x } ' ] \end{aligned}\ ] ] with @xmath423 & \equiv \frac{8 \vartheta^3}{1 + 4 \vartheta ^3 } \sqrt{\frac{1+\vartheta^3}{3(1 + 2\vartheta^3 ) - 2 \sqrt{\vartheta^3(1+\vartheta^3)}}},\end{aligned}\ ] ] and @xmath424 & \equiv \delta[\cos\theta-\cos\theta']\delta[\phi-\phi ' ] , \\ \mathfrak{g } & \equiv \det \mathfrak{g}_{\alpha\beta}[x ] , \ \mathfrak{g } ' \equiv \det \mathfrak{g}_{\alpha\beta}[x'].\end{aligned}\ ] ] we have denoted @xmath425 \nabla_{x^\alpha } \nabla_{x^\beta}$ ] and @xmath426 \nabla_{x'^\alpha } \nabla_{x'^\beta}$ ] to be the minimally coupled massless scalar wave operator in a curved spacetime geometry given by the metric @xmath427 observe that the @xmath428 $ ] ( eq . ) occurring in eq .
is @xmath429 in the small radii limit and unity at large radii .
this is the vainshtein effect at work : the magnitude of the point mass sourcing the green s function grows weaker the closer it gets to the central mass @xmath8 , but goes to unity far away from it . on the other hand ,
it is somewhat puzzling , in this curved spacetime picture , that a source located nearer and nearer to the spatial origin tends to zero strength ; for instance , we have already remarked , towards the end of section , that a static point source located at the origin must contribute to the background @xmath25 $ ] solution via a shifting of the mass , @xmath85 . to understand this we first re - express eq . in accordance with our decomposition in eq . ; this means that we drop the integration symbol @xmath430 , and replace @xmath104 $ ] and @xmath431 $ ] with , respectively , @xmath432 and the spherical harmonic completeness relation .
because placing the point source at the origin means we have a spherically symmetric problem , this means only the monopole @xmath174 term in eq .
is relevant . keeping the discussion general for now ,
let us merely assume the metric reads @xmath433 and is time independent .
our frequency space curved spacetime equation is then @xmath434 + \partial_r \left ( \sqrt{|g| } g^{rr } \partial_r \widetilde{g}_0[\xi,0 ] \right ) \right ) = \delta[r],\end{aligned}\ ] ] which allows us to integrate over an infinitesimally small neighborhood about @xmath435 to obtain the normalization condition : @xmath436 = 1 .\end{aligned}\ ] ] this indicates that , even though the measure @xmath437/r^2 $ ] ( see equations and ) tends to zero as @xmath438 , a point mass sitting at the spatial origin must nonethless produce a non - zero ( spherically symmetric ) galileon field , for otherwise eq .
can not be satisfied . that we are able to obtain an equivalent curved spacetime wave equation to eq .
, is in fact one way to justify the `` measure '' @xmath439 multiplying the @xmath76-functions on its right hand side . from general arguments due to hadamard @xcite
see @xcite for a review on green s functions in curved spacetime we know that solutions exist for the massless scalar green s function equation .
since equations and are equivalent , this means the @xmath83 is the correct measure .
moreover in this curved spacetime picture of the galileon green s function , we know that , when the observer at @xmath78 and the source at @xmath79 can be connected by a unique geodesic , the retarded green s function consists of the sum of two terms , namely @xmath440 = \frac{\theta[t - t']}{4\pi } \left ( \delta[\sigma_{x , x ' } ] \sqrt{\delta_{x , x ' } } + \theta[\sigma_{x , x ' } ] v_{x , x ' } \right).\end{aligned}\ ] ] here , @xmath441 is half the square of the geodesic distance from @xmath79 to @xmath78 . the first term after the equality describes propagation of galileon signals on the null cone of the geometry in eq .
; @xmath442 is related to the evolution of the cross sectional area of null rays emanating from @xmath79 to @xmath78 .
@xmath443 $ ] , known as the tail term , describes galileons traveling inside the future null cone of @xmath79 .
it is the solution to the homogeneous wave equation @xmath444 , obeying non - trivial boundary conditions on the null cone of @xmath79 .
we see that solving the retarded galileon green s function provides us information not only about the causal structure of galileon signals but also about the effective spacetime in eq . .
via equations through : we recover flat spacetime for the region well outside the vainshtein radius ( @xmath52 ) , @xmath445 and for the region well within vainshtein ( @xmath49 ) we have instead @xmath446 the @xmath447 in front of @xmath448 indicates that , deep within the vainshtein radius of the central object , galileon waves propagating in the radial direction are superluminal , since @xmath449 .
( we have already noted this , within the context of the wkb results , right after eq . . ) in the same vein , it is worth mentioning that , if some method can be found to evaluate the infinite mode sum of the green s function result for @xmath188 described by equations and through , we should obtain the hadamard form in eq . and this would give us a deeper insight into the superluminal properties of galileon signals near the matter source . note that defining @xmath450 transforms eq . into @xmath451
where the cartesian components of the spatial coordinates are @xmath452 this small radius curved spacetime metric provides an alternate means of deriving the inhomogeneous portion of the static green s function in eq . .
to see this , first re - scale the static green s function as @xmath453 \equiv \frac{\sqrt{3}}{\rho_v^3 } \rho \rho ' g^\text{(static)}[\rho,\theta,\phi;\rho',\theta',\phi ' ] \ .\end{aligned}\ ] ] next , compute the static analog of , namely @xmath454 = |\mathfrak{g}[x ] \mathfrak{g}[x']|^{-1/4 } \delta^{(3)}[\vec{x}-\vec{x}'],\end{aligned}\ ] ] using the small radii metric in .
thus , @xmath455 is the green s function to the laplacian in euclidean 3-space @xmath456 namely @xmath457 & = -\delta_{\vec{\rho } ' } g^\text{(static)}[\vec{\rho } , \vec{\rho } ' ] = \delta^{(3)}[\vec{\rho}-\vec{\rho}'].\end{aligned}\ ] ] remembering the rescaling performed in and requiring that @xmath171 = g^\text{(static)}[\vec{x}',\vec{x}]$ ] then fixes the general solution to take the form @xmath458 = \frac{\sqrt{3}}{4\pi \rho_v^3 } \left ( \frac{\rho \rho'}{|\vec{\rho}-\vec{\rho}'| } - \chi_0 - \chi_1 \left ( \rho + \rho ' \right ) -
\chi_2 \rho \rho ' \right),\end{aligned}\ ] ] where @xmath459 are spacetime constants ; these @xmath459 terms are homogeneous solutions , i.e. @xmath460 and @xmath461 applied on them give zero .
the @xmath462 may be fixed by placing the source at the spatial origin , @xmath463 or @xmath464 , and making sure that eq .
is recovered .
( the @xmath465 may also be determined using the condition derived in eq . .
first replace @xmath35 with @xmath466 ; remember @xmath467 = \lim_{\omega\to 0 } \omega \widetilde{g}_0[\xi,0]/(4\pi)$ ] ; and a short calculation yields @xmath468 .
altogether , @xmath469 . ) finally , @xmath470 can be fixed by demanding that , for @xmath268 , the green s function goes to zero .
( we do not compute @xmath459 in detail , since we have already obtained the exact result in the previous sections . ) in equations and we derive the minimally coupled scalar field generated by @xmath10 point masses in a generic curved spacetime in terms of the hadamard form in eq . .
the reason for doing so is that the galileon field @xmath67 generated by the @xmath10 body dynamics is , in fact , related to eq .
the curved geometry in question is eq . .
, whereas in this paper we are working in flat spacetime and the element of proper time is @xmath471 .
] one can not help but wonder if our mode expansion results in section can be utilized , at least in some limits , to extract the various portions ( @xmath472 , etc . ) of the hadamard form in eq . .
we leave these questions for possible future work .
in the following two sections we wish to examine the galileon waves produced by matter in motion , within two concrete scenarios . the first is the radiation made by acoustic waves propagating on the surface of the large central body of mass @xmath8 with radius @xmath473 .
the second is the radiation created by the movement of compact bodies orbiting this central body ; for instance , this could describe our solar system , with planets orbiting around the sun or a highly asymmetric mass ratio binary star system capable of also generating gravitational waves that could be observed by upcoming gravitational wave detectors .
we initiate the discussion by defining radiation to be the piece of the galileon field that carries a non - trivial energy - momentum flux to infinity .
quantitatively , we only wish to consider the portion of @xmath67 that contributes a non - zero power ( per unit solid angle ) at infinity , @xmath474 ( we have used eq . .
) as we will see more explicitly below , both @xmath475 and @xmath476 contain a power series in @xmath29 , beginning at @xmath29 . even though the full galileon field is @xmath477 , according to eq .
, the derivative of the static background galileon field goes as @xmath478 \sim 1/r^2 $ ] , it does not contribute to the asymptotic power we are currently after and hence would be neglected in the following discussion . to highlight the importance of the non - linear self interactions of the galileon field
, we will also compute the radiation signature in the same setups arising from the minimally coupled massless scalar , so that we can compare the prediction of the two theories .
that is , we will also consider the theory described by the following action @xmath479 compare this with the galileon action @xmath480 in equations , , and .
there , we had to first solve @xmath28 sourced by @xmath8 and then proceed to perturb around it . here , because the @xmath481-theory is linear , the complete field @xmath481 is gotten by superposing the field spawned by each individual source .
when we compute the radiation generated by the surface waves on @xmath8 , we will assume that these waves are driven by external ( non - gravitational ) forces , so that @xmath8 itself can be considered to be a static source of the scalar field(s ) ( falling off as @xmath482 ) , and do not produce any radiation due to backreaction .
the only source of radiation there is @xmath74 describing the waves themselves .
when we compute the radiation of the @xmath10-body system , however , we need to remember that , even though @xmath483 , and hence the location of @xmath8 never deviates far from the center of energy of the system the entire system is held together by ( largely ) conservative gravitational forces . in this case
, we shall see it is important to include @xmath8 in the radiation calculation so as to enforce the conservation of linear momentum .
the radiative limits of the green s function for @xmath67 ( equations through ) and for @xmath481 ( equations and ) both take the form @xmath484 & = \int \frac{{\text{d}}\omega}{2\pi } e^{-i\omega(t - t ' ) } \sum_{\ell , m } \widehat{\omega}_\ell^m[\theta,\phi ] ( \widehat{\omega}_\ell^m)^*[\theta',\phi ' ] \mathcal{r}^<_\ell[\xi _ < ] h_\ell^{(1)}[\xi _ > ] \ , \end{aligned}\ ] ] where we have chosen to write the orthonormal angular mode functions as a linear combination of the usual spherical harmonics @xmath485 ( the @xmath486 in @xmath487 represents complex conjugation . ) here @xmath488 is a unitary @xmath489 matrix , so @xmath490 is as good an orthonormal basis as the spherical harmonics . for the minimally coupled massless scalar @xmath481 , equations and tell us @xmath491
= i\omega j_\ell[\xi
_ < ] \ , \end{aligned}\ ] ] while eq
. gives , for the galileon @xmath67 , @xmath492 = \omega c_\ell^\text{(rad ) } \cdot \sqrt[4]{\xi _ < } j_{\frac{\sigma_\ell}{4}(2\ell+1)}\left [ \sqrt{3}\xi_</2 \right],\end{aligned}\ ] ] where @xmath493 and @xmath494 when @xmath216 .
if we decompose the matter source in the same way that we decomposed the green s function , @xmath495 = \int \frac{{\text{d}}\omega}{2\pi } e^{-i\omega t } \sum_{\ell , m } \widehat{\omega}_\ell^m[\theta,\phi ] \rho_\ell^m[\omega , r],\end{aligned}\ ] ] then by the orthonormality of both the exponentials and the angular mode functions , we may translate eq . into the following solution for the @xmath67 or @xmath481 field evaluated at ( @xmath496 ) , sourced by @xmath497 @xmath498,\xi[x ] & = \frac{1}{{m_{\text{pl } } } } \int \frac{{\text{d}}\omega}{2\pi } e^{-i\omega t } \sum_{\ell , m } \widehat{\omega}_\ell^m[\theta,\phi ] h_\ell^{(1)}[\xi ] \int_0^\infty { \text{d}}r ' r'^2 \mathcal{r}^<_\ell[\xi ' ] \rho_\ell^m[\omega , r'].\end{aligned}\ ] ]
the asymptotic behavior of the spherical hankel function then implies that the scalar field , for a fixed angular frequency @xmath92 , is indeed proportional to a finite power series in @xmath29 , and the series begins at @xmath29 .
this leading order @xmath29 piece of the time and radial derivatives is precisely what we call radiation , because when inserted into eq .
, the factors of @xmath35 cancel and what remains is a finite amount of energy transported to infinity .
any part of the fields containing a higher power than @xmath29 would make a contribution to @xmath499 that decays to zero at infinity .
moreover , since taking a time derivative brings down a @xmath500 and taking a @xmath35-derivative to lowest order in @xmath29 brings down a @xmath501 , we deduce that the radiative part of the fields obey the relationship , @xmath502 with @xmath481 respecting the same equation .
hence it suffices to display only the time derivatives , @xmath503 ( -i)^\ell\frac{e^{i\xi}}{r } \left(1+\mathcal{o}\left[\frac{1}{r}\right ] \right ) \int_0^\infty { \text{d}}r ' r'^2 \mathcal{r}^<_\ell[\xi ' ] \rho_\ell^m[\omega , r'].\end{aligned}\ ] ] because it contains an @xmath35-dependent exponential @xmath504 , one may worry that the fourier integral in eq .
would somehow introduce additional terms that go as @xmath29 ( after the @xmath92-integral has been performed ) , and therefore that dropping the higher powers of @xmath29 in eq . at this stage
is premature .
however , we will see from the integral in eq .
below that the @xmath504 drops out of the expression for total energy , and therefore @xmath35 does not take part in the resulting @xmath92-integral ; in particular , the higher powers of @xmath29 that have been discarded in eq .
would indeed decay away once we take the @xmath46 limit .
in general , we expect the fourier integral in eq . to be extremely difficult to evaluate .
we will instead consider what the spectrum of radiation emitted from a particular system is .
the spectrum is indeed a physical observable , since observations usually take place over enough cycles of the radiation field for a fourier analysis to be done .
the total energy per solid angle radiated to infinity is the integral @xmath505 ( the same expression holds for @xmath481 . )
a few standard fourier identities allow us to express the total scalar energy radiated per unit solid angle , from eq .
, as the following integral over all angular frequencies @xmath506}{{\text{d}}\omega } & = \frac{1}{{m_{\text{pl}}}^2 } \int \frac{{\text{d}}\omega}{2\pi } \bigg\vert \sum_{\ell , m } \widehat{\omega}_\ell^m[\theta,\phi ] ( -i)^\ell \int_0^\infty { \text{d}}r ' r'^2 \mathcal{r}^<_\ell[\xi ' ] \rho_\ell^m[\omega , r ' ] \bigg\vert^2 . \end{aligned}\ ] ] in the subsequent two sections , we perform the decomposition in eq . for surface waves on the spherical mass @xmath8 as well as that of @xmath10 compact bodies orbiting it , and proceed to apply them to eq . .
in this section we will describe surface ( acoustic ) waves propagating on the large central mass @xmath8 of radius @xmath473 by @xmath507 \equiv \frac{m}{v_\odot } \delta r \ \delta[r - r_0].\end{aligned}\ ] ] where @xmath508 is the volume of @xmath8 , @xmath509 denote by @xmath510 $ ] the radius of the mass @xmath8 at a given time @xmath169 and direction @xmath511 from the spatial center of the coordinate system . for a non - relativistic system , which we shall assume
is the case in this section , the trace of the stress - energy tensor @xmath74 primarily describes its mass density @xmath512 . then eq
. may be interpreted as describing surface waves of very small fluctuations @xmath513
\equiv r[t,\widehat{x}]-r_0 $ ] around the mean radius @xmath473 , on an otherwise perfectly spherical body .
we will decompose these undulations @xmath514 as @xmath515 = r_0 \sum_{\ell=1}^\infty \sum_{m=0}^\ell \left ( \widehat{a}_\ell^m[\widehat{x } ] a_\ell^m \cos\left [ \omega_\ell^m t + \phi_\ell^m \right ] + \widehat{b}_\ell^m[\widehat{x } ] b_\ell^m \cos\left [ \omega'^m_\ell t + \psi_\ell^m \right ] \right),\end{aligned}\ ] ] with @xmath516 .
the `` rotated '' spherical harmonics @xmath517 are defined as , whenever @xmath518 , @xmath519 whereas for @xmath520 , @xmath521 because @xmath522 , the @xmath523 and @xmath524 are real .
this is the primary reason for using them instead of the usual @xmath101 , because now @xmath525 and @xmath526 can directly be read off as dimensionless amplitudes of a particular mode of vibration , with respective oscillation frequencies @xmath527 , @xmath528 , and phases @xmath529 and @xmath530 . for technical convenience , we will assume all oscillation frequencies @xmath531 are distinct and positive .
it is also worthwhile to observe we have not allowed a monopole @xmath174 term in the infinite sum eq . :
in particular , @xmath532 = 0 $ ] , and the total mass is a constant .
while our scalar field theories do not enjoy any sort of symmetry giving rise to a conservation law for the associated scalar charge , the coupling to the scalar fields here is of ( sub-)gravitational strength , and thus very weak .
therefore the requirement from the known laws of physics that mass is a conserved quantity , in the minkowski spacetime we are working in , therefore takes precedence . by a direct calculation
, one can check that our surface waves have the following decomposition @xmath533 = \int \frac{{\text{d}}\omega}{2\pi } e^{-i\omega t } \frac{m r_0}{v_\odot } \delta[r_0-r ] \pi \\ & \times \sum_{\ell=1}^\infty \sum_{m=0}^\ell \left ( \widehat{a}_\ell^m[\widehat{x } ] a_\ell^m \left ( e^{-i\phi_\ell^m}\delta[\omega-\omega_\ell^m ] + e^{i\phi_\ell^m}\delta[\omega+\omega_\ell^m ] \right ) + \widehat{b}_\ell^m[\widehat{x } ] b_\ell^m \left ( e^{-i\psi_\ell^m}\delta[\omega-\omega'^m_\ell ] + e^{i\psi_\ell^m}\delta[\omega+\omega'^m_\ell ] \right ) \right ) .
\nonumber\end{aligned}\ ] ] here , the analog of the @xmath534s appearing within the formula in eq . , are the coefficients of @xmath535 and @xmath536 . when taking the square @xmath537 in eq .
, we encounter cross terms involving the @xmath76-functions . however , since we have assumed that all frequencies are distinct and positive , the arguments of the @xmath76-functions can not be simultaneously zero unless the frequencies are in fact the same .
this collapses the summations into a single @xmath129- and a single @xmath103-sum .
it remains to deal with squares of the form @xmath538)^2 $ ] , with @xmath539 being one of the @xmath527s or @xmath528s .
we will treat one of the @xmath540 $ ] as the total duration of time , since @xmath541 = \int { \text{d}}t \lim_{\omega\to 0 } e^{-i\omega t } = \text{total time elapsed}.\end{aligned}\ ] ] dividing both sides of eq . by total time , i.e. @xmath542 $ ] , then gives us back power radiated in scalar waves per unit solid angle .
_ minimally coupled massless scalar _ the result for @xmath481 , from eq .
, is then @xmath543}{{\text{d}}t{\text{d}}\omega } & = \frac{9 g_\text{n } m^2}{\pi r_0 ^ 2 } \sum_{\ell=1}^\infty \sum_{m=0}^\ell \bigg\ { \left\vert \widehat{a}_\ell^m[\widehat{x } ] a_\ell^m \omega_\ell^m r_0 j_\ell[\omega_\ell^m r_0 ]
\right\vert^2 + \left\vert \widehat{b}_\ell^m[\widehat{x } ] b_\ell^m \omega'^m_\ell r_0 j_\ell[\omega'^m_\ell r_0 ] \right\vert^2 \bigg\}.\end{aligned}\ ] ] ( we have exploited the fact that , because @xmath544 $ ] is @xmath545 times a power series in @xmath546 , @xmath547| = |j_\ell[z]|$ ] . ) for non - relativistic systems , a substantial subset of the oscillation periods @xmath548 ought to be much longer than the light crossing time @xmath549
. therefore we can treat the small dimensionless quantities @xmath550 as expansion parameters .
this leads us , in this non - relativistic limit , to @xmath551}{{\text{d}}t{\text{d}}\omega } = \frac{9 g_\text{n } m^2}{\pi r_0 ^ 2 } \sum_{\ell=1}^\infty \sum_{m=0}^\ell \frac{1}{((2\ell+1)!!)^2 } \bigg\ { \left\vert \widehat{a}_\ell^m[\widehat{x } ] a_\ell^m \left ( \omega_\ell^m r_0 \right)^{\ell+1 } \right\vert^2 + \left\vert \widehat{b}_\ell^m[\widehat{x } ] b_\ell^m \left ( \omega'^m_\ell r_0 \right)^{\ell+1 } \right\vert^2 \bigg\}.\end{aligned}\ ] ] _ galileons _ for the galileon @xmath67 , because the vainshtein radius depends on @xmath552 in eq .
a free parameter in this paper it is not necessary that @xmath553 .
this prompts us to write the energy output @xmath554 due to galileons as a function of the mode numbers @xmath93 instead .
we will assume that the surface waves are non - relativistic , and thus the ratio of the radius to that time of oscillation is a small number ( i.e. @xmath555 ) , so that we may replace the bessel and hankel functions with their small argument limits .
when @xmath556 and @xmath216 , @xmath557}{{\text{d}}t{\text{d}}\omega } & = \frac{288 g_\text{n } m^2}{r_0 ^ 2 } \frac{r_0}{r_v } \left ( \frac{\gamma \left[-\frac{2 \ell}{3}-\frac{1}{3}\right ] \gamma \left[\frac{5}{6}-\frac{\ell}{3}\right]}{2^{3\ell+\frac{7}{2 } } \gamma\left[\frac{\ell } { 2}+\frac{3}{4}\right ] \gamma\left[\frac{\ell}{2}+\frac{5}{4}\right ] \gamma[-\ell ] } \left(\frac{\cos\left[\frac{1}{6 } \pi ( 2 \ell + 1)\right]}{\sin [ \pi \ell ] } + 1\right ) \right)^2 \\ & \times \bigg\ { \left(\omega_\ell^m r_0\right)^{\ell+2 } \left ( \omega_\ell^m r_v \right)^\ell \left\vert \widehat{a}_\ell^m[\widehat{x } ] a_\ell^m \right\vert^2 + \left(\omega'^m_\ell r_0 \right)^{\ell+2 } \left ( \omega'^m_\ell r_v \right)^\ell \left\vert \widehat{b}_\ell^m[\widehat{x } ] b_\ell^m \right\vert^2 \bigg\}. \nonumber\end{aligned}\ ] ] notice that for the @xmath344 case , the vainshtein radius @xmath0 drops out ; low frequency oscillations are thus unscreened to the lowest order . at higher than dipole
order , @xmath144 , by comparing the @xmath558 in the galileon result in eq . to the analogous @xmath559 in eq .
for the non - interacting massless scalar , we find that galileon power is actually enhanced ( up to @xmath129-dependent numerical factors ) by the ratio @xmath560 .
when @xmath561 $ ] , and @xmath216 , @xmath562}{{\text{d}}t{\text{d}}\omega } & = \frac{g_\text{n } m^2}{r_0 ^ 2 } \frac{3^{\frac{\ell}{2}+\frac{9}{4}}}{4^{\ell+1 } \gamma\left [ \frac{\ell}{2 } + \frac{5}{4 } \right]^2 } \bigg\ { \left\vert \widehat{a}_\ell^m[\widehat{x } ] a_\ell^m \right\vert^2 \frac{\left ( \omega_\ell^m r_0 \right)^{\ell+3}}{(\omega_\ell^m r_v)^\frac{3}{2 } } + \left\vert \widehat{b}_\ell^m[\widehat{x } ] b_\ell^m \right\vert^2 \frac{\left ( \omega'^m_\ell r_0 \right)^{\ell+3}}{(\omega'^m_\ell r_v)^{\frac{3}{2 } } } \bigg\ } .\end{aligned}\ ] ] for a fixed @xmath129 mode , we may take the ratio of the @xmath481 results in eq . to the ones here for high frequency oscillations : up to numerical @xmath129-dependent factors , we obtain @xmath563 .
for @xmath344 , this ratio is @xmath564 , telling us the dipole term in eq .
is vainshtein screened . however , in this non - relativistic limit , @xmath565 , once @xmath129 is large enough that @xmath566 , we see that high multipole galileon radiation becomes vainshtein amplified relative to their non - interacting cousins .
we also note that , while low frequency oscillations in eq .
contain integer powers of frequency , here the radiation spectrum contain @xmath567 and @xmath568 .
the expressions in equations and came directly from the bessel and hankel mode functions satisfying the original wave equations of @xmath67 and @xmath481 .
together with the presence of the combinations @xmath569 and @xmath570 , these facts teach us that radiation generated by the surface vibrations of the massive object @xmath8 directly probes not only the vibrations themselves but also the dynamics of our field theories in the galileon case , it carries information about the theory operating deep within the vainshtein radius of @xmath8 , where the self interactions of the full @xmath11 theory are dominant . at small ratios of the vainshtein radius to oscillation time scale ( eq . ) , we see that the galileon radiation spectrum , like its @xmath481 cousin in eq . , contains only integer powers of the oscillation frequencies .
however , for very large vainshtein radius to oscillation time scale ratios ( eq . ) , the power emitted begins to contain fractional powers of angular frequencies ; this indicates there must be a change in the spectral index if one is able to probe galileon radiation over a broad bandwidth .
this can be traced to the galileon radial mode functions @xmath385{\xi } j_{(1/4)(2\ell+1)}[\sqrt{3}\xi/2]$ ] within the vainshtein radius , @xmath571 .
we will witness this phenomenon again in the following section on the @xmath10 body radiative problem . in this section
, we consider an arbitrary number of compact bodies of masses @xmath572 orbiting around the central object @xmath8 . body ( weak field ) gravitational problem has a long history ; see , for instance , @xcite and the references within .
] we will assume this system is held together by only gravitational and scalar forces , and we will further approximate these compact bodies as point masses , with spatial position vectors @xmath573 .
if we work within the non - relativistic ( i.e. , slow motion ) approximation , valid for a wide range of astrophysical dynamics , including that of our solar system the galileon and gravitational interactions are described by @xmath574 the effects of gravitation are encoded in the proper time element @xmath575 , which we have expanded in powers of velocities and the graviton field @xmath576 .
we have assumed gravity is weak , @xmath577 , where @xmath578 . the virial theorem tells us that the potential @xmath576 scales as the square of the typical velocities @xmath579 , so that the @xmath580 " in eq . can be understood to scale as @xmath581 and higher .
( the scalar potential @xmath582 would scale similarly ; but the static portion of @xmath583 would be considerably weaker because of vainshtein screening . ) because we are interested in galileon and not in gravitational radiation , we may `` integrate out '' the gravitational field @xmath584 . to the lowest order in the non - relativistic expansion ,
this amounts to replacing the @xmath584 evaluated on the @xmath585th point mass world line with the gravitational potential exerted on @xmath586 by @xmath8 ; the potentials due to the other compact bodies , as long as they are distant enough , scale as @xmath587 , @xmath588 , relative to that due to @xmath8 , and hence are subdominant .
( we ignore the possible self - force contribution . ) the newtonian energy per unit mass @xmath589 is a constant ( up to corrections of @xmath590 $ ] ) this means we may replace @xmath591 in eq . with @xmath592 . at this point , what is relevant for the galileon radiation problem is the interaction @xmath593 \frac{\varphi \text { or } \xi}{{m_{\text{pl}}}},\end{aligned}\ ] ] with the non - relativistic expansion denoted by @xmath594 \equiv 1 - \dot{\vec{y}}_a^2 + \frac{e_a}{m_a } + \mathcal{o}\left [ v_a^3 , v_a^2 \frac{m}{m } \right ] .\end{aligned}\ ] ] by a direct calculation , the trace of the associated stress - energy in eq
. is @xmath595 & = \sum_a m_a \int { \text{d}}t ' f_a[t ' ] \delta[t - x^0[t ' ] ] \delta^{(3)}[\vec{x}-\vec{y}_a[t ' ] ] .\end{aligned}\ ] ] we rewrite this as @xmath595 = \int \frac{{\text{d}}\omega}{2\pi } e^{-i\omega t } \sum_{\ell = 0}^\infty \sum_{m=-\ell}^\ell y_\ell^m[\widehat{x } ] \rho_\ell^m[\omega , r],\end{aligned}\ ] ] where , by a change of variables from proper time to coordinate time , @xmath596 & = \sum_a m_a \int { \text{d}}t ' f_a[t ' ] e^{i\omega t ' } \frac{\delta\left [ r - |\vec{y}_a[t']| \right]}{r |\vec{y}_a[t']| } \overline{y_\ell^m}[\widehat{y}_a[t']].\end{aligned}\ ] ] with this @xmath534 , the total scalar energy emitted by @xmath10 bodies orbiting around the parent body @xmath8 is therefore given by .
@xmath597}{{\text{d}}\omega } = 32\pi g_\text{n } \int \frac{{\text{d}}\omega}{2\pi } \bigg\vert \sum_{a=1}^n m_a \int { \text{d}}t ' f_a[t ' ]
e^{i\omega t ' } \sum_{\ell=0}^\infty \sum_{m=-\ell}^\ell y_\ell^m[\widehat{x } ] \overline{y_\ell^m}[\widehat{y}_a[t ' ] ] ( -i)^\ell \mathcal{r}^<_\ell\big [ \omega |\vec{y}_a[t']| \big ] \bigg\vert^2.\end{aligned}\ ] ] _ minimally coupled massless scalar _ using equations , and @xmath598 y_\ell^m[\widehat{k } ] \overline{y_\ell^m}[\widehat{x } ] \end{aligned}\ ] ] eq .
becomes @xmath599}{{\text{d}}\omega } = \frac{2 g_\text{n}}{\pi } \int \frac{{\text{d}}\omega}{2\pi } \bigg\vert \sum_{a=1}^n m_a \int { \text{d}}t ' e^{i\omega t ' } \partial_{t'}\left\ { f_a[t ' ] \exp\left[-i\omega \widehat{x}\cdot\vec{y}_a[t ' ] \right ] \right\ } \bigg\vert^2 \ , \end{aligned}\ ] ] where we have also used the fact that every integer power of @xmath92 occurring within our integrand may be replaced with a time derivative ( namely , @xmath600 ) acting on @xmath170 dependent factors .
let us now taylor expand @xmath601 \right]$ ] , and convert each additional power of @xmath92 into an additional @xmath602 .
roughly speaking , each time derivative should scale as @xmath603 where @xmath604 is the typical orbital radii of the compact bodies in motion .
we have previously highlighted that the @xmath10-body system under consideration is held together by ( largely ) conservative forces , and it is therefore important to include the central body @xmath8 in our radiative calculations to respect the conservation of linear momentum .
this is an appropriate place to consider how the backreaction on the central mass @xmath8 by the compact bodies orbiting it affects our analysis here and below .
what we shall do is to extend the sum @xmath605 to include that of @xmath8 itself , via the definitions @xmath606 for the minimally coupled scalar case at hand , this is nothing but the principle of superposition , since the theory is linear . for the galileon case ,
this amounts to including in the matter perturbations @xmath607 in eq . the time dependent multipole moments of the central mass @xmath8 , , where @xmath608 . ] induced by the gravitational and scalar forces of its @xmath10 planetary companions pushing it away from the center of the spatial coordinate system .
( that all our results in this section depend on at least two time derivatives of the spatial coordinates of @xmath8 and the @xmath10 compact bodies corroborates this interepretation . )
this is subtly different from the non - self - interacting @xmath481 case because , to capture the vainshtein mechanism , we first had to assume that @xmath8 was motionless so that we could solve for the background @xmath28 it generated , before proceeding to compute the galileon green s function .
it is for this reason we have phrased our discussion of including @xmath8 in the sum over @xmath585 in terms of a backreaction on the motion of @xmath8 .
more quantitatively , the stress energy of the central mass is given by @xmath609 $ ] , and we may taylor expand it about @xmath610 .
the lowest order term is @xmath611 $ ] , which is the source of the background field @xmath28 ; we are thus treating every higher term in the expansion ( which is necessarily proportional to powers of @xmath612 $ ] ) as part of the matter perturbation @xmath607 . by taylor expanding @xmath601 \right]$ ] , followed by converting all @xmath92 s into time derivatives , we find that the non - relativistic scalar energy loss per unit solid angle from our @xmath10-body system is given by the expansion : @xmath613}{{\text{d}}\omega } = \frac{2 g_\text{n}}{\pi } \int \frac{{\text{d}}\omega}{2\pi } \left\vert \sum_{a=0}^n m_a \int { \text{d}}t ' e^{i\omega t ' } \left ( \widehat{x}\cdot\frac{{\text{d}}^2 \vec{y}_a}{{\text{d}}t'^2 } + \frac{1}{2}\frac{{\text{d}}^3 \left ( \widehat{x}\cdot\vec{y}_a \right)^2}{{\text{d}}t'^3 } - \frac{{\text{d}}\dot{\vec{y}}^2_a}{{\text{d}}t ' } + \mathcal{o}\left [ v_a^4 , v_a^3 \frac{m}{m } \right ] \right ) \right\vert^2,\end{aligned}\ ] ] the importance of including @xmath8 in the sum over @xmath585 is now manifest , for the first term involving the sum of all forces must yield zero , @xmath614 ; in a spacetime translation symmetric background , linear momentum is conserved and newton s third law must be obeyed
. therefore @xmath481 radiation really begins at @xmath615 $ ] . as a check of the formalism here , in eq .
below we shall re - derive the analog of eq . directly from the position space green s function in eq . , but ( for simplicity ) without including the gravitational potential in the proper time element .. ] we see that , at this order in the non - relativistic expansion , including the gravitational potential merely changes the @xmath616 in eq . to @xmath617 in eq . .
before we proceed to the galileon case , let us note that we could have obtained eq . directly from the infinite @xmath129-sum in eq .
, if we taylor expand the spherical bessel functions and use the explicit polynomial expressions for the @xmath618s .
( the reason for collapsing the infinite @xmath129-sum into an exponential in eq . is to emphasize that , because @xmath601 \right]$ ] admits a taylor expansion in integer powers of @xmath92 , the radiation spectrum in depends on frequency solely through time derivatives acting on the @xmath619s ; this will not be the case for galileons . ) because we are seeking an answer accurate up to @xmath620 $ ] , by counting powers of @xmath92 , we may infer that up to the @xmath621 terms of the sum are needed .
specifically , the @xmath622 term comes from the leading order piece of the @xmath344 term ; whereas the @xmath623 from the leading order piece of the @xmath621 term ; and the @xmath624 comes from the @xmath174 term with @xmath625 included .
furthermore , there is cancellation between the @xmath174 first order correction term involving @xmath626 and that from the @xmath621 term .
_ galileons _ in parallel with the treatment for @xmath481 , we will assume that our astrophysical @xmath10-body system is non - relativistic , so that the orbital time scale @xmath627 is very small compared to the light crossing time @xmath628 , allowing us to replace the bessel and hankel functions with their small argument limits . for low frequencies , @xmath134
, we have @xmath629}{{\text{d}}\omega { \text{d}}(\omega/2\pi ) } = \frac{2 g_\text{n}}{\pi } \left\vert \sum_{a=1}^n m_a \int { \text{d}}t ' e^{i\omega t ' } \sum_{\ell=0}^\infty \partial_{t'}^{\ell+1 } \left\ { f_a[t ' ] \mathcal{m}_\ell[t ' ] \right\ } \right\vert^2,\end{aligned}\ ] ] where @xmath630 & = & \left\ { \begin{array}{lr } 1 -\frac{(\omega |\vec{y}_a[t']|)^2}{4 } + \dots , & \ell=0 \\ & \\ \sqrt{\frac{|\vec{y}_a[t']|}{r_v } } \left ( r_v |\vec{y}_a[t']| \right)^{\frac{\ell}{2 } } p_\ell\left [ \widehat{x}\cdot\widehat{y}_a[t ' ] \right ] \frac{\sqrt{\pi } ( 2\ell+1 ) \gamma \left[-\frac{2 \ell}{3}-\frac{1}{3}\right ] \gamma \left[\frac{5}{6}-\frac{\ell}{3}\right]}{2^{3 \ell + \frac{3}{2 } } \gamma\left[\frac{\ell } { 2}+\frac{3}{4}\right ] \gamma\left[\frac{\ell}{2}+\frac{5}{4}\right ] \gamma[-\ell ] } \left(\frac{\cos\left[\frac{1}{6 } \pi ( 2 \ell + 1)\right]}{\sin [ \pi \ell ] } + 1\right ) , & \ell > 0 \\ \end{array } \right .
\ , \end{aligned}\ ] ] and we have converted the spherical harmonics sum to one over legendre polynomials .
the presence of @xmath631 implies the magnitude of the @xmath129th channel is suppressed by @xmath632 relative to the lowest @xmath174 mode , where @xmath35 is the typical orbital radius and @xmath633 is the typical orbital time scale ; the factor of @xmath634 further suppresses the power loss because the motion is highly non - relativistic .
( we have already noted that the leading monopole and dipole terms in the radial galileon green s function in the non - relativistic limit matches that of its non - interacting cousin . )
developing the non - relativistic expansion up to @xmath620 $ ] requires the monopole , dipole and quadrupole terms ( @xmath635 ) : in the low frequency limit , we gather @xmath636}{{\text{d}}\omega { \text{d}}(\omega/(2\pi ) ) } = \frac{2 g_\text{n}}{\pi } \bigg\vert \sum_{a=0}^n m_a \int { \text{d}}t ' e^{i\omega t ' } \bigg ( \widehat{x}\cdot\frac{{\text{d}}^2 \vec{y}_a}{{\text{d}}t'^2 } \\ & \qquad + \frac{1}{4 } \frac{{\text{d}}^3 \vec{y}_a^2}{{\text{d}}t'^3 } - \frac{{\text{d}}\dot{\vec{y}}_a^2}{{\text{d}}t ' } + \frac{\sqrt{3}\gamma[-\frac{2}{3 } ] \gamma[\frac{1}{6}]}{40\pi^{3/2 } } \frac{{\text{d}}^3}{{\text{d}}t'^3}\left ( \sqrt{\frac{r_v}{|\vec{y}_a[t']|}}\left ( \vec{y}_a^2 - 3 \left(\widehat{x}\cdot\vec{y}_a\right)^2 \right ) \right ) + \mathcal{o}\left [ v_a^4 , v_a^3 \frac{m}{m } \right ] \bigg ) \bigg\vert^2 , \qquad |\xi_v| \ll 1 . \nonumber\end{aligned}\ ] ] comparing equations and informs us that very low frequency galileon waves travel essentially unscreened , even though they are generated deep within the vainshtein radius of the system .
the @xmath637 factor in eq . may even yield vainshtein enhancement relative to its cousin @xmath481 .
moreover , we shall now argue that unlike eq .
, the first term @xmath638 here in eq . and in eq .
below , is small it scales as @xmath639 but is no longer exactly zero .
the primary reason is that , while gravitational forces between any two objects obey newton s third law , galileon forces between the compact bodies ( in the non - relativistic limit ) do not .
this is because the background @xmath28 does not respect spatial translation symmetry ; this statement can even be checked explicitly by taking the gradients of , say , the static green s function in eq . with respect to both @xmath640 and @xmath641 , and noting they do not give equal and opposite spatial vectors .
it does turn out , however , that the force between each compact body and the central mass is equal and opposite : without loss of generality we may consider some small mass @xmath103 lying on the positive @xmath642-axis .
the problem is now cylindrically symmetric , which tells us the force on @xmath8 due to @xmath103 ; and the force on @xmath103 due to @xmath8 , must both point along the @xmath642-axis .
the @xmath643 coupling tells us the @xmath642-component of the force on @xmath103 due to @xmath8 , to leading order , is simply @xmath644 ( refer to @xmath645 in eq . ) , where @xmath97 is the radial location of @xmath103 .
the force on @xmath8 due to @xmath103 is , in the non - relativistic limit , given by first invoking the cylindrical symmetry to replace @xmath646 in eq . with @xmath647 , and then computing @xmath648 = m m/(2 \pi { m_{\text{pl}}}^2 r_v^{3/2 } \sqrt{r'})$ ] .
solution generated by @xmath103 , whose gradient is responsible for the force acting on @xmath8 , can perform the following order - of - magnitude check . replace the @xmath11 in the lagrangian density of the action in eq . with the total field of @xmath8 and @xmath103 , i.e. @xmath649 $ ] , with @xmath171 $ ] given by eq . .
after expanding about @xmath650 , divide the dominant piece of the quadratic - in-@xmath651 portion of the resulting lagrangian density by the dominant piece of the cubic - in-@xmath651 term . for fixed @xmath97 , the radial location of @xmath103 , one should find the ratio to go as @xmath652 i.e. the linear solution offered by @xmath651 should be an excellent description of the force of @xmath103 on @xmath8 .
in the same spirit , one may also expand this same quadratic - to - cubic ratio about @xmath653 and find that , in the @xmath188 limit , nonlinearities begin to render the solution offered by @xmath651 invalid at distances closer to @xmath103 than @xmath654 . ]
note that even though newton s third law is not obeyed between the compact bodies , because the theory we started with in eq .
was defined in flat spacetime , total linear and angular momentum must still be conserved .
what must happen is that the radiation generated at @xmath655 $ ] carries away some of the linear momentum . at high frequencies but non - relativistic orbital speeds , i.e. @xmath656 , @xmath657}{{\text{d}}\omega { \text{d}}(\omega/2\pi ) } = \frac{32 g_\text{n}}{|\omega r_v|^{3/2 } } \left\vert \sum_{a=1}^n m_a \int { \text{d}}t ' e^{i\omega t ' } \sum_{\ell=0}^\infty ( -i)^\ell \partial_{t ' } \left\ { f_a[t ' ] \mathcal{m}_\ell[t ' ] \right\ } \right\vert^2,\end{aligned}\ ] ] where @xmath630 & = & \left\ { \begin{array}{lr } \frac{(-3)^{7/8}}{16\gamma[\frac{7}{4}]}\left ( 1 - \frac{(\omega |\vec{y}_a[t']|)^2}{4 } + \dots \right ) , & \ell=0 \\ & \\ \frac{3^{\frac{\ell}{4}+\frac{1}{8}}}{2^{\ell+1 } \gamma[\frac{\ell}{2}+\frac{1}{4 } ] } e^{i\pi\frac{2\ell+5}{8 } } \left ( \omega |\vec{y}_a[t']| \right)^{\frac{\ell+1}{2 } } p_\ell\left[\widehat{x}\cdot\widehat{y}_a[t']\right ] , & \ell > 0 \\ \end{array } \right . \
.\end{aligned}\ ] ] here , the prefactor @xmath658 exhibits vainshtein screening of high frequency galileon radiation . counting powers of @xmath92
tells us that an @xmath620$]-accurate answer receives contributions from the @xmath659 terms : @xmath660}{{\text{d}}\omega { \text{d}}(\omega/(2\pi ) ) } = \frac{2 \cdot 3^{\frac{3}{4 } } g_\text{n}}{|\omega r_v|^{3/2 } \gamma[\frac{3}{4}]^2 } \bigg\vert \sum_{a=0}^n m_a \int { \text{d}}t ' e^{i\omega t ' } \bigg ( \widehat{x}\cdot\frac{{\text{d}}^2 \vec{y}_a}{{\text{d}}t'^2 } \\ & \qquad - \frac{\sqrt[4]{-3 } \gamma[\frac{3}{4}]}{\gamma[\frac{1}{4 } ] } \frac{{\text{d}}^3}{{\text{d}}t'^3 } \left ( \frac{\vec{y}_a^2 - 3 \left ( \widehat{x}\cdot\vec{y}_a\right)^2}{\sqrt{\omega|\vec{y}_a| } } \right ) + \frac{1}{2\sqrt{3 } } \left ( \frac{1}{2 } \frac{{\text{d}}^3 \vec{y}_a^2}{{\text{d}}t'^3 } - 2 \frac{{\text{d}}\dot{\vec{y}}_a^2}{{\text{d}}t ' } \right ) + \mathcal{o}\left [ v_a^{7/2 } , v_a^3 \frac{m}{m } \right ] \bigg ) \bigg\vert^2 , \qquad |\xi_v| \gg 1 . \nonumber\end{aligned}\ ] ] ( the @xmath622 came from @xmath344 ; the second group involving @xmath661 from @xmath621 ; and the final group from @xmath174 with @xmath625 included . ) comparing equations and , we find , as we did in the previous section , that the slope of the galileon power spectrum depends on the size of @xmath662 . at low frequencies
the radiation is very similar to that of the non - interacting case , while at high frequencies , the spectrum acquires an overall suppression factor of @xmath663 .
the phenomenology of galileon radiation appears to be richer than its minimally coupled massless counterpart due to the existence of the additional length scale @xmath0 in the problem .
also note the presence of fractional powers of @xmath92 that can not be associated with time derivatives this aspect of the galileon radiation has no analog in its non - interacting cousin nor in gravitational waves propagating on flat spacetime .
_ gravitational waves _ let us also record the power spectrum of gw emission .
arguments based on conservation of energy and the validity of newton s third law leads us to infer that the lowest order answer arises from the third time derivative of the quadrupole moment of the system .
misner , thorne and wheeler @xcite equations 36.1 and 36.2 tell us @xmath664 \right\vert^2,\end{aligned}\ ] ] where @xmath665 is the fourier transform of the triple time derivative of the quadrupole moment , @xmath666 & \equiv \int { \text{d}}t ' e^{i\omega t ' } \sum_{a=1}^n m_a \frac{{\text{d}}^3}{{\text{d}}t'^3 } \left ( y_{ai}[t ' ] y_{aj}[t ' ] - \frac{1}{3 } \delta_{ij } \vec{y}_a^2[t ' ] \right ) .\end{aligned}\ ] ] as we have seen , the forces acting between compact bodies orbiting around a central mass @xmath8 no longer obey newton s third law if galileons exist ( @xmath667 ) this in turn means , eq .
may no longer be the leading order answer to the gw spectrum .
we hope to return to this issue in the future .
within the next decade or so , gravitational wave detectors are expected to begin hearing signals from astrophysical systems such as inspiralling compact binaries . it is therefore an appropriate time to explore the possibility that there could be emission of radiation due to additional degrees of freedom coming from the various modifications of general relativity that has been proposed in the literature to date .
our work initiates such an investigation for galileons , a class of scalar field theories that exhibit what is known as the vainshtein mechanism , within the context where there is a large central mass @xmath8 .
we have constructed the galileon retarded green s function satisfying the linearized equations of motion about the background field of this central mass .
the main results are described in section . for the radiation problem , where the observer is situated at a very large distance from @xmath8 and
the source generating galileon waves is located well within the vainshtein radius of @xmath8 , the primary results can be found in equations through . we have also obtained the exact static green s function , the wkb limit of the retarded green s function , and the green s function evaluated both deep inside and far outside the vainshtein radius of @xmath8 .
we have used this radiative green s function to obtain the frequency spectrum of galileon radiation emitted from ( acoustic ) waves on the surface of the spherical mass @xmath8 , described by equations , , , and .
the power dissipated per unit solid angle per mode @xmath93 , in the non - relativistic limit , can be found in eq . for the case where the @xmath0-to - oscillation - time - scale ratio was much smaller than unity ; and in for that ratio very large . to illustrate the importance of the nonlinearities of the galileon interaction for the problem at hand , we have also calculated for comparison the power emitted if we replace the galileon with a non - interacting massless scalar , with the relevant results given in equations and .
a particularly interesting application of our results is to the radiation spectrum due to the motion of @xmath10 point masses gravitationally bound to the central mass @xmath8 .
this is the dissipative aspect of the galileon modified gravitational dynamics whose conservative portion we investigate in a separate paper @xcite .
we have focused on the non - relativistic limit , and have found the energy in galileon radiation lost to infinity in two regimes . for small @xmath0-to - orbital - time - scale ratios ,
this is given by eq .
, and the lowest order answer is . for large @xmath0-to - orbital - time - scale ratios ,
the answer is eq .
, with the lowest order result in eq . . for comparison , the non - interacting massless scalar result is eq .
and that of the quadrupole radiation formula for gravitational waves is eq . .
in these radiative processes , we find that the non - interacting massless scalar and galileon radiation are comparable in the non - relativistic , low frequency , and low multipole regime . moreover , in this limit galileon radiation is actually amplified relative to its non - interacting counterpart for higher multipoles these findings are a direct consequence of the structure of the galileon radial green s function , which we have already highlighted in the discussion surrounding equations and . in the high frequency limit , we confirm the anticipated vainshtein screening of the galileo radiation at low multipole orders ; demonstrating it to be of @xmath668 $ ] relative to its non - interacting counterpart . at high enough multipoles ,
however , high frequency galileon radiation becomes enhanced relative to its non - interacting counterpart .
moreover , for the astrophysical @xmath10 body system , where newton s third law is obeyed between each compact body and @xmath8 but not between the compact bodies themselves , the leading @xmath655 $ ] terms in the radiation formulas , equations and , are small ( scaling as @xmath669 $ ] ) but non - zero , in contrast with the non - self - interacting scalar case , where @xmath670 . having developed some quantitative understanding of the production of galileon radiation in this paper , let us remark that , if galileon waves exist , they are in principle detectable by gw detectors . about flat minkowski spacetime , the sum of the graviton - matter and galileon - matter coupling ( eq . ) is @xmath671 this implies that , if galileons are present , ordinary matter experiences an effective weakly curved metric of the form @xmath672 and the tidal forces experienced by the arms of the interferometers of gw detectors would now be due to both the transverse - traceless graviton @xmath673 and the galileon waves @xmath67 .
some of the calculations we have carried out have overlap with other recent work on the same topic @xcite . in this paper , we have sought to understand the galileon radiation spectrum in both the high and low frequency limits , and have found that the slope of the power spectrum of galileon radiation should be a non - trivial function of @xmath100 , the ratio of the vainshtein radius @xmath0 to the typical wavelength of the emitted waves .
( for instance , the radiative limit of our retarded green s function , namely , eq . and the coefficients @xmath124s described in and , have very different @xmath100-dependence for @xmath134 compared to @xmath409 . ) in @xcite , the focus was on the power loss from binary pulsar systems such as psr b1913 + 16 , and thus the authors carried out an analysis valid in the high frequency limit @xcite . , where @xmath674 ev is the current hubble parameter . for binary pulars like psr b1913 + 16 , with masses on the order of a few solar masses ,
the corresponding vainshtein radius is @xmath675 $ ] light years . because the period of typical binaries ( @xmath676 ) are of the order of a few hours , therefore @xmath409 . ] in the limit @xmath409 we find that the galileon power scales as @xmath677 , in agreement with the results of @xcite .
we note that , a priori , caution is required in interpreting the results both in this paper and in @xcite as describing a comparable mass binary pulsar system , because it is unclear if such a binary can be treated within the perturbative framework , given that the nonlinearities of galileons are very important when the motion is taking place deep within the collective vainshtein radius of the system itself . and
@xmath678 , is modeled by adding and subtracting to the binaries stress energies a monopole term with stress energy given by @xmath679 = ( m_1+m_2 ) \delta^{(3)}[\vec{x}]$ ] .
the field generated by @xmath680 is then used as a background , and the stress - energies of the pair of point masses themselves minus the stress energy of the central monopole ( see their eq .
( 2.6 ) ) are treated as perturbations . however , since , as in this paper , the linearized equation of motion about the background of the monopole are solved , the subtracted monopole really plays no role , and this scheme is really equivalent to the setup where there is one central mass @xmath681 and two masses in orbit around it , one of mass @xmath682 and the other @xmath678 .
but since @xmath682 and @xmath678 are comparable in magnitude to @xmath683 , and there is no small dimensionless ratio one may use as an expansion parameter , it is not evident for the general binary problem that the @xmath684 are mere perturbations on top of the @xmath8 .
more quantitatively , recall the mode functions evaluated deep within the vainshtein radius can be expressed in a separated form , reflecting the spherical symmetry of the background . in a comparable mass relativistic binary system , this spherical symmetry is completely absent .
thus , the separation of variables technique may not be useful in solving the general binary problem .
] however , the authors of @xcite have explicitly verified the validity of their perturbative scheme , and will present it in an upcoming publication @xcite . _ future work _ it would be worthwhile to convert the frequency space calculation in this paper to a real time one , in order to better understanding the physical meaning behind the fractional powers of angular frequency found in the galileon emission spectrum .
to this end , the identification of the correct contour prescriptions in the fourier integral of eq .
would be necessary .
we also have left unexplored a large range of @xmath662 , as we have only examined the extreme limits @xmath134 and @xmath409 .
furthermore , since the nonlinearities of the galileon theory play such a crucial role in its dynamics , we hope in the near future to go beyond the linear analysis about the background @xmath28 due to @xmath8 .
we could compute , say within the born approximation , the first correction arising from the cubic self - interactions in eq . to the wave solutions we have obtained here , so as to better understand the domain of validity of the results in this paper .
there are a number of interesting further directions for future work . in 4 dimensional flat spacetime
, one should introduce the quartic and quintic galileon terms and carry out an analogous analysis to that performed here .
one may also wish to develop an understanding of the backreaction of the power loss on the motion of the @xmath10 point masses .
in this context , there are other processes one could consider for example , one could carry out a calculation analogous to the one found in @xcite and @xcite for the gravitational case , in which one small mass scatters off @xmath8 , producing galileon bremsstrahlung radiation in the process .
we would like to thank melinda andrews , lasha berezhiani , claudia de rham , kurt hinterbichler , justin khoury , denis klevers , andrew matas , and andrew tolley for helpful discussions .
this work was supported in part by the us department of energy , nsf grant phy-0930521 , and nasa atp grant nnx11ai95 g .
we have made extensive use of mathematica @xcite to carry out the calculations in this paper .
the central goal of this section is an alternate derivation of eq . , but ( for technical simplicity ) without the inclusion of gravitational interactions .
we wish to compute the non - relativistic power emitted in non - interacting massless scalar radiation by the motion of @xmath10 compact bodies .
we will do so by finding the scalar field and its first derivatives generated by these point masses in flat spacetime ; and from these gradients construct @xmath499 .
however , we will first derive a expression in curved spacetime and proceed to specialize to minkowski spacetime .
the reason for doing so is that the galileon @xmath67 field generated by the @xmath10 body system we examined in section is related to the problem of a non - interacting scalar in the geometry given in eq . .
let the @xmath10 masses be @xmath572 and their spacetime locations @xmath685 .
the massless scalar theory in question is the curved spacetime generalization of the @xmath481-theory in eq . , namely @xmath686 with proper times @xmath687 from the hadamard form of the scalar green s function in eq . ,
@xmath688}{4\pi}\left ( \delta[\sigma ] u_{x , x ' } + \theta[\sigma ] v_{x , x'}\right),\end{aligned}\ ] ] with @xmath689 ( for an explanation of these symbols , see @xcite ) , the solution to @xmath481 is given by the integral @xmath690 & = \sum_{a=1}^n \frac{m_a}{4\pi { m_{\text{pl } } } } \int { \text{d}}s_a \theta[t - y_a^0[s_a ] ] \left ( \delta[\sigma_{x , y_a } ] u_{x , y_a } + \theta[\sigma_{x , y_a } ] v_{x , y_a } \right)\nonumber,\end{aligned}\ ] ] for a fixed @xmath78 , @xmath691 constant defines the geodesic joining @xmath78 ( the observation point ) to @xmath692 ( the source point ) , provided this geodesic is unique ( an assumption used when deriving the hadamard form in eq . ) .
thus it must be possible to invert @xmath693 for @xmath694 and vice versa .
in particular , in the following , we will need @xmath695 \
.\end{aligned}\ ] ] and hence @xmath696 the null cone piece of @xmath481 involves @xmath697 $ ] which we may then write @xmath698 = \left .
\frac{\delta[s_a-\tau_a]}{|{\text{d}}\sigma_{x , y_a}/{\text{d}}s_a| } \right\vert_{s_a = \tau_a},\end{aligned}\ ] ] where @xmath699 , the retarded time , is defined to be the proper time of the @xmath585th mass when it lies on the backward null cone of the observer at @xmath78 , @xmath700 \equiv 0.\end{aligned}\ ] ] the `` backward '' part of the requirement is reinforced by the step function @xmath701 $ ] in eq . .
the curved spacetime scalar field produced by the @xmath10 point masses is therefore @xmath702 & = \sum_{a=1}^n \frac{m_a}{4\pi{m_{\text{pl } } } } \bigg\ { \left ( \frac{\delta^{1/2}[x , y_a]}{|{\text{d}}\sigma_{x , y_a}/{\text{d}}s_a| } \right)_{s_a = \tau_a } + \int_{-\infty}^{\tau_a^- } { \text{d}}s ' v[x , y_a[s ' ] ] \bigg\}.\end{aligned}\ ] ] we may obtain the gradients of @xmath481 by differentiating the integral representation in eq . .
@xmath703 & = \sum_a \frac{m_a}{4\pi{m_{\text{pl } } } } \int { \text{d}}s_a \bigg ( \nabla_\alpha \sigma \delta'[\sigma ] u + \delta[\sigma ] \nabla_\alpha u + \delta[\sigma ] \nabla_\alpha \sigma v + \theta[\sigma ] \nabla_\alpha v \bigg).\end{aligned}\ ] ] the @xmath704 $ ] term may be re - written , holding @xmath78 fixed , @xmath705 = \frac{{\text{d}}s_a}{{\text{d}}\sigma_{x , y_a } } \frac{{\text{d}}}{{\text{d}}s_a } \frac{\delta[s_a-\tau_a]}{|{\text{d}}\sigma[s_a=\tau_a]/{\text{d}}s_a|},\end{aligned}\ ] ] which allows us to integrate by parts to obtain @xmath706 = \sum_{a=1}^n \frac{m_a}{4\pi{m_{\text{pl } } } } \bigg ( \bigg\ { \frac{\nabla_{\alpha } u_{x , y_a}}{|{\text{d}}\sigma_{x , y_a}/{\text{d}}s_a| } + \frac{\nabla_{\rho ' } \nabla_{\alpha } \sigma_{x , y_a } u_{x , y_a } + \nabla_{\rho ' } u_{x , y_a } \nabla_{\alpha } \sigma_{x , y_a}}{({\text{d}}\sigma_{x , y_a}/{\text{d}}s_a)^2 } \frac{{\text{d}}y_a^{\rho'}}{{\text{d}}s_a } \\ & \qquad + \frac{\nabla_{\alpha } \sigma_{x , y_a } u_{x , y_a}}{|{\text{d}}\sigma_{x , y_a}/{\text{d}}s_a|^3 } \left ( \frac{{\text{d}}^2 y_a^{\lambda'}}{{\text{d}}s_a^2 } \nabla_{\lambda ' } \sigma_{x , y_a } + \frac{{\text{d}}y_a^{\lambda'}}{{\text{d}}s_a } \frac{{\text{d}}y_a^{\kappa'}}{{\text{d}}s_a } \nabla_{\kappa ' } \nabla_{\lambda ' } \sigma_{x , y_a } \right ) + \frac{\nabla_{\alpha } \sigma_{x , y_a } v_{x , y_a}}{|{\text{d}}\sigma_{x , y_a}/{\text{d}}s_a| } \bigg\}\bigg\vert_{s_a=\tau_a } + \int_{-\infty}^{\tau_a^- } { \text{d}}s_a \nabla_\alpha v_{x , y_a } \bigg ) , \nonumber\end{aligned}\ ] ] with @xmath707 .
the primed derivatives are with respect to @xmath692 ( the @xmath585th point mass location ) and the unprimed ones with respect to @xmath78 ( the observer location ) .
_ minkowski spacetime _ in minkowski spacetime , @xmath708 and @xmath709 , while the world function reads @xmath710 so that its derivatives are @xmath711 and @xmath712 for convenience we shall use a dot to represent a derivative with respect to @xmath694 .
the gradient of @xmath481 in minkowski spacetime generated by @xmath10 point masses is given by @xmath713 & = -\sum_{a=1}^n \frac{m_a}{4\pi{m_{\text{pl } } } } \bigg\ { \frac{\dot{y}_a^\mu}{(\dot{y}_a^\kappa ( y_a - x)_\kappa)^2 } -\frac{(x - y_a)^\mu}{|\dot{y}_a^\kappa ( y_a - x)_\kappa|^3 } \left ( \ddot{y}_a^\lambda ( y_a - x)_\lambda + \dot{y}_a^2 \right ) \bigg\ } \bigg\vert_{s_a=\tau_a } \ , \end{aligned}\ ] ] where the retarded proper time @xmath699 is the solution of the equation @xmath714 = |\vec{x}-\vec{y}_a[\tau_a]|.\end{aligned}\ ] ] comparing equations and , we see that in flat spacetime , the field detected by an observer ( or felt by some test mass ) at @xmath78 is a function of the positions of the @xmath10 point masses evaluated at the appropriate retarded times that is how long it takes the signal to reach the observer from the source .
whereas in curved spacetime , where huygens principle no longer holds , the problem of motion is a richer one , because the proper time integral involving the tail term @xmath715 in eq . receives contributions from the point masses entire past histories .
( the electromagnetic counterpart to this statement can be found in @xcite , in which the @xmath716 and @xmath717 analogues of equations and are derived . )
we are now ready to determine the power radiated by the @xmath10 point masses moving in flat spacetime .
let us assume the motion of these @xmath10 bodies is confined within some finite spatial volume containing the center of the spatial coordinate system ; this will certainly be true when these @xmath10 bodies are bound together by their mutual gravity , which is the central theme of this paper .
we wish to extract the piece of @xmath481 that represents radiation .
as already explained earlier , in minkowski spacetime , the radiative piece of @xmath481 is the @xmath29 piece . from the result in eq .
, we see that this can be identified by counting powers of @xmath718 ; this isolates the acceleration @xmath719 term .
if we let the observer lie at some very large radius , we deduce @xmath720 and @xmath721 . recalling the notation @xmath722 and @xmath723 , @xmath724 & = -\frac{\left ( 1,\widehat{x } \right)}{4\pi r } \left.\sum_{a=1}^n \frac{m_a}{{m_{\text{pl } } } } \frac{\ddot{y}_a^\alpha \eta_{\alpha\beta } \left ( 1 , \widehat{x } \right)^\beta}{|\dot{y}_a^0 - \dot{\vec{y}}_a\cdot\widehat{x}|^3 } \right\vert_{s_a=\tau_a } \ , \end{aligned}\ ] ] and the power radiated to infinity per unit solid angle @xmath725 is @xmath726 to take the non - relativistic limit , we first recall that @xmath727 , where the retarded time @xmath170 satisfies eq .
, i.e. @xmath728|$ ] .
that retarded time depends on the trajectory @xmath619 means , upon carrying out the retarded time derivatives with respect to @xmath170 in eq .
, we still need to taylor expand every time dependent expression in powers of @xmath729|/r$ ] , because the latter expansion introduces further time derivatives .
it is at this point that terms containing the same number of time derivatives ( now evaluated at the approximate retarded time @xmath730 ) are considered to be of the same order in the non - relativistic expansion .
up to @xmath620 $ ] we have @xmath731 \right ) \right\vert_{t'=t - r } \right)^2.\end{aligned}\ ] ]
in this section we review the construction of the green s function for linear second order ordinary differential equations ( odes ) in terms of their homogeneous solutions .
consider the differential operator @xmath732 \frac{{\text{d}}^2}{{\text{d}}z^2 } + p_1[z ] \frac{{\text{d}}}{{\text{d}}z } + p_0[z ] \ , \end{aligned}\ ] ] where @xmath733 $ ] are smooth functions of @xmath642 .
the homogeneous solution @xmath734 to the corresponding linear second order ode satisfies @xmath735 = 0 $ ] , and the associated symmetric green s function obeys @xmath736 = \mathfrak{d}_{z ' } g[z , z ' ] = \lambda[z ] \delta[z - z ' ] = \lambda[z ' ] \delta[z - z ' ] \ .\ ] ] to solve this , first assume @xmath737 , so that @xmath738 = 0 $ ] .
the problem is thus reduced to solving the homogeneous equation @xmath739 = \mathfrak{d}_{z ' } g[z , z ' ] = 0 $ ] .
let @xmath740 $ ] be a pair of linearly independent homogeneous solutions , i.e. @xmath741 = \mathfrak{d}_z f_2[z ] = 0 , \quad { \sf wr}_{(z)}[f_1,f_2 ] \neq 0,\end{aligned}\ ] ] where the wronskian wr is defined to be @xmath742 \equiv f_1[z ] ( f_2)'[z ] - ( f_1)'[z ] f_2[z].\end{aligned}\ ] ] because @xmath739 = 0 $ ] , we must have @xmath743 = \alpha^\text{i } f_\text{i}[z ] , \qquad \text{i } = 1,2\end{aligned}\ ] ] where the @xmath744s are @xmath642-independent .
similarly , @xmath745 = 0 $ ] means @xmath746 , \qquad \text{j } = 1,2,\end{aligned}\ ] ] where @xmath747 is now a @xmath748 matrix of @xmath749-independent constants .
the same argument would hand us , for @xmath750 , @xmath751 = a_<^\text{ij } f_\text{i}[z ] f_\text{j}[z ' ] , \qquad \text{i},\text{j } = 1,2\end{aligned}\ ] ] if @xmath752 $ ] were not continuous at @xmath753 its second derivative with respect to @xmath642 or @xmath754 there , and hence the second line of eq .
, would involve a derivative of @xmath738 $ ] .
that implies we may assume @xmath752 $ ] is continuous at @xmath753 .
@xmath755 f_\text{j}[z ] = a_>^\text{ij } f_\text{i}[z ] f_\text{j}[z]$ ] imposes the conditions @xmath756 we now integrate @xmath739 = \lambda[z ] \delta[z - z']$ ] around the neighborhood of @xmath753 , applying integrating - by - parts as many times necessary to shift all the derivatives acting on @xmath752 $ ] onto the @xmath757 . by continuity , the @xmath758 term , @xmath759_{z = z'^-}^{z = z'^+}$ ] , @xmath760_{z = z'^-}^{z = z'^+}$ ] and the remaining integral involving @xmath761 ( with no derivatives acting it ) drop out assuming @xmath762 are smooth leaving us with @xmath763 = p_2[z ' ] \left [ \frac{\partial g[z , z']}{\partial z }
\right]_{z = z'^-}^{z = z'^+}.\end{aligned}\ ] ] employing the continuity conditions in equations , one would find the @xmath764 and @xmath765 do not contribute to @xmath766 $ ] .
we may use to eliminate , say @xmath767 , and find becomes @xmath768 = -p_2[z ' ] ( a_<^{21}-a_>^{21 } ) { \sf wr}_{(z')}[f_1,f_2]\end{aligned}\ ] ] because any `` rotation '' of the pair @xmath769 , i.e. the pair @xmath770 for any @xmath748 invertible @xmath771 , is an equally valid pair of linearly independent homogeneous solutions , without loss of generality we may choose @xmath772 , such that now @xmath773 = -p_2[z ] { \sf wr}_{(z)}[f_1 , f_2]\end{aligned}\ ] ] therefore , the general solution to eq . is @xmath774 & = \mathcal{c } f_1[z _ > ] f_2[z _ < ] - ( 1-\mathcal{c } ) f_1[z _ < ] f_2[z _ > ] \nonumber\\ & \qquad + c^{11 } f_1[z ] f_1[z ' ] + c^{22 } f_2[z ] f_2[z'],\end{aligned}\ ] ] where @xmath775 ( or @xmath776 ) is the greater ( or smaller ) of the pair @xmath777 , and @xmath778 , @xmath779 and @xmath780 are arbitrary constants .
these coefficients will be fixed by the boundary conditions of the given physical problem . by using the differential equation obeyed by the @xmath769
, one may readily show that @xmath781 = -\frac{p_1[z]}{p_2[z ] } { \sf wr}_{(z)}[f_1,f_2].\end{aligned}\ ] ] this in turn means the wronskian of two linearly independent solutions , and hence the measure @xmath782 $ ] , can be solved up to an overall constant , without first solving for the homogeneous solutions .
recalling eq .
, we gather @xmath783 = \chi \ p_2[z ] \exp\left [ -\int^z { \text{d}}z '' \frac{p_1[z'']}{p_2[z '' ] } \right],\end{aligned}\ ] ] where @xmath377 is the constant . in solving for @xmath752 $ ] , we will thus first choose a value for @xmath377 , and use this choice to fix the normalization of the product of the solutions @xmath784 f_2[z_<]$ ] in eq . .
we wish to emphasize , because we are constructing a symmetric green s function , one that is a green s function with respect to both variables @xmath642 and @xmath754 , we have just seen that one does not have a choice in picking the measure @xmath782 $ ] , but rather @xmath782 $ ] is fixed up to an overall numerical constant by @xmath785 $ ] in eq . and the wronskian of any two linearly independent homogeneous solutions
. to summarize , once a pair of homogeneous solutions @xmath769 are known , the symmetric green s function has the general solution given in equation .
the measure @xmath782 $ ] multiplying the @xmath76-functions in eq . is given by eq .
, and the overall constant @xmath377 there needs to be chosen . in a given problem
for this paper it is the solution of @xmath786 $ ] in eq .
the analog of @xmath778 , @xmath787 , and @xmath788 ( or the @xmath789 and @xmath790 ) will be fixed by appropriate regularity and boundary conditions . because @xmath782 $ ] has been computed , the overall normalization of the products @xmath784 f_2[z_<]$ ] is determined by the wronskian condition in eq . .
@xmath76-function measure _ let us conclude this section by justifying the @xmath83 measure on the right hand side of eq . .
equation informs us that we can determine this measure up to a constant , by integrating the ratio of @xmath791 to @xmath792 .
this may be achieved by using the explicit expressions in through yielding @xmath793}{\xi '' e_2[\xi'',\xi_v ] } { \text{d}}\xi '' \right ] = \left ( \xi(\xi^3+\xi_v^3)\right)^{-\frac{1}{2}}.\end{aligned}\ ] ] this immediately tells us that the measure multiplying the @xmath76-functions in eq .
is @xmath794 \exp\left [ -\int^\xi \frac{2 e_3[\xi'',\xi_v]}{\xi '' e_2[\xi'',\xi_v ] } { \text{d}}\xi '' \right ] = \frac{\chi}{\xi^2},\end{aligned}\ ] ] where @xmath377 is a constant . far away from the central mass @xmath8 , using equations and to keep only the most dominant terms in @xmath107 ( equations through ) , the left hand side of eq .
yields , as expected , the flat spacetime minimally coupled massless scalar wave equation @xmath795 & = \frac{\chi}{rr ' } \delta[t - t']\delta[r - r ' ] \delta[\cos\theta-\cos\theta ' ] \delta[\phi-\phi ' ] \ .\end{aligned}\ ] ] therefore choosing @xmath796 amounts to adhering to the usual convention .
a. nicolis , r. rattazzi and e. trincherini , phys .
d * 79 * , 064036 ( 2009 ) [ arxiv:0811.2197 [ hep - th ] ] .
g. r. dvali , g. gabadadze and m. porrati , phys .
b * 485 * , 208 ( 2000 ) [ hep - th/0005016 ] .
a. i. vainshtein , `` to the problem of nonvanishing gravitation mass , '' phys .
b * 39 * , 393 ( 1972 ) . c. deffayet , g. r. dvali , g. gabadadze and a. i. vainshtein , `` nonperturbative continuity in graviton mass versus perturbative discontinuity , '' phys .
d * 65 * , 044026 ( 2002 ) http://arxiv.org/abs/hep-th/0106001[[hep-th/0106001 ] ] .
m. a. luty , m. porrati and r. rattazzi , `` strong interactions and stability in the dgp model , '' jhep * 0309 * , 029 ( 2003 ) http://arxiv.org/abs/hep-th/0303116[[hep-th/0303116 ] ] .
a. nicolis and r. rattazzi , `` classical and quantum consistency of the dgp model , '' jhep * 0406 * , 059 ( 2004 ) http://arxiv.org/abs/hep-th/0404159[[hep-th/0404159 ] ] . c. deffayet , g. esposito - farese and a. vikman , `` covariant galileon , '' phys .
d * 79 * , 084003 ( 2009 ) http://arxiv.org/abs/0901.1314[[arxiv:0901.1314 [ hep - th ] ] ] . c. deffayet
, s. deser and g. esposito - farese , `` generalized galileons : all scalar models whose curved background extensions maintain second - order field equations and stress - tensors , '' phys .
d * 80 * , 064015 ( 2009 ) http://arxiv.org/abs/0906.1967[[arxiv:0906.1967 [ gr - qc ] ] ] . c. de rham and a. j. tolley , `` dbi and the galileon reunited , '' jcap * 1005 * , 015 ( 2010 ) http://arxiv.org/abs/1003.5917[[arxiv:1003.5917 [ hep - th ] ] ] . c. deffayet , s. deser and g. esposito - farese , `` arbitrary @xmath797-form galileons , '' phys .
d * 82 * , 061501 ( 2010 ) http://arxiv.org/abs/1007.5278[[arxiv:1007.5278 [ gr - qc ] ] ] .
a. padilla , p. m. saffin and s. -y .
zhou , `` bi - galileon theory i : motivation and formulation , '' jhep * 1012 * , 031 ( 2010 ) http://arxiv.org/abs/1007.5424[[arxiv:1007.5424 [ hep - th ] ] ] . c. deffayet , o. pujolas , i. sawicki and a. vikman , `` imperfect dark energy from kinetic gravity braiding , '' jcap * 1010 * , 026 ( 2010 ) http://arxiv.org/abs/1008.0048[[arxiv:1008.0048 [ hep - th ] ] ] . k. hinterbichler , m. trodden and d. wesley , `` multi - field galileons and higher co - dimension branes , '' phys . rev .
d * 82 * , 124018 ( 2010 ) http://arxiv.org/abs/1008.1305[[arxiv:1008.1305 [ hep - th ] ] ] .
g. l. goon , k. hinterbichler and m. trodden , `` stability and superluminality of spherical dbi galileon solutions , '' phys . rev .
d * 83 * , 085015 ( 2011 ) http://arxiv.org/abs/1008.4580[[arxiv:1008.4580 [ hep - th ] ] ] .
j. khoury , j. -l .
lehners and b. a. ovrut , `` supersymmetric galileons , '' phys . rev .
d * 84 * , 043521 ( 2011 ) http://arxiv.org/abs/1103.0003[[arxiv:1103.0003 [ hep - th ] ] ] .
g. goon , k. hinterbichler and m. trodden , `` symmetries for galileons and dbi scalars on curved space , '' jcap * 1107 * , 017 ( 2011 ) http://arxiv.org/abs/1103.5745[[arxiv:1103.5745 [ hep - th ] ] ]
. m. trodden and k. hinterbichler , `` generalizing galileons , '' class .
* 28 * ( 2011 ) 204003 http://arxiv.org/abs/1104.2088[[arxiv:1104.2088 [ hep - th ] ] ] .
g. goon , k. hinterbichler and m. trodden , `` a new class of effective field theories from embedded branes , '' phys .
lett . * 106 * , 231102 ( 2011 ) http://arxiv.org/abs/1103.6029[[arxiv:1103.6029 [ hep - th ] ] ] .
c. burrage , c. de rham and l. heisenberg , `` de sitter galileon , '' jcap * 1105 * , 025 ( 2011 ) http://arxiv.org/abs/1104.0155[[arxiv:1104.0155 [ hep - th ] ] ] . g. goon ,
k. hinterbichler and m. trodden , `` galileons on cosmological backgrounds , '' jcap * 1112 * , 004 ( 2011 ) http://arxiv.org/abs/1109.3450[[arxiv:1109.3450 [ hep - th ] ] ] . s. -y
. zhou and e. j. copeland , `` galileons with gauge symmetries , '' http://arxiv.org/abs/1112.0968[arxiv:1112.0968 [ hep - th ] ] .
g. goon , k. hinterbichler , a. joyce and m. trodden , `` gauged galileons from branes , '' http://arxiv.org/abs/1201.0015[arxiv:1201.0015 [ hep - th ] ] .
g. goon , k. hinterbichler , a. joyce and m. trodden , jhep * 1206 * , 004 ( 2012 ) [ arxiv:1203.3191 [ hep - th ] ] .
g. gabadadze , k. hinterbichler , j. khoury , d. pirtskhalava and m. trodden , arxiv:1208.5773 [ hep - th ] . c. de rham and g. gabadadze , `` generalization of the fierz - pauli action , '' phys .
d * 82 * , 044020 ( 2010 ) http://arxiv.org/abs/1007.0443[[arxiv:1007.0443 [ hep - th ] ] ] . c. de rham , g. gabadadze and a. j. tolley , `` resummation of massive gravity , '' phys .
lett . * 106 * , 231101 ( 2011 ) http://arxiv.org/abs/1011.1232[[arxiv:1011.1232 [ hep - th ] ] ] .
k. hinterbichler , `` theoretical aspects of massive gravity , '' http://arxiv.org/abs/1105.3735[arxiv:1105.3735 [ hep - th ] ] .
m. andrews , y .- z .
chu , k. hinterbichler , and m. trodden , ( in progress ) . r. v. wagoner and c. m. will , astrophys .
j. * 210 * , 764 ( 1976 ) [ erratum - ibid . * 215 * , 984 ( 1977 ) ] .
y. -z .
chu , phys .
d * 79 * , 044031 ( 2009 ) [ arxiv:0812.0012 [ gr - qc ] ] .
j. hadamard , `` _ lectures on cauchy s problem in linear partial differential equations _ '' ( yale university press , new haven , 1923 ) .
e. poisson , living rev .
* 7 * , 6 ( 2004 ) [ gr - qc/0306052 ] .
e. poisson , a. pound and i. vega , living rev .
* 14 * , 7 ( 2011 ) [ arxiv:1102.0529 [ gr - qc ] ] . c. w. misner , k. s. thorne and j. a. wheeler , san francisco 1973 , 1279p p. c. peters , phys .
d * 1 * , 1559 ( 1970 ) .
s. j. kovacs and k. s. thorne , astrophys .
j. * 224 * , 62 ( 1978 ) .
s. j. kovacs and k. s. thorne , astrophys .
j. * 217 * , 252 ( 1977 ) .
private communication with claudia de rham and andrew tolley . | motivated by the desire to test modified gravity theories exhibiting the vainshtein mechanism , we solve in various physically relevant limits , the retarded galileon green s function ( for the cubic theory ) about a background sourced by a massive spherically symmetric static body .
the static limit of our result will aid us , in a forthcoming paper , in understanding the impact of galileon fields on the problem of motion in the solar system . in this paper
, we employ this retarded green s function to investigate the emission of galileon radiation generated by the motion of matter lying deep within the vainshtein radius @xmath0 of the central object : acoustic waves vibrating on its surface , and the motion of compact bodies gravitationally bound to it .
if @xmath1 is the typical wavelength of the emitted radiation , and @xmath2 is the typical distance of the source from the central mass , with @xmath3 , then , compared to its non - interacting massless scalar counterpart , we find that the galileon radiation rate is suppressed by the ratio @xmath4 at the monopole and dipole orders at high frequencies @xmath5 . however , at high enough multipole order , the radiation rate is enhanced by powers of @xmath6 . at low frequencies
@xmath7 , and when the motion is non - relativistic , galileon waves yield a comparable rate for the monopole and dipole terms , and are amplified by powers of the ratio @xmath6 for the higher multipoles . |
in the present section , we derive inequalities that allow to detect entanglement between two subsystems @xmath1 and @xmath2 of a spin cluster , formed by @xmath10 and @xmath11 individual spins , respectively .
the approach for passing from individual to composite spins is outlined in subsection [ subsec1 ] .
different entanglement inequalities are typically used to detect entanglement in low- and in high - spin systems .
these two cases are thus treated separately ( subsections [ subsec2 ] and [ subsec3 ] , respectively ) .
a possible generalization of the approach to the case of more than two subsystems is sketched in subsection [ subsec4 ] .
our starting point is represented by a generic inequality which applies to the separable states of a two - spin system : @xmath12 here , @xmath13 is the entanglement witness and the lower bound @xmath14 depends on the spin values
. if @xmath8 and @xmath9 are composite , rather than individual spins , with @xmath15 ( and @xmath16 ) , then their lengths are no longer an intrinsic property of the system , but rather state - dependent quantities .
the above inequality can thus be used to detect entanglement between the subsystems @xmath1 and @xmath2 only if the system state @xmath17 is defined within a subspace with given values of the partial spin sums @xmath8 and @xmath9 . however , if the witness commutes with the partial spin sums , @xmath18 = [ w , { \bf s}_b^2 ] = 0 , \ ] ] then eq . [ ineq01 ] can be generalized as follows to an arbitrary state @xmath17 : @xmath19 where @xmath20 is the probability corresponding to eigenvalues @xmath21 and @xmath22 of @xmath23 and @xmath24 , respectively .
equation [ eq00 ] , which essentially results from the vanishing contribution to the expectation value @xmath25 of the coherences between subspaces corresponding to different values of @xmath26 , is proven as follows . in view of the above commutation relations ( eq . [ commut ] )
, one can define a complete basis formed by the states @xmath27 .
these are the simultaneous eigenstates of @xmath28 , @xmath29 , and @xmath13 , with the index @xmath30 denoting additional quantum numbers , if any , required to specify the state .
the expectation value of the witness can thus be written as : @xmath31 here , @xmath32 is the expectation value of the entanglement witness performed within the subspace @xmath33 , and @xmath34 is the density matrix projected onto the subspace @xmath33 and normalized : @xmath35 with @xmath36 . in the last line of eq .
[ eq77 ] , we made use of the intuitive fact that the separability of @xmath17 implies the separability of each @xmath34 .
equation [ eq00 ] can always be used to demonstrate theoretically the presence of entanglement between the subsystems @xmath1 and @xmath2 , for all the relevant quantities can be derived from the density matrix @xmath17 . from an experimental viewpoint ,
the use of eq .
[ eq00 ] for entanglement detection is limited to the cases where the lower bound @xmath37 can be derived from expectation values of accessible observables .
in particular , we consider here the case where these correspond to two - spin operators , such as @xmath38 ( @xmath39 ) . in general , the projectors @xmath40 , whose expectation values appear in the expression of @xmath41 , are of higher order in the single - spin operators .
however , as shown in the following subsection , @xmath42 can be written as a function of @xmath43 ( @xmath44 ) in a number of cases of interest , and specifically in the limit of a weak coupling between the subsystems @xmath1 and @xmath2 . for any separable density matrix @xmath17 of two spins @xmath8 and @xmath9 , the following inequality applies @xcite : @xmath45 here , the set of additional quantum numbers that define the basis states ( see eq .
[ eq77 ] ) is given by : @xmath46 , with @xmath47 and @xmath48 the total spin projection along @xmath49 .
the violation of this inequality implies entanglement between @xmath8 and @xmath9 , and can be typically observed in pairs of antiferromagnetically coupled spins , whose equilibrium density matrix at low temperatures ( @xmath50 ) is close enough to a low-@xmath51 state ( @xmath52 , if @xmath53 ) .
in the case where @xmath8 and @xmath9 are composite , rather than individual spins , eq . [ eqa01 ] can be generalized to @xmath54 along the lines defined in the previous subsection , being @xmath55 = [ { \bf s}_a \cdot { \bf s}_b , { \bf s}_b^2 ] = 0 $ ] ( eq .
[ commut ] ) . here
, the terms on the left - hand side of the inequality can be expressed as combinations of the quantities @xmath56 , while @xmath57 depends on the expectation values @xmath58 of the projector @xmath59 . in order to simplify the expression of the above lower bound , and to express it as a function of two - spin expectation values
, we consider the limit of weak coupling between the subsystems @xmath1 and @xmath2 . in such limit
, the interaction with @xmath2 ( @xmath1 ) induces a finite , but limited amount of @xmath51 mixing in the subsystem @xmath1 ( @xmath2 ) , and approximate expressions can be derived in two steps . as a first step , the average @xmath57 is majorized by a function of @xmath60 and @xmath61 ( @xmath16 ) .
this is performed by exploiting the cauchy - schwarz inequality @xcite : @xmath62 which applies to any joint probability distribution of two real variables , such as @xmath63 .
if the two subsystems @xmath1 and @xmath2 are identical , then @xmath64 and @xmath65 . the combination of eq . [ eqa01 ] and eq .
[ eq06 ] thus gives : @xmath66 with @xmath67 .
we note that , when the ground state of the spin hamiltonian is a singlet , the above inequality detects entanglement , if any .
in fact , for @xmath52 the values of @xmath8 and @xmath9 have to be perfectly correlated @xcite : this implies that @xmath68 .
the vanishing of the total spin @xmath51 also implies @xmath69 .
therefore , the inequality eq .
[ eq07 ] reduces to : @xmath70 with @xmath16 . if the singlet ground state results from the product of two singlet states ( @xmath71 ) , the inequality is fulfilled . in all the other cases , the subsystems @xmath1 and @xmath2
are entangled , and such entanglement is detected by the violation of eq .
[ eq09 ] , being @xmath72 . in the second step of our procedure ,
the averages @xmath73 are replaced by simple functions of the expectation values @xmath74 , and thus of @xmath75 .
the specific function to be used depends on whether the values of @xmath76 approach their theoretical minimum ( @xmath77 or @xmath78 ) or maximum ( @xmath79 ) .
these two cases are considered separately hereafter . [
[ low - spin - subsystems . ] ] low - spin subsystems . + + + + + + + + + + + + + + + + + + + + if the partial spin sum takes integer values , its smallest possible value is @xmath80 , with @xmath16 .
we consider the case where the only relevant terms in the probability distribution @xmath81 are @xmath82 , @xmath83 , and @xmath84 .
the corresponding expressions of the averages are @xmath85 , while @xmath86 . therefore , @xmath87 and eq .
[ eq07 ] becomes : @xmath88 if the partial spin sum has half - integer values , then its lowest value is @xmath89 .
one can proceed as in the previous case , and define : @xmath90 , @xmath91 , and @xmath92 .
the averages thus read @xmath93 and @xmath94 , while @xmath95 .
this results in the relation @xmath96 .
equation [ eq07 ] can thus be expressed in the form : @xmath97 [ [ high - spin - subsystems . ] ] high - spin subsystems .
+ + + + + + + + + + + + + + + + + + + + + if the partial spin sum tends to take the maximum possible value , @xmath98 , then the limit of small @xmath51-mixing for the subsystems corresponds to the probability distribution @xmath99 , @xmath100 , and @xmath101 . in this case , the averages and expectation values are @xmath102 and @xmath103 , while @xmath104 .
equation [ eq07 ] thus takes the form : @xmath105.\end{aligned}\ ] ] according to the spin - squeezing inequalities for arbitrary spins @xcite , the separability of the density matrix @xmath17 of @xmath8 and @xmath9 implies : @xmath106 this inequality is violated by those entangled states with a high ( i.e. close to @xmath107 ) value of the total spin @xmath51 and a small modulus of its projection along @xmath49 .
the simplest example of such a state might be represented by the spin triplet , with vanishing projection along @xmath49 , formed by two @xmath108 spins .
proceeding as in the previous subsection , we generalize eq .
[ eq21 ] to the case where @xmath8 and @xmath9 are not individual spins , but rather partial spin sums , corresponding to generic subsystems @xmath1 and @xmath2 : @xmath109 , \ ] ] where @xmath40 is the projector on the subspace @xmath110 .
we first consider the limit where the occupation of all subspaces but the ones corresponding to @xmath111 can be neglected . if @xmath1 and @xmath2 are formed by the same number of identical spins @xmath112 , then @xmath113 . the inequality eq .
[ eq22 ] thus becomes : @xmath114 any state violating such inequality is entangled .
if the system hamiltonian is invariant with respect to time - reversal symmetry , then @xmath115 for any thermal state .
the inequality is thus violated in all the cases where the @xmath49 components of @xmath116 and @xmath117 are anticorrelated .
if the state of each subsystem includes small contributions from the subspaces @xmath118 and the subsystems @xmath1 and @xmath2 are identical , then the approximate expressions of @xmath73 lead to the inequality : @xmath119.\end{aligned}\ ] ] the proocedure outlined in the previous subsections can be generalized to the case of @xmath120 composite spins @xmath121 .
the starting point is represented by a generic inequality which applies to the fully separable states of an @xmath120-spin system : @xmath122 where the lower bound @xmath14 depends on the spin values . in the case where the @xmath121 are composite , rather than individual spins , with @xmath15 ( and @xmath123 ) , the above inequality allows the detection of entanglement only if the system state is defined within a subspace @xmath33 with given values of the partial spin sums @xmath121 .
however , if @xmath124 = 0 \ ( i=1 , \dots , n),\ ] ] then eq . [ eqb01 ] can be generalized as follows : @xmath125 where @xmath126 is the probability corresponding to a subspace @xmath33 , with defined values of the partial spin sums . as a representative example , we consider the case of a spin cluster with a singlet ground state . at temperatures lower than the typical exchange constant , the density matrix of such a spin cluster typically violates the following inequality @xcite : @xmath127 which is instead fulfilled by any fully separable density matrix .
if the individual spins are replaced by composite ones @xmath128 , entanglement between the subsystems @xmath129 can be detected by observing the violation of the inequality : @xmath130 approximate solutions for expressing the averages @xmath131 in terms of expectation values can be derived in the small @xmath51-mixing limit , along the same lines as in the case of two subsystems .
-spin ring , with antiferromagnetic exchange interaction between nearest neighbors .
the two equivalent subsystems @xmath1 ( white circles ) and @xmath2 ( black ) are formed by @xmath132 consecutive ( a ) or alternate spins ( b ) each .
( c ) the system consists of a spin grid , with two inequivalent sublattices @xmath1 and @xmath2 .
( d ) dimer - like structure , where the subsystems @xmath1 and @xmath2 are given by exchange - coupled spin rings.,width=321 ] the inequalities discussed so far are suited for systems that can be partinioned in two complementary subsystems ( @xmath1 and @xmath2 ) , such that the partial spin sums @xmath8 and @xmath9 are characterized by finite but small fluctuations in the system ground state .
this might occur in clusters of exchange - coupled spins , if the typical exchange constant between the spins of each subsystem ( @xmath133 and @xmath134 ) is much larger than that between spins belonging to different subsystems ( @xmath135 ) .
molecular nanomagnets provide systems of this kind , such as supramolecular assemblies , formed by weakly coupled molecules . in subsection [ subsecdimer ]
we discuss a prototypical sistem of this kind , represented by a ring dimer .
the weak - coupling regime , defined by the inequality @xmath136 , does nt however represent a necessary condition for @xmath8 and @xmath9 to be approximately good quantum numbers .
this feature also shows up in spin clusters with a ring or a grid geometry .
the first of such systems is discussed in subsection [ subsecrings ] , where , depending on the partition of the ring , the intermediate ( @xmath137 ) or strong - coupling ( @xmath138 ) regimes apply .
the latter regime also characterizes the spin grid with nearest - neighbor interactions , which is discussed in subsection [ subsectiongrid ] .
molecular nanomagnets can act not only as individual quantum systems , but also as weakly coupled units within supramolecular structures @xcite .
hereafter , we consider a prototypical dimer - like structure ( fig .
[ figstruct](d ) ) , formed by two spin rings ( labelled @xmath1 and @xmath2 ) .
the spin hamiltonian of the dimer , @xmath139 , includes two kinds of contributions , corresponding to the intra- and inter - ring exchange interactions : @xmath140 in order for the ground state of @xmath141 to exhibit entanglement between the @xmath1 and @xmath2 rings , the inter - molecular exchange has to be antiferromagnetic ( @xmath142 ) . if @xmath143 and @xmath144 , the ground state of @xmath141 corresponds to a spin singlet .
we start by considering the case where the intra - ring exchange is antiferromagnetic ( @xmath145 , with @xmath146 ) , such that the ground state of @xmath147 corresponds to @xmath80 . in the limit @xmath148 , the partial spin sums @xmath8 and
@xmath9 are good quantum numbers , and the ground state tends to the product of the two singlet states @xmath71 . for larger values of @xmath149 ,
the interaction between @xmath150 and @xmath151 induces an increasing occupation of states with @xmath152 . in the limit @xmath153 , @xmath150 and @xmath151
form a singlet , while the remaining three spins of each ring form two uncoupled trimers . the comparison between @xmath154 ( fig .
[ figdimer](a ) , red curve ) and @xmath155 ( blue curve ) shows that , the ground state exhibits inter - ring entanglement in both these limits , and in all the intermediate cases .
besides , the averages @xmath155 are very well approximated by the expression reported in eq .
[ eq09 ] ( blue squares ) .
the ground state of the dimer thus presents entanglement between the subsystems @xmath1 and @xmath2 for arbitrary values of @xmath156 , and such entanglement can always be detected by the inequality eq .
[ eqa02 ] , expressed in terms of spin - pair correlation functions .
an analogous result ( not shown ) has been obtained for a ring dimer defined by the spin hamiltonian @xmath141 ( eq . [ eqdimer ] ) , formed by inequivalent spins : @xmath157 , @xmath158 ( @xmath5 ) .
this can be regarded as a simplified model of the ( cr@xmath159ni)@xmath160 dimers , that have been recently synthesized in a number of different derivatives @xcite . of @xmath154 ( red curves ) and @xmath161 ( blue curves ) for the ground state of the two antiferromagnetically coupled rings ( eq . [ eqdimer ] ) .
the blue squares correspond to the approximate expression of @xmath161 .
the two rings are identical , with @xmath145 .
( b ) same as above , but with ferromagnetic intra - ring interactions ( @xmath162 ) .
( c ) temperature dependence of @xmath154 and @xmath161 ( exact and approximate expressions ) , for @xmath163 .
( d ) same as above , but with ferromagnetic intra - ring interactions ( @xmath164).,width=321 ] in order to investigate the robustness of entanglement with respect to temperature , we compute the dependence on @xmath165 of @xmath154 ( panel ( c ) , red curve ) and @xmath161 ( blue curve ) , for @xmath166 .
the inequality eq .
[ eqa02 ] is violated ut to the threshold temperature @xmath167 , corresponding to about half of the gap @xmath168 between the ground state singlet and first excited triplet of the dimer . the approximate expression reported in eq .
[ eq09 ] ( blue squares ) slightly underestimates @xmath169 in the relevant temperature range , thus underestimating the threshold temperature .
the negativity @xcite has a small value ( @xmath170 ) for @xmath167 , and vanishes for @xmath171 .
an antiferromagetic interaction between the spins @xmath150 and @xmath151 ( eq .
[ eqdimer ] ) also tends to entangle the two rings in the presence of ferromagnetic intra - ring interactions ( @xmath162 ) . in this case
, the ground state in the limit @xmath172 corresponds to a singlet , with @xmath173 .
values of @xmath174 up to 10 lead to a limited amount of @xmath175mixing within each ring . as in the previous case ,
the inequality eq .
[ eqa02 ] is always violated , showing that the two rings are always entangled in the low - temperature limit ( fig .
[ figdimer ] , panel ( b ) ) . also , the approximate expression eq .
[ eq11 ] approaches the exact value of @xmath155 , thus enabling the detection of entanglement between the rings in terms of the spin - pair correlation functions .
the inequality eq .
[ eqa02 ] allows the detection of entanglement up to temperatures of about @xmath176 ( @xmath177 , panel ( d ) ) . in the relevant temperature range
, the average @xmath178 is very well approximated by the expression in eq .
[ eq11 ] ( blue squares ) .
both quantities however underestimate the temperature range where the rings are actually entangled , being the negativity finite up to @xmath179 .
the two - macrospin model also applies to spin clusters that do nt display any dimer - like structure , and thus do nt lend themselves to be naturally partitioned into two ( weakly - coupled ) subsystems . as a first example of this different kind of systems , we consider the case of single spin rings .
molecular nanomagnets provide a variety of ring - like structures , with different spin numbers ( @xmath180 ) and values ( @xmath181 ) @xcite .
hereafter , we focus on a ring formed by identical spins , with antiferromagnetic interaction between nearest neighbors .
we consider two possible partitions of such ring into two equivalent subsystems . in the first partition we consider ,
the two subsystems are formed by the odd- and even - numbered spins , respectively , and @xmath182 [ fig . [
figstruct](a ) ] .
the three terms of the spin hamiltonian @xmath183 thus read : @xmath184 where @xmath185 and @xmath186 . in the ground state of @xmath141
, the spins @xmath187 tend to be parallel to each other ( and antiparallel to the @xmath188 ) , and each partial spin sum tends to the theoretical maximum @xmath189 .
( red curves ) , and @xmath155 ( blue ) for the ring formed by @xmath182 rings @xmath108 .
the inequality detects entanglement between the subsystems @xmath1 and @xmath2 in the temperature range where @xmath190 .
the panels ( a ) and ( b ) refer to the partitions represented in fig .
[ figstruct ] , panels ( a ) and ( b ) , respectively .
the blue squares correspond to the approximate expressions of the thresholds in the large-@xmath4 ( a ) and small-@xmath4 limits ( b ) . ]
the calculations presented hereafter refer to the case of @xmath182 spins @xmath108 .
the average length of the intermediate spins in the system ground state is @xmath191 , corresponding to an entangled ground state @xcite . in order to verify the robustness of such entanglement with respect to temperature
, we report the dependence on @xmath165 of @xmath192 and @xmath161 ( fig .
[ figring](a ) , red and blue curves , respectively ) .
the inequality eq .
[ eqa02 ] is violated for @xmath193 , which is approximately twice as large as the gap @xmath194 between the ground state and the first excited triplet .
besides , in the considered temperature range , the threshold value @xmath161 remains close to the approximate expression of spin - pair correlation functions given in eq .
[ eq11 ] ( blue squares ) .
the threshold temperature derived by means of eq .
[ eqa02 ] compares well with the temperature dependence of the negativity @xmath195 @xcite .
in fact , @xmath196 at @xmath197 , whereas @xmath198 for @xmath199 .
we finally note that entanglement between sublattices in the spin ring persists to higher temperatures with respect to that between neighboring spins , which vanishes for @xmath200 . in the second partition we consider , the subsystems @xmath1 and
@xmath2 are formed by the first and second four spins , respectively [ fig . [
figstruct](b ) ] : @xmath201 where @xmath202 and @xmath203 .
the ground state of @xmath147 is thus a spin singlet ( @xmath204 ) , while @xmath205 mixes the @xmath71 state with those corresponding to finite values of the partial spin sums . in particular , for @xmath206 we obtain @xmath207 @xcite .
this results in a violation of eq .
[ eqa02 ] , even though a less prominent one with respect to the one obtained with the previous partition of the ring . as to the effect of temperature , entanglement between these two subsystems
is detected by the above criterion up to @xmath208 , where the value of @xmath209 ( fig .
[ figring](b ) , red curve ) becomes larger than the threshold @xmath210 ( blue ) . unlike the case of the previous partition , the approximate expression of eq .
[ eq09 ] underestimates the threshold value in the relevant temperature range ( the discrepancy arises mainly from the difference between @xmath211 and @xmath65 , not shown ) .
the use of such expression thus allows to detect entanglement in terms of expectation values of spin - pair operators , but only in the temperature range @xmath212 .
the above results suggest that entanglement between the two segments that form the ring is less robust with respect to temperature than that between even- and odd - numbered spins .
this is confirmed by the temperature dependence of the negativity @xmath195 , that is finite ( @xmath213 ) at @xmath214 and vanishes at @xmath215 .
antiferromagnetic wheels do nt represent the only nanomagnets whose ground state can be approximately described in terms of a two - macrospin model .
another class of systems that possess this property is represented by planar molecules such as the @xmath216 cu @xcite and mn @xcite grids . in the following we focus on the former molecule , where each ion corresponds to an @xmath108 spin , and
assume for simplicity a single exchange constant for all antiferromagnetic interactions between nearest neighbors .
within the spin grid , one can identify two sublattices , @xmath1 and @xmath2 , such that each spin belonging to @xmath1 is coupled only to spins of @xmath2 [ see fig .
[ figstruct](c ) ] .
unlike those of the spin ring , these subsystems are inequivalent , being @xmath217 .
the three terms of the spin hamiltonian @xmath183 read : @xmath218 where @xmath219 . within the @xmath220 ground doublet
, each spin tends to be parallel to the ones of the same subsystem and antiparallel to those of the other one .
in fact , the average values of the partial spin sums are given by @xmath221 and @xmath222 .
( red curve ) , and @xmath155 ( blue ) for the grid formed by @xmath223 spins @xmath108 .
the blue squares correspond to the approximate expressions of the thresholds in the large-@xmath4 limit .
inset : temperature dependence of the concurrence between the central spin @xmath224 and any of the @xmath225 ( filled squares ) , and between @xmath225 and a neighboring @xmath226 ( empty squares).,width=321 ] as in the previously considered spin clusters , the antiferromagnetic interactions between the spins @xmath227 and @xmath228 tend to entangle the two subsystems , and to induce the violation of the inequality eq .
[ eqa02 ] .
in particular , the expectation value of @xmath154 ( fig .
[ figgrid ] , red curve ) falls below the threshold @xmath229 ( blue ) in the temperature range @xmath230 . along the lines of sec .
[ subsec2 ] , we approximate the threshold values with simple functions of the partial - spin sums . in the small @xmath231mixing limit , one can disregard all the probabilities @xmath232 , but the following ones : @xmath233 , @xmath234 , and @xmath235 . expressing @xmath211 , @xmath23 and @xmath24 in terms of these probabilities , and eliminating @xmath30 and @xmath236 ,
one obtains : @xmath237 the temperature dependence of the above lower bound is reported in fig .
[ figgrid ] ( blue squares ) .
the approximate expression slightly underestimates the value of the @xmath211 .
however , due to the small slope of the curves in the relevant region , this results in a larger relative error concerning the temperature range where entanglement persists . in order to assess the suitability of inequality eq .
[ eqa02 ] to detect entanglement between the sublattices of the spin grid , we consider the temperature dependence of the negativity @xmath195 @xcite . at the threshold temperature @xmath238 , the negativity takes a finite , though small , value : @xmath239 .
the negativity vanishes for slightly higher temperatures , and more specifically at @xmath240 . from a physical point of view
, it might be instructive to compare the above temperature ranges with the ones that characterize spin - pair entanglement .
in fact , the antiferromagnetic interactions in @xmath141 also tend to entangle neighboring spins within the grid .
such spin - pair entanglement is here quantified by the concurrence @xcite , which is plotted in the figure inset as a function of temperature .
the concurrence vanishes below @xmath241 for both the spin pairs along the sides of the grid ( second contribution in eq .
[ hamgrid ] , empty squares ) and the pairs that involve the central spin @xmath242 ( first contribution in eq .
[ hamgrid ] , filled squares ) .
the pairs of uncoupled spins are unentangled at all temperatures .
entanglement between the subsystems @xmath1 and @xmath2 is thus significantly more robust with respect to temperature than that between pairs of individual spins .
the inequalities derived above allow the detection of entanglement through equal - time spin - pair correlation functions .
in fact , both the expectation values @xmath243 ( with @xmath244 ) and ( in a number of cases of interest ) the averages @xmath245 can be expressed as linear combinations of the @xmath43 .
these expectation values can be experimentally accessed by magnetic neutron scattering , for the neutron cross - section can be written in terms of dynamical correlation functions between pairs of individual spins @xcite .
in particular , it was recently shown @xcite that , if accurate data for the scattering function are available on a large - enough portion of the energy and momentum space , it is possible to extract equal - time two - spin correlation functions directly , without any prior knowledge of the system hamiltonian .
this allows the detection of entanglement between arbitrary subsystems also if the spin hamiltonian can not be reliably known , as is the case when the fit is not univocal , or even impractical , due to the size of the hilbert space .
it should also be noticed that neutron scattering requires samples containing a macroscopic number of identical spin clusters , and is thus well suited for the study of molecular nanomagnets .
four - dimensional inelastic neutron scattering has recently been applied to the cr@xmath246 molecule @xcite , essentially consisting in an octagon of @xmath247 spins .
the equal - time correlation functions that have been measured provide a direct experimental demonstration of entanglement between the ring subsystems considered in subection [ subsecrings ] . in the low - temperature limit ,
the state of cr@xmath246 approximately coincides with a spin singlet ( @xmath248 ) , whereas the partial spin sums @xmath249 approximtely correspond to 37 and 2 for the partitions represented in fig .
[ figstruct ] ( a ) and ( b ) , respectively .
these values imply the violation of the inequalities reported in eqs .
[ eq11 ] and [ eq09 ] . at low temperatures ,
the subsystems @xmath1 and @xmath2 of cr@xmath246 are thus entangled for both the above partitions .
such an experimental investigation of entanglement can also be applied to dimers of weakly - coupled molecules , such as ( cr@xmath159ni)@xmath160 .
these systems , that are of potential interest for quantum computation @xcite and simulation @xcite , have in fact been synthesized in large high - quality crystals , that are well suited for neutron - scattering . in particular , the inequalities eq . [ eqa02],[eq09 ] can be used to prove the existence of entanglement between the two cr@xmath159ni molecules , also if the inter - ring exchange @xmath250 is large enough to induce a small but finite amount of @xmath51-mixing in each ring , and the macrospin approximation for these does nt apply @xcite . in conclusion ,
the present approach allows one to apply entanglement witnesses and spin - squeezing inequalities to the detection of entanglement between collective spins .
the inequalities have been applied to a number of prototypical spin clusters , characterized by different geometries and coupling regimes . here , entanglement between complementary subsystems is systematically detected in the ground state .
the threshold temperature , above which thermal entanglement vanishes , is lower that the one obtained from the temperature dependence of the negativity , by a factor of the order of 2 .
approximate expressions have been derived for the lower bounds in the inequality , that allow the derivation of all the relevant quantities from spin - pair correlation functions .
the values of these expressions typically approaches from below that of the exact expressions in the parameter range of interest .
we note that , while this relation can not been taken from granted in general , the validity of the assumption underlying the approximation ( i.e. , small fluctuations of the partial spin sums ) can always be verified experimentally .
this work has been financially support by the firb project rbfr12rpd1 of the italian miur and by fondazione cariparma . | entanglement between individual spins can be detected by using thermodynamics quantities as entanglement witnesses .
this applies to collective spins also , provided that their internal degrees of freedom are frozen , as in the limit of weakly - coupled nanomagnets . here
, we extend such approach to the detection of entanglement between subsystems of a spin cluster , beyond such weak - coupling limit .
the resulting inequalities are violated in spin clusters with different geometries , thus allowing the detection of zero- and finite - temperature entanglement . under relevant and experimentally verifiable conditions ,
all the required expectation values can be traced back to correlation functions of individual spins , that are now made selectively available by four - dimensional inelastic neutron scattering .
a great effort has been devoted in the last years to the generation and detection of quantum entanglement in diverse physical systems @xcite .
one of the most practical means in the latter perspective is represented by entanglement witnesses @xcite .
these are observables that can be experimentally accessed in the system of interest , and whose expectation values can exceed given thresholds only in the presence of specific forms of entanglement .
the violation of the corresponding inequalities allows the detection of entanglement , without requiring the derivation of the systems state , and with a variable amount of knowledge of the system hamiltonian . in spin systems ,
routinely measured thermodynamic quantities , such as magnetic susceptibility , can be used as entanglement witnesses @xcite . in clusters of antiferromagnetically coupled spins ,
internal energy allows one to demonstrate the non full - separability of the equilibrium state at low temperatures @xcite .
exchange energy also allows the detection of multipartite entanglement in qubit systems @xcite and in clusters formed by @xmath0 spins @xcite . in the same spirit
, spin - squeezing inequalities can be used to demonstrate entanglement in the vicinity of relevant quantum states by means of collective observables only @xcite . in all these cases ,
the relevant subsystems are represented by individual spins , such as those carried by single ions within a molecular nanomagnet @xcite . here
we extend such approach to the detection of entanglement between subsystems ( hereafter labeled @xmath1 and @xmath2 ) of a spin cluster . each subsystem is formed by a finite number of individual spins ( @xmath3 ) and can be described in terms of a collective spins ( @xmath4 , with @xmath5 ) . in many cases of interest ,
these partial spin sums undergo non - negligible fluctuations ( @xmath6 ) in the low - energy eigenstates of the spin cluster .
typical examples are represented by dimers of weakly coupled molecular nanomagnets @xcite , where entanglement between the molecules has so far been discussed only in terms of a simplified two - macrospin model @xcite .
the approach developed hereafter provides the possibility of extending such investigations to the case where the internal degrees of freedom of each subsystem can not be disregarded ( @xmath7 ) . it is also shown that , if the fluctuations of the partial spin sums @xmath8 or @xmath9 are small , all the relevant quantities can be expressed in terms of correlation functions between pairs of individual spins .
this result potentially presents a practical relevance , for such correlation functions can be experimentally derived by four - dimensional inelastic neutron scattering @xcite .
the validity of the underlying physical assumption can be deduced from a qualitative knowledge of the system , and , most importantly , from the experimentally accessible spin - pair correlation functions .
molecular nanomagnets represent a varied class of spin clusters , whose physical properties can be widely tuned by chemical synthesis @xcite . in particular ,
low - spin nanomagnets with dominant antiferromagnetic interaction can be regarded as prototypical examples of highly correlated , zero - dimensional quantum systems .
these molecular spin clusters thus represent a suitable arena for the investigation of quantum entanglement .
the understanding and control of different forms of entanglement ( e.g. , that between single ions of a single molecule , or that between molecular spins within a dimer ) also represents a preliminary requirement for the use of molecular nanomagnets as spin cluster qubits @xcite .
the paper is organized as follows . in sec .
[ secent ] we derive the inequalities that allow to detect entanglement between the composite spins . in sec .
[ secpro ] such inequalities are applied to the investigation of entanglement within different bipartitions of prototypical spin clusters . in sec .
[ seccon ] we discuss the measurement of the relevant observables by means of inelastic neutron scattering and draw our conclusions . |
it is generally believed that cold dark matter ( dm ) dominates the mass in the universe and significantly affects both the process of galaxy formation and the large - scale distribution of galaxies .
here we present results of a study of the formation and the evolution of the dm component of galaxies , dm halos .
the structure of the halos depends on the environment ( e.g. , avila - reese et al .
1999 ) , so that the properties of galaxies are also expected to depend on the cosmological environment . for our analysis
we use a low - density flat cosmological model with cosmological constant @xmath0 , which have been proved to be very successful in describing most of the observational data at both low and high redshifts : @xmath1 , @xmath2 , @xmath3 km / s / mpc , @xmath4 gyrs .
in order to study the statistical properties of halos and to have a sufficient mass resolution we have chosen a simulation box of @xmath5 with @xmath6 cold dark matter particles ( particle mass of @xmath7 ) . using the adaptive refinement tree ( art ) code ( kravtsov , klypin & khokhlov 1997 ) we reached a force resolution of @xmath8 kpc in high density regions .
identification of halos in dense environments and reconstruction of their evolution is a challenge .
any halo finding algorithm has to deal with difficult `` decision - making '' situations , in particular when many gravitationally bound halos are moving within a large dense object ( a galaxy cluster or a group ) .
we use an algorithm described in klypin et al .
we characterize each halo by its mass and the maximum circular velocity @xmath9 .
this quantity is more meaningful observationally and can be numerically measured more easily and more accurately than mass .
the halo samples are complete for halos with @xmath10 km / s ( gottlber et al .
1999 ) . here
we consider only more massive halos with circular velocities @xmath11 km / s .
the limit on the circular velocity is relaxed at high redshifts to allow identification of small progenitors of present - day halos . at @xmath12
we have detected 4193 halos in our simulation , which corresponds to a halo number density of @xmath13 .
for each halo in our @xmath12 sample we have constructed a complete evolution tree over up to 23 time moments approximately regular in time distributed between @xmath14 and @xmath15 .
the procedure of progenitor identification is based on the comparison of lists of particles belonging to the halos at different moments both back and forward in time ( gottlber et al 1999 ) .
the epoch at which the halo has been identified for the first time depends on our assumption of thresholds of the circular velocity ( @xmath16km / s ) and mass ( minimum of 40 bound particles ) for the progenitors of our halos at @xmath17 .
in order to find the cosmological environment of each of the 4193 halos we run a friend - of - friend analysis over the dark matter particles with a linking length of 0.2 times the mean interparticle distance . by this procedure
we find clusters of dark matter particles with an overdensity of @xmath18 .
the virial overdensity in the @xmath0cdm model under consideration is @xmath19 , which corresponds to a linking length of 0.17 .
therefore , the objects which we find are slightly larger than the objects at the virial overdensity .
we have increased the linking length because we found that halos , which are outside , but close to a cluster , are affected by the cluster . for each of the halos at @xmath12 we find the cluster of dark matter particles to which the halo belongs .
we call the halo a cluster galaxy if the halo belongs to a particle cluster with a total mass larger than @xmath20 .
we call it an isolated galaxy if only one halo belongs to the object at overdensity 200 .
the rest of halos are called group galaxies . by this
definition pairs of galaxies are also handled as groups .
this increases slightly the number of galaxies in `` groups '' . with the procedure described above we find at @xmath14 that there are 401 cluster galaxies ( 9.6% ) , 1247 galaxies in groups ( 29.7% ) , and 2545 isolated galaxies ( 60.7% ) .
if we consider objects of virial overdensity , the number of cluster and group galaxies decreases to 6.5% and 25.6% and 67.9% of galaxies become `` isolated '' . for comparison see also avila - reese et al .
the mass of a halo increases due to accretion and merging .
however , interacting halos may exchange and lose mass .
we calculate the relative mass growth per time interval @xmath21 , where time is in units of @xmath22 years .
if it is larger than 0.35 , we call this a major - merging event . note that according to this definition , we calculate the _ total _ change of mass , not merging with a large halo . we found that 28% of the cluster halos , 29% of the group halos and 52% of the isolated halos never underwent a major - merging event
1 presents the number of major - merging events the merging rate of halos in different environments .
we show the merging rate averaged over three ( two at @xmath23 ) subsequent time intervals .
the error bars are @xmath24 errors for the number of events detected .
the three points for merging inside clusters at @xmath25 , and 0.5 come from 1 , 3 , and 4 events , respectively .
the points in the figure can be fitted by a curve @xmath26 with @xmath27 and @xmath28 for cluster , group , and isolated halos , respectively .
the exponential dilution is the same for all types , @xmath29 .
it is mainly due to the fact that at @xmath30 we are rapidly loosing the halo progenitors due to mass resolution .
the position and high of the maxima as well as the slope depend slightly on the chosen threshold for definition of major merging events , but the relative position of the curves remains practically constant .
the higher probability of major - merging that cluster and group halos had in the past is due to the higher density in regions , where cluster and groups were to form .
note that clusters have not yet existed at that time . as clusters with large internal velocities form ,
merging rate drastically decreases .
there are almost no major - merging events of cluster halos in the recent past .
those very few events have probably happened just outside the clusters before the mergers fell in the clusters .
the probability of recent major - merging is almost the same for isolated galaxies and for group galaxies .
it is very different in the past : the merging of galaxies , which end in clusters , was much higher than for isolated galaxies .
this work was funded by the nsf and nasa grants to nmsu , and collaborative nato grant crg 972148 .
sg acknowledges support from deutsche akademie der naturforscher leopoldina with means of the bundesministerium fr bildung und forschung grant lpd 1996 .
s. gottlber , a. a. klypin , a. v. kravtsov . .
in _ asp conference series vol .
176 , observational cosmology : the development of galaxy systems _ , eds . :
g. giuricin , m. mezetti , p. saluccip , 418429 , 1999 . | we discuss the impact of the cosmological environment on the evolution of dark matter halos using a high - resolution simulation within a spatially flat @xmath0cdm cosmology . |
successful realization of aneutronic fusion idea would greatly reduce problems associated with neutron radiation such as ionizing damage , neutron activation , and requirements for biological shielding , remote handling , and safety issues .
one of the reactions discussed in this respect is the reaction p+@xmath1b@xmath2 .
there are at least two proposals @xcite how to use this reaction for energy production .
the shape of @xmath0-particles spectrum , especially its low energy tail is of crucial importance for realisation of these projects .
existing measurements @xcite produce the spectra for fixed angles of @xmath0-particles that does not say much about integrated spectrum . in this paper
we propose a simple theoretical model for the reaction amplitude allowing to calculate with some assumptions the shape of @xmath0-particle spectrum .
the reaction p+@xmath1b@xmath4 has a well pronounced resonance at the proton energy 675 kev .
the resonance corresponds to excited state of @xmath5 nucleus with the energy 16.57 mev , and with quantum numbers @xmath6 and isospin @xmath7 .
the dominant decay mode is @xmath0-particle plus excited state of @xmath8 with the energy @xmath9 mev and quantum numbers @xmath10 and @xmath11 .
the cross section for decay into the ground state of @xmath8 is about two orders of magnitude lower @xcite .
the decay proceeds with isospin violation , therefore the width of the resonance is small enough . with
the @xmath12-value @xmath13 mev the total width is @xmath14 mev .
it leads to high cross section at the resonance peak @xmath15 b. @xcite .
the decaying state has negative parity , therefore the primary emitted @xmath0 must be in the state with an odd angular momentum . in our case
it can be @xmath16 , and @xmath17 .
the width of the excited @xmath8 state is 1.513 mev .
it corresponds to lifetime @xmath18 s. during this time the primary @xmath0-particle is still within strong interaction range .
it means that the decay is in fact a 3-body type .
immediate consequence of it is that the energy of any @xmath0-particle can vary from 0 to @xmath19 , where @xmath12 is total @xmath12-value for the reaction , @xmath20 mev . keeping in mind these features of the reaction
the amplitude can be presented in the following general form : @xmath21 where @xmath22 is the emission vertex of the primary @xmath0-particle with the momentum @xmath23 , @xmath24 is the propagator of the intermediate system @xmath0-particle plus @xmath3 , and @xmath25 is the decay vertex of @xmath3 depending on relative momentum of two @xmath0-particles .
the diagram corresponding to this process is shown in fig.1 .
-particles.,width=188 ] for unpolarized initial state an angular distribution of @xmath0-particle is isotropic .
for this reason we omit angular dependence in the vertexes @xmath26 and @xmath27 and will discuss only their energy dependence .
there are two factors determining momentum dependence of the vertexes at low momentum .
first , it is a coulomb barrier that suppresses emission of low energy particles .
second , it is a centrifugal barrier . with these factors
the decay vertex of @xmath3 can be presented as @xmath28 where @xmath29 is a relative velocity of @xmath0-particles in c.m .
frame , and @xmath30 is the euler gamma - function .
the value of the cut - off parameter @xmath31 is @xmath32 , where @xmath33 is the nuclear radius .
the shape of the @xmath0-particle spectrum appeared insensitive to exact value of @xmath31 . with this vertex
we can define the energy dependent width by equation @xmath34 here @xmath0 is the fine structure constant , and @xmath35 , where @xmath36 is the mass of @xmath0-particle , and @xmath37 is the c.m .
remaining parameter @xmath38 is determined by the width of the excited state of @xmath8 @xmath39 where @xmath40 mev .
the vertex @xmath41 has similar structure @xmath42 here the coulomb barrier is larger , @xmath43 . as for centrifugal barrier , strictly speaking there are two partial waves with @xmath16 and @xmath17 contributing to this vertex . however , since we are interested in low energy tail of the @xmath0 s spectrum , we retained only @xmath16 .
in addition , the total width of the excited carbon state is small enough , @xmath44 mev , therefore , the momentum dependence is smooth in this energy range and does not influence the shape of the spectrum . as in the previous vertex
the exact value of the cut - off parameter @xmath45 is not important .
the shape of the spectrum is insensitive to it .
the parameter @xmath46 is defined by normalization of the spectrum to one .
a propagator @xmath24 of the intermediate @xmath47 state is of the form @xmath48 where @xmath49 is the energy of the initial carbon state @xmath50 mev , and @xmath51 is the hamiltonian of @xmath47 system . omitting @xmath52 interaction and retaining only the width of @xmath53 state we obtain @xmath54 here @xmath55 and @xmath56 are the laboratory kinetic energies of @xmath0-particle and @xmath8 . introducing @xmath12-value for this transition we obtain , @xmath57 where we used the relation @xmath58 .
the center of mass energy @xmath37 in @xmath59 is defined by total @xmath12-value minus kinetic energy of @xmath0 and @xmath8 , @xmath60 . in the final state
we have three identical bose - particles .
therefore , the amplitude eq.([1 ] ) should be properly symmetrized .
the full amplitude is as follows @xmath61
with the amplitude eq.([9 ] ) we can calculate the spectrum of @xmath0-particles .
@xmath62 as usual , the line over the amplitude squared means averaging over initial spin projections and summation over angular momentum projections of @xmath0-particles .
after that the spectrum becomes isotropic . strictly speaking
, the interference terms in the amplitude squared may depend on the angle between @xmath23 and @xmath63 .
however , the scalar product @xmath64 is fixed by delta - function in eq.([10 ] ) . the delta - function in eq.([10 ] ) means that we neglected the small width of the initial excited carbon state .
integration over angles removes delta - function , we obtain @xmath65 integrating eq.([11 ] ) over @xmath66 we obtain the final equation for @xmath0-particle spectrum @xmath67 @xmath68 where @xmath69 , and @xmath70 .
the normalization factor @xmath71 defined by the condition @xmath72 the calculated spectrum is shown in fig.2 .
the spectrum has a well pronounced peak corresponding to emission of primary @xmath0-particle and a broad shoulder going down to very low energy .
this low energy tail contains 11% of @xmath0-particles with the energy less than 1 mev , and 17,5% of @xmath0-particles with the energy less than 1.5 mev .
it is worth to see how important are the coulomb and the centrifugal factors included in the vertexes in eq.([2 ] ) . to do this we calculated the spectrum using constant @xmath73 mev and constant @xmath74 and @xmath75 . in fig.3
we show the spectrum obtained in this way ( dashed line ) and the full spectrum from fig.2 .
the main difference is in the peak position which is shifted downward for @xmath76 mev and an increase of number of @xmath0-particles with the energies between 2.5 and 3.8 mev .
note that in this region all three @xmath0-particles have the energies close to each other . in this region
the effect of interference of three terms in eq.([9 ] ) is strong .
the low energy tail is almost insensitive to it . in summary
, using a simple phenomenological model for the reaction amplitude the shape of the @xmath0-particle spectrum has been calculated .
basing on the obtained shape we estimated the number of low energy @xmath0 s .
we found 17.5% of @xmath0 s below 1.5 mev and 11% below 1 mev .
99 norman rostoker , michl w. binderbauer , hendrik j. monkhorst , science , * 278*,1419 ( 1997 ) .
volosov , nucl . fusion , * 46 * , 820 ( 2006 ) . h. becker , c. rolfs , and h. trautwetter , z. phys . , * a327 * , 341 ( 1987 ) .
lin erh - kang _ et _ al .
, chin.phys.lett . ,
* 15 * , 796 ( 1998 ) . c. spitaleri
_ et _ al . , phis.rev
c * 69 * , 055806 ( 2004)/ | using a simple phenomenological parametrization of the reaction amplitude we calculated @xmath0-particle spectrum in the reaction p+@xmath1b@xmath2 at the resonance proton energy 675 kev .
the parametrization includes breit - wigner factor with an energy dependent width for intermediate @xmath3 state and the coulomb and the centrifugal factors in @xmath0-particle emission vertexes .
the shape of the spectrum consists of a well defined peak corresponding to emission of the primary @xmath0 and a flat shoulder going down to very low energy .
we found that below 1.5 mev there are 17.5% of @xmath0 s and below 1 mev there are 11% of them . |
from the gene regulatory networks that control organismal development @xcite and the coherent oscillations between brain regions responsible for cognition @xcite to the connected technologies that support critical infrastructure @xcite , systems at many scales of modern society rely on the coordination of the dynamics of interdependent systems . analyzing the mechanisms driving such complex networks presents serious challenges to dynamical systems , statistical mechanics , and control theory , including but not limited to the overtly high dimension of their state spaces .
this precludes directly identifying and visualizing their attractors and attractor - basin organization .
moreover , without knowledge of the latter large - scale architecture , predicting network behavior , let alone developing control strategies to maintain function or mitigate failure , is impossible . to shed light on these challenges , we explore limit - cycle attractors arising when rings of coupled nonlinear oscillators synchronize .
we demonstrate how synchronization emerges and stabilizes and identify the diversity of synchronized states .
we probe the global attractor - basin architecture by driving the networks with noise , capturing how the distinct synchronized states can be reached from each other via what we call attractor switching networks .
the analysis relies on the use of limit - cycle attractors to define coarse - grained units of system state space . in this way
, our study of attractor - basin architecture falls in line with the methods of qualitative dynamics introduced by poincare @xcite .
confronted with unsolvable nonlinear dynamics in the three - body problem , poincare showed that system behaviors are guided and constrained by invariant state space structures
fixed point , limit cycle , and chaotic attractors and their arrangement in state space basins of attraction and their separatrices .
the power of his qualitative approach came in providing a concise description of all possible behaviors of a system , without requiring detailed system solutions .
his architectural approach is more recently expressed in terms of smale basic sets @xcite and conley s index theory @xcite .
these show that any system decomposes into recurrent and wandering subspaces in which the behavior is a gradient flow . in short
, there is a kind of lyapunov function over the entire state space , underlying the architectural view of attractors and their basins .
this view is so basic to our modern understanding of nonlinear complex systems that it has been rechristened as the `` fundamental theorem of dynamical systems '' @xcite .
as we will see , our analytical study of oscillator arrays appeals to lyapunov functions to locally analyze limit cycle stability and noise robustness , while our numerical explorations allow us to knit together the stable attractors into a network of stable oscillations , connected by particular pathways that facilitate switching between them .
practically , complicated attractor - connectivity architectures can be probed via external controls or added noise .
we focus on the latter here , following recent successful explorations of noise - driven large - scale systems .
for example , the analysis of bistable gene transcription networks showed that attractor switches can be induced by periodic pulses of noise @xcite .
another recent study of networks of pulse - coupled oscillators showed that unstable attractors become prevalent with increasing network size and the attractors are closely approached by basin tendrils of other attractors .
thus , arbitrarily small noise can lead to switching between attractors @xcite .
our explorations illustrate the theoretical foundations and complements the newer works by focusing on the dynamics of synchronization .
synchronization between oscillators is itself an important and widely studied collective behavior of coupled systems @xcite , with examples ranging from neural networks @xcite to power grids @xcite , clapping audiences @xcite , and fireflies flashing in unison @xcite .
although different in scope and nature , all of these examples can be modeled as coupled oscillators .
decades of research has revealed that a system of coupled oscillators may produce a rich variety of behaviors ; in addition to full synchronization , more complex patterns may emerge , including chaos @xcite , chimera states @xcite , and cluster synchronization @xcite . here
, we study rings of oscillators a system that exhibits multiple stable synchronized patterns called rotating waves @xcite . rings of oscillators have been extensively studied @xcite ; our contribution in this respect focuses on _ reactively _ coupled amplitude - phase oscillators and the organization of their attractors , basins , and noise - driven basin transitions . reactive coupling , in the context of electromechanical oscillators , is that which does not dissipate energy , such as ideal elastic and electrostatic interactions between devices@xcite . a primary motivation of this work is to connect with experiment , using reactive coupling to characterize systems of nearest - neighbor coupled rings of nanoscale piezoelectric oscillators @xcite .
recent experiments investigated synchronous behavior of two such nanoelectromechanical systems ( nems ) @xcite , and it is expected that in the near future larger rings and more complex arrangements will be realized .
@xcite in the context of the complex values used to model these oscillators , reactive coupling means that the coefficient of the laplacian coupling terms is purely imaginary .
this coupling is captured , between landau - stuart oscillators , as a special case of the complex ginzburg - landau equation , which describes a wide range of physical phenomena @xcite .
if no noise is present , the system settles at one of its stable steady states .
exactly which stable state depends on initial conditions . in the presence of noise ,
the long - term behavior of the system is no longer characterized by deterministic attractors .
depending on the level of noise three possible scenarios may emerge : ( i ) if the noise is small , the system fluctuates around an attractor ; ( ii ) if the noise is strong , the system is randomly pushed around in the state space suppressing the deterministic dynamics ; and ( iii ) intermediate levels of noise cause the system to fluctuate around an attractor and occasionally jump to the basin of attraction of a different attractor .
the latter scenario suggests a coarse - grained description of the system s global organization : we specify the effective `` macrostate '' of the system by the attractor it fluctuates around , and we map out the likelihood of transitions to other attractors .
these transitions form an attractor switching network ( asn ) capturing the coarse - grained dynamics of the system .
noise and external perturbation induced jumps in the asn have been suggested as a feasible strategy to control large - scale nonlinear systems @xcite .
our goal is to study the fluctuations of the system and attractor switching in the presence of additive uncorrelated gaussian noise in the phases of oscillators .
setting up the analysis , we introduce the system in sec .
[ sec : deterministic ] , finding the available patterns of synchronization in sec .
[ sec : solutions ] and their local stability in sec .
[ sec : stability ] .
we introduce noisy dynamics in sec .
[ sec : stochastic ] . based on the linearized dynamics
we derive a closed - form expression that predicts the system s response to small noise in sec .
[ sec : covariance ] .
we demonstrate that the attractor switching occurs via a phase - crossing mechanism sec .
[ sec : phase crossing ] .
this motivates the coarse - graining of state space such that we can finally compile an asn for a network of @xmath0 oscillators in sec .
[ sec : the asn ] .
we study rings of reactively coupled oscillators that are governed by @xmath1,\end{aligned}\ ] ] where @xmath2 describes the amplitude and phase of the @xmath3 oscillator @xmath4 and @xmath5 .
the first three terms describe the oscillators in isolation : the first is the linear restoring force , the second term is the first nonlinear correction known as the duffing nonlinearity , and the third term is a saturated feedback that allows the system to sustain oscillatory motion .
the fourth term expresses the inter - oscillator feedback : the oscillators are diffusively coupled to their nearest neighbors with purely imaginary coefficient .
equation ( [ eq : envelope ] ) was derived to describe the slow modulation of rapid oscillations of a system of nems sometimes referred to as an envelope or modulational equation @xcite . although eq .
( [ eq : envelope ] ) presents a compact representation of the dynamics , it is in this case more insightful , and useful , to isolate the amplitude and phase components of the representative complex state .
we therefore separate the dynamics of the amplitudes @xmath6 in vector @xmath7 and those of the phases @xmath8 in vector @xmath9 according to @xmath10 .
the system then evolves according to @xmath11 , \\
\label{eq : phase } \frac{d\phi_i}{dt } & = \alpha a_i^2 + \frac{\beta}{2}\bigg[\frac{a_{i+1}}{a_i}\cos\left(\phi_{i+1}-\phi_i\right ) + \frac{a_{i-1}}{a_i}\cos\left(\phi_{i-1}-\phi_i\right)-2\bigg].\end{aligned}\ ] ] these equations make it clear that in the absence of coupling ( @xmath12 ) , each amplitude @xmath6 will settle to unity , and all phases oscillate with constant frequency @xmath13 .
this frequency , proportional to the square of the oscillator s amplitude , comes from the device s nonlinear restoring force the duffing nonlinearity .
this effect is accordingly referred to as nonlinear frequency pulling .
we now proceed to find solutions of the dynamics with nonzero coupling . to view self - organized patterns of synchronization of these nonlinear oscillators , we consider only the weak coupling regime , with positive nonlinear frequency pulling : @xmath14 .
this selection is heavily motivated by upcoming experimental realizations of the system @xcite and ensures that the internal nodal dynamics are not dominated by coupling terms . with zero coupling ( @xmath15 ) , each oscillator will follow its own limit cycle , and the composite attractor will have @xmath16 dimensions one corresponding to the phase of each oscillator . for small but nonzero coupling ( @xmath17 )
we expect the leading order effect to be in the dynamics of phases .
as these are limit cycles , displacements along phase are not restored except through the coupling edges .
solving eqs .
( [ eq : amplitude ] ) and ( [ eq : phase ] ) for sets of stationary phase differences with fixed unit amplitudes , @xmath18,\\ \left(\frac{d\phi_{i+1}}{dt } - \frac{d\phi_i}{dt}\right)\big|_{\mathbf{a}=\mathbf{1 } } = 0 & = \frac{\beta}{2}\big[\cos\delta_{i+1 } - \cos\delta_{i-1}\big],\end{aligned}\ ] ] where @xmath19 is the ( signed ) phase difference between adjacent oscillators @xmath20 and @xmath21 .
these conditions are satisfied if and only if every other phase difference is equal to some @xmath22 , where the other phase differences are either @xmath23 or also @xmath22 .
note that these conditions are independent of @xmath24 , so the solutions will be valid for all coupling strengths .
because the ring is a periodic lattice and the sum of all @xmath16 phase differences must be an integer multiple of @xmath25 , limit cycles that satisfy the @xmath23 condition for alternating phase differences may exist if and only if the number of nodes is an integer multiple of four . to ease comparison of attractors in systems with various numbers of nodes , we limit our subsequent discussion to the solutions defined wholly by a single phase difference @xmath22 supported across all edges , implying that @xmath16 may not be a multiple of four . for limit cycles where all phase differences are identical , _
@xmath26 , the periodic boundary condition requires @xmath22 to be an integer multiple of @xmath27 , giving precisely @xmath16 unique states of this sort .
these states follow the trajectory @xmath28 specific to a particular wavenumber @xmath29
. these are the expected rotating wave solutions .
each rotating wave has a fixed phase configuration , with phase differences of @xmath30 , represented in eq .
( [ eq : phasesolution ] ) as initial phases @xmath31 .
the form of reactive coupling causes the frequency of oscillation also to be dependent upon the wavenumber . noting that the phase difference @xmath22 is invariant under @xmath32
, we choose to make the restriction @xmath33 .
relative phase diagrams representing the @xmath16 unique configurations for @xmath34 and @xmath35 oscillator rings are shown in fig .
1 . in these ,
each oscillator is represented as a point on the unit circle in the complex plane , with edges connecting adjacent , coupled oscillators .
each edge connects oscillators with an arc length separation equal to the phase difference @xmath36 .
we see that , for instance , @xmath35 and @xmath37 or @xmath38 results in next nearest neighbors being closer in phase than nearest neighbors .
this is a general result ; as @xmath39 , neighboring oscillators will have a phase difference of @xmath40 and next nearest oscillators have nearly equivalent phases .
this is locally out - of - phase sychronization , in contrast to @xmath41 , which is completely in - phase synchronization , _
i.e. , _ zero phase difference between neighboring oscillators .
we also see a symmetry in wave numbers @xmath29 and @xmath42 .
these waves travel in opposite directions around the ring ; the phase configurations amount to a relabeling of oscillators , represented in the fig .
[ fig : wavediagrams - k ] by arrows indicating the direction of labeling .
just as the wavenumber represents the number of wavelengths of the rotating wave along the length of the ring , it may be interpreted as the winding number of the ring about the origin when represented in the complex plane as in fig .
[ fig : wavediagrams - k ] .
we have thus discovered @xmath16 synchronized states that are possible nodes of the global attractor switching network and which the system might visit once noise is included in the dynamics .
although motivated by the weak coupling limit , these rotating waves are valid solutions at all coupling magnitudes .
note that there can be solutions that do not converge to the unit amplitude states enumerated here . with sufficiently weak coupling ,
however , oscillator amplitudes in attractors are in fact confined to stay within a distance of order @xmath24 from unity .
this is shown in appendix [ app : attractoramplitudes ] using a lyapunov - like potential function .
having enumerated such candidate synchronized limit cycles , we need to determine their stability in order to identify those that we expect the noisy system to visit for extended times .
ring in each of four regions of nonlinearity / coupling @xmath43 parameter space , separated by solid black lines .
the @xmath41 wave is locally stable for all @xmath44 and for @xmath45 .
the @xmath46 waves are locally stable for all @xmath47 and for @xmath48 .
the blue dotted line indicates the parameters of likely experimental realizations .
our simulations were done at the endpoints @xmath49 . ] here we show that the stability of each rotating wave / pattern of synchronization to small perturbations is equivalent to finding the sign of @xmath50 .
we then characterize the linear response of these waves to uncorrelated , white gaussian noise on the oscillator phases and find that the @xmath51 and @xmath52 waves amplify noise least in their respective stable regimes .
linearizing eqs .
( [ eq : amplitude ] ) and ( [ eq : phase ] ) around any point on the limit cycle defined by wavenumber @xmath29 , we find the @xmath53 matrix @xmath54 that governs the linear evolution of small deviations from that limit cycle .
we write this matrix in block form , such that @xmath55 is the @xmath56 matrix corresponding to the dependence of deviations in oscillator @xmath20 on deviations in oscillator @xmath57 .
@xmath58 = \sum_jf_{ij}\left[\begin{array}{c } \delta a_j\\\delta\phi_j\end{array}\right ] \\
= \frac{1}{2}\sum_{j=0}^n\left[\begin{array}{cc } -i_{ij } - m_{ij}\beta\sin(2\pi k / n ) & l_{ij}\beta\cos(2\pi k / n ) \\
4\alpha i_{ij } - l_{ij}\beta\cos(2\pi k / n ) & -m_{ij}\beta\sin(2\pi k / n)\end{array}\right]\left[\begin{array}{c } \delta a_j\\\delta\phi_j\end{array}\right],\end{gathered}\ ] ] where @xmath59 is the @xmath60 identity matrix , @xmath61 is the @xmath60 unweighted ring laplacian matrix , and @xmath62 is an @xmath60 next - nearest - neighbor oriented incidence matrix of the ring , @xmath63 the local stability of each rotating wave is then determined by the signs of the eigenvalues of @xmath54 . while this is straightforward to do numerically , we find that exclusion of terms varying with @xmath62 will not effect any changes of sign , as detailed in appendix [ app : linearization ] .
we denote this simplified matrix @xmath64 and transform @xmath64 by a matrix @xmath65 to diagonalize the laplacian @xmath61 , leaving a @xmath66 linear dynamics for each laplacian mode .
the matrices @xmath61 and @xmath62 are not mutually diagonalizable , so this can not be done with the full linearization @xmath54 .
deviations in these laplacian modes are then governed by @xmath67.\ ] ] where @xmath68 are the eigenvalues of @xmath61 for the ring coupling topology ( and @xmath69 is the floor operation ) .
defining @xmath70 , we see that @xmath64 represents stable trajectories if and only if all its eigenvalues @xmath71 have negative real part .
that is , the rotating wave is stable if and only if @xmath72 for all laplacian modes .
the mode associated with @xmath73 , giving @xmath74 , may in fact be ignored .
this zero eigenvalue corresponds to the freedom of deviations along the limit cycle and is explicitly removed in appendix [ app : linearization ] by stabilizing this allowed nullspace of @xmath64 .
then , there are two regimes in which a mode of the modified dynamics is stable : @xmath75 and @xmath76 .
now , we see that all @xmath77 are strictly positive and therefore all @xmath78 will be of the same sign as @xmath50 . with a given sign of the coupling @xmath24 , all wavenumbers @xmath29 satisfying @xmath79 will correspond to stable rotating waves for all coupling magnitudes @xmath80 .
a wave solution is also stable if @xmath81 is large enough such that the smallest nonzero laplacian eigenvalue @xmath82 corresponds to @xmath83 .
this occurs when @xmath84 ( and requires @xmath85 ) .
this scenario clearly corresponds to large coupling magnitudes , which we are not considering here .
figure [ fig : regions ] portrays the above stability conditions in @xmath86 parameter space of the @xmath34 oscillator ring .
there are four distinct regions : large or small coupling - to - nonlinearity ratio , with positive or negative coupling .
the more nearly in- ( out - of- ) phase adjacent node oscillations are stable with small , negative ( positive ) coupling and become stable with positive ( negative ) coupling at some critical coupling magnitude proportional to the nonlinear coefficient @xmath13 . each rotating wave has a critical ratio @xmath87 , proportional to @xmath88 , above which the wave is stable for either sign of @xmath24 . as such , the required coupling magnitudes for this regime increase with @xmath16 .
these boundaries were found analytically as described above and corroborated by diagonalizing the original linearization @xmath54 numerically , validating the process of studying the modified dynamics in @xmath64 .
we have so far found attractors of the deterministic dynamics of rings of reactively coupled nonlinear oscillators .
this identifies orbits that may have some global importance in the system , but gives little indication of the higher - level state space architecture .
we investigate this organization by applying noise to the oscillators phases , first weakly and then strongly enough to induce distinct jumps between attracting limit cycles .
specifically , we focus on the analysis of one of the most ubiquitous and well - modelled forms of disturbances , namely white gaussian noise . the injection point is assumed to be an additive time - varying signal on the phases .
this generates a perturbed dynamics of the form @xmath89,\\ \label{eq : phasewithnoise } \frac{d\phi_i}{dt } & = \alpha a_i^2 + \frac{\beta}{2}\bigg[\frac{a_{i+1}}{a_i}\cos\left(\phi_{i+1}-\phi_i\right ) + \frac{a_{i-1}}{a_i}\cos\left(\phi_{i-1}-\phi_i\right)-2\bigg ] + w_i.\end{aligned}\ ] ] where @xmath90 is an element of @xmath91 : an uncorrelated zero mean i.i.d .
gaussian random process with covariance matrix @xmath92 .
having identified attractors the stable rotating waves , we begin to study the basin architecture by characterizing the system s response to weak noise at each of the stable rotating waves , finding that @xmath41 has the least amplification of noise when @xmath44 and @xmath93 has the least for @xmath47 .
close to an attractor , the dynamics can be predominately described by its linearization : @xmath94=f\;\left[\begin{array}{c } \delta a\\ \delta\phi \end{array}\right]+\left[\begin{array}{c } 0\\ i \end{array}\right]w,\label{eq : linearized dynamics with noise}\ ] ] where @xmath54 is the linearized state matrix of eq .
( [ eq : linearized dynamics without noise ] ) associated with the wavenumber @xmath29 and weak coupling @xmath24 , and the additive term describes injection of noise into the phase dynamics . the local amplification of the noise can be described by the steady state covariance .
specifically , the expectation of the outer product of deviations from the attracting limit cycles is @xmath95 \left[\begin{array}{c}\delta a\\\delta\phi\end{array}\right]^t\ , \right\}.\ ] ] small entries in @xmath96 indicate a good robustness of the attractor to noise as the steady state variances and cross - covariances of the dynamics are small , representing small deviations around the equilibrium .
the eigenvalues of @xmath96 represent the axis lengths of the covariance ellipse .
large eigenvalues are associated with directions of large noise amplification when compared to eigenvalues which are close to zero . for distinct pairs of rotating waves ,
@xmath29 and @xmath42 , the covariance and associated eigenvalues are the same , exhibiting a common robustness to noise .
this underlying symmetry indicates that equal time will be spent between attractors @xmath97 and @xmath98 when driven by basin switching noise , to be discussed in the following sections .
as observed in the deterministic linearization , eq .
( [ eq : linearized dynamics without noise ] ) , stable limit cycles have one neutrally stable mode that is undamped by the dynamics and appears , in the presence of noise , as a random walk along the limit cycle .
the absence of a restoring force in this mode manifests itself as an unbounded eigenvalue of the covariance matrix .
in addition to the infinite eigenvalue , there is a zero eigenvalue associated with the eigenvector @xmath99^{t}$ ] , which represents the average amplitude of the dynamics .
this indicates that near the attractor the average amplitude is invariant to noise .
this invariant feature is necessarily present wherever the dynamics are well approximated by an attractor s linear characterization . even in the presence of attractor switching behavior ,
the average amplitude remains largely unchanged .
the remaining @xmath100 eigenvalues and associated eigenmodes indicate the individual character of the attractor basin each proportional to @xmath50 .
the average of these eigenvalues @xmath101 is given in closed form as @xmath102\right),\ ] ] where @xmath103 $ ] , providing a metric of the attractors robustness ( see appendix [ app : covariance ] for details ) .
dependence on the wavenumber @xmath29 comes in as the inverse of @xmath50 , indicating that as @xmath104 , the basins are more robust to noise .
examining the metric as @xmath105 , with the total input variance @xmath106 , wave fraction @xmath107 , and assuming large constant frequency @xmath13 , _
i.e. , _ @xmath108
then @xmath109 for large @xmath16 , the attractor robustness scales linearly with oscillator frequency and total input variance while inversely with the coupling strength and cosine of the phase differences .
the covariance matrix may be used to construct a quadratic quasi - potential for each rotating wave , which is guaranteed to be decreasing along the deterministic system trajectories for some finite neighborhood of the rotating wave and can be used to place lower bounds on the basin boundaries .
we can build a global quasi - potential by piecewise stitching together the local potentials associated with each rotating wave , always selecting the one with the least value : @xmath110_k^t p_k \left[\begin{array}{c}\delta a\\\delta \phi\end{array}\right]_k\right).\ ] ] slices of level sets of this potential for rings of @xmath34 oscillators are plotted in fig .
[ fig : timetoconvergence ] .
negative coupling gives a single basin , but its locally motivated potential well is much larger than the two equal potential wells of positive coupling . to compare this to the full nonlinear system
, we indicate the time - to - convergence in color , which cleanly shows the two basins of negative coupling , with the basin separatrix covering the set of points where one phase difference is zero . the covariance matrix @xmath111 associated with edge states @xmath112 can be formed from the covariance matrix @xmath96 . due to the symmetry in the dynamics ,
the diagonal elements of @xmath111 are in common and correspond to the steady state variance @xmath113 of each edge state with @xmath114 . a probabilistic feature that follows
is the steady state probability @xmath115 that , for a single instance in time , all edge states remain in the interval @xmath116 $ ] .
appendix [ app : probabilityinterval ] includes an approximation of the probability of interval containment using the error function @xmath117 , namely @xmath118^{n},\ ] ] where @xmath119 .
similar noise robustness characteristics can be observed over the edge states as the full states @xmath120 and @xmath121 with wavenumbers associated with @xmath50 close to @xmath122 providing more noise robustness and so higher probabilities of maintaining interval containment .
extending this concept into the time domain , the probability @xmath123}(\varepsilon_{l},\varepsilon_{u})$ ] of any edge state first exiting the interval @xmath116 $ ] in time span @xmath124 $ ] given a sampling interval @xmath125 and the expected exit time @xmath126 of this interval are @xmath127}(\varepsilon_{l},\varepsilon_{u})=p(\varepsilon_{l},\varepsilon_{u})^{\left\lfloor t_{1}/\delta t\right\rfloor } -p(\varepsilon_{l},\varepsilon_{u})^{\left\lfloor t_{2}/\delta t\right\rfloor } , \,\mbox { and } \,\mathbb{e}_{t}(\varepsilon_{l},\varepsilon_{u})=-\frac{\delta t}{\log(p(\varepsilon_{l},\varepsilon_{u}))}.\ ] ] oscillator ring prepared with @xmath128 .
subfigures ( a ) and ( b ) show the three amplitudes at two different time scales .
the dashed lines in these plots indicate the amplitude bounds established in appendix [ app : attractoramplitudes ] .
subfigure ( c ) shows the phase difference across each of the three edges , showing groupings at @xmath129 and @xmath130 , with rapid switches between them .
subfigure ( d ) shows those same phase differences over the same time as subfigure ( b ) , revealing that one phase difference passed through zero and rejoined the others at the other attracting state , indicating a switching event .
this trajectory was generated with @xmath131 , @xmath132 , and @xmath133 . ]
so far we investigated the local properties of attractors and the response to small noise such that the system remains in the vicinity of stable rotating - wave attractors . in this section ,
we consider larger noise levels in eq .
( [ eq : phasewithnoise ] ) at which the system occasionally switches from the vicinity of one attractor to the vicinity of another .
our goal is to provide a coarse - grained description of the global dynamics ; we wish to define an attractor switching network ( asn ) in which each node represents the neighborhood of an attractor and the links connecting nodes represent the switches . to build an asn
, we must first be able to distinguish the vicinities of distinct attractors .
it is computationally infeasible to capture the precise deterministic basins of attraction , so we investigate the characteristics of an attractor switch in a network of @xmath34 oscillators to motivate a coarse - graining . throughout
, we employ numerical simulations using a fourth order runge - kutta algorithm with timestep @xmath134 . at the end of each runge - kutta step , we add a zero - mean , normally distributed random number with variance @xmath135 to the phase of each oscillator to capture the stochasticity of eq .
( [ eq : phasewithnoise ] ) . figure [ fig : switchingtraj ] plots a typical stochastic trajectory in a ring of @xmath34 oscillators with coupling @xmath136 , nonlinearity @xmath137 , and noise level @xmath131 . as discussed in sec .
[ sec : stability ] , positive @xmath24 on the three - oscillator ring supports two stable attractors : rotating waves with wavenumbers @xmath138 and @xmath139 ( phase differences @xmath140 and @xmath141 )
. figures [ fig : switchingtraj]a - b show the amplitudes of the three oscillators at different temporal resolutions ; although noise is only directly added to the phases of the oscillators , it causes fluctuations in the amplitudes through the deterministic dynamics .
however , as shown in appendix [ app : attractoramplitudes ] , the amplitudes remain bounded within @xmath142 $ ] ( dashed lines in figs . [
fig : switchingtraj]a - b ) . figure [ fig : switchingtraj]c shows the phase differences @xmath143 over time .
the phase differences initially fluctuate around @xmath140 and at the time of the first switch ( @xmath144 ) they rapidly reorganize around @xmath141 .
figure [ fig : switchingtraj]d zooms in on that first switch , revealing that one of the phase differences passes through @xmath145 .
indeed , such phase crossing necessarily happens if the system transitions from one rotating wave to another with a different wavenumber .
thus , the mechanism underlying the switching dynamics is associated with the phases of two neighboring oscillators crossing . for a ring of @xmath146 oscillators based on @xmath147 independent measurements with average @xmath148 where the error is the standard error of the mean .
the distribution has an exponential tail : according to the kolmogorov - smirnov test for @xmath149 the distribution is consistent with an exponential distribution with @xmath150 ( @xmath151 , @xmath152-value @xmath153 ) , where the @xmath154 is the maximum likelihood fit of the rate parameter and the error corresponds to the 95% confidence interval .
the variance of the noise is @xmath155 ; @xmath132 and @xmath133 .
the theoretical curve is the probability of a zero crossing @xmath123}(-2\pi k / n,2\pi-2\pi k / n)$ ] based on the linear analysis . ] indexing the limit cycles states of a ring of @xmath146 oscillators .
each point is an average over @xmath156 independent measurements , and the error bars represent the standard error of the mean .
these simulations were run with @xmath155 , @xmath132 , and @xmath133 .
the solid black curve shows the analytic prediction of the expected zero crossing time @xmath157 . ]
finally , we partition the state space into regions enclosing each limit cycle according to wavenumber @xmath29 and investigate attractor switching phenomena as characterized by these partition boundaries .
in particular , we study the distribution of time needed to escape an attractor , the average times for such a switch to occur , and the overall organization of the attractor switching network ( asn ) . driven by the observation that attractor switching is accompanied by a phase crossing , we choose to identify a switch as an event when any @xmath158 becomes @xmath145 .
more precisely we calculate @xmath159 where @xmath160 .
since the oscillators are organized in a ring , @xmath29 is an integer . if the system is on a deterministic attractor , @xmath29 is equal to the corresponding wavenumber . thus , @xmath29 changes value only when a @xmath158 passes through zero .
we therefore coarse grain the state space by assigning the system to be in rotating - wave `` state '' @xmath29 as defined by eq .
( [ eq : state_detector ] ) .
the magnitude of change in @xmath29 is precisely equal to the number of adjacent phase difference that pass through zero at a particular time .
note that this assigns different volumes of state space to different rotating wave states .
for example , @xmath41 only if all @xmath161 , and small fluctuations in the phase differences cause discrete fluctuations in @xmath29 . hence
, this choice of coarse - graining is natural only if @xmath41 is unstable , _
i.e. , _ @xmath47 .
we perform measurements of switching by preparing the system in a stable attractor of the deterministic dynamics , letting it evolve until the system switches to another state according to eq .
( [ eq : state_detector ] ) , and then recording the time taken to switch , @xmath162 , and the new state . in fig .
[ fig : switchtimehist ] , we show a histogram of @xmath162 based on @xmath147 independent runs for state @xmath163 of a ring of @xmath146 oscillators with @xmath132 and noise level @xmath155 .
we find that the tail of the histogram is consistent with an exponential distribution ; the typical time needed to switch is therefore well characterized by the average @xmath164 .
the linear analysis prediction of switching probabilities is described by @xmath123}(\varepsilon_{l},\varepsilon_{u})$ ] in eq .
( [ eq : probabilityinterval ] ) with the zero cross condition corresponding to @xmath116=\left(-2\pi k / n,2\pi-2\pi k / n\right)$ ] . a similar exponential tail
is noted between between both curves .
the discrepancy for small @xmath164 , can be attributed to the linear regime assumption within the @xmath123}(\varepsilon_{l},\varepsilon_{u})$ ] calculation , specifically the independence of edge states over time . for small @xmath164 ,
the simulation is exhibiting a distribution similar to the hitting time induced by brownian motion rather than the independent and identically distributed random variable sampling of the linear analysis . in fig .
[ fig : switchtime - k ] we show @xmath164 as a function of @xmath29 for @xmath146 , @xmath132 and @xmath155 .
we find that @xmath164 is sharply peaked at @xmath163 , and vanishes as the system approaches the fully synchronized state @xmath41 or , equivalently , @xmath165 .
we compare the nonlinear stability measure @xmath164 to the expected switching time @xmath126 based on the steady state covariance , where @xmath116=\left(-2\pi k / n,2\pi-2\pi k / n\right)$ ] , defined by eq .
( [ eq : probabilityinterval ] ) .
the general shape and scale of the curves agree with deviations occurring as the curves depart from @xmath163 . as
the dynamics are unstable about the rotating - wave states @xmath166 $ ] , the linear analysis indicates an instantaneous switch compared to the nonlinear case where some time is required to depart from the unstable limit cycle .
deviations in the stable regime @xmath167 $ ] , can be attributed to uncaptured higher order modes in the dynamics and variable size of the linear regime across @xmath29 .
finally , we construct the asn for a ring of @xmath0 oscillators , with @xmath132 and @xmath155 by preparing the system in each @xmath29 rotating - wave state @xmath156 times and recording to which state it switches .
we show the asn in fig .
[ fig : asn ] ; red nodes represent stable rotating - wave states and gray nodes unstable states .
we draw a directed link from node @xmath168 to node @xmath169 if we observed a switch from @xmath168 to @xmath169 .
the link weight is the count of observed switches .
it is unlikely that two @xmath158 s become zero simultaneously , therefore typically switching happens from state @xmath29 to neighboring states @xmath170 .
the most unstable state @xmath41 is an exception , because at @xmath41 each @xmath161 and this allows switching to any state . in the few other cases where this occurs ,
the system simply passed through the intermediate partitions within a single time step of simulation .
that is , multiple @xmath158 s became zero within one @xmath171 increment .
overall the system evolves towards states where the adjacent oscillators are most out of phase , @xmath172 and @xmath173 , and it rarely leaves these states . although we have not proven that our list of limit cycles captures all attractors of the deterministic system , the lack of cycles with low @xmath164 in the asn provide indication that any further attractors are contained within a single partition and are therefore associated with a single wavenumber .
this example demonstrates that dynamical coarse - graining of the state space is an informative and necessary approach when constructing attractor switching networks for systems with noisy dynamics .
moreover , asns provide an insightful description of the complex and high - dimensional dynamics of noisy , multistable systems .
our long - term goal is to understand the architecture of basins of attraction in large - scale complex dynamical systems and to develop methods that reveal how state - space structures can facilitate driving between basins .
here , we took several key steps toward these larger goals by analyzing in - depth synchronization phenomena in a system of coupled oscillators arranged in a ring topology . from the equations governing the evolution of the system ,
we first predict analytically the different patterns of synchronization that can exist ( i.e. , rotating wave solutions ) and analyze their local stability via the linearization of the governing equations .
we then analyze the covariance matrix of deviations around the attracting rotating waves and use this to construct a piecewise quadratic quasi - potential roughly describing the full attractor space .
we additionally use this covariance matrix to make predictions about the fluctuations of phase differences .
although the covariance analysis allows us to analytically calculate a metric for the robustness of each attractor to noise , we turn to simulation to deal with the impact of large noise . with this
, we can explore the mechanisms associated with attractor switching and develop the attractor switching network .
doing so reveals a clear and strong drive towards those rotating waves with wavenumber approximately half the number of nodes , such that adjacent oscillators are nearly out of phase .
the techniques developed here should generalize to other systems and provide an systematic and analytic advance for developing the underlying theory of attractor switching networks .
we intend to further this study by carefully investigating the dynamics of single switches in larger rings , extending our methods to complex networks with richer attractor types , and validating them in nems nanoscale device experiments .
we thank mike cross , leonardo dueas - osorio , warren fon , matt matheny , michael roukes , and sam stanton for helpful discussions .
this material is based upon work supported by , or in part by , the u.s .
army research laboratory and the u. s. army research office under multidisciplinary university research initiative award w911nf-13 - 1 - 0340 . 10 e davidson and m levin .
gene regulatory networks . , 102(14):49354935 , 2005 .
cm gray , ak engel , p knig , and w singer .
synchronization of oscillatory neuronal responses in cat striate cortex : temporal properties .
, 8(04):337347 , 1992 .
sm rinaldi , jp peerenboom , and tk kelly . identifying , understanding , and analyzing critical infrastructure interdependencies .
, 21(6):1125 , 2001 .
v rosato , l issacharoff , f tiriticco , s meloni , s porcellinis , and r setola .
modelling interdependent infrastructures using interacting dynamical models .
, 4(1 - 2):6379 , 2008 . ad gonzlez , l dueas - osorio , m snchez - silva , and al medaglia .
the interdependent network design problem for optimal infrastructure system restoration . , 2015 .
h poincar .
les nouvelles mthodes de la mcanique cleste . , 1892 .
j milnor . on the concept of attractor . in _ the theory of chaotic attractors
_ , pages 243264 .
springer , 1985 .
s smale . .
springer , 1980 .
c conley .
the gradient structure of a flow : i. , 8(8*):1126 , 1988 .
c conley . .
number 38 .
american mathematical soc . ,
de norton .
the fundamental theorem of dynamical systems .
, 36(3):585598 , 1995 .
j hasty , j pradines , m dolnik , and jj collins .
noise - based switches and amplifiers for gene expression . , 97(5):20752080 , 2000 .
m timme , f wolf , and t geisel .
prevalence of unstable attractors in networks of pulse - coupled oscillators .
, 89(15):154105 , 2002 . a pikovsky , m rosenblum , and j kurths . , volume 12 . cambridge university press , 2003 .
fc hoppensteadt and em izhikevich . .
springer , berlin , 1997 .
ae motter , sa myers , m anghel , and t nishikawa .
spontaneous synchrony in power - grid networks . , 9(3):191197 , 2013 .
z nda , e ravasz , y brechet , t vicsek , and a - l barabsi .
self - organizing processes : the sound of many hands clapping . , 403(6772):849850 , 2000 .
re mirollo and sh strogatz .
synchronization of pulse - coupled biological oscillators .
, 50(6):16451662 , 1990 .
pc matthews , re mirollo , and sh strogatz .
dynamics of a large system of coupled nonlinear oscillators .
, 52(2):293331 , 1991 .
dm abrams and sh strogatz .
chimera states for coupled oscillators .
, 93(17):174102 , 2004 .
am hagerstrom , te murphy , r roy , p hvel , i omelchenko , and e schll . experimental observation of chimeras in coupled - map lattices .
, 8(9):658661 , 2012 . a arenas , a daz - guilera , and cj prez - vicente .
synchronization reveals topological scales in complex networks .
, 96(11):114102 , 2006 .
k wang , x fu , and k li .
cluster synchronization in community networks with nonidentical nodes . , 19(2):023106 , 2009 .
gb ermentrout .
the behavior of rings of coupled oscillators .
, 23(1):5574 , 1985 .
pc bressloff , s coombes , and b de souza .
dynamics of a ring of pulse - coupled oscillators : group - theoretic approach . , 79(15):2791 , 1997 .
ja rogge and d aeyels .
stability of phase locking in a ring of unidirectionally coupled oscillators .
, 37(46):11135 , 2004 .
da wiley , sh strogatz , and m girvan .
the size of the sync basin .
, 16(1):015103 , 2006 .
c - u choe , t dahms , p hvel , and e schll . controlling synchrony by delay coupling in networks : from in - phase to splay and cluster states .
, 81(2):025205 , 2010 .
s - y ha and m - j kang . on the basin of attractors for the unidirectionally coupled kuramoto model in a ring .
, 72(5):15491574 , 2012 .
av shabunin .
phase multistability in a dynamical small world network .
, 25(1):013109 , 2015 .
r lifshitz and mc cross .
nonlinear dynamics of nanomechanical and micromechanical resonators . , 1:152 , 2008 .
mc cross , a zumdieck , r lifshitz , and jl rogers .
synchronization by nonlinear frequency pulling .
, 93(22):224101 , 2004 .
mh matheny , m grau , lg villanueva , rb karabalin , mc cross , and ml roukes .
phase synchronization of two anharmonic nanomechanical oscillators .
, 112(1):014101 , 2014 .
w fon , m matheny , j li , rm dsouza , jp crutchfield , and ml roukes .
modular nonlinear nanoelectromechanical oscillators for synchronized networks , under review .
y kuramoto .
chemical oscillations , waves and turbulence , 1984 . is aranson and l kramer .
the world of the complex ginzburg - landau equation .
, 74(1):99 , 2002 .
sp cornelius , wl kath , and ae motter .
realistic control of network dynamics . , 4 , 2013 .
an pisarchik and u feudel .
control of multistability .
, 540(4):167218 , 2014 .
dk wells , wl kath , and ae motter .
control of stochastic and induced switching in biophysical networks .
, 5(3):031036 , 2015 .
hk khalil . .
prentice hall , upper saddle river , 1996 .
r diestel . .
springer , berlin , 2000 .
ra horn and cr johnson . .
cambridge university press , new york , 1990 .
s skogestad and i postlethwaite . .
wiley , west sussex , 2005 .
consider the quasi - potential @xmath174 and assume that @xmath175 then @xmath176\\ & \leq & \sum_{i=1}^{n}-\frac{1}{2}\mbox{sgn}(a_{i}-1)\left(a_{i}-1\right)+\frac{\left|\beta\right|}{2}\big(\left|a_{i+1}\right|+\left|a_{i-1}\right|\big)\\ & = & \sum_{i=1}^{n}-\frac{1}{2}\left|a_{i}-1\right|+\left|\beta\right|\left|a_{i}\right|\end{aligned}\ ] ] and so for @xmath177 then @xmath178 .
hence , the dynamics will converge to the invariant set @xmath179 @xcite .
the smallest annulus containing @xmath180 is @xmath181 $ ] and so the dynamics will converge to this annulus .
the linearized dynamics state matrix can be formalized as a series of kronecker sums as @xmath182\otimes i+\left[\begin{array}{cc } 0 & \beta c\\ -\beta c & 0 \end{array}\right]\otimes l+\left[\begin{array}{cc } -\beta s & 0\\ 0 & -\beta s \end{array}\right]\otimes m\right),\ ] ] where @xmath183 and @xmath184 .
now @xmath185\otimes\boldsymbol{1}/\sqrt{n}$ ] is a right eigenvector of @xmath54 , with associated left eigenvector @xmath186\otimes\boldsymbol{1}/\sqrt{n}$ ] and unique eigenvalue 0 .
this follows from @xmath187 and @xmath188 and by examining the eigenvectors and eigenvalues of the matrix @xmath189 $ ] . denoting the eigenvalues of an arbitrary matrix @xmath190 as @xmath191 , where @xmath192 , then @xmath193 .
consider the matrices @xmath194\otimes\boldsymbol{1}\boldsymbol{1}^{t}/n,\label{eq : perturbed linearized dynamics}\\ f_{2 } & = & f_{1}-\frac{1}{2}\left[\begin{array}{cc } -\beta s & 0\\ 0 & -\beta s \end{array}\right]\otimes m,\mbox { and}\label{eq : removed m linearized dynamics}\\ \widetilde{f } & = & \lim_{y\rightarrow0}f_{2},\label{eq : perturbed removed m}\end{aligned}\ ] ] then @xmath195 [ prop : lyapunov equivalence]consider @xmath196 positive semidefinite , @xmath96 positive definite , @xmath197 and hurwitz matrix @xmath198 .
the matrix @xmath96 satisfies @xmath199 if and only if @xmath200 . consider the permutation matrix @xmath201 defined as @xmath202 , @xmath203 for @xmath204 and @xmath205 otherwise .
the action of the permutation matrix @xmath206 on a vector corresponds to a mapping of element @xmath20 to element @xmath207 ( mod @xmath208 ) , and corresponds to the rotational automorphism on an @xmath208 node ring graph @xcite . as @xmath206 represents an automorphism of the graph , @xmath209 and @xmath210 . from the lyapunov equation
@xmath211 as @xmath206 is a permutation matrix on @xmath61 , then @xmath212 and @xmath213 as @xmath198 is hurwitz , the solution to the lyapunov equation is unique @xcite . hence , @xmath214 and @xmath215 and @xmath216
therefore , @xmath217 and @xmath218\\ & = & -i+\frac{1}{2}\beta s\left[-\left(i\otimes m\right)p+\left(i\otimes m\right)p\right]=-i.\end{aligned}\ ] ] [ prop : eigenvalues of a2]the eigenvectors of @xmath219 are of the form @xmath220 and @xmath221 where @xmath222 is a unit eigenvector of @xmath61 and @xmath223 and @xmath224 are the eigenvectors of the matrix @xmath225+x_{i}\left[\begin{array}{cc } 0 & 1\\ -1 & 0 \end{array}\right]\right).\label{eq : base eigenvector matrix}\ ] ] the associated eigenvalues of @xmath64 are @xmath226 , @xmath227 , and @xmath228 for @xmath229 . here , @xmath230 where @xmath231 are the eigenvalues of @xmath61 .
this result follows as @xmath232\otimes i+\left[\begin{array}{cc } 0 & \beta c\\ -\beta c & 0 \end{array}\right]\otimes l\right)v_{1i}\otimes w_{i}\\ & = & \frac{1}{2}\left(\left[\begin{array}{cc } -1 & 0\\ 4\alpha & 0 \end{array}\right]v_{1i}\otimes w_{i}+\left[\begin{array}{cc } 0 & 1\\ -1 & 0 \end{array}\right]v_{1i}\otimes\left(\beta cl\right)w_{i}\right)\\ & = & \frac{1}{2}\left(\left[\begin{array}{cc } -1 & 0\\ 4\alpha & 0 \end{array}\right]v_{1i}\otimes w_{i}+x_{i}\left[\begin{array}{cc } 0 & 1\\ -1 & 0 \end{array}\right]v_{1i}\otimes w_{i}\right)\\ & = & \frac{1}{2}\left(\left[\begin{array}{cc } -1 & 0\\ 4\alpha & 0 \end{array}\right]+x_{i}\left[\begin{array}{cc } 0 & 1\\ -1 & 0 \end{array}\right]\right)v_{1i}\otimes w_{i}=\mu_{1i}v_{1i}\otimes w_{i}.\end{aligned}\ ] ] hence , by solving for the eigenvalues of matrix ( [ eq : base eigenvector matrix ] ) , the eigenvalues of @xmath219 in closed form follow .
the covariance matrix @xmath233 associated with noise driven dynamics ( [ eq : linearized dynamics with noise ] ) with @xmath234 can be found using the lyapunov equation @xmath235 where @xmath236 $ ] @xcite and then taking the limit of @xmath237 . from proposition
[ prop : lyapunov equivalence ] , @xmath96 satisfies @xmath238 let @xmath239 where @xmath240 represents the eigenvectors of @xmath198 and @xmath241 the diagonal matrix of its eigenvalues .
further , as @xmath61 is symmetric then @xmath242 .
hence , @xmath243 multiplying on the left and right by @xmath244 and @xmath245 , respectively , and applying the condition @xmath246 , we have @xmath247\left(v^{-t}\otimes w^{-t}\right)\lambda\left(v^{t}\otimes w^{t}\right)p\left[i\otimes w\right]\\ & & + \left[i\otimes w^{t}\right]p\left(v\otimes w\right)\lambda\left(v^{-1}\otimes w^{-1}\right)\left[i\otimes w\right]+\left[i\otimes w^{t}\right]q\otimes i\left[i\otimes w\right]\\ & = & \left(v^{-t}\otimes i\right)\lambda\left(v^{t}\otimes i\right)\left[i\otimes w^{t}\right]p\left[i\otimes w\right]\\ & & + \left[i\otimes w^{t}\right]p\left[i\otimes w\right]\left(v\otimes i\right)\lambda\left(v^{-1}\otimes i\right)+q\otimes i.\end{aligned}\ ] ] let @xmath248p\left[i\otimes w\right]$ ] then @xmath249 equivalently after row / column permutations then @xmath250^{t}d\left[\tilde{p}_{is}\right]+d\left[\tilde{p}_{is}\right]d\left[f_{is}\right ] & = & -q\otimes i,\end{aligned}\ ] ] where @xmath251=\left[\begin{array}{ccc } f_{1s } & & 0\\ & f_{2s}\\ 0 & & \ddots \end{array}\right]$ ] . from prop .
[ prop : eigenvalues of a2 ] , the eigenvectors of @xmath198 are @xmath220 and @xmath221 with @xmath252 .
consequently , for @xmath229 with @xmath253 then @xmath254 $ ] and for @xmath255 with @xmath256 then @xmath257 $ ] . for @xmath253 , then @xmath258\\ \frac{1}{2}\left[\begin{array}{cc } -1 & 4\alpha - x_{i}\\ x_{i } & 0 \end{array}\right]\left[\begin{array}{cc } p_{11 } & p_{12}\\ p_{12 } & p_{22 } \end{array}\right]+\frac{1}{2}\left[\begin{array}{cc } p_{11 } & p_{12}\\ p_{12 } & p_{22 } \end{array}\right]\left[\begin{array}{cc } -1 & x_{i}\\ 4\alpha - x_{i } & 0 \end{array}\right ] & = & -\left[\begin{array}{cc } 0 & 0\\ 0 & \sigma^{2 } \end{array}\right]\end{aligned}\ ] ] so @xmath259.\ ] ] similarly , for @xmath255 then @xmath260\ ] ] with @xmath261,$ ] and its associated eigenvalue is @xmath262 the trace of @xmath96 without the modes associated with @xmath263 denoted as @xmath264 is @xmath265\tilde{p}\left[i\otimes w^{t}\right]=\mbox{tr}_{*}\left[i\otimes w^{t}\right]\left[i\otimes w\right]\tilde{p}\\ & = & \mbox{tr}_{*}\tilde{p}=\sum_{i=2}^{n}\mbox{tr}(\tilde{p}_{is})=-\sum_{i=2}^{n}\frac{\sigma^{2}}{x_{i}}\left(4\alpha - x_{i}+\frac{1 - 4\alpha x_{i}+x_{i}^{2}}{4\alpha - x_{i}}\right)\\ & = & -\sum_{i=2}^{n}\frac{\sigma^{2}}{x_{i}}\left(4\alpha - x_{i}-x_{i}+\frac{1}{4\alpha - x_{i}}\right)=2\sigma^{2}\sum_{i=2}^{n}\left(1-\frac{2\alpha}{x_{i}}-\frac{1}{2x_{i}\left(4\alpha - x_{i}\right)}\right)\\ & = & 2\sigma^{2}\left(n-1-\sum_{i=2}^{n}\left(\frac{2\alpha}{x_{i}}+\frac{1}{2x_{i}\left(4\alpha - x_{i}\right)}\right)\right).\end{aligned}\ ] ] on the ring network , @xmath266 where @xmath267 . using the relation @xmath268 , this trace is further simplified as , @xmath269\right)\end{aligned}\ ] ] where @xmath270 . as @xmath271 $ ] , @xmath272 $ ] and @xmath273 then @xmath274\subseteq\left(16\alpha\right)^{-1}\left[\left(\alpha+\left|\beta\right|\right)^{-1},\left(\alpha\right)^{-1}\right]$ ] .
the perturbed edge states on a ring about an equilibrium defined by phase offsets @xmath97 is described by the states @xmath275 or compactly by @xmath276\left[\begin{array}{c } \delta a\\ \delta\phi \end{array}\right]$ ] , where @xmath206 is defined in appendix [ app : covariance ] .
consequently , the covariance matrix @xmath277 can be found by a projection of the covariance matrix @xmath278\left[\begin{array}{c } \delta a\\ \delta\phi \end{array}\right]^{t}\right)$ ] as @xmath279p'\left[\begin{array}{cc } 0 & e - i\end{array}\right]^{t}.\ ] ] the trace of @xmath111 without the mode associated with the undamped subspace spanned by @xmath280 is denoted as @xmath281 . from appendix [ app : covariance ] , noting that @xmath282 , @xmath283p'\left[\begin{array}{cc } 0 & e - i\end{array}\right]^{t})=\mbox{tr}_{*}(\left[\begin{array}{cc } 0 & e - i\end{array}\right]p\left[\begin{array}{cc } 0 & e - i\end{array}\right]^{t})$ ] and applying the closed form solution for @xmath96 then @xmath284p\left[\begin{array}{c } 0\\ e^{t}-i \end{array}\right]\right)=\mbox{tr}_{*}\left(\left[\begin{array}{cc } 0 & 0\\ 0 & \left(e^{t}-i\right)\left(e - i\right ) \end{array}\right]\left(i\otimes w\right)\tilde{p}\left(i\otimes w^{t}\right)\right)\\ & = & \mbox{tr}_{*}\left(\left[\begin{array}{cc } 0 & 0\\ 0 & l \end{array}\right]\left(i\otimes w\right)\tilde{p}\left(i\otimes w^{t}\right)\right)=\mbox{tr}_{*}\left(\left(i\otimes w^{t}\right)\left(\left[\begin{array}{cc } 0 & 0\\ 0 & 1 \end{array}\right]\otimes l\right)\left(i\otimes w\right)\tilde{p}\right)\\ & = & \mbox{tr}_{*}\left(\left(\left[\begin{array}{cc } 0 & 0\\ 0 & 1 \end{array}\right]\otimes\lambda\right)\tilde{p}\right)=\sum_{i=2}^{n}-\frac{\sigma^{2}}{x_{i}}\mbox{tr}_{*}\left(\left[\begin{array}{cc } 0 & 0\\ 0 & \lambda_{i } \end{array}\right]\left[\begin{array}{cc } 4\alpha - x_{i } & 1\\ 1 & \left(1 - 4\alpha x_{i}+x_{i}^{2}\right)/\left(4\alpha - x_{i}\right ) \end{array}\right]\right)\\ & = & \sum_{i=2}^{n}-\frac{\sigma^{2}}{x_{i}}\mbox{tr}\left(\left[\begin{array}{cc } 0 & 0\\ \lambda_{i } & \lambda_{i}\left(1 - 4\alpha x_{i}+x_{i}^{2}\right)/\left(4\alpha - x_{i}\right ) \end{array}\right]\right)=\sum_{i=2}^{n}-\frac{\sigma^{2}}{x_{i}}\frac{\lambda_{i}\left(1 - 4\alpha x_{i}+x_{i}^{2}\right)}{4\alpha - x_{i}}\\ & = & -\frac{\sigma^{2}}{\beta c}\sum_{i=2}^{n}\frac{1 - 4\alpha x_{i}+x_{i}^{2}}{4\alpha - x_{i}}=\frac{\sigma^{2}}{\beta c}\sum_{i=2}^{n}x_{i}-\frac{1}{4\alpha - x_{i}}\\ & = & \sigma^{2}\left(\sum_{i=2}^{n}\lambda_{i}-\frac{1}{\beta c}\sum_{i=2}^{n}\frac{1}{4\alpha-\lambda_{i}\beta c}\right).\end{aligned}\ ] ] for the ring graph due to the underlying symmetry in the @xmath285 states then @xmath286 and so @xmath287 . for a ring graph then @xmath288 and @xmath289 , so @xmath290 let the probability that the random variable @xmath291 remains in the bounded interval @xmath116 $ ] be @xmath292 .
this probability can be calculated using the cumulative distribution function @xmath293 of the gaussian distribution and the error function @xmath117 as @xmath294-\frac{1}{2}\left[1+\mbox{erf}\left(\frac{\varepsilon_{l}}{\bar{\sigma}_{k}\sqrt{2}}\right)\right]\\ & = & \frac{1}{2}\left[\mbox{erf}\left(\frac{\varepsilon_{u}}{\bar{\sigma}\sqrt{2}}\right)-\mbox{erf}\left(\frac{\varepsilon_{l}}{\bar{\sigma}\sqrt{2}}\right)\right].\end{aligned}\ ] ] assuming that cross - coupling between @xmath285 s are small , the probability of all edge states remaining bounded @xmath115 can be approximated as @xmath295^{n}.\ ] ] the probability of exiting the interval @xmath116 $ ] by time @xmath296 given a sampling interval @xmath125 is then @xmath297}(\varepsilon_{l},\varepsilon_{u } ) & = & \sum_{k=1}^{\left\lfloor t/\delta t\right\rfloor } p(\varepsilon_{l},\varepsilon_{u})^{k-1}p(\varepsilon_{l},\varepsilon_{u})\\ & = & 1-p(\varepsilon_{l},\varepsilon_{u})^{\left\lfloor t/\delta t\right\rfloor } , \end{aligned}\ ] ] and consequently the probability of first exiting in the time span @xmath124 $ ] is @xmath298}(\varepsilon_{l},\varepsilon_{u})=p_{\left[0,t_{2}\right]}(\varepsilon_{l},\varepsilon_{u})-p_{\left[0,t_{1}\right]}(\varepsilon_{l},\varepsilon_{u})=p(\varepsilon_{l},\varepsilon_{u})^{\left\lfloor t_{1}/\delta t\right\rfloor } -p(\varepsilon_{l},\varepsilon_{u})^{\left\lfloor t_{2}/\delta t\right\rfloor } .\ ] ] noting that the cumulative distribution function for this event is therefore @xmath299}(\varepsilon_{l},\varepsilon_{u})$ ] the expected switching time is @xmath300 | following the long - lived qualitative - dynamics tradition of explaining behavior in complex systems via the architecture of their attractors and basins , we investigate the patterns of switching between qualitatively distinct trajectories in a network of synchronized oscillators .
our system , consisting of nonlinear amplitude - phase oscillators arranged in a ring topology with reactive nearest neighbor coupling , is simple and connects directly to experimental realizations .
we seek to understand how the multiple stable synchronized states connect to each other in state space by applying gaussian white noise to each of the oscillators phases .
to do this , we first identify a set of locally stable limit cycles at any given coupling strength . for each of these attracting states ,
we analyze the effect of weak noise via the covariance matrix of deviations around those attractors .
we then explore the noise - induced attractor switching behavior via numerical investigations .
for a ring of three oscillators we find that an attractor - switching event is always accompanied by the crossing of two adjacent oscillators phases . for larger numbers of oscillators
we find that the distribution of times required to stochastically leave a given state falls off exponentially , and we build an attractor switching network out of the destination states as a coarse - grained description of the high - dimensional attractor - basin architecture .
= 1 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ proper functioning of large - scale complex systems , from metabolism to global economics , relies on the coordination of interdependent systems .
such coordination the emergence of synchronization in coupled systems is itself an important and widely studied collective behavior .
however , predicting system behavior and controlling it to maintain function or mitigate failure present severe challenges to contemporary science .
prediction and control depend most directly on knowing the architecture of the stable and unstable behaviors of such high - dimensional dynamical systems . to make progress , here we explore limit - cycle attractors arising when ring networks of nonlinear oscillators synchronize , demonstrating how synchronization emerges and stabilizes and laying out the combinatorial diversity of possible synchronized states .
we capture the global attractor - basin architecture of how the distinct synchronized states can be reached from each other via attractor switching networks .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ |
we first review the main results concerning energetics and entropy production in kinetic networks . consider a system interacting with a reservoir at temperature @xmath0 and with states @xmath1 that have energies @xmath2 and are occupied with probabilities @xmath3 .
the system dynamics is described by a markovian master equation @xmath4,\ ] ] where @xmath5 is the net probability current from site @xmath6 to site @xmath1 .
the rates @xmath7 describing the reservoir induced transitions from @xmath6 to @xmath1 satisfy local detailed balance @xmath8 where @xmath9 and @xmath10 is a non - conservative thermodynamic force pointing from @xmath6 to @xmath1 which could be induced for instance by a non - equilibrium chemical reservoir or a non - conservative mechanical force .
the state energies @xmath2 may change in time , @xmath11 , due to the action of an external agent . for this generic scenario an unambiguous formulation of non - equilibrium thermodynamics
ensues @xcite .
the first law of thermodynamics reads @xmath12 and expresses the fact that the average energy of the system , @xmath13 , changes due to three mechanisms : the driving work corresponding to the energy transferred to the system by the external agent @xmath14 the non - conservative work corresponding to the energy transferred to the system by the non - conservative forces @xmath15 and the heat corresponding to the energy transferred from the reservoir to the system which , using , can be written as @xmath16 the second law expresses the fact that the change in the total entropy or entropy production ( i.e. the sum of the change in the system shannon entropy @xmath17 plus the change in the entropy of the reservoir @xmath18 ) is always nonnegative @xmath19 : @xmath20 where @xmath21 is the nonnegative edge entropy production expressed as a flux times a force @xcite . the total entropy production may also be rewritten as @xmath22 where @xmath23 is the nonequilibrium free energy of the system whose change can in turn be split as @xmath24 where @xmath25 \ , dt.\ ] ] note that , in the absence of non - conservative forces , @xmath26 , @xmath27
we now introduce a generic reversible pumping mechanism that transfers probability between two states in a way that is indistinguishable from the poissonian transitions of a markovian dynamics .
poissonian transitions are characterized by the following properties : _
i ) _ the probability transferred during a small time interval @xmath28 is @xmath29 , where @xmath30 is the rate of the transition ; and _ ii ) _ the occurrence of a transition in a given time interval is independent from the transitions that occurred in the past .
when considering a pump induced by a periodic driving of very small period @xmath28 ( this condition will be made more precise below ) and transferring a probability @xmath29 during each cycle , then the pump will mimic poissonian transitions since the transitions that occurred in a given cycle are independent of those occurring in the other ones . to be precise ,
consider the setup depicted in fig .
[ fig1 ] a. the system is made of observable states @xmath31 ( @xmath32 not shown in the figure ) and two hidden states @xmath33 connecting states @xmath34 .
the transition rates satisfy local detailed balance .
the transitions between @xmath33 and @xmath34 can be turned on and off by an external agent without any expenditure of work ( this can be achieved for instance for arrhenius rates @xmath35 by instantaneously raising and lowering the energy barriers @xmath36 ) and do not involve any non - conservative forces @xmath26 .
the external agent also controls the two energies @xmath37 and @xmath38 .
the operations performed by the external agent are cyclic and of period @xmath28 chosen to be much shorter than any time scale between the observable states , i.e. @xmath39 for @xmath40 .
we first describe the process along path @xmath41 where the energy @xmath37 and the barriers between @xmath42 and @xmath43 and between @xmath42 and @xmath44 are modulated .
the protocol starts with the two barriers closed and an energy @xmath45 , and proceeds as follows ( see fig .
[ fig1 ] b ) : a ) the barrier @xmath46 is opened ; b ) the energy @xmath37 is quasistatically lowered to @xmath47 ; c ) the barrier @xmath46 is closed ; d ) the energy @xmath37 is changed to @xmath48 ; e ) the barrier @xmath49 is opened ; f ) the energy @xmath37 is quasistatically restored to its initial value @xmath50 ; g ) the barrier @xmath49 is closed to complete the cycle . we note that , while the barriers can be opened or closed instantaneously , the changes in @xmath37 are carried out quasistatically to minimize the entropy production , except for step d ) where state @xmath42 is not connected with states @xmath43 and @xmath44 and @xmath37 can be changed arbitrarily fast without compromising the reversibility of the cycle . and @xmath51 with respective energies @xmath37 and @xmath38 , which are modified by an external agent in a cyclic way . the agent can also open ( solid vertical bars ) and close ( dashed vertical bars ) the barriers connecting the network states @xmath43 and @xmath44 with the pump states @xmath42 and @xmath51 .
b ) protocol followed by pump @xmath42 ( upper figure ) and pump @xmath51 ( lower figure ) .
we have labeled the 7 steps of the @xmath42 protocol , according to the description in the main text .
notice that the superscript of the energies indicates the network state @xmath1 which is in contact with the pump state @xmath33 when @xmath52 .
the pump is reversible if the changes in the energies @xmath53 are carried out quasistatically with respect to the time scale of the transitions between network states and pump states and if @xmath54 and @xmath55 are appropriately chosen ( see and ) .
the opening and closing of the barriers can be done instantaneously without compromising the reversibility of the process . ]
the cycle is engineered in such a way that site @xmath42 is practically empty at the beginning and at the end of the cycle .
therefore , initially @xmath56 and the probability to be on state @xmath34 is denoted by @xmath57 . during step
a ) an irreversible probability leak occurs from @xmath56 and @xmath58 to @xmath59 we assume @xmath60 and neglect this leak : @xmath61 and @xmath62 ( see the discussion below on the different scales of energy and time in our model ) . during step
b ) , a quasistatic reversible transfer of probability from state @xmath43 to state @xmath42 is performed . because the two states are in equilibrium with respect to each other , the respective occupation probabilities after step b ) are @xmath63 since , after step b ) , the barrier @xmath46 remains closed for the rest of the cycle , @xmath64 is the probability that will be transferred from site 1 to site 2 after the cycle is completed . for the pump to mimic poisson transitions
, this probability must be of the order of the duration of the cycle @xmath28 . therefore
, we impose the following scaling relationship between @xmath28 and the energy @xmath47 @xmath65 where @xmath66 is a finite rate which we will soon prove to be the effective rate of transitions from state 1 to state 2 . for the transitions to be poissonian we have to further impose that @xmath67 , which amounts to impose @xmath68 . in the following
, we approximate all the expressions up to first order in @xmath28 , since this is the approximation that yields a markovian dynamics ruled by an effective master equation , once the pump is coarse grained .
notice also that the initial energy @xmath50 should be even bigger than @xmath47 since it must lead to @xmath69 in order to justify neglecting the irreversible leak of step a ) . using the scaling ,
the transferred probability from @xmath43 to @xmath42 to first order in @xmath28 reads @xmath70 we now impose the following relation on @xmath48 @xmath71 which is equivalent to @xmath72 as a result , the probabilities @xmath64 and @xmath73 are in equilibrium when the barrier @xmath49 is opened in step e ) .
hence , the entropy production is zero along this step as well as along step f ) and g ) . beside the initial probability leak in step
a ) , which can be made arbitrary small , the remaining steps are reversible . we also note that in all previous expressions , the shifts in @xmath58 and @xmath73 , due to transitions with other observable states @xmath32 or due to the dynamics along path @xmath74 , are of order @xmath28 .
they will therefore only affect terms of second order in @xmath28 and will not prevent the entropy production of the process to vanish up to first order in @xmath28 .
we now turn to evaluating the work performed by the external agent on the system when changing the energy @xmath37 along steps b ) , d ) , and f ) . the remaining steps
a ) , c ) , and e ) involve no work since only the barriers are changed . as every step involving work is quasistatic and reversible , the driving work can be calculated as a difference of equilibrium free energy .
we find @xmath75 \simeq -ktp_a '' \nonumber \\ & & \hspace{-0.7 cm } w_d = p''_a\left[e_a^{(2)}-e_a^{(1)}\right ] \\ & & \hspace{-0.7 cm } w_f = -p_2kt\left[\ln \left(e^{-\beta e_2}\right)-\ln \left(e^{-\beta e_a^{(2)}}+e^{-\beta e_2}\right)\right ] \simeq ktp_a '' \nonumber . \end{aligned}\ ] ] the overall work along path @xmath41 can thus be written as @xmath76 = p''_a\left[e_2-e_1-kt\ln\frac{p_1}{p_2}\right ] .\label{wa}\end{aligned}\ ] ] the l.h.s . of this equation
is the change of free energy in the system due to the probability @xmath77 transferred by the pump , confirming that the entropy production due to the pumping mechanism vanishes .
we now turn to the process affecting path @xmath74 .
the energy @xmath38 and the barriers between @xmath51 and @xmath43 and between @xmath51 and @xmath44 are changed in a similar way as along path @xmath41 ( see fig .
[ fig1 ] c ) .
the analysis for this part of the protocol is analogous to that of @xmath41 , and the resulting expressions are obtained by just swapping @xmath42 and @xmath51 as well as @xmath43 and @xmath44 . by combining the results obtained along the two paths , @xmath78 and @xmath79 , we find the first important result of this paper , namely that the effective rates @xmath66 and @xmath80 satisfy a local detailed balance relation @xmath81 which , contrary to the original rates , now contains an effective non - conservative force @xmath82 pointing from @xmath43 to @xmath44 .
furthermore the total work performed by the pump during a cycle is given by @xmath83+p_b''\left[e_b^{(1)}-e_b^{(2)}\right ] \nonumber \\ & & \hspace{0.3 cm } = ( p_a''-p_b'')\left[-kt\ln\frac{p_1}{p_2}+e_2-e_1\right ] \nonumber \\ & & \hspace{0.3 cm } = j_{21}\tau\,\left[-kt\ln\frac{p_1}{p_2}+e_2-e_1\right ] = \delta{\cal f}_{21}=\delta{\cal f}_{12 } , \label{wdrtot}\end{aligned}\ ] ] where we used with @xmath84 for the last equality .
we now turn to the comparison between the real entropy production of the full network which includes the pumping states and the coarse grained entropy production obtained by just considering the dynamics on the observable states . for simplicity , we assume no non - conservative force besides the effective force @xmath85 emerging at the coarse grained level . examples with non - conservative forces will be provided in the applications .
we consider pumping cycles of duration @xmath28 much smaller than the characteristic time of the dynamics of the coarse grained network . at the coarse grained level of description ,
the observed states are not driven and the only non - conservative force is the effective one induced by the pump .
the total work in a cycle is therefore @xmath86 and , using , the entropy production per cycle reads @xmath87 where the sum runs over the observable states @xmath40 and @xmath88 is given by eq . with @xmath84 . on the other hand , in the full network
all forces are conservative . using and
, the true entropy production is given by @xmath89 when calculating the differential @xmath90 over a cycle of the pump operation , the contributions to @xmath91 from the hidden states @xmath33 vanish since they are empty at the beginning and at the end of the cycle .
one finds @xmath92=\sum_{i < j}\delta{\cal f}_{ij}.\ ] ] inserting in , we get that @xmath93 comparing this result with , we observe that the link @xmath94 does not contribute to the entropy production , confirming the reversibility of the pumping mechanism .
our second important result is that the true entropy production overestimates the coarse grained one : @xmath95 this result follows from comparing with using the inequality @xmath96 of special interest is the entropy production rate when the system reaches a stationary state . in this case , @xmath97 in , and the entropy production in the coarse grained network is given by the non - conservative work , whereas the real entropy production is proportional to the driving work .
the respective entropy production rates are : @xmath98 \label{si } .\end{aligned}\ ] ] note that @xmath99 may vanish even for a finite current @xmath100 ( an example is provided below ) .
the driving protocol that we have introduced to pump reversibly between a pair of observable states can be designed for any system with preassigned effective rates @xmath101 and operating in the stationary regime .
indeed , the choice of @xmath66 and @xmath80 determines the effective force @xmath85 via , and along with the rest of the markov chain , also determines the stationary values of @xmath58 and @xmath73 . from these stationary values we set @xmath48 and @xmath102 using eq . and we set @xmath47 and @xmath103 using .
it should be clear that our procedure can be easily generalized to systems with pumps located between several pairs of observable states and/or to systems with non - conservative forces besides the ones induced by the pumps .
some examples of this are provided below .
as a first example , we consider a system of @xmath104 states @xmath105 connected as a ring ( see fig . [ fig2 ] a ) .
the states energies are all zero @xmath106 , no non - conservative forces act on the system , and a hidden pump is inserted between states @xmath43 and @xmath44 .
local detailed balance implies equal rates @xmath107 for all transitions except @xmath66 and @xmath80 which satisfy @xmath108 , where @xmath85 is the effective force induced by the pump .
the stationary state of the system is @xmath109j}{(w_{21}-w_{12})w } \nonumber \\ p_{n } & = & p_1+(n+1-n)\frac{j}{w } , \ \
n=2,\dots , n,\end{aligned}\ ] ] where the clockwise stationary current @xmath110 is given by @xmath111}{2nw+n(n-1)(w_{21}+w_{12})}.\ ] ] the real entropy production rate is proportional to the driving work performed by the pump and reads ( see ) @xmath112,\ ] ] whereas the coarse grained entropy production rate is given by ( see ) @xmath113 if the network transitions are much slower than the effective rates from 1 to 2 , i.e. if @xmath114 , then the coarse grained entropy coincides with the real one @xmath115 .
the same is true if @xmath116 . on the other hand if @xmath117 , the real entropy production vanishes despite the fact that the network exhibits a finite dissipative current @xmath110 giving rise to an apparent entropy production @xmath118 .
these results generalize to any conservative network obeying detailed balance and containing a hidden pump in the edge @xmath94 .
if the rates along the normal edges are much larger than the effective rates along the pumping edge ( @xmath80 and @xmath66 ) , the whole chain will be at equilibrium with respect to the energy landscape @xmath2 , @xmath119 .
then , according to , @xmath120 , whereas a finite current @xmath121 gives rise an apparent entropy production @xmath122 .
this remarkable result is not in contradiction with any fundamental law of thermodynamics .
the dissipationless finite current arises from the large separation of time scales : the current is finite over the time scales @xmath123 of the dynamics on the observable states but is induced quasistatically over the internal time scale of the pump .
a similar phenomenon has been previously discussed in the context of adiabatic pumps @xcite . in the direction of the black arrow .
the examples are : a ) a ring with a pump connecting two network states , 1 and 2 .
b ) a kinetic network that produces high free energy molecules @xmath124 from low free energy molecules @xmath125 ( @xmath126 ) .
c ) a ring with pumps at every link , working against a uniform force @xmath127.,title="fig:",width=275,height=249 ] [ fig2 ] we now consider an enzyme switching between two conformational states @xmath43 and @xmath44 with the same energy @xmath128 .
the enzyme jumps due to two different mechanisms with respective rates @xmath129 and @xmath130 .
the first is induced by a hidden pump generating an apparent effective force @xmath131 from @xmath43 to @xmath44 , and the second is induced by a chemical reaction @xmath132 such that @xmath133 ( see fig . [ fig2 ] b ) .
when operating alone , both mechanisms favor their respective transition towards state @xmath44 . however ,
when operating simultaneously with @xmath134 , the pump can revert the spontaneous direction of the chemical reaction and thus generate high free energy molecules @xmath124 at a rate given by the ( clockwise ) stationary current @xmath135 where the stationary probabilities read @xmath136 the current may be rewritten as @xmath137 } { w^{\rm pump}_{12}+w^{\rm reac}_{12}+w^{\rm pump}_{21}+w^{\rm reac}_{21}}.\ ] ] the coarse grained and the real entropy production are obtained by adding the non - conservative work @xmath138 to eqs .
and respectively : @xmath139 the real entropy production ranges from reversibility , @xmath120 , if @xmath140 , to @xmath141 if the pump transfers probability much faster than the reaction , @xmath142 . in the former case it is therefore possible to produce molecules of @xmath124 with very high efficiency since the synthetase can work at finite rate with a vanishing entropy production .
as mentioned before , this does no contradict the second law of thermodynamics since the current occurs on much slower time scale than the driving .
one can even show that , for fixed @xmath85 , the efficiency at maximum power ( over @xmath143 ) tends to 1 when @xmath144 . to demonstrate that the reversible behavior can be achieved for a large but reasonable separation of time scales
, we numerically solved the master equation of the synthetase using arrhenius rates for the pumping rates @xmath145 , where @xmath36 .
this will also help us to show in detail how to build a reversible pump to be inserted in a given network .
energy units are measured in @xmath146 and time units in @xmath147 .
we consider a reaction with a difference in chemical potential between species @xmath124 and @xmath125 of @xmath148 and with rates ranging between @xmath149 @xmath150 and @xmath151 6.065 .
our goal is to build a pump exerting a force @xmath152 with effective rates @xmath153 and @xmath154 so that the system will produce @xmath124 molecules at a rate @xmath110 ranging between 0 ( for @xmath149 ) and 0.3 ( for @xmath155 ) , as obtained from eq . .
to do so , we first set the cycle time to @xmath156 , i.e. small enough for the pump to generate poisson rates at the coarse - grained level .
according to and the equivalent equation for pump @xmath74 , this together with @xmath128 fixes the energy of the hidden states @xmath42 and @xmath51 at the end of step b ) to @xmath157 and @xmath158 .
we then fix the energies after step d ) according to to @xmath159 and @xmath160 which depend on the specific value of @xmath161 .
for instance for @xmath162 , we get that @xmath163 and @xmath164 .
finally , we set the time scale of the internal transitions in the pump by fixing the value of the open barriers @xmath165 ( @xmath166 and @xmath167 ) . in our numerical solution
we open and close the barriers using linear ramps ranging from @xmath168 to @xmath169 .
the protocol for the energies and barriers is depicted in the inset of figure [ figsim ] .
the entropy production of the system @xmath170 obtained by full numerical integration ( black points connected by blue lines ) is depicted in fig .
[ figsim ] .
it approaches , but still differs from , the entropy production of the ideal reversible pump @xmath171 ( red curve ) and is clearly below the coarse - grained entropy production @xmath172 ( black curve ) .
the irreversibility in the pump causing the discrepancy between @xmath170 and @xmath171 mainly occurs at the end of step b ) and the beginning of step f ) .
our final example is a @xmath104-state ring with energies @xmath173 .
each edge contains a hidden pump generating a force @xmath174 ( clockwise ) and is subjected to a constant external torque @xmath127 ( counterclockwise ) operating against the pumps ( see fig .
[ fig2 ] c ) .
if all the pumps are identical , then the stationary state is uniform @xmath175 and the ( clockwise ) current reads @xmath176.\ ] ] it is positive for @xmath177 meaning that the pumps generate a finite speed rotation against the torque . as in the previous example
, the coarse - grained entropy production can be derived by adding to eq . the non - conservative work performed on the @xmath104 edges of the motor @xmath178 : @xmath179 it is a non - negative quantity which only vanishes at zero power @xmath180 .
the calculation of the real entropy production is more subtle since , contrary to what happens for the synthetase , the external force @xmath127 affects the internal transitions of the pumps , @xmath46 , @xmath181 , @xmath182 , @xmath183 . in order to use the calculations for the driving work @xmath184 made in the previous section to the pump between site @xmath1 and @xmath185
, we must set @xmath186 .
notice that the actual energy of site @xmath1 is zero , because the effect of the torque is borne by the external force .
however , the work performed by the driving in the pump between site @xmath1 and @xmath185 is given by with @xmath186 and @xmath187 .
the total driving work obtained by summing over the @xmath104 pumps is therfore @xmath188 and the real entropy production rate in the stationary regime vanishes @xmath189 remarkably , this motor is able to operate reversibly against any external torque @xmath127 .
we have proposed a reversible time - dependent driving mechanism ( called reversible pump ) which can be inserted between any two states of a kinetic network .
when coarse grained , this pump gives rise to a forward and backward poissonian rate between the two states .
the ratio of these effective rates satisfies a local detailed balance displaying an emergent nonconservative force .
remarkably , these pumps can always be engineered in such a way to operate reversibly when inserted in any steady state kinetic network .
we found that contrary to common belief , the coarse grained markovian kinetics generated by our pumps exhibits an entropy production which is always larger than the real one .
we exploit this fact to propose several
hyper efficient " setups which produce finite currents ( and thus finite entropy production ) at the coarse grained level while the real entropy production vanishes .
the origin of this surprising phenomenon is that coarse graining the driving affects the symmetry of the system under time reversal .
entropy production is a measure of the probabilistic distinguishability between a process and its time reversal @xcite . to define the time - reversed process one must consider the time - reversed driving .
but if the information concerning the driving is lost during the coarse graining procedure as is the case here , the time - reversal operation at the coarse grained level does not relate anymore to the real time - reversal operation at the underlying level .
a similar phenomenon may occur if hidden variables which are odd under time reversal are considered , such as velocity , angular momentum , or magnetic moment @xcite .
in fact , an external driving can be implemented by a large mass with a non zero initial velocity @xcite .
our setup has also an intriguing relation with information engines that use the information gathered in a measurement to extract work , in the spirit of the celebrated maxwell demon . in ref .
@xcite a driven kinetic scheme that works as a maxwell demon was introduced . when the demon is coarse grained , the resulting dynamics is markovian and mimics the dynamics of a chemical motor .
the scheme is not able to always work reversibly and it is more restrictive than the one presented here , but the demon is able to run the motor with less entropy production than chemical fuel . in this case , the hidden states are long lifetime states ( with respect to the internal time scale of the motor ) featuring the strong correlation between the motor and the demon implied by the measurement @xcite
. it would be interesting to find whether our general scheme also admits an interpretation in terms information .
our pumping mechanism is based on two ingredients , the presence of time asymmetric driving ( an `` odd variables '' ) and a large separation of time scales .
these can yield a dramatic enhancement of the performance of a kinetic network .
it is an open question whether these two ingredients can be helpful for designing more efficient chemical motors and nanodevices or whether they are already present in protein motors and other biological processes .
is supported by the national research fund , luxembourg in the frame of project fnr / a11/02 . j.m.r.p .
acknowledges financial support from grant enfasis ( fis2011 - 22644 , spanish government ) .
this work also benefited from the cost action mp1209 .
puglisi a , pigolotti s , rondoni l , vulpiani a ( 20100 ) entropy production and coarse graining in markov processes .
_ journal of statistical mechanics : theory and experiment _
p stacks.iop.org/jstat/2010/p05015 .
leonard t , lander b , seifert u , speck t ( 2013 ) stochastic thermodynamics of fluctuating density fields : non - equilibrium free energy differences under coarse - graining . _ the journal of chemical physics _ 139:204109 . | we show that a reversible pumping mechanism operating between two states of a kinetic network can give rise to poisson transitions between these two states .
an external observer , for whom the pumping mechanism is not accessible , will observe a markov chain satisfying local detailed balance with an emerging effective force induced by the hidden pump . due to the reversibility of the pump
, the actual entropy production turns out to be lower than the coarse grained entropy production estimated from the flows and affinities of the resulting markov chain .
moreover , in presence of a large time scale separation between the fast pumping dynamics and the slow network dynamics , a finite current with zero dissipation may be produced .
we make use of these general results to build a synthetase - like kinetic scheme able to reversibly produce high free energy molecules at a finite rate and a rotatory motor achieving 100% efficiency at finite speed .
entropy production measures dissipation and is the key quantity to assess the performance of a kinetic process .
the efficiency of chemical motors or of biochemical processes such as metabolic cycles , replication , transcription , or proofreading , typically achieves its maximum value when the entropy production is minimal .
hill , in his classic work @xcite on the transduction of free energy in chemical reactions , provided the basic tools to calculate the entropy production of processes modeled by kinetic networks .
closely related results were also found by schnakenberg @xcite .
they showed that the entropy production in a network consists of a sum of positive edge contributions , each expressed as the product of a probability flux across the edge times an edge affinity ( or thermodynamic force ) .
their theoretical framework constitutes the basis of stochastic thermodynamics @xcite which has become central for the study of molecular machines @xcite . in many applications
the observer has only a partial access to the kinetic network .
some states are hidden and the resulting description of the system becomes `` coarse grained '' .
the thermodynamic implications of coarse graining is an active field of research @xcite .
the coarse grained entropy production is typically lower than the actual one since it misses the positive contribution of the hidden edges .
this result is true for autonomous systems if the hidden variables are even under time reversal .
it can be proved in various ways @xcite and remains true even when the coarse grained description is no longer markovian @xcite .
indications that odd hidden variables do not satisfy this result were analyzed in the context of langevin equations where velocities were coarse grained @xcite . in this paper
we prove that for systems driven by an external time - dependent force , the entropy production at a coarse grained level of description may overestimate the actual entropy production .
the driving plays the analogous role of an odd hidden variable when it is not invariant under time reversal .
we exploit this result to build hyper - efficient pumps .
to do so we consider systems with hidden states driven by an external cyclic time - dependent force generating currents between the apparent states .
the driving is based on adiabatic pumping previously introduced in the literature @xcite but gives rise in our case to a markovian dynamics at the coarse grained level .
we show that our reversible pumps can be used to generate currents again a bias with zero entropy production , without contradicting the second law of thermodynamics . |
axisymmetric spreading of an insoluble surfactant on a thin layer of incompressible fluid has been the subject of numerous experimental and mathematical studies @xcite . motivated by the biomedical application of aerosol medications delivered to the thin fluid lining the lung , gaver and grotberg derived a mathematical model , based on lubrication theory , that couples the height profile of the fluid surface @xmath2 to the local surfactant concentration @xmath3 .
this model captures the driving force associated with the marangoni surface stress induced by spatial variations in surfactant concentration , which in turn depends on an equation of state that specifies the dependence of surface tension @xmath4 on @xmath5 .
while the model was developed for monolayer applications of surfactant , it has come to be applied both above @xcite and near @xcite the critical monolayer concentration @xmath6 , the concentration above which a single layer of surfactant molecules can no longer form .
similar models have been used to study thin films in bronchial systems @xcite , ocular systems including blinking dynamics @xcite , bulk solute transport @xcite , drying of latex paint @xcite , ink - jet printing @xcite , and secondary oil recovery @xcite .
numerical simulations have been used to confirm several predictions of the model @xcite based on analysis of self - similar solutions .
the simulations have also yielded detailed information about the spatiotemporal evolution of the free surface height and surfactant concentration profiles .
one key observation is that the motion of the leading edge of the surfactant - covered region sets its spreading rate .
the decrease in concentration at the leading edge induces a surface stress ( marangoni stress ) , which drives a capillary ridge ( local maximum ) in the fluid free surface that propagates along with the leading edge . in spite of these advances in understanding
, experimental confirmation of the model has been more difficult to obtain due to the difficulty of measuring the surfactant concentration and its dynamics .
consequently , attention has primarily focused on the evolution of the surface height profile @xcite , with the location of the leading edge @xmath7 of the surfactant layer inferred from other dynamics @xcite . within the near - monolayer regimes
for which the model was developed , two experiments have observed @xmath8 spreading behavior on millimetric glycerin films , for both oleic acid @xcite and fluorescently - tagged phosphocholine ( nbd - pc ) @xcite . in experiments with initial concentrations @xmath9 of surfactant exceeding @xmath6 , spreading behavior consistent with the predicted @xmath10
were observed by dussaud et al . for oleic acid on a sub - millimetric water - glycerin mixture , and by fallest et al . for nbd - pc on millimeter - thick glycerin .
the latter experiments were able to simultaneously measure both the capillary ridge and the spatiotemporal dynamics of @xmath3 , to be analyzed in more detail below .
sec . ] in order to verify the model prediction for the time - dependent distribution of surfactant concentration , we explore stricter tests of the model than have been performed previously . using data from fallest et al .
( for which @xmath5 is well above @xmath6 ) , we make a detailed comparison between the model predictions and the measured surface height profiles @xmath2 ( from a laser line ) and measured surfactant concentration profiles @xmath3 ( from azimuthal averages of fluorescent intensity at each point @xmath11 ) .
figure [ f : exp - image ] provides a schematic of the apparatus and a sample image .
the well - specified physical parameters additionally allow us to evaluate the accuracy of the characteristic timescale predicted for the spreading rate , rather than just the exponent . while we find approximate agreement in the spreading exponent and the coincidence of the surfactant leading edge with the capillary ridge , we also find two significant inconsistencies .
first , there is a mismatch between the characteristic timescale between model and experiment .
second , the spatial distribution of the surfactant differs markedly from what is predicted in simulations . to account for the extent to which these discrepancies might be due to amounts of surfactant well beyond the monolayer regime , we perform new experiments , modified to allow for the detection of monolayer concentrations of surfactant ( @xmath12 ) . in these experiments ,
we observe a distribution of surfactant which differs from the model predictions .
in addition , we find that there is no spreading capillary ridge , and that the spreading exponent for the leading edge of the surfactant is less than @xmath1 .
this value is well below predictions of the theory .
the outline of the paper is as follows . in [ s : experiment ] , we provide details about the experimental setup and the materials used . in [
s : model ] , we review the gaver - grotberg model @xcite , including a discussion of the choice of equation of state relating surfactant concentration to surface tension , and outline the finite difference method used for numerical simulation . in [ s : highgamma ] , we describe comparisons between numerical simulations and experimental results at initial concentrations well above @xmath6 .
as described above , we find partial agreement but also two inconsistencies when comparing experimental observations to numerical simulations . in an attempt to address the latter problem , in
[ s : hybrid ] we describe a hybrid model that takes the experimentally - measured @xmath3 and uses that quantity ( instead of the model for surfactant evolution ) to generate the evolution of the surface height profile .
we find that this provides reasonable agreement with the experimental data for @xmath2 .
in particular , the timescale in the simulations is set by the experimental surfactant distribution , and the height profile dynamics therefore evolve according to the experimentally - observed timescale . in
[ s : lowgamma ] , we report new experiments with @xmath12 and find significant disagreement with the model predictions , as summarized above .
we conclude in [ s : discussion ] with a discussion of the results and their significance .
in our experiments , we simultaneously record the surface height profile @xmath2 of the underlying glycerin fluid layer , as well as the local fluorescence intensity , which corresponds to the local concentration @xmath13 of insoluble lipids ( surfactant ) spreading on the surface . the basic apparatus , described in and shown in figure [ f : exp - image ] , consists of an aluminum well , a black light for exciting the nbd fluorophores , an oblique red laser line to illuminate the profile of the fluid surface , and a digital camera positioned directly above the experiment to capture the laser line and fluorescence ; this apparatus is used for the data presented in [ s : highgamma ] .
new experiments , presented in [ s : lowgamma ] , are optimized to permit visualization of monolayer concentrations of surfactant .
the bottom of the aluminum well is covered with a plasma - cleaned silicon wafer for improved reflectivity , and the fluorescent excitation is provided by 467 nm ( blue ) leds which coincide with the 464 nm absorption peak of the nbd fluorophore .
a green laser line illuminates the profile of the fluid surface , so that both it and the fluorescently - emitted light pass through a band - pass filter centered at the emission peak ( 531 nm ) .
these improvements to the optics permit us to collect images of the spreading dynamics at a framerate of 3 hz and an integration time of @xmath14 second , using an andor luca r camera optimized for fluorescence measurements .
the signal - to - noise ratio now sets a lower limit of @xmath15 for the detection of surfactant . for all experiments , we deposit 1-palmitoyl-2-\{12-[(7-nitro-2 - 1 , 3-benzoxadiazol-4-yl)amino]lauroyl } -sn - glycero-3-phosphocholine ( abbreviated nbd - pc , from avanti polar lipids ) within a retaining ring which is lifted to begin the spreading process .
this lipid molecule has one 12-carbon chain and one 16-carbon chain ; the nbd fluorophore is attached to the 12-carbon chain .
experiments are conducted on a layer of 99.5% anhydrous glycerin , at an initial depth of @xmath16 mm , and held at room temperature ( @xmath17 @xmath18c ) .
the lipids are initially deposited while dissolved in chloroform , which is allowed to evaporate for at least @xmath19 min before the retaining ring is lifted by a motor at @xmath20 mm / min .
this allows sufficient time for the meniscus to drain before it detaches from the ring .
.key dimensional parameters . [ cols="^,^,^",options="header " , ] the key material parameters for nbd - pc and glycerin are summarized in table [ t : ndval ] .
the initial conditions ( ic ) for the experiment are distinguished by the initial concentration @xmath21 of surfactant deposited within the ring , defined by @xmath22 where @xmath23 is the mass of nbd - pc deposited and @xmath24 is the radius of the ring .
although a small amount of surfactant remains on the ring after it has been lifted , we nonetheless use the nominal concentration @xmath21 to describe the different initial conditions in table [ t : initcond ] .
the experiments presented in [ s : highgamma ] all begin from ic6 , with @xmath21 above the critical monolayer concentration @xmath6 .
the experiments of [ s : lowgamma ] employ initial conditions that probe the monolayer regime ( below @xmath6 ) .
sec , for experiments with initial conditions ( a ) ic6 and ( b ) ic4 . the dashed circles ( of radii 3.4 cm and 1.1 cm , respectively ) , highlight the corrugations in the leading edge .
] figure [ f : exp - image]b shows a sample image of both the laser line ( measures the height profile @xmath25 and the location @xmath26 of its maximum ) , and the fluorescence intensity ( measures the surfactant distribution @xmath27 after azimuthal averaging and the location @xmath7 of its leading edge ) .
figure [ f : idist ] shows sample images of the surfactant distribution alone , for a representative @xmath28 case and a @xmath29 case . in each image ,
a sharp interface between the surfactant - covered and bare glycerin is readily visible ; the location @xmath7 of this interface is determined by identifying the annulus of maximum fluorescence intensity gradient . while the surfactant distribution is uniform for @xmath30 , several heterogeneities are present when @xmath31 .
first , the central region contains a greater concentration of surfactant than the regions closer to the leading edge , an effect that we will explore in more detail below .
second , there are filamentary patches of high concentration which also propagate out from the central region , becoming more dilute during the spreading dynamics .
we also note that the outer edge of surfactant in both cases has corrugations ; in the figure , white circles are imposed to emphasize that the edges are not quite circular .
although the surfactant distributions are never precisely axisymmetric , we nonetheless record the distribution by averaging azimuthally .
moreover , in the model and simulations of the following sections , we assume that the surfactant distributions are axisymmetric .
we consider the model derived by gaver and grotberg for a single layer of surfactant molecules spreading on a thin liquid film .
the model is a coupled system of partial differential equations for the height @xmath2 of the fluid free surface and the concentration @xmath3 ( mass per unit area ) of surfactant .
we assume axisymmetric spreading , and the variables are nondimensionalized : @xmath32 , @xmath33 , @xmath34 , @xmath35 , where @xmath36 indicates the dimensional variable .
[ rfulleqn1 ] @xmath37 the nondimensional parameter groups @xmath38 , @xmath39 , and @xmath40 result from nondimensionalization , using values of physical parameters listed in table [ t : ndval ] . the parameter @xmath41 balances gravity and marangoni forces , @xmath42 is the ratio of the capillary driving forces to the forces from the surface tension gradient , and @xmath43 represents the surface diffusion of the surfactant molecules where @xmath44 is the pclet number .
the function @xmath45 expresses the dependence of surface tension @xmath4 on surfactant concentration @xmath5 .
it is specified by an equation of state , as discussed in the next subsection .
the timescale @xmath46 sec achieves a balance between the terms on the left side of ( [ rfulleqn1 ] ) . in
[ modelparams ] , we test this predicted timescale directly . with @xmath47 , the leading edge of the surfactant distribution
is not precisely defined , since @xmath48 for all @xmath49 and for all @xmath50 .
nonetheless , although @xmath40 is very small , it is retained in numerical simulations , apart from one case , in which @xmath40 is set to zero in order to track the leading edge of the surfactant distribution .
since the boundary at @xmath51 is not a physical boundary , boundary conditions there are natural ; for large @xmath49 , the free surface is undisturbed on the time scale of the experiment , and the surfactant concentration is expected to be identically zero : @xmath52 a finite difference method is used to simulate ( [ rfulleqn1 ] ) and is summarized in the appendix .
the initial condition @xmath53 is chosen to reflect the initial height profile in the experiment , as the fluid meniscus detaches from the ring , thereby releasing the surfactant to spread across the fluid surface .
the initial distribution @xmath54 of surfactant within the ring is unknown , and is varied in the simulations to test its effect on the spreading .
@xmath55 1,&\frac{11\pi}{12}<r < r_{max } , \end{cases}\nonumber\\ \gamma(r,0)&=&\begin{cases}\gamma_0(r),&\mbox 0\leq r<\frac{5\pi}{12}\\[6pt ] 0,&\frac{5\pi}{12}<r < r_{max}. \end{cases}\end{aligned}\ ] ] in the simulations , we set @xmath56 in order provide a match between the initial condition and the measured height profiles at later times . for the simulations , the location @xmath57 of the edge of the domain is taken large enough that the influence of the boundary conditions at @xmath58 is negligible over the time of the experiment .
we take @xmath59 , but display graphs of @xmath60 and @xmath5 over the smaller domain , @xmath61 . ) , g : langmuir ( [ langeos ] ) , e : measured ( [ bulleos ] ) , m : multilayer ( [ meos ] ) . ] in order to compare the model ( [ rfulleqn1 ] ) to the results of the experiments described in , we need to choose an appropriate equation of state relating the surfactant concentration @xmath5 to the surface tension @xmath4 .
however , the model is valid only for a single layer of surfactant molecules ( @xmath62 ) but the experiments are conducted with initial surfactant concentrations up to @xmath63 . in an attempt to extend the model to the regime of the experiment , we consider different equations of state @xmath64 that have been proposed in the literature . in figure
[ f : eos ] , we show the graphs of four such functions ; we argue below that only the graph labeled * m * is suitable for modeling the full range of surfactant concentrations we wish to consider .
the equation of state we seek should have the following properties : @xmath65 , expressing the effect that increasing surfactant concentration decreases surface tension ; @xmath66 , since this is the range of values of surface tension in our nondimensionalization , with @xmath67 .
as can be seen in the figure , only graph * m * has these properties .
the linear equation of state @xmath68 and has been used widely @xcite .
this equation is generally chosen for simplicity ; it is also a reasonable linear approximation to nonlinear equations of state at low concentration .
note that @xmath69 is a negative constant ( @xmath70 in the figure ) . in the case of more than a monolayer of surfactant
, this equation suggests that the surface tension decreases endlessly which is not physical , as surfactant concentration beyond a monolayer has little further effect in decreasing surface tension .
bull et al .
determined an equation of state for nbd - pc on glycerin by fitting the data obtained using a tensiometer . using the nondimensional parameters in table [ t : ndval ] ( @xmath71 @xmath72g/@xmath73 ) , the corresponding formula for @xmath45 is @xmath74 however , this formula is applicable only for surfactant concentrations below approximately @xmath75 ( @xmath76 in figure [ f : eos ] ) , as @xmath77 decreases sharply for larger values of @xmath5 .
the langmuir equation of state , used in and , is @xmath78 where @xmath79 and @xmath80 .
when only a small amount of surfactant is introduced , a large change in the surface tension occurs and as the surfactant comes close to saturation ( a monolayer ) then adding more surfactant does not alter the surface tension very much .
however , the range of @xmath45 is @xmath81 $ ] rather than @xmath82 $ ] .
the multiple layer equation of state used by borgas and grotberg is related to the langmuir equation of state : @xmath83 this formulation is based on properties of surface tension discussed by sheludko and by an experimental fit by foda and cox who worked with an oil layer on water . in addition , @xmath45 remains positive at large @xmath5 , allowing us to simulate much higher concentrations of surfactant .
therefore , we use eq .
( [ meos ] ) for the simulations .
the @xmath84 spreading behavior predicted by ( [ rfulleqn1 ] ) has already been observed in prior experiments with oleic acid on a water - glycerin mixture @xcite and nbd - pc on glycerin @xcite . in this section ,
we make a more detailed comparison between the results from experiments and numerical simulations , using data from fallest et al . . while we are able to obtain reasonable agreement in the height profile shapes , a comparison of the dynamics ( [ modelparams ] )
requires that we adjust the timescale . in [ compprofile ] , we show a significant discrepancy between the observed distribution of surfactant and the prediction from simulations , even though the location and time evolution of the leading edge of the surfactant layer agree well , as detailed in [ spreadingexp ] .
we also describe attempts to capture the experimentally observed surfactant distribution by varying the initial distribution @xmath54 in the simulations .
measured at @xmath85 sec in the experiment ( short dashed lines , red ) , compared to numerical solutions at nondimensional times .
long dashed lines : @xmath86 ( uses calculated @xmath87 sec ) ; solid line : @xmath88 ( uses better - fitting @xmath89 sec ) .
data are from fallest et al . , with initial condition
ic6 . ] in order to compare the model and experiment , we convert the simulation results from dimensionless time @xmath90 to dimensional time @xmath91 using the relation @xmath92 , where @xmath93 sec is calculated from the dimensional parameters listed in table [ t : ndval ] .
previous simulations @xcite have treated the lengthscale @xmath24 as a free parameter , effectively adjusting the timescale to agree with the experimental observations . however , for the experiments analyzed here , the ring radius @xmath94 cm is known and consequently the time scale @xmath95 is determined , with no free parameters . in figure
[ f : thalf ] , we observe that the simulated height profile and the experimental data are inconsistent at @xmath96 sec if the determined value @xmath87 sec is used : neither the peak location nor its width are in agreement with the model . however , if we use @xmath97 sec instead of @xmath87 sec , thereby comparing the simulation at the later time @xmath98 to the same experimental data at @xmath96 sec , then both the position and width of the ridge are in approximate agreement between the model and experiment .
this agreement between simulation and experiment using the timescale @xmath97 sec is observed to hold for all times beyond an initial transient .
although an explanation for the discrepancy between times scales awaits further investigation , we use the empirically - determined @xmath97 sec as the timescale for the remaining comparisons in the paper .
measured at @xmath96 seconds in the experiment ( dashed lines ) and @xmath98 in numerical solution ( solid line ) ; the location @xmath99 of the peak is marked with a @xmath100 .
( b ) the corresponding surfactant concentration profiles @xmath101 in the experiment ( noisy ) and numerical solution ( smooth ) ; the location @xmath102 of the leading edge of the surfactant is marked with a @xmath100 .
data are from fallest et al .
2010 , with initial condition ic6 . ] in figure [ f : profile1 ] we compare the simulated and measured surface profiles and surfactant distributions , using the model parameters described in [ modelparams ] and experimental data from fallest et al . .
as can be seen in figure [ f : profile1 ] , the height profiles @xmath103 are in approximate agreement : locations of the maximum and minimum are in approximate agreement between simulation and experiment , and the overall shapes are similar . in contrast , the measured surfactant distribution has quite a different shape from the distribution predicted by the simulations .
while the model predicts a smooth decrease in @xmath104 away from the central peak , experiments instead show an extended plateau over which the surfactant concentration is nearly constant , and which appears to be drawn out of a reservoir , near the peak concentration at @xmath51 .
for longer times , the plateau extends and decreases in height .
these features do not appear in the numerical simulations . and
( g - i ) surface height profiles @xmath105 from numerical solutions at @xmath98 .
the experimental height profiles at @xmath96 seconds are the same in each case , and are shown as dashed lines . ] in the experiment , the total surfactant mass is known , but its initial spatial distribution within the retaining ring is not measured .
consequently , there is some uncertainty about the appropriate initial condition @xmath106 .
we explore whether the choice of @xmath54 could change the simulations enough to replicate the expanding plateau in the experimentally observed surfactant distributions .
we tested three different functions @xmath107 , shown in figure [ f : surfic ] : ( a ) uniform distribution ; ( b ) surfactant more concentrated near the retaining ring ; ( c ) step distribution .
the initial free surface height profile @xmath108 is the same in each case , given by ( [ numic ] ) .
the results of these simulations at time @xmath96 sec ( after short - time transients have died out ) are shown in figure [ f : surfic ] . in the middle column
, we observe that the distribution @xmath5 of surfactant does not exhibit a plateau for any of the initial conditions , while in the final column we see that the height profiles show broad agreement with the experiment in each case . and ( b ) the location of the leading edge of surfactant in both experiment and simulation .
colored lines are from experiments with ic6 adapted from fallest et al .
2010 , black dots are from simulations .
dashed lines are ( a ) comparison to best - fit @xmath109 with @xmath110 and ( b ) comparison to @xmath0 . ] in spite of the disagreements above , we observe that the spreading dynamics of the model and experiment are in good agreement when the artificial choice of @xmath111 sec is used .
the surface diffusion term @xmath112 in ( [ rfulleqn1h ] ) smooths the surfactant profile , and guarantees @xmath48 for all @xmath113 and @xmath50 .
this means the surfactant distribution has no clearly defined leading edge .
since @xmath114 has a very small effect , we take @xmath115 in these simulations . the surfactant distribution
is then supported at each @xmath50 on a bounded interval @xmath116 , and the leading edge @xmath117 can be tracked using the numerical scheme described in with @xmath118 . in figure
[ f : spread](b ) , @xmath119 is shown as a dotted line ; on the log - log plot the numerical solution is compared to the experimental results and to the analytic form @xmath120 derived from the similarity solution of , in which @xmath121 .
we also tracked the capillary ridge @xmath122 , where the height profile @xmath2 has a maximum . in figure
[ f : spread](a ) we show the numerical solution as a dotted line , and compare to experimental results , with an approximate slope shown with a dashed line .
the data for the leading edge of the surfactant agrees with the @xmath123 prediction of the model .
the capillary ridge moves faster as it catches up to the surfactant leading edge ; it is best fit by @xmath124 over the duration of the experiment .
note that for @xmath121 in the model , the fluid surface experiences a discontinuity at the leading edge @xmath117 of the surfactant , and the fluid is undisturbed ahead of this front .
it is worth noting that with @xmath115 , the surfactant distribution still has a finite extent when @xmath38 or @xmath39 is non - zero , but the disturbance of the film does not .
in fact , the surface tension gradient induced at the leading edge @xmath117 of the surfactant generates fluid motion ahead of @xmath7 .
due to the failure of the model equations ( [ rfulleqn1 ] ) to capture the spatial distribution of the surfactant , as discussed in
[ s : highgamma ] , we consider whether there is a fundamental modeling problem with the equation ( [ rfulleqn1 g ] ) for the evolution of @xmath3 .
given that the height profile @xmath2 evolves in a quantitatively reasonable way , we leave ( [ rfulleqn1h ] ) intact , but take @xmath3 from the experiment and use it to determine @xmath2 numerically .
the experimental data for @xmath5 are noisy , and occur at discrete times ; we must smooth and interpolate the data for use in the numerical scheme to determine @xmath60 .
the first step is to smooth the experimental data at each recording time .
this is achieved using the matlab _ smooth _ function , which performs a moving spatial averages over a specified span of data points ; we found a low - pass filter with a span of 21 points to be effective .
the experimental data are recorded at one second intervals ; however , the numerical method requires the time step @xmath125 to be on the order of @xmath126 for stability . to remedy this inconsistency ,
we interpolate the smoothed surfactant profiles to obtain functions that can be used to represent the surfactant concentration @xmath3 at all values of @xmath49 and @xmath90 . from the numerical simulations ( solid blue ) and experimental data ( dashed red ) .
( e - h ) the corresponding surfactant profiles from the experiment , smoothed data ( smooth blue ) and raw data ( noisy green ) .
the profiles are shown at @xmath127 sec . ] first , we use the nonlinear fit function , _ nlinfit _ in matlab to fit the smoothed surfactant data at each experimental recording time to a function with a graph consistent with the two step structure of the experimental surfactant distributions .
as observed in figure [ f : exp4 ] , the experimental surfactant distribution @xmath5 is roughly constant in each of the two steps . with this in mind , we use the function @xmath128 to fit the surfactant profile at each time @xmath91 for which we have data , where @xmath129 sec .
this procedure generates coefficients @xmath130 .
next we create polynomial functions , @xmath131 from the discrete values @xmath132 using _ polyfit _ in matlab , but with a restriction that the mass of the surfactant must be conserved . since the numerical code requires nondimensional time @xmath133 , we define @xmath134 .
these functions define the surfactant concentration profiles that approximate the experimental data : @xmath135 we update the height profile using the finite difference scheme used in [ s : highgamma ] with boundary conditions ( [ bcsym ] ) and the surfactant concentration profile using ( [ gfit ] ) : @xmath136 the flux functions @xmath137 are described in the appendix . in figure [
f : exp4 ] , we show simulation results using initial condition ( [ numic ] ) and parameters @xmath138 .
note that this value of @xmath39 is larger than the value suggested by the nondimensional grouping .
this is in order to smooth the height profile , which otherwise would develop multiple persistent ridges , due to the steep gradient in the surfactant concentration .
multiple ridges are observed at very early time in the simulations discussed in [ compprofile ] but then the gradient in the surfactant is quickly smoothed . in this case , the ridge in the height profile from the surfactant gradient and the one from the initial condition ( due to the lifting of the ring ) combine at very early time and then propagate as one . by contrast , in the experiment the surfactant distribution does not experience this smoothing and consequently when the experimental values for the surfactant concentration are input into the height equation a second ridge develops and persists . by increasing the capillarity parameter @xmath39 , the height profile is smoothed , and the numerical height profile becomes more similar to the experimental height profile .
the surface height profile simulated using this hybrid model exhibits an evolution and shape that is comparable to the experiment .
we conclude that the equation modeling surfactant molecule motion through passive transport by the surface fluid is fundamentally flawed , and is missing some physical or chemical properties that would generate the surfactant distributions observed in the experiment .
because the original model ( [ rfulleqn1 ] ) was developed for use with monolayer concentrations of surfactants ( @xmath30 ) , we conduct new experiments in this regime . in addition , these experiments help elucidate the discrepancies between model and experiment at the higher concentrations . performing experiments at lower surfactant concentrations
requires improvements of our earlier techniques ( see [ s : experiment ] ) in order to visualize lower surfactant concentrations .
we perform experiments starting from four different initial concentrations , ic1-ic4 in table [ t : initcond ] , all of which result in similar spreading dynamics , described below . in no case
do we find that the agreement with the model is improved over the @xmath31 case : height profiles , surfactant distribution , and the spreading dynamics all significantly disagree with the model predictions . from the center of the surfactant region , shown at representative times , taken from an experiment with ic4 ( @xmath30 ) .
] as illustrated in figure [ f : idist ] , and shown quantitatively in figure [ f : fluoro ] , the spreading region has an approximately uniform surfactant distribution throughout the lipid - covered area .
the leading edge , located at @xmath7 , exhibits a sharp interface ( approximately @xmath139 mm wide , neglecting azimuthal corrugations ) that does not broaden as the surfactant spreads outward .
instead , the overall concentration decreases throughout the lipid - covered region .
these observations are in disagreement with the model , which predicts a monotonically decreasing profile ( see figure [ f : profile1 ] ) with a gradual and broadening transition to @xmath140 at @xmath7 . ) and ic2 ( f - j ) ( @xmath141 ) at five different times .
the uppermost bright line in each image is the reflection from the air glycerin interface , and is used to measure the free surface height profile @xmath142 . ]
figure [ f : mridge ] provides a comparison between the behavior of the fluid ridge in two regimes . for @xmath143 ,
the glycerin is pushed out by the surface tension gradient , and forms a capillary ridge at the surfactant edge ( panels f - j ) . as already discussed in [ compprofile ] , these two features travel outward together .
this is consistent with simulations at high or low surfactant concentrations , that show a fluid ridge propagating outwards due to marangoni forces .
in contrast , the experimental data in the @xmath144 regime ( panels a - e ) show an initial ridge ( due to the meniscus at the ring ) collapsing rapidly in place , and no discernible ridge propagating outwards . ) , scaled by the ring radius @xmath24 .
the dashed lines corresponding to @xmath0 and @xmath145 spreading behavior are shown for comparison . ]
given the significant differences between the observed and modeled @xmath2 and @xmath3 , it is unsurprising that we observe the spreading dynamics to be quite different as well . in figure
[ f : rvst ] , we plot the position of the leading edge @xmath7 as a function of time , for experiments in both the monolayer regime of ic1-ic4 and for higher initial concentrations ( ic5-ic6 ) . for @xmath30 ( ic1-ic4 )
the dynamics all follow a form @xmath146 , with @xmath147 . remarkably , this is reminiscent of tanner s law for fluid spreading on a solid @xcite . for slightly larger values of @xmath21 ( ic5 ) ,
we observe faster surfactant spreading dynamics initially , but @xmath146 with @xmath148 at later times .
this slow - spreading regime was not reached in the runs with much larger values of @xmath21 ( ic6 ) , for which @xmath149 remains close to @xmath150 , as predicted in @xcite .
the decrease in @xmath149 for low surfactant concentrations suggests a transition in the dynamics which is not covered by the model equations .
the differing prefactors for the various runs is probably a result of variations in @xmath21 due to some lipid molecules remaining on the ring after it lifts off , an effect that is more significant at lower concentrations .
current mathematical models that describe the dynamics of the free surface of a thin fluid layer subject to forces induced by variations in surface tension contain two key assumptions : ( 1 ) lubrication theory is valid , and ( 2 ) surfactant molecules are primarily passively - transported by the fluid motion along the surface , along with negligible molecular diffusion .
however , the dependence of surface forces on local variations in surfactant concentration is not completely settled , especially for larger concentrations . due to the coupling between the motion of the underlying fluid and the spreading of surfactant molecules ,
it is crucial to compare results from simulations and experiments for both fluid motion ( through surface deformations ) and the dynamics of surfactant distribution . in this paper
, we compare model predictions to experiments that include the simultaneous visualization of the fluid height profile and the distribution of surfactant , both above and below the critical monolayer concentration @xmath6 . in both cases ,
we find serious inconsistencies between the model and the experiments .
the aspect ratio for the experiments is @xmath151 , which justifies the use of the lubrication approximation ; we have not verified the magnitude of the vertical velocity profile . at all initial concentrations , both above and below @xmath6 , the spatial distribution of surfactants does not follow the smooth , monotonically - decreasing profiles predicted by the model . at low surfactant concentrations ( @xmath30 , for which the models were originally developed ) , the distribution is highly - uniform with a sharp interface at the leading edge .
second , spreading occurs much more slowly than is predicted by the model . for all experiments with @xmath30 ,
the spreading dynamics of the leading edge approximately follow a power law @xmath152 , with @xmath153 .
this is significantly smaller than the @xmath154 predicted by the natural scaling in the model .
interestingly , this exponent is also markedly different from the exponent of @xmath155 to @xmath156 observed by gaver and grotberg for oleic acid and by bull et al .
@xcite for nbd - pc . in the former case ,
the measurement technique relies on the model for interpretation of experimental results , while in the later the @xmath157 mm fluid layer thickness may be large enough that deviations from the lubrication approximation are significant . for @xmath158 , even though @xmath159 , there is a mismatch by a factor of two between the timescales predicted in the model and observed in experiment .
these inconsistencies are not resolved by changing assumptions concerning the initial distribution of surfactant . moreover , if a measured @xmath3 is incorporated directly into the lubrication model , then the timescale issue is no longer present , as the spreading ridge is simply driven by the spreading surfactant .
one untested assumption is the functional form of the equations of state which have been considered to date .
the lack of spreading ( @xmath153 for very low concentrations ) might indicate that the assumed @xmath45 equation of state is inadequate .
if there were a value of @xmath5 below which there were no longer a significant surface tension gradient , a lack of spreading would be expected .
in fact , in static surface - pressure measurements of diolein , oleyl alcohol , and lecithin on water , such an effect has been observed @xcite .
future work to make similar measurements for nbc - pc , potentially locating a second transition point , may clarify the reason for the reduction in spreading .
another possibility is that the passive transport model for surfactant distribution on the free surface is missing one or more effects that influence on the dynamics of insoluble surfactant spreading on thin liquid films .
we are grateful for support from the national science foundation under grant number dms-0968258 and research corporation under grant number 19788 .
in addition , we wish to thank rachel levy for valuable conversations concerning surfactant spreading .
00 ahmad , j. and r. hansen : 1972 , ` a simple quantitative treatment of the spreading of monolayers on thin liquid films ' .
( 3 ) , 601604 .
borgas , m. s. and j. b. grotberg : 1988 , ` monolayer flow in an thin film ' . , 151170 .
braun , r. : 2012 , ` dynamics of the tear film ' . , 267297 .
bull , j. l. , l. k. nelson , j. t. walsh , m. r. glucksberg , s. schurch , and j. b. grotberg : 1999 , ` surfactant - spreading and surface - compression disturbance on a thin viscous film ' .
, 8998 .
dussaud , a. d. , o. k. matar , and s. m. troian : 2005 , ` spreading characteristics of an insoluble surfactant film on a thin liquid layer : comparison between theory and experiment ' .
, 2351 .
edmonstone , b. d. , o. k. matar , and r. v. craster : 2004 , ` flow of surfactant - laden thin films down an inclined plane ' .
( 2 - 3 ) , 141156 .
evans , p. l. , l. w. schwartz , and r. v. roy : 2000 , ` a mathematical model for crater defect formation in a drying paint layer . ' .
( 1 ) , 191205 .
fallest , d. w. , a. m. lichtenberger , c. fox , and k. e. daniels : 2010 , ` fluorescent visualization of a spreading surfactant ' . , 73029 .
foda , m. and r. g. cox : 1980 , ` the spreading of thin liquid films on a water - air interface ' .
, 3357 .
gaver , d. p. and j. b. grotberg : 1990 , ` the dynamics of a localized surfactant on a thin - film ' . , 127148 .
gaver , d. p. and j. b. grotberg : 1992 , ` droplet spreading on a thin viscous film ' .
, 399414 .
grotberg , j. b. : 2011 , ` respiratory fluid mechanics ' .
( 2 ) , 021301 .
gundabala , v. r. , c. lei , k. ouzineb , o. dupont , j. l. keddie , and a. f. routh : 2008 , ` lateral surface nonuniformities in drying latex films ' .
( 12 ) .
gundabala , v. r. and a. f. routh : 2006 , ` thinning of drying latex films due to surfactant . ' .
( 1 ) , 30614 .
hanyak , m. , a. a. darhuber , and m. ren : 2011 , ` surfactant - induced delay of leveling of inkjet - printed patterns ' .
( 7 ) , 074905 .
hanyak , m. , d. k. n. sinz , and a. a. darhuber : 2012 , ` soluble surfactant spreading on spatially confined thin liquid films ' .
( 29 ) , 7660 .
henderson , d. m. : 1998 , ` effects of surfactants on faraday - wave dynamics ' .
( -1 ) , 89107 .
jensen , o. e. : 1994 , ` self - similar , surfactant - driven flows ' .
( 3 ) , 10841094 .
jensen , o. e. and j. b. grotberg : 1992 , ` insoluble surfactant spreading on a thin viscous film : shock evolution and film rupture ' .
, 259288 .
jensen , o. e. and j. b. grotberg : 1993 , ` the spreading of heat or soluble surfactant along a thin liquid film ' .
, 5868 .
jensen , o. e. , d. halpern , and j. b. grotberg : 1994 , ` transport of a passive solute by surfactant - driven flows ' .
( 8) , 11071117 .
levy , r. , m. shearer , and t. p. witelski : 2007 , ` gravity - driven thin liquid films with insoluble surfactant : smooth traveling waves ' .
, 679708 .
matar , o. k. and s. m. troian : 1999 , ` spreading of a surfactant monolayer on a thin liquid film : onset and evolution of digitated structures ' .
( 1 ) , 141153 .
peterson , e. r. and m. shearer : 2010 , ` radial spreading of surfactant on a thin liquid film ' . .
peterson , e. r. and m. shearer : 2012 , ` simulation of spreading surfactant on a thin liquid film ' .
( 9 ) , 5157 5167 .
sakata , e. k. and j. c. berg : 1969 , ` surface diffusion in monolayers ' .
, 570575 .
segur , j. b. and h. e. oberstar : 1951 , ` viscosity of glycerol and its aqueous solutions ' .
( 9 ) , 21172120 .
sheludko , a. : 1966 , _
colloid chemistry_. elsevier .
sinz , d. k. n. , m. hanyak , j. c. h. zeegers , and a. darhuber : 2011 , ` insoluble surfactant spreading along thin liquid films confined by chemical surface patterns . ' .
( 20 ) , 976877 .
tanner , l. h. : 1979 , ` the spreading of silicone oil drops on horizontal surfaces ' .
, 14731484 .
warner , m. r. e. , r. v. craster , and o. k. matar : 2004 , ` fingering phenomena created by a soluble surfactant deposition on a thin liquid film ' .
( 8) , 29332951 .
wulf , m. , s. michel , k. grundke , o. i. del rio , d. y. kwok , and a. w. neumann : 1999 , ` simultaneous determination of surface tension and density of polymer melts using axisymmetric drop shape analysis ' . ( 1 ) , 172 181 .
we summarize the finite difference method used to generate numerical results for system ( [ rfulleqn1 ] ) .
simulations are conducted on a large interval @xmath160 . for simplicity , we consider uniformly distributed grid points @xmath161 , where @xmath162 . at each time @xmath163 ,
let @xmath164 and @xmath165 .
we use the standard notation for spatial averages of @xmath166 , @xmath167 the numerical method is an implicit finite difference scheme in conservative form : in the hybrid model of [ s : hybrid ] , the values of @xmath172 are derived from the experimental data , as described in that section .
the height profile evolution can then be computed from equation and the companion equation is discarded . | the spreading dynamics of surfactant molecules on a thin fluid layer is of both fundamental and practical interest .
a mathematical model formulated by gaver and grotberg @xcite describing the spreading of a single layer of insoluble surfactant has become widely accepted , and several experiments on axisymmetric spreading have confirmed its predictions for both the height profile of the free surface and the spreading exponent ( the radius of the circular area covered by surfactant grows as @xmath0 ) . however , these prior experiments have primarily utilized surfactant quantities exceeding ( sometimes far exceeding ) a monolayer . in this paper
, we report that this regime is characterized by a mismatch between the timescales of the experiment and model , and additionally find that the spatial distribution of surfactant molecules differs substantially from the model prediction . for experiments performed in the monolayer regime for which the model was developed ,
the surfactant layer is observed to have a spreading exponent of less than @xmath1 , far below the predicted value , and the surfactant distribution is also in disagreement .
these findings suggest that the model is inadequate for describing the spreading of insoluble surfactants on thin fluid layers . |
the existence of yukawa - type long - range interactions between the atoms of macro - bodies is predicted by the modern unified gauge - field theories of fundamental interactions , from the principles of supersymmetry and supergravity .
these forces may be understood as a result of the exchange of light elementary particles ( such as the axion , scalar axion , dilaton , spin - one antigraviton etc .
they may be called _ long - range _ because their range of action is much larger than the typical nuclear sizes .
additional yukawa - type interactions may arise also from corrections to the newtonian gravitational law at small distances @xcite . at present , this subject is actively investigated .
there are a large number of papers from both theoreticians and experimentalists . an extended annotative collection of references can be found in @xcite .
the most reliable way to get constraints on the parameters of the yukawa - type long - range interactions is via laboratory experiments which are sensitive to the presence of additional forces .
examples are the etvs- , galileo- and cavendish - type experiments , casimir and van der waals force measurements , and measurements of transition probabilities in exotic atoms .
a collection of modern constraints following from the above - mentioned experiments can be found in @xcite .
casimir forces which arise between closely located surfaces of solid bodies due to the presence of electromagnetic vacuum fluctuations provide us with one of the most interesting possibilities to restrict the permissable strength of the hypothetical interaction .
a detailed analysis of the casimir effect , including those aspect under discussion here can be found in the monograph @xcite .
it was first suggested in @xcite that constraints on the yukawa - type interactions can be derived from the casimir effect .
the strongest constraints , which follow from the casimir force measurement between a quartz plate and a spherical lens , were obtained in @xcite ( see also @xcite ) .
they provide the strongest constraints for a range @xmath0 of the yukawa forces within @xmath1 . for @xmath2 the best constraints followed from the cavendish- and etvs - type experiments @xcite and for @xmath3 from van der waals force measurements and from measurements of transition probabilities in exotic atoms @xcite .
therefore new , more exact casimir force measurements are of prime importance in order to obtain stronger constraints on the constants of yukawa - type hypothetical interactions in the intermediate range of @xmath0 .
recently , a new experiment was performed @xcite , in which the casimir force was measured between two metallised surfaces , a plane disc and a spherical lens with a distance @xmath4 between them .
the theoretical value , obtained in @xcite , was confirmed within an accuracy of 5% in the distance region @xmath5 .
note that this was the first direct measurement of the casimir force for metallic surfaces since sparnaay @xcite , and that the accuracy has been increased in 20 times .
the metallic surface of the bodies ( a @xmath6 thick gold layer ) increased considerably the contribution of the hypothetical yukawa forces due to the high density of gold ( the yukawa forces are proportional to the density squared ) . also , the relatively large spacing between the bodies , up to @xmath7 , where the casimir force is very small , turned out to be essential for the strengthening of the constraints . in the present paper
new constraints on parameters of the yukawa - type hypothetical interactions , which follow from the experiment @xcite on casimir force measurement , are obtained .
they turn out to be stronger than those obtained previously from the measurement of the casimir force between dielectrics and from one of the versions of a cavendish - type experiment in the range @xmath8 . especially within @xmath9
, they are stronger than the previously known ones by about @xmath10 times and near @xmath11 , they are @xmath12 times stronger .
the corresponding constraints which result for the masses of the spin - one antigraviton and for the dilaton are also discussed .
in the recent experiment @xcite the casimir force was measured between a plane disc of diameter @xmath13 and thickness @xmath14 and a lens of height @xmath15 , diameter @xmath16 and curvature radius @xmath17 .
the surfaces of the plate and the lens were covered with copper and gold layers of @xmath6 thickness each ( see fig .
1 ) . for this configuration
the casimir force is attractive and given by @xcite @xmath18 this expression was later obtained by another method in ref .
@xcite ( see also refs .
@xcite ) . in fact
it is an approximation but , as shown in @xcite , for a large radius - of - curvature lens , the relative error is less than 0.01% .
the potential of the hypothetical interaction between the atoms of the disc and the atoms in the lens has the form @xmath19 where @xmath20 is the interatomic distance , @xmath21 is the ( dimensionless ) interaction constant , @xmath22 is the range of the interaction and simultaneously the compton wavelength of the hypothetical particles whose exchange leads to the hypothetical forces under discussion .
the coefficients @xmath23 are the numbers of nucleons in the atoms so that the constant @xmath21 is independent of the type of atoms , i.e. , on the nuclear charges . due to its smallness the hypothetical interaction between solid bodies
can be obtained by summing the potential ( [ 2 ] ) over their volumes @xmath24\nonumber \\ & & \times\left [ \tilde{\rho}_{1}\int_{\tilde{v}_{1}}{\rm d^{3}}{\bf r}_{2}+ \rho_{2}\int_{\tilde{v}_{2}}{\rm d^{3}}{\bf r}_{2}+ \rho_{3}\int_{\tilde{v}_{3}}{\rm d^{3}}{\bf r}_{2}\right ] { \exp(-r/\lambda)\over r}\ ; , \label{3}\end{aligned}\ ] ] where @xmath25 .
the vectors @xmath26 and @xmath27 belong to the volumes indicated in the lower integration limits in the operator brackets .
this formula reflects the composition of the disc and the lens , both having metal layers , used in the experiment . in ( [ 3 ] ) , by means of @xmath28 , where @xmath29 is the proton mass , we introduced the densities of the materials used in the experiment : @xmath30 and @xmath31 for the quartz in the disc and the lens , respectively and @xmath32 for the copper and gold layers ( @xmath33 being the _ atomic _ densities ) . using this formula ,
the hypothetical force between the disc and the lens can be expressed as @xmath34 the integrals in these expressions can not be taken explicitely . in general , numerical integration is required ( see below ) . however , if the part of the lens , which contributes most to the result is small compared with the disc , the integrals in ( [ 3 ] ) can easily be taken explicitely .
this approximation is in any case justified for @xmath35 .
in fact , its validity can be extended up to @xmath36 as the numerical calculations show . performing the integration in ( [ 9 ] )
we obtain the final expression for the force @xmath57 between an infinite disc and the lens : @xmath58\nonumber\\ & & \times\big\{2\lambda\rho_3(r-\lambda)-2\lambda(\rho_3-\rho_2 ) ( r-\delta-\lambda){\rm e}^{-\delta/\lambda } \nonumber\\ & & -2\lambda ( \rho_2-\tilde{\rho_1 } ) ( r-2\delta-\lambda){\rm e}^{-2\delta/\lambda } \nonumber\\ & & + \left[2r\lambda\tilde{\rho_1}\left(-1+{\lambda/ r}-{h / \lambda}+{h/ r}+ { h^2/ 2r\lambda}\right ) \right.\nonumber\\ & & \left.+\delta^2(\rho_3 + 3\rho_2 - 4\tilde{\rho_1})- 2r\delta(\rho_2+\rho_3 - 2\tilde{\rho_1})\right]{\rm e}^{-h/\lambda}\big\}\;. \label{15}\end{aligned}\ ] ]
the expression ( [ 1 ] ) for the casimir force acting between a flat plate and a spherical lens was confirmed experimentally in @xcite for a range of distance @xmath59 with a relative error of @xmath60 .
this means that the value of the hypothetical forces can not surpass five percents of the casimir force .
consequently the inequality @xmath61 where @xmath62 is given by eqs .
( [ 3 ] ) , ( [ 4 ] ) or , in the case of sufficiently small @xmath0 by eq . ( [ 15 ] ) , holds . from the inequality ( [ 16 ] ) the constraints on the parameters @xmath21 and @xmath0 of the hypothetical force @xmath62 follow .
we consider this in what follows .
note that for the range @xmath63 the force is well approximated by the simple formula following from ( [ 15 ] ) @xmath64 using ( [ 1 ] ) , ( [ 17 ] ) , the inequality ( [ 16 ] ) may be written in the form @xmath65 it is seen from ( [ 18 ] ) that the strength of the constraint in this range does not depend on @xmath66 , the radius of the lens . for the range
@xmath67 the constraints are obtained using formula ( [ 15 ] ) . for the range
@xmath68 it is impossible to consider the radius of the disc as large and a numerical integration in formula ( [ 3 ] ) was performed . for this purpose algorithm 698 from netlib ( http://www.netlib.org ) was used .
it is an adaptive multidimensional integration , the fortran program is called dcuhre. a large number of function calls ( about @xmath69 ) for @xmath70 was necessary in order to obtain reliable numerical values . for smaller @xmath0 the program does not work , but there the radius of the disc becomes unimportant and formula ( [ 15 ] ) can be used . note that , e.g. , for @xmath71 m the hypothetical force acting between the lens and the finite disc is 72.3% of the force for an infinite disc . for @xmath72
m it is already 99.96% .
the values of the geometrical parameters used in @xcite are given in sec .
the densities of the materials of the disc and of the lens are @xmath73 for the quartz in the lens and @xmath74 for the quartz in the disc , @xmath75 and @xmath76 for copper and gold , correspondingly .
the results of all calculations are collected into inequality ( [ 16 ] ) in order to obtain the new constraints on the constants of hypothetical long - range interactions .
the obtained constraints are shown by curve 1 in fig .
2 , where the region above the curve is ruled out by the cited experimental data .
thereby for every value of @xmath0 the value of @xmath4 was used which gives the best constraint .
this is @xmath77 for large @xmath0 and @xmath78 for small @xmath0 and correspondingly in between . for comparison , in the same fig .
2 the curve 2 shows the previously known @xcite constraints for the yukawa - type hypothetical interactions which follow from the experiment @xcite on casimir force measuring . in that experiment
the lens and the disc ( it was large as compared with the lens ) were made out from quartz only ( without metallic layers on the surfaces ) .
the dependence ( [ 1 ] ) with an appropriate correction factor due to the dielectricity was confirmed with a relative error of @xmath79 for distances @xmath80 . as a result , for extremely small @xmath81 the experiment @xcite gave stronger constraints than the experiment @xcite .
this constraint results from the smallest values @xmath82 .
this becomes clear when taking into account that the function @xmath83 appearing in ( [ 18 ] ) takes its minimal value for @xmath84 .
therefore the experiment @xcite with the smallest value of @xmath4 gives the best results in this case .
for @xmath85 , as can be seen from fig .
2 , the experiment under consideration gives much stronger constraints than that of @xcite .
in the range @xmath86 the new constraints surpass the old ones by more than a factor of 1000 and near @xmath87 , by 3000 . for @xmath88 , where the old casimir experiment has given the same constraints as one of the cavendish - type experiments , the increase is 400 times .
the constraints which follow from the cavendish- and etvs - type experiments are also shown in fig . 2 @xcite . curve 3 follows from the experiment @xcite , curve 4 from @xcite and curve 5 from @xcite , curve 6 from the etvs - type experiment @xcite . as shown in fig .
2 , in the range @xmath89 the new constraints from the casimir effect surpass in some points the results from the cavendish - type experiment in ref . @xcite , in some points they are of the same strength ( compare the curves 1 and 3 in fig .
the new constraints resulting from the recent casimir force measurement ( in distinction from the old ones ) give the possibility to get constraints for the masses of the spin - one antigraviton and the dilaton .
the exchange of spin - one antigravitons gives rise to repulsive forces described by ( [ 1 ] ) with @xmath90 , where @xmath91 is the sum of the masses of the current quarks belonging to a nucleus and of atomic electrons , and g is the gravitational constant .
this gives @xmath92 and from curve 1 of fig .
2 it follows that @xmath93 , or for the spin - one antigraviton mass , @xmath94 . this constraint is only slightly weaker than the previous one ( @xmath95 , @xmath96 , @xcite ) obtained from the cavendish - type experiment of ref .
@xcite ( shown as curve 4 in fig .
a similar constraint follows for the dilaton whose interaction constant is given by theory to be @xmath97 . as seen from fig .
2 , for such a constant @xmath98 or , in terms of the dilaton mass , @xmath99 follows .
quite the same constraint follows from the cavendish - type experiment of ref .
@xcite ( shown as curve 3 in fig .
therefore in the region @xmath100 the constraints following from the casimir effect are approximately of the same strength as those which follow from cavendish - type experiments .
as discussed in the preceeding sections , the new measurement of the casimir force presented in @xcite makes it possibile to strengthen the constraints of the constants of hypothetical long - range interactions by a factor of 1000 and more for a wide interaction range .
moreover , it it is also possible to widen ( by one and a half orders of magnitude ) the boundary of the @xmath0-interval for which the casimir effect gives good constraints for yukawa - type hypothetical interactions . in this way
the casimir effect becomes a strong competitor to the cavendish - type experiments in obtaining constraints for such interactions . to obtain even stronger constraints from the casimir effect it would be desirable to enlarge the distance interval over which the force is measured , perhaps up to @xmath101 . near the left boundary of this interval
, one must take into account the contributions to the casimir force resulting from the uneveness and imperfections of the disc and the lens @xcite .
near the right boundary , the contributions due to the temperature corrections @xcite should be considered too .
also , the use of materials with high density for the whole disc and lens would be helpful .
finally , a further reduction of the error in the force measurement would be important .
the authors are greatly indebted to s.k .
lamoreaux for sending the manuscript of his paper @xcite before publication and for several helpful discussions , in particular for communicating the densities @xmath102 and @xmath103 .
they also thank a.j .
sanders for his interest in this work and for facilitating communication .
one of us ( v.m.m . ) is indebted to the institute for theoretical physics of the university of leipzig , where this work was partly performed , for kind hospitality and to the dfg under grant 436 rus 17/88/94 for support .
99 j. kim , phys .
c * 150 * , 1 ( 1987 ) .
e.g. adelberger et al , annu .
sci . * 41 * , 269 ( 1991 ) .
gillies , metrologia suppl . *
24 * , 1 ( 1987 ) .
e. fischbach et al , metrologia * 29 * , 215 ( 1992 ) .
v.m . mostepanenko and i.yu .
sokolov , phys .
d * 47 * , 2882 ( 1993 ) .
v.m . mostepanenko and n.n .
trunov , _ the casimir effect and its applications_. clarendon press , oxford , 1997 .
kuzmin , i.i .
tkachev and m.e .
shaposhnikov , jetp letters ( usa ) * 36 * , 59 ( 1982 ) .
mostepanenko , i.yu .
sokolov , phys .
a132 * , 313 ( 1988 ) .
m. bordag , v.m .
mostepanenko and i.yu .
sokolov , phys .
a187 * , 35 ( 1994 ) .
lamoreaux , phys .
lett . * 78 * , 5 ( 1997 ) .
derjaguin , i.i .
abrikosova and e.m .
lifshitz , quart . rev . * 10 * , 295 ( 1968 ) .
sparnaay , physica * 24 * , 751 ( 1958 ) .
v.m . mostepanenko and i.yu .
sokolov , sov .
( usa ) * 33 * , 140 ( 1988 ) .
klimchitskaya and yu.v .
pavlov , int .
a11 * , 3723 ( 1996 ) .
v.p . mitrofanov and o.i .
ponomareva , sov .
jetp * 67 * , 1963 ( 1988 ) .
hoskins et al , phys .
d * 32 * , 3084 ( 1985 ) .
chen , a.h .
cook and a.j.f .
metherell , proc .
a394 * , 47 ( 1984 ) .
heckel et al , phys .
lett . * 63 * , 2705 ( 1989 ) .
fig . 1 + fig . 2 | we calculate the constraints on the constants of hypothetical long - range interactions which follow from the recent measurement of the casimir force . a comparison with previous constraints is given .
the new constraints are up to a factor of 3000 stronger in some parameter regions .
= 22.5 cm hep - th/9705101 new constraints on the yukawa - type hypothetical interaction from the recent casimir force measurement m. bordag , g.t .
gillies and v.m . |
noncovalent interactions of organic molecules play a fundamental role in biochemistry , solvation , surface science , and supermolecular chemistry .
an accurate description of the noncovalent interactions is thus very important for many applications @xcite .
actually , the `` golden standard '' for the simulation of noncovalently interacting complexes is the coupled cluster single and double with perturbative triple ( ccsd(t ) ) approach @xcite .
however , the ccsd(t ) method has a high computational cost , scaling as @xmath0 , and it is not easily applicable in general , especially when numerous single point calculations are required as in the determination of a potential energy surface ( pes ) or for geometry optimizations
. in various applications is thus necessary to recover to lower
level computational methods . however , most approaches ( including density functional theory ( dft ) and standard mller
plesset second order perturbation theory ( mp2 ) , are not fully satisfactory for an accurate description of noncovalent interactions and the corresponding pess @xcite . therefore , in the last years different computational schemes , mainly based on variants of mp2 , have been proposed to treat noncovalent interactions with sufficient accuracy and relative small computational effort @xcite .
one of such methods is the so called spin component scaled mp2 ( scs mp2 ) method @xcite which is based on the spin resolved mp2 formula for the correlation energy @xmath1 with @xmath2 } { \epsilon_i+\epsilon_j-\epsilon_a-\epsilon_b},\end{aligned}\ ] ] where @xmath3 denote occupied orbitals , @xmath4 denote virtual orbitals , @xmath5 denotes a two electron integral in the mulliken notation , @xmath6 is the energy of the @xmath7th orbital , and @xmath8 and @xmath9 are the scaling coefficients for the opposite spin ( os ) and same spin ( ss ) correlation , respectively .
this approach was shown to yield quite improved results with respect to the standard mp2 , when a proper choice of the scaling parameters is performed @xcite .
alternatively , a good performance was also obtained by limiting the method to consider only the opposite spin part ( i.e. setting @xmath10 and optimizing only @xmath8 ) , resulting in the scaled opposite
spin mp2 ( sos mp2 ) method @xcite , which allows , when properly implemented , to reduce the computational scaling to @xmath11 . however , the choice of the optimal scaling parameters in scs and sos
mp2 is not trivial and different proposals have been made @xcite , either based on theoretical considerations or on empirical fittings , showing that the `` best '' scaling parameters are somehow system and basis set dependent .
this issue may be not particularly relevant for covalent bonds , where the considered binding energies and improvements over mp2 are much larger that the inaccuracies due to a non
optimal choice of the parameters .
it becomes anyway relevant for noncovalently bonded complexes where the energies to be computed are much smaller .
for these latter cases , very good scs parameters were proposed by distasio and head gordon ( scs(mi)mp2 method ) @xcite by optimizing the scaling coefficients against benchmark noncovalent interaction energy data .
the same authors , as well as grant hill and platts in a separate paper , also showed that accurate results can even be achieved by just same spin scaling ( sss mp2 ; @xmath12 , @xmath131.75 ) @xcite methods .
additional investigations concerned the optimization of the sos
mp2 for long range interaction by the introduction in the sos
mp2 method of a distance
dependent scaling coefficient and a modified distance dependent two electron operator @xcite .
the effort spent to optimize the scaling parameters and the fact that the results of these works are not in an agreement with `` standard '' scs
mp2 ( @xmath14 , @xmath15 ) or sos mp2 ( @xmath16 ) values nor with the values suggested by theoretical arguments within wave function theory @xcite , indicates the difficulty to fix optimal values of the scaling for the proper description of noncovalent interactions .
moreover , all these approaches pay the prize of introducing a high level of empiricism .
therefore , a simple non empirical procedure to fix the value of the various scaling parameters appears highly desirable . in this paper
we address this issue and we show a simple non empirical scheme to fix the scaling factor in calculations of the dissociation curve of noncovalent bonded complexes . here , non empirical denotes the fact that scaling factors will be fixed by direct use of information from high level _ ab initio _ calculations and not from some empirical fitting procedure .
of course , some empiricism is still implied in the use of spin
resolved mp2 formulas .
our approach focuses on scaled mp2 calculations with a single parameter . to this end
, we introduce the sos(r)mp2 , sss(r)mp2 , and scaled mp2 ( s(r)mp2 , i.e. with @xmath17 ) methods . in acronym ( r )
stands for `` calculated from reference non empirical values '' , to avoid confusion and distinguish from empirically scaled versions of standard scs mp2 methods .
our procedure is based on the observation that there exist a well defined proportionality between scaled(same / opposite)mp2 energies and the correlation energy computed by high level methods ( e.g. ccsd(t ) ) , which is almost independent on the intermolecular distance .
thus , the scaling parameter can be fixed once by using the information from only one expensive high level calculations and successively the whole pes can be computed with high accuracy by performing only relatively cheap scaled(same / opposite)mp2 calculations . in particular , in this way the efficiency of the sos
mp2 ( sos(r)mp2 ) method ( @xmath11 scaling @xcite ) can be fully exploited for large scale explorations of dissociation potential energy surfaces ( pes ) without introducing empirical parameters and achieving almost the ccsd(t ) accuracy .
we acknowledge that a similar approach was already recently used in ref .
@xcite , to compute accurate interaction energies of an ar monolayer adsorbed on an mgo substrate .
however , in ref . @xcite
only local mp2 calculations were performed while spin - resolved mp2 calculations , which are the main topic of the present work , were not considered .
in addition , in the present work , a systematic assessment of the non - empirical procedure for fixing the scaling parameters is carried out for different systems and interactions .
the main quantity of interest in the present paper is the inter
fragment correlation ( ifc ) energy , defined as : @xmath18 where in our calculations @xmath19=ccsd(t ) , ccsd , mp2 , sos(r)mp2 , sss(r)mp2 or s(r)mp2 .
the ifc is the difference between the correlation energy ( @xmath20 ) of a complex ( @xmath21 ) and that of its constituting fragments ( a and b ) , computed with a method @xmath19 .
the ifc for two exemplary cases ( h@xmath22o dimer and ne@xmath22 ) is reported in fig .
[ fig_ifc ] as a function of the intermolecular distance ( @xmath23 ) for different methods .
the ifc is always negative and its absolute value rapidly decreases with @xmath23 .
inter fragment correlation ( ifc ) energy of ne@xmath22 and ( h@xmath22o)@xmath22 as computed with different methods . ]
it can be observed that for noncovalent interactions , with good approximation , the ifc energies computed at the mp2 ( or equivalently mp2(ss ) or mp2(os ) ) level and at the ccsd(t ) level are proportional to each other over a wide range of inter fragment distances @xmath23 .
this proportionality can be verified by direct comparison of the ifc energies calculated at different levels of theory for various noncovalent interacting complexes ( see fig . [ fig_ratio ] ) .
ratio between different spin resolved mp2 ifc energies and ccsd(t ) ones , for various systems at several distances .
@xmath24 denotes the equilibrium distance of each complex .
@xmath19 = mp2 [ continuous line ] , mp2(os ) [ dashed line ] , and mp2(ss ) [ dotted line ] . ] for mp2 the proportionality is very well satisfied for all systems and distances .
mp2(ss ) and mp2(os ) show instead some larger deviations at higher distances .
however such deviations are not very relevant , because the ifc energy is rapidly approaching zero for large distances . in fact , as shown in fig .
[ fig_ifc ] , for @xmath251.5 the ifc is already close or below subchemical accuracy .
the proportionality of the ( spin resolved ) mp2 and coupled cluster energies can be rationalized considering that in noncovalent complexes the variation of ifc with distance is mainly due to relaxation of intramolecular contributions , thus both mp2 and ccsd(t ) ifc energies display a very similar behavior with @xmath23 and their ratio stays almost constant .
this proportionality can be used to define single
parameter scaled ( spin resolved ) mp2 approaches suitable for the description of the dissociation of noncovalent complexes . to do that
, we define @xmath26 with @xmath27 where @xmath28 is some reference distance . with this definition we fix ( at the cost of a single expensive ccsd(t ) calculation )
the proportionality between ( spin resolved ) mp2 and ccsd(t ) results , by simply imposing that at some point , the two ifc energies , or analogously , the respective total binding energies ( @xmath29 ) are the same .
the choice of the reference point @xmath28 in eqs .
( [ e9])([e11 ] ) is not a major concern , since the value of @xmath30 is almost independent on it ( see fig .
[ fig_ratio ] and subsection [ sect_dist ] ) . in this work
we have chosen for the the reference point the equilibrium distance @xmath24 , computed at the ccsd(t ) level .
this choice is not mandatory , nevertheless it appears the best compromise between the need to avoid too short distances ( where the proportionality relation may be less accurate ) and the necessity to avoid the use of too small energies ( as would result for large @xmath23 values ) in the ratio of eqs .
( [ e9])([e11 ] ) to minimize numerical noise(see subsection [ sect_dist ] for a further discussion ) .
each of the equations ( [ e9 ] ) , ( [ e10 ] ) , ( [ e11 ] ) provides a simple non empirical procedure for fixing the ( system dependent ) scaling factor in one parameter ( scs)mp2 calculations when dissociation energies of noncovalent systems are of interest .
thus , a whole dissociation curve , comprising many single point results , can be simulated by performing only a single ccsd(t ) calculation and without the introduction of any empirical parameter .
the main question is if the accuracy of the such procedure is high enough . in section [ sec : res ]
we will show that deviations from the ccsd(t ) reference results are indeed well below subchemical accuracy ( i.e. 0.1 kcal / mol ) . to test our scaling approach we considered a representative set of small noncovalently interacting systems : he@xmath22 , ne@xmath22 , he
ne , ar@xmath22 ( dispersion interaction ) , ( h@xmath22o)@xmath22 , ( hf)@xmath22 ( hydrogen bond ) , ( h@xmath22s)@xmath22 , ( hcl)@xmath22 ( dipole dipole interaction ) . additionally , benzene
hcn , the stacked benzene dimer , and be@xmath22 have been also considered , as particular cases .
the former two are in fact relatively large systems with the first one also displaying a mixed electrostatic dispersion character , which changes with the bond distance @xcite .
the latter is a system where mp2 and ccsd completely fail even qualitatively @xcite .
the sizes of the molecules included in our test set were limited by the need to compute in all cases the full ccsd(t ) dissociation curves for reference purposes , except for benzene hcn and the stacked benzene dimer in which case the reference data were taken from ref .
@xcite .
calculations were performed with the aces ii @xcite and turbomole @xcite program packages .
for all systems an aug cc
pvqz basis set @xcite was used , except for ne@xmath22 ( uncontracted aug cc
pvtz ) , be@xmath22 ( cc pv5z ) , and ar@xmath22 ( aug cc
pv5z ) . to assess the performance of the different methods to reproduce the full dissociation curves we considered the mean absolute error : @xmath31 where @xmath19=ccsd , mp2 , sos(r)mp2 , ... , and @xmath32 note that because all the methods considered here use the hartree fock exchange , the difference between the ifc energies of different methods in eq .
( [ edelta ] ) are exactly the same as the differences in the total interaction energies .
the mae@xmath33 defined in eq .
( [ emae ] ) measures the average deviation of the results from the reference ccsd(t ) data over a given interval . in our work
we decided to remove from this evaluation the smallest values of @xmath23 , because our approach might be not fully justified at such small inter fragment distances .
hence , the lower bound for the integration was fixed at @xmath24 .
the upper bound of the integral was fixed instead to the value of @xmath23 beyond which the ccsd(t ) ifc energy is lower than 10@xmath34 hartree , in order to keep the normalization factor finite in eq .
( [ emae ] ) .
finally , we note that in all calculations no correction for the basis set superposition error ( bsse ) was included , for computational simplicity and because it is readsorbed in the scaling coefficient . for the water dimer ,
e.g. , we check that , including a bsse correction , the mae of eq .
( [ emae ] ) is 36.7 , 36.6 , 5.5 @xmath35 kcal / mol , for sos(r)-mp2 , sss(r)-mp2 and s(r)-mp2 respectively .
these values are in good agreement with the ones ( without bsse ) reported in tab .
the values of the @xmath36 , @xmath8 , and @xmath9 coefficients , obtained by applying eqs . ( [ e9 ] ) , ( [ e10 ] ) , and ( [ e11 ] ) to the systems considered in this work , are reported in table [ tablecos ] , together with the reference inter fragment separation @xmath24 .
.[tablecos ] values of the @xmath8 , @xmath9 , and @xmath36 coefficients obtained via eqs .
( [ e9 ] ) , ( [ e10 ] ) , ( [ e11 ] ) calculated at distance @xmath24 for all the systems presented in this paper . [ cols="<,^,^,^,^,^,^,^",options="header " , ] the tabulated results show that scaling coefficients change very little , as can be already inferred from fig .
[ fig_ratio ] .
this provides a validation of the working hypothesis and indicates the robustness of our simple non empirical scaling procedure . by inspecting the relative maes reported in tab .
[ tab3 ] we see that the accuracy of the methods is not only preserved but in many cases is also improved .
the improvement is anyway rather small and can be considered to lay within the numerical uncertainty related to the definition of the mae ( eq .
( [ emae ] ) ) .
this result is important because it remarks the possibility to utilize the present non
empirical scaling procedure also in all those cases where the equilibrium distance is not known , by simply performing a single ccsd(t ) calculations at any `` reasonable '' bond distance . in this subsection
we consider the application of our new scaling method to larger systems , namely the benzene
hcn t shaped complex and the stacked benzene dimer .
the first of these systems is particularly interesting because the interaction has a mixed character involving both electrostatic and dispersion contributions in a variable proportion along the dissociation curve @xcite .
its dissociation curves computed at the ( spin resolved ) mp2 level as well as the corresponding scaled results are reported in fig .
[ fig_benz ] .
reference ccsd(t ) data from ref .
@xcite are also reported . binding energy for the benzene
hcn complex as a function of the inter atomic distance @xmath23 , for different methods . ] the scaling coefficients have been computed at the equilibrium distance of 2.34 and are @xmath37 , @xmath38 , and @xmath39 .
consistently with the results discussed in the main body of section [ sec : res ] , the non empirical scaling methods all perform remarkably well , substantially improving over `` bare '' mp2 , mp2(os ) , and mp2(ss ) calculations and especially over empirically scaling methods ( e.g. scs(mi)mp2 ) .
s(r)mp2 , sos(r)mp2 and sss(r)mp2 show in fact deviations from the reference data below 0.5 kcal / mol along the whole range of distances considered .
similar results are found also by using scaling coefficients obtained considering a different reference distance @xmath28 . for @xmath40 ,
indeed , we found @xmath8=1.15 , @xmath9=1.14 , @xmath41=0.57 , and essentially the same performance for the binding energy ( or equivalently the ifc energy ) .
in fact , the scaling coefficients computed at the new distance are in good agreement , although slightly smaller , with the ones computed at the equilibrium distance .
we remark that the small difference between the coefficients computed at different reference distances shall , in this special case , be partially traced back to the computational noise originating from the fact that the reference ccsd(t ) values were not computed with the same set up as the mp2 ones , but rather extracted from the total binding energies ( hf+ccsd(t ) at the cbs limit ) reported in ref . @xcite .
despite this small issue , the non empirically scaled methods display an impressive robustness . as a further test we consider a typical @xmath42-stacking complex . in fig .
[ benzdimer ] we thus report the dissociation curve of the stacked benzene dimer as computed at the ( spin resolved ) mp2 level and with the various scaled approaches .
reference ccsd(t ) data from ref .
@xcite are also reported .
the scaling coefficients have been computed at the equilibrium distance of 3.765 and are @xmath43 , @xmath44 , and @xmath45 . binding energy for the stacked benzene dimer as a function of the inter
atomic distance @xmath23 , for different methods . ]
inspection of the figure shows that , similarly with the case of the benzene - hcn complex , all the scaled methods perform remarkably well yielding deviations within 0.2 kcal / mol from the reference ccsd(t ) data .
in particular , the non
empirical scaling procedure proposed here demonstrates to be able to correct well both the overbinding of the mp2 method and the strong underbinding of the mp2(os ) and mp2(ss ) approaches .
overall , the dissociation curves computed with our scaling approach agree very well with reference data and show that the method is very effective even for such `` large '' and `` difficult '' systems as the present ones .
we note that this performance is remarkable especially for sos(r) and sss(r)mp2 , because the mp2(os ) and mp2(ss ) binding energies are almost degenerate , so that the corresponding scaling coefficient are almost identical .
thus , the same scaling is obtained for the two methods . however , this has only a minor effect on the sss(r)mp2 results at relatively small distances , while a good performance is observed for bond distances larger than the equilibrium one .
this shows once more that the here presented scaling procedure is rather robust in many different situations .
finally , we consider in fig .
[ fig_be2 ] the beryllium dimer , which is a very difficult system where mp2 and ccsd fail to give even an approximate quantitative description of the dissociation curve @xcite . for this system
even ccsd(t ) shows some limitations @xcite .
however , this high - level method can at least capture most of the features of the dissociation curve .
therefore , it will be used anyway as reference in the present example , although caution must be paid to the quantitative analysis of the results . in this case , the computed scaling coefficients are @xmath46 , @xmath47 , and @xmath48 . however , an accurate description of the dissociation curve is found only using the sos(r)mp2 method . on the contrary ,
methods based on the same spin correlation fail quite evidently
. this may be due to the fact that in this system , where static correlation is important , the stretching of the bond is not only promoting polarization and induction phenomena in each atom , but is also changing the actual `` multireference '' description of the system , influencing the same spin correlation which includes antisymmetrized integrals .
this fact indicates that the use of mp2(ss ) correlation energy in scaling procedures has strong limitations for this peculiar system . due to this limitations
also the s(r)mp2 method , which includes important contributions of same spin correlation , can not be accurate .
nevertheless , we remark that s(r)mp2 and even sss(r)mp2 can compensate the limitations of the underlying mp2(ss ) energy relatively well , thanks to appropriate values of the scaling factor granted by eqs .
( [ e9 ] ) , ( [ e10 ] ) , ( [ e11 ] ) .
however , the independence of the scaling factor from the reference separation is partially lost and only a limited portion of the dissociation curve can be properly described when same spin correlation is considered .
the sos(r)mp2 method with coefficients calculated in our scaled procedure provides for the be@xmath22 the best dissociation curve ( as compared to ccsd(t ) ) significantly outperforming mp2 , ccsd or sos(mi)mp2 methods . binding energy and ifc energy for be@xmath22 as a function of the inter
atomic distance @xmath23 , for different methods .
@xmath24 denotes the equilibrium distance . ]
we have proposed a simple non empirical procedure which can be used to calculate optimal scaling coefficients for ( spin resolved ) mp2 calculations of the dissociation of noncovalent complexes .
the applicability of the proposed method was demonstrated for a series of test systems , considering equilibrium and non equilibrium reference distances , as well as for a notoriously difficult case as be@xmath49 .
the obtained results show that the presented method works well in most cases , confirming the robustness of our hypothesis .
the proposed method is especially attractive because it allows to obtain a full dissociation curve of almost ccsd(t ) quality at the cost of just a single ccsd(t ) calculation .
moreover , it is conceivable to replace the costly ccsd(t ) calculation with a focal point analysis @xcite ( the @xmath50ccsd(t ) procedure ) to obtain an estimate of the ccsd(t ) correlation energy from a cheaper calculation . in this way the computational cost is much reduced , losing only little on accuracy .
thus , future applications on large systems can be foreseen .
moreover , it might be also possible to replace the ccsd(t ) reference with an alternative accurate treatment of the correlation , e.g. the ccsd[t ] @xcite or the fno ccsd(t ) @xcite method , to optimize the accuracy or the computational effort .
finally we remark that , although in this paper we focused only on single parameter scaled mp2 methods , the methodology presented here can be extended in the future to treat the more general case of spin component scaled ( scs ) mp2 calculations with two parameters ( @xmath8 and @xmath9 ) . to this end eqs .
( [ e6])-([e8 ] ) will be generalized to @xmath51 and an additional constraint will be needed for the scaling parameters ( e.g. fixing the sum or the ratio of the parameters from theoretical considerations ) .
we also have to stress that this approach can be useful for the verification of the quality of scs coefficients used in different scs mp2 methods .
* acknowledgments * this work was supported by the polish committee for scientific research mnisw under grant no .
n n204 560839 and by the erc
stg fp7 project dedom ( no . 207441 ) .
we thank turbomole gmbh for providing the turbomole program package and m. margarito for technical support . | we present a simple and non empirical method to determine optimal scaling coefficients , within the ( spin component)scaled mp2 approach , for calculating intermolecular potential energies of noncovalently interacting systems .
the method is based on an observed proportionality between ( spin component ) mp2 and ccsd(t ) energies for a wide range of intermolecular distances and allows to compute with high accuracy a large portion of the dissociation curve at the cost of a single ccsd(t ) calculation .
the accuracy of the present procedure is assessed for a series of noncovalently interacting test systems : the obtained results reproduce ccsd(t ) quality in all cases and definitely outperform conventional mp2 , ccsd and scs
mp2 results .
the difficult case of the beryllium dimer is also considered . |
the objective of the escape dynamics out of a metastable state has been thoroughly investigated since the seminal work of kramers @xcite .
a quantity of primary interest to establish the time scale of the escape dynamics of a classical brownian particle in the presence of a potential barrier is the mean first passage time ( mfpt ) ; i.e. the average time it takes for a particle driven by noise to reach a target position beyond an intervening barrier top @xcite .
the topic of evaluating the mfpt in the presence of external modulations of either stochastic or also deterministic nature has ample applications , among others , in neuronal models which are characterized by a time - varying voltage threshold ; e.g. see refs .
@xcite . + a minimum occurring in the mfpt versus increasing frequency scale of the modulation is known in the literature as resonant activation .
the phenomenon may emerge when the time scale of the barrier modulation matches the characteristic time scale of the escape dynamics .
the phenomenon was originally predicted for a confining potential composed of a stylized piecewise linear , fluctuating barrier in ref .
@xcite . soon after , the objective for the corresponding reaction rate dynamics in presence of general modulations of a metastable potential landscape has been investigated with a pioneering work in ref .
@xcite ; cf .
also the surveys on escape over fluctuating barriers @xcite and investigations in ref .
a similar phenomenon occurs if periodically varying modulations are acting : resonant activation emerges then due to the interplay between the nonstationary , deterministic barrier modulation and ambient environmental noise @xcite .
+ resonant activation constitutes therefore an archetypical feature for escape under deterministic modulations or fluctuations of a potential barrier .
the general features of the mfpt as a function of the characteristic modulation frequency scale are a saturation to a maximal value for very slow modulations , where the highest barrier configuration dominates the barrier passage , followed by a decreasing behavior towards an intermediate nonadiabatic minimum the resonant activation minimum , and then by an increase towards a limiting high - frequency behavior , as determined by the corresponding averaged potential configuration @xcite .
+ with the present work we investigate the phenomenon of resonant activation for the archetype of the widely studied quantum dissipative two - state system ( tss ) @xcite , here driven by dichotomous noise and/or by a deterministic coherent field .
this setup allows for a detailed investigation of the regime in which the barrier is not surmounted thermally but instead rather passed by dissipative quantum tunneling connecting the left- and right - well states of a double - well exhibiting a lowest energy doublet of energy separation @xmath1 . throughout the following
we assume that this lowest doublet is well separated from higher lying quantum energy levels ; put differently , the presence of applied modulation is assumed not to excite appreciably higher lying quantum energies .
+ the use of externally applied dichotomous noise modulations of the tunneling amplitude allows for an exact averaging over the noise realizations .
this in turn yields a generalized non - markovian master equation that has been invoked previously to investigate the effects of correlated noise in electron transfer problems and tight - binding models @xcite .
+ given a quantum context , the concept of the mfpt generally presents , however , an unsolved subtle issue @xcite .
this has its origin in quantum mechanics , where position and momentum knowingly can not be sharply defined simultaneously .
this issue is overcome , however , when investigating the incoherent tunneling regime . in the latter limit ,
being realized by coupling the tss strongly to an environment , one is able to describe the tunneling dynamics in terms of a generally non - markovian quantum master equation for the left / right state probabilities with well - defined quantum transition rates .
the resulting treatment then mimics ( in its markovian limit ) a classical discrete process @xcite .
nevertheless , the dynamics is governed by the quantum tunneling mechanism ; the effect of the environment being a renormalization of the bare tunneling amplitude . in such a situation a sensible statement of the problem of an absorbing boundary state in presence of generally time - dependent driving is feasible @xcite . + in practical terms , the population dynamics in the tss may be realized when the passage of the system to the target state triggers a sudden modification of the potential , rendering the target state an absorbing state .
an alternative setting , with the system composed of a particle coupled to a detector and a bath continuum of electromagnetic or phonon modes , has been discussed in ref .
strong dissipation is experimentally also attained for tss systems characterized by macroscopic quantum tunneling such as occurring in superconducting flux qubits @xcite .
+ the strategy adopted in this study is based on the use of the corresponding , periodically modulated master equation approach to the time - dependent escape - dynamics @xcite . in this master equation for the modulated tss we introduce the appropriate boundary conditions for reflection and absorption .
particularly , we consider the case with the particle initially prepared in the ( reflecting ) left - well state and set as an absorbing state the neighboring right - well state . as a result
, we end up with an equation for the so adjusted decay of the survival population @xmath2 of the left metastable state in terms of explicit time - dependent rates .
the negative of the rate of change of @xmath2 then defines the first passage time density , correctly obeying the boundary conditions at all times , whose first moment yields the searched for mfpt .
+ in the case of dichotomous noise driving , it becomes possible to solve analytically the resulting master equation for the noise - averaged left - well population , at least in the case in which no time - periodic modulations are present .
the latter situations does require a first passage analysis with arbitrary time - dependent transition rates entering the corresponding quantum master equation ; a situation that can be treated by numerical means only .
+ our approach extends the amply studied case of the dissipative quantum dynamics for a tss to the situation with a quantum resonant activated escape regime .
our findings display similar features as those observed for the mean residence time statistics occurring in modulated classical double - well systems where the barrier is thermally surmounted @xcite .
the obtained results therefore corroborate our expectation that the general phenomenon of resonant activation occurs likewise in the deep incoherent quantum regime .
note also that this used setup distinctly differs from the stochastic schrdinger equation approach @xcite and , alike , the approximated semiclassical approach @xcite .
indeed , the study of quantum resonant activation for a tss involves a dependence on a wide set of parameters ; it involves , besides the various driving parameters , also the strength and type of quantum dissipation , the temperature , and a suitably chosen dissipative high - frequency cut - off @xcite .
this latter value is system - specific as it depends on the type of physics addressed with such a quantum tss @xcite .
as a model of driven dissipative quantum dynamics being composed of two metastable quantum states , we consider the archetypical spin - boson model @xcite in which a quantum two - state system @xmath3 is coupled to a heat bath @xmath4 , with the latter made up of independent bosonic modes of frequencies @xmath5 ( caldeira - leggett model ) @xcite . the coupling to the bath occurs via a scaled position operator which , in the localized basis @xmath6 of a truncated double - well system , is represented by @xmath7 .
the total hamiltonian of system coupled to the heat bath reads @xcite @xmath8 \nonumber\\ & -&\frac{\hbar}{2}\sigma_z\sum_i c_i ( a_i^{\dag}+a_i)+\sum_i\hbar\omega_i a_i^{\dag}a_i\;,\end{aligned}\ ] ] where @xmath9 denotes the tss tunneling matrix element , modulated around its bare value @xmath10 , and @xmath11 stands for a modulated bias energy of vanishing average . here , @xmath12 .
in the following we will consider both , deterministic and stochastic modulations of the tunneling amplitude @xmath9 and a periodically driven bias @xmath11 of the form @xmath13 , wherein @xmath14 denotes an initially chosen phase offset .
+ the bosonic environment , with creation and annihilation operators @xmath15 and @xmath16 , interacts with the tss system via the set of coupling constants @xmath17 .
this system - bath interaction is fully characterized by the spectral density function @xmath18 , whose continuum limit we assume to be of ohmic form @xcite ; i.e. , @xmath19 where @xmath20 characterizes the dimensionless , dissipative coupling strength and @xmath21 marks the suitably chosen high - frequency cutoff . assuming a factorizing initial preparation with a total density operator for the combined setup of the tss interacting with the thermal bath , i.e. , @xmath22 ( with the bath being prepared in the thermal state at temperature @xmath23 ) , the exact dynamics of the tss can be cast into the form of a generalized master equation ( gme ) for the population difference @xmath24 . here , the population @xmath25 is the probability to find the system in the localized state @xmath26 ( @xmath27 ) .
the resulting non - markovian gme assumes the form @xcite , @xmath28 \;,\end{aligned}\ ] ] being formally valid for any coupling and temperature regime , spectral density function , and time dependence of the modulation . within the non - interacting blip approximation ( niba ) , which is valid for strong coupling and not too low temperatures ,
these kernels @xmath29 take on the explicit expressions @xcite : @xmath30\cos[\zeta(t , t ' ) ] \nonumber\\ k^{a}(t , t')&=&\delta(t)\delta(t ' ) e^{-q'(t - t')}\sin[q''(t - t')]\sin[\zeta(t , t ' ) ] \nonumber,\\ & \quad&\end{aligned}\ ] ] where the function @xmath31 is defined by @xmath32 the kernels @xmath33 and @xmath34 in eq .
( [ k ] ) are symmetric and antisymmetric , respectively , under the change @xmath35 .
this implies that for the static unbiased case we obtain @xmath36 .
+ the functions @xmath37 and @xmath38 in eq .
( [ k ] ) denote the real and imaginary part of the thermal bath correlation function , respectively @xcite . for the chosen ohmic spectral density in eq .
( [ g ] ) and in the so - called scaling limit ( @xmath39 ) , @xmath40 reads @xcite @xmath41\nonumber\\ & & + i2\alpha\arctan(\omega_{c}t),\end{aligned}\ ] ] where @xmath42 . in the incoherent tunneling regime , occurring at finite temperatures and _ strong _ coupling ( i.e. , @xmath43 for the symmetric tss ) @xcite , the nondriven dynamics of the population difference is well approximated by the markovian limit to eq .
( [ gme ] ) with time - independent transition rates .
this is so because the memory time of the kernels in eq .
( [ k ] ) constitutes the smallest time scale . in the driven case
, using the definition of @xmath44 and the conservation of total probability , i.e. , @xmath45 , the following quantum master equation with time - dependent forward ( @xmath46 ) and backward ( @xmath47 ) quantum transition rates is then valid for general modulations of @xmath9 and @xmath11 @xcite ; reading @xmath48 where @xmath49\end{aligned}\ ] ] are the quantum transition rates from the left to the right well ( forward ) and vice versa ( backward ) . in this markovian limit
, the rates generally vary in time @xmath50 , but are independent of the population themselves . these transition rates incorporate implicitly both the quantum dissipation and the shape of the double - well potential and depend as well locally only on the externally applied modulation .
+ we note that setting the upper integration limit to infinity in eq .
( [ w - pm ] ) constitutes a further approximation within this type of markovian limit .
however , limiting the integration to the latest acting physical time @xmath51 , yields an improvement which allows us to analyze the short time behavior of the pdf .
the resulting improved markovian quantum transition rates read @xmath52\;.\end{aligned}\ ] ] the upper integration limit at @xmath53 in eq .
( [ w - pm ] ) yields time - independent rates in the absence of deterministic driving .
this modification is of relevance for the regime of very short passage times only . in practice , using @xmath53 as the upper integration limit yields , however , indistinguishable numerical results away from the very short time regime . in the main panels of figs . [ fig2 ] , [ fig6 ] , [ fig9 ] , and [ fig11 ] , shown in the result section @xmath54 below ,
we consistently have used the improved rate expression ( [ w - pm - exact ] ) .
we confirmed numerically that noticeable differences occur only at small times @xmath50 .
this short - time regime is shown in the insets .
the quantum master equation eq .
( [ gmeplr ] ) can be understood as describing a discrete stochastic process , randomly switching between two _ reflecting _ states ; meaning that the rates to go leftward at the left state and rightward at the opposite , right - placed state are both vanishing @xcite . to perform a first passage time analysis
, we consider the situation in which the particle is initially prepared at time @xmath55 in the left quantum state ( @xmath56 ) .
we next calculate the passage time statistics to become detected ( absorbed ) at the right state ( @xmath57 ) while the left state is kept reflecting @xcite .
this requirement is implemented upon introducing an absorbing boundary conditions at the state @xmath57 and reflecting boundary condition at the state @xmath56 .
given these two generally time - dependent `` birth and death '' quantum transition rates this amounts to setting for all times @xmath58 @xcite @xmath59 in eq .
( [ gmeplr ] ) ( see fig . [ fig1 ] ) .
moreover , given the initial condition that @xmath60 , the left well population @xmath2 must be interpreted as a conditional survival probability @xmath61 .
this conditional survival probability in the left state , with @xmath57 being an absorbing state , is thus governed by @xmath62 with the initial condition @xmath60 and the forward rate @xmath63 detailed with eq .
( [ w - pm ] ) .
note that this so obtained survival conditional probability distinctly differs from the usual conditional probability of the two - state discrete master equation dynamics in eq .
( [ gmeplr ] ) .
+ just alike in a classical situation @xcite , the negative rate of change of this so tailored conditional passage time probability to find the particle still in state @xmath56 yields the first - passage time probability density function ( pdf ) .
specifically , the first passage time ( fpt ) pdf is therefore given by @xmath64 with @xmath65 determined from eq .
( [ eqpl ] ) . with positive - valued forward rates and starting out at @xmath66
we have , with absorption occurring at state @xmath57 , that @xmath67 .
this fpt pdf @xmath68 in eq .
( [ g ] ) yields the manifest first passage time pdf , satisfying @xmath69 and with proper unit normalization , i.e. , @xmath70 .
moreover , using the improved expression ( [ w - pm - exact ] ) for the rate , the first passage time pdf @xmath68 then starts out physically consistently at @xmath71 .
+ the mfpt to the state @xmath57 of the tss can be obtained in commonly known way @xcite , namely as the first moment @xmath72 of the fpt pdf @xmath68 in eq .
( [ g ] ) ; reading @xmath73 up to here , the theory has been general in regard to the choice for the shape of the temporal modulation . here and in the following sections we specify the various specific forms of modulations used in evaluating both the fpt pdf @xmath68 and its first mean , the mfpt @xmath72 .
+ we start with the case where one of the two parameters of the tss ( either the bare tunneling matrix element @xmath10 or the bias ) is periodically modulated in time , while the other is held fixed . to be specific ,
consider the following two forms of periodically driven settings : @xmath74 for a vanishing amplitude of the driving on the tunneling matrix element , i.e. , @xmath75 in i ) , and as well for the bias , i.e. , @xmath76 in ii ) , the static case with @xmath77 is recovered , wherein @xmath78\cos[q''(\tau)]$ ] .
+ for both the driving settings , the fpt pdf depends explicitly on the initial driving phase @xmath14 .
consequently , the mfpt is evaluated as an average over a uniform distribution of this initial driving phase , yielding @xmath79 with @xmath80 , and rhs evaluated by eq .
( [ eqpl ] ) in terms of the phase - dependent quantum transition rate @xmath63 . next ,
consider a situation in which the system is driven with a deterministic modulation of the bias ; i.e. , @xmath81 and the tunneling amplitude is driven by stationary , exponentially correlated dichotomous noise ( also known as telegraphic noise ) of vanishing average around its bare value @xmath10 .
explicitly , we set @xmath82 where @xcite @xmath83 with @xmath84 a poissonian counting process with parameter @xmath85 , yielding that @xmath86 . here , the amplitude @xmath87 is a two - state random variable @xmath88 which is evenly distributed , i.e. , @xmath89 $ ] , thus yielding a vanishing average , while the poisson parameter @xmath85 determines the noise correlation of the two - state dichotomous process @xmath90 , i.e. , @xmath91 where the subscript @xmath92 stands for average over the noise realizations .
+ in the extreme limit @xmath93 and @xmath94 this two - state noise approaches white gaussian noise of vanishing mean @xcite . keeping the noise amplitude fixed
, however , the intensity of this noise vanishes identically with @xmath95 , as can be seen by writing the noise correlation function as @xmath96 $ ] , where the term inside the square brackets approaches a dirac delta - function whereas its strength ( i.e. the prefactor ) vanishes .
this accounts for the behavior observed in sec .
[ fluctuating - delta ] below , where the limit @xmath95 of dichotomous fluctuations indeed coincides with the mfpt for the noiseless case .
+ dichotomous noise allows for an exact averaging over the noise realizations of the dynamics given by eq .
( [ eqpl ] ) .
as detailed in appendix [ derivation ] , the noise - averaged population @xmath97 is obtained by solving the set of equations in which @xmath97 is coupled to the correlation expression @xmath98 .
the rate of change for @xmath97 is then given by @xmath99 y(t).\qquad\end{aligned}\ ] ] the value of @xmath100 at @xmath101 gives the initial correlation between the position of the particle and the state of the noise @xmath102 . in what follows
we assume uncorrelated initial condition , i.e. , eq .
( [ eqplfluct ] ) is solved with initial conditions @xmath103 and @xmath104 . the rates appearing in eq .
( [ eqplfluct ] ) are obtained as @xmath105 \nonumber\\ & & \end{aligned}\ ] ] with @xmath106 and where @xmath107 because the noise amplitude appears as the prefactor in the function @xmath108 , it follows readily that , for vanishing noise driving @xmath109 , @xmath110 .
the averaged probabilities then decouple from @xmath100 . in this case
the first line of eq .
( [ eqplfluct ] ) reduces to an equation formally identical to eq .
( [ eqpl ] ) with @xmath111 .
also note that , as stated before for the rates ( [ w - pm ] ) , here too the time - dependent rates must be properly defined with the upper integration limit set at @xmath50 , i.e. , @xmath112\nonumber\\ & & \end{aligned}\ ] ] and again i=0,1 .
+ the mfpt is calculated by using the following fpt pdf averaged over the two state noise realizations and also over the deterministic initial driving phase of the driving , yielding @xmath113\rangle_{\eta},\end{aligned}\ ] ] where @xmath114\rangle_{\eta}=-\langle \dot{p}_l(t;\phi)\rangle_{\eta}$ ] , with @xmath115 given by eq .
( [ eqplfluct ] ) with phase - dependent rates .
this section reports the findings for the resonant activation occurring in an incoherent spin - boson system with modulated tunneling matrix element and/or oscillating bias .
specifically , we consider for the tunneling matrix element @xmath9 separate modulations : either an unbiased two - state noise @xmath116 or a deterministic periodical driving .
only afterwards we consider a more general case for which the tunneling matrix element is fluctuating while an unbiased oscillating field drives the bias @xmath11 . throughout the remaining parts
all quantities are scaled in terms of the bare tunneling frequency @xmath10 ; i.e. , * frequencies @xmath117 and noise switching rates @xmath85 are given in units of @xmath10 .
time @xmath50 is measured in units of @xmath118 * temperature is measured in units of @xmath119 , with @xmath120 denoting the boltzmann constant . as can be deduced from hamiltonian ( [ h ] ) , the noise and the driving amplitudes are given in terms of frequencies .
moreover , in what follows cutoff frequency and temperature are held fixed , assuming the values @xmath121 and @xmath122 . finally , with the exception of the results in figs . [ fig5 ] and [ fig8 ] , the dimensionless coupling strength is always set to the value @xmath123 .
this situation with stationary telegraphic noise modulations can be treated analytically .
we first address the setting described in sec .
[ noise - av - me ] in this analytically solvable case in which the bias is vanishing , i.e. , @xmath76 , and the tunneling matrix element fluctuates around its static value @xmath10 , namely @xmath124 the noise @xmath125 denotes the markovian two - state noise of vanishing average as discussed in sec .
[ noise - av - me ] .
this setting corresponds to a tss - version of the classical brownian particle in a piecewise linear fluctuating potential considered by doering and gadoua in ref .
@xcite , but here the dynamics is governed by quantum tunneling transition rates rather than by classical ( arrhenius - like ) over - barrier escape rates .
+ in this case eq . ( [ eqplfluct ] ) for the noise - averaged population reads explicitly @xmath126 where the time - independent transition rates are given by eq .
( [ w - pm - i ] ) with @xmath127 . using eqs .
( [ w - pm - i ] ) and ( [ s ] ) , the rates @xmath128 are explicitly given by @xmath129 where @xmath130\end{aligned}\ ] ] and @xmath131 .
+ due to the fact that the transition rates are time - independent , eq . ( [ me_noise_delta ] ) can be solved analytically with the boundary conditions @xmath132 and @xmath133 .
the solution of eq .
( [ me_noise_delta ] ) for the noise - averaged population of the left state is [ solution ] @xmath134 where @xmath135 and further @xmath136 note that for vanishing noise amplitude , i.e. , @xmath109 , the rate @xmath137 vanishes identically . in this latter case eq .
( [ solution ] ) renders a strictly single - exponential decay with @xmath138 .
+ + the fpt pdf , i.e. , @xmath139\rangle_{\eta}\nonumber\\ & = & -\langle \dot{p}_l(t)\rangle_{\eta}\end{aligned}\ ] ] assumes the form of a bi - exponential decay , as it follows from taking the time derivative of the solution in eq .
( [ solution ] ) . in fig .
[ fig2 ] we depict @xmath68 for two values of the poisson parameter @xmath85 . for the lower , adiabatic rate ( red solid line )
g(t ) overrides the corresponding nonadiabatic curve ( blue dashed line ) at long times , having a larger tail .
this gives rise to a mfpt @xmath72 whose value , in the adiabatic case , exceeds the one assumed in the intermediate ( nonadiabatic ) regime , as can be seen in fig .
[ fig3 ] . in fig .
[ fig2 ] we plot the numerically evaluated pdf @xmath68 , using the time - dependent rates given by eq .
( [ w - pm - i - exact ] ) with @xmath127 .
the inset in fig .
[ fig2 ] shows that evaluations for @xmath68 , using either the time - dependent expression or the time - independent form ( [ w - pm - i ] ) , yield results that differ at very short times only , but otherwise become indistinguishable ! for this reason the mfpt evaluated alternatively with those time - independent transition rates provides an excellent approximation .
this feature also holds true alike for the fpt pdfs calculated in subsequent sections .
+ the mfpt @xmath72 as a function of the poisson rate @xmath85 can be calculated analytically by using the solution ( [ solution ] ) in eq .
( [ g - noise ] ) and the definition for @xmath72 given in eq .
( [ mfpt ] ) .
we find that @xmath140 where the dependence on the poisson parameter @xmath85 is made explicit . from this analytic result
three important limits can be investigated using eqs .
( [ rates2 ] ) and ( [ a ] ) .
first , the static case is recovered upon setting @xmath109 , yielding @xmath141 this same value is assumed for the mfpt in the extreme large noise switching limit @xmath142 ; i.e. , @xmath143 finally , in the adiabatic limit @xmath144 the mfpt emerges as @xmath145 in fig .
[ fig3 ] the mfpt @xmath72 , evaluated according to eq .
( [ mfpt - analytical ] ) , is depicted as a function of the poisson rate @xmath85 for different values of the noise amplitude @xmath87 .
the curves display adiabatic , low switching rate maxima whose values , for different values of @xmath87 , approach the analytical limit ( [ t1-low - rate ] ) .
the resonantly activated regime occurs at intermediate noise switching time scales . as described by eqs .
( [ t1-static ] ) and ( [ t1-high - rate ] ) , at large noise switching rates @xmath85 the mfpt converges to the results for the _ average _ configuration which , in our case , coincides with the unmodulated , static case .
these general features are shared with the predictions obtained in refs .
@xcite when using a classical brownian motion escape dynamics .
+ fig .
[ fig3 ] also depicts yet another intriguing feature : the different curves seemingly cross exactly the horizontal line ( static case ) at a switching rate which surprisingly depends very weakly on the noise amplitude @xmath87 .
interestingly , a similar behavior has also been observed numerically in ref .
@xcite for classical brownian particle dwelling a piecewise linear fluctuating barrier and in experiments @xcite . +
our detailed analytical evaluations of the mfpt for a wider range of amplitudes indicate that this _ crossing point _ for entering the resonant activation regime in the nonadiabatic regime is in fact mathematically _ not _ exact ( see the filled circles in fig .
[ fig4 ] below ) .
this near - exact nonadiabatic crossing switching frequency @xmath0 , at which the mfpt crosses with the static value , cf .
[ fig3 ] , is determined by the transcendental solution to the equation @xmath146 this relation results from equating the analytical expression of the mfpt in eq .
( [ mfpt - analytical ] ) with the static value eq .
( [ t1-static ] ) as given by the dotted line in fig .
+ the relation in eq .
( [ equation ] ) can be solved approximately in analytical terms by assuming that the rates @xmath128 are nearly _ independent _ of the poisson rate @xmath85 , when restricted to a narrow regime around the _ a posteriori _ chosen numerical value @xmath147 .
put differently , we substitute @xmath148 with the value @xmath149 in eq .
( [ rates2 ] ) ( with the present choice of parameters , numerical values for the coefficients are @xmath150 and @xmath151 ) .
the solution of eq .
( [ equation ] ) for the crossing point then reads @xmath152 this shows that the leading contribution to @xmath0 is quadratic in the amplitude @xmath87 .
this explains that for @xmath153 the crossing point depends only weakly on this noise amplitude .
the analytic crossing rate @xmath0 obtained from eq .
( [ crossing ] ) as a function of @xmath87 is shown in fig .
[ fig4 ] as the solid ( violet ) line .
note the excellent quantitative agreement between this approximate evaluation of the crossing rate @xmath0 and the numerically precise evaluation at selected noise amplitudes , as marked by the filled circles .
+ to provide a deeper insight for the interplay between the characteristic time scale of the dynamics , essentially dictated by @xmath20 , and that of the noise dynamics as encoded with the poisson parameter @xmath85 , we show in fig .
[ fig5 ] a comparison among the mfpt results _ vs. _
@xmath85 at different values of the dissipation strength @xmath20 .
the crossing frequency @xmath0 assumes a lower value ( corresponding to larger noise correlation time ) for stronger bath coupling , where the bare tunneling amplitude becomes dissipation - renormalized towards a lower value @xcite , meaning that the tunneling passage to the rightward well occurs on a larger time scale .
we also observe that the regime of noise switching rates for resonant activation spans a wider regime with increasing dissipation strength .
see also in fig .
[ fig8 ] below , where the same features are obtained when the tunneling amplitude is being modulated periodically instead .
+ here we consider a situation in which the tunneling matrix element is subjected to a periodically varying , deterministic driving of the form @xmath154 in fig .
[ fig6 ] the fpt pdf @xmath80 , with @xmath155 given by eq .
( [ eqpl ] ) and a phase - dependent rate @xmath63 , is considered for three values of the phase @xmath14 ( see in eq .
( [ driven - delta ] ) ) .
the presence of periodic driving causes a modulation on the fpt pdf similar in spirit to the fpt pdf obtained for a periodically driven leaky integrate - and - fire model for neural spiking ; there , the fpt pdf peaks ( for an initial driving phase @xmath156 ) seemingly tend to synchronize with the driving oscillation period in the adiabatic limit @xmath157 .
moreover , this oscillating behavior is similar to that observed for the switching time probability in a long josephson junction @xcite .
+ the mfpt is obtained by solving eq .
( [ eqpl ] ) with time - dependent rate @xmath63 determined by using eq .
( [ w - pm ] ) with @xmath127 and @xmath9 from eq .
( [ driven - delta ] ) .
the mfpt @xmath72 versus the angular driving frequency @xmath158 , for different values of the amplitude @xmath159 , is shown in fig .
[ fig7 ] . for each value of @xmath158 the average over the phase @xmath14 of the driving in eq .
( [ g - driving - only ] ) is realized by uniformly sampling the interval @xmath160 at @xmath161 intermediate values .
the results for the mfpt display essentially the same features as for the noise - driven tunneling matrix element in fig .
namely , there are the low frequency saturation to a maximal value at slow driving , the resonant activated regime occurring at intermediate driving frequencies , where @xmath72 underruns the static value , while at high frequency driving the mfpt converges to that of the average configuration .
the latter coincides with the static configuration case . also in this case , results for a larger driving amplitudes range ( not shown ) display a nearly exact crossing .
this implies that the crossing frequency @xmath162 , where @xmath72 enters the resonant activation regime ( i.e. the crossing with the horizontal line in fig .
[ fig7 ] ) , also here depends weakly on the amplitude @xmath159 .
+ in fig .
[ fig8 ] we compare the obtained mfpts versus the oscillation angular frequency for different values of dissipation strength @xmath20 .
the results show the same features observed already in fig .
[ fig5 ] for the noise - modulated tunneling matrix element . as a second configuration with purely deterministic modulation we consider the case where the tunneling matrix element is held constant , @xmath111 , while a periodic driving modulates the bias @xmath11 according to @xmath163 the population of the left state satisfies formally the same equation as for the periodically driven tunneling matrix element , eq .
( [ eqpl ] ) , with forward transition rate @xmath63 given by eq .
( [ w - pm ] ) and fixed tunneling strength @xmath111 .
+ in fig . [ fig9 ]
the fpt pdf @xmath80 is depicted for three values of the initial driving phase @xmath14 .
the fpt pdf displays multiple peaks also in this case .
interestingly , using different initial driving phases @xmath14 , the fpt pdf exhibits peak locations at time points @xmath50 that differ from those obtained by using a periodically driven tunneling matrix element ( see fig . [ fig6 ] ) .
+ results for the mfpt @xmath72 versus the angular frequency @xmath164 , for different values of the driving amplitude @xmath165 , are shown in fig . [ fig10 ] . as in the previous subsection , also in this case , the average over @xmath14 prescribed by eq . ( [ g - driving - only ] )
is performed by uniformly sampling the interval @xmath160 at @xmath161 intermediate values .
+ the mfpt results versus angular driving frequency in fig .
[ fig10 ] overall share the same features with those for noise - driven and periodically driven tunneling matrix element shown in figs .
[ fig3 ] and [ fig7 ] , respectively . in this subsection
we consider the combined action of dichotomous noise and a periodic driving .
specifically , we consider the mfpt @xmath72 as a function of the noise switching rate @xmath85 of two - state noise on the tunneling matrix element , detailed by eq .
( [ delta_noise ] ) , while simultaneously rocking periodically the bias at angular frequency @xmath164 , according to eq .
( [ epsilon - t ] ) .
+ fig .
[ fig11 ] depicts the dynamics of the fpt pdf @xmath166\rangle_{\eta}$ ] for a fixed initial driving phase @xmath167 and for two values of the poisson parameter of the telegraphic noise , modulating the tunneling matrix element .
as in figs . [ fig6 ] and [ fig9 ] , @xmath168 is modulated due to the presence of the deterministic periodic driving .
this time however , the additional presence of two - state noise , plotted for the same two switching rate parameters @xmath85 , as done in fig .
[ fig2 ] , affects the average behavior ; it does , however , not wash out the multi - peak behavior imposed by the applied periodic forcing .
+ in the setting considered here , the mfpt @xmath72 is obtained by solving eq .
( [ eqplfluct ] ) with transition rates given by eq .
( [ w - pm - i ] ) .
our findings are shown in fig .
[ fig12 ] for two angular driving frequencies of the bias .
these two chosen values for @xmath164 are marked by filled circles in fig .
[ fig10 ] ( see curve of mfpt at @xmath169 ) .
a further comparison is made with the noise - only case , i.e. with the periodic driving being switched off . also here
the average over the initial driving phase @xmath14 detailed by eq .
( [ g - driving - noise ] ) is performed by uniformly sampling the interval @xmath160 at @xmath161 points .
+ while the overall behavior of @xmath170 exhibits the same features observed as in subsection iv.a , the role of introducing a periodically driven bias with frequency @xmath164 consists in shifting downwards to smaller values the curves of the mfpt versus the switching rate @xmath85 .
specifically , @xmath72 assumes systematically lower values with a bias @xmath11 periodically driven at @xmath171 . at this angular driving frequency
, @xmath170 converges in the limit @xmath95 to the value highlighted by the full circle located at the minimum of @xmath72 in fig .
[ fig10 ] , as to be expected .
likewise , for the case of a large angular driving frequency , i.e. , @xmath172 , the line @xmath170 virtually coincides with the analytical result obtained by setting @xmath173 ( i.e. only dichotomous noise is driving the tunneling matrix element ) .
this is due to the fact that , for such a large deterministic driving frequency , one approaches the situation discussed in fig .
[ fig3 ] ( green line ) for @xmath174 .
with this work we studied , by use of analytical and numerical means , the phenomenon of resonant activation , occurring for a dissipative two - state quantum system ( spin - boson system ) which is modulated by either periodic deterministic driving or also via telegraphic two - state noise . at strong dissipative system - bath coupling
the quantum dynamics proceeds incoherently so that an effective classical description in terms of a master equation with incoherent quantum rates becomes feasible .
this in turn allows for studying the detailed first passage time statistics when initially starting out at one of the two metastable states and with absorption occurring at the neighboring quantum state .
+ here we studied the complete first passage time probability density for general time - dependent driving of the two energy parameters characterizing the spin - boson system .
specific driving mechanisms involve a modulation in terms of a stationary two - state process with exponentially correlated noise or also an external deterministic periodic driving of those parameters , including combinations of both driving mechanisms .
in contrast to the case of stationary noise driving , the passage time dynamics for deterministic driving is cumbersome as it involves explicit time - dependent transition rates with corresponding time - dependent boundary conditions for reflection and absorption .
particularly , the role of periodic driving results in a decaying first passage time probability densities which exhibits multiple peaks .
those peaks reflect an initial - phase dependent quantum synchronization feature @xcite .
this latter feature is absent when the transition rates are time - independent ( stationary noise driving ) , resulting now in a monotonic decay of the first passage time pdf .
+ this first passage time statistics allows for the evaluation of all its moments .
of particular interest is its first mean , the mfpt .
this quantity displays the typical signatures of resonant activation , i.e. the existence of an intermediate modulation regime where the mfpt undrerruns the values assumed in the opposite limits of adiabatic slow driving and high frequency modulation . in the limit of very high frequency modulation one approaches the non - driven mfpt value .
+ our findings for various modulation settings corroborate the _ universal _ behavior @xcite found for classical over - the - barrier - resonant - activation , where ( i ) at low frequencies the mfpt is dominated by the adiabatic configuration , with the largest possible passage time ruling the overall escape , while ( ii ) for high frequency modulations the mfpt is governed by the value of the time - averaged energy profile yielding typically the static mfpt value ; ( iii ) for barrier modulations driven at intermediate time - scales ( of the order of the system dynamics time scale ) the regime with minimal mfpt values emerges ( resonant activation regime ) where the mfpt underruns both limits .
this so modulated tss possesses a wide parameter region for the two quantum transition rates .
this allows one to engineer the regime of resonant activation towards either smaller or also ( more interestingly ) much wider regimes .
this feature is illustrated with fig .
[ fig3 ] versus fig .
[ fig5 ] , or alike with figs .
[ fig7 ] and [ fig10 ] versus fig .
[ fig8 ] . + a further interesting feature we detected with this study is the approximate , although nearly exact , crossing behavior ( as demonstrated analytically and validated numerically in sec .
iv.a ) of the nonadiabatic mfpt entering the resonant activation regime at some critical frequency @xmath0 , being only weakly dependent on the driving amplitude . +
while the present approach readily can be generalized to situations with many intermediate quantum states ( overdamped tight - binding systems ) open challenges remain .
a particularly difficult objective to be addressed in the future is its extension to the regime of quantum coherence ; i.e. to the case in which modulations act on weakly damped quantum systems . in this latter regime
the very concept of a ( quasi-)classical mfpt analysis is doomed to fail .
the authors would like to thank prof .
p. talkner for helpful discussions .
p.h . acknowledges support by the russian science foundation ; grant no .
15 - 12 - 20029 .
b. s. and d. v. acknowledge support by ministry of education , university and research of italian government .
using a reasoning put forward in ref .
@xcite , a dichotomous noise allows for an exact averaging of the dynamics of the population difference @xmath44 , which results in a set of equations where @xmath175 is coupled to the correlation expression @xmath176 . + along the same lines we derive eq .
( [ eqplfluct ] ) via an averaging of the equation for @xmath2 , with @xmath57 being an absorbing state , over the noise realizations @xmath116 of the dichotomous two - state process @xmath177 detailed in sec .
[ noise - av - me ] .
we start out from @xmath178 where @xmath179.\end{aligned}\ ] ] substituting eq .
( [ noise ] ) into eq .
( [ w - p ] ) and performing the average over the noise we obtain @xmath180 where @xmath181 and the rates @xmath182 are given in eq .
( [ w - pm - i ] ) . in passing from first to second line of eq .
( [ eq - av ] ) we made use of two results in ref .
the first is @xmath183 \rangle_{\eta}&=&\langle \eta(t)\eta(t_1)\rangle_{\eta}\langle \varphi[\eta(\ ; ) ] \rangle_{\eta}\qquad \end{aligned}\ ] ] with @xmath184 , where @xmath185 $ ] is a functional of the dichotomous noise involving times @xmath186 .
choosing @xmath185\equiv \eta(t_1)p_l(t_1)$ ] and using the properties @xmath86 and @xmath187 , we obtain one of the identities necessary to derive eq .
( [ eq - av ] ) , namely @xmath188 ( @xmath189 ) .
the second result in ref .
@xcite reads @xmath190 \eta(t)\eta(t_1)\chi[\eta(\ ; ) ] \rangle_{\eta}=\langle \varphi[\eta(\ ; ) ] \eta(t)\rangle_{\eta}\langle \eta(t_1)\chi[\eta(\ ; ) ] \rangle_{\eta}&\nonumber\\ & + \langle \varphi[\eta(\ ; ) ] \rangle_{\eta } \langle \eta(t)\eta(t_1)\langle\chi[\eta(\ ; ) ] \rangle_{\eta}&\end{aligned}\ ] ] with @xmath184 , where @xmath185 $ ] and @xmath191 $ ] are two functionals of the dichotomous noise involving times @xmath192 and @xmath186 , respectively .
taking @xmath185\equiv p_l(t)$ ] and @xmath191\equiv1 $ ] , and using the property @xmath193 , we get the identity @xmath194 with @xmath189 .
( [ result2-a ] ) is used , along with eq .
( [ result1-a ] ) , to obtain eq .
( [ eq - av ] ) .
+ next , the equation for @xmath181 can be derived analogously starting from the shapiro - loginov theorem @xcite , which states that @xmath195 using eq .
( [ eq ] ) for @xmath65 on the rhs , calculating the noise averages by means of eqs .
( [ result1-a ] ) and ( [ result2-a ] ) , and observing again that @xmath86 , we find the following equation for @xmath196 @xmath197y(t).\quad\end{aligned}\ ] ] 46ifxundefined [ 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 http://link.aps.org/doi/10.1103/revmodphys.62.251 [ * * , ( ) ] link:\doibase 10.1016/b978 - 0 - 12 - 287460 - 4.50005 - 9 [ _ _ ] ( , ) http://link.aps.org/doi/10.1103/physrevlett.51.2242 [ * * , ( ) ] @noop _ _ ( , ) http://dx.doi.org/10.1038/nature18272 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.93.048102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.108103 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.69.2318 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.73.2772 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.51.3849 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.52.3301.2 [ * * , ( ) ] `` , '' in link:\doibase 10.1007/bfb0105605 [ _ _ ] , , vol . , ( , ) pp
. http://link.aps.org/doi/10.1103/physreve.69.061103 [ * * , ( ) ] http://stacks.iop.org/0295-5075/74/i=6/a=937 [ * * , ( ) ]
link:\doibase 10.1103/revmodphys.59.1 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.67.725 [ * * , ( ) ] @noop _ _ , ed .
( , ) \doibase http://dx.doi.org/10.1063/1.470630 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.81.649 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.81.2837 [ * * , ( ) ]
link:\doibase 10.1080/00018730500429701 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/s0370157300000478 [ * * , ( ) ] http://ptp.oxfordjournals.org/content/102/4/707.abstract [ * * , ( ) ] link:\doibase 10.1103/physrevlett.46.211 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.66.810 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys3905 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.60.r1 [ * * , ( ) ] http://stacks.iop.org/0295-5075/52/i=3/a=264 [ * * , ( ) ] http://scitation.aip.org/content/aip/journal/jcp/123/22/10.1063/1.2135781 [ * * , ( ) ] link:\doibase 10.1016/0003 - 4916(83)90202 - 6 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(98)00022 - 2 [ * * , ( ) ] link:\doibase 10.1103/physreve.53.r2033 [ * * , ( ) ] link:\doibase 10.1006/aphy.2001.6174 [ * * , ( ) ] `` , '' ( , ) pp . ,
http://stacks.iop.org/1367-2630/1/i=1/a=314 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.97.210601 [ * * , ( ) ] in link:\doibase 10.1016/b978 - 0 - 12 - 287460 - 4.50005 - 9
[ _ _ ] ( , ) pp .
link:\doibase 10.1016/j.physa.2004.12.020 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed .
( , ) http://link.aps.org/doi/10.1103/physrevlett.71.1649 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.83.041122 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevb.82.033303 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevb.89.214510 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0031-8914(73)90347-9 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/037843717890198x [ * * , ( ) ] | resonant activation in temporally driven spin - boson systems subjected to strong dissipation is researched by means of both , analytical and extensive numerical investigations .
the phenomenon of resonant activation emerges in the presence of either fluctuating or periodically varying driving fields . addressing the incoherent regime ,
a characteristic minimum emerges in the mean first passage time ( mfpt ) to reach an absorbing neighboring state whenever the intrinsic time scale of the modulation matches the characteristic time scale of the system dynamics . for the case of deterministic periodic
driving the first passage time statistics displays a complex , multi - peaked first passage time probability density , which depends crucially on the details of the initial driving phase , the driving frequency and its strength . as an interesting feature we find that the mfpt enters the resonant activation regime at a critical frequency @xmath0 which depends very weakly on the strength of the driving . |
the last two decades or so have seen an explosion of application areas of evolutionary algorithms ( eas ) in diverse scientific or engineering disciplines .
an ea is a random search heuristic , using evolutionary mechanisms such as crossover and mutation , for finding a solution that often aims at maximizing an objective function .
eas are proved to be extremely useful for combinatorial optimization problems because they can solve complicated problems with reasonable efficiency using only basic mathematical modeling and simple operators ; see @xcite for more information .
although eas have been widely applied in solving practical problems , the analysis of their performance and efficiency , which often provides better modeling prediction for practical uses , are much less addressed ; only computer simulation results are available for most of the eas in use .
see for example @xcite .
we are concerned in this paper with a precise probabilistic analysis of a simple algorithm called ( @xmath1)-ea ( see below for more details ) .
a typical ea comprises several ingredients : the coding of solution , the population of individuals , the selection for reproduction , the operations for breeding new individuals , and the fitness function to evaluate the new individual .
thus mathematical analysis of the total complexity or the stochastic description of the algorithm dynamics is often challenging .
it proves more insightful to look instead at simplified versions of the algorithm , seeking for a compromise between mathematical tractability and general predictability .
such a consideration was first attempted by bck @xcite and mhlenbein in @xcite in the early 1990 s for the ( @xmath1)-ea , using only one individual with a single mutation operator at each stage .
an outline of the procedure is as follows .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * algorithm ( @xmath1)-ea * 1 .
choose an initial string @xmath2 uniformly at random 2 .
repeat until a terminating condition is reached * create @xmath3 by flipping each bit of @xmath4 independently with probability @xmath5 * replace @xmath6 by @xmath3 iff @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ step 1 is often realized by tossing a fair coin for each of the @xmath8 bits , one independently of the others , and the terminating condition is usually either reaching an optimum state ( if known ) or by the number of iterations .
mhlenbein @xcite considered in detail the complexity of ( @xmath1)-ea under the fitness function onemax , which counts the number of ones , namely , @xmath9 .
the expected time needed to reach the optimum value , which is often referred to as the expected _ optimization time _
, for onemax , denoted for convenience by @xmath10 , was argued to be of order @xmath11 , indicating the efficiency of the ( @xmath1)-ea .
bck @xcite derived expressions for the transition probabilities .
finer asymptotic approximation of the form @xmath12 was derived by garnier et al . in @xcite , where @xmath13 when the mutation rate @xmath14 .
they went further by characterizing the limiting distribution of @xmath15 in terms of a log - exponential distribution ( which is indeed a double exponential or a gumbel distribution ) .
however , some of their proofs , notably the error analysis , seem incomplete ( as indicated in their paper ) .
thus a strong result such as has remained obscure in the ea literature .
more recent attention has been paid to the analysis of the ( @xmath1)-ea ; see for example @xcite .
we briefly mention some progresses .
neumann and witt @xcite proved that a simple ant colony optimization algorithm behaves like the ( @xmath16)-ea and all results for ( @xmath1)-ea translate directly into those for the aco algorithm .
sudholt and witt @xcite showed a similar translation into particle swarm optimization algorithms .
moreover , variants such as ( @xmath17)-ea in @xcite and ( @xmath1)-ea over a finite alphabet in @xcite were investigated . the expected optimization time required by
( @xmath1)-ea has undergone successive improvements , yet none of them reached the precision of garnier et al.s result ; we summarize in the following table some recent findings . [ cols="^,^,^,^ " , ]
recall that we start from the initial state @xmath19 and that @xmath20 denotes the number of steps used by ( @xmath1)-ea before reaching @xmath21 .
we derive first a recurrence relation satisfied by the probability generating function @xmath22 of @xmath20 .
the probability generating function @xmath23 satisfies the recurrence @xmath24 for @xmath25 , with @xmath26 , where @xmath27 start from the state @xmath28 and run the two steps inside the loop of algorithm ( @xmath1)-ea .
the new state becomes @xmath3 with @xmath29 if @xmath30 bits in the group @xmath31 and @xmath32 bits in the other group @xmath33 toggled their values , where @xmath34 and @xmath35 .
thus , the probability from state @xmath6 to @xmath3 is given by @xmath36 which is identical to .
we then obtain @xmath37 and this proves the lemma .
while this simple recurrence relation seems not new in the ea literature , tools have been lacking for a direct asymptotic resolution , which we will develop in detail in this paper . for convenience ,
define @xmath38 in particular , when @xmath39 , @xmath40 so that @xmath41 this is a standard geometric distribution @xmath42 with probability @xmath43 ( assuming only positive integer values ) . obviously ,
taking @xmath44 , we obtain @xmath45 as @xmath46 , uniformly for @xmath47 , implying , by curtiss s convergence theorem ( see @xcite ) , the _ convergence in distribution _
@xmath48 where @xmath49 denotes an exponential distribution with parameter @xmath50 .
equivalently , this can be rewritten as @xmath51 for @xmath52 .
such a limit law indeed extends to the case when @xmath18 , which we formulate as follows . for @xmath53 ( in left to right order ) and @xmath54 , and their corresponding limit laws._,title="fig:",width=132 ] for @xmath53 ( in left to right order ) and @xmath54 , and their corresponding limit laws._,title="fig:",width=132 ] for @xmath53 ( in left to right order ) and @xmath54 , and their corresponding limit laws._,title="fig:",width=132 ] for @xmath53 ( in left to right order ) and @xmath54 , and their corresponding limit laws._,title="fig:",width=132 ] let @xmath55 denote the @xmath56-th order harmonic numbers and @xmath57 .
for convenience , we define @xmath58 . [
thm : yo1 ] if @xmath18 , the time used by ( @xmath1)-ea to reach the optimum state @xmath59 , when starting from @xmath28 , converges , when normalized by @xmath60 , to a sum of @xmath61 exponential random variables @xmath62 with mean asymptotic to @xmath63 and variance asymptotic to @xmath64 .
the convergence in distribution can be expressed alternatively as @xmath65 by the sum definition of @xmath66 , we see that @xmath67 where the @xmath68-term holds uniformly for @xmath69 and @xmath70 . in particular , for each fixed @xmath18 , we then have @xmath71 so that @xmath72 where both @xmath73-terms are uniform for @xmath47 . now take @xmath74 .
then @xmath75 uniformly for @xmath47 .
this and curtiss s convergence theorem ( see @xcite ) imply .
the asymptotic mean and the asymptotic variance can be computed either by a similar inductive argument or by following ideas used in the quasi - power framework ( see @xcite or @xcite ) that relies on the _ uniformity _ of the estimate @xmath76 p_{n , m}\left(e^{\frac{s}{en}}\right ) \sim [ s]\prod_{1\le r\le m}\frac1{1-\frac{s}{r } } = h_m,\ ] ] where @xmath77f(s)$ ] denotes the coefficient of @xmath78 in the taylor expansion of @xmath79 , and @xmath80 \log p_{n , m}\left(e^{\frac{s}{en}}\right)\\ & \sim 2 [ s^2 ] \sum_{1\le r\le m } \log\frac1{1-\frac sr}\\ & = h_m^{(2)}.\end{aligned}\ ] ] we will derive more precise expansions below by a direct approach .
the simple inductive argument fails when @xmath81 and we need more uniform estimates for the error terms .
sums of the form @xmath83 appear frequently in our analysis .
we thus digress in this section to develop tools for deriving the asymptotic behaviors of such sums . for technical simplicity
, we define the sequence @xmath84 and the normalized sum @xmath85 let also @xmath86 throughout this paper , we use the abbreviation @xmath87 [ [ asymptotics - of - a_nm . ] ] asymptotics of @xmath88 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + observe that most contribution to @xmath89 comes from small @xmath90 , say @xmath91 , provided that @xmath92 does not grow too fast ; see .
we formulate a more precise version as follows .
assume that @xmath93 is a given sequence such that @xmath94 has a nonzero radius of convergence in the @xmath95-plane .
then @xmath96 where @xmath97 and @xmath98 are entire functions of @xmath99 defined by @xmath100 and ( @xmath101 ) @xmath102 the first term on the right - hand side of can be readily obtained as follows . if @xmath25 , then @xmath103 the more precise approximation in can be obtained by refining all estimates , but the details are rather messy , notably the error analysis .
we resort instead to an analytic approach .
observe that the sum on the left - hand side of is itself a convolution .
our analytic proof then starts from the relation
@xmath104\left(z+\frac1n\right)^m \left(1+\frac zn\right)^{n+1-m}\nonumber \\ & = \frac1{2\pi i}\oint_{|z|=c } z^{\ell-1 } \left(1+\frac1{nz}\right)^m \left(1+\frac zn\right)^{n+1-m } \dd z , \label{lnml}\end{aligned}\ ] ] where @xmath105 .
the relation holds _ a priori _ for @xmath106 , but the right - hand side becomes zero for @xmath107 .
it follows that @xmath108 where @xmath109 , @xmath110 being the radius of convergence of @xmath111 . by the expansion @xmath112 uniformly for @xmath95 on the integration path , and the integral representations @xmath113 we deduce .
the expression is then obtained by straightforward term - by - term integration . for
, we apply the relation @xmath114 and integration by parts , and then obtain @xmath115 substituting the series expansion @xmath116 and integrating term by term , we get .
when @xmath99 tends to the two boundaries @xmath117 and @xmath118 , we have @xmath119 where @xmath120 is the smallest integer such that @xmath121 . [ [ asymptotics - of - sum_1le - ell - le - m - ellrlambda_nmell ] ] asymptotics of @xmath122 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we now discuss special sums of the form @xmath123 which will be repeatedly encountered below .
define @xmath124 so that @xmath125 . for convenience
, we also write @xmath126 let @xmath127 denote the modified bessel functions @xmath128 uniformly for @xmath25 [ cor : ur ] @xmath129 for @xmath130 , where both @xmath131 and @xmath132 are entire functions given by @xmath133 and @xmath134 in particular , @xmath135 these are sufficient for our uses .
we start with the integral representation ( see ) @xmath136 where @xmath137 .
when @xmath138 , we have @xmath139 .
thus @xmath140 note that @xmath141 thus @xmath142 which proves for @xmath138 . for @xmath143 , we have @xmath144 from this and the relation , we obtain .
note that the coefficient of @xmath145 is zero .
the corollary implies specially that @xmath146 uniformly for @xmath25 and @xmath147 .
since @xmath148 as @xmath149 , we have the uniform bound @xmath150 meaning that the ratio of @xmath151 remains bounded away from zero and infinity for all @xmath61 in the specified range .
we also have the limiting behaviors @xmath152 and @xmath153 without the first term , the right - hand side is , up to @xmath154 , the bell numbers ( all partitions of a set ; sequence a000110 in sloane s encyclopedia of integer sequences ) . the following expansions for @xmath155 and @xmath156 as @xmath157 will be used later @xmath158 for @xmath130 .
see also appendix a for other properties of @xmath159 .
consider the mean @xmath160 .
it satisfies the recurrence @xmath161 for @xmath25 with @xmath162 . from theorem [ thm : yo1 ]
, we already have @xmath163 when @xmath18 , while for @xmath81 and @xmath164 we expect that ( recalling @xmath165 ) @xmath166 see section [ sec : ae - small - m ] for how such a form arises .
we will indeed derive in this section a more precise expansion .
the uniform appearance of the harmonic numbers @xmath167 may be traced to the asymptotic estimate ; see also lemma [ lmm - at ] .
the expected value of @xmath20 satisfies the asymptotic approximation [ thm : mu ] @xmath168 uniformly for @xmath25 , where @xmath169 is defined in and @xmath170 is an analytic function defined by @xmath171 for simplicity , we consider @xmath172 where @xmath173 , and we will prove that @xmath174 for @xmath25 , which is identical to ; see figure [ fig : mu - diff ] for a graphical rendering .
more figures are collected in appendix b. for @xmath25 ( normalized to the unit interval ) and @xmath175 ( left in top - down order ) , and the normalized differences @xmath176 for @xmath175 ( right).__,title="fig:",width=226 ] for @xmath25 ( normalized to the unit interval ) and @xmath175 ( left in top - down order ) , and the normalized differences @xmath176 for @xmath175 ( right).__,title="fig:",width=226 ] our analysis will be based on the recurrence @xmath177 or alternatively @xmath178 with @xmath179 , where @xmath180 and @xmath181 are defined in and , respectively . in particular , this gives @xmath182 , @xmath183 in general , the @xmath184 are all rational functions of @xmath8 but their expressions become long as @xmath61 increases . in section [ sec : ae - small - m ] , we give asymptotic expansions for @xmath184 for small values of @xmath61 as @xmath185 , which are also required as initial values for obtaining more refined asymptotic expansions for other ranges of @xmath61 in section [ sec : mu - ae ] . starting from the asymptotic estimate @xmath186 when
@xmath18 , we first postulate an ansatz approximation of the form @xmath187 for some smooth function @xmath188 .
then we will justify such an expansion by an error analysis relying on lemma [ lmm - at ] after a proper choice of @xmath188 .
this same procedure can then be extended and yields a more precise expansion ; see section [ sec : mu - ae ] . instead of starting from a state with a fixed number of ones
, the first step of the algorithm ( @xmath1)-ea described in introduction corresponds to the situation when the initial state @xmath189 ( the number of @xmath118s ) is not fixed but random .
assume that this input follows a binomial distribution of parameter @xmath190 ( each bit being @xmath118 with probability @xmath191 and @xmath117 with probability @xmath192 ) .
denote by @xmath193 the number of steps used by ( @xmath1)-ea to reach the optimum state .
such a situation can also be dealt with by applying theorem [ thm : y - all ] and we obtain the same limit law .
the following result describes precisely the asymptotic behavior of the expected optimization time .
the expected value of @xmath193 satisfies [ thm : eyn ] @xmath194 note that @xmath195 is an increasing function of @xmath192 , which is consistent with the intuition that it takes less steps to reach the final state if we start with more @xmath118s ( small @xmath192 means @xmath191 closer to @xmath118 , or @xmath118 occurring with higher probability ) .
also @xmath196 the constant @xmath197 in can now be computed and has the value @xmath198 numerically , to compute the value of @xmath199 for @xmath200 $ ] , the most natural way consists in using the taylor expansion @xmath201 and after a term - by - term integration @xmath202
@xmath203 _ has an infinity number of zeros on @xmath204 . _
while @xmath203 is an entire functions with rapidly decreasing coefficients , such an expansion converges slowly when @xmath205 , the main reason being that the smallest @xmath206 for which @xmath207 occurs when @xmath208 , implying that the radius of convergence of this series is slightly larger than unity .
note that @xmath209 but the simple pole is removed by subtracting @xmath210 .
a better idea is then expanding @xmath211 at @xmath212 and integrating term - by - term @xmath213 this expansion is numerically more efficient and stable because of better convergence for @xmath214 $ ] .
the same technique also applies to the calculation of @xmath170 and other functions in this paper .
our asymptotic approximation to @xmath184 was largely motivated by intensive symbolic computations for small @xmath61 .
we briefly summarize them here , which will also be crucial in specifying the initial conditions for the differential equations satisfied by functions ( @xmath215 ) involved in the full asymptotic expansion of @xmath184 ; see . starting from the closed - form expressions ,
we readily obtain @xmath216 , @xmath217 , and @xmath218 similarly , we have @xmath219 from these expansions , we first observe that the leading sequence is exactly @xmath167 ( @xmath220 ) @xmath221 these also suggest the following ansatz @xmath222 for some functions @xmath223 of @xmath61 . using this form and the above expansions to match the undetermined coefficients of the polynomials ( in @xmath61 ) , we obtain successively @xmath224 so we observe the general pattern @xmath225 for some explicitly computable sequence @xmath226 and coefficients @xmath227 .
a crucial complication arises here : the general form for each @xmath223 holds only for @xmath228 , and correction terms are needed for smaller @xmath61 .
for example , @xmath229 where we use the iverson bracket notation @xmath230 if @xmath111 holds , and @xmath117 , otherwise .
it is such a complication that makes the determination of smaller - order terms more involved .
all the expansions here hold only for small @xmath61 .
when @xmath61 grows , we see that @xmath231 and it is exactly this form that motivated naturally our choice of the ansatz . we develop here some other asymptotic tools that will be used in proving theorem [ thm : mu ] .
the following lemma is very helpful in obtaining error estimates to be addressed below .
it also sheds new light on the occurrence of the harmonic numbers @xmath167 in .
consider the recurrence [ lmm - at ] @xmath232 where @xmath233 is defined for @xmath25 and @xmath234 .
assume that @xmath235 for @xmath236 , where @xmath237 .
if @xmath238 holds uniformly for @xmath25 and @xmath234 , where @xmath105 , then @xmath239 the result is true for @xmath240 . for @xmath241 ,
we start from the simple inequality @xmath242 because all terms in the sum expression are positive and taking only one term ( @xmath243 and @xmath244 ) gives the lower bound
. then , by the induction hypothesis , @xmath245 proving the lemma . applying this lemma to the recurrence , we then get a simple upper bound to @xmath184 . for @xmath246 ,
the inequality @xmath247 holds . if @xmath188 is a @xmath248$]-function , then [ lmm - phi ] @xmath249 uniformly for @xmath25 .
a direct taylor expansion with remainder gives @xmath250 uniformly for @xmath251 , since @xmath252 for @xmath253 $ ] .
the lemma follows from the estimates .
the approximation can be easily extended and refined if more smoothness properties of @xmath188 are known , which is the case for all functions appearing in our analysis ( they are all @xmath254 $ ] ) .
another standard technique we need is stirling s formula for the factorials @xmath255 where @xmath256 denotes euler s gamma function .
[ [ formal - calculus . ] ] formal calculus .
+ + + + + + + + + + + + + + + + applying formally and lemma [ lmm - phi ] using @xmath257 and @xmath258 , we have @xmath259 by .
thus we see that @xmath188 satisfies @xmath260 we now specify the initial condition @xmath261 . since the postulated form holds for @xmath25 ( indeed also true for @xmath240 ) , we take @xmath39 and see that @xmath262 because @xmath182 .
this implies that @xmath263 .
the first few terms in the taylor expansion of @xmath199 read as follows .
@xmath264 which can then be checked with the explicit expressions of @xmath184 for small @xmath61 ( see section [ sec : ae - small - m ] ) .
[ [ error - analysis . ] ] error analysis .
+ + + + + + + + + + + + + + + to justify the form ( with @xmath265 ) , we consider the difference @xmath266 which satisfies the recurrence @xmath267 where @xmath268 by the asymptotic relation with @xmath269 and the definition of @xmath169 , we have @xmath270 and thus @xmath271 by lemma [ lmm - phi ] , we see that @xmath272 uniformly for @xmath25 . on the other hand
, we have the upper bounds @xmath273 note that the first estimate is only uniform for @xmath274 . when @xmath90 is close to @xmath61 , say @xmath275 , the left - hand side blows up with @xmath61 but the right - hand side @xmath276 remains bounded .
thus we split the sum at @xmath277 and then obtain ( @xmath278 when @xmath244 ) @xmath279 now , by , @xmath280 and @xmath281 by stirling s formula , the last @xmath68-term is of order @xmath282 for @xmath283 . combining these estimates ,
we then obtain @xmath284 uniformly for @xmath25 .
thus @xmath285 satisfies a recurrence of the form @xmath286 with @xmath287 .
it follows , by applying lemma [ lmm - at ] , that @xmath288 , and we conclude that , uniformly for @xmath289 , @xmath290 this proves the first two terms of the asymptotic approximation to @xmath184 in .
the more refined expansion is obtained by refining the same calculations and justification , which we carry out the main steps subsequently .
[ [ refined - computations . ] ] refined computations .
+ + + + + + + + + + + + + + + + + + + + + we consider now the difference @xmath291 and will determine the constant @xmath292 and the function @xmath293 such that @xmath294 uniformly for @xmath295 , which then proves theorem [ thm : mu ] . by ,
@xmath296 satisfies , for @xmath295 , the recurrence @xmath297 where @xmath298 in particular , @xmath299 .
the hard part here is to derive an asymptotic expansion for @xmath300 that holds uniformly for @xmath295 as @xmath301 . to that purpose , we first extend lemma [ lmm - phi ] by using a taylor expansion of third order for a @xmath302$]-function @xmath303 , which then gives , uniformly for @xmath295 , @xmath304 where we used corollary [ cor : ur ] .
we now examine weighted sums involving the difference of the harmonic numbers .
we start with the following identity whose proof is straightforward . for a given function @xmath305 , let @xmath306 denote the backward difference operator @xmath307 . then for @xmath308 @xmath309 note that the sum vanishes for @xmath310 .
take @xmath311 .
then we obtain @xmath312 for @xmath308 .
this relation implies that , for @xmath313 and @xmath241 , @xmath314 by the same argument we used above for , we get the expansion @xmath315 which holds uniformly for @xmath295 .
note that for @xmath39 a correction term is needed ; more correction terms have to be introduced in more refined expansions ( see section [ sec : mu - ae ] ) . combining the estimates ( with @xmath316 ) and
, we see that @xmath317 uniformly for @xmath25 , where @xmath318 obviously , @xmath319 because @xmath320 . to determine @xmath321 and @xmath170
, we observe that @xmath322 where we used the relation @xmath323 ( see ) . in order that @xmath324 uniformly for @xmath25 , we need @xmath325 , so that @xmath326 .
now the equation @xmath327 also implies , by , that @xmath328 with these choices of @xmath292 and @xmath293 , we have @xmath329 uniformly for @xmath295 . the exact solution to the differential equation
requires the constant term @xmath330 , which we have not yet specified . to specify this value ,
we take @xmath39 in and then obtain , by the recurrence , @xmath331 this entails the choice @xmath332 in order that @xmath333 .
thus we obtain the integral solution for @xmath334 .
in particular , the first few terms of @xmath293 in the taylor expansion are given as follows .
@xmath335 as a function in the complex plane , the region where @xmath334 is analytic is dictated by the first zeros of @xmath336 , which exceeds unity . to complete the proof of
, we require a variation of lemma [ lmm - at ] , since the assumption on @xmath337 given there is not satisfied here .
[ lmm : rec_variation ] consider the recurrence [ lmm - at - variation ] @xmath338 where @xmath233 is defined for @xmath25 and @xmath234 .
assume that @xmath339 for @xmath236 , and @xmath340 for @xmath236 .
if there exists a @xmath105 such that @xmath238 holds uniformly for @xmath341 and @xmath234 , then @xmath342 the inequality holds when @xmath39 by assumption . for @xmath343 ,
we write the recurrence as follows @xmath344 by induction hypothesis and the two inequalities ( see lemma [ lmm - at ] ) @xmath345 we obtain @xmath346 and this proves the lemma . in view of the estimates
@xmath347 , @xmath348 and @xmath349 , for @xmath350 , there exists a constant @xmath105 such that the quantity @xmath351 satisfies the assumptions of lemma [ lmm : rec_variation ] , which implies the bound @xmath352 , or , equivalently @xmath353 , uniformly for @xmath295 .
this completes the proof of theorem [ thm : mu ] .
the above procedure can be extended to get more smaller - order terms , but the expressions for the coefficients soon become very involved .
however , it follows from the discussions in [ sec : ae - small - m ] that we expect the asymptotic expansion @xmath354 in the sense that the truncated asymptotic expansion @xmath355 holds uniformly for @xmath356 and introduces an error of order @xmath357 .
this asymptotic approximation may not hold when @xmath358 because additional correction terms are needed in that case .
technically , the correction terms stem from asymptotic expansions for sums of the form @xmath359 ; see and the comments given there .
we propose here an easily codable procedure for the coefficients in the expansion , whose justification follows the same error analysis as above .
we start with the formal expansion and expand in all terms for large @xmath360 in decreasing powers of @xmath8 , match the coefficients of @xmath361 on both sides for each @xmath362 , and then adjust the initial condition @xmath363 by taking into account the extremal case when @xmath364 ( for @xmath365 the expansion up to that order may not hold ) . with this algorithmic approach it is possible to determine the coefficients @xmath366 and the functions @xmath367 successively one after another .
observe first that @xmath368 where ( @xmath369 ) @xmath370 the @xmath371 representing the bernoulli numbers . on the other hand , @xmath372 thus @xmath373 then , by , @xmath374 \left(1+\frac1{nt}\right)^m \left(1+\frac tn\right)^{n+1-m } f_r(t),\ ] ] where @xmath375 now @xmath376 a direct expansion using bell polynomials @xmath377 ( see @xcite ) then gives @xmath378 where @xmath379 , @xmath380 and @xmath381 is a polynomial of degree @xmath382 .
collecting these expansions , we get @xmath383e^{\frac\alpha t+(1-\alpha)t } \frac{\tilde{b}_r(\mathbf{t})}{r!}\,f_{k+1-r}(t).\end{aligned}\ ] ] all terms now have the form @xmath384e^{\frac\alpha t+(1-\alpha)t } f(t ) & = \sum_{\ell\ge0}\frac{\alpha^\ell}{\ell ! } [ t^{\ell+2r-1}]e^{(1-\alpha)t}f(t ) \\ & = \sum_{\ell\ge0}\frac{\alpha^\ell}{\ell ! } \sum_{0\le j<2r+\ell } \frac{(1-\alpha)^j}{j ! } \cdot \frac{f^{(\ell+2r-1-j)}(0 ) } { ( \ell+2r-1-j)!}.\end{aligned}\ ] ] since we are solving the recurrence @xmath385 we have the relations @xmath386\frac{e^{\frac\alpha t+(1-\alpha)t } } { ( 1-t)^2 } = 1,\\ \displaystyle \sum_{0 \le
r \le k } [ t^{2r-1}]e^{\frac\alpha t+(1-\alpha)t } \frac{\tilde{b}_r(\mathbf{t})}{r!}\,f_{k+1-r}(t ) = 0 , \quad ( k\ge1 ) .
\end{cases}\ ] ] by induction , each @xmath387 satisfies a differential equation of the form @xmath388(\alpha),\ ] ] for some functional @xmath389 . since @xmath390 as @xmath391
, we also have the relation @xmath392(0).\ ] ] once the value of @xmath366 is determined , we can then write @xmath393(x ) } { s_{1}(x ) } - \frac{b_{k}}{x}\right)\dd x,\ ] ] and it remains to determine the initial value @xmath363 , which is far from being obvious .
the crucial property we need is that the truncated expansion holds when @xmath356 , and particularly when @xmath364 .
so we compute with @xmath364 and drop all terms of order smaller than or equal to @xmath361 .
then we match the coefficient of @xmath394 with that in the expansion of @xmath395 obtained by a direct calculation from the recurrence .
we illustrate this procedure by computing the first two terms in .
first , we have @xmath396 which implies @xmath397 and @xmath398 .
moreover , substituting the initial value @xmath399 , we get @xmath400 entailing @xmath401 , which is consistent with what we obtained above . the next - order term when @xmath402 is ( after substituting the relations @xmath397 , @xmath403 and @xmath404 ) @xmath405\frac{e^{\frac \alpha t+(1-\alpha)t } } { ( 1-t)^2 } \\ & \qquad = [ t^{-1}]e^{\frac \alpha t+(1-\alpha)t } \left(\frac1{2\alpha^2(1-t)^2 } -\frac{(1+t)s_1'(\alpha)}{2(1-t)^3s_1(\alpha ) } + \frac{\alpha-2t^3+(1-\alpha)t^4}{2 t^2(1-t)^2s_1(\alpha)}\right),\end{aligned}\ ] ] implying that @xmath406 as @xmath407 , the right - hand side of has the local expansion @xmath408 , forcing @xmath326 , and , accordingly , we obtain the same differential equation . substituting the value @xmath409 in yields @xmath410 implying , by using , @xmath411 , which is consistent with theorem [ thm : mu ] . although the expressions become rather involved for higher - order terms , all calculations ( symbolic or numerical ) are easily coded .
for example , we have @xmath412 we now give an outline of the proof of theorem [ thm : eyn ] concerning the asymptotics of @xmath10 .
the method of proof relies on standard normal approximation to the binomial distribution .
we begin with @xmath413 where @xmath414 ( @xmath415 ) . from this expression
, we see that @xmath416 write @xmath417 . by stirling s formula
, we have @xmath418 uniformly for @xmath419 , where , here and throughout the proof , the @xmath420 are polynomials of @xmath4 containing only powers of the same parity as @xmath30 . on the other hand , by theorem [ thm : mu ] , we have in the same range of @xmath61 @xmath421 with these expansions , the asymptotic evaluation of @xmath422 is reduced to sums of the form @xmath423 for any @xmath424 by an application of the euler - maclaurin formula .
thus polynomials of odd indices ( containing only odd powers of @xmath4 ) will lead to asymptotically negligible terms after integration . outside the range where @xmath419 ,
the binomial distribution is smaller than any negative power of @xmath8 , so the contribution from this range is also asymptotically negligible . except for this part
, all other steps are easily coded .
we prove in this section that the variance @xmath425 of @xmath20 is asymptotically quadratic . for @xmath25 , the variance of @xmath20 satisfies [ thm : var1 ] @xmath426 where @xmath427 and @xmath428 similar to the mean
, we work on the sequence @xmath429 and prove that ( see figure [ fig : var - diff ] and appendix d ) @xmath430 rhs of @xmath431 for @xmath341 ( normalized to the unit interval ) and @xmath175 ( left in top - down order ) , and the absolute normalized differences @xmath432 rhs of @xmath431 for @xmath175 ( right).__,title="fig:",width=226 ] rhs of @xmath431 for @xmath341 ( normalized to the unit interval ) and @xmath175 ( left in top - down order ) , and the absolute normalized differences @xmath432 rhs of @xmath431 for @xmath175 ( right).__,title="fig:",width=226 ] the variance of @xmath193 is computed by the relation @xmath433 where @xmath434 , @xmath435 and @xmath436 . the variance of @xmath193 satisfies asymptotically [ thm : var2 ] @xmath437 where ( @xmath415 ) @xmath438 and @xmath439 [ [ recurrences - for - the - centered - moment - generating - function . ] ] recurrences for the centered moment generating function .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to compute the variance , one may start with the second moment and then consider the difference with the square of the mean ; however , it is computationally more advantageous to study directly the recurrence satisfied by the variances themselves . from
, we have , by substituting @xmath440 , @xmath441 which can be rewritten as @xmath442 or @xmath443 this is a simpler recurrence to start as fewer terms are involved for the moments .
we now consider the moment generating function for the centered random variables @xmath444 @xmath445 which then satisfies the recurrence @xmath446 for @xmath25 with @xmath447 .
[ [ variance . ] ] variance .
+ + + + + + + + + let @xmath448 be the variance of @xmath20 .
then @xmath449 satisfies the recurrence @xmath450 in terms of @xmath451 , we have @xmath452 , and for @xmath295 @xmath453 where @xmath454 in particular , this gives @xmath455 the expressions become very lengthy as @xmath61 increases . in appendix
c we give asymptotic expansions for @xmath456 for a few small @xmath61 as @xmath185 .
based on these expansions , a suitable ansatz for the asymptotic behavior of @xmath456 can be deduced ( assisted again by computer algebra system ) , which then can be proven analogous to the method of proof presented in section [ sec : exp ] .
[ [ proof - of - asymptotics - with - error - analysis . ] ] proof of asymptotics with error analysis .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + by the same procedure used for @xmath184 , we start from computing the asymptotic expansions for @xmath456 for small @xmath61 .
these expansions suggest the more uniform ( for @xmath25 ) asymptotic expansion @xmath457 for some constants @xmath458 , @xmath459 and @xmath460 , and some function @xmath461 .
such an asymptotic form can be justified by the same approach we used above for @xmath184 .
more precisely , we now prove that @xmath462 uniformly for @xmath463 and @xmath236 , where @xmath461 is given in .
our proof start from considering the difference @xmath464 and specify the involved coefficients and @xmath461 such that @xmath353 . by ,
@xmath296 satisfies , for @xmath295 , the recurrence @xmath465 with the initial value is @xmath466 , where ( @xmath467 being defined in ) @xmath468 we will derive an asymptotic expansion for @xmath469 . for that purpose
, we use the expansions , as well as theorem [ thm : mu ] in section [ sec : exp ] , and apply the same error analysis used for @xmath184 .
a careful analysis then leads to @xmath470 and @xmath471 both holding uniformly for @xmath295 as @xmath185 . collecting the expansions , , and
, we obtain @xmath472 uniformly for @xmath295 .
we can now specify all the undetermined constants and @xmath473 such that all terms except the last will vanish and @xmath474 .
this entails first the choices @xmath475 and @xmath476 .
it remains only the @xmath477-term .
we consider the limit when @xmath99 tends to zero using the taylor expansions , and deduce that @xmath478 .
these values give the equation satisfied by @xmath479 @xmath480 which in view of leads to the differential equation @xmath481 thus with the choices @xmath475 , @xmath482 , @xmath483 , and the function @xmath484 by , we get the bound @xmath485 uniformly for @xmath295 . accordingly , by , the sequences @xmath486 satisfy the recurrence @xmath487 with @xmath488 .
choose now the initial value @xmath489 , so that @xmath287 and lemma [ lmm - at ] can be applied .
this implies that @xmath490 , and consequently @xmath353 .
also @xmath461 is indeed given by .
in particular , the first few terms in the taylor expansion of @xmath461 are given as follows .
@xmath491 this completes the proof of the asymptotic expansion for @xmath456 .
the more refined approximation follows the same line of proof but with more detailed expansions .
we show in this section that the distribution of @xmath20 , when properly normalized , tends to a gumbel ( or extreme - value or double exponential ) distribution , as @xmath81 , @xmath164 .
the proof consists in showing that the result when @xmath18 extends to all @xmath164 but requires an additional correction term @xmath169 coming from the linear part of the random variables , which complicates significantly the proof . the standard gumbel distribution @xmath492 ( with mode zero , mean @xmath493 ) is characterized by the distribution function @xmath494 and the moment generating function @xmath495 , respectively .
note that if @xmath496 , then @xmath497 , which was the description used in @xcite . the genesis of the gumbel distribution is easily seen as follows .
let @xmath498 , where the @xmath61 exponential random variables are independent .
then @xmath499 converges in distribution to the gumbel distribution [ lmm - gumbel ] @xmath500 we have @xmath501 uniformly for @xmath47 .
here we used the infinite - product representation of the gamma function @xmath502 the lemma then follows from the asymptotic estimate @xmath503 and curtiss s theorem ( see @xcite ) .
unlike the case when @xmath18 , we need to subtract more terms to have the limit distribution . for @xmath25
, we have the uniform asymptotic approximation [ prop - y - all ] @xmath504 for @xmath47 , where @xmath169 is defined in .
note that @xmath505 as @xmath506 , and thus @xmath507 when @xmath18 . in this case , the proposition re - proves theorem [ thm : yo1 ] ( with an explicit error term ) . a combination of lemma [ lmm - gumbel ] and proposition [ prop - y - all ] leads to the limit law for @xmath20 in the remaining range .
if @xmath508 with @xmath8 and @xmath164 , then [ thm : y - all ] @xmath509 where @xmath169 is defined in . the number @xmath193 of steps used by the ( @xmath1)-ea to reach the final state @xmath21 , when starting from the initial state @xmath510 , satisfies [ thm : y ] @xmath511 from figure [ fig : yn ] , we see the fast convergence of the distribution to the limit law . for @xmath512 , and the limiting gumbel curve.__,width=302 ] [ [ outline - of - proofs
] ] outline of proofs .
+ + + + + + + + + + + + + + + + + + we focus on the proof of proposition [ prop - y - all ] for which we introduce the following normalized function @xmath513 here the probability generating function @xmath514 of @xmath20 satisfies the recurrence and the function @xmath515 is any @xmath516$]-function with @xmath262 ( because in the proof we will require a taylor expansion of order two ) .
it turns out that if we choose @xmath517 , where @xmath169 ( see ) appears as the second - order term in the asymptotic expansion of the mean ( see ) , then @xmath518 uniformly for all @xmath295 , @xmath519 , and @xmath520 , where @xmath521 is independent of @xmath522 .
indeed , our induction proof here does not rely on any information of the mean asymptotics and entails particularly the right choice of @xmath515 .
this is why we specify @xmath188 only at a later stage .
[ [ the - recurrence - satisfied - by - f_nm . ] ] the recurrence satisfied by @xmath523 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + by , @xmath524 satisfies the following recurrence @xmath525 for @xmath526 , with @xmath527 , where @xmath528 . [ [ an - auxiliary - sum . ] ] an auxiliary sum .
+ + + + + + + + + + + + + + + + + since we expect @xmath524 to be close to @xmath118 , we replace all occurrences of @xmath529 on the right - hand side by @xmath118 and consider the following function @xmath530 the following lemma is the crucial step in our proof .
[ lem : gmns_general ] let @xmath515 be a @xmath531-function on the unit interval satisfying @xmath262 .
then @xmath532 where the @xmath68-terms hold uniformly for @xmath295 , and @xmath533 .
the proof consists in a detailed inspection of all factors , using estimates and we derived earlier for @xmath534 .
we consider first the case when @xmath18 . in this case ,
@xmath535 and the numerator and the denominator of both have the form @xmath536 which can be readily checked by using the estimates and @xmath537 from now on , we assume @xmath538 , where @xmath539 is sufficiently large , say @xmath540 . throughout the proof ,
all @xmath68-terms hold uniformly for @xmath47 and @xmath541 and @xmath8 large enough .
we begin with the denominator of @xmath542 , which satisfies @xmath543 where we used the estimate @xmath544 ; see . by and ,
the second - order term on the right - hand side satisfies @xmath545 thus we obtain @xmath546 now we turn to the numerator of @xmath542 and look first at the exponential term @xmath547 uniformly for @xmath251 , where we used the twice continuous differentiability of @xmath188 .
consider now the finite product @xmath548 .
obviously , for @xmath549 , we have the uniform bound @xmath550 on the other hand , we also have the finer estimates @xmath551 uniformly for @xmath552 .
combining these two estimates , we obtain the following approximation @xmath553 which holds uniformly for @xmath554 . thus the numerator , up to the factor @xmath555 , satisfies @xmath556 each of the sums can be readily estimated as in and , and we have @xmath557 similarly , @xmath558 finally , for @xmath283 , @xmath559 collecting these estimates , we get @xmath560 by applying . by , and the simple estimate @xmath561 we conclude .
[ lem : gmns_special ] let @xmath562 .
then @xmath563 where the @xmath68-term holds uniformly for @xmath295 , @xmath8 large enough and @xmath47 . to obtain the error term @xmath564
, we choose @xmath188 in a way that the two middle terms in the fraction of are identical , which means @xmath565 observe that @xmath566 for @xmath52 .
this , together with @xmath401 , implies @xmath263 , which is not only a @xmath567-function but also analytic in the unit circle .
[ [ proof - of - propositionprop - y - all . ] ] proof of proposition [ prop - y - all ] .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we now prove proposition [ prop - y - all ] by induction .
[ lemma : fnms ] let @xmath568 .
then @xmath569 uniformly for @xmath463 , @xmath8 large enough and @xmath570 , @xmath571 .
we use induction on @xmath61 and show that there exists a constant @xmath572 , such that @xmath573 for all @xmath295 , @xmath574 large enough and @xmath575 . when @xmath240 , the lemma holds , since @xmath576 .
assume that the lemma holds for all functions @xmath577 for @xmath578 and @xmath574 .
by corollary [ lem : gmns_special ] , there exists a constant @xmath579 such that for all @xmath580 , @xmath581 large enough and @xmath582 , @xmath583 , @xmath584 now @xmath585 the first term on the right - hand side can be re - written as @xmath586 since we assume @xmath587 , the product involved in the sum on the right - hand side is nonnegative and we have , by the induction hypothesis , @xmath588 it follows that @xmath589 choose first @xmath590 such that @xmath591 , which implies that @xmath592 for @xmath295 and @xmath593 .
then choose @xmath594 .
we then have @xmath595 and thus @xmath596 note that apart from requiring @xmath587 the only restriction on @xmath597 comes from @xmath542 , thus we may choose @xmath598 .
this completes the proof .
[ [ the - gumbel - limit - laws - for - x_nm - mtoinfty . ] ] the gumbel limit laws for @xmath20 ( @xmath81 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we prove theorem [ thm : y - all ] by proposition [ prop - y - all ] . since @xmath508 , we have @xmath599 thus theorem [ thm : y - all ] follows from another application of curtiss s theorem ( see @xcite ) . [ [ the - gumbel - limit - law - for - x_n . ] ] the gumbel limit law for @xmath193 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we now prove theorem [ thm : y ] , starting from the moment generating function ( @xmath600 ) @xmath601 then @xmath602 where @xmath603 since the binomial distribution is highly concentrated around the range @xmath604 where @xmath419 , we see that @xmath605 for @xmath61 in this central range . by a standard argument ( gaussian approximation of the binomial and exponential tail estimates ) using the expansion
, we then deduce that @xmath606 this proves theorem [ thm : y ] .
we consider the complexity of the ( @xmath1)-ea when the underlying fitness function is the number of leading ones .
this problem has been examined repeatedly in the literature due to the simple structures it exhibits ; see @xcite and the references therein .
the strongest results obtained were those by ladret @xcite ( almost unknown in the ea literature ) where she proved that the optimization time under leadingones is asymptotically normally distributed with mean asymptotic to @xmath607 and variance to @xmath608 , where @xmath609 , @xmath105 .
we re - visit this problem and obtain similar type of results by a completely different approach , which can be readily amended for obtaining the convergence rate . throughout this section
, the probability @xmath5 still carries the same meaning from algorithm @xmath0-ea and @xmath610 .
let @xmath611 denote the conditional optimization time when beginning with a random input ( each bit being 1 with probability @xmath612 ) that has @xmath613 leading ones .
then the moment generating function @xmath614 of @xmath611 satisfies the recurrence relation @xmath615 for @xmath25 , where @xmath610 .
the probability of jumping from a state with @xmath613 leading ones to another state with @xmath616 leading ones is given by @xmath617 which corresponds to the situation when the first @xmath613 bits do not toggle their values , the @xmath618st bit toggles ( from @xmath117 to @xmath118 ) , together with the following @xmath619 bits also being 1 .
when @xmath620 , the probability becomes @xmath621 we thus obtain the recurrence relation @xmath622 which implies .
the most interesting case is when @xmath623 ( roughly , @xmath624 is linear , giving rise to polynomial bounds for the cost ) , all other cases when @xmath625 lead to higher - order complexity .
[ [ small - m . ] ] small @xmath61 .
+ + + + + + + + + + + + + + + + + + + + + + we start with the simplest case when @xmath39 and obtain , by , @xmath626 then the mean of @xmath627 is simply given by @xmath628 which , by substituting @xmath609 , yields @xmath629 note that this estimate holds as long as @xmath630 .
similarly , the variance is given by @xmath631 which satisfies , when @xmath609 , @xmath632 uniformly when @xmath630 .
we then consider the normalized random variables @xmath633 . by the expansion @xmath634 uniformly when @xmath597 is away from @xmath118 and @xmath635 , we obtain @xmath636 implying that the limit law is an exponential distribution with the density @xmath637 . while shows that @xmath638 when @xmath14 , they behave differently when @xmath639 . for each @xmath640 ,
the limit distribution of @xmath641 is a binomial mixture of gamma distributions ; more precisely , [ thm : xo1 ] @xmath642 as @xmath46 .
the mean and the variance satisfy @xmath643 note that when @xmath39 , degenerates to the exponential distribution . on the other hand ,
the normalizing factor @xmath644 is not asymptotically equivalent to the mean . by induction and
, we see that @xmath645 when @xmath18 .
since @xmath646 we then deduce . the mean and
the variance then follows from straightforward calculations .
[ [ mean - and - the - variance - of - y_nm-1le - mle - n . ] ] mean and the variance of @xmath611 : @xmath25 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the recurrence is much simpler than and we can indeed obtain very precise expressions and approximations for the mean and the variance .
the mean @xmath647 and the variance @xmath648 of @xmath611 are given explicitly as follows .
for @xmath25 [ thm : xmv ] @xmath649 and @xmath650 with these closed - form expressions , we easily obtain , assuming @xmath609 , where @xmath105 , @xmath651 and @xmath652 uniformly for @xmath653 .
we see that the asymptotic equivalents indeed hold in the wider range @xmath653 .
more uniform approximations have the following forms .
assume that @xmath609 , where @xmath630 .
then , uniformly for @xmath654 , @xmath655 and @xmath656 [ [ proof - of - theoremthmxmv . ] ] proof of theorem [ thm : xmv ]
. + + + + + + + + + + + + + + + + + + + + + + + + + + + our approach is based on and it turns out that all moments satisfy the same simple recurrence of the following type . the solution to the recurrence relation @xmath657 is given by the closed - form expression @xmath658 the corresponding generating functions @xmath659 and @xmath660 satisfy the equation @xmath661 or @xmath662 this proves . from ( by taking derivative with respect to @xmath597 and then substituting @xmath663 ) , we see that the mean @xmath647 satisfies the recurrence @xmath664 substituting @xmath665 into , we obtain .
similarly , for the second moment @xmath666 , we have the recurrence @xmath667 by the same procedure , we obtain @xmath668 implying .
this proves theorem [ thm : xmv ] .
the proofs of the two corollaries are straightforward and omitted .
[ [ a - finite - product - representation - for - q_nms . ] ] a finite - product representation for @xmath614 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the recurrence relation can indeed be solved explicitly as follows .
the moment generating function @xmath614 of @xmath611 has the closed - form @xmath669 for @xmath670 .
let @xmath671 .
we start with the recurrence ( from ) @xmath672 which , by , has the alternative form @xmath673 from , we see that the bivariate generating function @xmath674 of @xmath614 satisfies @xmath675 which implies the simpler functional equation @xmath676 multiplying both sides by @xmath677 gives @xmath678 implying the relation @xmath679 accordingly , we obtain the closed - form expression .
let @xmath680 denote the probability generating function of a geometric distribution @xmath681 with parameter @xmath682 and support @xmath683 .
the random variables @xmath611 can be decomposed as the sum of @xmath61 independent random variables @xmath684 } + \cdots + z_{n , m}^{[m-1]},\end{aligned}\ ] ] where @xmath685 } \sim \text{geo}(pq^{n - m})$ ] and the @xmath686}$ ] are mixture of @xmath687 @xmath688}}\right ) = \frac12\cdot\frac{1-(1 - 2pq^{n - j})t}{1-(1-pq^{n - j})t } = \frac12 + \frac{r_j(t)}2 \qquad(j=1,\dots , m-1).\ ] ] thus the mean of @xmath611 is given by @xmath689}\right ) = \frac1{pq^{n - m}}+ \frac12\sum_{1\le j < m } \frac1{pq^{n - j}},\ ] ] which is identical to .
similarly , the variance of @xmath611 satisfies @xmath690}\right ) = \frac{3 - 2pq^{n - m}}{8(pq^{n - m})^2}+ \frac12\sum_{1\le j < m } \frac{1-pq^{n - j}}{(pq^{n - j})^2},\ ] ] which is also identical to .
the distributions of @xmath691 are asymptotically normal [ thm : x - all ] @xmath692 uniformly as @xmath81 ( with @xmath8 ) and @xmath164 , where @xmath693 denotes the standard normal distribution function . again from the decomposition , we derive the following expression for the third central moment @xmath694 similarly , the fourth central moment satisfies @xmath695 which implies that @xmath696 uniformly for @xmath25 .
we then deduce a central limit theorem by , say lyapounov s condition , or by levy s continuity theorem ; see , for example , @xcite .
we can indeed derive an optimal berry - esseen bound by more refined fourier argument , details being omitted here .
in particular , we have @xmath697 when @xmath81 and @xmath164 .
this will be needed later .
[ [ random - input . ] ] random input .
+ + + + + + + + + + + + + now consider the cost @xmath698 used by algorithm @xmath0-ea when starting from a random input ( each bit being @xmath118 with probability @xmath612 ) .
then its moment generating function satisfies @xmath699 the random variables @xmath698 are asymptotically normally distributed @xmath700 with mean @xmath701 and variance @xmath702 asymptotic to @xmath703 respectively .
in particular , we also have , by replacing the exact mean and variance by the corresponding asymptotic approximations @xmath704 by , we have @xmath705 and then the first estimate in follows . similarly , by , @xmath706 and
the second estimate in also follows . for the asymptotic normality , we consider the characteristic function @xmath707 we split the sum into two parts : @xmath708 and @xmath709 .
observe that when @xmath710 , we have the uniform estimate @xmath711 by and .
we then have the local expansion ( see ) @xmath712 thus @xmath713 uniformly in @xmath61 .
consequently , @xmath714 the remaining part is negligible since @xmath715 and @xmath716 we conclude that @xmath717 which implies the convergence in distribution of @xmath718 to the standard normal distribution
. 99 a. auger and b. doerr ( eds . ) , _ theory of randomized search heuristics : foundations and recent developments _ , world scientific , 2011 .
t. bck , the interaction of mutation rate , selection , and self - adaptation within a genetic algorithm , in proceeding of the 2nd conference on parallel problem solving from nature ( ppsn - ii ) , brussels , belgium , september 2830 , 1992 .
beyer , h .-
schwefel and i. wegener , how to analyse evolutionary algorithms , _ theoretical computer science _ , * 287*:1 ( 2002 ) , 101130 .
s. bttcher , b. doerr and f. neumann , optimal fixed and adaptive mutation rates for the leadingones problem , _ ppsn xi _
, part i , 2010 , lncs 6238 , 110 . c. a. coello coello ,
evolutionary multiobjective optimization : a historical view of the field , _ ieee computational intelligence magazine _ , * 1*:1 ( 2006 ) , 2836 .
l. comtet , _ advanced combinatorics , the art of finite and infinite expansions _ ,
revised and enlarged edition , d. reidel publishing co. , dordrecht , 1974 .
k. deb , _ multi - objective optimization using evolutionary algorithms _ , john wiley & sons , 2001 .
b. doerr , m. fouz and c. witt , quasirandom evolutionary algorithms , in _ proc .
gecco10 _ ( 2010 ) , pp.14571464 , acm press .
b. doerr , m. fouz and c. witt , sharp bounds by probability - generating functions and variable drift , in _ proc .
gecco11 _ ( 2011 ) , pp .
20832090 .
b. doerr , d. johannsen and c. winzen , multiplicative drift analysis , _ algorithmica _ , * 64 * ( 2012 ) , 673697 . b. doerr , d. johannsen , t. kotzing , f. neumann , and m. theile , more effective crossover operators for the all - pairs shortest path problem , _ theoretical computer science _ * 471 * , ( 2013 ) 1226 .
b. doerr and s. pohl , run - time analysis of the ( @xmath719 ) evolutionary algorithm optimizing linear functions over a finite alphabet , in _ proc . gecco12 _ ( 2012 ) ,
13171324 .
s. droste , t. jansen and i. wegener , a rigorous complexity analysis of the ( @xmath16 ) evolutionary algorithm for linear functions with boolean inputs , in _ proc .
ieee international conference on evolutionary computation icec98 _ ( ieee press , piscataway , nj , 1998 ) , 499504 .
s. droste , t. jansen and i. wegener , on the analysis of the ( @xmath1 ) evolutionary algorithm , _ theoretical computer science _ , * 276*:1 - 2 ( 2002 ) , 5181 .
p. flajolet and r. sedgewick , _ analytic combinatorics _ , cambridge university press , cambridge , 2009 .
j. garnier , l. kallel and m. schoenauer , rigorous hitting times for binary mutations , _ evolutionary computation _ , * 7*:2 ( 1999 ) 167203 .
j. he and x. yao , drift analysis and average time complexity of evolutionary algorithms , _ artificial intelligence _
, * 127 * ( 2001 ) , 5785 . j. he and x. yao , from an individual to a population : an analysis of the first hitting time of population - based evolutionary algorithms , _ ieee transactions on evolutionary computation _ , * 6*:5 ( 2002 ) , 495511 .
r. v. hogg , j. w. mckean and a. t. craig , _ introduction to mathematical statistics _ , 7th edition , pearson education inc .
, ny , 2013 . j. horn , multicriterion decision making , _ handbook of evolutionary computation _ , volume 1 , f1.9:1f1.9:15 , iop publishing ltd . and oxford university press , 1997 .
j. horn , d. e. goldberg and k. deb , long path problems , in _ parallel problem solving from nature ppsn iii _ , edited by y. davidor , h .-
schwefel , r. mnner , _ lecture notes in computer science _ , vol .
866 , springer , berlin , 1994 , pp .
149158 .
hwang , on convergence rates of central limit theorems for combinatorial structures , _
european journal of combinatorics _ , * 19 * ( 1998 ) , 329343 j. jagerskupper , combining markov - chain analysis and drift analysis : the ( @xmath1 ) evolutionary algorithm on linear functions reloaded , _ algorithmica _ * 59 * ( 2011 ) , 409424 .
t. jansen and i. wegener , the analysis of evolutionary algorithms a proof that crossover really can help , _ algorithmica _ , * 34 * ( 2002 ) , 4766 .
t. jansen and i. wegener , real royal road functions where crossover provably is essential , _
discrete applied mathematics _ , * 149 * ( 2005 ) , 111125 .
v. ladret , asymptotic hitting time for a simple evolutionary model of protein folding , _
_ , * 42 * ( 2005 ) , 3951 .
h. mhlenbein , how genetic algorithms really work : mutation and hill - climbing , in _ proc .
ppsn92 _ , 1526 , 1992 .
f. neumann and c. witt ( 2009 ) , runtime analysis of a simple ant colony optimization algorithm , _ algorithmica _ , * 54 * ( 2009 ) 243255 .
f. neumann and c. witt , _ bioinspired computation in combinatorial optimization algorithms and their computational complexity _ , springer , 2010 .
v. v. petrov , _ sums of independent random variables _ , springer - verlag , new york - heidelberg , 1975 ( translated from the russian by a. a. brown ) .
g. rudolph , how mutation and selection solve long - path problem in polynomial expected time , _ evolutionary computation _ * 4*:3 ( 1997 ) , 195205 .
d. sudholt , general lower bounds for the running time of evolutionary algorithms , in _ proc .
ppsn10 _ ,
lecture notes in computer science , volume 6238 , pp .
124133 ( 2010 ) .
d. sudholt , c. witt ( 2010 ) , runtime analysis of a binary particle swarm optimizer , _ theoretical computer science _ , * 411 * ( 2010 ) 20842100 . c. witt , runtime analysis of the ( @xmath720 ) ea on simple pseudo - boolean functions , _ evolutionary computation _ ,
* 14 * ( 2006 ) , 6586 . c. witt , tight bounds on the optimization time of a randomized search heuristic on linear functions , _ combinatorics , probability and computing _ , * 22*:2 ( 2013 ) , 294318
we collected here some interesting expressions for @xmath155 . we begin with proving that all @xmath131 can be expressed in terms of @xmath721 and the two modified bessel functions @xmath722 the starting point is the obvious relation ( @xmath723 ) @xmath724 applying the integral representation and integration by parts
, we have @xmath725 by the same argument used for corollary [ cor : ur ] , we deduce the recurrence @xmath726 for @xmath727 with @xmath728 a closed - form expression can be obtained for the recurrence but it is very messy .
more precisely , let @xmath729 .
then @xmath730 satisfies the first - order differential equation @xmath731 the solution to the differential equation with the initial condition @xmath732 is given by @xmath733 this implies that @xmath159 has the general form @xmath734}(\alpha ) { \accentset{\rule{.4em}{1pt}}{i}}_0(\alpha ) + p_r^{[1]}(\alpha ) { \accentset{\rule{.4em}{1pt}}{i}}_1(\alpha)+p_r^{[2]}(\alpha ) s_0(\alpha ) \qquad(r\ge1),\ ] ] where the @xmath735}$ ] are polynomials of @xmath99 of degree @xmath736 .
closed - form expressions can be derived but are less simpler than the recurrence for small values of @xmath736 . on the other hand
, the same argument also leads to @xmath737 in particular , @xmath738 .
note that @xmath739 implying that @xmath740 this in turn gives @xmath741 this expression can be further simplified by taking second derivative with respect to @xmath99 of the integral representation @xmath742 giving @xmath743 which implies that ( with @xmath744 ) @xmath745 similarly , since @xmath746 , we have @xmath747 these expressions show not only the intimate connections of @xmath131 to bessel functions but also their rich algebraic aspects .
we now consider @xmath748 . by the same integral representation and a change of variables
, we see that , for @xmath143 , @xmath749 e_r(1-z ) e^{\frac{\alpha}{1-z } + ( 1-\alpha)(1-z)}.\ ] ] now @xmath750 \frac{e^w}{1-(1-z)e^w } = \sum_{0\le j\le r } ( -1)^{r+j}j!\ , \text{stirling}_2(r , j ) z^{-j-1}.\ ] ] thus we deduce the identity ( for @xmath143 ) @xmath751 or @xmath752 note that for @xmath138 @xmath753 in particular , this gives @xmath754 . for @xmath143 @xmath755
the successive improvements attained by adding more terms on the right - hand side of can be viewed in figures [ app : fig - mu ] and [ app : fig - mu2 ] . for @xmath25 and @xmath756 ;
right : the difference between @xmath757 for @xmath758 in the same ranges.__,title="fig:",width=226 ] for @xmath25 and @xmath756 ; right : the difference between @xmath757 for @xmath758 in the same ranges.__,title="fig:",width=226 ] ( left ) and @xmath759 ( right ) for @xmath25 and @xmath756.__,title="fig:",width=226 ] ( left ) and @xmath759 ( right ) for @xmath25 and @xmath756.__,title="fig:",width=226 ]
recall that @xmath760 this sequence satisfies @xmath452 , and for @xmath295 , @xmath761 where @xmath762 from this recurrence , we obtain the following expansions . @xmath763 observe that the leading constant terms are exactly given by @xmath764 these expansions suggest the general form @xmath765 with this form using the technique of matched asymptotics , we are then led to the following explicit expressions . @xmath766 the above expansions for small @xmath61 suggest the more uniform asymptotic expansion for @xmath456 for @xmath25 @xmath767 in the sense that when omitting all terms with indices @xmath768 introduces an error of order @xmath769 ; furthermore , the expansion holds uniformly for @xmath770 .
we elaborate this approach by carrying out the required calculations up to @xmath771 , which then characterizes particularly the constant @xmath772 and the function @xmath773 .
we start with the formal expansion and expand in recurrence all terms for large @xmath774 in decreasing powers of @xmath8 ; we then match the coefficients of @xmath775 on both sides for each @xmath776 . to specify the initial condition @xmath777 we incorporate the information from the asymptotic expansion for @xmath778 ( obtained by exact solution ) . with this algorithmic approach it is possible to determine the coefficients @xmath779 and @xmath780 and the functions @xmath781 successively one after another .
we remark that a formalization of this procedure at the generating function level as carried out for the expectation in section [ sec : mu - ae ] could be given also , but here we do not pursue this any further .
we use the expansions @xmath782 as well as those for @xmath184 and @xmath783 in and , respectively .
the expansion of the right - hand side of then starts as follows .
@xmath784 where @xmath785 for the left - hand side of , the asymptotic form leads to @xmath786 where @xmath787 observe that all functions @xmath788 , @xmath789 have a simple pole at @xmath790 .
we match the terms in the expansion and consider @xmath791 .
first we compare the first two terms of the laurent expansions of both functions . using , we get @xmath792 and by matching the two constant terms , we see that @xmath793 .
the equation @xmath794 characterizes then the function @xmath484 of the form @xmath795 next we consider @xmath796 and obtain @xmath797 and thus , by matching the terms and using the values already computed in the first - order approximation for @xmath456 , @xmath798 and @xmath799 .
then the function @xmath800 can be characterized by equating @xmath796 , which then gives @xmath801 all constants and functions here match with those obtained earlier in previous paragraphs , and we can pursue the same calculations further and obtain finer approximations .
for example , we have @xmath802 .
but the calculations are long and laborious .
finally , it remains to determine the constant terms in the taylor expansion of the functions @xmath781 by adjusting them to the expansion of @xmath456 for small @xmath61 .
this yields @xmath803 , and @xmath804 this characterizes the function @xmath773 in theorem [ thm : var1 ] as follows .
@xmath805 \dd t.\end{aligned}\ ] ] in particular , the first few terms in the taylor expansion of @xmath773 are given by @xmath806
for @xmath341 and @xmath756 ; right : the difference between @xmath807 for @xmath758 in the same ranges.__,title="fig:",width=226 ] for @xmath341 and @xmath756 ; right : the difference between @xmath807 for @xmath758 in the same ranges.__,title="fig:",width=226 ] | we give a detailed analysis of the cost used by the @xmath0-evolutionary algorithm .
the problem has been approached in the evolutionary algorithm literature under various views , formulation and degree of rigor .
our asymptotic approximations for the mean and the variance represent the strongest of their kind .
the approach we develop is also applicable to characterize the limit laws and is based on asymptotic resolution of the underlying recurrence .
while most approximations have their simple formal nature , we elaborate on the delicate error analysis required for rigorous justifications . |
due to their layered structure , transition - metal dichalcogenides ( tmdcs ) show a variety of anisotropic properties , such as thermal expansion , sound velocity and thermal as well as electrical conductivity @xcite .
the lowered dimensionality often also manifests itself in the occurrence of a charge - density wave ( cdw ) phase transition within the @xmath2-plane of the crystal . in the present example of
, this cdw superstructure is incommensurate below and becomes commensurate going below @xcite .
a detailed study of the phase transition containing an incommensurate superstructure followed by a lock - in transition to a commensurate ordering vector has for example been realised by neutron scattering @xcite or x - ray diffraction @xcite .
this transition as well as the cdw state itself were also investigated by angle - resolved photoemission measurements @xcite , clearly showing the opening of a gap at the fermi level when going to the ordered phase . however , the reasoning that the fermi surface is partially gapped already for temperatures far above the phase transition and the link of this fact to a possible nesting mechanism for the cdw order is supported by reflectivity as well as resistivity measurements @xcite .
recent spectroscopic investigations on showed results seemingly contradicting the conventional description of a metal
apart from the observation of the partial gapping
namely , a negative slope of the plasmon energy dispersion , which can be assigned to an interplay of the charge density fluctuations and the plasma resonance @xcite . while the @xmath2-planes of the crystals of the tmdcs consist of hexagonal layers , the @xmath0-direction is formed by sandwich - like structures of these layers , bound to each other through
_ _ van - der - waals-__interaction .
the gaps between the building blocks are the typical place for all kinds of intercalates to arrange within the crystal @xcite .
thereby , the range of possible doping substances is as wide as the spectrum of physical properties to tune with intercalation . from the electronic point of view
it is reported that the band structure changes extensively upon alkali metal ( na with na / ta@xmath3 and cs with cs / ta varying between 0.3 and 0.6 ) doping in @xcite and related compounds , such as @xmath4-rb@xmath5tase@xmath6 @xcite and @xmath4-cs@xmath5tase@xmath6 ( up to @xmath7 ) @xcite .
structural changes upon intercalation can be observed in various polytypes of the tmdcs . primarily the @xmath0-axis is widened , while the spacing within the @xmath2-plane remains nearly unchanged @xcite .
slight changes in the in - plane crystal structure indicated by a change of super structure are observed for different compounds such as the @xmath4-cs@xmath5tase@xmath6 mentioned before @xcite as well as in k@xmath5tis@xmath6 @xcite .
because of the suppression of the cdw , as for example in @xmath1-k@xmath5tas@xmath6 @xcite , the observed super structures can be assumed to result from the intercalated atoms .
structural changes are also discussed in terms of staging effects in the class of silver doped tmdcs in close analogy to intercalated graphite compounds .
the added silver atoms are supposed to slightly disarrange the chalcogenide atoms and with it opening new scattering channels .
the observed stages are at doping rates of one and two thirds , respectively , for ag@xmath5tas@xmath6 as well as ag@xmath5tis@xmath6 @xcite .
other examples also show super structure changes in ag@xmath5nbs@xmath6 @xcite .
we report on the _ in situ _
intercalation into the bulk sample of with the alkali metal potassium , reaching a comparatively high stable doping rate .
electron energy - loss spectroscopy is applied to investigate changes in the electronic as well as crystal structure .
we demonstrate that doping levels up to k@xmath8tase@xmath6 can be reached which are characterized by well defined plasmon excitations , while the cdw superstructure is suppressed upon doping .
single crystals of were grown from ta metal and se ( and purity , respectively ) by iodine vapor transport in a gradient of , the crystals growing in the cooler end of the sealed quartz tubes .
a very slight excess of se was included ( typically of the charge ) to ensure stoichiometry in the resulting crystals .
each experimental run lasted for .
this procedure yielded single crystals with maximum size of @xmath9 .
appropriate thin films ( thickness of about ) were prepared either by the use of an ultramicrotome or by cleaving with adhesive tape , whereby the crystal was repeatedly cleaved and the glue layer dissolved afterwards .
the films were mounted onto standard electron microscopy grids and transferred to the spectrometer .
as depicted in fig.[fig : superstruct ] , a very good sample quality is achieved , since discrete bragg reflections as well as clear superstructures of the cdw arising at the predicted transition temperature ( ) are visible .
potassium intercalation was achieved by evaporation from commercial saes ( saes getters s.p.a .
, italy ) getter sources under ultra high vacuum conditions ( base pressure below ) .
the intercalation levels were achieved by several discrete intercalation steps , of which each took about .
the relation between intercalation time and corresponding increase in potassium content could not be shown to be linear due to several reasons , such as varying getter quality and inhomogeneous saturation of the sample .
furthermore , the saturation potassium content was found to be stable over a time span of several months at room temperature and ultra - high vacuum conditions .
all measurements were carried out with an electron energy - loss spectrometer equipped with a he flow cryostat .
the resolution of the spectrometer is and for momentum and energy - loss , respectively .
the spectrometer setup can be found in ref.@xcite .
for the present studies , the access to elastic scattering processes by measuring at zero energy - loss is of great importance . showing a red shift of the plasma resonance for increasing potassium content ( left panel ) .
the right panel depicts the plasmon energy as well as the width of the plasmon peak , corresponding to the left panel , with respect to potassium content @xmath10 per tase@xmath6 unit as estimated with the help of lda calculations ( cf .
fig.[fig : theoretical - results :- a:]c).,scaledwidth=80.0% ] to gauge and analyse the experimental results we performed density functional theory ( dft ) calculations within the local - density - approximation ( lda ) .
we used the scalar relativistic mode of the full potential local orbital code ( fplo)@xcite in version 9.09 . in order to model the intercalation we used two approaches . in the virtual crystal approximation ( vca ) k@xmath11-layers are inserted between adjacent tase@xmath12 layers , where @xmath13 denotes the number of protons and electrons on the vca - pottasium atom , which models the vca equivalent of k@xmath14tase@xmath12 . in this way
the filling effect of the tase bands upon doping can be described . for judging the quality of vca and for incorporating more realistic fermi surface reconstruction effects we constructed several supercells ( sc ) with a corresponding k / ta ratio .
guided by the experimentally observed superstructure patterns we focus on simple supercells with at least a threefold axis .
for @xmath15 we also considered an orthorhombic cell . especially for small concentrations
it is not obvious that pottasium intercalates evenly into all interlayers . to cover such effects two series of supercells were created ,
one with potassium distributed as evenly as possible and one , where only every second interlayer contains the dopant ( for @xmath16 ) .
details about the superstructures are found in the supplementary material .
we did not try to determine the true structural ground state , since this could become rather involved and the purpose of the supercells is merely to supplement and check the vca calculations . for all vca and sc calculations we relaxed all lattice and wyckoff parameters to accomodate breething of the lattice due to k - intercalation .
the @xmath17-integration for the self consistent calculations was performed via the linear tetrahedron method with a @xmath18 mesh for the simple vca k@xmath19ta@xmath12se@xmath20 unit cell . for the supercell
calculations the number of mesh points was adjusted accordingly to obtain an equivalent mesh density .
optical properties were calculated with a @xmath21 mesh within the whole brillouin zone , which was checked for convergence . for the undoped compound we performed full - relativistic calculations to check the influence of spin orbit coupling on the plasmon frequencies .
it turns out that the spin - orbit - splitting of the bands at the fermi level is quite sizable ( ) but the splittings ( which reduce the @xmath22-dispersion ) occur in such a way that their effects on the @xmath23-plane plasmon energies average out over the brillouin zone .
the resulting changes of in - plane @xmath24 are negligible , while the out - of - plane plasmon energy gets reduced .
we only focus on the in - plane plasmon in this work and hence used scalar relativistic calculations throughout this work .
figure[fig : pot_loss ] ( left panel ) shows loss spectra of @xmath1-k@xmath5tase@xmath6 for increasing potassium content @xmath10 .
all spectra are normalized in the region around and are furthermore corrected by fitting and subtracting the elastic line , contributing at around zero energy - loss @xcite . the undoped sample shows a dominating peak at around that , with increasing potassium content , shifts to lower energies .
furthermore , the normalized peak maximum as well as the peak width are reduced .
this spectral feature represents the charge carrier plasmon of the investigated samples , and therefore contains information on the charge carrier density .
the reduction of the plasmon energy is a direct consequence of the charge carrier density variation induced by the potassium doping .
undoped is characterized by half filled conduction bands , and addition of either electrons or holes will lower the plasmon energy as has been discussed previously based on rigid band considerations @xcite . in general ,
the plasmon energy therefore can be also used to analyze the doping level . since the band structure of is rather complex and variations beyond a rigid band approach
can not be excluded , we compared calculated screened plasmon energies ( fig.[fig : theoretical - results :- a : ] , below ) to the measured ones and deduced the @xmath10-axis in the right panel of fig.[fig : pot_loss ] .
we used the thus defined concentrations in the rest of the paper .
note , however , that the calculated and measured screened plasmon energies are remarkably similar , which makes such a gauging procedure feasible ( see also discussion below ) .
it can be concluded that saturation of potassium doping is likely to be reached at about @xmath1-k@xmath8tase@xmath6 .
older reports on intercalation of transition - metal dichalcogenides were able to find also higher dopant rates up to @xmath25 .
@xcite however , rouxel @xcite also emphasized that a homogeneous structure in the selenides could only be obtained for rates of @xmath26 to @xmath27 .
furthermore , the spectral width of charge carrier plasmons in most cases is predominantly determined by damping via interband transitions @xcite . as a consequence ,
we attribute the reduction of the plasmon width upon doping mainly to its shift to lower energies , where less interband transitions are available for a damping of the plasmon excitation .
in addition , mid - infrared transitions as observed for undoped are suppressed upon doping @xcite , which could also reduce the plasmon damping .
-plane with increasing potassium content from top left to bottom right showing cdw superstructures for low temperatures and low k contents and arising dopant superstructures for increasing k contents . ] in fig.[fig : struct_doped_30020 ] intensity maps , achieved by elastic scattering in eels , are shown .
they show congruent sections of the @xmath2-plane with increasing potassium content from top ( @xmath28 ) to bottom ( @xmath29 ) at a temperature of .
the bright spots in the highlighted @xmath30 and @xmath31 direction ( [ 100 ] and [ 110 ] , respectively ) represent the basic structure of the compound .
additional reflections at one - third and two - third of the [ 100 ] vector are observable in the undoped phase ( topmost left part of fig.[fig : struct_doped_30020 ] ) , which can be assigned to the occurring cdw at temperatures below ( see also fig.[fig : superstruct ] ) .
@xcite -k@xmath8tase@xmath6 ) around the [ 110 ] zero reflection ( left panel ) showing the suppressed intensity for increasing angles of the scattering direction with respect to the @xmath2-plane . the right panel shows the according [ 11@xmath32 reflections ( see text ) , that emerge with increasing angles , allowing a calculation of the @xmath0-axis length , which is widened compared to the undoped crystal structure . ] during the successive doping process several changes in the scattering image can be observed .
the most prominent superstructure is a hexagon like ring at around doping ( bottom left part of fig.[fig : struct_doped_30020 ] ) . upon increasing potassium addition , in this state
the sample first shows a mixed superstructure of the cdw spots at one - third and two - third as well as new superstructures occuring at one - half ( for @xmath33 ) and one and three quarter ( for @xmath34 ) , respectively .
however , the latter doping state is no longer showing the original triple cdw superstructure . at the highest possible doping level ( for @xmath29 ) again a ring like structure around the [ 100 ] reflection can be found .
these findings lead to the conclusion that in the cdw is suppressed by k intercalation , as it was also found for the related compound @xmath1-tas@xmath6 @xcite . at the same time new features arise that must be assigned to an increasing potassium filling of the crystal structure . at this point
a comparison between alkali metal intercalated graphite and tmdcs shows striking similarities .
both material classes show a van - der - waals coupled hexagonal layer structure and metallic behavior in their room temperature phase @xcite .
furthermore , a transition to a superconducting phase could be induced or supported , respectively , by means of alkali metal doping @xcite . upon intercalation , most commonly rb or k , graphite is able to form distinct phases , which can be distinguished as stage 1 and 2 ( c@xmath35x and c@xmath36x , respectively , with x being the alkaline metal ) . with increasing content of the dopant it is showing a diversity of superstructures that can be assigned to the intercalant @xcite .
however , the formation of distinct phases present in intercalated graphite could not be observed in .
in particular the continuous variations in line shape and position of the charge carrier plasmon do not signal the simultaneous presence of two or more phases at any doping level .
nevertheless , the appearance of new superstructure peaks at one - half of the basic unit length of would be consistent with the potassium arrangement for a doping rate of @xmath37 as used in our calculations ( see supplementary informations fig.2 left ) .
our theoretically assumed structures ( see supplementary information ) would also be in agreement with these diffraction peaks to persist up to relatively high doping levels . at very high doping the doping induced superstructure would then lead to diffraction peaks at one - third and two - third of the original positions in agreement with fig.[fig : struct_doped_30020 ] ( lower right panel ) .
a quantitative analysis of our diffraction data is not possible because of multiple scattering effects .
since the momentum is transfered within the sample plane @xcite , it is not possible in the present case to directly access the @xmath0-direction of the crystal . with the possibility to rotate the sample around an axis perpendicular to the beam direction , measurements are possible in a range of with respect to this axis .
figure[fig : polar_angle ] shows elastic scattering measurements ( fully doped sample ) in a plane rotated relative to the sample plane by the denoted angle .
while the left panel shows a clear decrease of the [ 110]-peak intensity ( around ) with increasing angles , the right panel shows rising features at higher momentum transfers , corresponding to reflections with a finite @xmath38 component , such as representing [ 114 ] , [115 ] and [116 ] .
the corresponding @xmath38 component of around is in agreement with the value of an earlier x - ray investigation of doped @xcite , where this value was found for a maximum doping rate of @xmath39 , which is in very good agreement to our stoichiometry analysis ( see below ) . from the interband contributions ;
b ) the @xmath0-axis lattice parameter ; c ) the plasmon peak position of the calculated loss function , the solid line shows a fit of the supercell frequencies , used for fixing the potassium content of the measured sample ; d ) the bare plasmon frequency .
_ es _ denotes cells with alternatingly empty interlayers . ] in order to quantitatively determine substantial structural changes , the behavior of the lattice parameters in the lda calculation can be summarized by noting that the planar lattice constant stays rather constant upon doping , while the @xmath0-axis expands with k - intercalation .
figure [ fig : theoretical - results :- a:]b ( black solid symbols ) shows that in the supercell calculations there is a rather sudden widening of the lattice even on small amount of added potasium , due to the rigid ion cores , which need some space and due to the rigidity of the covalently bound tase@xmath12 prisms . for the cells , where only every second interlayer has k the @xmath0-axis is smaller ( light(red ) filled symbols ) , which follows from the same reasoning . with increasing k - content the @xmath0-axis shrinks because of the increasing electrostatic interaction provided by an increasing electron content in the interlayer .
the behavior of vca and sc coincide for larger concentrations .
however , towards smaller @xmath10 the @xmath0 lattice parameter of vca ( open black and light(red ) symbols ) rises above the sc result , which can be understood from the fact that vca contains the k@xmath40-ion cores at a 100@xmath41 filling rate for all concentrations , which keeps the tase@xmath12 layers appart , while for supercells all k - atoms vanish for @xmath42 .
this larger @xmath0-axis spacing however seems to mostly influence the plasma frequency along the @xmath22-direction , which is not considered in this investigation .
however , the @xmath0-axis value of around for @xmath43 and with it the corresponding @xmath38 value ( ) are slightly higher than the ones measured and from the literature @xcite , respectively . the planar bare plasmon frequencies of vca and sc calculations fig.[fig : theoretical - results :- a:]d form a consistent trend to lower energy values for increasing potassium content .
we emphasize that for undoped the agreement between our measurements and the calculations is very good .
interestingly , the plasmon frequencies jump , when going from undoped to slightly doped samples , which we attribute to the immediate lattice widening upon small amounts of intercalated potassium . in order to gain some insight into the screening due to interband transitions and with it the link to the measured plasma frequencies we also calculated the real part of the dielectric function in dependence on @xmath10 .
in general the real part tends to flatten out into a quasi constant ( @xmath44 ) for @xmath45ev .
unfortunately , there are some oscillations for small energies due to a low energy feature in @xmath46 , whose strength is concentration dependent .
this makes a clear extraction of @xmath44 rather approximate .
fig.[fig : theoretical - results :- a:]a shows the trend of @xmath44 as a function of @xmath10 .
the screening is of the same order as extracted from experiment and also exhibits the sudden jump , upon the doping - onset .
since @xmath44 is not a constant and since the real part has low energy features we used the theoretically calculated loss function to extract a screened plasmon peak position ( fig.[fig : theoretical - results :-
a:]c ) and to estimate the k - content of the samples via the measured plasmons peak position .
the values obtained in this manner are used throughout this contribution .
to summarize , we performed electron energy - loss spectroscopy as well as lda calculations on the transition - metal dichalcogenide and thereby investigated the influence of in situ intercalation of potassium into the sample . we demonstrate that our intercalation procedure can be used to achieve rather high doping levels , which allows to study the physical properties of these structures .
furthermore , we present a detailed theoretical analysis of the structure and electronic properties of intercalated compounds . upon doping the charge carrier plasmon continuously
shifts to lower energies , a behavior supported by the theoretical investigations and ruling out a possible formation of doped phases .
the bands at the fermi level , which get filled by potassium doping are of ta@xmath47 character .
their bandwidth @xmath48 is rather small ( below ) , which could hint towards the presence of sizable correlation effects .
however , the wannier functions for these band states as calculated in ref .
are rather extended , which results in a comparatively small value for the hubbard @xmath49 .
hence , we expect a ( possibly much ) larger ratio @xmath50 compared to @xmath51 transistion metal compounds and no pronounced mott - hubbard physics .
structurally , our thin film samples show a transition to a charge - density wave state below in the undoped case and a suppression of the according superstructure upon potassium intercalation . moreover ,
new superstructures with a different periodicity arise that can be assigned to the increasing potassium content in the samples .
the changes in the basic structure with respect to the @xmath0-axis can be consistently found in the electron energy - loss spectroscopy measurements as well as in the calculations .
the spectral width of the charge carrier plasmon significantly decreases upon potassium addition , i.e. , the plasmon excitation is much less scattered at higher doping levels .
furthermore , as our data indicate the suppression of the charge density wave ground state upon electron doping , it is tempting to speculate about the possibility of an increasing superconducting transition temperature in potassium doped ( in the undoped compound @xcite ) following the observation that often there is competition between a superconducting and a charge density wave state in transition metal dichalcogenides @xcite .
we thank r.schnfelder , s.leger and r.hbel for technical assistance .
this work was supported by the deutsche forschungsgemeinschaft ; grant number kn393/12 and kn393/13 .
43ifxundefined [ 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 10.1107/s0365110x65000063 [ * * , ( ) ] link:\doibase 10.1021/j100873a038 [ * * , ( ) ] link:\doibase 10.1016/0375 - 9601(67)91005 - 5 [ * * , ( ) ] http://www.informaworld.com/10.1080/00018736900101307 [ * * , ( ) ] link:\doibase 10.1080/00018737500101391 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.34.734 [ * * , ( ) ] link:\doibase 10.1103/physrevb.16.801 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.233101 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.196402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.125112 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.81.453 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.176404 [ * * , ( ) ] http://www.informaworld.com/10.1080/00018738700101951 [ * * , ( ) ] `` , '' ( , , ) pp .
http://stacks.iop.org/0953-8984/13/9879 [ * * , ( ) ] link:\doibase 10.1016/0038 - 1098(92)90459-m [ * * , ( ) ] link:\doibase 10.1103/physrevb.67.125107 [ * * , ( ) ]
\doibase doi : 10.1016/s0039 - 6028(00)00717 - 2 [ * * , ( ) ] link:\doibase 10.1142/s0217984900000628 [ * * , ( ) ] \doibase doi : 10.1016/0022 - 5088(70)90096 - 2 [ * * , ( ) ] link:\doibase 10.1016/s0039 - 6028(01)00849 - 4 [ * * , ( ) ] link:\doibase 10.1080/00268948508074851 [ * * , ( ) ] , link:\doibase 10.1016/0025 - 5408(80)90188 - 9 [ * * , ( ) ] link:\doibase 10.1002/pssa.2211070237 [ * * , ( ) ] link:\doibase 10.1103/physrevb.43.9420 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/s0065253908609476 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.1743 [ * * , ( ) ] http://scitation.aip.org/getabs/servlet/getabsservlet?prog=normal&id=prbmdo000079000004045134000001&idtype=cvips&gifs=yes [ * * , ( ) ] link:\doibase 10.1103/physrevb.20.2217 [ * * , ( ) ] link:\doibase 10.1002/pssb.19700380245 [ * * , ( ) ] http://dx.doi.org/10.1007/bf01420885 [ * * , ( ) ] , \doibase doi : 10.1016/s0379 - 6779(98)00685 - 7 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.216404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.045103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.014534 [ * * , ( ) ] link:\doibase 10.1080/00018730110113644 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.14.225 [ * * , ( ) ] link:\doibase 10.1002/zaac.19542770307 [ * * , ( ) ] http://stacks.iop.org/0022-3727/1/i=3/a=303 [ * * , ( ) ] link:\doibase 10.1103/physrevb.21.3491 [ * * , ( ) ] \doibase doi : 10.1016/0378 - 4363(80)90276 - 4 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.026406 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys360 [ * * , ( ) ]
in order to check and complement the virtual crystal approximation ( vca ) calculations we constructed a few supercells .
no attempt was made to actually determine precise minimum energy structure for each concentration , which can get rather involved due to possible disorder effects or large wavelength structural modulations .
we picked cells , which seemed reasonable and have a threefold axis ( except for @xmath15 , see below ) .
for the smallest cell ( k ) we tried three structures in order to get a feeling for the spread of the plasmon parameters @xmath52 and @xmath53 due to structural and fermi surface reconstruction effects .
for the same reason we additionally considered cells with every second interlayer being empty for @xmath37 . for @xmath54
we only have an alternatingly empty supercell , due to cell size considerations .
it turns out that the energies of the alternatingly empty cells are higher but only by a rather small amount of about per tase@xmath6 formula unit .
such structures could be realized and stable in nature due to kinetic effects during intercalation and the large barrier for k to move through a tase@xmath6 layer . .
atoms at the corner of the prisms are se , the atoms inside the prisms are ta and the largest atoms on top are k. the difference between@xmath54 and @xmath55 is the alternation of empty and filled k - layers for @xmath54.,scaledwidth=90.0% ] .
note , that the second supercell has two different alternating k - layers .
the first cell has orthorhombic symmetry .
the first two cellsare energetically degenerate with a difference below , while the energy of the alternatingly empty third cell os higher by per tase@xmath6 formula unit.,scaledwidth=90.0% ] | we have investigated the influence of potassium intercalation on the formation of the charge - density wave ( cdw ) instability in by means of electron energy - loss spectroscopy and density functional theory .
our observations are consistent with a filling of the conduction band as indicated by a substantial decrease of the plasma frequency in experiment and theory . in addition
, elastic scattering clearly points to a destruction of the cdw upon intercalation as can be seen by a vanishing of the corresponding superstructures .
this is accompanied by a new superstructure , which can be attributed to the intercalated potassium .
based on the behavior of the @xmath0-axis upon intercalation we argue in favor of interlayer - sites for the alkali - metal and that the lattice remains in the @xmath1 modification . |
the essential assumption of general relativity is that the spacetime is given by a curved manifold with a metric @xmath0 of the lorentzian signature .
@xmath1 denotes the inverse of @xmath2 .
the curvature of the spacetime is given by the riemann tensor @xmath3 .
the metric lifts and lowers the tensor indices .
a vector is timelike , spacelike and null if @xmath4 , @xmath5 and @xmath6 , respectively .
a hypersurface is called timelike , spacelike and null , if its normal vector is spacelike , timelike and null , respectively .
we use the abstract index notation @xcite in this article .
the field equation for the metric is given by einstein s equations
@xmath7 where @xmath8 is the ricci tensor , @xmath9 is the scalar curvature and @xmath10 is the stress - energy tensor of matter fields .
we adopt the units in which @xmath11 .
einstein s equations were proposed so that it has the limit to newtonian gravity in a weak - field and slow - motion regime .
the conservation law @xmath12 follows from the bianchi identity , where @xmath13 denotes the covariant derivative associated with @xmath2 .
see @xcite for more complete description about the formulation of general relativity .
because of the lorentzian signature of the metric , the spacetime can admit a time function @xmath14 and then `` 3 + 1 '' decomposition , i.e. foliation with spacelike hypersurfaces labelled by @xmath14 .
this is called time foliation or time slicing of the spacetime and this enables us to regard einstein s equations as the combination of the evolution and constraint equations for the induced metric and the extrinsic curvature on the spacelike hypersurface .
if the spacetime admits a timelike killing vector , we can choose @xmath14 so that the induced metric does not depend on @xmath14 .
such a spacetime is called stationary .
general relativity is a self - consistent theory of gravity but not written in a completely closed form .
we need to specify the physics of matter fields by giving the action or the stress - energy tensors of matter fields , which provide the source term on the right - hand side of einstein s equations . given the action of matter fields @xmath15 , the stress - energy tensor is defined as a functional derivative in the following : @xmath16 where @xmath17 denotes the determinant of @xmath2 .
the simplest source term is vacuum , i.e. @xmath18 . if @xmath19 for a constant @xmath20 , this is called a cosmological constant .
a perfect fluid is often used , which is given by @xmath21 where @xmath22 is the four - velocity of the fluid element , satisfying @xmath23 .
this provides a very good approximation for gases , liquids and solids in many circumstances .
the equation which gives @xmath24 is called the equation of state . if @xmath25 , the perfect fluid is called a dust .
if @xmath26 , it is called a radiation fluid .
another common example is a scalar field , for which the stress - energy tensor is given by @xmath27 where @xmath28 is called a potential .
if @xmath29 , the scalar field is called massless . the matter fields evolve according to their equations of motion .
if there is only a single matter component , the conservation law ( [ conservation_law ] ) gives the equation of motion .
the metric is a solution of einstein s equations with the matter fields as the source . because of the conservation law
, the distribution of matter fields in space and time can not be put arbitrarily by some `` external force '' which does not contribute to the stress - energy tensor . in other words , we can not determine the metric simply after we assume the distribution of matter fields as we often do to determine electric and magnetic fields in electromagnetism .
we need to solve the metric and matter fields in a consistent manner . moreover , einstein s equations are highly nonlinear with respect to @xmath2 .
these properties make it very difficult to get exact solutions in dynamical and generic situations with or without matter fields . to obtain the general properties of spacetimes ,
it is important to know the general properties of matter fields . among such conditions for matter fields
are energy conditions , which impose the energy density being positive in some sense .
the strong energy condition is one of them , assuming @xmath30 for any timelike vector @xmath31 .
this implies @xmath32 and @xmath33 for a perfect fluid .
almost all known exact solutions to einstein s equations have been obtained under strong assumption on symmetry .
a very powerful method to obtain more or less general solutions is to establish numerical solutions .
this is called numerical relativity in a broad sense .
it has been achieving great success in recent days , not only in numerical simulations of relativistic astrophysical phenomena but also in the numerical experiments of phenomena with nonlinearly strong gravity .
there are exact solutions to einstein s equations which have spacetime singularities .
see @xcite for the precise notions of causal structure and spacetime singularities . here
we look at several examples .
the first one is the schwarzschild solution , in which the line element is written in the following form : @xmath34 this is the unique solution for spherically symmetric vacuum spacetimes .
this describes a black hole for @xmath35 . in this case
, @xmath36 is a coordinate singularity , corresponding to an event horizon , while @xmath37 is a genuine spacetime singularity , towards which the scalar curvature polynomials tend to diverge .
these features are well understood in the penrose diagram or conformal diagram shown in fig .
[ schwarzschild_kruskal ] . in this figure
, the straight line with forty five degrees denotes a null ray and the the physical spacetime is compactified through the conformal transformation .
suppose that an observer is in region i and she tries to send a signal to future null infinity @xmath38 by emitting a light .
we can see that the future - directed outgoing null rays earlier than @xmath36 , i.e. in region i , can reach @xmath38 , while those later than @xmath36 , i.e. in region ii , can not .
@xmath37 is described by the future and past boundaries of the spacetime , which are black hole and white hole singularities , respectively .
no future - directed null ray can emanate from the black hole singularity .
the penrose diagram of the schwarzschild solution for @xmath35.,width=264 ] the second example is the friedmann solution which describes a homogeneous and isotropic universe .
the line element is then given by @xmath39,\ ] ] where the scale factor @xmath40 obeys the friedmann equation @xmath41 the matter energy density @xmath42 obeys the energy conservation law @xmath43 and the dot denotes the derivative with respect to @xmath14 .
@xmath44 denotes the curvature of the @xmath45const spacelike hypersurface .
note that the source term for the friedmann solution must be a perfect fluid .
if the strong energy condition is satisfied , or @xmath32 and @xmath33 in this case , @xmath46 then begins with 0 at @xmath47 , implying the divergence of scalar curvature polynomial where @xmath48 . this is a spacelike singularity and usually called big bang singularity or initial singularity .
figure [ friedmann ] shows the evolution of the scale factor and the causal structure . [
cols="^,^ " , ] the simplest model for gravitational collapse is the oppenheimer - snyder solution , which describes the complete collapse of a uniform dust ball .
this is given by matching the interior and exterior solutions .
the interior solution is the time - reversed friedmann universe with a dust .
the exterior is spherically symmetric and vacuum , which is given by the schwarzschild solution .
they are matched on a timelike hypersurface , which is generated by timelike radial geodesics .
the penrose diagram of the resultant spacetime is shown in fig .
[ os ] , which is given by cutting and pasting those of the friedmann solution and the schwarzschild solution .
we can see that the singularity is spacelike and hidden behind the event horizon from the external observer .
there exists an achronal , i.e. spacelike or null , three - dimensional hypersurface such that the whole spacetime is the domain of dependence of this surface .
this surface is called the cauchy surface and the spacetime with the cauchy surface is called globally hyperbolic .
the penrose diagram of the oppenheimer - snyder solution.,width=188 ] the geodesic deviation equation implies @xmath49 along a geodesic congruence , where @xmath31 is the normalised tangent vector , @xmath50 is an affine parameter , and @xmath51 , @xmath52 and @xmath53 are respectively the expansion , shear and twist .
this equation is called the raychaudhuri equation .
this implies that the timelike geodesic congruence tends to focus , in other words , @xmath51 tends to diverge to @xmath54 if @xmath55 or matter fields satisfy the strong energy condition .
based on the above properties , it was proved that there is at least one incomplete timelike or null geodesic in generic expanding universe and generic gravitational collapse @xcite .
this is called singularity theorems .
the geodesic incompleteness implies the existence of spacetime singularities .
this is a great achievement in the studies of classical general relativity .
however , singularity theorems would not reveal the properties of generic singularities .
then , we see two important conjectures on this issue . we first consider initial singularities . the kasner solution is given by @xmath56 where @xmath57 , @xmath58 and @xmath59 are three indices , satisfying @xmath60 and @xmath61 .
this solution describes a homogeneous and anisotropic vacuum universe and has initial singularity at @xmath47 .
the dynamics of the bianchi type ix homogeneous universe was shown to be the successive series of kasner regimes replacing the indices one another @xcite . in each regime the solution is well approximated by the kasner solution but the different regime has different indices @xmath62 .
the transition is oscillatory and chaotic .
subsequently , it was conjectured that the process of approach to initial singularity in the general inhomogeneous case tends to be local , oscillatory and chaotic and the dynamics of nearby observers decouple from each other near the singularity @xcite .
this conjecture is called the belinski - khalatnikov - lifshitz conjecture and has been recently strongly supported by numerical relativity experiment of initial singularity with no symmetry @xcite . as for spacetime singularities formed in gravitational collapse ,
the cosmic censorship conjecture was proposed by penrose .
for the weak version , he conjectured , `` a system which evolves , according to classical general relativity with reasonable equations of state , from generic non - singular initial data on a suitable cauchy - hypersurface , does not develop any spacetime singularity which is visible from infinity '' @xcite .
for the strong version , he conjectured , `` @xmath63 a physically reasonable classical spacetime @xmath64 ought to have the property @xmath63 @xmath64 is globally hyperbolic @xmath63 '' @xcite .
a singularity which is censored by this conjecture is called a naked singularity . in other words ,
the cosmic censorship claims that there is no naked singularity in physical spacetimes .
obviously , there are undefined terms in this conjecture , such as reasonable equations of state and generic initial data .
the cosmic censorship is a basic assumption to prove theorems on the properties of black holes , such as no bifurcation , area increase and the existence of an event horizon outside or coinciding to an apparent horizon @xcite .
the proof for the cosmic censorship is still very limited . on the other hand
, it has been revealed that there are a lot of solutions of einstein s equations , which satisfy energy conditions and have naked singularities .
we do not know how to apply known physics at spacetime singularities .
hence , if there is a naked singularity , it would spoil the the predictability for our future within classical theory . on the other hand ,
naked singularities may be regarded as a window into physics beyond general relativity @xcite .
see @xcite for naked - singular solutions in gravitational collapse and possible physical processes in naked singularity formation .
general relativity has no characteristic scale of its own , which implies the existence of self - similar solutions .
self - similar solutions are easier to obtain than more general solutions because partial differential equations reduce to ordinary differential equations for spherically symmetric self - similar spacetimes . for self - similar solutions , the energy density @xmath42 , for example , is written as @xmath65 with appropriate time and radial coordinates @xmath14 and @xmath66 .
it is found that in some spatially homogeneous models self - similar solutions can describe the asymptotic behaviour of more general solutions @xcite .
it was also conjectured that spherically symmetric fluctuations might naturally evolve via einstein s equations from a complex initial conditions to a self - similar form @xcite .
see @xcite for a recent review of self - similar solutions in general relativity .
more precisely , if a vector field @xmath67 satisfies @xmath68 this is said to be a homothetic killing vector .
if there exists a homothetic killing vector in a spacetime , this spacetime is said to be self - similar or homothetic . for a spherically symmetric self - similar spacetime
, introducing coordinates @xmath69 such that @xmath70 a nondimensional quantity @xmath71 satisfies @xmath72 for any @xmath73 and hence @xmath74 .
the line element is then given by @xmath75 where @xmath76 .
on the other hand , if there exists a positive @xmath77 such that @xmath78 holds only for @xmath79 , this is called discretely self - similar . to have self - similar spacetimes , matter fields are strongly restricted .
both a perfect fluid with the equation of state @xmath80 and a massless scalar field @xmath81 are still compatible with self - similar spacetimes . for self - similar spacetimes , einstein
s equations reduce to a set of ordinary differential equations .
note that a perfect fluid with @xmath80 ( @xmath82 ) has a sound wave at the speed @xmath83 , while a massless scalar field has a scalar wave at the speed of light .
this property results in a critical point of the ordinary differential equations .
this is called a sonic point for the perfect fluid case .
critical points are classified through dynamical systems theory .
no information propagates inwardly beyond the sonic point .
we focus on self - similar solutions with a perfect fluid with @xmath80 ( @xmath82 ) . because of the critical point in the ordinary differential equations
, the smoothness condition strongly restricts the class of solutions .
we assume analytic initial data for self - similar perfect fluid solutions .
analyticity in the present context means the taylor - series expandability of the energy density with respect to the riemannian normal coordinates .
in particular , we impose analyticity both at the centre and at the sonic point .
then we have only a discrete set of solutions .
they are the flat friedmann solution , general relativistic larson - penston ( grlp ) solution , general relativistic hunter ( a ) ( grha ) solution and so on .
these are obtained numerically except for the flat friedmann solution . for these self - similar solutions , a set of analytic initial data
is prepared at @xmath84 and singularity appears at @xmath47 .
it is found that , except for the flat friedmann solution , the singularity appears at @xmath47 only at the centre and we can analytically extend the solution beyond @xmath47 to positive @xmath14 for @xmath85 .
if it is extended , the grlp solution describes the formation of naked singularity from analytic and therefore regular initial data for @xmath86 @xcite .
see fig .
[ penrose_grlp ] for the penrose diagram of this spacetime .
we can see that the singularity is not completely hidden behind the event horizon . the penrose diagram of the grlp solution for @xmath86 .
the dashed line denotes the event horizon.,width=188 ] since the cosmic censorship conjecture censors the generic occurrence of naked singularity , it is important whether the naked - singular solution is stable or not . the self - similar solution is given representatively as @xmath87 .
we assume that a linearly perturbed solution from this is given in the following form : @xmath88 where @xmath89 and @xmath90 .
we can get equations for linear perturbation from einstein s equation .
we impose regularity condition both at the centre and the sonic point .
then , we can determine the value of @xmath91 as an eigenvalue problem
. then it is found that the grlp has no unstable mode , the grha has one unstable mode and other solutions except the flat friedmann have more than one unstable modes @xcite .
since the sonic point is a one - way membrane for sound waves , there can appear a different kind of instability @xcite .
if we inject density gradient discontinuity at the sonic point , we can find this discontinuity evolves locally at the sonic point .
this perturbation mode is called a kink mode . the stability against this discontinuity
is completely determined by the class to which the sonic point belongs as a critical point .
then , it is found that , against this specific mode , the flat friedmann solution is unstable for @xmath92 and stable for @xmath93 , the grlp is stable for @xmath94 and unstable for @xmath95 , and the grha is stable for @xmath96 and unstable for @xmath97 . as seen above , we have a self - similar solution which has no unstable mode .
this is the grlp solution for @xmath98 .
in fact , a numerical relativity experiment strongly suggests that this is a global attractor @xcite . in the numerical simulation
, the collapse ends in singularity formation at the centre for a certain subset of initial data sets , where the initial data sets were prepared without fine - tuning .
then , the profile of the density profile tends to evolve in a self - similar manner and agree very well with the grlp solution .
it was confirmed that this convergence to the self - similar attractor solution does not depend on the detailed choice of the initial density profile .
the above is the results of numerical simulations for @xmath99 .
although the first numerical simulation was done by the simple misner - sharp scheme , this result has been recently confirmed by a much more elaborated numerical scheme code with high resolution shock capturing , adaptive mesh refinement and innovative treatment of vacuum exterior @xcite .
this strongly suggests that the cosmic censorship is violated within spherical collapse .
this is because the grlp solution describes naked singularity formation for @xmath86 .
hence , this example is one of the strongest counterexamples against the cosmic censorship . on the other hand
, the grlp solution would be unstable against nonspherical perturbation for @xmath100 , as the direct consequence of the linear perturbation analysis @xcite .
there remains much to study in nonspherical collapse .
suppose we have a generic one - parameter family of initial data sets parametrised by @xmath24
. then , there generally exists a threshold value @xmath101 for black hole formation .
a near - critical collapse first approaches a self - similar solution and deviates away eventually .
this self - similar solution that sits at the threshold is called a critical solution .
the scaling law for the formed black hole for supercritical collapse is given as @xmath102 for @xmath103,where @xmath104 is called a critical exponent .
this is observed only as a result of fine - tuning the parameter @xmath24 to be @xmath103 .
the critical solution and critical exponent do not depend on the prepared one - parameter family of initial data sets , which is called universality .
the above phenomena were first observed in numerical simulation by choptuik @xcite for the spherical collapse of a massless scalar field and are called critical phenomena .
it is found that critical phenomena are seen in a variety of systems , such as a perfect fluid with @xmath80 @xcite .
see @xcite for an extensive review on this subject .
the critical behaviour turns out to be well understood by renormalisation group approach @xcite .
we consider the space of functions of @xmath105 . we can regard this as the space of initial data sets .
there the critical solution @xmath87 is characterised as a fixed point in this space with a single unstable mode .
this means that the fixed point has a stable manifold of codimension one .
let @xmath106 and @xmath107 be the eigenvalue and eigenfunction of this unstable mode .
the one - parameter family of initial data sets corresponds to a curve in the space of initial data sets and therefore generically has intersection with the stable manifold .
the intersection is actually the initial data for the exact critical collapse , i.e. @xmath108 , which we denote as @xmath109 .
the initial data set for near critical collapse is then given by @xmath110 where @xmath50 is chosen to be 0 for the initial time .
for large @xmath50 , the deviation from the critical solution is dominated by the unique unstable mode of the fixed point .
the near - critical solution is then approximated as @xmath111 assuming that the deviation becomes of order unity at the black hole formation , we get the scaling law for the black hole mass as @xmath112 where @xmath113 .
in fact , for the system of a perfect fluid with @xmath80 , the grha solution has a single unstable mode and acts as a critical solution .
the observed critical exponent in the numerical experiment of gravitational collapse agrees very well with the above expected value obtained from the eigenvalue analysis . from this point of view ,
the grlp solution , which has no unstable mode , corresponds to an attractor of codimension zero , i.e. a global attractor .
the critical behaviour itself has important implication to cosmic censorship . in the limit of @xmath114 from the supercritical regime
, we will have a black hole of arbitrarily small mass , which can be regarded as a naked singularity because the curvature strength near the black hole horizon scales as @xmath115 .
more directly , it is found that the choptuik critical solution , which is discretely self - similar , actually has a naked singularity at the centre @xcite .
however , it should be noted that this naked singularity is realised as a consequence of exact fine - tuning and hence nongeneric .
this is very different from the naked - singular grlp solution , in which naked singularity generically appears because the solution acts as a global attractor .
lastly , we briefly review the newtonian collapse of isothermal gas in the present context .
this system is much simpler than but still very similar to the general relativistic system of a perfect fluid .
there are a discrete set of spherically symmetric self - similar solutions with analytic initial data , including a homogeneous collapse , larson - penston solution , hunter ( a ) solution and so on @xcite .
the kink mode was also studied in this system @xcite .
the numerical experiment of gravitational collapse and the normal mode analysis show that there exist both convergence and critical phenomena @xcite .
the larson - penston solution acts as a global attractor solution , while the hunter ( a ) solution acts as a critical solution .
this example shows that general relativity is not very essential to the appearance of critical behaviour .
figure [ flow ] schematically shows the renormalisation group flow for this system .
the global attractor solution ( the larson - penston solution ) has no unstable mode , while the critical solution ( the hunter ( a ) solution ) has a single unstable mode . a generic initial data set approaches the global attractor solution .
on the other hand , only if we tune the parameter near the critical value , the supercritical initial data set first approaches the critical solution , deviates away from it and then approaches the global attractor solution .
this interplay of the critical and global attractor solutions was first numerically observed in the newtonian collapse of an isothermal gas @xcite and subsequently in the general relativistic system of a perfect fluid @xmath80 with sufficiently small positive @xmath116 @xcite .
the interplay of convergence and critical phenomena.,width=264 ]
the spacetime singularity has been a central issue in general relativity over several decades . recently numerical relativity has been developed and can reveal the properties of generic spacetime singularities .
we have focused on the role of self - similar solutions in singularity formation in gravitational collapse .
the general theory of relativity as well as newtonian gravity admits self - similar solutions . this is due to the scale - invariance of the theory .
the self - similar solutions are important not only because they are dynamical and inhomogeneous solutions easier to obtain but also because they may play important roles in the asymptotic behaviour of more general solutions in certain circumstances .
the numerical relativity experiments of gravitational collapse and the semi - analytical studies on self - similar solutions reveal that there is a self - similar solution which acts as a global attractor in the spherical collapse of a perfect fluid and then the cosmic censorship will be violated if it is formulated within spherical symmetry .
we have also seen that the stability analysis of self - similar solutions gives a unified picture of the convergence to an attractor and the critical behaviour in gravitational collapse in terms of attractors of codimension zero and one , respectively .
the interplay of these two behaviours has been recently observed in numerical simulation of gravitational collapse both in newtonian gravity and general relativity . since both the critical and convergence behaviours are not only in general relativity but also in newtonian gravity , these two are considered to be common characteristics of gravitational physics and possibly of a wider class of scale - free nonlinear systems .
wald , _ general relativity _ ,
( university of chicago press , chicago , 1984 ) .
s.w . hawking and g.f.r .
ellis , _ the large scale structure of space - time _ , ( cambridge university press , cambridge , 1973 ) .
clarke , _ the analysis of space - time singularities _ , ( cambridge university press , cambridge , 1993 ) .
belinski , e.m .
lifshitz and i.m .
khalatnikov , adv . in physics
* 19 * , 525 ( 1970 ) .
belinski , i.m .
khalatnikov and e.m .
lifshitz , adv . in physics
* 31 * , 639 ( 1982 ) .
d. garfinkle , phys .
lett . * 93 * , 161101 ( 2004 ) .
r. penrose , riv .
nuovo cim . * 1 * , 252 ( 1969 ) .
r. penrose , in _ general relativity , an einstein centenary survey _ , edited by s.w .
hawking and w. israel ( cambridge university press , cambridge , england , 1979 ) , p.581 .
t. harada and k. nakao , phys .
rev . d*70 * , 041501(r ) ( 2004 ) .
joshi , pramana * 55 * , 529 ( 2000 ) .
t. harada , h. iguchi and k. nakao , prog .
theor . phys . *
107 * , 449 ( 2002 ) .
t. harada , pramana * 63 * , 741 ( 2004 ) .
j. wainwright and g.f.r .
ellis , _ dynamical systems in cosmology _ , ( cambridge university press , cambridge 1997 ) .
carr , unpublished ( 1993 ) .
b.j . carr and a.a .
coley , gen .
* 37 * , 2165 ( 2005 ) .
a. ori and t. piran , phys .
lett . * 59 * , 2137 ( 1987 ) .
a. ori and t. piran , phys .
d * 42 * , 1068 ( 1990 ) .
t. koike , t. hara and s. adachi , phys .
lett . * 74 * , 5170 ( 1995 ) .
t. koike , t. hara and s. adachi , phys .
d * 59 * , 104008 ( 1999 ) .
d. maison , phys .
b * 366 * , 82 ( 1996 ) .
t. harada and h. maeda , phys .
d * 63 * , 084022 ( 2001 ) .
brady , m.w .
choptuik , c. gundlach and d.w .
neilsen , class .
quantum grav .
* 19 * , 6359 ( 2002 ) .
m. snajdr , class . quantum grav . * 23 * , 3333 ( 2006 ) .
t. harada , class .
quantum grav . * 18 * , 4549 ( 2001 ) . c. gundlach , phys .
d*65 * , 084021 ( 2002 ) .
choptuik , phys .
lett . * 70 * , 9 ( 1993 ) .
neilsen and m.w .
choptuik , class .
quantum grav .
* 17 * , 761 ( 2000 ) .
evans and j.s .
coleman , phys .
lett . * 72 * , 1782 ( 1994 ) . c. gundlach , phys
. rept . * 376 * , 339 ( 2003 ) .
j.m . martin - garcia and c. gundlach , phys .
rev . d*68 * , 024011 ( 2003 ) .
larson , mon . not .
. soc . * 145 * , 271 ( 1969 ) .
penston , mon . not .
. soc . * 144 * , 425 ( 1969 ) . c. hunter , astrophys .
j. * 218 * , 834 ( 1977 ) .
a. whitworth and d. summers , mon . not .
. soc . * 214 * , 1 ( 1985 ) .
a. ori and t. piran , mon . not .
* 234 * , 821 ( 1988 ) .
h. maeda and t. harada , phys .
d * 64 * , 124024 ( 2001 ) .
t. harada , h. maeda and b. semelin , phys .
d*67 * , 084003 ( 2003 ) . | einstein s field equations in general relativity admit a variety of solutions with spacetime singularities . numerical relativity has recently revealed the properties of somewhat generic spacetime singularities .
it has been found that in a variety of systems self - similar solutions can describe asymptotic or intermediate behaviour of more general solutions .
the typical example is the convergence to an attractor self - similar solution in gravitational collapse .
this is closely related to the cosmic censorship violation in the spherically symmetric collapse of a perfect fluid .
the self - similar solution also plays an important role in critical phenomena in gravitational collapse .
the critical phenomena are understood as the intermediate behaviour around a critical self - similar solution .
we see that the convergence and critical phenomena are understood in a unified manner in terms of attractors of codimension zero and one , respectively , in renormalisation group flow . |
the spontaneous breakdown of symmetries in the early universe that produce linear discontinuities in the field is called cosmic string ( kibble 1976 ) .
the cosmic strings are also common in modern string cosmologies . in 2003 , general interest in cosmic strings was heightened by the discovery of what seemed at first to be a plausible candidate for lensing by a cosmic string . a pair of images of elliptical galaxies separated by 1.8 arc seconds was found to have the same redshift , @xmath1 , and the same spectra .
sazhim et al ( 2003 ) confirmed that these images were not distorted in the way that would be expected for lensing of a single galaxy but are consistent with lensing by cosmic string . in the literature ,
the line like structure of cosmic strings with particle attached to them are considered as possible seeds for galaxy formation at the early stages of the evolution of universe . in the past time , stachel ( 1980 ) , letelier ( 1983 ) and vilenkin et al ( 1987 ) have studied different aspects of string cosmological models in general relativity .
reddy ( 2003 ) , reddy and naidu ( 2007 ) and recently yadav ( 2013 ) have investigated anisotropic string cosmological models in scalar - tensor theory of gravitation . + harko et al ( 2011 ) proposed @xmath0 gravity theory by taking into account the gravitational lagrangian as the function of ricci scalar r and of the trace of energy - stress tensor t. they have obtained the equation of motion of test particle and the gravitational field equation in metric formalism both .
myrzakulov ( 2011 ) presented point like lagrangian s for @xmath0 gravity .
the @xmath0 gravity model that satisfy the local tests and transition of matter from dominated era to accelerated phase was considered by houndjo ( 2012 ) .
recently chaubey and shukla ( 2013 ) , naidu et al ( 2013 ) and ahmad and pradhan ( 2013 ) have investigated anisotropic cosmological model in @xmath0 gravity . in this paper ,
our aim is to study the role of strings in @xmath0 gravity and bianchi - v space time .
since bianchi - v models are natural generalization of frw models hence the bianchi - v cosmological models permit one to obtain more general cosmological model , in comparison to frw model .
in the recent years several authors ( yadav 2011 , kumar and yadav 2011 , yadav 2012 , pradhan et al 2005 , singh et al 2008 ) have studied bianchi - v cosmological models in different physical context .
+ in this paper , we establish the existence of bianchi - v string cosmological model in @xmath0 gravity and examine a cosmological scenario by proposing power law expansion .
we observe that strings do not survive for long time and eventually disappear from universe .
the organization of the paper is as follows : the model and basic theory of @xmath0 gravity are presented in section 2 .
the field equations are established in section 3 and their solution is presented in section 4 .
section 5 deals with the cosmological parameters and classical potential of derived model .
finally conclusions are presented in section 6 .
we consider the bianchi - v metric in following form @xmath2 here , @xmath3 , @xmath4 & @xmath5 are scale factors in the x , y & z direction respectively and @xmath6 is constant .
+ the action of @xmath0 gravity is given by @xmath7 where @xmath8 , @xmath9 and @xmath10 are the ricci scalar , the trace of the stress - energy tensor of matter and the matter lagrangian density respectively .
+ the stress - energy tensor of matter is given by @xmath11 the gravitational field equation of @xmath0 gravity is given by @xmath12 where @xmath13 , @xmath14 , @xmath15 and @xmath16 denotes the covariant derivative .
+ the stress - energy tensor is given by @xmath17 in general , the field equations depend through the tensor @xmath18 , on the physical nature of the matter field . hence we obtain several theoretical models for different choice of f(r , t ) depending on the nature of the matter source . in the literature , chaubey @xmath19 shukla ( 2013 ) , reddy et al .
( 2012a , 2012b ) and naidu et al ( 2013 ) have been studied the cosmological models , assuming @xmath20 .
recently ahmad @xmath19 pradhan studied consequence of bianchi - v cosmological model in f(r , t ) gravity by considering @xmath21 .
they have assumed perfect fluid as source of matter while in this paper , we assumed the string fluid as source of matter to describe the physical consequences of early universe .
thus our paper is all together different from the paper of ahmad and pradhan ( 2013 ) .
assuming @xmath22 and @xmath23 where @xmath24 is arbitrary parameter .
+ now equation ( [ eq4 ] ) can be rewritten as @xmath25 throughout the paper , we use units @xmath26 .
+ the expansion scalar @xmath27 and shear scalar @xmath28 have the form @xmath29 @xmath30 - \frac{\theta^{2}}{3}\ ] ]
the einstein s field equation ( [ eq6 ] ) for the line - element ( [ eq1 ] ) leads to the following system of equations @xmath32
@xmath33 @xmath34 @xmath35 @xmath36
we have system of five equation ( [ eq9])@xmath37([eq13 ] ) involving six unknown variables , namely @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 & @xmath43 .
thus , in order to completely solve the fields equations , we need at least one physical assumption among the unknown parameters . in the literature , it is common to use the law of variation of hubble s parameter which yields the constant value of deceleration parameter .
this law deals with two type of cosmology ( i ) power law cosmology ( ii ) exponential law cosmology . it is well known that the exponential law projects the dynamics of future universe and such type of model does not have consistency with present day observations .
since we are looking for a model , describing the late time acceleration of universe therefore we choose the average scale factor in following form @xmath44 where @xmath45 , is positive constant .
( [ eq14 ] ) can be easily obtain by law of variation of hubble s parameter .
+ we define average scale factor @xmath46 and mean hubble s parameter @xmath47 of the bianchi - v model as @xmath48 @xmath49 where @xmath50 , @xmath51 and @xmath52 are directional hubble s parameters in the direction of x , y , and z- direction respectively .
+ from equation ( [ eq10 ] ) and ( [ eq11 ] ) , we obtain the following relation @xmath53 where @xmath54 and @xmath55 are constant of integrations . + integrating eq .
( [ eq13 ] ) , we get @xmath56 from equations ( [ eq14 ] ) , ( [ eq15 ] ) , ( [ eq17 ] ) and ( [ eq18 ] ) , the metric functions can be explicitly written as @xmath57 @xmath58\ ] ] @xmath59\ ] ] provided @xmath60 .
the average hubble s parameter @xmath47 , expansion scalar @xmath27 , spatial volume @xmath61 , deceleration parameter ( dp ) and shear scalar @xmath28 are given by @xmath62 @xmath63 @xmath64 @xmath65 @xmath66 for accelerating universe , we impose the restriction on the value of n @xmath67 . if we take @xmath68 then the value of dp is @xmath69 which exactly matches with the observed value of dp at present epoch ( cunha et al .
hence we constrain @xmath68 in graphical representations and discussion of physical parameters of derived model .
+ the isotropic pressure @xmath70 , proper energy density @xmath71 , string tension density @xmath72 and particle energy density @xmath73 are found to be @xmath74\ ] ] @xmath75\ ] ] @xmath76\ ] ] @xmath77\ ] ] [ cols= " > , < " , ] from eq .
( [ eq17 ] ) , we obtain @xmath78 following saha and boyadjiev ( 2004 ) and yadav(2013 ) , we have the following equation of motion of a single particle with unit mass under force @xmath79 @xmath80 where @xmath81 and @xmath82 are the classical potential and amount of energy respectively .
( [ eq29 ] ) and ( [ eq30 ] ) lead to @xmath83 the classical potential in terms of hubble s parameter is given by @xmath84 the age of universe in connection with dp is given by @xmath85 from eq .
( [ eq33 ] ) , it is clear that the value of @xmath86 in the range @xmath87 increase the age of universe .
+ also we know that the speed of sound @xmath88 is less than the speed of light @xmath89 .
in gravitational unit we take @xmath90 . therefore for a physical viable model
, @xmath88 lies between 0 and 1 .
+ from eq .
( [ eq25 ] ) and ( [ eq26 ] ) , the speed of sound is given by @xmath91 where + @xmath92 + @xmath93 + @xmath94 + in figure panel 1 , we graphed the parameters of derived model against @xmath95 for @xmath68 .
the behaviour is quite evident ; the scale factors along axial direction satisfies the anisotropic nature of universe ; @xmath96 shows that the particles dominate over the strings with the evolution of universe hence the strings are not observed today ; the classical potential @xmath97 is positive and speed of sound @xmath98 is less the speed of light throughout the expansion of universe from big bang to present epoch .
in the present paper , we have considered @xmath0 gravity model with an arbitrary coupling between matter and geometry in bianchi - v space - time .
we have derived the gravitational field equations for string fluid corresponding to @xmath0 gravity model .
we observed that string tension density @xmath72 decreases with time and it approaches to zero at present epoch .
therefore strings could not survive with the evolution process of universe .
that is why the strings are not observed today but it play significant role in early universe .
the classical potential @xmath97 is positive and it decreases with time .
the scale factors vanish at @xmath99 .
thus the model has point type singularity at @xmath99 . as @xmath100 the scale factors diverge and the physical parameters such as expansion , scalar @xmath27 , energy density @xmath71 and hubble s parameters @xmath47 tend to zero .
therefore in the derived model , all matter and radiation are concentrated in the big bang .
it is important to note that @xmath101 and @xmath102 for @xmath100 in the derived model which implies the fastest rate of expansion of universe .
so , the derived model can be utilized to describe the dynamics of universe at present epoch . since @xmath103 ,
thus the model approaches isotropy at late times . | * abstract : * in this paper , we search the existence of bianchi - v string cosmological model in @xmath0 gravity with power law expansion .
einstein s field equations have been solved by taking into account the law of variation of hubble s parameter that yields the constant value of deceleration parameter ( dp ) .
we observe that the massive strings dominate the early universe but they do not survive for long time and finally disappear from the universe .
we examine the nature of classical potential and also discuss the physical properties of universe . + * keywords : * early universe , @xmath0 gravity and cosmological parameters . |
the neon - sodium cycle of hydrogen burning ( nena cycle ) converts hydrogen to helium through a chain of proton induced reactions involving neon and sodium isotopes .
the nena cycle contributes negligibly to the energy budget , but it is of great importance for stellar nucleosynthesis , because it affects the abundances of the elements between @xmath9ne and @xmath10al @xcite .
predicting the abundances of nena cycle elements in different astrophysical objects has become highly topical since the discovery of the anticorrelation between sodium and oxygen abundances in red giant stars of globular clusters ( * ? ? ?
* a review ) . in principle ,
material showing such a na - o anticorrelation may be produced by hydrogen burning when the temperature is so high that not only the cno cycle , but also the nena cycle is activated @xcite . in order to explain how such material can be brought to the stellar surface where it can be detected ,
mainly one scenario is invoked : pollution of the interstellar medium with the ashes of hydrogen burning from a previous generation of stars .
several candidates have been proposed for such a pollution : intermediate - mass asymptotic giant branch ( agb ) stars or super agb stars @xcite , fast rotating massive stars @xcite , supermassive stars @xcite , massive stars in close binary systems @xcite , stellar collisions @xcite , and classical novae @xcite . a detailed understanding of the nuclear physics of the nena cycle is needed in order to sharpen quantitative predictions for the nucleosynthetic output of these objects , which , in turn , are necessary to evaluate their respective impact on the na - o anticorrelation @xcite . the @xmath0ne(p,@xmath1)@xmath2na reaction is the most uncertain reaction of the cycle @xcite .
the uncertainty is due to a large number of predicted but hitherto unobserved resonances at proton energies refers to the proton beam energy in the laboratory system , @xmath11 to the center - of - mass energy , and @xmath8 to the excitation energy in the @xmath2na nucleus . ] below 400 kev @xcite , and to a number of observed but poorly constrained resonances at 400 - 1300kev @xcite , see fig .
[ fig : ls ] for a partial level scheme of @xmath2na including the astrophysically relevant resonances .
the resonances above 400 kev have been studied several times before , most recently at the 3mv tandetron of helmholtz - zentrum dresden - rossendorf ( hzdr ) ( * ? ? ? * and references therein ) . for the resonances below 400 kev ,
however , data are scarce .
the @xmath2na states between @xmath8 = 8822 - 9171 kev have been studied with indirect techniques @xcite , but in many cases the spin and parity assignments to those levels are uncertain and their @xmath1 decay modes are poorly known , if at all .
direct @xcite and indirect @xcite experiments have been performed to derive the strength of @xmath0ne(p,@xmath1)@xmath2na resonances below 400 kev , but both approaches only provided upper limits ( tab .
[ tab : wg_trr ] ) .
moreover , three states at @xmath8 = 8862 , 8894 and 9000 kev have been reported as tentative in an indirect measurement @xcite .
( color online ) partial level scheme of @xmath2na , level energies taken from @xcite .
the @xmath12-value and the resonance energies in the laboratory system of the @xmath0ne(p,@xmath1)@xmath2na reaction are also shown .
the resonances highlighted in red have been studied at luna ( * ? ? ? * and present work ) .
the resonances shown in blue have recently been investigated at hzdr @xcite . on the right side of the plot ,
relevant energy ranges are given for three astrophysical scenarios.,width=264 ] the upper limits for the unobserved resonances are treated differently in the literature . in the nacre compilation @xcite ,
usually 1/10 of the upper limit is adopted as value , with an error bar from zero to the full upper limit . in the iliadis / starlib compilation @xcite ,
several unobserved resonances were completely excluded from consideration , therefore also the uncertainty stemming from these cases was left out @xcite . for the other resonances , iliadis / starlib adopted a monte carlo sampling technique . as a result of the different treatment of the unobserved resonances between nacre , on the one hand , and iliadis / starlib ,
on the other hand , the difference between the adopted rates in the two compilations is as high as a factor of 700 at a temperature of 0.08 gk @xcite .
sensitivity studies focusing on nucleosynthesis in agb stars experiencing hot bottom burning ( hbb ) @xcite and on classical novae explosions @xcite show how the uncertainty in the @xmath0ne(p,@xmath1)@xmath2na rate affects the abundances of the nuclides between @xmath9ne and @xmath10al , leading to abundance variations as high as two orders of magnitude . in order to clarify the situation of the @xmath0ne(p,@xmath1)@xmath2na resonances below 400 kev , a new direct experiment has been performed deep underground at the laboratory for underground nuclear astrophysics ( luna ) @xcite . the present data on the @xmath13 = 156.2 , 189.5 and 259.7 kev resonances has been published in abbreviated form @xcite . here , full details of the experiment and data analysis are given , together with the decay branching ratios of the newly observed resonances .
the thermonuclear reaction rate is given both in tabular and parameterized forms for use in stellar evolution codes .
the present work is organized as follows .
the experimental method and the setup are described in sec .
[ sec : experiment ] .
section [ sec : results ] discusses the data analysis technique .
the results obtained for each of the resonances investigated are shown in secs .
[ subsec:156]-[subsec : ul ] . the astrophysical reaction rate and a brief discussion are given in sec .
[ sec : astro ] .
the experiment has been conducted at the luna 400 kv accelerator @xcite , deep underground at the gran sasso national laboratories ( lngs ) , operated by the italian national institute for nuclear physics ( infn ) .
owing to its deep underground location , luna affords a uniquely low background @xcite , allowing to investigate proton capture on @xmath0ne with unprecedented sensitivity @xcite .
a proton beam with a typical intensity of 200 @xmath14a was delivered to a differentially pumped , windowless gas target ( * ? ? ? * for details ) .
the target chamber was filled with 1.5 mbar of neon gas , enriched to 99.9@xmath15 in @xmath0ne .
a recirculation and purification system collected the gas exhaust from the pumping stages , removed possible contaminants through a chemical getter and finally recycled the gas to the target chamber , keeping the purity of @xmath0ne constant throughout the measurements .
the beam entered the target chamber through a water cooled copper collimator ( denoted ap1 in fig .
[ fig : chamber ] ) of 4.3 cm length and 0.7 cm diameter and was stopped on the hot side of a beam calorimeter , used to measure the beam intensity with 1@xmath15 precision @xcite .
the gas density as a function of the position inside the chamber and the beam heating correction have been studied in details @xcite , giving the areal number density of the target atoms with an overall uncertainty of 1.3@xmath15 .
the possible presence of an air contamination in the gas was checked periodically using the intense @xmath16n(p,@xmath1)@xmath17o resonance at @xmath13 = 278 kev . the nitrogen level in the gas
was always found to be below 0.1@xmath15 .
emitted @xmath1 rays were detected by two high - purity germanium ( hpge ) detectors ( fig .
[ fig : chamber ] ) : one , of 90@xmath15 relative efficiency @xcite , was mounted on top of the target chamber and was collimated at 90@xmath18 ( ge90 ) .
the other , of 135@xmath15 relative efficiency , was mounted below the target chamber and collimated at 55@xmath18 ( ge55 ) .
both detectors were enclosed in a copper and lead shield of 15 - 25 cm thickness , absorbing @xmath1 rays from the decay of environmental radioactive isotopes @xcite .
two independent data acquisition systems were used in parallel : one with a 100 mhz , 14-bit caen n1728b digital adc providing list mode data and the other with a standard analog signal amplification chain and 16k channel histogramming ortec 919e adc unit .
the detection efficiency for @xmath1-ray energies between 478 and 1836 kev was measured with point - like @xmath19cs , @xmath20co and @xmath21y activity standards from physikalisch - technische bundesanstalt , calibrated with @xmath22 error , and one @xmath23be radioactive source produced at mta atomki ( hungary ) and calibrated with @xmath24 uncertainty . by moving each of the radioactive sources , in turn , along the beam axis , the peak detection efficiency was measured as a function of the distance from the ap1 collimator ( see the left panel of fig . [
fig : efficiency ] for the efficiency profile at @xmath25 = 478kev , measured with the @xmath23be source ) .
the shape of the profile is dominated by the collimation geometry . for both detectors ,
the efficiency profile is maximal at a distance @xmath26 = 6.2 cm from ap1 ( fig.[fig : efficiency ] , left panel ) .
the efficiency curve as a function of @xmath1-ray energy was then extended up to @xmath25 = 6791 kev using the well - known @xmath16n(p,@xmath1)@xmath17o resonance at @xmath13 = 278 kev @xcite ( @xmath8 = 7556 kev ) .
this resonance de - excites either directly to the ground state or to one of the excited states at 5180 , 6172 , and 6791 kev .
each of these three states , in turn , decays with 100@xmath15 probability directly to the ground state . in the cascade via the 6172 kev ( 6792 kev ) level , a 1384 kev ( 765 kev ) @xmath1 ray is emitted , lying in the energy range where the efficiency can be determined using radioactive sources .
this allows to determine the efficiency for 6 mev @xmath1 rays using a relative approach : the efficiency at 1384 kev ( 765 kev ) is derived from the fit of the radioactive sources data , then the efficiency at higher energies was calculated normalizing the 6172 kev ( 6791 kev ) counting rate to the counting rate of the 1384 kev ( 765 kev ) peak .
given the short beam - detectors distance , true coincidence summing had to be taken into account in the evaluation of the detection efficiency .
the summing - out correction @xcite depends both on the full - energy peak efficiency and on the total detection efficiency .
a geant4 @xcite simulation of the setup was used to determine the peak - to - total ratio and to infer the total detection efficiency for @xmath20co , @xmath21y and @xmath16n(p,@xmath1)@xmath17o @xmath1 rays .
the summing out correction was always between 0.028 and 0.040 , depending on @xmath1-ray energies .
a conservative uncertainty of @xmath27 on the summing - out correction was adopted .
[ fig : efficiency ] ( right ) shows the full - energy peak efficiency as a function of the @xmath1-ray energy , measured at 6.2 cm from the ap1 collimator .
fits of the efficiency curves with second - order polynomials in the double - logarithmic plane @xcite are also shown , together with the fit residuals . the total uncertainty on the detection efficiency determined through the fit functions in fig .
[ fig : efficiency ] is always at the 1 - 2@xmath15 level .
further details on the experimental setup , and on the data analysis , can be found in thesis works @xcite .
for each resonance studied here , as a first step a run was performed at the beam energy where the maximum yield was expected according to the literature resonance energy , the beam energy loss and the efficiency profile of the detectors .
if a signal was observed , a resonance scan was performed , changing the beam energy in 1 - 2 kev energy steps and measuring the yield of the @xmath25 = 440 kev transition from the first excited state to the ground state of @xmath2na .
finally , a long run was performed at the beam energy where the maximal yield was observed , aiming for a statistical uncertainty below 10@xmath15 for each detector .
the branching ratio @xmath28 for each transition @xmath29 was then derived using the following relationship : @xmath30 where @xmath31 is the number of counts observed for each @xmath1 line and @xmath32 its peak detection efficiency . as the @xmath1-ray spectra
were always dominated by the resonance under study , the contribution from possible unobserved transitions was assumed to be negligible in the branching ratio analysis .
only one detector , ge55 , was used for the branching ratios determination .
the second order legendre polynomial vanishes at 55@xmath33 , minimizing possible angular distribution effects . in order to determine the energy of the three newly discovered resonances , each resonance scan
was compared to the efficiency profile measured with the @xmath34be source .
the maxima of the efficiency profile and of the resonance scan are expected to be in the same position along the target chamber , because the two @xmath1-ray energies are very close by ( @xmath35=478kev for @xmath34be , and @xmath35=440kev for the resonance scan ) .
the resonance energies were then calculated using the relationship : @xmath36 where : * @xmath13 is the proton energy corresponding to the maximum in the yield profile of the resonance , * @xmath29 is the distance between the place of the efficiency maximum ( fig.[fig : efficiency ] ) and the right edge of the ap1 collimator ( fig.[fig : chamber ] ) , * @xmath37 is the effective length of the ap1 collimator ( fig.[fig : chamber ] ) , * @xmath38 is the proton energy loss per unit length in neon gas , and * @xmath39@xmath40 @xmath41 0.5kev is a correction introduced to account for the proton energy loss in the tube that connects the first pumping stage with the target chamber @xcite . the uncertainty of @xmath42 was estimated taking into account the systematic uncertainty on the proton beam energy ( 0.3kev @xcite ) and on the proton energy loss in neon gas ( 1.7% @xcite ) .
the overall uncertainty is two to three times smaller than the uncertainty reported in the literature @xcite .
the results on the resonance energies are shown in table [ tab : res_en ] , together with a comparison with previously reported values @xcite .
the values by jenkins _
@xcite for the excitation energy @xmath8 have been reported with 1kev uncertainty . for comparison ,
the jenkins @xmath8 data are converted to resonance energies @xmath42 using the @xmath12-value of the @xmath0ne(p,@xmath1)@xmath2na reaction , which is precisely known , @xmath12 = ( [email protected] ) kev @xcite . also , new recommended excitation energies @xmath8 are computed from the present @xmath42 values ( table [ tab : res_en ] ) .
[ tab : res_en ] .resonance energy @xmath42 in kev obtained here ( luna ) , compared to the literature values by iliadis _
_ @xcite and by jenkins _
the values for the excitation energy @xmath8 from the present work ( luna ) and from jenkins _ et al .
_ @xcite are also shown . [ cols="^,^,^,^,^",options="header " , ] ( color online ) regions of interest for the 440 kev peak ( pink ) and for the background estimation ( light green lines ) in the spectrum taken on the tentative 215 kev resonance . ]
the tentative resonances at 71 , 105 and 215 kev were searched for by setting the beam energy to populate each resonance at the target position where both detectors had maximum efficiency . because of the width of the efficiency plateau , for a given resonance energy @xmath42 this search covers a beam energy range of [ @xmath42 - 2kev;@xmath42 + 2kev ] .
the existence of the three excited states corresponding to these states had been reported as tentative in an early @xmath0ne(@xmath43he , d)@xmath2na experiment @xcite .
a later study with the same method re - investigated the 71 and 105 kev resonances but could only find upper limits @xcite .
the nuclear data sheet lists the existence of all three levels as tentative @xcite .
however , for each of the first two levels , corresponding to the 71 and 105 kev resonances , a definitive spin - parity assignment of @xmath44 = 1/2@xmath45 is given @xcite .
this value is justified @xcite with the assumption of @xmath46=0 angular momentum transfer in the @xmath0ne(@xmath43he , d)@xmath2na reaction .
however , these levels were only tentatively or not at all detected in the two @xmath0ne(@xmath43he , d)@xmath2na experiments @xcite , so the assumed @xmath44 assignment for the 71 and 105 kev resonances @xcite seems speculative and is not followed here . for the tentative 215 kev resonance , no spin - parity assignment is given @xcite . also in the present experiment
, none of the three resonances was detected . instead
, upper limits for the resonance strengths were derived assuming @xmath47 branching ratio for the 440@xmath480 kev transition from the first excited to the ground state . for the evaluation of the upper limits
the rolke method has been used @xcite , which applies the profile likelihood technique in the case of a poisson - distributed signal , taking into account the uncertainty on the background determination . in the present case
the background estimate is based on experimental data , i.e. on the number of counts in two spectral regions on the left and on the right side of the 440 kev peak .
the background thus has similar uncertainty as the signal , which is why the rolke method is used here .
[ fig : ul_spec ] shows the regions of interest adopted for the signal and for the background estimation in the case of the tentative 215 kev resonance .
the regions of interest were chosen using spectra from different , observed resonances where the 440 kev peak was clearly in evidence , and then used without change for the tentative resonances .
[ tab : wg_trr ] d | c | r |
r | r | c & & + & & & & & + & direct & & & & + 29 & - & @xmath492.6@xmath5010@xmath51@xcite & - & @xmath492.6@xmath5010@xmath51 + 37 & - & ( [email protected])@xmath5010@xmath52@xcite & - & ( [email protected])@xmath5010@xmath52 + 71 & @xmath493.2@xmath5010@xmath7@xcite & @xmath491.9@xmath5010@xmath53@xcite & @xmath491.5@xmath5010@xmath54 & @xmath491.5@xmath5010@xmath54 & 1.266 + 105 & @xmath490.6@xmath5010@xmath7@xcite & @xmath491.4@xmath5010@xmath6@xcite & @xmath497.6@xmath5010@xmath54 & @xmath497.6@xmath5010@xmath54 & 1.140 + 156.2 & @xmath491.0@xmath5010@xmath7@xcite & ( [email protected])@xmath5010@xmath54@xcite & ( [email protected]@[email protected]@xmath56)@xmath5010@xmath6 & ( [email protected])@xmath5010@xmath6 & 1.074 + 189.5 & @xmath492.6@xmath5010@xmath7@xcite & 3.4@xmath5010@xmath7@xcite & ( [email protected]@[email protected]@xmath56)@xmath5010@xmath7 & ( [email protected])@xmath5010@xmath7 & 1.055 + 215 & @xmath491.4@xmath5010@xmath7@xcite & - & @xmath492.8@xmath5010@xmath57 & @xmath492.8@xmath5010@xmath57 & 1.045 + 259.7 & @xmath492.6@xmath5010@xmath7@xcite & @xmath491.3@xmath5010@xmath6@xcite & ( [email protected]@[email protected]@xmath56)@xmath5010@xmath7 & ( [email protected])@xmath5010@xmath7 & 1.034 + 291 & @xmath492.2@xmath5010@xmath7@xcite & - & - & @xmath492.2@xmath5010@xmath7 + 323 & @xmath492.2@xmath5010@xmath7@xcite & - & - & @xmath492.2@xmath5010@xmath7 + 334 & @xmath493.0@xmath5010@xmath7@xcite & - & - & @xmath493.0@xmath5010@xmath7 + 369 & - & @xmath496.0@xmath5010@xmath58@xcite & - & @xmath496.0@xmath5010@xmath58 + 394 & - & @xmath496.0@xmath5010@xmath58@xcite & - & @xmath496.0@xmath5010@xmath58 + 436 & 0.079 @xmath4 0.006 @xcite & & - & 0.079 @xmath4 0.006 + 479 & 0.594 @xmath4 0.038 @xcite & & - & 0.594 @xmath4 0.038 + 638.5 & 2.45 @xmath4 0.18 @xcite & - & & 2.45 @xmath4 0.18 + 661 & 0.032 @xmath4 0.015 @xcite & & - & 0.032 @xmath4 0.015 + the present new upper limits are summarized in table [ tab : wg_trr ] .
owing to the ultra - low background at luna , the new limits are orders of magnitude lower than the upper limits from the previous direct experiment @xcite .
it should be noted that the limits calculated are purely experimental , different from indirect upper limits available from the literature which depend on the spin - parity assumed and on the normalization for spectroscopic factors @xcite .
an updated reaction rate has been calculated using the results from the present experiment and the most recent literature data @xcite .
the input parameters for the reaction rate calculation are listed in table [ tab : wg_trr ] and explained in the following .
the effect of cross section enhancement due to electron screening @xcite has been taken into account in the reaction rate calculation .
the resonance strength is proportional to the proton partial width @xmath59 , which , in turn , depends on the coulomb barrier penetrability @xmath60 as a function of partial wave @xmath46 and energy @xmath11 .
the enhancement of the cross section , hence the resonance strength , measured in the laboratory due to the electron screening effect @xcite can be expressed by the enhancement factor @xmath61 given by : @xmath62 where @xmath11 is the resonance energy in the center - of - mass system , @xmath63 is the screening potential , and @xmath64 is the resonance strength for a bare nucleus without atomic electrons . in the adiabatic limit , the screening potential @xmath63 is given by the difference between the total electron binding energy of neutral @xmath2na and the combined binding energies of neutral @xmath0ne and neutral @xmath65h @xcite . using tabulated values @xcite ,
@xmath63 = 0.895 kev is found . for
@xmath46 = 0 , which is usually the case at low energy , @xmath61 can be approximated as @xcite : @xmath66 \label{eq : screening_2}\ ] ] where @xmath67 is the sommerfeld parameter @xcite .
for the resonances studied here , the calculated screening enhancement factors @xmath61 in the adiabatic limit are included in table [ tab : wg_trr ] . in order to obtain the bare resonance strength @xmath64
, the strength values listed in table [ tab : wg_trr ] have to be divided by the respective @xmath61 value . for the intermediate - energy resonances at @xmath68 = 436 , 479 , 638.5 , and 661kev ,
recent experimental data @xcite have been used . for the resonances at even higher energies ,
@xmath69 661 kev ( not listed in table [ tab : wg_trr ] ) , and for the 29 and 37 kev resonances , energy and strength data from the starlib compilation @xcite are used here . using these input data ,
a monte carlo technique has been adopted to calculate the reaction rate : a random sampling of the input parameters that enter the reaction rate calculation was performed , assuming a gaussian distribution for the resonance energy , and a lognormal distribution for the resonance strength @xcite .
the upper limits on the strengths of the tentative resonances at 71 , 105 and 215 kev were also included using a monte carlo approach but taking into account the probability density function of the present experimental data : for each tentative resonance , poisson - distributed random counts in the region of interest and in the background were extracted , then a number of net counts was derived subtracting the background from the roi counts .
the resonance strength was then calculated using eq .
( [ eq : wg ] ) . for each of the temperatures in table [ tab : trr ] , the random sampling of all parameters was repeated 10@xmath70 times and the total resonant reaction rate @xmath71 was calculated for a grid of temperatures @xmath72 with the relationship @xcite : @xmath73\ ] ] where @xmath72 is the temperature in gk , @xmath74 the reduced mass in amu ( atomic mass units ) , @xmath75 the strength of resonance @xmath76 in ev , and @xmath77 the center - of - mass energy of resonance @xmath76 in mev .
the total reaction rate @xmath78 is then given by : @xmath79 with @xmath80 the direct capture contribution , assuming a constant @xmath81-factor of 62 kev@xmath5b with a relative uncertainty of 40% @xcite .
at each given temperature , the ` low ' , ` median ' and ` high ' value of the thermonuclear reaction rate were then determined using the 0.16 , 0.50 , and 0.84 quantiles of the probability density function for the total reaction rate .
the three tentative resonances at 71 , 105 and 215 kev were treated in a special manner for the ` low ' rate . for this rate ,
they were simply set to zero , with uncertainty zero , as had been done previously by iliadis / starlib @xcite . in this way , the remaining uncertainty due to the non - zero experimental upper limit led to a relatively large error band especially at @xmath72 @xmath82 0.07 , where the influence of the 105 kev resonance is greatest .
this error is much larger than the one by iliadis / starlib @xcite , but more realistic given the strong differences between indirect and direct results observed here for the 156.2 and 259.7 kev resonances .
the reaction rate is provided in tabular form in table [ tab : trr ] . in the temperature range
0.02@xmath49@xmath72@xmath492 , it can be approximated to @xmath44% by the following analytical formula : @xmath83 with coefficients : @xmath84 the contribution of the various resonances to the total reaction rate is shown in fig .
[ fig : rate_allreso ] and discussed in the following paragraphs .
at the very lowest temperatures , @xmath72 @xmath49 0.05 , the reaction rate is almost completely dominated by the 37kev resonance , whose energy and strength are so low that a direct strength measurement can not be conceived of with present technology .
it has to be noted that at these low temperatures , the nena cycle is very inefficient and hardly contributes to nucleosynthesis , when compared to the cno cycles . at somewhat higher temperatures ,
@xmath72 = 0.05 - 0.1 , the present median rate is dominated by the tentative resonance at 105kev . even though it is not observed ,
the little rate contributed by the nonzero part of the probability distribution function derived from the present upper limit is enough to dominate the recommended rate , and the error bar .
it is clear that this resonance warrants further study .
just such a measurement is presently underway at luna .
it should be noted that the 1@xmath85 lower limit of the thermonuclear reaction rate has been calculated setting this resonance to zero . at even higher temperatures ,
@xmath72 = 0.08 - 0.2 , each of the two newly observed resonances at 156.2 and 189.5 kev contributes more to the rate than the total recommended rate in the previous iliadis / starlib compilations .
the total rate enhancement is up to a factor of three , even when the tentative resonance at 105 kev is left out .
the relatively strong new resonance at 259.7 kev contributes much less to the reaction rate , only up to 10% of the previous total .
the direct capture component , assumed to be described by a constant @xmath81-factor of 62kevbarn @xcite , contributes in this temperature range , as well .
for even higher temperatures , 0.2 @xmath49 @xmath72 , a number of resonances above 400 kev play a role , and in particular three resonances recently studied at hzdr @xcite are included in the present rate .
the tentative resonances at 71 and 215 kev have been omitted from fig .
[ fig : rate_allreso ] , because they contribute less than 1% to the previous total rate at all temperatures shown here .
d l l l & low rate & median rate & high rate + 0.01 & 2.936@xmath86 & 6.915@xmath86 & 1.643@xmath87 + 0.011 & 1.135@xmath88 & 2.553@xmath88 & 5.756@xmath88 + 0.012 & 2.402@xmath89 & 5.058@xmath89 & 1.059@xmath90 + 0.013 & 3.036@xmath90 & 6.088@xmath90 & 1.226@xmath91 + 0.014 & 2.671@xmath91 & 5.233@xmath91 & 1.023@xmath92 + 0.015 & 1.762@xmath92 & 3.338@xmath92 & 6.471@xmath92 + 0.016 & 9.133@xmath92 & 1.687@xmath93 & 3.092@xmath93 + 0.018 & 1.358@xmath94 & 2.463@xmath94 & 4.432@xmath94 + 0.02 & 1.173@xmath95 & 2.053@xmath95 & 3.576@xmath95 + 0.025 & 5.423@xmath96 & 9.089@xmath96 & 1.545@xmath97 + 0.03 & 6.538@xmath97 & 1.086@xmath98 & 1.836@xmath98 + 0.04 & 1.335@xmath99 & 2.337@xmath99 & 4.519@xmath99 + 0.05 & 7.723@xmath99 & 1.811@xmath100 & 1.094@xmath101 + 0.06 & 2.450@xmath100 & 1.589@xmath101 & 2.209@xmath102 + 0.07 & 8.370@xmath100 & 1.631@xmath102 & 1.550@xmath103 + 0.08 & 5.892@xmath101 & 9.926@xmath102 & 8.124@xmath103 + 0.09 & 4.624@xmath102 & 4.544@xmath103 & 2.744@xmath104 + 0.1 & 2.783@xmath103 & 1.654@xmath104 & 7.338@xmath104 + 0.11 & 1.257@xmath104 & 4.575@xmath104 & 1.695@xmath105 + 0.12 & 4.503@xmath104 & 1.160@xmath105 & 3.348@xmath105 + 0.13 & 1.337@xmath105 & 2.690@xmath105 & 6.410@xmath105 + 0.14 & 3.413@xmath105 & 5.857@xmath105 & 1.135@xmath106 + 0.15 & 7.740@xmath105 & 1.173@xmath106 & 1.980@xmath106 + 0.16 & 1.585@xmath106 & 2.205@xmath106 & 3.259@xmath106 + 0.18 & 5.327@xmath106 & 6.688@xmath106 & 8.602@xmath106 + 0.2 & 1.466@xmath107 & 1.732@xmath107 & 2.038@xmath107 + 0.25 & 1.653@xmath108 & 1.797@xmath108 & 1.917@xmath108 + 0.3 & 2.160@xmath109 & 2.295@xmath109 & 2.429@xmath109 + 0.35 & 1.807@xmath110 & 1.910@xmath110 & 2.015@xmath110 + 0.4 & 9.285@xmath110 & 9.793@xmath110 & 1.032@xmath111 + 0.45 & 3.332@xmath111 & 3.510@xmath111 & 3.697@xmath111 + 0.5 & 9.281@xmath111 & 9.767@xmath111 & 1.028@xmath112 + 0.6 & 4.335@xmath112 & 4.549@xmath112 & 4.773@xmath112 + 0.7 & 1.320@xmath113 & 1.382@xmath113 & 1.448@xmath113 + 0.8 & 3.088@xmath113 & 3.232@xmath113 & 3.383@xmath113 + 0.9 & 6.067@xmath113 & 6.366@xmath113 & 6.680@xmath113 + 1 & 1.059@xmath114 & 1.112@xmath114 & 1.175@xmath114 + 1.25 & 2.992@xmath114 & 3.189@xmath114 & 3.430@xmath114 + 1.5 & 6.240@xmath114 & 6.749@xmath114 & 7.373@xmath114 + 1.75 & 1.081@xmath115 & 1.179@xmath115 & 1.301@xmath115 + 2 & 1.657@xmath115 & 1.814@xmath115 & 2.014@xmath115 + 2.5 & 3.046@xmath115 & 3.342@xmath115 & 3.729@xmath115 + 3 & 4.585@xmath115 & 5.032@xmath115 & 5.586@xmath115 + 3.5 & 6.113@xmath115 & 6.689@xmath115 & 7.397@xmath115 + 4 & 7.526@xmath115 & 8.187@xmath115 & 8.967@xmath115 +
a new direct study of the @xmath0ne(p,@xmath1)@xmath2na reaction has been performed deep underground at luna .
three resonances at 156.2 kev , 189.5 kev and 259.7 kev have been observed for the first time . for these resonances , new resonance strengths
@xmath3 have been measured , superseding the previous upper limits . moreover ,
new @xmath1-ray transitions and the corresponding branching ratios are provided .
two of the three new resonances observed here ( 156.2 and 259.7 kev ) have an experimental strength that is more than a factor of ten higher than a previous indirect upper limit , underlining the uncertainties involved when using indirect data .
the three new resonances lie directly in the gamow window for the hot bottom burning process in agb stars , and they also address the lower part of the gamow window for classical novae explosions .
the full astrophysical implications will be addressed in a forthcoming publication
. improved , direct upper limits for three tentative resonances at 71 , 105 and 215 kev have also been determined .
the new limits are two to three orders of magnitude lower than the limits from previous work and do not rely on spin - parity assumptions that may significantly affect the upper limits determined with indirect approaches . a new study of the @xmath0ne(p,@xmath1)@xmath2na reaction is currently ongoing at luna using a high - efficiency @xmath116 bgo detector .
in that experiment , the tentative resonances and the direct capture contribution will be investigated with unprecedented sensitivity .
44ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1086/146296 [ * * , ( ) ] link:\doibase 10.1007/s00159 - 012 - 0050 - 3 [ * * , ( ) ] \doibase 10.1051/0004 - 6361:20077205 [ * * , ( ) ] link:\doibase 10.1093/mnras / stt444 [ * * , ( ) ] http://stacks.iop.org/2041-8205/761/i=2/a=l30 [ * * , ( ) ] link:\doibase 10.1093/mnras / stt1877 [ * * , ( ) ] \doibase 10.1051/0004 - 6361:20066013 [ * * , ( ) ] link:\doibase
10.1093/mnras / stv211 [ * * , ( ) ] link:\doibase 10.1051/0004 - 6361/200913205 [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2010.16876.x [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2011.20328.x [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9474(99)00030 - 5 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2010.04.009 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2010.04.010 [ * * , ( ) ] link:\doibase 10.1088/0067 - 0049/207/1/18 [ * * , ( ) ] link:\doibase 10.1103/physrevc.92.045807 [ * * , ( ) ] link:\doibase 10.1103/physrevc.4.2030 [ * * , ( ) ] link:\doibase 10.1103/physrevc.65.015801 [ * * , ( ) ] link:\doibase 10.1103/physrevc.87.064301 [ * * , ( ) ] link:\doibase 10.1016/0375 - 9474(82)90489 - 4 [ * * , ( ) ] link:\doibase 10.1016/j.nds.2007.01.002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.252501 [ * * , ( ) ] \doibase 10.1051/0004 - 6361:20066903 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1146/annurev.nucl.012809.104526 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(03)01435 - 9 [ * * , ( ) ] link:\doibase 10.1088/0034 - 4885/72/8/086301 [ * * , ( ) ] link:\doibase 10.1140/epja / i2014 - 14179 - 5 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(02)00577 - 6 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1016/j.physletb.2004.03.092 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed .
( , , ) _ _ , @noop ph.d . thesis , ( ) _ _ , @noop ph.d . thesis , ( ) link:\doibase 10.1016/j.nimb.2010.02.091 [ * * , ( ) ] http://stacks.iop.org/1674-1137/36/i=12/a=002 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1016/0375 - 9474(83)90588 - 2 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.nima.2005.05.068 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/0092 - 640x(76)90027 - 9 [ * * , ( ) ] @noop _ _ ( , ) \doibase http://dx.doi.org/10.1016/j.nuclphysa.2010.04.008 [ * * , ( ) ] , | * background : * the @xmath0ne(p,@xmath1)@xmath2na reaction is the most uncertain process in the neon - sodium cycle of hydrogen burning . at temperatures relevant for nucleosynthesis in asymptotic giant branch stars and classical novae , its uncertainty is mainly due to a large number of predicted but hitherto unobserved resonances at low energy .
* purpose : * a new direct study of low energy @xmath0ne(p,@xmath1)@xmath2na resonances has been performed at the laboratory for underground nuclear astrophysics ( luna ) , in the gran sasso national laboratory , italy .
* method : * the proton capture on @xmath0ne was investigated in direct kinematics , delivering an intense proton beam to a @xmath0ne gas target .
@xmath1 rays were detected with two high - purity germanium detectors enclosed in a copper and lead shielding suppressing environmental radioactivity .
* results : * three resonances at 156.2 kev ( @xmath3 = ( [email protected])@xmath510@xmath6 ev ) , 189.5 kev ( @xmath3 = ( [email protected])@xmath510@xmath7 ev ) and 259.7 kev ( @xmath3 = ( [email protected])@xmath510@xmath7 ev ) proton beam energy , respectively , have been observed for the first time . for the levels at @xmath8
= 8943.5 , 8975.3 , and 9042.4 kev excitation energy corresponding to the new resonances , the @xmath1-decay branching ratios have been precisely measured .
three additional , tentative resonances at 71 , 105 and 215 kev proton beam energy , respectively , were not observed here . for the strengths of these resonances ,
experimental upper limits have been derived that are significantly more stringent than the upper limits reported in the literature .
* conclusions : * based on the present experimental data and also previous literature data , an updated thermonuclear reaction rate is provided in tabular and parametric form . the new reaction rate is significantly higher than previous evaluations at temperatures of 0.08 - 0.3 gk . |
auxetic materials are rather special and unusual elastic solids .
their poisson s ratio @xmath2 is negative , which means that it is easy to change their volume while fixing their shape , but it is hard to change their shape while fixing their volume @xcite .
this behavior is just opposite to that of an ideal liquid @xcite and to that of an ideal pentamode metamaterial @xcite . in general , auxetic materials
can be anisotropic , in which case the poisson s ratio turns into a poisson s matrix @xcite .
there are no fundamental bounds for the values of the elements of the general poisson s matrix @xcite . in sharp contrast , there are established bounds for stable elastic isotropic media . here , the poisson s ratio is connected to the ratio of bulk modulus @xmath3 ( the inverse of the compressibility ) and shear modulus @xmath4 via @xcite @xmath5 for a stable elastic solid , neither bulk nor shear modulus can be negative .
for example , exerting a hydrostatic pressure onto a material with @xmath6 would lead to an expansion , further increasing the pressure , further increasing the volume , etc .
this non - negativity together with equation ( [ b_g ] ) immediately translates into the well - known interval of possible poisson s ratios of @xmath7 $ ] .
effectively isotropic auxetic materials with @xmath8 composed of disordered polymer- or metal - based foams have extensively been studied in the literature , for a recent review see ref.@xcite .
it is not clear though how one would systematically approach , along these lines , the ultimate limit of @xmath9 .
such ultimate extreme auxetic materials are called `` dilational '' because they support strictly no other modes than dilations .
intuitively , for example , if one exerts a force onto a statue of liberty made of a dilational material at any point and along any direction , one can change its volume , but it will always maintain exactly the shape of the statue of liberty . obviously , this behavior is very much different from that of a regular elastic solid .
as an impact would be distributed throughout the entire elastic structure , dilational materials can , for example , be used as shock absorbers @xcite .
early three - dimensional auxetic metamaterials with anisotropic poisson s ratio have recently been presented @xcite .
it is again not clear though how this approach @xcite could be brought towards an isotropic behavior with @xmath10 . in the literature , several conceptual models for dilational metamaterials
have been discussed @xcite .
these , however , contain elements like `` perfect joints '' and `` rigid rods '' that still need to be translated to a three - dimensional continuous microstructure composed of one constituent material ( and vacuum in the voids ) that can be fabricated with current technology . in this paper , inspired by the two - dimensional conceptual model of ref.@xcite
, we introduce a three - dimensional blueprint for such a dilational material .
several questions arise .
does this microstructure support unwanted easy modes other than the wanted dilations ? can this microstructure at all be described by a simple elasticity tensor and a constant mass density ?
for so - called cosserat materials or for materials with anisotropic mass - density tensors @xcite , the answer would be negative .
our blueprint contains small internal connections mimicking the mentioned `` ideal joints '' .
how small do these connections have to be ? the blueprint uses a simple - cubic translational lattice .
do we really get an isotropic poisson s ratio ?
in general cubic elastic solids , the answer would be negative . to address all of these questions , we start by presenting calculated phonon band structures for our blueprint .
next , we compare these with static continuum - mechanics calculations .
these can then directly be compared with our static experiments on macroscopic polymer - based metamaterials made by three - dimensional printing .
finally , we show that also microscopic versions can be fabricated by recent advances in galvo - scanner dip - in direct - laser - writing optical lithography .
a related but different idea has recently been demonstrated by a rubbery chartreuse ball with 24 carefully spaced round dimples @xcite .
these buckliballs can also be arranged into bucklicrystals @xcite .
our three - dimensional blueprint depicted in fig.1 is based on a recently published two - dimensional conceptual model @xcite .
this model contains `` ideal joints '' and `` rigid bars '' . in our blueprint ,
the ideal joints are implemented by small connections between the square and the triangular elements . upon pushing from one side , for example from the top , the inner squares rotate and the triangular outer connection elements get pulled inwards . thus , ideally , the structure contracts laterally by the same amount as it contracts vertically .
the poisson s ratio would thus be @xmath10 .
we will have to investigate though to what extend we approach this ideal for a finite connection size @xmath11 compared to the cubic lattice constant @xmath12 .
furthermore , the very thin rigid bars in @xcite have been eliminated in our blueprint because they can not be implemented using a single constituent material . as a result , it is not clear whether unwanted easy modes of deformation might occur . indeed , in preliminary simulations
, we have found that using only one sense of rotation of the squares , the squares do not only rotate around their center , but rather also translate . to eliminate this unwanted easy mode
, we use a three - dimensional checkerboard arrangement with the discussed motif alternating with its mirror image . the small cubes with a side length identical to the thickness of all squares and triangles are not necessary for the function of the metamaterial .
they are , however , crucial as markers in our measurements of the poisson s ratio ( see below ) .
they are hence considered in all our band structure and static calculations to allow for direct comparison . .
the other parameters are : block size @xmath13 , width of the holding element @xmath14 , layer thickness @xmath15 , holder length @xmath16 , and connection size @xmath17 . ]
the phonon band structure reveals all modes of the elastic metamaterial , possibly including unwanted easy modes other than dilations ( see above )
. the long - wavelength limit of the band structure can be the starting point for a description in terms of effective elastic metamaterial parameters ( see next section ) . in our numerical band - structure calculations for the dilational metamaterial structure in fig.1
, we solve the usual elastodynamic equations @xcite for the displacement vector @xmath18 containing the time - independent rank-4 elasticity tensor @xmath19 and the scalar mass density @xmath20 , i.e. , @xmath21 by using a commercial software package ( comsol multiphysics , mumps solver ) .
we impose bloch - periodic boundary conditions onto the primitive cubic real - space cell shown in fig.[fig1 ] .
we have carefully checked that all results presented in this paper are converged .
typically , convergence is achieved by using several ten thousand tetrahedra in one primitive real - space cell .
we choose an isotropic polymer as constituent material with young s modulus @xmath22 , poisson s ratio @xmath23 , and mass density @xmath24 .
these values are chosen according to the below experiments . due to the scalability of the elastodynamic equations , our results
can easily be scaled to isotropic constituent materials with any different young s modulus and density .
the poisson s ratio of the constituent material influences the results only to a very minor degree .
the voids in the polymer are assumed to be vacuum .
the lattice constant is chosen as @xmath25 , according to our below experiments on macroscopic polymer structures .
however , the results can again easily be scaled to any other value of @xmath12 .
we represent the results in a simple - cubic brillouin zone corresponding to the underlying simple - cubic translational lattice . versus wave vector for the usual tour through the simple - cubic brillouin zone
( see inset in ( a ) ) .
the red straight lines are fits assuming a homogeneous cubic - symmetry effective medium .
( a ) connection size @xmath26 ( b ) @xmath27 ; @xmath28 .
the grey area in ( a ) highlights a complete three - dimensional elastic band gap . ]
examples of calculated band structures are depicted in fig.[fig2 ] for two different values of the ratio @xmath29 . shown
are the six lowest eigenmodes .
it becomes immediately clear that the slope of the low - frequency or long - wavelength acoustic modes is not the same for all directions , it is anisotropic . along the @xmath30-direction ( i.e. , along the principal cubic axes ) , the velocity of the longitudinally polarized mode is smaller than that of the transversely polarized modes ( see fig.[fig3 ] ) . in isotropic
elastic media , the opposite holds true . as to be expected , the phase velocities are larger for larger @xmath29 . we note in passing that the band structure in fig.[fig2 ] for @xmath26 exhibits a complete three - dimensional elastic ( i.e. , not only acoustic ) band gap between normalized frequencies of @xmath31 and @xmath32 .
this region with zero bulk phonon density of states corresponds to a gap - to - midgap ratio of @xmath33 .
this complements other possibilities reported in the literature @xcite . to further emphasize the anisotropy , we also plot the phase velocity in polar diagrams in fig.[fig3 ] .
constant phase velocity would lead to circles in the two - dimensional cuts depicted . clearly , the curves shown in fig.[fig3 ] ( a ) and ( b ) are not circular at all , neither for the small nor for the large value of @xmath29 .
they do show four - fold symmetry though . this statement is not trivial , because the three orthogonal axes are not strictly equivalent in terms of the geometrical structure .
this can be seen when comparing a view on the unit cell shown in fig.[fig1 ] onto the @xmath34- and the @xmath35-planes .
nevertheless , the band structures show that wave propagation is equivalent for the three cubic axes . ,
i.e. , the phase velocity in a particular direction is given by the radial length .
the cut on the left is for the @xmath34 plane , the cut on the right for a plane spanned by the [ 111 ] and the [ 110 ] directions .
the cut on the left shows the fourfold rotational symmetry expected for a cubic structure .
the connection size is ( a ) @xmath36 and ( b ) @xmath27 .
all other geometrical parameters are as quoted in fig.1 . the blue dots are derived from the phonon band structure , the red curves are the result of an effective - parameter description of a cubic - symmetry medium .
( c ) selected eigenmodes for a fixed wave number of @xmath37 and for @xmath36 . the corresponding eigenfrequencies increase from the left to the right .
the black arrows indicate the direction of the wave vector @xmath38 , the red arrows the directions of the displacement vector @xmath39 . shown
are the @xmath40 direction identical to the principal cubic axes , the @xmath41 direction parallel to the cubic face diagonals , as well as for an oblique direction . for the latter ,
the modes are no longer purely transversely or longitudinally polarized . ]
the anisotropic behavior of the phase velocity is connected to rather complex underlying eigenmodes that are illustrated by the examples shown in fig.[fig3 ] ( c ) . for waves propagating along the principal cubic axes or along the face diagonals
, we find pure longitudinal or pure transverse polarization . for arbitrary oblique propagation directions with respect to the principal cubic axes , the eigenmodes are complicated mixtures of transverse and longitudinal polarization . in contrast , in an ideal isotropic elastic medium
, the polarizations would be purely transverse or longitudinal .
the phonon band structures presented in the previous section have shown pronounced anisotropies resulting from the cubic symmetry of the underlying translational lattice . in this light
, one might suspect that the poisson s ratio is anisotropic as well , whereas we aim at an isotropic poisson s ratio .
we thus derive a poisson s ratio from the phonon band structure .
to do so , we compare the band structures with the expectation from continuum mechanics of an effective cubic - symmetry medium . for crystals obeying simple - cubic symmetry @xcite ,
the rank-4 elasticity tensor @xmath42 has the three different non - zero elements @xmath43 . here
, the elements with only two indices refer to voigt notation @xcite .
all other elements are zero .
furthermore , we assume a constant scalar effective metamaterial mass density @xmath44 , which is simply given by the volume filling fraction @xmath45 of the constituent material times its own bulk mass density ( see fig.1 ) . for @xmath17 ( @xmath46 )
we get @xmath47 ( @xmath48 ) . on this basis
, we can now calculate the phonon band structure in the long - wavelength limit .
to do so , one needs to connect the phase velocities with the elements of @xmath42 and with @xmath49 . here
, it is convenient to inspect the @xmath41 or [ 110 ] direction with three different phase velocities @xmath50 and three orthogonal eigenmodes that are either purely longitudinally ( l ) or purely transversely ( t ) polarized ( see above ) .
the latter either lie in the @xmath34 plane or along the @xmath51-direction . following @xcite ,
the connections are given by @xmath52 @xmath53 @xmath54 the three different elements of @xmath42 can immediately be computed from the three different phase velocities .
one has to make sure though that the polarizations of the corresponding eigenmodes are the same as for the numerical band structure calculations .
we have checked this aspect ( not depicted ) .
furthermore , one needs to make sure that the elastic behavior is described for all other propagation directions as well . to this end
, we compare in figs.[fig2 ] and [ fig3 ] the results from the phonon band structure ( blue dots ) with those of the effective - medium description ( red lines ) .
obviously , we obtain excellent overall agreement for all conditions in the long - wavelength limit .
this means that a description of the elastic metamterial in terms of an elasticity tensor @xmath42 for a cubic - symmetry effective medium is adequate . as discussed in our introduction ( see cosserat and anisotropic - mass - density metamaterials ) , this finding itself is not trivial .
having derived all non - zero elements of the effective metamaterial elasticity tensor , we can now apply established analysis to extract the young s modulus @xmath55 , the shear modulus @xmath4 , the bulk modulus @xmath3 @xcite , and the poisson s ratio ( or poisson s matrix ) @xcite .
we have @xmath56 @xmath57 @xmath58 examples are given in table 1 .
.examples of retrieved effective parameters .
the three non - equivalent non - zero elements of the elasticity tensor @xmath59 , @xmath60 , @xmath61 , the young s modulus @xmath55 , and the bulk modulus @xmath3 are given for selected values of @xmath29 .
+ [ cols="^,^,^,^,^,^",options="header " , ] the poisson s ratio for pushing along the principal cubic axes is given by @xcite @xmath62 however , the poisson s ratio might still be different for arbitrary oblique pushing directions . here , we use the more general expressions as given in ref .
@xcite , which are based on averaging along the directions normal to the pushing direction : @xmath63 } \label{eq : nuang}\ ] ] introducing the compliance tensor @xmath64 ( two indices refer to voigt notation ) , the abbreviations in ( [ eq : nuang ] ) are given by @xmath65,\label{subeqn-3:nu_definitions}\\ b&=&-11 + 4\rm { cos}{(2 \theta)}+7 cos{(4 \theta)}+8 cos{(4\phi ) } sin^4{(\theta ) } , \label{subeqn-4:nu_definitions}\\ c&=&8 \rm cos^4 { ( \theta)}+6 sin^4 { ( \theta)}+2 cos{(4\phi ) } sin^4{(\theta ) } , \label{subeqn-5:nu_definitions}\\ d&=&2\left [ \rm sin^2 { ( 2 \theta ) } + sin^4 { ( \theta ) } sin^2 { ( 2 \phi)}\right].\label{subeqn-6:nu_definitions}\end{aligned}\ ] ] here , @xmath66 and @xmath67 are the usual angles in spherical coordinates .
the resulting direction dependence of the poisson s ratio is visualized in two different ways in fig.[fig4 ] and fig.[fig5 ] .
fig.[fig4 ] is the generalized polar representation , i.e. , the poisson s ratio is proportional to the length of the vector from the origin to the depicted surface .
the poisson s ratio is also visualized by the false - color scale . for large connections , e.g. , for @xmath27 , the behavior is obviously far from isotropic .
furthermore , the effective metamaterial poisson s ratio is far from @xmath0 . for a decreasing connection size @xmath29 ,
the effective metamaterial poisson s ratio @xmath2 becomes more negative and reaches a nearly isotropic behavior at @xmath68 .
this means that experiments need to realize values of @xmath29 below @xmath69 or better .
fig.[fig5 ] depicts the derived minimum and maximum values of @xmath2 versus @xmath29 .
for the smallest numerically accessible values of @xmath29 , the poisson s ratio comes close to @xmath0 .
the data do not appear to extrapolate to exactly @xmath0 though .
we suspect that this aspect is due to the small cubes in our blueprint that we have introduced for experimental reasons ( see above ) .
we expect that the effective metamaterial poisson s ratio would converge to @xmath0 in the hypothetical limit of no cubes and @xmath70 which can not be realized experimentally though .
, i.e. , the length of the vector from the origin to the surface is proportional to modulus of the poisson s ratio .
the poisson s ratio including its sign is also indicated by the false - color scale . as the @xmath29 ratio decreases , @xmath2 becomes more negative and more isotropic , eventually approaching the ultimate limit of @xmath0 for an isotropic elastic material .
these results are derived from the band structures like exemplified in fig.[fig2 ] and for the other geometrical parameters as in fig.[fig1 ] . ] ) versus @xmath29 ratio .
the red symbols are the minima and maxima derived from the phonon band structures , the blue dots are obtained from static continuum - mechanics calculations for pushing along one of the principal cubic axes . ]
often , the poisson s ratio is measured in ( quasi-)static experiments . to derive the poisson s ratio ,
one pushes along one direction , e.g. , the @xmath51-direction , leading to a certain strain ( or relative displacement ) @xmath71 , observes or calculates the displacement along the orthogonal @xmath72-direction ( or the orthogonal @xmath73-direction ) , hence the element of the strain tensor @xmath74 , and computes the poisson s ratio according to its definition @xmath75 furthermore , experiments are based on finite - size samples , also containing a finite number of unit cells only . for metamaterials , the number of unit cells may be rather small .
we thus also investigate the question to what extend measurement artifacts are to be expected for accessible - size samples .
the numerical calculations to be presented in this section have been performed with comsol multiphysics using the structural mechanics module .
the structure s geometry is created using the cad comsol kernel .
the mesh is created within comsol multiphysics using the preset parameter values called `` normal '' meshing with settings(referring to a @xmath76 geometry size ) : maximum element size = 0.1 , minimum element size = 0.018 , maximum element growth rate = 1.5 , resolution of curvature = 0.6 , resolution of narrow regions = 0.5 .
for example , for a connection size of @xmath77 , this leads to about 90000 tetrahedral elements .
we use the mumps solver . in convergence tests
, we have verified that the derived effective metamaterial poisson s ratios are accurate to within 0.01 .
all geometrical parameters and constituent material parameters are as given above . to mimic a fictitious infinitely extended crystal
, we assume that all unit cells behave the same way ( analogous to zero wave vector in the previous section ) . for convenience ,
we choose our coordinate system such that the crystal center of mass is fixed . for pushing along one principal cubic axis
these conditions can be implemented by imposing anti - symmetric boundary conditions onto a single cubic unit cell .
this means that the normal component of the displacement vector on one surface of the unit cell is constant on this surface and equal to the negative of the normal component on the opposing surface . to investigate the linear regime , we choose strains along the pushing direction of @xmath78 .
the resulting behavior is illustrated in fig.[fig6](a ) .
the length of the ( black ) arrows is exaggerated and indicates the local displacement vectors .
the false - color scale shows the modulus of the local displacement vector .
note that the corners of the unit cell move nearly diagonally towards the center .
the poisson s ratio can immediately be computed from the components of these displacement vectors .
for example , for @xmath79 corresponding to our below experiments , we obtain @xmath80 .
we note in passing that the other points within the unit cell generally move in different directions than the corners .
this has important implications for our below experiments in that we must not evaluate the movement of all points within the unit cell but rather only of the corners . to have a finite region for imaging and tracking
, we have introduced the small cubes in fig.[fig1 ] .
the results of @xmath2 versus @xmath29 are shown in blue in fig.[fig5 ] . obviously , the agreement of these ( static ) values with those derived from the ( dynamic ) phonon band structures is good , providing further faith into the validity of our results .
the most negative dynamic results tend to be more negative than the static ones .
for example , for @xmath79 , the minimum dynamic value is @xmath81 , the static one @xmath80 .
static calculations have also been performed for a finite metamaterial sample containing @xmath82 unit cells ( see fig.[fig1 ] ) as shown in fig.[fig6](b ) to directly compare these with experiments . here ,
we assume sliding boundary conditions parallel to the pushing stamp interfaces . good agreement together with calculations for infinite crystals as above will allow us to extract poisson s ratios under these extreme conditions .
fig.[fig6](b ) is an example for a finite crystal composed of @xmath83 unit cells . to compare to the experiments to be discussed below
, one can measure the lateral strain of the left and right outer corners in the middle of the horizontal direction ( see small circles ) and divide these by the relative axial shift of the stamps .
we obtain a poisson s ratio for the finite crystal of @xmath84 , which is not too far from the one for the fictitious infinitely extended crystal of @xmath80 in panel ( a ) of the same figure .
these parameters have been chosen to match those of the experiments to be discussed next .
unit cells ( compare fig.[fig1 ] ) out of an infinite crystal pushed upon along the @xmath51-direction .
the resulting in - plane strain for an axial strain of 0.01 is depicted by the false - color scale projected onto the front surface of the cube as well as by the black arrows .
( b ) same , but for a finite crystal with @xmath82 unit cells . ]
we have fabricated macroscopic ( this section ) and microscopic ( next section ) versions of the blueprint shown in fig.[fig1 ] .
the macroscopic samples are fabricated with the printer `` objet30 '' sold by former objet , now stratasys , usa . for the metamaterial , we have used the basic polymer ink `` fullcure850 verogray '' . during the fabrication ,
however , one also needs a support material . the default is a mixture of `` fullcure850 verogray '' and `` fullcure705 support '' that we have not been able to remove from the composite .
thus , we have chosen exclusively `` fullcure705 support '' , which can be etched out in a bath of naoh base after hand cleaning .
the structure files are exported in the stl file format directly from the geometry used in the comsol multiphysics calculations and are imported into the objet printer .
the printing is done automatically in a standard process .
an example of a fabricated macroscopic metamaterial structure is shown in fig.[fig7](a ) .
the geometrical parameters are as defined in fig.[fig1 ] with @xmath79 and @xmath85 .
the constituent polymer material has a measured young s modulus of about @xmath86 .
unit cells fabricated by 3d printing , following the blueprint and the parameters given in fig.[fig1 ] .
( b ) measured lateral versus axial strain ( solid red curve ) as obtained from an image correlation approach upon pushing along the vertical @xmath51-direction .
the green circles on straight lines correspond to a poisson s ratio of @xmath84 and @xmath87 , respectively , the black circles to numerical calculations for @xmath79 .
extrapolation to an infinite three - dimensional crystal ( see previous section ) delivers a poisson s ratio of @xmath80 . ]
the measurement setup for the macroscopic samples consists of two metallic stamps and a linear stage containing a force cell .
the sliding boundary conditions ( see previous section ) are achieved by placing the watered sample between the stamps .
we have alternatively attempted to implement fixed boundary conditions by gluing the sample to the stamp by double - sided tape .
this has led to the same lateral displacements of the sample indicating very strong forces parallel to the stamps .
this suggests to us that assuming sliding boundary conditions is adequate .
we gradually push onto one stamp by moving the linear stage while fixing the other stamp and recording the images of one of the sample surfaces .
these images are taken with a canon eos 550d camera in full hd ( @xmath88 pixels ) resolution and 24 frames per second . the objective lens ( tamron sp 70 - 300 mm f/4 - 5,6 di vc usd ) is located at a distance of approximately @xmath89 to the sample
we have checked that image distortions ( e.g. , barrel - type aberrations ) are sufficiently small to not influence our experiments .
the displacements of the unit cells corners are tracked using an autocorrelation approach used previously @xcite .
multiple measurements with increasing maximum strain for a crystal composed of @xmath90 unit cells ( see fig.[fig7](a ) ) are depicted in fig.[fig7](b ) .
the graph shows the strain along the horizontal @xmath72-direction versus the strain along the axial pushing direction ( @xmath51 ) .
the solid red curve corresponds to two measurement cycles , i.e. , the sample is pushed , released and pushed and released again .
clearly , the four parts are hardly distinguishable , indicating a nearly reversible elastic behavior . from fits with straight lines
( see green dots ) we deduce a poisson s ratio of @xmath91 .
the corresponding numerically calculated strains for a finite sample with @xmath90 unit cells are shown as black dots . using the identical structure parameters but assuming an infinite crystal ( see previous section ) , we get a poisson s ratio of @xmath80 .
we note in passing that we have simultaneously measured the axial force from a load cell ( not depicted ) . comparing with theory
, we obtain a metamaterial young s modulus of @xmath92 with the above bulk young s modulus @xmath55 .
unit cells ( and a smaller one on the right - hand side ) @xmath93 .
( b ) electron micrograph of two microstructure samples with overall aspect ratios of 1:1 and 2:1 , respectively @xmath94 .
( c ) magnified view onto one unit cell of the structure , revealing details within the unit cell ( compare fig.[fig1 ] ) , @xmath95 . ]
the structures shown in the preceding section have validated our theoretical blueprint of a three - dimensional dilational metamaterial but they hardly qualify as a `` material '' in the normal sense .
thus , it is interesting to ask whether corresponding structures with lattice constants @xmath12 that are two to three orders of magnitude smaller are in reach .
also , it would be highly desirable to obtain structures containing a larger total number of unit cells .
we have thus also fabricated microscopic structures based on the same blueprint ( without the small blocks for tracking ) . for fabricating such microscopic dilational metamaterial samples ,
photoresist samples are prepared by drop - casting the commercially available negative - tone photoresist `` ip - dip '' ( nanoscribe gmbh , germany ) on diced silicon wafers ( @xmath96 ) .
we use the commercial direct laser writing ( dlw ) system photonic professional gt ( nanoscribe gmbh , germany ) . in this instrument ,
the liquid photoresist is polymerized via two - photon absorption using a @xmath97 frequency - doubled erbium fiber laser with a pulse duration of @xmath98 . to avoid depth - dependent aberrations ,
the objective lens ( with numerical aperture @xmath99 or @xmath100 , carl zeiss ) is directly dipped into the resist .
the laser focus is scanned using a set of pivoted galvo mirrors .
structural data are again created in stl file format using the open - source software blender and comsol multiphysics . due to the demanding critical distances of the mechanical metamaterials ,
the scan raster is set to @xmath101 ( @xmath102 ) laterally and @xmath103 ( @xmath104 ) axially for the @xmath99 ( @xmath100 ) objective lens .
each individual layer is scanned in the so - called skywriting mode , i.e. , while the laser focus is scanned continuously , the laser power is switched between @xmath105 ( no exposure ) and about @xmath106 or higher ( exposure ) to build up the fine features of the metamaterial .
the writing speed is set to @xmath107 .
after dlw of the preprogrammed pattern , the exposed sample is developed for 20 minutes in isopropanol and acetone .
the process is finished in a supercritical point dryer to avoid capillary forces during drying .
optical and electron micrographs of different samples are depicted in fig.[fig8 ] .
the sample in panel ( a ) has overall dimensions of @xmath108 ( @xmath109 unit cells ) yet , at the same time , minimum feature sizes in the sub - micron range .
due to the smaller lattice constants , it is not possible though to resolve the details within the unit cell to track the positions of the small marker cubes for measuring the poisson s ratio as done for the macroscopic samples .
we have designed , fabricated , and characterized a three - dimensional microstructure based on a simple - cubic translational lattice that effectively acts as an auxetic , converging for small internal connections @xmath70 to the ultimate limit of an isotropic three - dimensional dilational metamaterial of @xmath10 .
our experiments approach this limit .
interestingly , the poisson s ratio becomes isotropic in the limit @xmath70 , whereas the acoustic phase velocity and other elastic properties remain anisotropic .
if fabricated in larger volumes and composed of different constituent materials , such dilational metamaterials might find applications in terms of shock absorbers . in our treatment
, we have derived all elements of the effective metamaterial elasticity tensor and hence all elastic parameters by comparing the phonon band structure in the long - wavelength limit with continuum mechanics of homogeneous media .
this parameter retrieval could be of interest for other cubic - symmetry elastic metamaterials beyond the specific example discussed here .
we thank the dfg - center for functional nanostructures ( cfn ) , the karlsruhe school of optics @xmath110 photonics ( ksop ) and the national science foundation through grant dms-1211359 for support . | dilational materials are stable three - dimensional isotropic auxetics with an ultimate poisson s ratio of @xmath0 .
we design , evaluate , fabricate , and characterize crystalline metamaterials approaching this ideal .
to reveal all modes , we calculate the phonon band structures . on this basis , using cubic symmetry , we can unambiguously retrieve all different non - zero elements of the rank-4 effective metamaterial elasticity tensor , from which all effective elastic metamaterial properties follow . while the elastic properties and the phase velocity remain anisotropic , the effective poisson s ratio indeed becomes isotropic and approaches @xmath0 in the limit of small internal connections .
this finding is also supported by independent static continuum - mechanics calculations . in static experiments on macroscopic polymer structures fabricated by three - dimensional printing , we measure poisson s ratios as low as @xmath1 in good agreement with theory .
microscopic samples are also presented .
_ keywords _ : mechanical metamaterials , acoustics , auxetics , dilational metamaterials |
when an assembly of particles are placed in an asymmetric potential , a net dc particle flow can arise due to a ratchet effect that occurs when an ac drive is applied or when the substrate is periodically switched on and off in the presence of a thermal bath @xcite .
ratchet effects on asymmetric substrates have been extensively studied in colloidal systems @xcite , granular matter @xcite , and polymers @xcite .
ratchet effects also appear in ac - driven vortices in type - ii superconductors in the presence of an asymmetric substrate @xcite , such as a quasi - one - dimensional periodic array produced by asymmetrically modulating the sample thickness @xcite , etching funnel - shaped channels for vortex flow @xcite , introducing asymmetry to the sample edges @xcite , or adding periodic pinning arrays in which the individual pinning sites have some form of intrinsic asymmetry @xcite . at lower vortex densities when collective interactions between vortices are weak , the ratchet effect produces a dc flow of vortices in the easy flow direction of the asymmetric substrate ; however , when collective effects are present it is possible to have reversals of the ratchet effect where for one set of parameters the vortices move in the easy direction while for another set of parameters they move in the hard direction @xcite .
a ratchet effect can also be produced by a pinning array containing symmetric pinning sites arranged with a density gradient .
@xcite first studied vortex ratchet effects for random gradient array pinning geometries and found that the vortices undergo a net dc flow in the easy direction .
experiments and simulations later showed that for a square array of pinning sites with constant pinning density but with a gradient in pinning site size , a variety of forward and reverse vortex ratchet behaviors occur @xcite .
experiments on triangular pinning arrays with a density gradient also revealed a forward ratchet effect at low fields with a reversal at higher fields @xcite . .
pinning gradients run along the @xmath0 direction .
( a ) conformal pinning ( conf ) .
( b ) random pinning with a periodic gradient ( randg ) .
( c ) square pinning with a gradient ( squareg ) .
for each sample we apply an ac drive along the @xmath0 direction and measure the average net vortex displacements @xmath1 in the @xmath0 direction .
, width=336 ] recently a new type of pinning geometry was proposed that is constructed by conformally transforming a triangular pinning lattice to create what is called a conformal pinning array , abbreviated conf in this work @xcite . as in the original triangular lattice ,
each pinning site in the transformed array has six neighbors separated by 60@xmath2 ; however , the distance to each neighbor is no longer constant , producing a density gradient in the pinning sites @xcite . in fig .
[ fig:1](a ) we show a periodic lattice composed of three conformal pinning arrays with the same orientation .
experimental structures with nearly conformal geometries have been observed for magnetically interacting particles subjected to a gravitational force , and due to the arching nature of the conformal array , the magnetic conformal crystals were dubbed gravity s rainbow structures @xcite .
similar conformal structures have also been studied in foams @xcite and in charged particle ordering in confined geometries @xcite . in the superconducting system , an arrangement of two conf arrays placed with their minimum pinning density regions in the center of the sample produces an enhanced critical current or depinning force compared to an equivalent number of pinning sites placed in a uniform periodic , uniform random , or random density gradient array @xcite .
the enhancement results both from the natural density gradient formed by the vortices as they enter the sample from the edges and form a bean state , and from the preservation of local six - fold ordering in the conf array @xcite .
the suppression of easy vortex flow channels which arise in periodic and random arrays by the arching conformal structure also plays a role in the enhancement . only at the integer matching fields do the periodic pinning arrays produce higher critical currents than the conf array @xcite .
subsequent experimental studies confirmed that the conf array produces enhanced pinning over a wide range of fields compared to uniform periodic pinning or uniform random pinning arrangements @xcite .
other experiments have shown enhanced critical currents in systems with periodic arrays when a gradient in the pinning density is introduced @xcite .
there have also been studies of hyperbolic - tesselation arrays which have a gradient in the pinning density @xcite .
since conformal pinning arrays have an intrinsic asymmetry , it is natural to ask whether a ratchet effect can occur under application of an ac drive , and if so , whether this ratchet effect would be enhanced compared to other pinning array geometries with density gradients , or whether ratchet reversals could be possible . in fig . [ fig:1 ] we illustrate three examples of the gradient pinning array geometries we consider in this work : conformal pinning ( conf ) [ fig .
[ fig:1](a ) ] , random pinning with a gradient ( randg ) [ fig .
[ fig:1](b ) ] , and a square pinning array containing a gradient along the @xmath0 direction ( squareg ) [ fig .
[ fig:1](c ) ] . in each case , both the gradient and the applied ac drive are along the @xmath0 direction , while the easy - flow direction for vortex motion is in the negative @xmath0 direction .
in addition to superconducting vortex realizations of these geometries , similar pinning arrangements could also be created in colloidal systems using optical trap arrays . vs time measured in ac drive cycles for the systems in fig . 1 at @xmath3 , @xmath4 , and @xmath5 bottom black curve : conf array ; middle green curve : squareg array ; upper red curve : randg array . here
the conf array produces a ratchet that is four times more effective than the squareg array and 20 times more effective than the randg array .
, width=336 ]
we consider a two - dimensional system with periodic boundary conditions in the @xmath0 and @xmath6 directions .
the sample size is @xmath7 with @xmath8 , where distance is measured in units of the london penetration depth @xmath9 .
the applied magnetic field is perpendicular to the system in the @xmath10 direction .
our results apply to the london limit regime in which the vortices can be treated as rigid objects , when the coherence length @xmath11 is much smaller than @xmath9 .
the pinning sites are modeled as in previous studies of conformal pinning arrays @xcite by non - overlapping parabolic circular traps with radius @xmath12 and a maximum pinning force of @xmath13 .
we place the @xmath14 pinning sites in a conformal array ( conf ) , as described in previous work @xcite , in a random arrangement with a gradient ( randg ) , or in a square array with a density gradient along the @xmath0 direction ( squareg ) .
the width of each pinning array segment is @xmath15 , and the segments are repeated three times across the sample as shown in fig .
[ fig:1 ] . the total density of the pinning sites is @xmath16 .
the sample contains @xmath17 vortices and we measure the magnetic field in units of @xmath18 , where @xmath19 is the matching field at which there is one vortex per pinning site .
we obtain the initial vortex configuration by annealing from a high temperature molten state and cooling to @xmath20 or to a low but finite fixed temperature .
after annealing , we apply an ac driving force to all the vortices .
the dynamics of an individual vortex @xmath21 is obtained by integrating the following overdamped equation of motion : @xmath22 here @xmath23 is the damping constant which is set equal to 1 .
the repulsive vortex - vortex interaction force is given by @xmath24 , where @xmath25 is the location of vortex @xmath21 , @xmath26 is the modified bessel function , @xmath27 , @xmath28 , @xmath29 , @xmath30 is the flux quantum , and @xmath31 is the permittivity .
the vortex - pinning interaction force is @xmath32 , where @xmath33 is the heaviside step function , @xmath34 is the pinning radius , @xmath13 is the pinning strength , @xmath35 is the location of pinning site @xmath36 , @xmath37 , and @xmath38 .
all forces are measured in units of @xmath39 and lengths in units of @xmath9 .
thermal forces are represented by langevin kicks @xmath40 with the properties @xmath41 and @xmath42 , where @xmath43 is the boltzmann constant
. the ac driving force is @xmath44 where @xmath45 is the ac amplitude . to characterize the ratchet effect we measure the average net displacement of all vortices from their starting positions as a function of time , @xmath46 , where @xmath47 is the @xmath0 position of vortex @xmath21 at time
@xmath48 and @xmath49 is an initial reference time .
this measure produces a sinusoidal signal , as shown in fig .
[ fig:2 ] ; the presence of a net drift indicates that a ratchet effect is occurring .
we condense this information into a single number @xmath50 , the value of @xmath1 at @xmath51 ac drive cycles .
except where otherwise noted , we consider a fixed ac frequency of @xmath52 and a time step of @xmath53 , so that a single drive cycle has a period of 8000 simulation time steps .
in fig . [ fig:2 ] we plot the average net displacement @xmath1 versus time for conf , randg , and squareg arrays with @xmath54 , @xmath55 , and @xmath56 during 100 ac drive cycles . here
the overall drift of each curve indicates that all the arrays produce a ratchet effect with the vortices translating in the negative @xmath0 direction .
the conf array generates the largest ratchet effect , with the vortices translating distances up to @xmath57 per drive cycle .
the ratchet effect for the conf array is about four times larger than that of the squareg array and 20 times larger than that of the randg array . , @xmath58 , and @xmath59 , highlighting the enhanced effectiveness of the ratchet mechanism in the conf array . (
a ) the trajectories for the positive half of the ac drive cycle in the randg array showing the formation of disordered flow channels .
( b ) the negative half of the ac drive cycle in the randg array has a similar pattern and density of flow channels .
( c ) in the positive half of the ac drive cycle for the conf array , the vortices can not move past the densely pinned region .
( d ) in the negative half of the ac drive cycle in the conf array , the vortices can easily funnel between the arches in the conformal array .
( e ) in the positive half of the ac drive cycle for the squareg array , vortices can slip through the interstitial regions between pinned vortices .
( f ) similar interstitial motion occurs in the negative half of the ac drive cycle for the squareg array .
, width=336 ] the relative effectiveness of the different arrays can be more clearly understood by plotting the trajectories of the vortices during the positive and negative portions of a single ac cycle .
figure [ fig:3](a , b ) shows the trajectories for both halves of the ac cycle in a a randg array with @xmath3 , @xmath58 and @xmath59 . under both positive and negative drive ,
the vortices form disordered flow paths with a similar density that is independent of the driving direction .
in contrast , in fig . [ fig:3](c ) during the positive portion of the ac driving cycle for the conf array , almost no vortices can cross the densely pinned regions of the sample ; instead , the vortices either become trapped at pining sites or remain localized in interstitial cages formed by the pinned vortices .
figure [ fig:3](d ) shows that in the negative portion of the cycle for the conf array , numerous vortices move into the interstitial regions and funnel through the conformal arch structures , producing significant vortex motion in the negative @xmath0 direction . in the squareg system ,
[ fig:3](e , f ) show that vortex motion is strongly suppressed for both directions of drive and occurs only when interstitial vortices manage to squeeze between equally spaced occupied pinning sites .
the barrier to this type of vortex motion is the same in each half of the cycle .
in contrast , for the conf array the perpendicular spacing between pinned vortices in the sparse portion of the array is larger than the equivalent spacing between pinned vortices in the dense portion of the array , so the interstitial vortices experience much different effective caging barriers when entering the sparse side of the array than when entering the dense side of the array . in the randg arrays , channels of easy vortex flow occur somewhere in the sample with equal probability for both the positive and negative portions of the ac drive cycle . , the average net displacement per vortex after 50 ac drive cycles , vs @xmath13 for the conf ( red circles ) , randg ( yellow squares ) , and squareg ( green diamonds ) arrays . here
@xmath60 and @xmath61 . in general , the ratchet effect is suppressed for weak pinning and for strong pinning .
( a ) at @xmath3 , the conf array exhibits the strongest ratchet effects , followed by the squareg array .
the randg array has the weakest ratchet effect .
( b ) at @xmath62 the ratchet effect extends to higher values of @xmath13 in all the systems .
the conf ratchet is still the most effective .
( c ) at @xmath63 , the squareg ratchet is more effective than the conf or randg ratchets .
, width=336 ] in fig .
[ fig:4](a ) we plot @xmath50 , the average net displacement per vortex after 50 ac drive cycles , versus @xmath13 for conf , randg , and squareg samples with @xmath3 and @xmath60 . for weak pinning @xmath64 , the vortices move elastically and easily slide over the pinning sites so that there is no ratchet effect in any of the arrays . for @xmath65
most of the vortices become increasingly pinned and the ratchet effect is reduced .
the optimal ratchet effect occurs for the conf array at @xmath66 , where there is a mixture of pinned vortices coexisting with vortices that move temporarily through the interstitial regions as illustrated in fig .
[ fig:3](c , d ) .
the squareg array has a weaker ratchet effect in the range @xmath67 , with a relatively sharp cutoff at the upper end of this range that occurs when the ability of the pinned vortices to shift inside the pinning sites is reduced , preventing the interstitial vortices from slipping between occupied pinning and causing the motion to become localized , as shown in fig .
[ fig:3](e , f ) .
there is a weak ratchet effect for the randg array with an extremum at @xmath68 where the combination of the ac drive and the vortex - vortex interactions causes a portion of the vortices to depin .
the maximum magnitude of the ratchet effect for the randg array is smaller than that for the squareg array ; however , the effect occurs over a wider range of @xmath13 .
figure [ fig:4](b ) shows @xmath50 versus @xmath69 for @xmath62 , where there are more interstitial vortices . here the range of @xmath13 over which the ratchet effect occurs for the conf and randg arrays extends up to @xmath70 , with the ratcheting for @xmath71 completely dominated by the flow of interstitial vortices .
for the squareg array the ratchet effect is lost for @xmath72 , the point at which the interstitial vortices can no longer slip through the one - dimensional interstitial channels of the array . in fig .
[ fig:4](c ) we plot @xmath50 versus @xmath69 at @xmath63 , where there are few interstitial vortices . here most of the motion occurs when vortices jump from one pinning site to another .
the ratchet effect for all three arrays vanishes for @xmath73 when vortex hopping is suppressed . at this vortex density ,
the ratchet effect is most pronounced for the squareg array , where the vortices are able to hop along one - dimensional channels of pinning sites . vs ac amplitude @xmath45 for conf ( red circles ) , randg ( yellow squares ) , and squareg ( green diamonds ) arrays with @xmath4 and @xmath61 .
( a ) at @xmath3 , the ratchet effect is reduced at low @xmath45 when the vortices are pinned as well as at higher @xmath45 when the vortices move rapidly over the pinning array .
( b ) @xmath62 .
( c ) at @xmath63 the squareg array produces the most effective ratchet .
, width=336 ] in fig .
[ fig:5](a ) we plot @xmath50 versus the ac drive amplitude @xmath45 for conf , randg , and squareg samples with @xmath4 and @xmath3 . for @xmath74 , the vortices are mostly pinned and the ratchet effect is absent for all three of the pinning geometries . at intermediate @xmath45
the conf array has the strongest ratchet effect , with an extremum in @xmath50 at @xmath75 .
the squareg array has the next most effective ratchet effect , with an optimal magnitude at @xmath76 . for higher values of @xmath45 ,
the vortices are all in motion during some portion of the driving cycle and the ratchet effect gradually decreases to zero with increasing @xmath45 . we observe a similar trend at @xmath62 as shown in fig .
[ fig:5](b ) . here
the ratchet effect for the conf array extends up to much larger values of @xmath45 ; however , the maximum value of @xmath77 is slightly smaller than for the @xmath3 case . for @xmath63 in fig .
[ fig:5](c ) , the dominant motion is hopping of vortices from pinning site to pinning site . here
the ratchet effect is strongest for the squareg array , similar to what is shown in fig .
[ fig:4](c ) .
vs @xmath18 for conf ( red circles ) , randg ( yellow squares ) , and squareg ( green diamonds ) arrays with @xmath60 .
( a ) at @xmath4 the ratchet effect is negative for the entire range of @xmath18 .
( b ) at @xmath58 there is a reversal in the ratchet effect for the conf array for @xmath78 .
, width=336 ]
in fig . [ fig:6](a ) we plot @xmath50 versus @xmath18 for conf , randg , and squareg arrays with @xmath4 and @xmath60 . here
the conf array outperforms the randg array for all fields and the squareg array for @xmath79 .
for @xmath80 the vortex - vortex interactions become dominant and the ratchet effect is suppressed in all the arrays . in the squareg array , due to the periodic ordering along the @xmath6 direction , some commensuration effects occur , such as enhanced pinning near @xmath3 which locally suppresses the ratchet effect . vs @xmath18 for the conf array in fig .
[ fig:6](b ) highlighting the vortex ratchet reversal effect from negative for @xmath81 to positive for @xmath82 to negative again at higher fields .
( b ) @xmath1 vs time in ac drive cycle numbers for the system in ( a ) at @xmath83 ( lower blue curve ) where the vortex motion is in the negative @xmath0 direction and at @xmath84 ( upper red curve ) where the motion is in the positive @xmath0 direction .
, width=336 ] figure [ fig:6](b ) shows @xmath50 versus @xmath18 for @xmath58 . in this case , the ratchet effect for the squareg array is lost for @xmath85 when the vortices become strongly pinned at the pinning sites . in general
, the ratchet effect for the conf array is stronger than that for the squareg and randg arrays , with a local extremum for the ratchet effect in the negative or normal direction occurring at @xmath86 .
the squareg array has a local extremum in @xmath50 in the negative direction at @xmath62 , followed by a sharp drop in @xmath50 for @xmath87 .
we find a ratchet reversal in the conf array , where @xmath50 switches from negative to positive over the range @xmath88 .
there is a local maximum in the positive ratchet effect at @xmath84 . in fig .
[ fig:7](a ) we show a highlight of @xmath50 versus @xmath18 from fig .
[ fig:6](b ) for the conf array indicating that two reversals in the ratchet effect occur . figure [ fig:7](b ) illustrates @xmath1 vs time in ac drive cycles for the system in fig .
[ fig:7](a ) at @xmath83 , where the motion is in the negative @xmath0 direction , and at @xmath84 , where the motion is in the positive @xmath0 direction , showing more clearly the change in the net direction of vortex motion .
vs @xmath45 for conf arrays with @xmath58 .
( a ) at @xmath84 , there is a transition from a negative ratchet effect at low @xmath45 to a positive ratchet effect , followed by a second transition back to a negative ratchet effect .
( b ) at @xmath89 the ratchet effect is always negative ; however , there is a local minimum and a local maximum of the ratchet effect .
( c ) at @xmath90 , the ratchet effect is always negative and has few features .
, width=336 ] in fig .
[ fig:8](a ) we plot @xmath50 versus @xmath45 for a conf array with @xmath91 at @xmath84 , where there are multiple reversals in the ratchet effect . for @xmath92 there is no ratchet effect since the vortices move only small distances .
a negative ratchet effect occurs for @xmath93 , while for @xmath94 there is a positive ratchet effect with a maximum amplitude at @xmath60 .
there is another transition to a weaker negative ratchet effect for @xmath95 , and @xmath50 gradually approaches zero for high values of @xmath45 .
figure [ fig:8](b ) shows that at @xmath89 , the ratchet effect is always negative ; however , there are still local features in the response such as at @xmath96 where the negative ratchet effect is strongly reduced . in fig .
[ fig:8](c ) at @xmath90 , the ratchet effect is strongly negative with an extremum in @xmath50 near @xmath97 .
the ratchet effect goes to zero for increasing @xmath45 .
vs @xmath13 for conf arrays with @xmath84 .
( a ) at @xmath60 there are multiple reversals as @xmath13 increases .
( b ) at @xmath98 there are again multiple reversals and the positive ratchet effect extends over a wider range of @xmath13 .
( c ) at @xmath99 there is a weak negative ratchet effect .
, width=336 ] in fig .
[ fig:9](a ) we show @xmath50 versus @xmath13 for a conf array at @xmath84 and @xmath60 .
there is a negative ratchet effect for @xmath100 , a positive ratchet effect for @xmath101 , and a much larger negative ratchet effect for @xmath102 . at intermediate @xmath69
when there is a positive ratchet effect , vortices can be temporarily trapped by pinning sites . the negative ratchet effect for large
@xmath69 arises from the interstitial flow of vortices , and @xmath50 saturates at large @xmath69 since the caging barrier experienced by interstitial vortices from the neighboring pinned vortices does not increase with increasing @xmath69 . in fig .
[ fig:9](b ) , at @xmath98 there is a negative ratchet effect for @xmath103 , a positive ratchet effect for @xmath104 , and another negative ratchet regime for @xmath105 .
the vortices at the pinning sites remain permanently pinned for @xmath105 .
the positive ratchet effect is larger and extends out to higher values of @xmath69 for the @xmath106 system compared to the @xmath107 system .
figure [ fig:9](c ) shows that at @xmath99 , there is a weak negative ratchet effect for all values of @xmath13 .
although we focus here on the conf array , we also found that some weak ratchet reversals are possible in the squareg array ; however , we did not observe a vortex ratchet reversal for the randg array . , the fraction of sixfold - coordinated particles , vs time in ac drive cycles for conf arrays from fig .
[ fig:9](a ) with @xmath108 and @xmath107 .
light orange lines indicate the phase of the drive cycle .
mp is the maximum positive drive and mn is the maximum negative drive .
( a ) at @xmath109 the ratchet effect is negative .
the system is most ordered whenever the magnitude of the ac drive is maximum ; however , the ordering peaks for the negative portions of the drive cycle are slightly higher than those for the positive portions of the drive cycle .
( b ) at @xmath58 the ratchet effect is positive .
the system is most ordered whenever the magnitude of the ac drive is close to zero , but the net motion is determined by the relatively larger ordering at mp points compared to mn points .
, width=336 ] in order to better understand the vortex dynamics and ordering during an individual ac cycle , in fig .
[ fig:10 ] we plot the time series of the fraction of sixfold - coordinated vortices , @xmath110 , versus time . here
@xmath111 , the coordination number of vortex @xmath21 , is obtained from a voronoi construction .
superimposed over the plot is a curve showing the phase of the ac drive , and the points at which the drive reaches its maximum positive value are marked mp while those at which the drive reaches its maximum negative value are marked mn .
figure [ fig:10](a ) shows @xmath112 versus time for the system from fig .
[ fig:9](a ) with @xmath84 and @xmath113 at @xmath109 where there is a negative ratchet effect . here @xmath114 at the start of each drive cycle when the drive magnitude is zero , deceases slightly when the drive becomes positive and the system disorders , then reaches its highest values of @xmath115 in the mp portions of the drive cycle and @xmath116 in the mn portions of the drive cycle .
when the magnitude of the ac drive is maximum , all the vortices move elastically , and since they are slightly more ordered during the negative cycle of the drive , they can slide slightly further in the negative @xmath0 direction than in the positive @xmath0 direction , giving a negative ratchet effect . at @xmath58 in fig .
[ fig:10](b ) , the ratchet effect is positive and the vortex ordering is reversed .
the vortices are now the most ordered when the magnitude of the ac drive is close to zero , and they are disordered when the ac drive magnitude reaches a maximum . during the mp portion of the drive cycle , @xmath117 , while in the mn portion of the drive cycle the system is more disordered with @xmath118 .
the more ordered vortices are able to slide slightly further in the positive @xmath0 direction , resulting in a net positive ratchet effect .
there is also an asymmetry in the ordering at the zero force portions of the drive cycle .
the value of @xmath119 at cycle times of @xmath120 and @xmath121 is smaller than that at times of @xmath122 and @xmath123 ; however , since the vortices are not moving during this portion of the cycle , this asymmetry does not produce a preferred direction of motion .
in general we find that the ordering of the vortices at the mp and mn points of the drive determines the direction of the ratchet motion , with the net ratchet effect occurring in whichever drive direction generates the most ordered vortex arrangement .
vs @xmath124 for conf ( red circles ) , randg ( yellow squares ) , and squareg ( green diamonds ) arrays with @xmath3 and @xmath60 .
( a ) at @xmath4 , thermal effects reduce the ratchet effect .
( b ) at @xmath125 thermal effects can increase the ratchet effect over a range of @xmath124 .
, width=336 ] we next consider thermal effects on the ratchet response . for weak
pinning , the addition of thermal fluctuations monotonically decreases the ratchet effect for all three geometries , as shown in fig .
[ fig:11](a ) for @xmath4 , @xmath60 , and @xmath3 . as before ,
the ratchet effect is most pronounced for the conformal array . as @xmath13 increases ,
the vortices become more strongly pinned , and the addition of thermal fluctuations can increase the ratchet effect by permitting vortices to escape from pinning sites or interstitial caging sites via thermal activation . in fig . [
fig:11](b ) we plot @xmath50 versus @xmath124 in samples with @xmath69 increased to @xmath126 , showing a strong ratchet effect in the conf array . here
, the ratchet effect is lost at small @xmath124 since the vortices are strongly pinned , and the ratchet effect also disappears for high values of @xmath124 when the thermal fluctuations become so strong that the vortices enter a molten state that interacts too weakly with the substrate for an asymmetry in the response to positive and negative drives to be noticeable .
the largest ratchet signatures appear for intermediate @xmath124 . in general ,
when @xmath13 increases , the point at which the magnitude of the ratchet effect is largest shifts to higher values of @xmath127 . versus ac cycle period in simulation time steps for conf ( red circles ) , randg ( yellow squares ) , and squareg ( green diamonds ) arrays with @xmath3 , @xmath60 , and @xmath4 .
@xmath50 increases linearly with the drive cycle period .
, width=336 ] we find that @xmath50 increases linearly as the period of the ac driving cycle increases , as illustrated in fig .
[ fig:12 ] for conf , randg , and squareg arrays with @xmath3 , @xmath4 , and @xmath60 .
as the other parameters are varied , we find a robust increase in the magnitude of the ratchet effect with decreasing ac frequency .
vs time measured in ac drive cycles for colloidal particles interacting with a conf array ( bottom black curve ) , a squareg array ( middle green curve ) , and a randg array ( upper red curve ) with @xmath128 , @xmath4 , @xmath129 , and @xmath130 . as in the vortex case shown in fig .
[ fig:2 ] , the conf array produces the strongest ratchet effect . , width=336 ] vs colloid - colloid interaction strength @xmath131 for the conf colloid system in fig .
[ fig:13 ] with @xmath128 , @xmath4 , and @xmath129 .
( b ) the corresponding fraction of sixfold coordinated colloids @xmath112 vs @xmath131 . here
the maximum ratchet effect occurs when @xmath132 , indicating that although colloid - colloid interactions remain important , the system is in a disordered state .
when the system forms a crystalline state with @xmath133 , the ratchet effect disappears .
, width=336 ] ratchet effects can be generated in systems of colloidal particles interacting with various types of periodic arrays of traps that are created using optical means @xcite .
the ability to make structures similar to conformal lattices has been demonstrated by xiao _
@xcite , who examined a colloidal ratchet effect on optical traps forming fibonacci spirals . in that case
the ratchet effect is induced by rotating the potential through a three - step cycle .
here we consider the ac - driven motion of colloidal particles over a conf array .
the equation of motion for colloids is similar to that given in eqn . 1 for vortices , except that the pairwise repulsive colloid - colloid interaction potential has the form @xmath134 , where @xmath135
, @xmath136 is the dimensionless interaction strength , @xmath137 is the effective charge of the colloidal particles , @xmath138 is the solvent dielectric constant , and @xmath139 is the screening length which we set equal to @xmath121 .
the number of colloids in the sample is @xmath140 . for our parameters ,
the interactions between colloids for @xmath141 is much larger than the interactions between vortices separated by the same distance , while for @xmath142 the colloidal interaction strength falls off much more rapidly than the vortex - vortex interaction strength , so that nearest neighbor interactions are dominant in the colloidal system .
we have conduced a series of simulations for colloidal particles moving through conf , randg , and squareg pinning landscapes under an ac driving force , and find results very similar to those obtained in the vortex system .
for example , in fig .
[ fig:13 ] we plot @xmath143 versus time for colloids interacting with conf , squareg , and randg arrays for @xmath128 , @xmath144 , @xmath129 , and @xmath130 , where we observe that just as in the vortex case , the conf array produces the most pronounced ratchet effect , the squareg array shows a weak ratchet effect , and the randg array does not exhibit a ratchet effect . since the effective charge on the colloids
can be changed experimentally , it is possible to hold the substrate strength fixed and modify how strongly the colloids interact with one another . in fig .
[ fig:14](a ) we plot @xmath50 versus @xmath131 for the conf array from fig . [ fig:13 ] , and in fig .
[ fig:14](b ) we show the corresponding fraction of sixfold coordinated colloids @xmath112 versus @xmath131 . for @xmath145
the colloids all become pinned in the pinning sites since @xmath146 .
as @xmath131 increases , the colloid - colloid interactions become important and a ratchet effect arises with a maximum amplitude near @xmath147 .
the largest ratchet effect is associated with a sixfold ordering fraction of @xmath132 , indicating that the colloids are still disordered , with some colloids trapped in pinning sites and others occupying interstitial regions between pins . for @xmath148 the ratchet effect begins to diminish with increasing @xmath149 while simultaneously @xmath112 increases , indicating an increase in the ordering of the colloids . for @xmath150 ,
the colloids form a rigid triangular lattice as indicated by the fact that @xmath133 , and the ratchet effect disappears .
these results show that in order for a ratchet effect to appear in the gradient pinning arrangements , it is generally necessary for plasticity or defects in the colloid or vortex lattice to appear .
our results with the colloidal system indicate that pronounced ratchet effects should be realizable in a variety of systems where assemblies of interacting particles are driven with an ac drive over conformal array substrates .
we examine ratchet effects for ac driven vortices interacting with a conformal pinning array and with square and random pinning arrays containing a gradient along one direction . in general
, the conformal pinning array produces the most pronounced ratchet effect , particularly for fields greater than the first matching field .
the enhanced effectiveness of the conformal ratchet results in part from the fact that the pinning sites in the low density portion of the array are widely spaced not only parallel to but also perpendicular to the net pinning gradient direction , permitting the easy flow of interstitial vortices through the sparse portion of the array .
in contrast , for the square pinning array with a gradient , the perpendicular distance between pinning sites is constant throughout the array , producing the same barrier for interstitial motion in both the sparse and dense portions of the array and reducing the relative magnitude of the ratchet effect for fields at which interstitial vortices are present . for the random pinning array with a gradient , channels of easy vortex flow form for driving in either direction , significantly reducing the effective asymmetry of the array .
we find that the conformal array exhibits a series of vortex ratchet reversals as a function of vortex density , ac drive amplitude , and pinning strength , and we show that the direction of the ratchet is determined by the amount of order present in the vortex lattice at different phases of the ac driving cycle . finally , we demonstrate that the conformal array also produces a larger ratchet effect compared to square and random pinning arrays with a gradient in systems of colloidal particles , suggesting that pronounced ratchet effects should be a general feature of particles moving over conformal arrays .
p. pieranski , in _ phase transitions in soft condensed matter _ , edited by t. riste and d. sherrington ( plenum , new york , 1989 ) , p. 45 ; f. rothen , p. pieranski , n. rivier , and a. joyet , eur . j. phys .
* 14 * , 227 ( 1993 ) . | a conformal transformation of a uniform triangular pinning array produces a structure called a conformal crystal which preserves the six - fold ordering of the original lattice but contains a gradient in the pinning density .
here we use numerical simulations to show that vortices in type - ii superconductors driven with an ac drive over gradient pinning arrays produce the most pronounced ratchet effect over a wide range of parameters for a conformal array , while square gradient or random gradient arrays with equivalent pinning densities give reduced ratchet effects . in the conformal array , the larger spacing of the pinning sites in the direction transverse to the ac drive permits easy funneling of interstitial vortices for one driving direction , producing the enhanced ratchet effect . in the square array ,
the transverse spacing between pinning sites is uniform , giving no asymmetry in the funneling of the vortices as the driving direction switches , while in the random array , there are numerous easy - flow channels present for either direction of drive .
we find multiple ratchet reversals in the conformal arrays as a function of vortex density and ac amplitude , and correlate the features with a reversal in the vortex ordering , which is greater for motion in the ratchet direction .
the enhanced conformal pinning ratchet effect can also be realized for colloidal particles moving over a conformal array , indicating the general usefulness of conformal structures for controlling the motion of particles . |
the evolution of cooperation and altruism are fundamental scientific challenges highlighted by their role in the major transitions in life s history , when natural selection acted simultaneously on several competing levels @xcite . in this context , the relevance of basic concepts , including group selection and hamilton s rule remain controversial @xcite . here
we address these problems by studying a framework for evolution in group structured populations that incorporates inter- and intra - group competition and migration . combining group - centric with gene - centric perspectives in a constructive group / kin selection approach ,
we build methodology that allows for the analysis of arbitrary non - linear fitness functions , resulting from complex multi - individual interactions across life cycles .
we obtain the conditions for a rare social allele to invade the population .
this is obtained in a mathematically rigorous way , by analyzing the stability of the equilibrium in which this allele is absent .
this analysis is done for arbitrary strength of selection , but when selection is weak and groups are large the condition for invasion simplifies significantly into a form that is easy to apply and provides substantial intuition . in the case of linear fitness functions ,
the condition for invasion is identical to hamilton s rule , and it is natural to regard the more general non - linear cases as generalizations of that rule . our results also show that one of the most widely used approaches to analyzing kin selection models , @xcite , @xcite ( condition ( 6.7 ) ) , and @xcite ( box 6 ) , yields incorrect results in some biologically relevant situations .
our results reveal biologically realistic conditions under which altruism can evolve when rare , but genetic relatedness in groups is modest . in this way we challenge a common understanding according to which inter - group selection favoring altruism could only override intra - group selection favoring selfishness under exceptional conditions , namely small group size and very low migration rates @xcite .
we identify the emphasis on linear public goods games in the literature , including most of these papers , as having supported this belief .
in contrast , we show that for iterated public goods games , in which altruists cooperate or not in each round based on previous outcomes @xcite , altruism can spread even when groups are large , selection is weak and migration rates are substantially larger than the inverse of group size .
this result corrects @xcite , who predicted that large group size would not allow cooperation to spread when rare in this model . for species that live in groups ,
several vital group activities repeat themselves periodically and behavior changes as feedback is obtained from previous iterations .
the iterated public goods game that we study is therefore often more realistic than a simple one shot public goods game .
a proper analysis of this model fills therefore an important gap in the literature . to obtain our result
we show that in the absence of selection , when groups are large , the fraction of group members that are close relatives of a randomly chosen individual has a non - gaussian distribution with a fat tail . as a consequence ,
even when altruistic alleles are rare in the population , they have a significant probability of concentrating in some groups , accruing substantial reproductive gains through multi - individual synergy .
when members of a species live in groups , their reproductive success depends on the behavior of all group members .
more efficient groups may grow faster and split , outcompeting the less efficient ones that die out . on the other hand
, individuals may free ride on the cooperation of other members of their group , and in this way outcompete them .
this picture is further complicated by migration among groups . the _ two - level fisher - wright framework with selection and migration _ ( 2lfw ) captures all these elements , in a simplified fashion . in 2lfw haploid individuals
live in a large number @xmath0 of groups of size @xmath1 , and are of two genetically determined phenotypic types , a or n. generations do not overlap , reproduction is asexual and the type is inherited by the offspring .
( mutations will be considered briefly later . ) the relative fitness ( @xmath2 ) of a type a , and that of a type n , in a group that has @xmath3 types a , are , respectively , @xmath4 and @xmath5 , with the convention that @xmath6 , i.e. , @xmath7 . the quantities @xmath8 and @xmath9 represent life - cycle payoffs derived from behavior , physiology , etc .
the parameter @xmath10 indicates the strength of selection .
[ fig1 ] describes the creation of a new generation in the 2lfw through inter- and intra- group competition , followed by migration at rate @xmath11 .
cases in which types a behave in some altruistic fashion are of particular interest @xcite .
most of the literature concerns the very special case of a _ linear _ public goods game ( pg ) , defined by @xmath12 , @xmath13 , @xmath14 , in which each type a cooperates , at a cost @xmath15 to herself , providing a benefit @xmath16 shared by the other members of her group .
the need to consider more complex intra - group interactions and non - linear payoff functions is , nevertheless , well known .
non - linearities appear naturally whenever activities involve many group members simultaneously .
they result from threshold phenomena , increasing returns to scale , saturation , etc .
for instance , to hunt large prey may require a large minimum number of hunters , the likelihood of success may first increase rapidly with the number of hunters , but it may plateau when this number becomes very large . allowing for
the analysis of such synergistic multi - individual interactions and activities is a central feature of our approach , distinguishing it from theoretical frameworks based on pairwise interactions , or single actors benefiting a group @xcite . .
* ( center ) fw intragroup competition : * if a group descends from a group with @xmath3 types a , then it will have @xmath17 types a with probability @xmath18 , where the binomial probability @xmath19 is the probability of @xmath17 successes in @xmath1 independent trials , each with probability @xmath20 of success . *
( right ) migration : * once the new @xmath0 groups have been formed according to the two - level competition process , a random fraction @xmath11 of the individuals migrates .
migrants are randomly shuffled .
* note : * the assignment of relative fitness to the groups in the fashion done above is a necessary and sufficient condition @xcite for individuals in the parental generation to have each an expected number of offspring proportional to their personal relative fitness .
, scaledwidth=50.0% ] the 2lfw framework can be seen as a generalization of the trait - group framework ( see sec .
2.3.2 of @xcite ) , which corresponds to the case @xmath21
. one can interpret @xmath22 as an assortment parameter . because migration is completely random in 2lfw
, this assortment represents a worst case scenario , abstracting away additional assortment caused by kin recognition , greenbeard effects , selective acceptance of migrants , joint migration of individuals , etc .
it is well known @xcite that even when @xmath21 non - linearities in fitness functions allow for coexistence of cooperators and defectors . but under the strong altruism condition
@xmath23 ( meaning that each type a would be better off mutating into a type n ) , this is not the case @xcite .
one of our goals is to determine the level of migration compatible with invasion by rare strong altruists .
the 2lfw also generalizes the `` typical kin selection model '' of @xcite , where the payoffs were those of pg , and the analysis relied on this and on the assumption of weak selection . that paper was a response to @xcite , where group selection was argued to be an important mechanism for the evolution of cooperation , and a multilevel selection model based on moran s model was introduced .
our analysis of 2lfw with non - linear fitness functions highlights the importance of combining group - centric with gene - centric perspectives , and shows that group selection can be an important force in evolution under realistic conditions .
it also shows that mathematically rigorous analysis can be carried out even when selection is strong and fitness functions are non - linear . and
it shows that one has to be very careful in applying mathematically non - rigorous methodology , as it can produce substantially incorrect results , even when selection is weak .
non - linearities in life - cycle payoffs can result from activities repeating themselves during a lifetime , and behavior being contingent on previous outcomes .
a basic example is the iterated public goods game ( ipg ) @xcite . in ipg
a pg is repeated an average of @xmath24 times in a life - cycle .
we will suppose that types n never cooperate , while types a cooperate in the first round and later cooperate only if at least a fraction @xmath25 of group members cooperated in the previous round .
mathematically , this model generalizes the iterated prisoner dilemma and tit - for - tat , from the dyadic setting of @xcite and @xcite to the multi - individual setting .
but while direct or indirect reciprocity requires the identification of individuals in the group , this is not the case here .
the behavior of types a in the ipg can be triggered by individuals simply discontinuing cooperative behavior when previous cooperation produced negative feedback to them .
in other words , allele a can predispose individuals to cooperate , but as they do it and obtain feedback from that behavior , they may continue it or discontinue it .
the ipg is in this sense closely related to generalized reciprocity mechanisms @xcite with low cognitive requirements .
negative feedback from cooperation should occur if the fraction of group members that cooperated was less than @xmath26 , but not if it was larger than that threshold , since in the former case the payoff to a cooperator is negative , while in the latter case it is positive .
this gives a special role to the value @xmath27 . if @xmath28 , the behavior of types a is altruistic in the strong sense that each type a individual would increase its fitness if it behaved as a type n , everything else being equal , i.e. , @xmath29 ( see si appendix ( section 8) for a detailed discussion ) . moreover , types n always free ride and have greater fitness than types a in the same group , regardless of the values of @xmath25 and @xmath24 , i.e. , @xmath30 .
[ fig2 ] displays a detailed analysis of some instances of the ipg , giving conditions for allele a to spread when rare . for many species that live and interact in groups for many years
, several vital activities , including collective hunting and food sharing , can repeat themselves hundreds or thousands of times in a life - cycle , giving plausibility to the values of @xmath24 in fig .
( the assumption that individuals discontinue behavior after a single unsuccessful participation is a simplification . when this is not a realistic assumption , one can interpret the parameter @xmath24 as the ratio between the typical number of repetitions of the activity and the typical number of unsuccessful attempts before cooperation is discontinued by a type a. ) panel c , in which selection is weak and groups are large , shows two important contrasting results .
when @xmath31 , and the ipg is identical to the pg , allele a can only invade under hamilton s condition @xmath32 .
but as @xmath24 increases , the level of relatedness needed for invasion drops substantially , so that for modest values of @xmath33 , allele a can invade under @xmath34 significantly lower than @xmath35 , compatible with levels observed in several species , including humans @xcite ( table 8.3 ) , @xcite , @xcite ( tables 6.4 and 6.5 ) , @xcite , @xcite ( table 4.9 ) . the corresponding number of migrants per group per generation , @xmath36 , can be of the order of 10 .
further examples showing the spread of altruism and cooperation under high levels of gene flow and low levels of relatedness are provided in figs . 5 , 6 , 7 , 8 , 20 and 21 in the si appendix .
times in a life - cycle . in each round
each individual can cooperate at a cost @xmath15 to herself , producing a benefit @xmath16 shared by the other members of the group .
types n never cooperate , while types a cooperate in the first round and later cooperate only if at least a fraction @xmath25 of group members cooperated in the previous round .
in all panels @xmath27 ( types a are strongly altruistic ) and curves correspond to @xmath37 ( black , this case is identical to pg ) , 10 ( blue ) , 100 ( magenta ) , 1000 ( green ) ( bottom to top in panel a , top to bottom in panels b and c ) .
* panel a : * @xmath38 , @xmath39 , @xmath40 .
curves give the critical migration rate @xmath41 below which types a proliferate when rare , and that solves @xmath42 , or equivalently @xmath43 in ( [ gen ] ) .
( the subscript ` s ' stands for ` survival ' . )
the dependence of @xmath41 on the strength of selection @xmath44 indicates the relevance of studying both weak and strong selection .
short horizontal red lines indicate value of @xmath41 under weak selection , obtained from setting @xmath43 in ( [ ws ] ) ( note the excellent agreement ) .
* panel b : * again , @xmath38 , @xmath39 , @xmath40 .
curves give the critical relatedness @xmath45 above which types a proliferate .
here @xmath46 is the relatedness obtained from neutral genetic markers .
short horizontal red lines are again from @xmath47 in ( [ ws ] ) .
* panel c : * limit of large @xmath1 under weak selection .
critical values of relatedness @xmath48 , as function of @xmath26 .
solid lines provide the solution to the equation @xmath49 derived from setting @xmath47 in ( [ wsln ] ) .
dashed lines give its approximation ( [ rcblnt ] ) .
red vertical line corresponds to @xmath50 , while horizontal red lines are at the same level of those from panel b. their intersections illustrate the fact that both the solid and dashed lines in panel c are good approximations to weak - selection values of critical relatedness , @xmath51 , when @xmath52 . ,
to analyze the 2lfw , denote by @xmath53 , @xmath54 the fraction of groups in generation @xmath55 that have exactly @xmath3 types a. denote by @xmath56 the frequency of types a in the population .
the state of the population in generation @xmath55 is described by the vector @xmath57 , since @xmath58 .
we will suppose that @xmath59 , so that , by the law of large numbers , @xmath60 evolves as a deterministic ( non - linear ) dynamical system in dimension @xmath1 . here
we will study its linearization close to the fixed point @xmath61 , with no types a. this means that we are restricting ourselves to the case in which @xmath62 , and studying the conditions for allele a to invade the population when rare .
with notation introduced in fig .
[ fig1 ] , we have then @xmath63 , where @xmath64 , @xmath65 and @xmath66 , if @xmath67 , @xmath68 if @xmath69 .
matrix @xmath70 represents the production of groups in the new generation , through the two - level competition .
matrix @xmath71 represents the effect of types a migrating out of groups , and matrix @xmath16 represents the effect of these migrant types a joining groups that previous to migration had no types a. ( when @xmath62 , the migrant types a are a small fraction of the migrant population , and therefore each one is likely to settle in a different group that had no types a before migration . )
a standard application of the perron - frobenius theorem implies that when @xmath72 , we have , in good approximation , @xmath73 , where @xmath15 is a constant that depends on @xmath74 , @xmath75 is the leading eigenvalue of @xmath76 and @xmath77 is its corresponding left - eigenvector normalized as a probability vector .
this means that , regardless of the initial distribution @xmath74 , with @xmath78 , demographics and natural selection drive @xmath60 towards multiples of @xmath77 , in what can be seen as self - organization of copies of a in the optimal stable way for them to spread .
once this has happened , @xmath79 grows at rate @xmath80 .
consequently , allele a will proliferate , when rare , if the _ viability condition _
@xmath81 holds , and it will vanish if @xmath82 .
( see fig .
[ fig2 ] and si appendix ( sections 1 and 2 ) for applications , illustrations and further explanations . )
when @xmath81 , even if allele a is initially absent , a small rate of mutation will introduce it , allowing it to then invade the population . in the terminology of evolutionary game theory ( see , e.g. , @xcite chapter 7 ) , phenotype n is an evolutionary stable strategy ( ess ) when @xmath82 and n is not an ess when @xmath81 .
the viability condition @xmath81 has a gene - centric ( kin - selection ) interpretation in terms of average ( neighbor modulated ) fitnesses .
for this purpose , define @xmath83
. then it is well known that @xmath84 , where @xmath85 and @xmath86 are the average fitnesses of types a and n , and @xmath87 is the average fitness of all individuals . if we choose a random type a , it will have probability proportional to @xmath88 of being in a group with exactly @xmath3 types a ( bayesian sampling bias ) .
therefore @xmath89 . when @xmath62 , if we choose a random individual , it is likely to be in a group with no types a. therefore , in good approximation , @xmath90 .
since @xmath60 is driven towards multiples of @xmath77 , we obtain p = p , [ gen ] provided @xmath62 and @xmath72 ( the error term is of order @xmath91 + @xmath92 , with @xmath93 ) .
the viability condition @xmath81 can also be stated as @xmath94 , in ( [ gen ] ) .
it is important to observe that @xmath79 does not need to be monotone , and that @xmath94 is the proper condition for invasion only when , as in ( [ gen ] ) , one is considering the stationary regime , @xmath72 .
if selection is weak , i.e. , @xmath95 , migration acts much faster than selection , providing a separation of time scales @xcite .
this allows us to replace @xmath77 in ( [ gen ] ) with @xmath96 , obtained by assuming @xmath97 .
algebraic simplifications ( presented in si appendix ( section 5 ) ) allow us then to rewrite the neighbor modulated fitness relation ( [ gen ] ) in the form p = p _ k=1^n _
k v^a_k [ ws ] ( the error term is of order @xmath98 ) , where @xmath99 and @xmath100 , with @xmath101 the markov transition matrix @xmath102 and its invariant distribution @xmath103 have natural interpretations in terms of identity by descent ( ibd ) under neutral genetic drift , as we explain next when we provide a second derivation of ( [ ws ] ) .
two individuals are said to be ibd if following their lineages back in time , they coalesce before a migration event affects either one .
the separation of time scales implies that when selection acts , the demographic distribution is well approximated by that obtained in equilibrium with @xmath97 .
this means that in good approximation @xmath104 , where @xmath105 is the @xmath106 equilibrium probability that in the group of a randomly chosen focal type a there are exactly @xmath3 types a ( focal included ) .
but because we are supposing that types a are rare , the only individuals that are type a in this group are those that are ibd to the focal , so that @xmath105 is also the probability that exactly @xmath3 individuals in this group are ibd to the focal . as in the derivation of ( [ gen ] ) , since types a are rare , we have @xmath107 and hence @xmath108 , which is ( [ ws ] ) . to compute @xmath103 , we will use the standard kronecker notation @xmath109 if @xmath110 and @xmath111 if @xmath112 . now , the probability @xmath113 that the focal is ibd to exactly @xmath114 other members of her group is @xmath115 if the focal is a migrant ( probability @xmath11 ) , while if she is not a migrant ( probability @xmath22 ) , then we have to consider how many individuals in her mother s group were ibd to her mother . if , counting her mother , that number was @xmath17 ( probability @xmath116 , assuming demographic equilibrium ) then the probability that the focal is ibd to exactly @xmath114 other members of her group is equal to the probability that of the @xmath117 other members of her group , exactly @xmath114 are non - migrants who chose for mother one of the @xmath17 candidates ( among @xmath1 possibilities ) that were ibd to the focal s mother ( probability @xmath118 ) . combining these pieces , we have @xmath119 this is exactly the same as the set of equations @xmath120 .
the ibd distribution @xmath103 contains all the relevant information about genetic relatedness in the groups , including and exceeding that given by the average relatedness between group members , @xmath121 , obtained from lineages , regression coefficients , or wright s @xmath122 statistics .
specifically , we can define @xmath123 as the probability that a second member chosen from the focal s group is ibd to the focal and then obtain ( from linearity of expected values ) that @xmath124 is a linear function of @xmath103 s first moment .
when @xmath8 is a non - linear function of @xmath3 , more information contained in @xmath103 , including its higher moments , are needed to decide whether @xmath94 in ( [ ws ] ) .
it is important to also stress that ( [ ws ] ) can be easily used for applications in which even the knowledge of all the moments of @xmath103 ( see @xcite ) would be cumbersome to apply , as for instance in the computation of the short horizontal red lines in fig .
[ fig2 ] , panels a and b.
the stationarity condition @xmath125 allows for a recursive computation of all the moments of @xmath103 ( see si appendix ( section 5 ) ) .
these moments can then be used to show the powerful result that if @xmath1 is large and @xmath11 is small , then @xmath103 , when properly rescaled , is close to a beta distribution , with mean @xmath34 ( see si appendix ( section 6 ) ) . in this case , if in addition to the assumption of weak selection , also @xmath8 is well approximated by @xmath126 , for some piecewise continuous function @xmath127 , then ( [ ws ] ) takes the easy to apply form p = p ( - 1 ) _
0 ^ 1 ( 1-x)^ - 2 v^a_x dx , [ wsln ] where @xmath128 . equations
( [ gen ] ) and ( [ wsln ] ) play complementary roles in the analysis of 2lfw .
both provide the condition for invasion by allele a ; ( [ gen ] ) holds in full generality , while ( [ wsln ] ) requires special assumptions ( small @xmath44 , large @xmath1 ) , but is computationally much simpler and provides a great deal of intuition , as we discuss next .
equation ( [ wsln ] ) should be contrasted with what @xcite predicted by supposing that the number of individuals in a group that are ibd to a focal individual would be well approximated by a binomial with @xmath117 attempts and probability @xmath123 of success .
that would lead to a normal distribution , narrowly concentrated close to its mean @xmath123 , in place of the beta distribution above .
our result reveals a strong dependency structure among lineages , producing the beta distribution , with a standard deviation comparable to its mean , and a tail that decays slowly compared to a gaussian distribution . as a consequence ,
fitness functions that are large only when the fraction of types a in a group is above a threshold value , as in the ipg , will allow for proliferation of types a under levels of relatedness substantially lower than that predicted under the assumption in @xcite .
we will refer to the fact that the fraction of group members that are ibd to a focal individual has a non - vanishing standard deviation , even when selection is weak and groups are large , as _ persistence of variability_. this phenomenon poses a severe limitation to the applicability of covariance - regression methods in which regression of fitness on genotype is replaced with derivatives , as in @xcite , @xcite ( box 6 ) , @xcite ( condition ( 6.7 ) ) . both the assumptions in @xcite , or in @xcite applied to the ipg would have implied incorrectly that when selection is weak and groups are large , types a could only invade the population when rare if @xmath129 ( these computations are presented in the si appendix ( sections 8 and 9 ) ) . in a companion paper @xcite we show that methodologies in which one expresses the fitness of a focal individual in terms of partial derivatives with respect to the focal individual s phenotype and the phenotype of the individuals with whom the focal interacts , as in @xcite , require @xmath130 to be a linear function of @xmath131
. for the pg , ( [ ws ] ) and ( [ wsln ] ) clearly reduce to the well known @xmath132 .
the same is also true for the more general ( [ gen ] ) , as was shown in @xcite , where in case of strong selection the relatedness @xmath123 depends on the payoff functions .
in contrast , if we are under the conditions of ( [ wsln ] ) with @xmath133 , then p = p ( -c + br + b_2 r_2 + ... + b_l r_l ) , [ wspowers ] where @xmath134 $ ] is the @xmath135-th moment of the beta distribution . for the ipg , @xmath136 , if @xmath137 , and @xmath138 , if @xmath139 . the viability condition derived from ( [ wsln ] ) can be analyzed in detail , by simple , but long , computations , presented in the si appendix ( section 8) . in the case
@xmath27 , the viability condition reads @xmath140 .
when @xmath24 is large , this yields the following approximation for the critical relatedness @xmath34 : r = .
[ rcblnt ] if also @xmath141 , then r = = .
[ rcblnt+ ] the simplicity and transparency of ( [ rcblnt ] ) and ( [ rcblnt+ ] ) illustrate the power of ( [ wsln ] ) , and fig . [ fig2 ] shows how well they compare to the more general , but less transparent ( [ gen ] ) .
note also how ( [ rcblnt ] ) and ( [ rcblnt+ ] ) provide a direct grasp on the effect of the number of repetitions in the game , and a nice comparison between the pg and the ipg . both fig .
[ fig2 ] and ( [ rcblnt+ ] ) show that alleles that promote contingent cooperative behavior , which is discontinued when participation is low , can spread under levels of genetic relatedness ( @xmath142 ) more than 5 times smaller than @xmath26 .
this mechanism should , therefore , be seriously investigated as a possible route for the proliferation of altruistic / cooperative behavior .
1 . natural selection in group
structured populations is best analyzed by a combination of group - centric and gene - centric perspectives and methods .
both shed light , carry intuition and provide computational power , in different ways .
rigorous mathematical analysis of models is necessary , especially when fitness functions are non - linear , to assess the validity of non - rigorous approaches .
contingent forms of group altruism that are discontinued when participation is low can proliferate under biologically realistic conditions .
their role in the spread of altruism should be empirically investigated .
3 . natural selection can promote traits that ( in net terms over a full life - cycle ) are costly to the actors and beneficial to the other members of their group , under demographic conditions that are not stringent .
this can happen in large groups and with realistically high levels of gene flow .
excessive focus on one - shot linear public goods games in the literature has obscured this fact .
the authors thank rob boyd for many hours of stimulating and informative conversations on the subjects in this paper .
thanks are given to marek biskup for assistance in exploring the nature of the distribution @xmath103 .
we are also grateful to clark barrett , maciek chudek , daniel fessler , kevin foster , willem frankenhuis , bailey house , laurent lehmann , glauco machado , sarah mathew , diogo meyer , cristina moya , peter nonacs , karthik panchanathan , susan perry , joan silk , jennifer smith and ming xue for nice conversations and feedback on various aspects of this project and related subjects .
this project was partially supported by cnpq , under grant 480476/2009 - 8 .
wenseleers t , gardner a , foster , kr ( 2010 ) in _ social behaviour : genes , ecology and evolution _ , eds .
szekely t , moore aj , komdeur j ( cambridge university press , cambridge , uk , cambridge , uk ) , pp . 132158 .
lessard s ( 2009 ) diffusion approximations for one - locus multi - allele kin selection , mutation and random drift in group - structured populations : a unifying approach to selection models in population genetics .
, 659696 . | the ways in which natural selection can allow the proliferation of cooperative behavior have long been seen as a central problem in evolutionary biology .
most of the literature has focused on interactions between pairs of individuals and on linear public goods games .
this emphasis led to the conclusion that even modest levels of migration would pose a serious problem to the spread of altruism in group structured populations . here
we challenge this conclusion , by analyzing evolution in a framework which allows for complex group interactions and random migration among groups .
we conclude that contingent forms of strong altruism can spread when rare under realistic group sizes and levels of migration .
our analysis combines group - centric and gene - centric perspectives , allows for arbitrary strength of selection , and leads to extensions of hamilton s rule for the spread of altruistic alleles , applicable under broad conditions . |
quantum theory ( qt ) is in conflict with noncontextual realism @xcite , that is , with the assumption that the results of measurements reveal pre - existing properties that are not affected by compatible measurements .
noncontextual realism is a legitimate assumption in so far as the statistics of the measurement outcomes are not perturbed by other measurements , or , in other words , whenever measurements can not be used to communicate information .
no signaling holds when the measurements are spacelike separated . if this is the case , noncontextual realism is called local realism .
it is well known that qt is also in conflict with local realism @xcite .
however , most quantum violations of noncontextual realism do not occur in scenarios in which parties perform spacelike separated measurements .
still , they can be reveled in experiments @xcite .
this shows that contextuality , besides being `` a class of ` paradoxes ' which result from counterfactual logic '' @xcite , is a phenomenon that can be experimentally tested @xcite . among the violations of noncontextual realism that can not be observed in bell - like tests
there are many cases of interest : qutrit contextuality @xcite , quantum - state - independent contextuality @xcite , simple fully contextual correlations @xcite , contextuality associated to universal quantum computation via magic states @xcite , and absolute maximal contextuality @xcite .
this variety of examples motivates the challenge of whether there is a unified universal way to test any form of contextuality .
certifying contextuality requires observing the violation of a noncontextuality ( nc ) inequality @xcite , which is an inequality for correlations satisfied by any noncontextual hidden - variable theory ( nchvt ) , i.e. , for any theory assuming noncontextual realism . testing a nc inequality entails considerable difficulty and requires solving some problems .
the first difficulty is related to the fact that in a test of a nc inequality every observable must be measured using the same device in any context .
an alternative would be , before conducting the experiment , certifying that different devices measure the same observable since they produce the same distributions of probability for any possible state . however , this is unfeasible using a finite sample of states . without such certification
it seems unreasonable to assume that measurement outcomes do not depend on the context when measurements themselves do @xcite .
this problem can be avoided by designing devices that exactly measure each of the required observables in any possible context .
the second difficulty comes from the fact that , except for the case of bell inequalities ( which are nc inequalities ) , testing other nc inequalities requires performing sequential measurements of quantum observables @xcite , as illustrated in fig .
[ fig1 ] ( a ) .
this is challenging . recall that the measurement of a quantum observable , as defined by von neumann @xcite ( i.e. , a projective measurement ) , should not only produce an outcome ( that should be `` indelibly recorded '' @xcite before any further measurement ) but also transform the quantum state according to lders s rule @xcite .
the problem is that only a few experiments have successfully implemented sequential projective measurements ( e.g. , ref .
the typical measurements on quantum systems are actually demolition measurements which absorb the system , preventing subsequent measurements , or fail to transform the state of the system according to lders s rule .
nevertheless , the main difficulty is achieving the no - signaling condition with sequential measurements . recall that in a contextuality test each measurement should not affect the probabilities of the outcomes of any subsequent measurement .
that is , if the correlations between @xmath0 and @xmath1 and between @xmath2 and @xmath1 both appear in the nc inequality and @xmath1 is measured in the second place , then @xmath3 must be zero for any @xmath4 .
the problem is that it is impossible to have @xmath5 in actual experiments since the number of trials is finite , long - lasting experiments suffer drifts , and there are inevitable imperfections @xcite .
one could argue that the condition of perfect no signaling is not even achieved in experiments with spacelike separated measurements ( e.g. , @xmath6 in ref .
. however , in these experiments no signaling can be assumed on physical grounds and a nonzero @xmath7 can be attributed to statistical fluctuations .
in contrast , in experiments with sequential measurements , a nonzero @xmath7 should be taken into account in the analysis @xcite . a practical issue
, however , is that these analysis are in general difficult or are not well defined beyond the case of two - point correlations . despite these problems ,
little effort has been made for designing contextuality tests with small @xmath7 that are easy to analyze . from this perspective , a main challenge is to have tests in which the influence of past measurements on future measurements is experimentally indistinguishable from the influence of future measurements on past measurements .
more precisely , experiments in which @xmath7 is zero within the same experimental error that @xmath8 .
assuming causality , @xmath9 is , by definition , zero .
however , when calculated from experimental data , this zero will be inevitably affected by an error . only if the experimental value of @xmath7 is compatible with zero and is affected by a similar error can one reasonably assume that no signaling is satisfied within the experimental precision . in this case
, the violation of the nc inequality can be taken as a compelling evidence of contextuality .
another difficulty comes from the infeasibility of implementing projective sequential measurements on most quantum systems , particularly on high - dimensional quantum systems @xcite which are necessary for some forms of quantum contextuality @xcite .
one solution is implementing sequentially the unitary transformations corresponding to the desired measurements and detect the systems afterwards .
this is the approach adopted in some experiments with photons @xcite , neutrons @xcite , and molecular nuclear spins @xcite . other solution , closer to the ideal contextuality experiment in fig .
[ fig1 ] ( a ) , consists of encoding the outcome of each measurement in an extra degree of freedom ( e.g. , extra spatial modes @xcite , time bin @xcite , or polarization @xcite ) before performing the next one
. however , this solution has one disadvantage : ( i ) the complexity of the setup grows exponentially with the number of sequential measurements .
this does not only entail practical problems but also introduces a conceptual issue since these extra degrees of freedom could provide the memory classical systems need to simulate quantum contextuality @xcite .
moreover , both solutions have other disadvantage : ( ii ) the observers can not delay the choice of the next measurement to the moment in which the result of the previous measurement is indelibly recorded . as a result ,
the system is not really forced to answer before the context is fixed @xcite .
this indicates that other interesting challenge is designing contextuality tests for high dimensional quantum systems and free of disadvantages ( i ) and ( ii ) . in this paper
we show how to reveal any form of quantum contextuality in an experiment requiring sequences of only two measurements and allowing a better control on the experimental imperfections and a simple analysis .
in addition , we show how to design contextuality tests for quantum systems in which sequential projective measurements are unfeasible and free of disadvantages ( i ) and ( ii ) .
in this section we show that two - point correlation experiments between binary compatible observables are sufficient to test any form of quantum contextuality . in the next section
we will explain why this is important . to understand what we mean
by `` any form of quantum contextuality , '' we start by recalling that any given nc inequality can be associated to graph @xcite .
this association requires converting the linear combination of probabilities in the initial nc inequality into a positive linear combination @xmath10 of probabilities of yes - no tests ( represented in qt by rank 1 projectors ) .
@xmath10 can then be associated to a graph @xmath11 in which each probability is represented by a vertex while edges connect vertices corresponding to probabilities of exclusive events ( i.e. , corresponding to alternative outcomes of a sharp measurement @xcite ) .
the maximum of @xmath10 for nchvts and qt are two characteristic numbers of @xmath11 : the independence number , @xmath12 , and the lovsz number , @xmath13 , respectively @xcite .
reciprocally , any graph @xmath11 can be converted into a nc inequality such that the bound for nchvts is @xmath12 and the maximum in qt is @xmath13 @xcite .
this establishes a one - to - one connection between any possible quantum violation of any possible nc inequality and a graph .
it is in this sense that any form of quantum contextuality can be represented by a graph .
here we show that any nc inequality represented by a graph @xmath11 can be further transformed into a nc inequality involving only two - point correlations , while still having @xmath12 as upper bound for nchvts and @xmath13 as maximum in qt .
_ theorem .
_ for any graph @xmath11 with vertex set @xmath14 and edge set @xmath15 the following inequalities are tight : @xmath16 where @xmath17 is the probability of obtaining result @xmath18 when observable @xmath19 is measured and @xmath20 is the joint probability of obtaining result @xmath18 for @xmath19 and result @xmath18 for @xmath21 .
the quantum maximum @xmath13 can be attained by preparing a quantum state @xmath22 and measuring a set of rank 1 projectors @xmath23 such that each projector is associated to a vertex of @xmath11 , adjacent vertices are associated orthogonal projectors , and @xmath24 .
the set @xmath25 is called a lovsz - optimum - orthogonal representation of @xmath26 @xcite .
its existence is guaranteed by the definition of @xmath13 .
_ we can rewrite @xmath27 as @xmath28 \equiv { \cal s ' } - |e(g)| , \end{split}\ ] ] where @xmath29 is the cardinal of @xmath15 .
then , @xmath30 where @xmath31 is the graph of the events in @xmath32 .
therefore , @xmath33 now we have to prove that @xmath34 and @xmath35 .
the maximum of @xmath32 for any nchvt is always attained by assigning probability @xmath18 to some events and @xmath36 to the others . for every @xmath37 , there are five vertices in @xmath31 : those corresponding to the probabilities of the events @xmath38 , @xmath39 , @xmath40 , @xmath41 , and @xmath42 .
see fig .
[ fig2 ] ( b ) .
if @xmath43 and @xmath44 , then @xmath45 , @xmath46 , and @xmath47
. thus if we do this assignment for all edges in @xmath11 we obtain @xmath48 ( i.e. , the number of vertices of @xmath11 ) , which is less than or equal to @xmath49 . on the other hand , if @xmath50 and @xmath51 , then @xmath52 , @xmath46 , and @xmath47 .
thus if we do this assignment for all edges in @xmath11 , we obtain @xmath53 , which is strictly smaller than @xmath49 . if @xmath43 and @xmath51 , then @xmath45 , @xmath46 , and @xmath54 .
however , in general , we can not do this assignment for all edges in @xmath11 .
an assignment maximizing @xmath32 for nchvts is one maximizing @xmath27 ( recall that @xmath55 ) and assigning probability @xmath18 to at most one of the vertices that are adjacent in @xmath11 .
this is exactly accomplished by any assignment leading to @xmath56 .
then , for every edge @xmath57 in @xmath11 with probabilities @xmath18 and @xmath36 , the sum of the probabilities of the corresponding five - vertex subgraph of @xmath31 is @xmath58 , while for every @xmath37 with probabilities both @xmath36 , the sum of the probabilities of the corresponding five - vertex subgraph of @xmath31 is @xmath18 . therefore , the sum of the probabilities of the events in @xmath31 is @xmath49 .
this proves the upper bound of @xmath27 for nchvts . for the maximum of @xmath32 in qt consider the assignment of probabilities to the vertices of @xmath31 corresponding , via born s rule , to a lovsz - optimum - orthogonal representation of @xmath26 . since , for every edge @xmath57 of @xmath11 , @xmath59 , then @xmath60
to prove that this is actually the quantum maximum notice that @xmath27 ( and @xmath32 ) contains only probabilities which are well defined within qt ( irrespective of the order in which measurements are performed ) only if all pairs @xmath37 correspond to compatible observables .
then , maximizing @xmath61 is equivalent to maximizing @xmath62 under the restriction that @xmath63 for @xmath37 .
consequently , @xmath35 . for clarity
s sake we have not addressed the case of vertex - weighted graphs .
however , the result presented in this section can be extended to vertex - weighted graphs with weights that are natural numbers ( i.e. , to the graphs corresponding to inequalities with rational coefficients ) by noticing that any of these graphs can be converted into a standard graph by copying @xmath64 times each vertex with weight @xmath64 .
the observation that any nc inequality can be converted into a nc inequality of the form ( [ main ] ) may change the way contextuality is tested in the laboratory .
this is so because testing ( [ main ] ) only requires measuring sequences of two observables , which allows us to simultaneously overcome many obstacles that , so far , had prevented experiments on most forms of contextuality . more specifically : \(i ) on one hand , the fact that only two sequential measurements are required will help to improve the quality of the experimental results by reducing the imperfections and making easier to identify and correct them .
\(ii ) on the other hand , the fact that only two - point correlations are required allows us to apply existing tools for analyzing contextuality experiments with imperfections which are only defined in this case @xcite .
\(iii ) furthermore , it allows us to design tests free of problems ( i ) and ( ii ) using systems in which sequential nondemolition measurements are not feasible .
if only sequences of two measurements are needed , then the complexity of setup for testing each sequence is constant and thus problem ( i ) is solved . on the other hand ,
the first measurement can be simulated by a demolition measurement followed by a preparation that depends on the outcome .
for example , the projective measurement of @xmath19 in fig .
[ fig1 ] ( b ) can be simulated by a demolition measurement of @xmath19 followed by a preparation of a new system in state @xmath65 if the outcome of @xmath19 is @xmath18 , or in the state predicted by lders s rule for the state @xmath22 if the outcome of @xmath19 is @xmath36 ; see fig .
[ fig1 ] ( c ) .
actually , only the first part is needed for obtaining the probabilities in eq .
( [ main ] ) , which can then be expressed as @xmath66 , where @xmath67 denotes the probability @xmath67 for the quantum state @xmath68 .
then , the choice of the second measurement is made after the result of the first measurement is indelibly recorded , thus solving problem ( ii ) .
nevertheless , it is important to point out that the experiments using the scheme in fig .
[ fig1 ] ( c ) present some conceptual disadvantages with respect to those using the schemes in figs .
[ fig1 ] ( a ) and [ fig1 ] ( b ) .
for example , the scheme in fig .
[ fig1 ] ( c ) provides an operational justification for the assumption of outcome noncontextuality , but only in experiments with sequences of two measurements on @xmath22 .
there , it is possible to check that , when @xmath22 is prepared , @xmath19 does not affect the outcome probabilities for any @xmath21 that is measured afterwards .
however , this is not guaranteed for other possible preparations because the way the measurement of @xmath19 in fig .
[ fig1 ] ( c ) transforms the state depends on @xmath22 ( and , in particular , fails to imitate a projective measurement when the outcome is @xmath69 ) .
in contrast , experiments using the schemes in figs .
[ fig1 ] ( a ) and [ fig1 ] ( b ) provide a much more compelling operational justification of the assumption of outcome noncontextuality , since they allow us to check repeatability and no signaling for any preparation and any sequence of compatible measurements .
notice also that the experiments in figs .
[ fig1 ] ( b ) and [ fig1 ] ( c ) are not thermodynamically equivalent . the measurement of @xmath19 in fig .
[ fig1 ] ( c ) , a demolition measurement followed by a preparation , dissipates more heat than the projective measurement of @xmath19 in fig .
[ fig1 ] ( b ) .
since classical simulations of quantum contextuality can be distinguished from truly quantum contextuality by the fact that the former produces more heat @xcite , then to certify quantum contextuality it is more convenient to use tests like those in fig .
[ fig1 ] ( b ) than those in fig .
[ fig1 ] ( c ) .
in addition to the advantages for designing better experiments , the fact that any form of contextuality can be revealed in experiments measuring only two compatible observables also has some conceptual advantages : \(iv ) it allows for a novel black box approach to contextuality , extending the black box approach to nonlocality @xcite .
the basic elements would be pairs of boxes with respective inputs @xmath70 such that @xmath71 and respective outputs @xmath72 . unlike the case of nonlocality , here we are not limited by the condition that each box should be attributable to a party spacelike separated from the others .
still , we can use tools like postselection , composition , and wiring of boxes .
this approach may help to understand why certain sets of nonlocal correlations @xcite are not realizable .
\(vi ) it permits translating any nc inequality into a communication complexity scenario involving only two - point communication .
this may help in elucidating the connection between the principle that constraints quantum contextuality @xcite and information - theoretic principles that are essentially bipartite @xcite .
\(vii ) it allows for building theory - independent sets of events with exclusivity relations leading to noncontextual and quantum bounds on demand . for any graph @xmath11
, one can assign unit vectors to the vertices of @xmath11 such that adjacent vectors are orthogonal .
this means that , according to qt , adjacent vertices correspond to alternative outcomes of a sharp measurement and , therefore , to exclusive events .
however , these exclusivities can not be justified in a theory - independent way unless we identify each of the vertices in @xmath11 with an operational procedure that explains why adjacent vertices correspond to exclusive events .
the construction used in inequality ( [ main ] ) avoids the problem of finding such procedures and allows us to construct theory - independent sets of events with noncontextual and quantum bounds equal to those of @xmath11 .
notice that the only assumption in the transition from @xmath11 to @xmath31 is that edges in @xmath11 correspond to jointly measurable observables .
\(viii ) the recent observation that contextuality is a necessary resource for fault - tolerant universal quantum computation via magic states @xcite depends on a definition of contextuality based on the violation of inequalities involving probabilities of noncompatible observables ( following ref .
the fact that any of these inequalities can be converted into a nc inequality of the type ( [ main ] ) allows us to reformulate the contextuality - computation connection in terms of the traditional definition of contextuality as the violation of inequalities involving probabilities of compatible observables @xcite .
we thank l. aolita , m. arajo , g. caas , j. chen , v. dambrosio , e. s. gmez , j .- .
larsson , g. lima , a. j. lpez - tarrida , s. pdua , j. r. portillo , f. sciarrino , s. van dam , and a. winter for useful conversations .
this work was supported by the fqxi large grant project `` the nature of information in sequential quantum measurements '' and project no .
fis2014 - 60843-p , `` advanced quantum information '' ( mineco , spain ) , with feder funds .
e. p. specker , die logik nicht gleichzeitig entscheidbarer aussagen , http://dx.doi.org/10.1111/j.1746-8361.1960.tb00422.x[dialectica * 14 * , 239 ( 1960 ) ] [ the logic of non - simultaneously decidable propositions , http://arxiv.org/abs/1103.4537 [ ] . ]
a. a. klyachko , m. a. can , s. biniciolu , and a. s. shumovsky , simple test for hidden variables in spin-1 systems , http://dx.doi.org/10.1103/physrevlett.101.020403[phys .
* 101 * , 020403 ( 2008 ) . ]
p. badziag , i. bengtsson , a. cabello , and i. pitowsky , universality of state - independent violation of correlation inequalities for noncontextual theories , http://dx.doi.org/10.1103/physrevlett.103.050401[phys .
lett . * 103 * , 050401 ( 2009 ) . ] n. delfosse , p. a. guerin , j. bian , and r. raussendorf , wigner function negativity and contextuality in quantum computation on rebits , http://dx.doi.org/10.1103/physrevx.5.021003[phys .
x * 5 * , 021003 ( 2015 ) . ]
e. amselem , m. bourennane , c. budroni , a. cabello , o. ghne , m. kleinmann , j .- .
larsson , and m. wieniak , comment on ` state - independent experimental test of quantum contextuality in an indivisible system ' , http://dx.doi.org/10.1103/physrevlett.110.078901[phys .
* 110 * , 078901 ( 2013 ) . ]
larsson , m. kleinmann , o. ghne , and a. cabello , violating noncontextual realism through sequential measurements , in _ advances in quantum theory _ , edited by g. jaegger , a. y. khrennikov , m. schlosshauer , and g. weihs , http://dx.doi.org/10.1063/1.3567466[aip conf .
proc . * 1327 * , 401 ( 2011 ) . ] j. von neumann , _ mathematische grundlagen der quantenmechanik _
( springer - verlag , berlin , germany , 1932 ) [ _ mathematical foundations of quantum mechanics _ ( princeton university press , princeton , nj , 1955 ) ] .
w. a. miller and j. a. wheeler , delayed - choice experiments and bohr s elementary quantum phenomenon , in _ foundations of quantum mechanics in the light of new technology ( kokubunji , tokyo , 1983 ) _ , edited by s. kamefuchi et al .
( physical society of japan , tokyo , 1984 ) , p. 140 .
g. lders , ber die zustandsnderung durch den meproze , http://www.physik.uni-augsburg.de/annalen/history/historic-papers/1951_443_322-328.pdf[ann . phys .
( leipzig ) * 8 * , 322 ( 1951 ) ] [ concerning the state - change due to the measurement process , http://dx.doi.org/10.1002/andp.200610207[ann .
( leipzig ) * 15 * , 663 ( 2006 ) ] . ]
g. kirchmair , f. zhringer , r. gerritsma , m. kleinmann , o. ghne , a. cabello , r. blatt , and c. f. roos , state - independent experimental test of quantum contextuality , http://dx.doi.org/doi:10.1038/nature08172[nature ( london ) * 460 * , 494 ( 2009 ) . ]
o. ghne , m. kleinmann , a. cabello , j .- .
larsson , g. kirchmair , f. zhringer , r. gerritsma , and c. f. roos .
compatibility and noncontextuality for sequential measurements , http://dx.doi.org/10.1103/physreva.81.022121[phys .
a * 81 * , 022121 ( 2010 ) . ]
b. hensen , h. bernien , a. e. drau , a. reiserer , n. kalb , m. s. blok , j. ruitenberg , r. f. l. vermeulen , r. n. schouten , c. abelln , w. amaya , v. pruneri , m. w. mitchell , m. markham , d. j. twitchen , d. elkouss , s. wehner , t. h. taminiau , and r. hanson , loophole - free bell inequality violation using electron spins separated by @xmath73 kilometres , http://dx.doi.org/10.1038/nature15759[nature ( london ) * 526 * , 682 ( 2015 ) . ]
j. v. kujala , e. n. dzhafarov , and j .- .
larsson , necessary and sufficient conditions for maximal noncontextuality in a broad class of quantum mechanical systems , http://dx.doi.org/10.1103/physrevlett.115.150401[phys .
lett . * 115 * , 150401 ( 2015 ) . ] v. dambrosio , f. bisesto , f. sciarrino , j. f. barra , g. lima , and a. cabello , device - independent certification of high - dimensional quantum systems , http://dx.doi.org/10.1103/physrevlett.112.140503[phys .
lett . * 112 * , 140503 ( 2014 ) . ] g. caas , s. etcheverry , e. s. gmez , c. e. saavedra , g. b. xavier , g. lima , and a. cabello , experimental implementation of an eight - dimensional kochen - specker set and observation of its connection with the greenberger - horne - zeilinger theorem , http://dx.doi.org/10.1103/physreva.90.012119[phys .
a * 90 * , 012119 ( 2014 ) . ]
g. caas , m. arias , s. etcheverry , e. s. gmez , a. cabello , g. b. xavier , and g. lima , applying the simplest kochen - specker set for quantum information processing , http://dx.doi.org/10.1103/physrevlett.113.090404[phys .
* 113 * , 090404 ( 2014 ) . ]
m. michler , h. weinfurter , and m. ukowski , experiments towards falsification of noncontextual hidden variable theories , http://dx.doi.org/10.1103/physrevlett.84.5457[phys .
* 84 * , 5457 ( 2000 ) . ]
r. apkiewicz , p. li , c. schaeff , n. langford , s. ramelow , m. wieniak , and a. zeilinger , experimental non - classicality of an indivisible quantum system , http://dx.doi.org/10.1038/nature10119[nature ( london ) * 474 * , 490 ( 2011 ) . ] y. hasegawa , r. loidl , g. badurek , m. baron , and h. rauch , violation of a bell - like inequality in single - neutron interferometry , http://dx.doi.org/10.1038/nature01881[nature ( london ) , * 425 * , 45 ( 2003 ) . ]
h. bartosik , j. klepp , c. schmitzer , s. sponar , a. cabello , h. rauch , and y. hasegawa , experimental test of quantum contextuality in neutron interferometry , http://dx.doi.org/10.1103/physrevlett.103.040403[phys .
lett . * 103 * , 040403 ( 2009 ) . ] o. moussa , c. a. ryan , d. g. cory , and r. laflamme , testing contextuality on quantum ensembles with one clean qubit , http://dx.doi.org/10.1103/physrevlett.104.160501[phys .
lett . * 104 * , 160501 ( 2010 ) . ] e. amselem , m. rdmark , m. bourennane , and a. cabello , state - independent quantum contextuality with single photons , http://dx.doi.org/10.1103/physrevlett.103.160405[phys .
rev . lett . * 103 * , 160405 ( 2009 ) . ] j. ahrens , e. amselem , a. cabello , and m. bourennane , two fundamental experimental tests of nonclassicality with qutrits , http://www.nature.com/srep/2013/130710/srep02170/full/srep02170.html[sci . rep . * 3 * , 2170 ( 2013 ) . ] b. marques , j. ahrens , m. nawareg , a. cabello , and m. bourennane , experimental observation of hardy - like quantum contextuality , http://dx.doi.org/10.1103/physrevlett.113.250403[phys .
* 113 * , 250403 ( 2014 ) . ]
d. frustaglia , j. p. baltans , m. c. velzquez - ahumada , a. fernndez - prieto , a. lujambio , v. losada , m. j. freire , and a. cabello , classical physics and the bounds of quantum correlations , http://arxiv.org / abs/1511.08144[. ] a. peres and a. ron , cryptodeterminism and quantum theory , in _ microphysical reality and quantum formalism _ , edited by a. van der merwe , f. selleri , and g. tarozzi ( kluwer academic , dordrecht , the netherlands , 1988 ) , vol . 2 , p. 115
m. pawowski , t. paterek , d. kaszlikowski , v. scarani , a. winter , and m. ukowski , information causality as a physical principle , http://dx.doi.org/10.1038/nature08400[nature ( london ) * 461 * , 1101 ( 2009 ) . ] | contextuality provides a unifying paradigm for nonclassical aspects of quantum probabilities and resources of quantum information .
unfortunately , most forms of quantum contextuality remain experimentally unexplored due to the difficulty of performing sequences of projective measurements on individual quantum systems . here
we show that two - point correlations between binary compatible observables are sufficient to reveal any form of contextuality .
this allows us to design simple experiments that are more robust against imperfections and easier to analyze , thus opening the door for observing interesting forms of contextuality , including those requiring quantum systems of high dimensions .
in addition , it allows us to connect contextuality to communication complexity scenarios and reformulate a recent result relating contextuality and quantum computation .
= 1 |
in the hot and dense system of strongly interacting matter created in heavy ion collisions the mean free path of the particles is much smaller than the size of the fireball .
this fact together with the assumption of fast thermal equilibration allows to apply relativistic hydrodynamics for the dynamical description of heavy ion collisions . at later times of the evolution and at its edges
the system is dilute , the mean free path is larger than system size , and hydrodynamics is not applicable .
an adequate description of the system in those circumstances is provided by kinetic ( transport ) equations , such as the boltzmann equation or its modifications for the quantum case .
state of the art simulations of heavy ion collisions couple hydrodynamics for the early stage of the evolution to hadron transport for the late stage .
such approaches are called hybrid approaches @xcite . the transition from hydrodynamics to transport ( so - called particlization ) is non - trivial , because hydrodynamics contains no microscopic information that is needed for transport . currently , most of the models implement particlization in the following way .
hydrodynamical equations are solved in the whole forward light cone , including those regions , where hydrodynamics is not applicable .
then the particlization hypersurface is found from the hydrodynamical evolution .
usually a hypersurface of constant time , temperature or energy density is taken .
particle distributions to be fed into the kinetic model are generated on this hypersurface according to the cooper - frye formula @xcite : @xmath2 here @xmath3 is a spectrum of particles emerging from an element of the hypersurface , @xmath4 is the normal four - vector to the hypersurface , its length being equal to the area of hypersurface element . in ideal fluid calculations
the distribution function @xmath5 is taken as an equilibrium fermi or bose distribution : @xmath6^{-1}$ ] .
an advantage of the cooper - frye formula is that it respects conservation laws .
the disadvantage is that for space - like elements of the hypersurface ( _ i.e. _ where @xmath7 ) there exist particle momenta @xmath8 such that @xmath9 and thus @xmath10 in the cooper - frye formula .
this is clearly unphysical , because the number of particles must be positive in every element of phase space .
however , in all practical cases after integration over the hypersurface one gets positive definite spectra .
one can split the integrated cooper - frye formula into positive and negative parts : @xmath11 the second term is called cooper - frye negative contributions and the first one - positive . for the applicability of the cooper - frye formula the negative contributions must be much smaller than the positive ones .
this is usually true for simulations of high energy collisions .
negative contributions of around 9% relative to the positive ones were reported for rhic top energy @xcite . at the same time negative contributions at @xmath12gev were found to be around 13% .
this suggests that at lower collision energies negative contributions might become large and cooper - frye formula will be inapplicable .
does it really happen ? to answer this question we systematically investigate the negative cooper - frye contributions differentially in rapidity and transverse momentum against collision energy , centrality and particle species .
several ways to circumvent the problem of negative contributions were suggested ( see @xcite for a short summary ) .
one practical way to do it in hybrid models would be to simultaneously neglect negative contributions and remove the particles from the transport calculation , if they fly to the hydrodynamic region .
if the distributions in hydrodynamics and cascade are the same , then conservation laws will be fulfilled .
such an approach has never been implemented . to check
if it is feasible , we compare negative contributions from the cooper - frye formula to distributions of backscattered particles from the cascade .
we take advantage of the coarse - grained approach , which allows to calculate both within one framework .
we aim at two goals : systematic estimation of cooper - frye negative contributions and comparing them to distributions of transport particles flying inwards to the hydrodynamics region . as a transport model
we take ultra - relativistic quantum molecular dynamics ( urqmd 3.3p2 ) @xcite .
urqmd allows to simulate heavy ion collisions as a sequence of elementary particle collisions .
included processes are 2 to 2 scattering , resonance formation and decays , string excitation and fragmentation .
we generate an ensemble of urqmd au+au collision events and average them on a rectangular grid to obtain the energy momentum tensor @xmath13 and the baryon current @xmath14 in each cell . in each cell
we find the landau rest frame ( lrf , a frame , where energy flow is zero : @xmath15 ) .
we obtain the energy density @xmath16 , flow velocity @xmath17 and baryon density in the lrf @xmath18 .
knowing @xmath19 for each grid cell we construct the hypersurface @xmath20 of constant @xmath19 and find the normal vectors @xmath4 for each piece of @xmath20 .
the latter is done using the cornelius subroutine @xcite , that provides a continuous surface without holes and avoids double counting of hypersurface pieces .
the hypersurface @xmath20 mimics the transition hypersurface in hybrid models .
when @xmath20 is obtained we perform a cooper - frye calculation on it and compare to distributions of underlying urqmd particles that cross @xmath20 .
we perform our calculations with time step @xmath21 = 0.1 fm / c , grid spacing in the beam direction @xmath220.3 fm , and grid spacings in transverse direction @xmath23 = 1 fm . for collision energy @xmath12 gev we take @xmath220.1 fm , and @xmath23 = 0.3 fm .
we have checked that for such a choice of grid spacing conservation laws on the surface are fulfilled with an accuracy better than 1% . in other words
@xmath24 and total energy flowing out of the hypersurface calculated by particles differ by no more than 1% . to create a smooth hypersurface and obtain reproducible results we employ a gaussian smearing procedure . for the construction of the hypersurface every urqmd particle with coordinates @xmath25 and 4-momentum @xmath8
is substituted by 300 marker particles with coordinates distributed with the probability density @xmath26 , where @xmath27 . in this way
every particle contributes to @xmath13 and @xmath14 not only of the cell , where it is located , but also to the adjacent cells .
we take as number of events @xmath28 and the gaussian width @xmath29 fm .
choice of @xmath30 , @xmath31 , grid spacing and sensitivity of results to these choices are discussed in @xcite .
if the distribution of urqmd particles is thermal on some closed hypersurface and the system is in chemical equilibrium , then the cooper - frye formula should give results identical to explicit particle counting .
this would allow to compensate negative contributions by removing particles from urqmd , if they cross the hypersurface inwards .
in such a treatment conservation laws on the surface would be respected .
we check , if this can be done on the hypersurface @xmath20 of constant energy density in landau frame , @xmath32 gev/@xmath33 .
( red circles ) are compared to distribution of urqmd pions crossing @xmath20 ( blue crosses ) .
left panel : negative contributions and inward crossings .
right panel : positive contributions and outward crossings .
collision energy @xmath34
@xmath35gev , central collisions . note the very different scale of negative and positive contributions.,title="fig:",height=181 ] ( red circles ) are compared to distribution of urqmd pions crossing @xmath20 ( blue crosses ) .
left panel : negative contributions and inward crossings .
right panel : positive contributions and outward crossings .
collision energy @xmath34
@xmath35gev , central collisions . note the very different scale of negative and positive contributions.,title="fig:",height=181 ] ( blue crosses ) , @xmath36 ( grey bars ) , nucleons ( magenta rectangles ) and @xmath37 ( green triangles ) .
cooper - frye calculation in central au+au collisions at @xmath38gev.,height=181 ] when calculating the net number of pions passing through the surface , one finds that the number of pions in urqmd is larger than in the equilibrated cooper - frye calculation @xcite
. it might be possible to explain this as a sign of chemical non - equilibrium in urqmd , but when one looks at the positive and negative contributions to the pion distributions shown in fig .
[ fig : pos_contr_rap ] , one sees that a difference in the pion density only is not sufficient to explain the differences in the contributions .
the positive contribution is much larger in urqmd than in the cooper - frye calculation , whereas the negative contribution depicts the opposite behaviour : in urqmd it is much smaller than in the cooper - frye scenario .
this kind of distributions may indicate that the collective flow velocity of pions is much larger than the collective velocity of other particles @xcite , or that the dissipative corrections are very large .
since neither negative contributions coincide with inward crossings , nor positive contributions coincide with outward crossings , the above mentioned idea of compensating negative contributions will not work in our case .
instead we concentrate on finding when negative contributions play the most prominent role , and evaluate the ratio of negative contributions to positive ones to estimate the error they bring into hybrid calculations . for
that we vary hadron sort , collision energy and centrality . as a relevant variable we consider the ratio of negative to positive contributions integrated over the hypersurface , @xmath39 . from fig .
[ fig : neg_contr_pmass ] one can see that negative contributions become smaller for larger particle mass .
it is simple to understand this result in the rest frame of the fluid element where the surface is moving .
if the hypersurface @xmath20 moves inwards , as is usually the case in fluid - dynamical calculations , a particle must be faster than the surface to cross it inwards .
for larger particle mass the average velocities of the thermal motion are smaller and the probability to catch up the hypersurface is also smaller .
therefore , to find maximal negative contributions it is sufficient to consider pions only . for them
we find out that if negative contributions are binned according to rapidity , then the largest contribution occurs at midrapidity .
we show the negative to positive contribution ratio for the pion yield at midrapidity for collision energies @xmath40@xmath41 gev in fig .
[ fig : neg_contr_ecoll ] .
the ratio in urqmd calculation is much smaller than in the equilibrated cooper - frye calculation at all collision energies , but the maximum lies in the region of @xmath42 gev in both approaches . the value of the maximum is about 13% in cooper - frye calculation , and about 4% in urqmd calculation .
gev/@xmath33 surface for pions at midrapidity .
red circles depict the ratio in a cooper - frye calculation assuming thermal equilibrium , and blue crosses in explicit calculation of urqmd particles.,height=196 ] gev at various centralities : @xmath43 ( red circles ) , @xmath44 fm ( blue crosses ) and @xmath45 fm ( green rectangles).,height=196 ] the dependency of negative to positive contribution ratio on collision energy is non - monotonous .
this is not obvious , because several factors influence the ratio : the temperature on the hypersurface , the relative velocities between the flow and the surface , and the relative amounts of volume and surface emission , _
i.e. _ emission from the time- and space - like parts of the surface .
larger temperature results in larger negative contributions because the thermal velocities increase .
larger relative velocity leads to smaller negative contributions .
the larger the relative amount of volume - emission , the smaller the negative contributions . with increasing collision energies the temperature saturates and
thus the changes in the last two factors make negative contributions fall with increasing collision energy .
the decrease of negative contributions at higher energy is predictable , because the relative amount of volume emission and relative velocity between flow and surface increase with collision energy .
however , the behaviour of negative contributions at lower energy is caused by the interplay of all three factors .
we plot negative contributions for pions versus collision centrality in fig .
[ fig : neg_contr_b ] , and find that for peripheral collisions negative contributions are smaller than for central collisions . the relative amount of surface and volume emission plays the most prominent role here . for peripheral collisions volume emission dominates due to the short lifetime of the system and negative contributions
are small . after varying the collision energy and the centrality
, we can conclude that in the worst scenario negative contributions can hardly exceed 15% . changing the criterion of hypersurface to @xmath46 gev/@xmath33
does not change this conclusion ( see ref .
we have investigated negative cooper - frye contributions and backscattering using a coarse - grained molecular dynamics approach .
au+au collisions at @xmath40@xmath47 gev energies have been simulated using urqmd , and a hypersurface @xmath20 of constant landau rest frame energy density has been constructed . on this surface
we have calculated two quantities : the ratio of cooper - frye negative to positive contributions ( @xmath48 ) , which assumes local thermal equilibrium , and the ratio of urqmd particles crossing @xmath20 inward to crossing @xmath20 outward ( @xmath49 ) , which does not assume equilibrium .
we found that at all collision energies @xmath50 .
we explain this by a deviation of pions in urqmd simulation from equilibrium .
a non - monotonous dependency of @xmath48 and @xmath49 on collision energy was found with a maximum at 10 - 20 @xmath35 gev , maximal @xmath48 being around 13% .
the size of the negative contributions is a result of an interplay of several factors : the temperature on the hypersurface , the relative velocities between flow and surface , and the relative amounts of volume and surface emission .
this work was supported by the helmholtz international center for the facility for antiproton and ion research ( hic for fair ) within the framework of the landes - offensive zur entwicklung wissenschaftlich - oekonomischer exzellenz ( loewe ) program launched by the state of hesse .
do and hp acknowledge funding of a helmholtz young investigator group vh - ng-822 from the helmholtz association and gsi , and ph by bmbf under contract no .
06fy9092 . do acknowledges support of hgs - hire .
computational resources have been provided by the center for scientific computing ( csc ) at the goethe university of frankfurt . | many models of heavy ion collisions employ relativistic hydrodynamics to describe the system evolution at high densities .
the cooper - frye formula is applied in most of these models to turn the hydrodynamical fields into particles . however , the number of particles obtained from the cooper - frye formula is not always positive - definite .
physically negative contributions of the cooper - frye formula are particles that stream backwards into the hydrodynamical region .
we quantify the cooper - frye negative contributions in a coarse - grained transport approach , which allows to compare them to the actual number of underlying particles crossing the transition hypersurface .
it is found that the number of underlying inward crossings is much smaller than the one the cooper - frye formula gives under the assumption of equilibrium distribution functions .
the magnitude of cooper - frye negative contributions is also investigated as a function of hadron mass , collision energy in the range @xmath0 gev , and collision centrality .
the largest negative contributions we find are around 13% for the pion yield at midrapidity at @xmath1 gev collisions . |
modern calculations @xcite based on relativistic equations of state indicate that the neutron star cores consist of neutrons with the admixture of protons , electrons , muons and some exotic particles ( including hyperons , k - mesons , quarks and so on ) .
the composition is governed by the charge neutrality and equilibrium of the medium under weak processes @xmath1 , @xmath2 , where @xmath3 and @xmath4 are baryons ( or quarks ) , and @xmath5 is a lepton , either an electron or a muon .
these reactions , widely known as the direct urca processes , are a central point of any modern scenarios of evolution of neutron stars .
neutrino energy losses caused by the direct urca processes lead to a rapid cooling of degenerate neutron star cores @xcite .
the corresponding reactions on nucleons , @xmath6 , @xmath7 , are the most powerful sources of neutrinos and antineutrinos in cooling neutron stars . in spite of
widely adopted importance of these reactions , the corresponding neutrino energy losses are not well investigated yet . a simple formula suggested by lattimer et al .
@xcite more than ten years ago has been derived in a non - relativistic manner .
actually , however , the superthreshold proton fraction , necessary for the direct urca processes to operate in the degenerate nuclear matter , appears at large densities , where the fermi momenta of participating nucleons are comparable with their effective mass .
moreover , according to modern numerical simulations , the central density of the star can be up to eight times larger than the nuclear saturation density @xcite .
this implies a substantially relativistic motion of nucleons in the superdense neutron star core .
the appropriate equation of state for such a matter is actually derived in the relativistic approach , and the relevant neutrino energy losses must be consistent with the used relativistic equation of state .
some aspects of this problem was studied by leinson and prez @xcite , who have estimated relativistic effects of baryon recoil and parity violation in the direct urca processes including also effects of the baryon mass difference .
this approach is useful , for example , for the direct urca reactions changing the baryon strangeness because , in this case , the contribution of the weak magnetism into the matrix element of the beta - decay is not well known even for free particles . for nucleons , the relativistic regime should incorporate the effects of weak magnetism and pseudoscalar interaction , which drastically influence the neutrino emissivity of the corresponding reactions . to demonstrate this , in the present paper we consider the totally relativistic direct urca process on nucleons .
we utilize the walecka - type relativistic model of baryon matter @xcite , where the baryons interact via exchange of @xmath8 , @xmath9 , and @xmath10 mesons , and perform the calculation of the neutrino energy losses in the mean field approximation .
this approximation is widely used in the theory of relativistic nuclear matter , and allows to calculate in a self - consistent way the composition of the matter together with energies , and effective masses of the baryons . in section
ii we begin with considering the relativistic kinematics of the neutron beta decay in the medium under beta equilibrium .
we consider a free gas model and some field theoretical models of nuclear matter to demonstrate that the direct neutron decay in neutron stars is open only due to strong interactions caused by isovector mesons .
we shortly discuss the energies and the wave functions of nucleons in the mean field approximation and demonstrate the nonconservation of the charged vector current of nucleons in this model .
the matrix element of the neutron beta decay is derived in section iii . in section
iv we calculate the neutrino energy losses caused by the direct urca on nucleons in the degenerate nuclear matter under chemical and thermal equilibrium . in section
v we inspect the non - relativistic limit of the neutrino energy losses in order to compare this with the expressions earlier obtained in @xcite and @xcite .
efficiency of the relativistic approach is numerically studied in section vi .
we evaluate the neutrino energy losses due to the direct urca processes on nucleons in the multicomponent baryon matter under beta equilibrium and compare the relativistic result with that predicted by the known non - relativistic formula .
we specially discuss the contributions of weak magnetism and pseudoscalar interaction .
summary and conclusion are in section vii . in appendix , we discuss some details of the model used for the nuclear matter . in what follows
we use the system of units @xmath11 and the boltzmann constant @xmath12 .
summation over repeated greek indexes is assumed
before calculating the neutrino energy losses caused by the direct urca processes , let us examine relativistic kinematics of the reaction , @xmath13 , in the degenerate matter under beta equilibrium . in what follows we consider massless neutrinos of energy and momentum @xmath14 with @xmath15 the energy - momentum of the final lepton @xmath16 of mass @xmath17
is denoted as @xmath18 with @xmath19 .
thus , the energy and momentum conservation in the beta decay is given by the following equations@xmath20where @xmath21 and @xmath22 are the in - medium energies of the neutron and the proton respectively .
the energy exchange in the matter goes naturally on the temperature scale @xmath23 , which is small compared to typical kinetic energies of degenerate particles
. therefore the momenta of in - medium fermions can be fixed at their values at fermi surfaces , which we denote as @xmath24 , @xmath25 for the nucleons and @xmath26 for leptons respectively .
since the antineutrino energy is @xmath27 , and the antineutrino momentum @xmath28 is much smaller than the momenta of other particles , we can neglect the neutrino contributions in eqs .
( [ cons ] ) . then the momentum conservation , @xmath29 , implies the well known triangle
condition , @xmath30necessary for the urca processes to operate .
however , the momentum conservation is necessary but insufficient condition for the direct beta decay to occur .
the energy conservation requires a one more condition@xmath31squaring of both sides of this equation gives@xmath32if we denote the energy - momentum transfer from the nucleon as @xmath33 the eq .
( [ ec ] ) can be readily recognized as @xmath34 .
this condition naturally arises from the time - like momentum of the final lepton pair , @xmath35 , and can be satisfied only in the presence of the energy gap between spectrums of protons and neutrons .
consider now the condition of beta equilibrium , @xmath36 , where @xmath37 , @xmath38 , and @xmath39 are the chemical potentials of neutrons , protons , and leptons respectively .
the chemical potentials of degenerate particles can be approximated by their individual fermi energies , yielding @xmath40generally ) is valid but @xmath41 .
] this equation is not the same as eq .
( econs ) . however , if the triangle condition ( [ triangle ] ) is fulfilled then the direct neutron decay is open , and one has @xmath42 . in this case
( [ chem ] ) becomes identical to the energy consrvation ( [ econs ] ) .
thus under beta equilibrium , the triangle
condition is the necessary and sufficient for the direct neutron decay to occur .
it should be emphasized however that the stated conditions are consistent only in the presence of the energy gap between protons and neutrons .
therefore the possibility of the direct neutron decay depends substantially on the model of nuclear matter .
consider , for example , a degenerate free gas consisting of neutrons , protons , and electrons under beta equilibrium . in this case
the energy gap exists only due to the mass difference of a neutron and a proton .
if we denote the masses as @xmath43 and @xmath44 respectively , then the corresponding fermi energies are @xmath45 @xmath46 , and @xmath47 @xmath48 .
due to charge neutrality the number density of electrons , @xmath49 , equals to the number density of protons , @xmath50 .
this implies @xmath51 and the equation of chemical equilibrium ( [ chem ] ) becomes of the form@xmath52solution of this equation @xmath53@xmath54 gives the proton fermi momentum as a function of the neutron fermi momentum in the beta - equilibrated gas of protons and neutrons . as required by the triangle
condition , @xmath55 , the direct neutron decay in such a medium is open when @xmath25 is larger than @xmath56 .
so , solution of the equation@xmath57@xmath54gives the critical value of the neutron fermi momentum above which the direct neutron decay is forbidden .
we find @xmath58thus the direct neutron decay is forbidden if the number density of neutrons is larger than @xmath59this number density is much smaller than that typical for neutron star cores , therefore the direct neutron decay in the cooling neutron stars can occur only due to strong interactions .
notice , some of models of strong interaction also forbid the direct neutron decay .
consider , for example , a simple model for the baryon matter wa74 , which contains fields for baryons and neutral scalar @xmath60 and vector @xmath61 mesons .
this model reproduces in a simple way the empirical nn scattering amplitude and the bulk nuclear properties .
however , the neutral scalar and vector mesons equally interact with protons and neutrons , which therefore have identical energy spectrums . in this case , the energy and momentum in the direct neutron decay can not be conserved simultaneously . to allow the direct neutron decay ,
the model of nuclear matter must be generalized to include some additional degrees of freedom and couplings , which are able to create a large energy gap between possible energies of the proton and neutron . for this purpose , besides isoscalar mesons @xmath8 and @xmath9 , the model should include also isovector mesons .
the isovector meson couples differently to protons and neutrons , thus creating the energy gap necessary for the direct neutron decay . in the following we consider a self - consistent relativistic model of nuclear matter in which baryons interact via exchange of isoscalar mesons @xmath8 and @xmath9 and an isovector meson @xmath10 ( see appendix ) . in the mean field approximation ,
when the contribution of mesons reduce to classical condensate fields @xmath62 , @xmath63 , @xmath64 , only the baryon fields must be quantized .
this procedure yields the following linear dirac equation for the nucleon @xmath65here and below we denote as @xmath66 , and @xmath67 the components of isospin operator , which act on the isobaric doublet @xmath68 of nucleon field ; @xmath69 is the bare nucleon mass .
the stationary and uniform condensate fields equally shift the effective masses @xmath70but lead to different potential energies of the proton and neutron @xmath71thus creating the energy gap @xmath72 between possible energies of protons and neutrons . the exact solutions of eq .
( [ diraceq ] ) can be found separately for protons and for neutrons . in our case of a stationary , uniform system , solution for the neutron
is a spinor plane wave @xmath73where the neutron energy is given by @xmath74 , and the spinor @xmath75 should be found from the following equation@xmath76it is easily to understand that solution of this equation @xmath77 is a free - like spinor constructed from the neutron _
momentum @xmath78for the proton , one has@xmath79where the spinor @xmath80 obeys the equation@xmath81and depends respectively on the proton kinetic momentum @xmath82 in what follows we denote by @xmath83 the kinetic energy of the neutron and the proton respectively .
so the normalization factors are of the form @xmath84and the single - particle energies are @xmath85 , and @xmath86 . at the fermi surfaces the single - particle energies are @xmath87 , and @xmath88 . in this case
( [ ec ] ) takes the form@xmath89typically @xmath90 , so this equation can be readily satisfied simultaneously with the momentum conservation @xmath91 .
the lagrangian density ( [ lagr ] ) ensures a conserved isovector current @xcite @xmath92where @xmath93 is the @xmath10 meson field strength tensor .
besides the directly nucleon contribution the conserved current ( [ cic ] ) includes also the contribution of the isovector field @xmath94 , which obeys the field equations @xmath95by the use of eq .
( [ beq ] ) the condition @xmath96 may be transformed as@xmath97 in the mean field approximation this gives @xmath98by introducing the covariant derivative @xmath99we can recast the eq .
( [ ncnc ] ) to the following form@xmath100at the level of matrix elements this can be written as@xmath101where @xmath102 is the kinetic momentum transfer @xmath103thus the matrix element of the transition current is orthogonal to the _ kinetic _ momentum transfer but not to the total momentum transfer from the nucleon .
note that this effect originates not from a special form ( [ cic ] ) of the conserved isovector current in the medium but is caused by the energy gap between the proton and neutron spectrums .
this follows directly from the dirac equation ( [ diraceq ] ) , which ensures the eq .
( [ orthq ] ) with @xmath104which coincides with eq .
( [ q ] ) because @xmath105 .
in the lowest order in the fermi weak coupling constant @xmath106 , the matrix element of the neutron beta decay is of the form@xmath107were @xmath108 is the cabibbo factor .
the effective charged weak current in the medium consists of the polar vector and the axial vector , @xmath109 .
our goal now is to derive the nucleon matrix element of the charged weak current in the medium .
consider first the polar - vector contribution . by the use of the isovector current ( [ cic ] ) one
can construct the conserved electromagnetic current in the medium@xmath110the last term in eq .
( [ jem ] ) is the pauli contribution , where @xmath111 and @xmath112 in the mean field approximation , we replace the magnetic formfactors of the nucleon with anomalous magnetic moments of the proton and the neutron , @xmath113 and @xmath114 . by the conserved - vector - current theory ( cvc ) , the nucleon matrix element of the charged vector weak current is given by @xmath115for the electromagnetic transitions one has@xmath116@xmath117with @xmath118 , as given by eq .
( [ q ] ) .
thus , in the mean field approximation , we obtain @xmath119 u_{{\rm n}}\left ( p\right ) . \label{wcv}\]]the second term in eq .
( [ wcv ] ) , describes the weak magnetism effects . by the use of the dirac eqs .
( [ uneq ] ) and ( [ upeq ] ) for the nucleon spinors we find @xmath120 in accord with eq .
( [ orthq ] ) .
consider now the axial - vector charged current .
this current is responsible for both the @xmath121 transitions and the pion decay . in the limit of chiral symmetry , @xmath122
, the axial - vector current must be conserved . in the medium with @xmath10 meson condensate , this implies @xmath123where the covariant derivative @xmath124 is defined by eq .
( [ cd ] ) . at the finite mass of a pion , @xmath125
, the axial - vector charged current is connected to the field @xmath126 of @xmath127 meson . for a free space ,
this relation is known as the hypothesis of partial conservation of the axial current ( pcac ) . in the medium
the pcac takes the form @xmath128where @xmath129 is the mass of @xmath130-meson , and @xmath131 is the pion decay constant . with allowing for interactions of the pions with nucleons and @xmath10 mesons
the lagrangian density for the pion field is of the form @xcite@xmath132 + ig_{\pi } \bar{\psi}\gamma _ { 5}{\vec \tau \cdot \vec \pi } \psi \text{. } \label{lpi}\]]in the mean field approximation this results in the following equation for the field of @xmath127 meson@xmath133where @xmath134 is the pion - nucleon coupling constant .
for the nucleon transition of our interest , the eq .
( [ pcac ] ) gives @xmath135here the right - hand side can be calculated by the use of eq .
( [ pifield ] ) .
we obtain@xmath136this equation allows to derive the nucleon matrix element of the axial - vector charged current . really , to construct the axial - vector matrix element of the charged current , caused by the nucleon transition
, we have only two independent pseudovectors , consistent with invariance of strong interactions under @xmath137 isospin transformation , namely : @xmath138 , and @xmath139 .
this means that the matrix element of the axial - vector charged current is of following general form @xmath140here , in the mean field approximation , we set @xmath141 , while @xmath142 is the form - factor to be chosen to satisfy the eq .
( [ mepcac ] ) , which now reads@xmath143to obtain this we used the dirac eqs .
( [ uneq ] ) and ( [ upeq ] ) for the nucleon spinors .
thus@xmath144 in the mean field approximation , we assume that the coupling constants are independent of the momentum transfer . by setting @xmath145 in eq .
( [ feq ] ) we obtain the goldberger - treiman relation@xmath146by inserting this in ( [ feq ] ) we find@xmath147thus , with taking into account eqs .
( [ wcv ] ) , ( [ ameg ] ) , and ( [ fq ] ) , the total matrix element of the neutron beta decay is found to be@xmath148 u_{{\rm n}}\left ( p\right ) , \nonumber\end{aligned}\]]where , in the mean field approximation , we assume @xmath149note that the matrix element obtained is of the same form as that for the neutron decay in a free space , but with the total momentum transfer replaced with the kinetic momentum transfer . due to the difference in the neutron and proton potential energy ,
the kinetic momentum transfer @xmath150to be used in the matrix element ( [ mend ] ) differs from the total momentum of the final lepton pair @xmath151this ensures @xmath152 , while @xmath153 .
the square of the matrix element of the reaction summed over spins of initial and final particles is found to be .
the author is grateful to m. prakash and s. ratkovi who have pointed out this error . ] : @xmath154 \nonumber \\ & & -\frac{c_{m}^{2}}{4m^{2}}\left [ m^{\ast 2}\left ( k_{1}p_{2}\right ) \left ( 3\left ( k_{2}p_{2}\right ) -\left ( k_{2}p_{1}\right ) \right ) \right .
\nonumber \\ & & + m^{\ast 2}\left ( k_{1}p_{1}\right ) \left ( 3\left ( k_{2}p_{1}\right ) -\left ( k_{2}p_{2}\right )
\right ) + \left ( k_{1}k_{2}\right ) \left ( p_{1}p_{2}-m^{\ast 2}\right ) ^{2 } \nonumber \\ & & \left . -\left ( k_{1}p_{1}+k_{1}p_{2}\right ) \left ( k_{2}p_{1}+k_{2}p_{2}\right ) \left ( p_{1}p_{2}\right ) \right ] \nonumber \\ & & + c_{a}^{2}f_{q}\left ( 2m^{\ast } + f_{q}\left ( m^{\ast 2}-\left ( p_{1}p_{2}\right ) \right ) \right ) \left [ \left ( k_{1}k_{2}\right ) \left ( m^{\ast 2}-\left ( p_{1}p_{2}\right ) \right ) \right .
\nonumber \\ & & \left .
-\left ( k_{1}p_{1}-k_{1}p_{2}\right ) \left ( k_{2}p_{1}-k_{2}p_{2}\right ) \right ] \right ] \label{matrel}\end{aligned}\]]with @xmath155 and @xmath156 .
we consider the total energy which is emitted into neutrino and antineutrino per unit volume and time . within beta equilibrium
, the inverse reaction @xmath157 corresponding to a capture of the lepton @xmath5 , gives the same emissivity as the beta decay , but in neutrinos .
thus , the total energy loss @xmath158 for the urca processes is twice more than that caused by the beta decay . taking this into account by fermi s golden rule we have@xmath159antineutrinos are assumed to be freely escaping .
the distribution function of initial neutrons as well as blocking of final states of the proton and the lepton @xmath5 are taken into account by the pauli blocking - factor @xmath160 .
the fermi - dirac distribution function of leptons is given by @xmath161while the individual fermi distributions of nucleons are of the form @xmath162@xmath163 by neglecting the chemical potential of escaping neutrinos , we can write the condition of chemical equilibrium as @xmath164 . then by the use of the energy conservation equation , @xmath165 , and taking the total energy of the final lepton and antineutrino as @xmath166 we can recast the blocking - factor as @xmath167where @xmath168 .
furthermore , since the antineutrino energy is @xmath27 , and the antineutrino momentum @xmath169 is much smaller than the momenta of other particles , we can neglect the neutrino contributions in the energy - momentum conserving delta - functions@xmath170and perform integral over @xmath171 to obtain @xmath172 in the next integrals .
nucleons and leptons in the neutron star core are strongly degenerate , therefore the main contribution to the integral ( [ gold ] ) comes from narrow regions near the corresponding fermi momenta .
thus we can set @xmath173 , @xmath174 in all smooth functions under the integral .
the energy of the final lepton is close to its fermi energy @xmath175 . here
the chemical potentials of nucleons can be approximated by their individual fermi energies @xmath176 , @xmath177 .
this allows us to transform the energy - conserving @xmath178-function as @xmath179where @xmath180 is the angle between the momentum @xmath181 of the initial neutron and the momentum @xmath182 of the final lepton .
notice , when the baryon and lepton momenta are at their individual fermi surfaces , the @xmath178- function ( [ delfun ] ) does not vanish only if @xmath183 .
further we use the particular frame with @xmath184-axis directed along the neutron momentum @xmath181 .
then@xmath185the energy - momentum of the final proton is defined by conservation laws:@xmath186insertion of ( [ pk ] ) and ( [ p2 ] ) in the square of the matrix element ( [ matrel ] ) yields a rather cumbersome expression , which , however , is readily integrable over solid angles of the particles .
since we focus on the actually important case of degenerate nucleons and leptons , we may consider the neutrino energy losses to the lowest accuracy in @xmath187 .
then the remaining integration reduces to the factor @xmath188finally the neutrino emissivity is found to be of the form)]:@xmath189 \nonumber \\ & & + c_{v}c_{m}\frac{m^{\ast } } { 2m}\left [ 2\left ( \varepsilon _ { { \rm n}% }
-\varepsilon _ { { \rm p}}\right ) p_{l}^{2}-\left ( 3\left ( \varepsilon _ { { \rm n% } } -\varepsilon _ { { \rm p}}\right ) ^{2}-p_{l}^{2}\right ) \mu _ { l}\right ] \nonumber \\ & & + c_{a}\left ( c_{v}+\frac{m^{\ast } } { m}c_{m}\right ) \left ( \varepsilon _ { % { \rm n}}+\varepsilon _ { { \rm p}}\right ) \left ( p_{l}^{2}-\left ( \varepsilon _ { % { \rm n}}-\varepsilon _ { { \rm p}}\right ) ^{2}\right ) \nonumber \\ & & + c_{m}^{2}\frac{1}{16m^{2}}\left [ 8m^{\ast 2}\left ( \varepsilon _ { { \rm n}% }
-\varepsilon _ { { \rm p}}\right ) \left ( p_{l}^{2}-\left ( \varepsilon _ { { \rm n}% } -\varepsilon _ { { \rm p}}\right ) \mu _ { l}\right ) \right .
\nonumber \\ & & + \left ( p_{l}^{2}-\left ( \varepsilon _ { { \rm n}}-\varepsilon _ { { \rm p}% } \right ) ^{2}\right ) \left ( 2\varepsilon _ { { \rm n}}^{2}+2\varepsilon _ { { \rm p% } } ^{2}-p_{l}^{2}\right ) \mu _ { l } \nonumber \\ & & \left .
-\left ( p_{l}^{2}-\left ( \varepsilon _ { { \rm n}}-\varepsilon _ { { \rm p% } } \right ) ^{2}\right ) \left ( \varepsilon _ { { \rm n}}+\varepsilon _ { { \rm p}% } \right ) ^{2}\left ( 2\varepsilon _ { { \rm n}}-2\varepsilon _ { { \rm p}}-\mu _ { l}\right ) \right ] \nonumber \\ & & \left .
-c_{a}^{2}m^{\ast 2}\phi \left ( 1+m_{\pi } ^{2}\phi
\right ) \left [ \mu _ { l}\left ( \left ( \varepsilon _ { { \rm n}}-\varepsilon _ { { \rm p}}\right ) ^{2}+p_{l}^{2}\right ) -2\left ( \varepsilon _ { { \rm n}}-\varepsilon _ { { \rm p}% } \right ) p_{l}^{2}\right ] \right\ } \label{qmfa}\end{aligned}\ ] ] with @xmath190 if @xmath191 and zero otherwise . in the above , the last term , with @xmath54 @xmath192represents the contribution of the pseudoscalar interaction .
the triangle
condition @xmath193 , required by the step - function , is necessary for conservation of the total momentum in the reaction and exhibits the threshold dependence on the proton concentration .
under beta - equilibrium , the superthreshold proton fraction in the core of neutron stars appears at large densities , when fermi momenta of nucleons are of the order of their effective mass .
therefore the triangle
condition is inconsistent with the non - relativistic limit . nevertheless , it is necessary to show how the expression given by lattimer et al .
@xcite can be formally obtained from eq .
( [ qmfa ] ) . for this purpose
we consider the non - relativistic limit of eq .
( [ qmfa ] ) by neglecting the triangle condition .
the non - relativistic approximation is valid when @xmath194 .
however , the smallness of the particle momenta is not enough to point out unambiguously the leading terms in eq .
( [ qmfa ] ) .
the relative contributions of various terms depend also on electron abundance in the medium .
if the electron fraction is as large that @xmath195 then we obtain the result of lattimer et al .
@xcite : @xmath196widely used in simulations of neutron stars . however , when @xmath197 , what actually takes place in the non - relativistic nucleon matter under beta - equilibrium , we obtain@xmath198 \theta \left ( p_{l}+p_{{\rm p}}-p_{{\rm n}}\right ) . \label{qnr}\]]here the additional term is due to the nucleon recoil . when the electron ( proton ) fraction is small , this contribution is comparable to the terms of eq .
( [ ql ] ) .
neutrino energy losses caused by the direct urca on nucleons depend essentially on the composition of beta - stable nuclear matter .
therefore , in order to estimate the relativistic effects , we consider the model of nuclear matter , which includes nucleon and hyperon degrees of freedom ( see appendix ) .
the parameters of the model are chosen as suggested in ref .
gm to reproduce the nuclear matter equilibrium density , the binding energy per nucleon , the symmetry energy , the compression modulus , and the nucleon effective mass at saturation density @xmath199 @xmath200 .
the composition of neutrino - free matter in beta equilibrium among nucleons , hyperons , electrons and muons is shown in fig .
1 versus the baryon number density @xmath201 , in units of @xmath202 ( left panel ) .
the right panel represents the results of self - consistent calculation of the effective baryon masses and individual fermi momenta of the baryons ) to relativistic nucleons the authors @xcite suggest replacing the effective dirac mass with the relativistic fermi energy , @xmath203 . the latter is termed as the effective landau mass .
this trick can not be justified from the theoretical point of view .
firstly because the mass term and the energy term in the dirac equation are of different matrix structure . secondly , by the correct definition , the effective landau mass has no relation to relativism but is intended to take into account interactions between particles near the fermi surface . in this meaning the effective mass ( [ mef ] ) is actually the effective landau mass . ] . at number densities smaller than the saturation density
the matter consists mostly of neutrons with a small admixture of equal amounts of protons and electrons , however , the proton fraction grows along with the matter density .
near saturation density the electron chemical potential reaches the muon mass , and the muon fraction appears also growing along with the matter density together with the electron fraction up to @xmath204 @xmath200 . above this density of nuclear matter
, constituents other than neutrons , protons , electrons , and muons appear .
the @xmath205 hyperons appear when their lowest energy state first lies below @xmath206 . due to the charge neutrality ,
increasing of the @xmath205 fraction suppresses the lepton abundance . 0.3 cm fig .
1 . the left panel shows individual concentrations for matter in beta equilibrium among nucleons , hyperons , electrons and muons as a function of the density ratio @xmath207 .
the right panel represents the self - consistent effective baryon masses ( solid lines ) and individual fermi momenta of the baryons ( dashed lines ) . 0.3 cm above the density @xmath208 @xmath200 , beta equilibrium requires appearance of @xmath209 hyperons because the lowest energy state for a @xmath210 lies lower than @xmath37 . at densities larger than @xmath211 @xmath200 fraction of @xmath212 hyperons also exists . in fig .
2 we compare the relativistic neutrino emissivity ( [ qmfa ] ) with the non - relativistic energy losses , as given by formula ( [ ql ] ) . due to a steady decrease of the effective mass of the nucleon the non - relativistic formula ( short - dashed curve ) predicts a decreasing of the emissivity along with the density increase .
appearance of hyperons in the system suppresses the nucleon fractions and lepton abundance .
therefore at densities , where the number of hyperons is comparable with the number of protons , the relativistic emissivity ( solid curve ) reaches the maximum and then also has a tendency to decrease .
the relativistic emissivity , however , is found to be substantially larger than that predicted in the non - relativistic approach . to inspect the contribution of weak magnetism and the pseudoscalar interaction
we demonstrate two additional graphs .
the long - dashed curve demonstrates the energy losses obtained from eq .
( [ qmfa ] ) by formal setting @xmath213 .
this eliminates the pseudoscalar contribution .
the dot - dashed curve is obtained by formal replacing @xmath213 and @xmath214 , which eliminates both the weak magnetism and pseudoscalar contributions .
a comparison of these curves demonstrates the weak magnetism effects .
the contribution of the pseudoscalar interaction can be observed by comparing the total neutrino energy losses ( solid curve ) with the long - dashed curve , which is calculated without this contribution .
we see that , due to weak magnetism effects , relativistic emissivities increase by approximately 40 - 50% , while the pseudoscalar interaction only slightly suppresses the energy losses , approximately by 5% . 0.3 cm fig .
2 . the left panel shows the neutrino emissivity of the direct urca processes among nucleons and electrons .
the right panel is the same , but for the muon reactions .
the emissivity is shown versus the density ratio @xmath207 for the matter composition represented in fig .
1 . the curves begin at the threshold density .
solid curves represent the total relativistic emissivity , as given by eq .
the short - dashed curves are the non - relativistic emissivity given by eq .
( [ ql ] ) .
the dot - dashed curves show the relativistic emissivity without the contributions of weak magnetism and pseudoscalar interaction , and the long - dashed curves are the emissivity without the pseudoscalar contribution .
all the emissivities are given in units @xmath215 @xmath216 , where the temperature @xmath217 .
in the mean field approximation , we have studied the neutrino energy losses caused by the direct urca processes on nucleons in the degenerate baryon matter under beta equilibrium .
we have shown that the direct urca processes in a superdense matter of neutron star cores are kinematically allowed only due to isovector mesons , which differently interact with protons and neutrons . by creating the energy gap between proton and neutron
spectrums the isovector mesons support a time - like total momentum transfer from the nucleon , as required by kinematics of the reaction . in the mean field approximation , we derived the matrix element of the nucleon transition current , which is found to be a function of the space - like kinetic momentum transfer .
we have calculated the neutrino energy losses caused by the direct urca processes on nucleons .
( [ qmfa ] ) for neutrino energy losses exactly incorporates the effects of nucleon recoil , parity violation , weak magnetism , and pseudoscalar interaction . to quantify the relativistic effects we consider a self - consistent relativistic model , widely used in the theory of relativistic nuclear matter .
the relativistic energy losses are up to four times larger than those given by the non - relativistic approach . in our analysis
, we pay special attention to the effects of weak magnetism and pseudoscalar interaction in the neutrino energy losses .
we found that , due to weak magnetism effects , relativistic emissivities increase by approximately 40 - 50% , while the pseudoscalar interaction only slightly suppresses the energy losses , approximately by 5% . the mean field eq .
( qmfa ) may be considered as a starting point for studying of the correlation effects .
this work was partially supported by the russian foundation for fundamental research grant 00 - 02 - 16271
we employ a self - consistent relativistic model of nuclear matter in which baryons , @xmath218 , interact via exchange of @xmath8 , @xmath9 , and @xmath10 mesons @xcite . in principle , the pion fields should be also included in the model .
however , the expectation value of the pion field equals zero , giving no contribution to the mean fields .
therefore , only non - redundant terms are exhibited in the lagrangian density@xmath219 b \nonumber \\ & & -\frac{1}{4}f_{\mu \nu } f^{\mu \nu } + \frac{1}{2}m_{\omega } ^{2}\omega _ { \mu } \omega ^{\mu } -\frac{1}{4}{\vec b}_{\mu \nu } { \vec b}^{\mu \nu } + \frac{1% } { 2}m_{\rho } ^{2}{\vec b}_{\mu } { \vec b}^{\mu } \nonumber \\ & & + \frac{1}{2}\left ( \partial _ { \mu } \sigma \partial ^{\mu } \sigma -m_{\sigma } ^{2}\sigma ^{2}\right ) -u\left ( \sigma \right ) + \bar{l}\left ( i\gamma _ { \mu } \partial ^{\mu } -m_{l}\right ) l , \label{lagr}\end{aligned}\]]which includes the interaction of baryon fields @xmath220 with a scalar field @xmath221 , a vector field @xmath222 and an isovector field @xmath223 of @xmath10-meson . in the above , @xmath220 are the dirac spinor fields for baryons , @xmath224 is the isovector field of @xmath10-meson .
we denote as @xmath225 the isospin operator , which acts on the baryons of the bare mass @xmath226 .
the leptons are represented only by electrons and muons , @xmath16 , which are included in the model as noninteracting particles .
the field strength tensors for the @xmath9 and @xmath10 mesons are @xmath227 and @xmath228 , respectively .
the potential @xmath229 represents the self - interactions of the scalar field and is taken to be of the form @xmath230 in what follows we consider the mean field approximation widely used in the theory of relativistic nuclear matter . in this approximation , the meson fields are replaced with their expectation values@xmath231 in this case only the baryon fields must be quantized .
this procedure yields the following linear dirac equation for the baryon field @xmath232the effective baryon mass is @xmath233 , and the single - particle energies and self - consistent potentials are given by @xmath234here @xmath235 , @xmath236 , and @xmath237 are , respestively , nonzero average values of the @xmath9- , @xmath10- , and @xmath8-meson fields , and @xmath238 , @xmath239 , and @xmath240 are the strong interaction couplings of the baryons to different meson fields , and @xmath241 is the third component of isospin for the baryons . the number density for baryon species @xmath220
is related to the corresponding fermi momentum as @xmath242thus , the average meson fields are solutions of the following equations @xmath243for a fixed total baryon number density @xmath244the additional conditions needed to obtain a solution are given by the charge neutrality requirement@xmath245and the beta equilibrium relations @xmath246for nucleons , the strong interaction constants @xcite @xmath247with @xmath248 mev,@xmath249 mev,@xmath250 mev , are determined by reproducing the nuclear matter equilibrium density @xmath199 @xmath200 , and the binding energy per nucleon ( @xmath251mev ) , the symmetry energy ( @xmath252mev ) , the compression modulus ( @xmath253mev@xmath254mev ) , and the nucleon effective mass @xmath255mev at @xmath202 . for hyperons
we take @xmath256 . | we use the field theoretical model to perform relativistic calculations of neutrino energy losses caused by the direct urca processes on nucleons in the degenerate baryon matter . by our analysis ,
the direct neutron decay in the superdense nuclear matter under beta equilibrium is open only due to the isovector meson fields , which create a large energy gap between protons and neutrons in the medium .
our expression for the neutrino energy losses , obtained in the mean field approximation , incorporates the effects of nucleon recoil , parity violation , weak magnetism , and pseudoscalar interaction . for numerical testing of our formula
, we use a self - consistent relativistic model of the multicomponent baryon matter .
the relativistic emissivity of the direct urca reactions is found substantially larger than predicted in the non - relativistic approach .
we found that , due to weak magnetism effects , relativistic emissivities increase by approximately 40 - 50% , while the pseudoscalar interaction only slightly suppresses the energy losses , approximately by 5% . #
1@xmath0#1 psfig.sty |
dwarf galaxies have come to play an increasingly important role in understanding how galaxies form and evolve . as the smallest ,
least luminous , and most common systems in the universe , dwarf galaxies span a wide range in physical characteristics and occupy a diverse set of environments ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , making them excellent laboratories for direct studies of cause and effect in galaxy evolution .
the low average masses and metallicities of dwarf galaxies suggest they may be the best available analogs to the seeds of hierarchical galaxy formation in the early universe .
a cohesive picture of the evolution of dwarf galaxies remains elusive .
historically , evolutionary scenarios have often been considered in the context of a dual morphological classification , namely dwarf spheroidals ( dsphs ; we include dwarf ellipticals in this general category ) and dwarf irregulars ( dis ) .
the former are classified based on a smooth morphology , with no observed knots of sf , and are generally found to be gas - poor . the latter exhibit morphological evidence for current / recent sf activity and have a high gas fraction . a third , rarer class of
so - called ` transition ' dwarf galaxies , have a high gas fraction yet little to no recent sf activity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these galaxies may be an evolutionary link between dis and dsphs ( e.g. , * ? ? ? * ) or simply could be ordinary dis witnessed between massive star forming events ( e.g. , * ? ? ?
understanding the relationship between these three types of dwarf galaxies , specifically determining if and how dis evolve into dsphs , is among the most pressing questions in dwarf galaxy evolution ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. resolved stellar populations have proven to be an incredibly powerful tool for observationally constraining scenarios of dwarf galaxy evolution .
past patterns of sf and chemical evolution are encoded in a galaxy s optical color - magnitude diagram ( cmd ) . to extract this information , a number of sophisticated algorithms ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have been developed to measure the star formation history ( sfh ) , i.e. , the star formation rate ( sfr ) as a function of time and metallicity , by comparing observed cmds with those generated from models of stellar evolution .
the robustness of this method has become increasingly solidified with a variety of techniques and stellar models converging on consistent solutions for a range of galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
analysis of cmd - based sfhs have become particularly prevalent in studies of the formation and evolution of the local group ( lg ) .
the lg contains @xmath2 80 dwarf galaxies ranging from highly isolated to strongly interacting ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
results from cmd - based sfhs of dwarf galaxies in the lg have revealed that both dsphs and dis feature complex sfhs , typically with dominant stellar components older than 10 gyr ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the sfhs of individual lg dwarf galaxies , as well as aggregate compilations , now serve as the basis for our understanding of the evolution of dwarf galaxies and have significantly advanced our knowledge of galactic group dynamics ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . beyond the lg , only a small subset of dwarf galaxies
have explicitly measured cmd - based sfhs ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
ground - based observations of resolved stellar populations in more distant galaxies are challenging , due to the faintness of individual stars and the effects of photometric crowding .
however , observations from the hubble space telescope ( hst ) , particularly the advanced camera for surveys ( acs ; * ? ? ?
* ) , have revolutionized the field of resolved stellar populations , producing stunning cmds of dwarf galaxies both in and beyond the lg . outside the lg ,
past hst observations of resolved stellar populations have been fairly piecemeal , with individual or small sets of galaxies as the typical targets .
subsequent analysis often employed different methodologies or sought different science goals .
this lack of uniformity also makes it challenging to place the results from lg studies in the context of the broader universe , as lg dwarf galaxies may not be representative of the larger dwarf galaxy population ( e.g. , * ? ? ?
the acs nearby galaxy survey treasury ( angst ; * ? ? ?
* ) was designed to help remedy this situation .
using a combination of new and archival imaging taken with the hst / acs and wide field planetary camera 2 ( wfpc2 ; * ? ? ?
* ) , angst provides a uniformly reduced , multi - color photometric database of the resolved stellar populations of a volume - limited sample of nearby galaxies ( d @xmath1 4 mpc ) that are strictly outside the lg . of the @xmath2 70 galaxies in this sample , 60 are dwarf galaxies that span a range of @xmath2 10 in m@xmath5 and that reside in both isolated field and strongly interacting group settings , providing an unbiased statistical sample in which to study the detailed properties dwarf galaxy formation and evolution . in this paper
, we present the uniformly analyzed sfhs of 60 dwarf galaxies based on observations , photometry , and artificial star tests produced by the angst program .
the focus of this study is on the lifetime sfhs of the sample galaxies , with the recent ( @xmath6 1 gyr ) sfhs the subject of a separate paper @xcite .
we first briefly review the sample selection , observations , and photometry in [ data ] .
we then summarize the technique of measuring sfhs in [ sfhs ] . in [ results ] , we discuss and compare the resultant sfhs in the context of both dwarf galaxy formation and evolution .
we then explore our results with respect to the morphology
density relationship in [ morph ] .
finally , we discuss the evolution of dwarf galaxies in the context of cosmology in [ cosmic ] .
cosmological parameters used in this paper assume a standard wmap-7 cosmology as detailed in @xcite .
in this section , we briefly summarize the selection of the angst dwarf galaxy sample along with the observations and photometry .
a more detailed discussion of the angst program can be found in @xcite .
we constructed the initial list of angst galaxies based on the catalog of neighboring galaxies @xcite , consisting exclusively of galaxies located beyond the zero velocity surface of the lg @xcite .
we selected potential targets with @xmath7 @xmath8 @xmath9 to avoid observational difficulties associated with low galactic latitudes . by simulating cmds and crowding limits
, we found that a maximum distance of @xmath2 3.5 mpc provides the optimal balance between observational efficiency and achieving the program science goals . because the sample of galaxies within 3.5 mpc contains predominantly field galaxies
, we chose to extend the distance limits in the direction of the m81 group .
this extension added to the diversity of galaxies in the sample , while maintaining the goal of observational efficiency , due to the m81 group s close proximity to the fiducial distance limit ( d@xmath10 @xmath2 3.6 mpc ) and low foreground extinction values .
similarly , we also included galaxies in the direction of the ngc 253 clump ( d@xmath11 @xmath2 3.9 mpc ) in the sculptor filament @xcite , further extending the range of environments probed by the sample , while maintaining strict volume limits for galaxy selection . in this paper
, we analyze a sample of 60 angst dwarf galaxies from both new and comparable quality archival imaging .
the galaxies range in m@xmath5 from @xmath128.23 ( kk 230 ) to @xmath1217.77 ( ngc 55 ) and in distance from 1.3 mpc ( sex a ) to 4.6 mpc ( ddo 165 ) .
we have included most galaxies considered dwarf galaxies in the literature in our analysis , although the upper mass / luminosity cutoff for what constitutes a dwarf galaxy is somewhat ambiguous ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for example , @xcite chose to exclude the lmc ( @xmath13 @xmath1217.93 ) and smc ( @[email protected] ) from the ` dwarfs ' category .
analogous to @xcite , we have not included ngc 3077 ( @[email protected] ) , ngc 2976 ( @xmath13 @xmath1216.77 ) , and ngc 300 ( @xmath13 @xmath1217.66 ) , which would be among the brightest and most massive galaxies in the sample .
detailed studies of the sfhs of ngc 2976 @xcite and ngc 300 @xcite are available in the literature . a full list of sample galaxies and their properties are listed in table [ tab1 ] with the distances and blue luminosities of the sample shown in figure [ sample ] .
although the sample of angst dwarf galaxies is extensive , it is not complete within a fixed distance limit .
galaxies at low galactic latitudes have been intentionally excluded from the original volume selection to avoid complications associated with high degrees of reddening . while sfhs can still be derived from cmds of such galaxies ( e.g. , ic 4662 ; * ? ? ?
* ; * ? ? ?
* ) , the effects of extreme reddening can lead to larger uncertainties and can require special analysis techniques ( e.g. , individual stellar line of sight reddening corrections ) , which detracts from a uniform approach to the data reduction .
in addition there have been a number of recently discovered dwarf galaxies in the m81 group @xcite , which were not discovered in time to be included in the original angst sample .
the recent distance reassignment of ugc 4879 to the periphery of the lg @xcite meant this galaxy has also been excluded from the angst sample .
further , the angst sample is likely to be missing any faint dsphs which might be located in close proximity to m81 , i.e. , analogs to milky way satellites such as draco and ursa minor , as such galaxies have not yet been detected due to their inherent faintness . at a distance of m81 ( 3.6 mpc )
draco and ursa minor would have apparent blue luminosities of 18.98 and 20.13 , respectively . in comparison ,
sc22 ( @xmath14 @xmath15 17.72 ) has the faintest apparent magnitude in the angst sample . in this paper
, we initially divide the sample of dwarf galaxies according to morphological type , @xmath16 , @xcite resulting in 12 dwarf galaxies with @xmath16 @xmath17 0 ( dsphs ) , 5 with @xmath16 @xmath15 8 or 9 ( dwarf spirals ; dspirals ) , and 43 with @xmath16 @xmath15 10 ( dis ) .
for ease of direct comparison with well studied lg dwarf galaxies , we have adopted nomenclature consistent with @xcite , in which dwarf ellipiticals ( e.g. , ngc 147 , ngc 185 ) are found to be rare , and dsphs are more common . among the dis , there could be some ambiguity between bright dis and dspirals . for consistency , we defer to the @xmath16-type morphological classification scheme , but note that the distinction between a bright di and dspiral is not always clear .
morphological type @xmath16 @xmath15 10 further includes transition dwarf galaxies ( dtrans ) , galaxies with reduced recent star formation but high gas fractions ( e.g. , * ? ? ?
* ) , and tidal dwarf galaxy candidates ( dtidals ) , which appear to be condensing out of tidally disturbed gas .
we classify subtypes , i.e. , dtidal and dtrans , as follows : the three dtidals are holmberg ix , a0952 + 069 , and bk3n , and are all located in the m81 group . for dtrans , we adopt the definition of @xcite , namely , that a galaxy has detectable gas but very little or no h@xmath18 flux .
the final sample of dtrans was classified based on hi and h@xmath18 measurements in the literature @xcite .
we find 12 angst dwarf galaxies that satisfy the dtrans criteria : kk 230 , antlia , kkr 25 , kkh 98 , kdg 73 , eso294 - 10 , eso540 - 30 , eso540 - 32 , kdg 52 , eso410 - 005 , ddo 6 , and ugca 438 , leaving the final tally of true dis at 28 .
lccccccccccc kk230 & kkr3 & m31 & -8.49 & 1.3 & 0.04 & 10 & -1.0 & 12.0 & f606w , f814w & @xmath190.62 & 9771 + bk3n & & m81 & -9.23 & 4.0 & 0.25 & 10 & 1.0 & 18.0 & f475w , f814w & -0.32 & 10915 + antlia & & n3109 & -9.38 & 1.3 & 0.24 & 10 & -0.1 & 1.20 & f606w , f814w & @xmath191.75 & 10210 + kkr25 & & m31 & -9.94 & 1.9 & 0.03 & 10 & -0.7 & 2.50 & f606w , f814w & -0.18 & 11986 + fm1 & f6d1 & m82 & -10.16 & 3.4 & 0.24 & -3 & 1.8 & 5.00 & f606w , f814w & @xmath190.11 & 9884 + kkh86 & & m31 & -10.19 & 2.6 & 0.08 & 10 & -1.5 & 5.20 & f606w , f814w & -1.3 & 11986 + kkh98 & & m31 & -10.29 & 2.5 & 0.39 & 10 & -0.7 & 5.40 & f475w , f814w & @xmath190.46 & 10915 + bk5n central & & n3077 & -10.37 & 3.8 & 0.20 & -3 & 2.4 & 3.80 & f606w , f814w & -0.77 & 6964 + bk5n outer & & & & & & & & 3.80 & f606w , f814w & -0.70 & 5898 + sc22 & sc - de1 & n253 & -10.39 & 4.2 & 0.05 & -3 & 0.9 & 5.70 & f606w , f814w & @xmath190.30 & 10503 + kdg73 & & m81 & -10.75 & 3.7 & 0.06 & 10 & 1.3 & 12.0 & f475w , f814w & -0.38 & 10915 + ikn & & m81 & -10.84 & 3.7 & 0.18 & -3 & 2.7 & 0.58 & f606w , f814w & -0.82 & 9771 + e294 - 010 & & n55 & -10.86 & 1.9 & 0.02 & -3 & 1.0 & 4.60 & f606w , f814w & @xmath191.80 & 10503 + a0952 + 69 & & n3077 & -11.16 & 3.9 & 0.26 & 10 & 1.9 & 1.20 & f475w , f814w & -0.23 & 10915 + e540 - 032 & & n253 & -11.22 & 3.4 & 0.06 & -3 & 0.6 & 2.30 & f606w , f814w & @xmath190.10 & 10503 + kkh37 & & i342 & -11.26 & 3.4 & 0.23 & 10 & -0.3 & 3.70 & f475w , f814w & @xmath190.12 & 10915 + kdg2 & e540 - 030 & n253 & -11.29 & 3.4 & 0.07 & -1 & 0.4 & 2.70 & f606w , f814w & @xmath190.32 & 10503 + ua292 & cvni - dwa & n4214 & -11.36 & 3.1 & 0.05 & 10 & -0.4 & 5.10 & f475w , f814w & -0.37 & 10915 + kdg52 & m81-dwarf - a & m81 & -11.37 & 3.5 & 0.06 & 10 & 0.7 & 2.30 & f555w , f814w & @xmath190.37 & 10605 + kk77 & f12d1 & m81 & -11.42 & 3.5 & 0.44 & -3 & 2.0 & 0.83 & f606w , f814w & @xmath190.28 & 9884 + e410 - 005 & & n55 & -11.49 & 1.9 & 0.04 & -1 & 0.4 & 2.80 & f606w , f814w & @xmath191.70 & 10503 + hs117 & & m81 & -11.51 & 4.0 & 0.36 & 10 & 1.9 & 2.70 & f606w , f814w & -0.81 & 9771 + ddo113 & ua276 & n4214 & -11.61 & 2.9 & 0.06 & 10 & 1.6 & 1.80 & f475w , f814w & @xmath190.08 & 10915 + kdg63 & u5428,ddo71 & m81 & -11.71 & 3.5 & 0.30 & -3 & 1.80 & 1.40 & f606w , f814w & @xmath190.31 & 9884 + ddo44 & ua133 & n2403 & -11.89 & 3.2 & 0.13 & -3 & 1.7 & 0.59 & f475w , f814w & @xmath190.16 & 10915 + gr8 & u8091,ddo155 & m31 & -12.00 & 2.1 & 0.08 & 10 & -1.2 & 3.30 & f475w , f814w & @xmath190.82 & 10915 + e269 - 37 & & n4945 & -12.02 & 3.5 & 0.44 & -3 & 1.6 & 3.80 & f606w , f814w & -1.8 & 11986 + ddo78 & & m81 & -12.04 & 3.7 & 0.07 & -3 & 1.8 & 0.95 & f475w , f814w & -0.27 & 10915 + f8d1central & & m81 & -12.20 & 3.8 & 0.33 & -3 & 3.8 & 0.42 & f555w , f814w & -0.77 & 5898 + f8d1outer & & & & & & & & 0.42 & f606w , f814w & -0.90 & 5898 + u8833 & & n4736 & -12.31 & 3.1 & 0.04 & 10 & -1.4 & 5.00 & f606w , f814w & -0.23 & 10210 + e321 - 014 & & n5128 & -12.31 & 3.2 & 0.29 & 10 & -0.3 & 2.20 & f606w , f814w & -0.84 & 8601 + kdg64 & u5442 & m81 & -12.32 & 3.7 & 0.17 & -3 & 2.5 & 1.10 & f606w , f814w & @xmath190.57 & 11986 + ddo6 & ua15 & n253 & -12.40 & 3.3 & 0.05 & 10 & 0.5 & 3.00 & f475w , f814w & -0.02 & 10915 + ddo187 & u9128 & m31 & -12.43 & 2.3 & 0.07 & 10 & -1.3 & 1.60 & f606w , f814w & @xmath190.40 & 10210 + kdg61 & kk81 & m81 & -12.54 & 3.6 & 0.23 & -1 & 3.9 & 1.10 & f606w , f814w & @xmath190.33 & 9884 + u4483 & & m81 & -12.58 & 3.2 & 0.11 & 10 & 0.5 & 2.20 & f555w , f814w & -1.36 & 8769 + ua438 & e407 - 18 & n55 & -12.85 & 2.2 & 0.05 & 10 & -0.7 & 1.00 & f606w , f814w & -1.7 & 8192 + ddo181 & u8651 & m81 & -12.94 & 3.0 & 0.02 & 10 & -1.3 & 1.20 & f606w , f814w & -0.31 & 10210 + u8508 & izw60 & m81 & -12.95 & 2.6 & 0.05 & 10 & -1.0 & 2.10 & f475w , f814w & @xmath190.39 & 10915 + n3741 & u6572 & m81 & -13.01 & 3.0 & 0.07 & 10 & -0.8 & 1.60 & f475w , f814w & -0.24 & 10915 + ddo183 & u8760 & n4736 & -13.08 & 3.2 & 0.05 & 10 & -0.8 & 2.30 & f475w , f814w & -0.46 & 10915 + ddo53 & u4459 & m81 & -13.23 & 3.5 & 0.12 & 10 & 0.7 & 1.60 & f555w , f814w & -0.05 & 10605 + hoix & u5336,ddo66 & m81 & -13.31 & 3.7 & 0.24 & 10 & 3.3 & 0.71 & f555w , f814w & @xmath190.13 & 10605 + ddo99 & u6817 & n4214 & -13.37 & 2.6 & 0.08 & 10 & -0.5 & 0.58 & f606w , f814w & -0.95 & 10210 + sexa & ddo75 & mw & -13.71 & 1.3 & 0.14 & 10 & -0.6 & 0.06 & f555w , f814w & @xmath190.80 & 7496 + n4163 & u7199 & n4190 & -13.76 & 3.0 & 0.06 & 10 & 0.1 & 1.20 & f475w , f814w & -0.04 & 10915 + sexb & u5373 & mw & -13.88 & 1.4 & 0.10 & 10 & -0.7 & 0.10 & f606w , f814w & @xmath190.04 & 11986 + ddo125 & u7577 & n4214 & -14.04 & 2.5 & 0.06 & 10 & -0.9 & 0.18 & f606w , f814w & -1.3 & 11986 + e325 - 11 & & n5128 & -14.05 & 3.4 & 0.29 & 10 & 1.1 & 0.52 & f606w , f814w & -0.58 & 11986 + ddo190 & u9240 & m81 & -14.14 & 2.8 & 0.04 & 10 & -1.3 & 1.20 & f475w , f814w & -0.01 & 10915 + hoi & u5139,ddo63 & m81 & -14.26 & 3.8 & 0.15 & 10 & 1.5 & 0.34 & f555w , f814w & @xmath190.23 & 10605 + ddo82 & u5692 & m81 & -14.44 & 4.0 & 0.13 & 9 & 0.9 & 0.53 & f475w , f814w & -0.32 & 10915 + ddo165 & u8201 & n4236 & -15.09 & 4.6 & 0.08 & 10 & 0.0 & 0.50 & f555w , f814w & -0.7 & 10605 + n3109-deep & ddo236 & antlia & -15.18 & 1.3 & 0.20 & 9 & -0.1 & 0.05 & f606w , f814w & @xmath190.50 & 10915 + n3109-wide2 & & & & & & & & 0.05 & f606w , f814w & @xmath190.09 & 11307 + i5152 & & m31 & -15.55 & 2.1 & 0.08 & 10 & -1.1 & 0.10 & f606w , f814w & -1.38 & 11986 + n23661 & u3851 & n2403 & -15.85 & 3.2 & 0.11 & 10 & 1.0 & 0.39 & f555w , f814w & -0.16 & 10605 + n23662 & & & & & & & & 0.39 & f555w , f814w & -0.05 & 10605 + ho ii1 & u4305 & m81 & -16.57 & 3.4 & 0.10 & 10 & 0.6 & 0.12 & f555w , f814w & -0.32 & 10605 + ho ii2 & u4305 & & & & & & & 0.12 & f555w , f814w & -0.21 & 10605 + n4214 & u7278 & ddo113 & -17.07 & 2.9 & 0.07 & 10 & -0.7 & 0.03 & f606w , f814w & -0.48 & 11986 + i2574sgs & u5666,ddo81 & m81 & -17.17 & 4.0 & 0.11 & 9 & 0.9 & 0.05 & f555w , f814w & -0.26 & 9755 + i25741 & u5666,ddo81 & & & & & & & 0.05 & f555w , f814w & -0.55 & 10605 + i25742 & u5666,ddo81 & & & & & & & 0.05 & f555w , f814w & -0.15 & 10605 + e383 - 87 & & n5128 & -17.41 & 3.45 & 0.24 & 8 & -0.8 & 0.25 & f606w , f814w & -2.14 & 11986 + n55central & & n300 & -17.77 & 2.1 & 0.04 & 8 & 0.4 & 0.01 & f606w , f814w & -1.55 & 9765 + n55disk & & & & & & & & 0.01 & f606w , f814w & -0.42 & 9765 [ tab1 ] throughout this paper , we adopt the tidal index , @xmath20 @xcite , as a measure of a galaxy s isolation . @xmath20 describes the local mass density around galaxy @xmath21 as : @xmath22 + c , \phantom{1 } i = 1,2,\cdots , n\ ] ] where @xmath23 is the total mass of any neighboring galaxy separated from galaxy @xmath21 by a distance of @xmath24 .
the values of @xmath20 we use in this paper have been taken from @xcite .
negative values correspond to more isolated galaxies , and positive values represent typical group members ( see table [ tab1 ] ) .
hst observations of new angst targets were carried out in two phases due to the failure of acs in 2007 .
prior to the failure , we observed new targets using acs with wfpc2 in parallel mode .
galaxies observed post - acs failure were imaged with wfpc2 alone , as part of a ` supplemental ' hst program ( proposal ids 11307 and 11986 in table [ tab1 ] ; * ? ? ? * ) .
new acs observations used the f475w ( sdss @xmath25 ) and f814w ( @xmath26 ) filter combination , to optimize both photometric depth and temperature ( color ) baseline .
a third filter in f606w ( wide @xmath27 ) is also available for most galaxies .
the low throughput in the bluer filters of wfpc2 led us to only use f606w and f814w for wfpc2 observations .
as described in @xcite , both new and archival observations were processed uniformly beginning with image reduction via the standard hst pipeline . using the angst data reduction pipeline , we performed photometry on each image with hstphot @xcite , designed for wfpc2 images , and dolphot , running in its acs - optimized mode , for acs observations , providing for a uniform treatment of all data .
the resultant photometry for each data set was filtered to ensure the final photometric catalogs excluded non - stellar objects such as cosmic rays , hot pixels , and extended sources .
for the purposes of this paper , we considered a star well measured if it met the following criteria : a signal - to - noise ratio @xmath8 4 in both filters , a sharpness value such that ( @xmath28 @xmath19 @xmath29)@xmath30 @xmath17 0.075 , and a crowding parameter such that ( @xmath31 @xmath19 @xmath32 ) @xmath17 0.1 . to characterize observational uncertainties , we performed 500,000 artificial star tests on each image .
both the full and filtered ( i.e. , the ` gst ' files ) photometric catalogs , hst reference images , and cmds are publicly available on mast .
definitions and detailed descriptions of the filtering criteria and the observational strategies can be found in @xcite , @xcite , and @xcite . because of the variety of distances in the sample
, the acs / wfpc2 field of view does not subtend the same physical area for each galaxy . for certain comparisons among the sample ( e.g. , integrated stellar masses ) ,
it is important to account for these differences in coverage area . to that effect
, we employ a simple areal normalization factor based on each galaxy s apparent blue surface brightness . from the measurements listed in table 4 of @xcite
, we consider an effective elliptical area computed using the angular diameter and angular axial ratios at a blue surface brightness level of @xmath2 25 mag arcsec@xmath33 ( @xmath2 26.5 mag arcsec@xmath33 in the case of the faintest galaxies ) .
we then calculate the normalization factor , @xmath34 , for each galaxy by taking the ratio of the angular area subtended by the acs/ wfpc2 field of view to the angular area computed from @xcite .
this normalization has been specifically applied to the integrated stellar masses throughout this study .
the @xmath34 normalizations are typically @xmath8 1 ( i.e. , the hst field of view exceeds the area computed by @xcite ; see table [ tab1 ] ) , and the most and least covered galaxies are bk3n ( 18.0 ) and ngc 55 ( 0.02 ) , respectively . in most cases ,
a single hst field was sufficient to cover the main optical body of a galaxy , ensuring the sfhs are representative of the whole galaxy . however , several of the sample galaxies required multiple observations to cover a reasonable fraction of the optical body ( ngc 2366 , holmberg ii , ic 2574 , ngc 55 , bk5n , f8d1 , ngc 3109 ) . to derive the sfhs for each of these galaxies
, we first checked to see that the 50% completeness limits for each of the fields were similar ( i.e. , within @xmath350.2 mag ) .
this condition was met for all galaxies except ic 2574 , ngc 55 , and ngc 3109 , which have large angular sizes and wide ranges of surface brightnesses within each galaxy . for galaxies with similar completeness limits , the photometry and false stars were first combined , and then the sfh code was run on the combined data . for galaxies with overlapping fields ( ngc 2366 , holmberg ii ) we carefully removed duplicate stars before merging the photometry .
ic 2574 was a special case as it has three overlapping fields with significantly different completeness limits .
in this instance , we first removed the duplicate stars from overlapping fields , ran the sfhs on each field , and then combined the results .
ngc 55 and ngc 3109 have many new and archival hst observations taken over multiple visits . in each case
, we selected representative non - overlapping fields , one in the center of the galaxy and one in the disk .
the completeness functions were sufficiently different such that we did not combine photometry prior to computing the sfh . instead , we derived the sfhs for each field separately , and combined the resultant sfhs from each field .
the fractional coverage listed in table [ tab1 ] takes into account the combined areas for galaxies with multiple fields .
to ensure uniformity in the sfhs of the angst dwarf galaxies , we selected one sfh code @xcite and one set of stellar evolution models ( padova ; * ? ? ? * ) .
the sfh code of @xcite provides the user with robust controls over critical fixed input variables , e.g. , imf , binary fraction , time and cmd bin sizes , as well as the ability to search for the combination of metallicity , distance , and extinction values that produce a model cmd that best fits the observed cmd .
the models of @xcite combine updated asymptotic giant branch ( agb ) evolution tracks with the models of @xcite ( @xmath36 @xmath37 7 m@xmath38 ) and @xcite ( @xmath36 @xmath6 7 m@xmath38 ) . as with any models , there may be inevitable systematic biases associated with these particular choices ( e.g. , at some metallicities the padova model rgbs have red offsets from observations ; e.g. , * ? ? ?
however , these systematic effects will be shared by all galaxies in the sample , making for a robust relative comparison within the sample .
we discuss systematic uncertainties due to choice of isochrones in appendix [ systematics ] . here , we briefly summarize the technique of measuring a sfh based on the full methodology described in @xcite .
the user specifies an assumed imf and binary fraction , and allowable ranges in age , metallicity , distance , and extinction .
photometric errors and completeness are characterized by artificial star tests . from these inputs ,
many synthetic cmds are generated to span the desired age and metallicity range .
for this work , we have used synthetic cmds sampling stars with age and metallicity spreads of 0.1 and 0.1 dex , respectively .
these individual synthetic cmds are then linearly combined along with a model foreground cmd to produce a composite synthetic cmd .
the linear weights on the individual cmds are adjusted to obtain the best fit as measured by a poisson maximum likelihood statistic ; the weights corresponding to the best fit are the most probable sfh .
this process can be repeated at a variety of distance and extinction values to solve for these parameters as well .
monte carlo tests were used to estimate uncertainties due to both random and systematic sources . for
each monte carlo run , a poisson random noise generator was used to create a random sampling of the best - fit cmd .
this cmd was then processed identically to the original solution , with additive errors in @xmath39 and @xmath40 introduced when generating the model cmds for these solutions .
single shifts in @xmath39 and @xmath40 ) are used for each monte carlo draw , and the errors themselves were drawn from normal distributions with 1-@xmath41 values of 0.41 ( @xmath39 ) and 0.03 ( @xmath40 ) .
these distributions were designed to mimic the scatter in sfh uncertainties obtained by using multiple isochrone sets to fit the data , though we prefer to use this technique due to the small number of isochrone sets that fully cover the range of ages , metallicities , and evolutionary states required to adequately model our cmds .
we found that the uncertainties were stable after 50 monte carlo tests , and thus conducted 50 realizations for each galaxy .
this technique of estimating error on sfhs is described in greater detail in @xcite . to minimize systematic effects for comparisons among the sample , we selected consistent parameters for measuring sfhs of all galaxies .
all sfhs were measured using a single slope power law imf with a spectral index of @xmath121.30 over a mass range of 0.1 to 120 m@xmath38 , a binary fraction of 0.35 with a flat secondary mass distribution , 71 equally spaced logarithmic time bins ranging from @xmath42 @xmath15 6.610.15 , color and magnitude bins of 0.05 and 0.1 mag , and the padova stellar evolution models @xcite .
we note that the difference between our selected imf and a kroupa imf @xcite is negligible , as the angst cmds are limited to stellar masses @xmath4 0.8 m@xmath38 .
we designated the faint photometric limit to be equal to the 50% completeness limit in each filter ( see table [ tab1 ] ) as determined by @xmath2 500,000 artificial star tests run for each galaxy .
the sfh program was initially allowed to search for the best fit distance and extinction values without constraints .
initial values for the distances were taken from the trgb distances measured in @xcite , while foreground extinction values were taken from the galactic maps of @xcite .
we found no significant discrepancies between the assumed values and the best cmd fit distance and foreground extinction values , i.e. , all were consistent within error .
final solutions were computed using the trgb distances from @xcite and @xcite and foreground extinction values from @xcite .
we placed an additional constraint on the cmd fitting process , namely that the mean metallicity in each time bin must monotonically increase toward the present .
the deepest angst cmds do not reach the ancient ms turnoff , a requisite feature for completely breaking the age - metallicity degeneracy of the rgb only using broadband photometry ( e.g. , * ? ? ?
* ; * ? ? ?
as a result , sfhs derived from shallow cmds without a metallicity constraint can often have accompanying chemical evolution models that are unphysical ( e.g. , a drop in metallicity of several tenths of a dex over sub - gyr time scales ) .
a more robust analysis of the chemical enrichment of dwarf galaxies would need to include the ancient ms , measured gas phase abundances , and/or individual stellar spectra ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we therefore consider analysis of the metallicity evolution of the angst sample beyond the scope of this paper . as an example of a typical cmd fit made by the sfh code , we compare the observed and synthetic cmds of a representative angst dtrans , ddo 6 ( figure [ fit ] ) .
@xcite determined the trgb distance of ddo 6 to be [email protected] mpc while the foreground extinction maps of @xcite give values of a@xmath5 @xmath15 0.07 and a@xmath43 @xmath15 0.06 . allowing the sfh code to search for the best fit cmd , we find best fit values of d @xmath15 [email protected] mpc and a@xmath44 @xmath15 [email protected] , both in excellent agreement with the independently measured values .
examining the residual significance cmd , i.e. , the difference between the data and the model weighted by the variance ( panel ( d ) of figure [ fit ] ) , we see a good , although not perfect fit . notably , the area between the blue helium burning stars and young ms appears to be too cleanly separated in the model , which could be due to differential extinction affecting young stars in the observed cmd .
additionally , the model red helium burning stars are too blue compared to the data , likely due to uncertainties in the massive star models ( e.g. , * ? ? ?
* ) . however , even the most discrepant regions are fit within @xmath35 5@xmath41 , which are indicated by black or white points .
overall , the model cmd appears to be in good agreement with the observed cmd , indicating that we have measured a reliable sfh .
see @xcite for a full discussion of the quality measures of this cmd fitting technique . in general , uncertainties in the absolute sfhs
generally are somewhat anti - correlated between adjacent time bins , such that if the sfr is overestimated in one bin , it is underestimated in the adjacent bin . however , cumulative sfhs , i.e. , the stellar mass formed during or previous to each time bin normalized to the integrated final stellar mass , do not share this property , and thus present a more robust way of analyzing sfhs . in
the following sections , we consider sfhs plotted versus two different time binning schemes . for the individual galaxies ( figure [ sfh1 ] ) , we present the cumulative sfhs at the highest time resolution possible . errors in the cumulative sfhs presented at this time resolution indicate both the uncertainty in the fraction of total stellar mass formed prior to a given time , and the inherent time resolution for the sfh of that particular galaxy .
a coarser time binning scheme is used for the absolute sfhs of the individual galaxies as well as the mean absolute and cumulative sfhs across the angst sample . in appendix
[ appena ] , we determine that six broad time bins of of 0 - 1 , 1 - 2 , 2 - 3 , 3 - 6 , 6 - 10 , and 10 - 14 gyr provide an optimal balance between photometric depths and age leverage contained in the cmds of the angst dwarfs .
the broader time bins allow us to securely average the best fit sfhs from individual galaxies .
for the absolute sfhs of individual galaxies , broader time bins tend to minimize co - variance between adjacent time bins . thus applying a broader time binning scheme to the absolute sfhs of individual galaxies
provides for more secure measurements of the sfrs .
lccccccccc kk230 & 0.20 & 0.006 & 0.15 & 0.76 & 0.76 & 0.86 & 0.89 & 0.94 & 1.00 + bk3n & 1.40 & 0.03 & 0.37 & 0.41 & 0.41 & 0.95 & 0.95 & 0.95 & 1.00 + antlia & 0.38 & 0.10 & 1.30 & 0.16 & 0.41 & 0.83 & 0.94 & 0.95 & 1.00 + kkr25 & 0.19 & 0.03 & 0.19 & 0.58 & 0.59 & 0.90 & 0.90 & 0.97 & 1.00 + fm1 & 1.8 & 0.13 & 0.69 & 0.89 & 0.89 & 0.95 & 0.95 & 0.97 & 1.00 + kkh86 & 0.31 & 0.02 & 0.12 & 0.70 & 0.81 & 0.94 & 0.94 & 0.94 & 1.00 + kkh98 & 0.64 & 0.41 & 0.20 & 0.19 & 0.62 & 0.83 & 0.83 & 0.93 & 1.00 + bk5n & 2.0 & 0.10 & 0.46 & 0.93 & 0.93 & 0.97 & 0.97 & 0.97 & 1.00 + sc22 & 1.26 & 0.08 & 0.35 & 0.72 & 0.73 & 0.94 & 0.96 & 0.98 & 1.00 + kdg73 & 0.65 & 0.02 & 0.06 & 0.30 & 0.30 & 0.87 & 0.89 & 0.94 & 1.00 + ikn & 7.9 & 4.8 & 14.0 & 0.02 & 0.95 & 0.98 & 0.98 & 9.98 & 1.00 + e294 - 010 & 1.16 & 0.09 & 0.26 & 0.80 & 8.86 & 0.92 & 0.96 & 0.99 & 1.00 + a0952 + 69 & 0.46 & 0.13 & 0.29 & 0.40 & 0.40 & 0.68 & 0.68 & 0.68 & 1.00 + e540 - 032 & 0.24 & 0.36 & 0.76 & 0.86 & 0.86 & 0.91 & 0.91 & 0.98 & 1.00 + kkh37 & 1.64 & 0.16 & 0.31 & 0.45 & 0.48 & 0.87 & 0.87 & 0.95 & 1.00 + kdg2 & 0.34 & 0.04 & 0.09 & 0.02 & 0.02 & 0.60 & 0.68 & 0.88 & 1.00 + ua292 & 0.47 & 0.03 & 0.06 & 0.45 & 0.45 & 0.77 & 0.77 & 0.87 & 1.00 + kdg52 & 1.60 & 0.24 & 0.44 & 0.93 & 0.93 & 0.93 & 0.94 & 0.97 & 1.00 + kk77 & 4.6 & 1.94 & 3.4 & 0.47 & 0.74 & 0.96 & 0.96 & 0.98 & 1.00 + e410 - 005 & 1.42 & 0.18 & 0.30 & 0.62 & 0.79 & 0.87 & 0.89 & 0.98 & 1.00 + hs117 & 0.33 & 0.04 & 0.07 & 0.83 & 0.83 & 0.83 & 0.83 & 0.89 & 1.00 + ddo113 & 1.30 & 0.24 & 0.36 & 0.58 & 0.58 & 0.71 & 0.85 & 0.98 & 1.00 + kdg63 & 3.4 & 0.85 & 1.10 & 0.43 & 0.64 & 0.84 & 0.86 & 0.98 & 1.00 + ddo44 & 2.20 & 1.30 & 1.5 & 0.34 & 0.34 & 0.83 & 0.83 & 0.98 & 1.00 + gr8 & 0.90 & 0.10 & 0.1 & 0.59 & 0.69 & 0.88 & 0.88 & 0.95 & 1.00 + e269 - 37 & 2.00 & 0.20 & 0.2 & 0.94 & 0.94 & 0.96 & 0.96 & 0.97 & 1.00 + ddo78 & 4.75 & 1.77 & 1.7 & 0.56 & 0.56 & 0.79 & 0.79 & 0.99 & 1.00 + f8d1 & 8.35 & 3.8 & 3.2 & 0.65 & 0.65 & 0.73 & 0.89 & 0.99 & 1.00 + u8833 & 1.54 & 0.11 & 0.08 & 0.71 & 0.72 & 0.79 & 0.79 & 0.94 & 1.00 + e321 - 014 & 1.66 & 0.29 & 0.22 & 0.77 & 0.82 & 0.89 & 0.92 & 0.95 & 1.00 + kdg64 & 3.35 & 0.56 & 0.42 & 0.47 & 0.56 & 0.82 & 0.82 & 0.98 & 1.00 + ddo6 & 1.74 & 0.20 & 0.14 & 0.55 & 0.55 & 0.87 & 0.87 & 0.93 & 1.00 + ddo187 & 0.97 & 0.21 & 0.14 & 0.45 & 0.47 & 0.60 & 0.60 & 0.87 & 1.00 + kdg61 & 4.35 & 1.4 & 0.88 & 0.63 & 0.63 & 0.76 & 0.77 & 0.98 & 1.00 + u4483 & 1.57 & 0.27 & 0.16 & 0.00 & 0.91 & 0.93 & 0.93 & 0.96 & 1.00 + ua438 & 2.74 & 1.0 & 0.48 & 0.63 & 0.92 & 0.98 & 0.98 & 0.98 & 1.00 + ddo181 & 3.11 & 0.92 & 0.39 & 0.72 & 0.72 & 0.74 & 0.74 & 0.93 & 1.00 + u8508 & 2.33 & 0.39 & 0.16 & 0.58 & 0.58 & 0.73 & 0.73 & 0.93 & 1.00 + n3741 & 2.63 & 0.57 & 0.23 & 0.68 & 0.68 & 0.84 & 0.84 & 0.91 & 1.00 + ddo183 & 3.31 & 0.50 & 0.19 & 0.66 & 0.66 & 0.86 & 0.86 & 0.94 & 1.00 + ddo53 & 4.75 & 1.0 & 0.34 & 0.41 & 0.56 & 0.75 & 0.75 & 0.95 & 1.00 + ho ix & 4.04 & 1.15 & 0.35 & 0.23 & 0.72 & 0.83 & 0.83 & 0.89 & 1.00 + ddo99 & 3.12 & 4.0 & 1.20 & 0.75 & 0.93 & 0 .
94 & 0.94 & 0.95 & 1.00 + sexa & 0.59 & 3.53 & 0.75 & 0.30 & 0.46 & 0.65 & 0.67 & 0.84 & 1.00 + n4163 & 8.93 & 2.65 & 0.54 & 0.45 & 0.92 & 0.92 & 0.92 & 0.96 & 1.00 + sexb & 1.17 & 4.38 & 0.78 & 0.56 & 0.56 & 0.58 & 0.58 & 0.87 & 1.00 + ddo125 & 7.26 & 15.6 & 2.40 & 0.42 & 0.97 & 0.97 & 0.97 & 0.98 & 1.00 + e325 - 11 & 3.59 & 2.61 & 0.40 & 0.58 & 0.68 & 0.81 & 0.87 & 0.93 & 1.00 + ddo190 & 4.29 & 1.21 & 0.17 & 0.32 & 0.41 & 0.63 & 0.77 & 0.86 & 1.00 + ho i & 6.65 & 6.83 & 0.87 & 0.00 & 0.00 & 0.62 & 0.82 & 0.84 & 1.00 + ddo82 & 25.1 & 16.8 & 1.80 & 0.47 & 0.48 & 0.87 & 0.87 & 0.96 & 1.00 + ddo165 & 15.1 & 9.77 & 0.58 & 0.54 & 0.55 & 0.74 & 0.82 & 0.85 & 1.00 + n3109 & 2.64 & 20.2 & 1.10 & 0.79 & 0.79 & 0.79 & 0.79 & 0.91 & 1.00 + i5152 & 8.50 & 29.6 & 1.10 & 0.30 & 0.91 & 0.01 & 0.91 & 0.95 & 1.00 + n2366 & 28.5 & 26.8 & 0.79 & 0.67 & 0.67 & 0.74 & 0.74 & 0.89 & 1.00 + ho ii & 23.4 & 61.6 & 0.93 & 0.81 & 0.81 & 0.82 & 0.85 & 0.89 & 1.00 + n4214 & 7.02 & 82.3 & 0.79 & 0.71 & 0.95 & 0.95 & 0.95 & 0.97 & 1.00 + i2574 & 74.3 & 175 & 1.50 & 0.86 & 0.86 & 0.86 & 0.86 & 0.93 & 1.00 + e383 - 87 & 28.5 & 168 & 1.90 & 0.69 & 0.91 & 0.92 & 0.93 & 0.97 & 1.00 + n55 & 138.5 & 1210 & 6.10 & 0.63 & 0.67 & 0.91 & 0.91 & 0.97 & 1.00 [ tab2 ]
the angst dwarf galaxies exhibit a wide variety of complex sfhs . in figure [ sfh1 ]
we show the absolute sfhs , i.e. , sfr(@xmath45 ) , and the cumulative sfhs of the individual galaxies , sorted in order of increasing blue luminosity .
a cursory inspection of the 60 sfhs reveals that often galaxies with similar luminosities , morphologies , or chemical compositions , do not have consistent sfhs , confirming the complexity of dwarf galaxy sfhs previously found in studies of the lg ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the wide variety of sfhs underlines the importance of having a large sample ; conclusions based on the sfh of a single galaxy may not be be representative of the population as a whole .
this paper leverages the size of the angst sample to explore general trends seen in dwarf galaxies , e.g. , differences among the morphological types .
while the sfhs of individual galaxies can provide useful insight into a particular system s evolutionary state , we generally do not consider individual galaxies in this paper .
however , in table [ tab2 ] we provide data on the sfhs of individual galaxies considered in this sample , which may be useful for specific studies . lcccccccc dsph & [email protected] & 0.65@xmath46 & 0.71@xmath47 & 0.87@xmath48 & 0.89@xmath49 & 0.98@xmath50 & 1.00 + di & [email protected] & 0.54@xmath51 & 0.67@xmath52 & 0.80@xmath49 & 0.82@xmath53 & 0.92@xmath54 & 1.00 + dtrans & [email protected] & 0.53@xmath55 & 0.63@xmath56 & 0.86@xmath57 & 0.89@xmath57 & 0.96@xmath54 & 1.00 + dspiral & 74@xmath3535 & 0.59@xmath58 & 0.71@xmath59 & 0.88@xmath49 & 0.89@xmath49 & 0.95@xmath54 & 1.00 + dtidal & [email protected] & 0.38@xmath60 & 0.64@xmath61 & 0.83@xmath62 & 0.84@xmath63 & 0.85@xmath64 & 1.00 [ tab3 ] we first consider the unweighted mean specific sfhs ( i.e. , a sfh divided by the total integrated stellar mass formed ) of the angst sample grouped by morphological type ( figure [ sfh_avg ] ) .
comparison among the specific sfhs shows general consistency for times @xmath4 1 gyr ago , with the exception of dtidals ( note that dtidals will be discussed in greater detail in the context of recent sfhs in @xcite ) .
qualitatively , we find that a typical dwarf galaxy exhibits dominant ancient sf ( @xmath8 10 gyr ago ) , and lower levels of sf at intermediate times ( 110 gyr ago ) . the only consistent difference among the morphological types is within the last 1 gyr , where dsphs exhibit a significant drop in sfrs relative to the other types .
this suggests that many of the dsphs in the angst sample could have been gas - rich as recently as 1 gyr ago , implying that the process of gas loss can occur relatively quickly and at late times .
it is tempting to associate the sharp drop in the sfr of a typical dsph @xmath2 1 gyr ago with rapid gas loss .
however , broad uncertainties in the agb models ( e.g. , * ? ? ?
* ; * ? ? ?
* ) make the precise age for the drop in dsph sfrs and the degree of synchronization uncertain .
in addition , subsamples of the dsphs do appear to have dramatically lower sfrs at more intermediate ages .
cumulative sfhs allow us to readily compare galaxies of different masses . like the absolute sfhs ,
the cumulative sfhs plotted in figure [ cum1 ] show significant diversity within each morphological class , yet converge on mean values that are broadly consistent among the morphological types .
perhaps the most striking result is that the average dwarf galaxy formed @xmath4 50% of its total stellar mass by z @xmath2 2 ( 10 gyr ago ) and @xmath4 60% by z @xmath2 1 ( 7.6 gyr ago ) , independent of morphological type ( table [ tab3 ] ) . although dsphs appear to have formed a slightly larger percentage of total stellar mass than dis by z @xmath2 2 , additional consideration of systematic uncertainties , as discussed in appendix [ systematics ] ,
indicate that the angst data is not deep enough to make such fine distinctions .
more importantly , the amplitude of the difference is significantly smaller than typical models , which assume that dsphs are dominated by ancient stellar populations and dis are consistent with constant sf over their lifetimes ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . the mean cumulative sfhs hint at divergence among the morphological types within the last few gyr ( figure [ cum1 ] )
however , the exact characteristics of differences in the sfhs during this time period are ambiguous
. broad uncertainties in the agb star models ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , our modest time resolution , and systemic effects on the mean cumulative sfh ( see appendix [ systematics ] ) all affect our ability to quantitatively examine the details of intermediate age sf . we do identify secure measurements of differences within the most recent 1 gyr . at these times ,
luminous ms and core blue and red helium burning stars provide excellent age leverage ( e.g. , * ? ? ?
* ) , and the sfhs clearly illustrate differences among the morphological types . specifically , the typical dsph , di , dtrans , and dspiral formed @xmath2 2% , 8% , 4% , and 5% of their total stellar mass within the most recent 1 gyr .
the combined findings from intermediate and recent sfhs suggest that morphological differences and complete gas loss in dwarf galaxies can be relatively recent phenomena , at least within the local volume .
dsphs that are strictly old , i.e. , have no agb star populations , are known to exist in the lg ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but appear to only represent a minority of all known dsphs in the larger volume we analyze here .
this is at least in part due to the detection limit of the angst sample as discussed in [ selection ] . to further explore the differences among the morphological types
, we consider the fractional sfh , i.e. , the fraction of stellar mass formed in each time bin , as a function of the total normalized stellar mass , @xmath65 ( computed by integrating the sfh over time and applying the @xmath34 area normalization ) .
we show values for both the individual galaxies ( grey symbols ) and the mean values per morphological type , with error bars representative of the uncertainty in the mean in figure [ mbf_avg ] . the mean stellar masses and cumulative sfhs for each of the morphological types are listed in table [ tab3 ] .
the mean stellar masses show that dspirals are typically the most massive galaxies , dtrans are the least massive , and dis and dsphs have similar mean masses .
there are not strong links between patterns of sf , total stellar mass , and morphological type , confirming that sf processes among the different types of dwarf galaxies are not dramatically different .
we continue analysis involving the integrated stellar masses in [ gascomp ] .
galaxies in clusters , groups , and the field follow similar morphology density relationships .
namely , gas - poor galaxies , i.e. , ellipticals , are generally found to be less isolated than gas - rich galaxies , i.e. , spirals ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this same morphology
density relationship has been found for dwarf galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the change in the relative fraction of gas - rich and gas - poor galaxies with environment , provides a simple test for various models of dwarf galaxy evolution .
as illustrated in figures [ morphden ] and [ thf_avg ] , the angst dwarf galaxies clearly adhere to the morphology
density relationship , when using tidal index , @xmath20 , as a proxy for local density . typically , dis are significantly more isolated than dsphs , in spite of having similar mean total stellar masses .
dtrans , on average , have intermediate tidal indices between dis and dsphs ( see [ transition ] ) , and have a lower mean stellar mass than either .
dspirals are the most massive galaxies , and are typically located in regions of intermediate isolation .
these findings are in general agreement with earlier studies of lg dwarf galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . when combined with measured sfhs , the morphology
density relationship can be used to gauge the reliability of dwarf galaxy evolution models , particularly processes that induce gas loss .
scenarios favoring internal mechanisms ( i.e. , stellar feedback ) as the primary driver of gas loss ( e.g. , * ? ? ?
* ; * ? ? ?
* ) can reproduce a number of observed dwarf galaxy properties ( e.g. , surface brightness , rotation velocities , metallicities , etc ; * ? ? ?
however , such models are generally unable to account for the morphology
density relationship ( e.g. , * ? ? ?
* ) and often predict that gas - rich and gas - poor dwarf galaxies may have different patterns or efficiencies of sf ( e.g. , * ? ? ?
* ; * ? ? ?
in contrast , some models that additionally factor in external effects ( e.g. , gravitational interactions , ram pressure stripping ) have been able to reproduce a wide range of dwarf galaxy properties , including a canonical morphology density relationship ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ?
* ) . in the following sections , we discuss results from the angst sfh analysis within the context of physical processes that can affect the evolution of dwarf galaxies .
the relative masses of the gas and stellar components can provide clues to dwarf galaxy evolution .
for the angst sample , we consider the gas mass and baryonic gas fraction as functions of total stellar mass and tidal index .
the gas masses are based on the hi masses in @xcite , corrected by a factor of 1.4 to account for helium content , while the baryonic gas fractions , @xmath66/@xmath67 , is defined to be @xmath66/ ( @xmath66+@xmath65 ) , and does not account for the contribution due to warm / hot baryons .
the stellar masses , @xmath65 , are derived from integrating the sfhs over time and applying the areal normalization , @xmath34 .
we first consider the relationship between total gas mass and integrated stellar mass as shown in figure [ masses ] , where the dot - dashed line denotes @xmath66 @xmath15 @xmath65 . in this context ,
the most massive galaxies occupy the upper right portion of the plot , while lower mass dwarfs are located to the left .
galaxies without detectable gas are located in the lower portion of the plot , and have been placed at @xmath68 3 for convenience . to illustrate the range of stellar mass covered by the angst dsphs relative to the lg dsphs
, we have included select lg dsphs in grey ( placed using stellar mass - to - light values from @xcite ) .
figure [ masses ] provides a concise snapshot of the current evolutionary state of nearby dwarf galaxies . for large stellar masses
, this view illustrates the morphological ambiguity between dis and dspirals ; galaxies from both classes are gas - rich and can have stellar masses @xmath4 10@xmath69 m@xmath38 .
hi surface density maps of some dis even reveal hints of spiral gas structure in dis ( e.g. , * ? ? ?
in this same mass regime , however , there is a conspicuous absence of dsphs . in the angst sample ,
we not do find dsphs more massive than @xmath2 10@xmath69 m@xmath38 , which is in agreement with the lack of massive gas - poor galaxies in the lg ( e.g. , * ? ? ?
* ; * ? ? ?
dtrans are predominantly located at low stellar masses , yet have relatively large gas supplies .
these properties suggest that some dtrans may not be significantly different from low mass dis ; we return to this point in [ transition ] .
the baryonic gas fraction provides a more detailed view of the current evolutionary state of the angst dwarf galaxies ( figure [ bary ] ) .
the general trend for dwarf galaxies mirrors that of massive counterparts on the hubble sequence ( e.g. , * ? ? ?
* ) , namely that spiral galaxies typically have low gas fractions , while irregulars have higher gas fractions .
this perspective also reinforces the view that many dtrans could be low mass dis in between episodes of massive sf ( e.g. , * ? ? ?
* ) . there appear to be several outliers to the general trends in figure [ bary ] .
the two most conspicuous outliers , hs 117 and ddo 113 , may simply be morphological misclassifications .
hs 117 is classified as a di , but has a very low gas content , with an upper limit of @xmath70 @xmath2 10@xmath71 m@xmath38 @xcite and error bars consistent with zero .
interestingly , @xcite detect low levels of h@xmath18 in hs 117 , which reinforces the di classification , but based on the lack of hi , they deem this a dsph .
morphologically , inspection of the hst image further reveals that it appears to be superimposed on an hii region , leading to an erroneous di classification .
we therefore suggest hs 117 is best classified as a dsph or possibly a dtrans .
ddo 113 is likewise classified as a di with negligible gas content , but has a possible h@xmath18 detection @xcite .
inspection of the hst based cmd suggests that ddo 113 resembles a prototypical dsph , with a handful of relatively faint blue stars , suggesting it is also likely a dsph or a dtrans .
additional putative outliers include three dtrans ( eso294 - 010 , eso410 - 005 , eso540 - 032 ) and three dis ( ngc 4163 , ic 5152 and ddo 125 ) .
two of the dis ( ic 5152 and ddo 125 ) are isolated and relative massive , yet have moderate gas fractions .
the three dtrans have confirmed blue horizontal branch populations @xcite , which helps constrain their ancient sfhs .
however , none of these galaxies show any unique features in their sfhs that would explain their outlier status .
in contrast , ngc 4163 is a starburst dwarf galaxy , which may have consumed a significant amount of gas due to an intense burst of sf within the last @xmath2 1 gyr ( e.g. , * ? ? ? * ) .
examining the gas fraction as a function of isolation ( figure [ thbary ] ) , we see that isolated galaxies tend to have high gas fractions , while those in high density environments have low gas fractions .
interestingly , there appear to be no galaxies with _ any _ appreciable gas in dense environments ( @xmath20 @xmath4 1.5 ; hs 117 and ddo 113 have uncertain gas measurements as described above ) .
we also see some evidence for separation of dtrans into groups of isolated ( @xmath20 @xmath6 0 ) and moderate / high ( @xmath20 @xmath1 0 ) density environments , which we further discuss in [ transition ] .
how gas - rich galaxies _ completely _ lose their gas has long been an outstanding question in dwarf galaxy evolution ( e.g. , * ? ? ?
* and references therein ) . within the angst sample
, we can test the viability of various gas loss mechanisms using the observed properties of gas - poor dsphs .
the angst dsphs share several notable characteristics including : ( 1 ) little sf in the most recent 1 gyr compared to gas - rich dwarfs , ( 2 ) similar total stellar masses , ( 3 ) sfhs that are generally extended and indistinguishable from dis , and ( 4 ) are located exclusively in high density environments . in
what follows , we consider the impact of putative gas loss mechanisms on the evolution of a typical gas - rich dwarf galaxy ( e.g. , with both @xmath65 and @xmath66 @xmath2 10@xmath72 m@xmath38 ) in the context of figure [ masses ] .
the first mechanism for gas removal is consumption through sf .
this process will increase the stellar mass and decrease the gas mass of our prototypical di , moving it to the right on figure [ masses ] . though sf can consume large amounts of a galaxy s gas reservoir
, the gas densities will eventually become too low to continue to form stars ( e.g. , * ? ? ?
this suggests that if consumption was the only mechanism , we should observe trace amount of gas in dsphs , which is not typically the case . in the event that gas density was not a limiting factor , characteristically low sfrs and sf efficiencies in dwarf galaxies
( @xmath2 1% ; e.g. , * ? ? ?
* ) imply long gas consumption timescales . in this event
, gas supplies would typically not be exhausted for more than a hubble time , which suggests that dsphs would be quite rare .
in addition , gas consumption does not have a dependence on the environment external to a galaxy .
thus , it can not account for the observed morphology
density relationship .
although sf is a critical process to galaxy evolution , it can not be responsible for the complete removal of gas from dwarf galaxies .
the second possible mechanism for removing gas is stellar feedback .
mechanical energy due to stellar winds and supernovae provide an appealing explanation for gas removal from a galaxy via expulsion ( e.g. , * ? ? ?
for the prototypical gas - rich di , this process removes gas with minimal effect on stellar mass , moving the galaxy straight down in figure [ masses ] .
several feedback models can reproduce observed dwarf galaxy relationships between mass , luminosity , and rotational velocities ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
additionally , these models are generally able explain the observed mass metallicity relationship for dwarf galaxies ( e.g. , * ? ? ?
although these models hold promise , there are two significant challenges to gas removal in dwarf galaxies due to feedback .
first , a number of simulations have demonstrated that the energy due to stellar feedback is insufficient to completely expel cold gas ( e.g. , hi ) from the gravitational potential of a typical dwarf galaxy ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
second , stellar feedback alone can not explain the morphology density relationship .
similar to the rationale for gas consumption through sf , stellar feedback is not dependent on the environment surrounding a galaxy .
thus , while sf and stellar feedback are primary drivers of dwarf galaxy evolution , it is unlikely that either or both are able to account for the final transition from a gas - rich to a gas - poor state .
the third mechanism we consider is ram pressure stripping .
@xcite first proposed that dsphs were once gas - rich dwarfs whose gas supply has been removed by ram pressure stripping from the intergalactic medium ( igm ) .
this mechanism has negligible impact on stellar mass loss , thus moving our prototypical di straight down in figure [ masses ] .
satellite galaxies in higher density environments would likely encounter a denser igm , which would lead to more efficient gas loss .
this provides a feasible explanation for the observed morphology - density relationship . despite the appeal of this mechanism
, there is a distinct lack of observational evidence of systematic ram pressure stripping of lg dwarf galaxies .
for example , @xcite and @xcite find that ram pressure stripping appears to be a localized phenomena , and only evident for a small minority of lg dwarfs .
however , given that the lg is in a state of relatively passive evolution , it could be that signatures of ram pressure stripping have been erased in many satellites .
an additional challenge comes from the magnitude of ram pressure stripping .
@xcite demonstrate the ram pressure stripping is unlikely effective enough to completely remove the gas supply of a dwarf galaxy in the lg .
it appears that ram pressure stripping is not able to account for complete gas loss in dwarf galaxies .
we next consider tidal effects on gas removal from dwarf galaxies . the close passage of a dwarf galaxy to a massive companion has strong gravitational effects on the gas , stellar , and dark matter contents of the smaller galaxy . as shown in @xcite , the magnitude of the tidal force during an interaction does not appear to be enough to remove the gas content from a gas - rich di . instead
, @xcite advocate a more complex approach in which a combination of ram pressure stripping and stellar mass loss due to tidal effects provide a plausible model , ` tidal stirring ' , for the transformation of gas - rich dis into gas - poor dsphs .
in this scenario , the prototypical gas - rich di is able to lose both stellar and gas mass , allowing it to move down and left in figure [ masses ] , meaning a di with a high stellar mass could be transformed into a less massive dsph .
predictions from tidal stirring models provide qualitative explanations for a number of trends seen in the angst sample .
foremost , tidal stirring naturally produces a morphology density relationship .
galaxies in lower density environments have had few ( or no ) interactions with massive galaxies , and are able to maintain high gas fractions .
this is in general agreement with the trends seen in figure [ thbary ] . of the galaxies
that do interact with a massive companion , tidal stirring predicts that stellar and gas mass loss happens progressively over several gyr ( a single complete orbit in the lg is typically 1 - 2 gyr ; * ? ? ?
therefore , these galaxies spend the majority of their life cycling between sf and stellar feedback . by extension , this implies that many dsphs and dis should not have drastically different sfhs , in general agreement with the sfhs of angst dwarf galaxies .
while the general qualitative agreement between tidal stirring predictions and the angst sample is encouraging , a quantitative comparison between predictions and observed dwarf galaxy sfhs are needed to place precise constraints on the gas loss processes in dwarf galaxies . with little evidence of recent sf , yet detectable amounts of hi , dtrans may hold clues to the transformation of gas - rich to gas - poor galaxies . synthesizing several previous studies
, there are two favored scenarios for the origins of dtrans , namely that they are either in the last throes of sf or are observed during temporary lulls between episodes of massive star formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
dtrans in the angst sample have characteristics consistent with both scenarios . those with low gas fractions ( @xmath66/@xmath67 @xmath1 0.4 ; eso294 - 010 , es410 - 005 , and eso540 - 032 ) typically have positive tidal indices , indicating a higher likelihood of gas loss via interactions ( figure [ thbary ] ) .
these three galaxies may be examples of galaxies genuinely transitioning from gas - rich to gas - poor states .
in contrast , the most isolated dtrans ( kk 230 , kkh 98 , kkr 25 , and ua 438 ) typically have high gas fractions ( @xmath4 0.9 ) and stellar masses among the lowest in the angst sample ( figure [ bary ] ) . these dtrans have properties consistent with other low mass dis , and could simply be low mass dis lacking h@xmath18 , which can be attributed to a wide range of effects including temporary episodes between massive star formation ( e.g. , * ? ? ? * ) , a stochastically sampled or systematically varying imf ( e.g. , * ? ? ? * ; * ? ? ? * ) , and/or leakage of ionizing photons ( e.g. , * ? ? ? * ) .
the remaining dtrans ( antlia , kdg 2 , ddo 6 , kdg 52 , kdg 73 ) appear to be have masses and tidal indices close to the genuinely transitioning group , but gas fractions similar to the low mass dis .
it is possible that these galaxies may have had few interactions , resulting in less gas loss .
detailed multi - wavelength studies of these galaxies could provide insight into gas loss mechanisms in dwarf galaxies .
as the smallest and most pristine galaxies in the universe , dwarf galaxies occupy a critical , but poorly constrained role in cosmological models of galaxy formation and evolution ( e.g. , * ? ?
* ; * ? ? ?
because of their intrinsic faintness , dwarf galaxies have been difficult to directly detect at cosmologically significant redshifts ( e.g. , * ? ? ?
* ) . as an alternate approach to observing high redshift dwarf galaxies , we present a comparison between the mean cumulative sfhs of the angst sample dwarfs and the cosmic sfh , as derived from observed uv fluxes in high redshift galaxies ( e.g. , * ?
* ) . in the top panel of figure [ cum2 ]
, we see good agreement between the cosmic sfh ( grey shaded region ) and the mean cumulative sfhs of the angst dwarf galaxies at z @xmath2 2 ( 10 gyr ago ) . at more recent times ,
the cosmic sfh increases sharply , while the cumulative sfhs of dwarf galaxies increase at a slower rate .
this comparison suggests that the largest differences between sfrs in dwarf galaxies and massive galaxies studied at higher redshifts are at intermediate and recent epochs .
this finding is generally consistent with the expectations of galaxy ` downsizing ' , where high mass galaxies preferentially stop forming stars at higher redshifts ( e.g. , * ? ? ?
. the precise role of dwarf galaxies in theories of downsizing is highly uncertain , making this result difficult to interpret in the context of current models ( e.g. , * ? ? ?
we also caution that the physical significance of this finding is still ambiguous due to a several factors including uncertain agb star models and redshift dependent dust corrections applied to the cosmic sfh ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . a comparison of sfhs from the entire angst sample ( including the most massive galaxies ) and the cosmic sfh is presented in @xcite .
the cumulative sfhs can also provide insight into the validity of model sfhs often used to describe the evolution of dwarf galaxies . as an illustrative example
, we have selected a simple exponentially declining sf model ( sfr @xmath73 @xmath74 ; a @xmath75 model ) , and show the cumulative sfhs for a range of @xmath75 values ( grey shading ) in the lower panel of figure [ cum2 ] .
visual inspection suggests that the measured sfhs show significant deviations from the predicted smooth curve of single @xmath75 valued models .
a @xmath76@xmath30 test between the measured cumulative sfhs and exponentially declining sfh models ( with @xmath75 varying from 0.1 to 14.1 gyr ) confirms that any single value of @xmath75 does not accurately represent the data .
similarly , the mean measured sfhs also are not well matched with a constant sfh ( i.e. , @xmath75 @xmath77 @xmath78 gyr ) or a single ancient epoch of sf followed ( @xmath75 @xmath1 0.1 gyr ) by passive galaxy evolution .
instead more complex and multi - component models may better describe the lifetime sfhs of dwarf galaxies .
we uniformly analyzed sfhs of 60 dwarf galaxies in the nearby universe based on observations and data processing done as part of the angst program . while the sfhs of individual galaxies are quite diverse , we find that the mean sfhs of the different morphological types are generally indistinguishable outside the most recent @xmath2 1 gyr . on average ,
the typical dwarf galaxy formed @xmath2 50% of its stellar mass by z @xmath2 2 and 60% by z @xmath2 1 . among the morphological types , the sfhs hint at divergence within the past few gyr ,
although assigning a precise time to this phenomena is challenging due to uncertainties in agb star modeling and the modest time resolution afforded by the data .
the clearest differences between the morphological types can been seen in the most recent 1 gyr , where the typical dsph , di , dtrans and dspiral formed @xmath2 2% , 8% , 4% , and 5% of their total stellar mass .
the dwarf galaxies in the angst sample show a strong morphology density relationship .
this suggests that internal mechanisms , e.g. , stellar feedback , can not solely account for gas - loss in dwarf galaxies , as the corresponding models are unable to produce this observed relationship .
instead , we find qualitative consistency with the model of ` tidal stirring ' ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which can broadly explain the extended sfhs as well as the observed morphology density relationship .
a comparison with the cosmic sfh reveals general good agreement between the independently derived sfhs . the slower rate of sf for dwarf galaxies at intermediate epochs may be tentative evidence of galaxy downsizing
. however , broad uncertainties in extinction corrections and agb star models may be able to explain the offset .
further , the mean measured sfhs are inconsistent with single valued exponential models of sf ( i.e. , @xmath75 models ) , and may require more complex or multi - component models .
we also identify 12 dtrans in the sample , based on the literature definition of present day gas fraction and sf as measured by h@xmath18 ( e.g. , * ? ? ?
* ) . within this sample of dtrans
, we find that galaxies with high gas fractions are associated with more isolated galaxies , while those with lower gas fractions are less isolated .
this suggests that there are two mechanisms that can produce the observed dtrans characteristics : the isolated gas - rich galaxies are simply between episodes of sf due to the stochastic nature of sf in low mass galaxies , while the less isolated galaxies could be in the process of interacting with a more massive companion .
drw is grateful for support from the university of minnesota doctoral dissertation fellowship and penrose fellowship .
idk is partially supported by rfbr grant 10 - 02 - 00123 .
the authors would like to thank stephanie ct for fruitful discussions on the nature of transition dwarf galaxies and oleg gnedin for his insightful suggestions on measures of photometric quality .
this work is based on observations made with the nasa / esa hubble space telescope , obtained from the data archive at the space telescope science institute .
support for this work was provided by nasa through grants go-10915 , dd-11307 , and go-11986 from the space telescope science institute , which is operated by aura , inc .
, under nasa contract nas5 - 26555 .
this research has made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this research has made extensive use of nasa s astrophysics data system bibliographic services .
_ facility : _ widest - label aparicio , a. , & hidalgo , s. l. 2009 , , 138 , 558 baade , w. 1951 , publications of michigan observatory , 10 , 7 begum , a. , chengalur , j. n. , karachentsev , i. d. , sharina , m. e. , & kaisin , s. s. 2008 , , 386 , 1667 bertelli , g. , bressan , a. , chiosi , c. , fagotto , f. , & nasi , e. 1994 , , 106 , 275 binggeli , b. 1986 , star - forming dwarf galaxies and related objects , 53 binggeli , b. 1994 , panchromatic view of galaxies . their evolutionary puzzle , 173 blanton , m. r. , eisenstein , d. , hogg , d. w. , schlegel , d. j. , & brinkmann , j. 2005 , , 629 , 143 bouchard , a. , da costa , g. s. , & jerjen , h. 2009 , , 137 , 3038 chiboucas , k. , karachentsev , i. d. , & tully , r. b. 2009 , , 137 , 3009 cignoni , m. , & tosi , m. 2010 , advances in astronomy , 2010 cole , a. a. , tolstoy , e. , gallagher , j. s. , iii , & smecker - hane , t. a. 2005 , , 129 , 1465 cole , a. a. , et al .
2007 , , 659 , l17 ct , s. , freeman , k. c. , carignan , c. , & quinn , p. j. 1997 , , 114 , 1313 ct , s. , draginda , a. , skillman , e. d. , & miller , b. w. 2009 , , 138 , 1037 cowie , l. l. , songaila , a. , hu , e. m. , & cohen , j. g. 1996 , , 112 , 839 da costa , g. s. , rejkuba , m. , jerjen , h. , & grebel , e. k. 2010 , , 708 , l121 dalcanton , j. j. , et al .
2009 , , 183 , 67 de vaucouleurs g. , de vaucouleurs a. , corwin h.g . , buta r.j .
, paturel g. , & fouque p. , 1991
, third reference catalog of bright galaxies ( new york : springer ) dekel , a. , & silk , j. 1986 , , 303 , 39 dekel , a. , & woo , j. 2003 , , 344 , 1131 dellenbusch , k. e. , gallagher , j. s. , iii , knezek , p. m. , & noble , a. g. 2008 , , 135 , 326 dercole , a. , & brighenti , f. 1999 , , 309 , 941 dohm - palmer , r. c. , et al .
1998 , , 116 , 1227 dolphin , a. e. 2000 , , 112 , 1383 dolphin , a. e. 2002 , , 332 , 91 dolphin , a. e. , et al .
2003 , , 126 , 187 dolphin , a. e. , weisz , d. r. , skillman , e. d. , & holtzman , j. a. 2005 , arxiv : astro - ph/0506430 dolphin , a. e. 2011 , in prep .
dotter , a. , chaboyer , b. , jevremovi , d. , kostov , v. , baron , e. , & ferguson , j. w. 2008 , , 178 , 89 dressler , a. 1980 , , 236 , 351 einasto , j. , saar , e. , kaasik , a. , & chernin , a. d. 1974 , , 252 , 111 eldridge , j. j 2010 , arxiv:1008.1352 ferguson , a. m. n. , wyse , r. f. g. , gallagher , j. s. , iii , & hunter , d. a. 1996 , , 111 , 2265 ford , h. c. , et al . 1998 , , 3356 , 234 gallart , c. , aparicio , a. , bertelli , g. , & chiosi , c. 1996 , , 112 , 1950 gallart , c. , zoccali , m. , & aparicio , a. 2005 , , 43 , 387 geha , m. , blanton , m. r. , masjedi , m. , & west , a. a. 2006 , , 653 , 240 gilbert , k. , et al .
2011 , in prep .
girardi , l. , bertelli , g. , bressan , a. , chiosi , c. , groenewegen , m. a. t. , marigo , p. , salasnich , b. , & weiss , a. 2002 , , 391 , 19 girardi , l. , et al .
2010 , , 724 , 1030 gogarten , s. m. , et al .
2010 , , 712 , 858 grebel , e. k. , gallagher , j. s. , iii , & harbeck , d. 2003 , , 125 , 1926 harris , j. , & zaritsky , d. 2001 , , 136 , 25 hernandez , x. , valls - gabaud , d. , & gilmore , g. 1999 , , 304 , 705 hodge , p. w. 1971 , , 9 , 35 hodge , p. 1989
, , 27 , 139 holtzman , j. a. , burrows , c. j. , casertano , s. , hester , j. j. , trauger , j. t. , watson , a. m. , & worthey , g. 1995 , , 107 , 1065 holtzman , j. a. , et al .
1999 , , 118 , 2262 huchtmeier , w. k. , karachentsev , i. d. , karachentseva , v. e. , & ehle , m. 2000 , , 141 , 469 hunter , d. a. , brinks , e. , elmegreen , b. , rupen , m. , simpson , c. , walter , f. , westpfahl , d. , & young , l. 2007 , bulletin of the american astronomical society , 38 , 895 ikuta , c. , & arimoto , n. 2002 , , 391 , 55 jacobs , b. a. , tully , r. b. , rizzi , l. , karachentsev , i. d. , chiboucas , k. , & held , e. v. 2010 , arxiv:1010.1325 jarosik , n. , et al .
2010 , arxiv:1001.4744 karachentsev , i. d. , et al .
2003 , , 404 , 93 karachentsev , i. d. , karachentseva , v. e. , huchtmeier , w. k. , & makarov , d. i. 2004 , , 127 , 2031 karachentsev , i. d. , & kaisin , s. s. 2007 , , 133 , 1883 kazantzidis , s. , lokas , e. l. , callegari , s. , mayer , l. , & moustakas , l. a. 2010 , arxiv:1009.2499 kennicutt , r. c. , jr .
1989 , , 344 , 685 kirby , e. n. , cohen , j. g. , smith , g. h. , majewski , s. r. , sohn , s. t. , & guhathakurta , p. 2010 , arxiv:1011.5221 koo , d. c. , & kron , r. g. 1992 , , 30 , 613 kopylov , a. i. , tikhonov , n. a. , fabrika , s. , drozdovsky , i. , & valeev , a. f. 2008 , , 387 , l45 klypin , a. , kravtsov , a. v. , valenzuela , o. , & prada , f. 1999 , , 522 , 82 kniazev , a. y. , brosch , n. , hoffman , g. l. , grebel , e. k. , zucker , d. b. , & pustilnik , s. a. 2009 , , 400 , 2054 kroupa , p. 2001
, , 322 , 231 lee , h. , skillman , e. d. , cannon , j. m. , jackson , d. c. , gehrz , r. d. , polomski , e. f. , & woodward , c. e. 2006 , , 647 , 970 leroy , a. k. , walter , f. , brinks , e. , bigiel , f. , de blok , w. j. g. , madore , b. , & thornley , m. d. 2008 , , 136 , 2782 lin , d. n. c. , & faber , s. m. 1983 , , 266 , l21 mac low , m .- m .
, & ferrara , a. 1999 , , 513 , 142 marcolini , a. , dercole , a. , brighenti , f. , & recchi , s. 2006 , , 371 , 643 marigo , p. , et al .
2010 , iau symposium , 262 , 36 marigo , p. , girardi , l. , bressan , a. , groenewegen , m. a. t. , silva , l. , & granato , g. l. 2008 , , 482 , 88 mateo , m. l. 1998 , , 36 , 435 mathews , w. g. , chomiuk , l. , brighenti , f. , & buote , d. a. 2004 , , 616 , 745 mayer , l. , governato , f. , colpi , m. , moore , b. , quinn , t. , wadsley , j. , stadel , j. , & lake , g. 2001 , , 547 , l123 mayer , l. , governato , f. , colpi , m. , moore , b. , quinn , t. , wadsley , j. , stadel , j. , & lake , g. 2001 , , 559 , 754 mayer , l. , mastropietro , c. , wadsley , j. , stadel , j. , & moore , b. 2006 , , 369 , 1021 mayer , l. 2010 , advances in astronomy , 2010 marzke , r. o. , & da costa , l. n. 1997 , , 113 , 185 mcconnachie , a. w. , venn , k. a. , irwin , m. j. , young , l. m. , & geehan , j. j. 2007 , , 671 , l33 mcquinn , k. b. w. , skillman , e. d. , cannon , j. m. , dalcanton , j. j. , dolphin , a. , stark , d. , & weisz , d. 2009 , , 695 , 561 mcquinn , k. b. w. , et al .
2010 , , 721 , 297 mighell , k. j. 1997 , , 114 , 1458 miller , b. w. , dolphin , a. e. , lee , m. g. , kim , s. c. , & hodge , p. 2001
, , 562 , 713 monelli , m. , et al .
2010 , , 720 , 1225 monelli , m. , et al .
2010 , , 722 , 1864 moore , b. , ghigna , s. , governato , f. , lake , g. , quinn , t. , stadel , j. , & tozzi , p. 1999
, , 524 , l19 mould , j. r. , cannon , r. d. , frogel , j. a. , & aaronson , m. 1982 , , 254 , 500 mouri , h. , & taniguchi , y. 2006 , , 459 , 371 oemler , a. , jr .
1974 , , 194 , 1 orban , c. , gnedin , o. y. , weisz , d. r. , skillman , e. d. , dolphin , a. e. , & holtzman , j. a. 2008 , , 686 , 1030 puche , d. , westpfahl , d. , brinks , e. , & roy , j .-
1992 , , 103 , 1841 reddy , n. a. , steidel , c. c. , pettini , m. , adelberger , k. l. , shapley , a. e. , erb , d. k. , & dickinson , m. 2008 , , 175 , 48 reddy , n. a. , erb , d. k. , pettini , m. , steidel , c. c. , & shapley , a. e. 2010 , , 712 , 1070 revaz , y. , et al .
2009 , , 501 , 189 ricotti , m. , & gnedin , n. y. 2005 , , 629 , 259 roberts , m. s. , & haynes , m. p. 1994
, , 32 , 115 & mountain , m. 2008 , , 676 , l113 sandage , a. , & hoffman , g. l. 1991 , , 379 , l45 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 skillman , e. d. , & bender , r. 1995 , revista mexicana de astronomia y astrofisica conference series , 3 , 25 skillman , e. d. 1996 , the minnesota lectures on extragalactic neutral hydrogen , 106 , 208 skillman , e. d. , & gallart , c. 2002 , observed hr diagrams and stellar evolution , 274 , 535 skillman , e. d. , ct , s. , & miller , b. w. 2003 , , 125 , 593 skillman , e. d. , tolstoy , e. , cole , a. a. , dolphin , a. e. , saha , a. , gallagher , j. s. , dohm - palmer , r. c. , & mateo , m. 2003 , , 596 , 253 tosi , m. , greggio , l. , & focardi , p. 1989
, , 156 , 295 tolstoy , e. , & saha , a. 1996 , , 462 , 672 tolstoy , e. , hill , v. , & tosi , m. 2009 , , 47 , 371 van den bergh , s. 2000 , the galaxies of the local group , by sidney van den bergh
. published by cambridge ,
uk : cambridge university press , 2000 cambridge astrophysics series series , vol no : 35 , isbn : 0521651816 .
weidner , c. , & kroupa , p. 2005
, , 625 , 754 weisz , d. r. , skillman , e. d. , cannon , j. m. , dolphin , a. e. , kennicutt , r. c. , jr . ,
lee , j. , & walter , f. 2008 , , 689 , 160 weisz , d. r. , et al .
, 2011b , in prep .
weisz , d. r. , et al .
, 2011c , in prep .
williams , b. f. , et al .
2010 , , 709 , 135 williams , b. f. , et al . , 2011 , in prep .
woo , j. , courteau , s. , & dekel , a. 2008 , , 390 , 1453 young , l. m. , skillman , e. d. , weisz , d. r. , & dolphin , a. e. 2007 , , 659 , 331 yuk , i .- s . , & lee , m. g. 2007 , , 668 , 876 zwicky , f. 1957 , berlin : springer , 1957
meaningful comparison of sfhs across the angst sample requires a uniform time binning scheme .
this necessitates balancing the wealth of information about the ancient and intermediate epochs from the deepest cmds ( e.g. , antlia ; 50% completeness of @xmath79 @xmath2 @xmath191.75 ) with the course leverage of the shallower cmds ( e.g. , ugc 4483 ; 50% completeness of @xmath79 @xmath2 @xmath121.7 ; see table [ tab2 ] ) . as a compromise between the two extremes
, we fixed a sample - wide time binning scheme using a galaxy with a 50% completeness value near the median of the sample ( ddo 6 ; @xmath79 @xmath2 0 ) . to test for the optimal time resolution , we constructed cmds of single age stellar populations , convolved the simulated photometry with the artificial stars and photometric limits of ddo 6 , and ran the sfh recovery program on each of these simulated cmds .
this process was conducted on single age cmds of 0.5 , 1.5 , 2.5 , 4.5 , 8 , and 12 gyr in age .
the purpose of this exercise is to provide a simple test for the robustness of the final time bins .
while coarser time bins encompass a larger fraction of the input sf , there are generally fewer of them , which provides less information about pattens of sf over time . in this test
, we deemed a good recovery as one in which the input sfh was within the uncertainties of the recovered sfh . in figure
[ frac_sf ] , we show the cumulative sfhs , where the input sf is indicated by the dashed magenta line and the recovered sfh by the solid black line .
uncertainties were computed using 50 monte carlo realizations as described in
[ sfhs ] , including identical shifts in @xmath39 and @xmath80 .
initial tests suggested that six broad time bins of 0 - 1 , 1 - 2 , 2 - 3 , 3 - 6 , 6 - 10 , 10 - 14 gyr would provide suitable time resolution across the angst sample .
as shown in figure [ frac_sf ] , the general agreement between the input and recovered sfhs is good and falls within the uncertainties .
clearly , the best recovery is seen in the 0 - 1 gyr bin , where the young ms and massive core and blue helium burning stars provide secure leverage on recent sf .
the oldest bin also shows similarly good results .
the results in the intermediate bins are generally reasonable , although not as reliable as the youngest and oldest bins .
much of this can be attributed to uncertainties in agb star models ( e.g. , * ? ? ?
* ; * ? ? ?
although , we used the same models to generate and recover the single age populations , photometric depth , systematic stellar model uncertainties , and blurring of distinct features due to observational effects all contribute to uncertainties in the recovered sfh ( see appendix [ systematics ] ) .
agb populations of different ages often have similar colors and magnitudes on optical cmds ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , thus it can be difficult to confidently determine the epoch of sf at intermediate ages ( 110 gyr ago ) .
this limitation is clearly illustrated in figure [ frac_sf ] , which shows broad uncertainties on the recovered sfhs during intermediate times .
however , we still find that the input sfhs are generally within the recovered sfh error bars .
additionally , @xmath4 50% of the sf is recovered in the correct time interval . although it would be possible to combine bins at intermediate times to increase the accuracy of the recovered sfrs
, we believe that this scheme will permit future comparisons with sfhs derived using improved agb star models ( e.g. , * ? ? ?
for this time binning scheme , we find that reported uncertainties in the sfrs typically decrease for cmds deeper than ddo 6 , and increase for shallower cmds , in agreement with expectations .
we thus conclude that this scheme is adequate for comparison of sfhs across the angst sample .
photometric depth is an important consideration when interpreting the accuracy of a measured sfh . varying photometric depths impact both the number of stars on a cmd and
determine the presence of age sensitive cmd features ( e.g. , ancient ms turn - off , horizontal branch ) .
intuitively , a cmd with a brighter photometric limit has less information available than a deeper cmd , and the derived sfh is thus more uncertain .
however , quantifying the precise impact of these uncertainties , such as the amplitude and ages affected , is challenging as varying photometric depths can amplify effects of uncertainties in the stellar models used to measure sfhs . in this section
, we explore the effects of photometric depth on the accuracy of sfh recovery .
we demonstrate the effects in two regimes : one where the only variable is photometric depth , and the other where we vary both photometric depth and stellar evolution models .
we first analyze the accuracy of recovered sfhs as a function of _ only _ photometric depth .
that is , we construct synthetic cmds at select photometric depths , then attempt to recover the input sfh using identical parameters ( e.g. , imf , binary fraction , filter combination ) and , in this case , the same stellar evolution models .
more concretely , we constructed cmds at six different photometric depths ( @xmath81 @xmath194 , @xmath192 , @xmath191 , 0 , @xmath121 , @xmath122 ) , using a constant sfh ( @xmath827.4 - 10.15 , with a time resolution of @xmath2 0.1 dex ) , a fixed metallicity , and the basti stellar evolution models .
each cmd was populated with @xmath2 10@xmath83 stars in order to minimize the contribution of random uncertainties due to the number of stars used to measure the sfr in each time bin ; for sfhs measured from observed cmds , the uncertainties for poisson sampling of the cmd are already well characterized by our monte carlo tests .
we then recovered the sfh of each cmd using the basti stellar evolution models , to ensure that the only variable being tested is photometric depth . in figure [ basbas_sf ] , we compare the input ( black lines ) and recovered ( colored lines ) cumulative sfhs at each photometric depth .
overall , the recovered cumulative sfhs are in excellent agreement with the input sfhs at all photometric depths .
the maximum deviation between the input and recovered sfhs at any photometric depth is @xmath2 4% , which is consistent with the expected poisson precision of 1/@xmath84 , where @xmath85 is the number of stars used to measure the sfr in a given time bin .
this exercise demonstrates that if all the underlying stellar models are known exactly , then the accuracy of the recovered sfh only depends on the number of stars in the cmd , and not the photometric depth .
the same results are found when using different stellar evolution models , e.g. , padova , dartmouth @xcite , to conduct this exercise .
systematic uncertainties are introduced into sfh measurements by uncertainties in the selected stellar models .
the luminosity , color , and number density of specific cmd features provide leverage on the sfr at different epochs .
however , stellar models are not always self - consistent when trying to model multiple observed features ( e.g. , reproducing colors consistent with observations for both the horizontal branch and rgb ; e.g. , * ? ? ?
the effect on a measured sfh is that the sfr may be systematically shifted into a particular time bin , depending on the stellar model used and photometric depth of the cmd ( i.e. , which particular age sensitive cmd features are available ) .
we refer the reader to @xcite for a comprehensive review on the adequacy of stellar evolution models for reproducing cmd features . to test for systematic effects on
measured sfhs , one ideally wants to test for differences between the stellar models and ` truth ' , that is , stellar population characteristics based on the exact physics governing stellar evolution in nature .
however , current stellar evolution models represent the best physical descriptions we have of nature . by measuring how different stellar models impact a measured sfh
, we can get a sense of the amplitude of systematic uncertainties for sfh recovery . to examine the magnitude of systematic effects
, we again construct six cmds at selected photometric depths each containing @xmath2 10@xmath83 stars , assuming a constant sfh , fixed metallicity , and the basti stellar evolution models .
however , the sfhs are now recovered with the padova stellar models .
thus , in this case the differences between the input and recovered sfhs are indicative of the systematic effects introduced by choice in stellar model . in figure [ padbas_sf ]
, we see the differences in the input ( black lines ) and recovered ( colored lines ) mean cumulative sfhs are more substantial than when identical models were used . for the deepest cmd considered ( @xmath86@xmath194 ) , the agreement between
the input and recovered cmd is excellent . here
, the ancient ms turnoff provides a reliable constraint on the ancient sfh , meaning the systematics in progressively younger time bins are also known more precisely . for shallower cmds that do not include the ancient ms turn - off , we see significant deviations between the recovered and input sfhs .
these difference arise primarily in the treatment of such features of the horizontal branch , red clump , and rgb ( see @xcite for a more detailed discussion ) .
we quantify the magnitude of the discrepancies between the padova and basti models by examining the absolute value of the differences in input and recovered cumulative sfhs ( figure [ diff_sf ] ) .
of interest are the general order of magnitude variations , and not a specific point by point analysis . in the case of the shallowest cmds ( @xmath86@xmath122 , @xmath121 , 0 ) , the ancient sf
is preferentially recovered at intermediate ages , with a typical magnitude in the difference of input and recovered sfh of @xmath2 20% .
for the deeper cmds ( @xmath86@xmath191 , @xmath192 ) , sf at the oldest times tends to be over - estimated by up to @xmath2 40% . at all photometric depths , the systematic effects within the last @xmath2 2 gyr are @xmath1 10% .
the exercises we have conducted demonstrate the importance of systematic uncertainties in sfh recovery . we caution that our results are actually measuring
the systematic differences between the padova and basti models , and only serve a a proxy for the difference between a given model and ` truth ' , i.e. , observed cmds .
that being said , the methodology of this type of analysis provides a framework for exploring the effects of systematic uncertainties .
the primary limitation to this type of analysis is the number of models available that span the entire age / metallicity range needed to derive sfhs in nearby galaxies . as the number of models that sample a broader parameter space increases
, it will be possible to gain more leverage on the effects of systematic uncertainties . | we present uniformly measured star formation histories ( sfhs ) of 60 nearby ( @xmath0 @xmath1 4 mpc ) dwarf galaxies based on color - magnitude diagrams of resolved stellar populations from images taken with hubble space telescope and analyzed as part of the acs nearby galaxy survey treasury program ( angst ) .
this volume - limited sample contains 12 dsph / de , 5 dwarf spiral , 28 dirr , 12 dsph / dirr ( transition ) , and 3 tidal dwarf galaxies .
the sample spans a range of @xmath2 10 in @xmath3 and covers a wide range of environments , from highly interacting to truly isolated . from the best fit sfhs we find three significant results : ( 1 ) the average dwarf galaxy formed @xmath4 50% of its stars by z @xmath2 2 and 60% of its stars by z @xmath2 1 , regardless of current morphological type ; ( 2 ) the mean sfhs of dis , dtrans , and dsphs are similar over most of cosmic time , and only begin to diverge a few gyr ago , with the clearest differences between the three appearing during the most recent 1 gyr ; and ( 3 ) the sfhs are complex and the mean values are inconsistent with simple sfh models , e.g. , single bursts , constant sfrs , or smooth , exponentially declining sfrs .
the mean sfhs are in general agreement with the cosmic sfh , although we observe offsets at intermediate times ( z @xmath2 1 ) that could be evidence that low mass systems experienced delayed star formation relative to more massive galaxies .
the sample shows a strong density - morphology relationship , such that the dsphs in the sample are less isolated than dis .
we find that the transition from a gas - rich to gas - poor galaxy can not be solely due to internal mechanisms such as stellar feedback , and instead is likely the result of external mechanisms , e.g. , ram pressure and tidal stripping and tidal forces .
the average transition dwarf galaxy is slightly less isolated and less gas - rich than the typical dwarf irregular .
further , the transition dwarfs can be divided into two groups : interacting and gas - poor or isolated and gas - rich , suggesting two possible evolutionary pathways . |
the characterization of entanglement and the elucidation of its role in quantum computation remain formidable challenges , in spite of the conspicuous presence of this concept in quantum physics since the fundamental and instigating papers published by einstein , podolski , and rosen , as well as schrdinger , in 1935 @xcite .
motivation for the understanding of this subtle concept is stimulated not only by its fundamental character , but also by its applications in quantum information @xcite and , much specially , by the perspective that entanglement could be the key ingredient in the increased efficiency of quantum computing compared to classical computation for certain quantum algorithms @xcite . in an entangled state
, classical and quantum correlations may coexist .
indeed , for instance , in the singlet state @xmath0 there is a perfect classical correlation between the first and the second qubit , namely if the first is in state @xmath1 the second is in state @xmath2 , and vice - versa .
several measures of entanglement have been proposed , like the concurrence @xcite and the negativity @xcite . also , different criteria for the existence of classical correlations have been proposed @xcite . in ref .
@xcite , the existence of genuine classical correlations was associated to the non - vanishing of @xmath3-party correlation functions involving local observables of the system .
based on this definition , it was shown that it is possible to have multiparty entangled states with no genuine classical correlations , as long as the number of parties is larger than two @xcite .
more recently , it has become clear that entanglement does not exhaust the realm of quantum correlations .
indeed , separable states can exhibit quantum correlations , which seem to play a role in the explanation of the power of some schemes of quantum computation @xcite .
several quantifiers have been proposed for these quantum correlations , starting with the work by ollivier and zurek @xcite , who introduced the quantum discord " .
intuitively , this measure quantifies , in a bipartite system , the minimum change in the state of the system and on the information of one of its parts induced by a measurement on the other part . for a state with zero quantum discord , it is possible to measure any of its parts without changing the state of the system .
an example would be the state @xmath4 for which there exist one - dimensional complete projective measurements on both the first and second subsystem namely , measurements on the basis @xmath5 for the first qubit and @xmath6 for the second qubit that do not perturb the overall quantum state , nor the state of each part .
the quantum discord of this state is zero : the state has only classical correlations . on the other hand , for the state @xmath7
it is clear that any projective measurement performed on the second qubit disturbs the state of the first qubit . in this case , quantum discord is different from zero .
the same is true for the maximally - entangled singlet state , which is transformed into a statistical mixture upon measurement of any of the two qubits .
an interesting implication of this concept is the demonstration that vanishing quantum discord is necessary and sufficient for completely positive maps @xcite , that is , it is not necessary to assume , in the usual master equation description , that the initial state of system+environment is a product state . it could be a more general separable state , as long as the quantum discord vanishes .
quantum discord has been calculated for several families of quantum states and compared with the entanglement @xcite .
modified versions have been proposed , with different physical meanings @xcite .
other measures of quantum correlations , involving non - orthogonal measurements ( positive operator - valued measurements - povm s ) and measurements on both qubits , have been introduced in the literature @xcite , and have been studied and compared with quantum discord in different situations : the dqc1 model of mixed - state quantum computation @xcite , and within the context of accelerated frames @xcite .
we present in sec .
[ sec : quantum and classical correlations ] these other quantifiers , as well as a precise definition of quantum discord .
the existence of quantum correlations in the absence of entanglement is another subtle trait of quantum mechanics , still to be fully understood . in this paper , we show that when two qubits interact with each other through a common environment and also through dipole forces , there is a peculiar dynamical behavior intertwining quantum correlations , classical correlations , and entanglement . for two qubits
initially excited , quantum correlations show up as precursors of entanglement , growing up and then shrinking as entanglement belatedly appears . at this moment , classical correlations , even though non - vanishing , become practically negligible .
the dynamics of quantum correlations is studied in terms of two distinct quantifiers , the above - mentioned quantum discord and the measured - induced disturbance ( mid ) , introduced by luo @xcite .
in particular , we show that the mid coincides with the quantum discord for part of the evolution .
the dynamics of multiparty entangled systems interacting with independent environments that act on each of its parts is quite different from the time - dependent behavior of each individual component of the system . quite generally , the decay of entanglement is non - exponential , and it may lead to separability at finite times , before each part reaches its final state @xcite .
the finite - time disappearance of entanglement , sometimes called `` sudden - death of entanglement '' @xcite was experimentally demonstrated by almeida et al .
the dynamics of quantum correlations other then entanglement has also been analyzed , both theoretically @xcite and experimentally @xcite .
as the two qubits get closer , at a distance comparable to the radiated wavelenght , a different physical situation arises .
the model of individual and independent environments does not correspond anymore to the physical reality : one must consider that the qubits interact with the same environment , and furthermore the interaction between the qubits , which depends on their physical characterization , must also be taken into account .
this has important consequences for the dynamical behavior of the system .
a common environment may entangle initially separable systems , even in the absence of direct interaction @xcite .
this can be easily understood from the fact that the product state @xmath8 may be expressed as a sum of a singlet and a triplet component : @xmath9\,,\ ] ] and , since the singlet component does not decay under a common environment , a residual entanglement remains at asymptotic times @xcite .
this is not valid anymore for two initially excited atoms , since then the singlet component does not appear . in this case
, one can show that the system remains separable throughout its evolution @xcite .
the situation changes dramatically , however , when dipole interactions between the two initially excited atoms are taken into account . in this case , it was shown that entanglement does emerge , but only after a time lag @xcite .
here we study the dynamic evolution of classical and quantum correlations corresponding to this peculiar behavior of entanglement . as mentioned before , three main results emerge from our work : ( i ) the belated appearance of entanglement is preceded by the build up of strong classical and quantum correlations , which so to speak prepare the scenario for the emergence of entanglement ; ( ii ) the interaction between the two initially excited qubits , indirectly through the environment and directly through dipole forces , leads to an entangled state that has extremely small classical correlations ; and ( iii ) although stemming from very different definitions , different quantifiers of quantum correlations coincide within some ranges of time as the system evolves : this is the case for the mid and the quantum discord .
this paper is organized as follows . in sec .
[ sec : quantum and classical correlations ] .
we review the main quantifiers of quantum and classical correlations , with special attention to quantum discord and the measure - induced disturbance " ( mid ) introduced by luo @xcite . in sec .
[ sec : dynamics of correlations and entanglement ] .
we present the master equation that describes the interaction of two qubits with a common environment and the main results concerning the dynamical behavior of several quantities that have been introduced to characterize entanglement and quantum correlations : concurrence , quantum mutual information , quantum discord , and measure - induced disturbance .
we also consider in the same section the evolution of the classical correlations .
the conclusions are presented in sec .
[ sec : conclusions ] , while the appendices contain detailed demonstrations of some of the assertions made in the article .
the total classical correlations between two random variables x , y are given by the classical mutual information : @xmath10 where @xmath11 is the shannon entropy , with the probability distributions calculated from the joint one , @xmath12 : @xmath13 , @xmath14 and @xmath15 is the joint entropy with @xmath16 the probability of both outcomes @xmath17 and @xmath18 happening . using the bayes rule , @xmath19 , where @xmath20 is the conditional probability that the event @xmath21 occurs once the event @xmath22 has already occurred , the classical mutual information can be expressed equivalently as : @xmath23 where @xmath24 is the conditional entropy of @xmath21 given @xmath25 , with @xmath26 in this form , it becomes clear that the classical mutual information describes the difference in the ignorance about the subsystem @xmath21 before and after performing a measurement on subsystem @xmath25 .
ollivier and zurek generalized these two equivalent expressions of the classical mutual information to quantum systems . equation ( [ infoclassica1 ] ) is easily generalized by replacing the probability distributions by density matrices and the shannon entropy by the von neumann entropy @xmath27 , thus obtaining : @xmath28 it is not straightforward to generalize eq .
( [ infoclassica2 ] ) , since the definition of the conditional entropy @xmath29 involves specifying the state of the subsystem @xmath30 , knowing the state of the subsystem @xmath31 . in quantum theory
, this implies that a measurement must be applied to subsystem @xmath31 . the generalization proposed by ollivier and zurek
is done by assuming a complete unidimensional projective measurement made on system @xmath31 , corresponding to the projectors @xmath32 , such that @xmath33 .
the state of @xmath30 after this measurement is implemented is @xmath34 , with probability @xmath35 .
then , the conditional entropy is defined as @xmath36 , and the quantum equation corresponding to eq .
( [ infoclassica2 ] ) can be expressed as @xmath37 the quantum discord is the minimum of the difference between eq .
( [ infoquantica1 ] ) and eq .
( [ infoquantica2 ] ) : @xmath38 the quantity @xmath39 is the information gained about system a when measurements are performed on system b. through the process of minimization over all possible measurements on system b we search the measurement that disturbs least the total quantum system and allows to obtain as much information as possible from system a. ollivier and zurek demonstrated that @xmath40 that is , a zero quantum discord implies that the complete quantum system is not disturbed by the measurement .
furthermore , the information about the system @xmath30 is not perturbed by the measurement of system @xmath31 . however
, a nonzero quantum discord implies that the measurement disturbs the state and part of the information about system @xmath30 that exists in the correlations between the subsystems @xmath30 and @xmath31 is lost .
equation ( [ nonpert ] ) shows that the concept of quantum discord admits a simple physical explanation .
however , this measure has an annoying feature : it is asymmetric under exchange of systems @xmath30 and @xmath31 , which should not be expected from a quantifier of quantum correlations .
this is made very clear by going back to the example given by eq .
( [ quantumcor ] ) . for any complete one - dimensional projective measurement performed on the second qubit
, the quantum state of the first one is disturbed and therefore information about it is lost .
the quantum discord for this case is different from zero .
but if we perform the measurements on the first qubit , the quantum state is not perturbed and all the information on the second qubit , initially present in the state , is recovered . in this case
the quantum discord would be zero .
this example also highlights the difference between separability and classicality , and clearly associates the existence of quantum correlations in a separable state with the presence of nonorthogonal states in one of the subsystems , as already mentioned by @xcite .
as shown in @xcite , the quantum discord is always greater or equal than zero , and is zero if and
only if the state has only classical correlations . given any bipartite state @xmath41 , one may associate with it , by means of local measurements , another state , interpreted as the classical part of the former .
the quantum correlations present in the state @xmath41 are then determined by quantifying the difference between these two states .
consider any complete set of one - dimensional orthogonal projections @xmath42 , @xmath43 acting on each party _ a _ and _ b_. the state after the measurement is : @xmath44 if for some measurement @xmath45 , then the state is called classsical state , otherwise the state is truly quantum .
luo demonstrated in @xcite that , if @xmath41 is classical , @xmath46 , @xmath47 and @xmath48 are the eigenprojectors of @xmath49 , @xmath50 and @xmath41 respectively .
this implies that the definition of classical states is unambiguous , since there is a unique measurement that leads to @xmath45 .
when a complete set of projective measurements @xmath51 is performed on a system with state described by the density matrix @xmath41 , the entropy of the final state @xmath52 is greater or equal than the entropy of the initial state , @xmath53 , and equality is attained only when the projective measurements are the eigenprojectors of the matrix @xmath41 @xcite . in order to quantify the quantum correlations in a quantum state @xmath41 , luo @xcite chose the measurement @xmath54 induced by the eigenprojectors of the reduced subsystems . with that choice ,
the reduced states remain invariant , and then also the corresponding entropies : the information on each subsystem is not changed and , in that sense , this measurement is the least disturbing .
the corresponding measure of quantum correlations , named measurement - induced disturbance " , is defined as : @xmath55 , where @xmath56 is the quantum mutual information and @xmath57 , with @xmath58 and @xmath59 the eigenprojectors of the reduced subsystems , @xmath60 and @xmath61 , respectively . in particular
when @xmath41 is a pure bipartite state , @xmath62 : this measure coincides with the quantum discord , and is equal to the entropy of the reduced system , which is a measure of entanglement . the mid presents a great advantage with respect to the quantum discord
: it is easily calculable , since it does not involve any minimization , one has only to find the eigenvectors of the density matrices corresponding to the subsystems .
however , it is not applicable in all cases as wu _
@xcite have remarked : when the local density matrices have degenerate eigenvalues , the mid is not uniquely defined .
a quantifier for classical correlations was proposed by henderson and vedral @xcite . the classical mutual information given by eq .
( [ infoclassica2 ] ) was generalized to quantum mechanics by replacing the shannon entropy by the von neumann entropy and the classical probability distributions by density matrices .
the corresponding quantifier is given by : @xmath63 where @xmath64 is the state of the subsystem @xmath31 after performing the povm @xmath65 on @xmath30 .
hamieh _ et al .
_ showed in @xcite that , for the case of two qubits , the povm that maximizes eq .
( [ classicalcorr ] ) is a complete set of unidimensional orthogonal projectors . with this result
it is easy to see that @xmath66 , thus the total correlations of the system , quantified by the quantum mutual information ( [ infoquantica1 ] ) , are separated in quantum correlations ( measured by the quantum discord ) and classical correlations .
one should note that this definition also presents the problem of asymmetry : the quantification of classical correlations depends on which subsystem is measured .
wu _ et al . _
@xcite proposed an alternative definition for a quantifier of quantum correlations , which is symmetric and , as opposed to mid , unique even when the states of the subsystems have degenerate eigenvalues .
they proposed to quantify the classical correlations between two systems @xmath30 and @xmath31 by performing a povm locally on each subsystem . with the detection records they calculate then the classical mutual information for the chosen povm .
they define @xmath67 as the maximal classical mutual information available over all choices of possibles povm s . in order to quantify the quantum correlations in a bipartite state ,
they define @xmath68 . when @xmath69 is a pure state @xmath70 , and in the case of a general mixed state , @xmath71 in this work we refrain from studying the dynamical behavior of the quantity @xmath72 , because it requires an optimization over all possible measurements , and an analytical expression for it is still unknown . for the quantification of entanglement we use the _ concurrence _
@xcite , defined as : @xmath73 where : @xmath74 @xmath75 being the eigenvalues in decreasing order of the matrix : @xmath76 where @xmath41 is the two - qubit density matrix , @xmath77 is the second pauli matrix and the conjugation is performed in the computational basis .
concurrence ranges from 0 , which corresponds to a separable state , to 1 , which corresponds to a maximally entangled state .
we consider a system of two identical qubits interacting with all modes of the electromagnetic field , assumed in the vacuum state .
the state of each qubit is represented in the basis @xmath78 ( @xmath79 excited , @xmath80 ground state ) .
the two qubits interact via dipole forces , associated to the dipole transition moments @xmath81.the total hamiltonian of the atoms plus the electromagnetic field , in the electric dipole approximation , is @xmath82,\end{aligned}\ ] ] where @xmath83 and @xmath84 are the ladder operators , @xmath85 is the energy operator of the ith qubit , @xmath86 are the transition frequencies ( in what follows we will consider all frequencies equal , @xmath87 ) , @xmath88 and @xmath89 the annihilation and creation operators corresponding to the field mode @xmath90 , with wave vector @xmath91 , frequency @xmath92 and index of polarization @xmath93 .
@xmath94 is the coupling constant , @xmath95 is the position of the ith qubit , @xmath96 is the normalization volume , and @xmath97 is the unit polarization vector of the field . when the length of the system is small compared to the radiated wavelength @xmath98 , we may neglect the spatial variation of @xmath99 , so that , in the rotating - wave approximation , the hamiltonian reduces to @xcite : @xmath100,\end{aligned}\ ] ] where @xmath101 and @xmath102 are collective spin operators defined by : @xmath103 and @xmath104 .
this hamiltonian describes the dicke model .
the other limit , when the length of the system is much greater than the resonant wavelength , is easily obtained from the master equation approach that we considerer in the following .
the dynamical evolution of the qubit system is given by the following master equation @xcite : @xmath105- i \omega_{12}\sum_{i\ne j}^{2}[s_{i}^{+}s_{j}^{-},\rho ] \nonumber \\ & & -\frac{1}{2 } \sum_{i , j=1}^{2 } \gamma_{ij}(\rho s_{i}^{+}s_{j}^{-}+s_{i}^{+}s_{j}^{-}\rho-2s_{j}^{-}\rho s_{i}^{+})\ , , \nonumber\\ \label{mastereq}\end{aligned}\ ] ] where @xmath106 are the spontaneous emission rates of the qubits , assumed to be identical , and @xmath107 and @xmath108 are respectively the collective damping and the dipole - dipole interaction defined by @xmath109 \frac{\sin(k_{0}r_{12})}{k_{0}r_{12 } } \nonumber \\ & + \big [ 1 - 3 \left ( \hat{\mu } .
\hat{r}_{12 } \right)^{2 } \big]\bigg [ \frac{\cos(k_{0}r_{ij})}{(k_{0}r_{12})^2}- \frac{\sin(k_{0}r_{12})}{(k_{0}r_{12})^3 } \bigg ] \bigg\ } , \end{aligned}\ ] ] and @xmath110 \frac{\cos(k_{0}r_{12})}{k_{0}r_{12}}\nonumber \\ & + \big[1 - 3(\hat{\mu}.\hat{r}_{12})^{2}\big ] \bigg[\frac{\sin(k_{0}r_{12})}{(k_{0}r_{12})^{2}}+\frac{\cos(k_{0}r_{12})}{(k_{0}r_{12})^{3 } } \bigg ] \bigg\ } , \end{aligned}\ ] ] where @xmath111 , @xmath112 is the distance between the qubits , and @xmath113 as before is the dipole transition moment .
it follows from eq .
( [ gamma ] ) that @xmath114 .
when the distance between the qubits is much greater than the resonant wavelenght , it is easy to see that @xmath115 and @xmath116 , so that the master equation becomes @xmath117\\ & -\frac{1}{2}\gamma \sum_{i}^{2 } ( \rho s_{i}^{+}s_{i}^{-}+s_{i}^{+}s_{i}^{-}\rho-2s_{i}^{-}\rho s_{i}^{+})\ , , \label{mastereqsep}\end{aligned}\ ] ] which corresponds to the case of independent environments . in order to study the dynamical evolution of quantum and classical correlations ,
the master equation ( [ mastereq ] ) was solved analytically ; we refrain from showing the details of the solution here since our results coincide with those in ref . @xcite . from the analytical solution , we calculated the relevant quantities for our purposes : concurrence , quantum discord , mid and classical correlation , for several initial states .
+ the dynamical evolution of these quantities when both qubits are initially excited , interacting with a common reservoir and with each other via dipole - dipole interaction , is shown in fig .
[ correl ] . ) ] we see that initially all correlations , with the exception of the concurrence that remains zero , rise to a maximum value and then decrease . one should also note that , for a finite period of time , the mid and the quantum discord coincide .
this can be understood by using the results in @xcite ; for a subset of density matrices with structure x , @xmath118 the quantum discord @xmath119 is given by : @xmath120 , where @xmath121 and @xmath122 with @xmath123 + the density matrix describing the dynamical evolution of the two - qubit system here considered , when both qubits are initially excited , has the same form as the one in eq .
( [ rhocaldeira ] ) .
indeed , we get @xmath124 where @xmath125(1 -\gamma)}{2(1 + \gamma ) } \nonumber \\ & + & \frac{[-e^{-2\tau}+e^{-(1+\gamma)\tau}](1 + \gamma)}{2(1 -\gamma ) } \nonumber \\
c(\tau)&=&-\frac{[-e^{-2\tau}+e^{-(1-\gamma)\tau}](1 -\gamma)}{2(1 + \gamma)}\nonumber \\ & + & \frac{[-e^{-2\tau}+e^{-(1+\gamma)\tau}](1 + \gamma)}{2(1 -\gamma)}\ , , \label{coeff}\end{aligned}\ ] ] with @xmath126 as a function of the renormalized time @xmath127 , these expressions depend on a single parameter @xmath128 .
it is easy to show from thie above equations that the analytic expression for the mid coincides with the expression for @xmath129 ( see appendix a ) , so whenever @xmath130 , the mid coincides with the quantum discord .
this condition is fulfilled for a finite time domain , that depends on the parameter @xmath128 as shown in fig .
[ d1d2 ] . at a later time
@xmath131 , the system becomes entangled and the concurrence starts increasing .
this feature was previously found by ficek and tana in @xcite .
they showed that , for two qubits initially in their excited states , under the conditions considered in this paper ( interacting with a common environment and through dipole forces ) , and for small distance @xmath132 between them ( as compared to the resonance wavelength ) , entanglement appears only after some finite period of time .
this feature is independent of the values of the parameters , as shown in appendix b. after the time interval where mid and quantum discord match , the numerical value of the quantum discord remains close to the mid , but always a little lower . on the other hand , after the emergence of entanglement , the difference between mid and concurrence continuously decreases , going to zero when both vanish at infinite times .
classical correlations are smaller than quantum correlations for most of the dynamic evolution of the state , a result that was previously found in @xcite . after they reach their maximum value , they go to zero asymptotically
. however , for most of the time of the evolution , they have a very small value , about a hundred times smaller than the concurrence .
this behavior is quite general : we have demonstrated it for several values of the parameter @xmath133 , as shown in appendix c. in ref .
@xcite , it was shown that only with three or more qubits can one have states with genuine quantum correlations but no classical correlations .
our example does not contradict that general result , since for the two - qubit system here considered the classical correlations , as defined in @xcite , are indeed different from zero , but it also shows that the same correlations , even though not vanishing , may be much smaller than the quantum correlations , even for a two - qubit system .
ficek and tana also showed in ref .
@xcite that , for the dicke model ( evolution of two two - level atoms in a common reservoir with no dipole - dipole interaction ) , two qubits initially excited do not get entangled .
a question naturally arises as to whether that is the case for the other quantum correlations in the dicke model .
this question is answered by taking the limit @xmath134 in the expressions for @xmath133 and @xmath108 . in that case , @xmath135 , so that @xmath136 , and @xmath137[1 - 3(\vec{\mu}.\vec{r_{12}})^{2}]$ ] .
although @xmath108 diverges in this limit , eq . ( [ coeff ] ) shows that the evolution of the initial quantum state considered here is independent of the dipole - dipole interaction .
therefore , in this limit we do recover , for the considered initial state , the evolution predicted by the dicke model .
the density matrix for this model is easily obtained from eqs .
( [ rhonossa ] ) and ( [ coeff ] ) taking the limit @xmath136 : @xmath138 where @xmath139 figure [ correldicke ] displays the evolution of all the above - mentioned correlations , within the framework of the dicke model , when the two qubits are initially excited . in this case , the initial behavior of classical and quantum correlations ( except entanglement ) is the same as in the model that includes dipole - dipole interaction : an initial period of growth to a maximum and then a decay .
however , one does not have , as in the former situation , a revival of these correlations .
the lack of entanglement for an initially doubly excited two - qubit state @xmath140 in the absence of dipole - dipole interaction stems from the fact that the singlet state @xmath141 does not show up in the evolution , as discussed by ficek and tana @xcite : this is a decoherence - free state in the dicke model , and its constant contribution leads to asymptotic entanglement when the initial state is @xmath142 , since this state can be written as a sum of the anti - symmetric singlet @xmath141 and the symmetric state @xmath143 .
the dipole - dipole interaction , on the other hand , leads to the emergence of the singlet state for the initial state @xmath140 , to the survival of quantum correlations , and to the belated appearance of entanglement .
indeed , the analytical expressions for the mid and quantum discord show that , for @xmath144 , the decay time of the symmetric state , both mid and quantum discord decay as @xmath145 , which is the decay law for the population of the singlet state .
this is shown in figure [ decayasimetrico ] , which makes clear the essential role of the singlet state in explaining the remarkable behavior of the entanglement and the quantum correlations in the presence of dipole - dipole interactions . ) ] when the distance between the qubits @xmath132 becomes much larger than the resonant wavelength , we recover the expected behavior for two initially excited qubits evolving in independent environments : the state remains separable and no correlations are created between the qubits .
in this work , we have studied the subtle dynamics of quantum and classical correlations for two qubits , initially in a pure product state , coupled through dipole forces and interacting with all modes of the electromagnetic reservoir .
when the two qubits are initially excited , entanglement has a peculiar behavior @xcite : it remains zero for a finite - time interval , and then it builds up .
our work is aimed at clarifying what happens during the dormant " time , that is , what kind of dynamic changes prepare the system for the late appearance of entanglement . with this aim
, we have analyzed the dynamic behavior of quantifiers of quantum and classical correlations .
we show that , for two excited qubits , the reservoir creates initially classical correlations and quantum correlations between the qubits , which remain in a separable state .
these correlations evolve from zero to a maximum value , and then decay .
this overall behavior does not depend on the presence of dipole - dipole interactions .
however , this interaction is fundamental for keeping up the quantum correlations and building up entanglement after a time lag .
it delays the decay of quantum correlations , thus allowing for the build up of entanglement .
therefore , in the presence of dipole - dipole interactions , quantum correlations can be considered as precursors of entanglement . in the absence of these interactions ,
they build up and decay to zero , the state remaining separable for all times .
the dipole - dipole interaction helps preserving quantum correlations , and this seems to fire up entanglement in this case .
while entanglement is still zero , there is a time span for which mid and quantum discord coincide .
even before entanglement appears , classical correlations become negligible , and remain so throughout the evolution . therefore , the very evolution of the system generates a peculiar entangled state , with very small classical correlations .
even though there is still considerable controversy over the proper definition of quantum correlations and their role in quantum computation , their dynamics under the action of the environment seems to be , in the present context , intimately related to the generation of entanglement .
further studies in this direction might help to elucidate this subtle dynamic connection between quantum correlations and entanglement .
the density matrix corresponding to the initial state is given by eq .
( [ rhonossa ] ) . to calculate the mid
, we apply projective measurements on the density matrix , corresponding to the eigenvectors of the reduced density matrices . after these measurements , which amount to keeping only the diagonal elements in eq .
( [ rhonossa ] ) , the density matrix is : @xmath146 then @xmath147\nonumber \\ & + [ b(\tau)-c(\tau)]\log_{2}[b(\tau)-c(\tau ) ] \nonumber \\ & + [ b(\tau)+c(\tau)]\log_{2}[b(\tau)+c(\tau ) ] , \label{mididiscordia}\end{aligned}\ ] ] expression that coincides with @xmath129 , eq .
( [ d1 ] ) , throughout the evolution . according to @xcite
the quantum discord is the minimum of @xmath148 .
therefore , whenever @xmath149 , the mid coincides with the quantum discord .
the expressions for @xmath129 and @xmath150 do not allow getting an analytical solution for the interval of time where @xmath149 .
[ d1d2 ] exhibits the results of numerical calculations made for several values of @xmath128 , displaying time intervals ( reduced to a single point for small @xmath128 ) where @xmath149 . as mentioned above , in the calculation of the mid
there is an ambiguity when the reduced matrices have degenerate eigenvalues , or equivalently , are multiples of the identity matrix . in this case ,
any two orthogonal vectors are eigenvectors of the reduced matrices , so there are infinite ways to choose the local projective measurements , which give rise to different values of the mid . here
the reduced density matrices are : @xmath151 and the mathematical condition for the reduced matrices to become proportional to the identity is : @xmath152 with @xmath153\right . \\ & & \left .
\times(1+\gamma^2)-\gamma \sinh ( \gamma\tau ) \right\}\,.\end{aligned}\ ] ] this expression decreases monotonically from its initial unitary value to zero , as @xmath154 see fig .
[ ab ] . therefore , there can only be a single instant of time for which eq .
( [ degenerados ] ) holds . at this point ,
the mid is not uniquely defined and depends on the eigenvectors chosen to make the measurements . however , since , for all other times , the eigenvectors are uniquely defined and coincide with the computational basis , it is natural to choose the same basis for the degeneracy points , in which case the mid does not present discontinuities . the concurrence @xmath156 is expressed in terms of the square roots of the eigenvalues of the matrix in eq .
( [ lamatriz ] ) : @xmath157},\ , \sqrt{a(\tau)\left[1-a(\tau)-2b(\tau)\right]},\nonumber\\ & \left|b(\tau)-c(\tau)\right|,\ , \left|b(\tau)+c(\tau)\right|\}\,.\end{aligned}\ ] ] first we will show that @xmath158 . this is immediate since , from eq .
( [ coeff ] ) , @xmath159 we distinguish three different cases : ( 1 ) @xmath160}$ ] is the largest value in eq .
( [ values ] ) , in which case @xmath161 , ( 2 ) @xmath162 is the largest value , in which case @xmath163}\right)\bigg\}$ ] , and ( 3 ) @xmath164 is the largest value , in which case @xmath165}\right)\bigg\}$ ] . in the second case
, we must investigate the behavior of @xmath169}$ ] .
first , we note that this case is realized only in the interval where @xmath170 , so that @xmath171 .
we show now that @xmath172 , which implies that @xmath173}$ ] .
this is equivalent to proving that @xmath174 .
since @xmath175 , the equality is verified at @xmath176 .
furthermore , it is easy to show that @xmath177 for all @xmath127 , independently of the value of the parameter .
therefore , @xmath178 is a monotonically decreasing function of @xmath127 with @xmath175 , so @xmath179 for all @xmath127 , implying that @xmath173}$ ] .
thus , we conclude that for this second case the concurrence is also zero . in the third case
, we must investigate the behavior of @xmath180}$ ] . because the equation @xmath181 contains the time in a non - algebraic expression
, it is not possible to determine analytically the time when the state becomes entangled ; however we can show , through some simple mathematical arguments , that , whatever the value of the parameter , this time always exists .
this requires to show that there is a time interval for which two conditions are satisfied : ( i ) @xmath182 ; and ( ii ) @xmath164 is the largest value in eq .
( [ values ] ) condition ( i ) is demonstrated by noting that @xmath183 , and @xmath184 .
then between zero and infinite there exists an instant of time @xmath185 for which @xmath181 .
this implies the existence of a time interval from @xmath186 to infinity where @xmath187 is negative , which would imply entanglement , as long as condition ( ii ) is satisfied within at least part of this time interval .
we show now that condition ( ii ) holds for sufficiently large @xmath127 .
this follows from @xmath188 , which implies that @xmath189 for sufficiently large times .
therefore , for times sufficiently large , one has @xmath182 and @xmath190 .
it remains to show that , for sufficiently large times , @xmath191 } $ ] .
this is shown by noting that there is a time @xmath192 such that , for all @xmath193 , @xmath194 , which is immediate .
therefore , for sufficiently large times , both conditions ( i ) and ( ii ) above are fulfilled , and the state becomes entangled . as mentioned above
, classical correlations are much smaller than the concurrence for most of the evolution . depending on the analytical expression of the quantum discord ( @xmath129 or @xmath150 )
, we have two expressions for the classical correlations , each valid in a different time interval : @xmath195\log_{2}[a(\tau)+b(\tau)]-\nonumber \\ & 2[1-a(\tau)-b(\tau)]\log_{2}[1-a(\tau)-b(\tau)]+\nonumber \\ & [ 1-a(\tau)-2b(\tau)]\log_{2}[1-a(\tau)-2b(\tau)]+\nonumber \\ & a(\tau)\log_{2}a(\tau)+2b(\tau)\log_{2}b(\tau)\end{aligned}\ ] ] and , @xmath196\log_{2}[a(\tau)+b(\tau)]+ \nonumber \\ & a(\tau)\log_{2}a(\tau)+ [ b(\tau)-c(\tau)]\log_{2}[b(\tau)-c(\tau)]+ \nonumber \\ & [ b(\tau)+c(\tau)]\log_{2}[b(\tau)+c(\tau)]+ \nonumber \\ & [ 1-a(\tau)-2b(\tau)]\log_{2}[1-a(\tau)-2b(\tau)]+ \nonumber \\ & [ a(\tau)+b(\tau)]\log_{2}[1-a(\tau)-b(\tau)]+ \nonumber \\ & \frac{1}{2}(1+\alpha)\log_{2}\left[\frac{1}{2}(1+\alpha)\right]\nonumber\\ & + \frac{1}{2}(1-\alpha)\log_{2}\left[\frac{1}{2}(1-\alpha)\right]\end{aligned}\ ] ] where @xmath197 , @xmath166 , @xmath198 are given in eq .
( [ coeff ] ) and @xmath199 ^ 2 + 4|c(\tau)|^2 $ ] .
figure [ classical ] displays the classical correlations and the concurrence for several values of the parameter @xmath128 .
typically , classical correlations are of the order of 1% of the concurrence . | we show that for two initially excited qubits , interacting via dipole forces and with a common reservoir , entanglement is preceded by the emergence of quantum and classical correlations . after a time lag ,
entanglement finally starts building up , giving rise to a peculiar entangled state , with very small classical correlations .
different measures of quantum correlations are discussed , and their dynamics are compared and shown to lead to coincident values of these quantifiers for several ranges of time . |
the role of graph analysis is becoming increasingly important in science and industry because of the prevalence of graphs in diverse scenarios such as social networks , the web , power grid networks , and even scientific collaboration studies .
massive graphs occur in a variety of situations , and we need to design better and faster algorithms in order to study them
. however , it can be difficult to access to informative large graphs in order to test our algorithms .
companies like netflix , aol , and facebook have vast arrays of data but can not share it due to legal or copyright issues .
moreover , graphs with billions of vertices can not be communicated easily due to their sheer size . as was noted in @xcite , good _ graph models
_ are extremely important for the study and algorithmics of real networks .
such a model should be fairly easy to implement and have few parameters , while exhibiting the common properties of real networks .
furthermore , models are needed to test algorithms and architectures designed for large graphs .
but the theoretical and research benefits are also obvious : gaining insight into the properties and processes that create real networks . the _ stochastic kronecker graph _ ( skg ) @xcite , a generalization of the _ recursive matrix _
( r - mat ) model @xcite , has been proposed for these purposes .
it has very few parameters and can generate large graphs quickly . indeed , it is one of the few models that can generate graphs fully in _
parallel_. it has been empirically observed to have interesting real - network - like properties .
we stress that this is not just of theoretical or academic interest this model has been chosen to create graphs for the graph500 supercomputer benchmark @xcite .
it is important to know how the parameters of this model affect various properties of the graphs .
we stress that a mathematical analysis is important for understanding the inner working of a model .
we quote mitzenmacher @xcite : i would argue , however , that without validating a model it is not clear that one understands the underlying behavior and therefore how the behavior might change over time .
it is not enough to plot data and demonstrate a power law , allowing one to say things about current behavior ; one wants to ensure that one can accurately predict future behavior appropriately , and that requires understanding the correct underlying model . "
we explain the skg model and notation .
our goal is to generate a directed graph @xmath0 with @xmath1 nodes and @xmath2 edges .
the general form of the skg model allows for an arbitrary square generator matrix and assumes that @xmath3 is a power of its size . here , we focus on the @xmath4 case ( which is equivalent to r - mat ) , defining the generating matrix as @xmath5 we assume that @xmath6 for some integer @xmath7 .
_ for the sake of cleaner formulae , we assume that @xmath8 is even in our analyses . _
each edge is inserted according to the probabilities defined by @xmath9 where @xmath10 denotes the kronecker product operation . in practice ,
the matrix @xmath11 is never formed explicitly . instead
, each edge is inserted as follows .
divide the adjacency matrix into four quadrants , and choose one of them with the corresponding probability @xmath12 , or @xmath13 .
once a quadrant is chosen , repeat this recursively in that quadrant .
each time we iterate , we end up in a square submatrix whose dimensions are exactly halved .
after @xmath8 iterations , we reach a single cell of the adjacency matrix , and an edge is inserted .
it should be noted that here we take a slight liberty in requiring the entries of @xmath14 to sum to 1 .
in fact , the skg model as defined in @xcite works with the matrix @xmath15 , which is considered the matrix of probabilities for the existence of each individual edge ( though it might be more accurate to think of it as an expected value ) .
note that all edges can be inserted in parallel .
this is one of the major advantages of the skg model and why it is appropriate for generating large supercomputer benchmarks . for convenience ,
we also define some derivative parameters that will be useful in subsequent discussions .
we let @xmath16 denote the _ average degree _ and let @xmath17 denote the _
skew_. the parameters of the skg model are summarized in .
.9 primary parameters * @xmath18 generating matrix with @xmath19 * @xmath20 number of levels ( assumed even for analysis ) * @xmath21 number of edges derivative parameters * @xmath22 number of nodes * @xmath23 average degree * @xmath24 skew our overall contribution is to provide a thorough study of the properties of skg and show how the parameters affect these properties .
we focus on the degree distribution , the number of ( non - isolated nodes ) , the core sizes , and the trade - offs in these various goals .
we give rigorous mathematical theorems and proofs explaining the degree distribution of skg , a noisy version of skg , and the number of isolated vertices . * degree distribution : * we provide a rigorous mathematical analysis of the degree distribution of skg .
the degree distribution has often been claimed to be power - law , or sometimes lognormal @xcite .
kim and leskovec @xcite prove that the degree distribution has some lognormal characteristics .
gror et al .
@xcite give exact series expansions for the degree distribution , and express it as a mixture of normal distributions .
this provides a qualitative explanation for the oscillatory behavior of the degree distribution ( refer to ) .
since the distribution is quite far from being truly lognormal , there has been no simple closed form expression that closely approximates it .
we fill this gap by providing a complete mathematical description .
we prove that skg _ can not _ generate a power law distribution , or even a lognormal distribution .
it is most accurately characterized as fluctuating between a lognormal distribution and an exponential tail .
we provide a simple formula that approximates the degree distribution .
* noisy skg : * it has been mentioned in passing @xcite that adding noise to skg at each level smoothens the degree distribution , but this has never been formalized or studied .
we define a specific noisy version of skg ( nskg ) .
we prove theoretically and empirically that nskg leads to a lognormal distribution .
( we give some experimental results showing a naive addition of noise does _ not _ work . )
the lognormal distribution is important since it has been observed in real data @xcite .
one of the major benefits of our enhancement is that only @xmath8 additional random numbers are needed in total . using graph500 parameters , plots the degree distribution of a ( standard ) skg and nskg for two levels of ( maximum ) noise .
we can clearly see that noise dampens the oscillations , leading to a lognormal distribution .
we note that though the modification of nskg is straightforward , the reason why it works is not .
it involves an intricate mathematical analysis , which may be of theoretical interest in itself . from the graph500 benchmark parameters with @xmath25 .
] * isolated vertices : * an isolated vertex is one that has no edges incident to it ( and hence is not really part of the output graph ) .
we provide a formula that accurately estimates the fraction of isolated vertices .
we discover the surprising result that in the graph500 benchmark graphs , 50 - 75% vertices are isolated ; see .
this is a major concern for the benchmark , since the massive graph generated has a much reduced size .
furthermore , the average degree is now much higher than expected .
.expected percentage of isolated vertices and repeat edges , along with average degree of _ non - isolated _
nodes for the graph500 benchmark . excluding the isolated vertices results in a much higher average degree than the value of 16 that is specified by the benchmark . [ cols="^,^,^,^",options="header " , ] [ tab : core ] we focus our attention on the max core number of skg
. how does this number change with the various parameters ?
the following summarizes our observations .
[ emp : core ] for skg with symmetric @xmath14 , we have the following observations .
the max core number increases with @xmath26 . by and large ,
if @xmath27 , max core numbers are extremely tiny .
max core numbers grow with @xmath8 only when the values of @xmath26 are sufficiently large .
even then , the growth is much slower than the size of the graph . for smaller @xmath26 ,
max core numbers exhibit essentially negligible growth .
max core numbers increase essentially linearly with @xmath28 .
large max core numbers require larger values of @xmath26 . as mentioned in , increasing @xmath26 increases the number of isolated vertices .
hence , there is an inherent tension between increasing the max core number and decreasing the number of isolated vertices . for the sake of consistency , we performed the following experiments on the max core after taking a symmetric version of the skg graph .
our results look the same for in and out cores as well . in ,
we show how increasing @xmath26 increases the max core number .
we fix the values of @xmath29 and @xmath30 .
( there is nothing special about these values . indeed the results are basically identical , regardless of this choice . )
then , we fix @xmath31 ( or @xmath32 ) to some value , and slowly increase @xmath26 by increasing @xmath32 ( resp .
@xmath31 ) .
we see that regardless of the fixed values of @xmath31 ( or @xmath32 ) , the max core consistently increases .
but as long as @xmath27 , max core numbers remain almost the same . in , we fix matrix @xmath14 and average degree @xmath28 , and only vary @xmath8 . for webnotredame
is not symmetric , we can still define @xmath26 .
also , the off diagonal values are @xmath33 and @xmath34 , so they are almost equal . ] , we have @xmath35 and for ca - hep - ph , we have @xmath36 . for both cases , increasing @xmath8 barely increases the max core number . despite increasing the graph size by @xmath37 orders of magnitude
, the max core number only doubles .
contrast this with the graph500 setting , where @xmath38 , and we see a steady increase with larger @xmath8 .
this is a predictable pattern we notice for many different parameter settings : larger @xmath26 leads to larger max core numbers as @xmath8 goes up .
finally , in , we see that the max core number is basically linear in @xmath28 .
our general intuition is that nskg mainly redistributes edges of skg to get a smooth degree distribution , but does not have major effects on the overall structure of the graph .
this is somewhat validated by our studies on isolated vertices and reinforced by looking at @xmath39-cores . in , we plot the core decompositions of skg and two versions on nskg ( @xmath40 and @xmath41 ) .
we observe that there are little changes in these decompositions , although there is a smoothening of the curve for graph500 parameters .
the problem of tiny cores of skg is not mitigated by the addition of noise .
for a true understanding of a model , a careful theoretical and empirical study of its properties in relation to its parameters is imperative .
this not only provides insight into why certain properties arise , but also suggests ways for enhancement .
one strength of the skg model is its amenability to rigorous analysis , which we exploit in this paper .
we prove strong theorems about the degree distribution , and more significantly show how adding noise can give a true lognormal distribution by eliminating the oscillations in degree distributions .
our proposed method of adding noise requires only @xmath8 random numbers all together , and is hence cost effective .
we want to stress that our major contribution is in providing _ both _ the theory and matching empirical evidence .
the formula for expected number of isolated vertices provides an efficient alternative to methods for computing the full degree distribution . besides requiring fewer operations to compute and being less prone to numerical errors
, the formula transparently relates the expected number of isolated vertices to the skg parameters .
our studies on core numbers establish a connection between the model parameters and the cores of the resulting graphs .
in particular , we show that commonly used skg parameters generate tiny cores , and the model s ability to generate large cores is limited .
we are grateful to david gleich for the matlab bgl library as well as many helpful discussions .
we thank todd plantenga for creating large skg and nskg instances , and for generating .
we also thank jon berry for checking our graph500 predictions against real data , and also david bader and richard murphy for discussions about the graph500 benchmark .
we acknowledge the inspiration of jennifer neville and blair sullivan , who inspired us with their different work on skg during recent visits to sandia . , kirshner , s. , neville , j. , and vishwanathan , s. v. n. 2010 .
tied kronecker product graph models to capture variance in network populations . in _ proc .
48th annual allerton conf . on communication , control , and computing_. 11371144 . | graph analysis is playing an increasingly important role in science and industry . due to numerous limitations in sharing real - world graphs , models for generating massive graphs are critical for developing better algorithms . in this paper
, we analyze the stochastic kronecker graph model ( skg ) , which is the foundation of the graph500 supercomputer benchmark due to its favorable properties and easy parallelization .
our goal is to provide a deeper understanding of the parameters and properties of this model so that its functionality as a benchmark is increased .
we develop a rigorous mathematical analysis that shows this model _ can not _ generate a power - law distribution or even a lognormal distribution . however , we formalize an enhanced version of the skg model that uses random noise for smoothing .
we prove both in theory and in practice that this enhancement leads to a lognormal distribution .
additionally , we provide a precise analysis of isolated vertices , showing that the graphs that are produced by skg might be quite different than intended .
for example , between 50% and 75% of the vertices in the graph500 benchmarks will be isolated .
finally , we show that this model tends to produce extremely small core numbers ( compared to most social networks and other real graphs ) for common parameter choices .
[ complexity measures , performance measures ] [ graphs and networks ] this work was funded by the applied mathematics program at the united states department of energy and performed at sandia national laboratories , a multiprogram laboratory operated by sandia corporation , a wholly owned subsidiary of lockheed martin corporation , for the united states department of energy s national nuclear security administration under contract de - ac04 - 94al85000 .
author s addresses : c. seshadhri , a. pinar , t. kolda , sandia national laboratories , livermore , ca 94551 . |
it is believed that galaxies form in dark matter halos , which grow through gravitational instability from the density fluctuations seeded in the early universe .
the baryonic matter is then accreted into the dark matter halos and cools radiatively to assemble galaxies .
although the structure and the evolution of dark matter halos are fairly well understood with the use of analytic modeling @xcite and @xmath1-body simulations @xcite , galaxy formation still remains a challenging problem as it is a complex process involving simultaneous actions of many physical mechanisms .
one of the long - standing problems in galaxy formation is that , in the absence of heating sources , the majority of baryonic matter in the universe is predicted to have cooled into dark matter halos by the present time ( e.g. @xcite ) , while in reality only a small fraction of the baryonic matter associated with galaxies is observed in the cold ( stars plus cold gas ) phase ( @xcite ) .
this `` over - cooling '' problem implies that feedback must have played a very important role in shaping galaxies .
energetically , the ongoing galactic feedback appears to be sufficient to balance the cooling .
indeed , much of the expected mechanical energy from supernovae ( sne ) feedback is `` missing '' at least from x - ray observations @xcite . in general
, the observed x - ray emission consists of several components : 1 ) relatively bright x - ray binaries with neutron star and black hole companions ( @xmath2 ) , which may be detected individually in nearby galaxies ; 2 ) numerous cataclysmic variables ( cvs ) and coronally active binaries ( abs ) ( @xmath3 ) , which are too faint to be detected individually , but can be constrained based on near - ir emission @xcite ; 3 ) diffuse hot gas heated chiefly by type ia sne .
the typical observed diffuse x - ray emission only accounts for a few percent of the expected sn energy input ( e.g. , @xcite ) , the majority of which is undetected .
naturally , one would expect that the missing energy in such galaxies is gone with a galactic outflow , which could have a profound impact on the halo gas and the accretion of the igm .
of course , this impact should evolve with time and depend on the formation process of the galaxy .
for example , if a bulge forms primarily in a starburst ( sb ) , it can induce an initial blastwave propagating into the surrounding medium .
the heated circum - galactic medium ( cgm ) can then affect the impact of the later feedback from stellar mass loss and type ia supernovae ( sne ) as well as from star formation and agn .
therefore , one needs to study the interplay of the stellar feedback and the gas accretion of the galaxies to understand their evolution .
existing studies have shown the importance of the stellar feedback , though they are limited by the over - simplified treatments or are for some specific targets .
on the one hand , existing models for galactic bulge outflows / winds include little consideration for the evolving gravitational potential and gaseous circum - galactic environment ( e.g. , @xcite ) . on the other hand , in simulations of large - scale structure formation
, the treatment of the stellar feedback is limited by the spatial resolution . a more sophisticated model that incorporates the stellar feedback in an evolving galaxy has been investigated by @xcite with 1-d hydrodynamic simulations .
they found that the core - collapse sne in an initial sb can substantially affect the gas dynamics beyond 1mpc .
but the work was done specifically to explain the hot gas properties of the luminous elliptical galaxy ngc 4472 . in this high - mass galaxy ,
later feedback is found to be unimportant , partly because the authors adopted a low sn rate ( 0.03 snu ; one snu unit denotes one sn per @xmath4 per century ) . in this case , the gas dynamics is dominated by a cooling flow . in the present paper , we introduce a stellar feedback model in which the feedback is closely associated with the formation history of the galactic bulges . in this scenario ,
we divide the evolution of a bulge into three stages : 1 ) . at high redshifts , as the host dark matter halo assembles , the central galaxy builds up a bulge through frequent major mergers .
2 ) . as a consequence of the violent mergers , a sb in the bulge is triggered .
the intensive starburst powers a blastwave to heat and rarefy the cgm .
3 ) . at low redshift ,
the galaxy is in a quiescent phase as its host halo accretes mass gently .
however , as the low - mass bulge stars die off from the main sequence , type ia sne provide a long - lasting energy input to the cgm and this gradual feedback maintains a hot gaseous halo .
this scenario naturally connects the stellar feedback and the general formation history of a galaxy .
we find that such a feedback can profoundly affect the evolution of the model galaxies .
in particular , the overcooling and missing - energy problems can be naturally explained .
here we demonstrate the effects of the feedback based on a 1-d model of galaxies , although limited 2- and 3-d simulations have also been carried out to check multi - dimensional effects .
our aim here is to capture qualitatively the feedback effects on the structure , physical state , and dynamics of the hot gas around bulge - dominated galaxies on scales greater than galactic bulges / disks .
the organization of the paper is as follows .
we first describe the main physical ingredients of our model and numerical methods ( 2 ) .
we then present the results of a reference galaxy model and its variations ( 3 ) . in 4 ,
we discuss how the results provide a potential solution to the overcooling and missing energy problems , as well as the caveats of the present model .
we summarize our results and conclusions in 5 .
we set up hydrodynamical simulation models to investigate the effects of the feedback from a galactic stellar bulge on galaxy formation .
we focus our study on the early - type spirals and the typical low@xmath5@xmath0 elliptical galaxies in fields or in small groups ( such as the mw and m31 ) .
our fidicuial galaxy has an intermediate / massive galactic bulge and a dark matter halo with mass @xmath6 at the present time . in the following ,
we first describe how such a galaxy assembles mass , and then introduce the recipe to model the stellar bulge formation and its feedback .
finally we describe the numerical codes and methods . throughout the paper ,
we adopt the @xmath7cdm cosmology with @xmath8 , @xmath9 and @xmath10km / s / mpc .
we assume the universal cosmic baryon fraction to be @xmath11 .
it has been shown by high resolution @xmath1-body simulations that within the virial radius @xmath12 the density profile of a collapsed dark matter halo can be well described by the nfw profile @xcite , @xmath13 where @xmath14 is a characteristic radius and @xmath15 a characteristic density of the halo that is related to the collapse time of the halo .
the shape of such a profile is usually characterized by the concentration parameter , defined as @xmath16 . since a virialized halo
is defined as a spherical region with a mean density that is @xmath17 times the mean density of the universe , the characteristic density can be written as @xmath18 where @xmath19 is the critical density of the universe .
for the cosmology we adopted here , @xmath20 at @xmath21 @xcite . for a given cosmology , a virial mass , and a given time , the density profile of a dark matter halo is determined by the single parameter , @xmath22 .
simulations have demonstrated that dark matter halos assemble their mass through an early fast accretion phase followed by a slow one @xcite .
this two - phase assembly history can be fairly well described by an exponential function @xcite , @xmath23 where @xmath24 is the halo mass at the present time , @xmath25 is the expansion scale factor [ i.e. , @xmath26 , @xmath27 is the redshift ] , and @xmath28 is the scale factor at the halo `` formation time '' that separates the two accretion phases .
both @xmath1-body simulations @xcite and analytical model @xcite have shown that the halo concentration parameter is closely related to the formation time @xmath28 . for a halo that is still in its fast accretion regime , @xmath29 ; for a halo identified at @xmath30 , @xmath31 .
we make use of the 1-d simulation code developed by @xcite to reproduce the mass distribution of a growing dark halo embedded in the cosmic density field .
following @xcite and @xcite , we setup the initial perturbation at @xmath32 .
the perturbation profile is set up in such a way that the subsequent accretion history of the halo follows exactly the average mass accretion history given by eq.([eq : mah ] ) .
the halo s virial mass is @xmath33 at the present time . to be consistent with the statistical properties of dark halos in cosmological simulations , we choose the formation time @xmath34 .
the resulting halo has a concentration of about 13 , which is typical for halos of such a mass .
the advantage of adopting this approach is that , at any redshift , the density profiles both within and outside the current virial radius are modeled realistically .
[ fig : dmprofile ] shows the density profiles of the simulated halo at four different redshifts .
we further find that the evolution of our simulated profile can be fitted with the following formula : @xmath35,\quad x = r / r_s { \rm~for~ } r \leq r_v \\
\rho(r_v , t)\lbrack\frac{8}{9}\left(\frac{r}{r_v}\right)^{\alpha(t ) } + \frac{1}{9}\left(\frac{r}{r_v}\right)^{-1 } \rbrack,\quad { \rm~for~ } r > r_v \end{array}\right.\ ] ] where @xmath36 is the cosmic time in gyr ; @xmath37 is the halo virial radius at time @xmath36 and @xmath14 is the scale radius of the nfw profile at @xmath27=0 ; @xmath38 is a time - dependent index obtained from fitting the simulated profiles at @xmath39 .
this formula ( illustrated as the solid red lines in fig .
[ fig : dmprofile ] ) characterizes the simulated profiles to an accuracy of @xmath40 in the interested region .
we adopt this fitting formula to simplify the gravity calculation of the dark matter in eulerian - based hydro - dynamical simulations .
although the deposition of the cold gas in the halo center may make the the halo contract @xcite , our test with the same 1-d simulation code but including the gravity of the cool gas shows that the change of the dark matter distribution is only in the central several kpc , and that our main results on the halo gas properties are not expected to be affected significantly by such contraction .
lr halo mass @xmath41&@xmath42 + sb time @xmath43 ( gyr ) ( @xmath44 ) & 2.5 + sb gas mass @xmath45&@xmath46 + sb energy deposit region @xmath47 ( kpc ) & @xmath48 + bulge stellar mass @xmath49&@xmath50 + bulge scale @xmath51 ( kpc ) & @xmath52 + age of galactic bulge @xmath53 ( gyr ) & 10 + bulge blue band luminosity ( @xmath54 ) & @xmath55 we approximate the stellar feedback into two phases : ( 1 ) a blastwave generated by the initial sb that forms the stellar bulge and ( 2 ) a gradual injection of mass and energy through stellar mass - loss and type ia sne in the bulge .
additional feedback may arise from the galactic disk , if present ( although cool gas there may consume much of the input energy ) , and from agns .
we do not deal with these complications here , which may be incorporated into the above two phases approximately within the uncertainties in our adopted feedback parameters .
the key parameters are the initial sb energy feedback @xmath56 and the evolution of the type ia sn rate and mass loss rate .
we assume that the bulk of bulge stars forms in a single sb at the time @xmath57 right after the fast - accretion regime of the dark matter halo , when the inner halo is well established . in this regime
the gas cools in a time - scale much less than the halo dynamical time - scale ; dense gas clouds are thus expected to form rapidly , leading to the sb @xcite .
the sb converts a fraction of the cooled mass @xmath58 into bulge stars with a total mass @xmath59 comparable to that observed in our mw or m31 . the adopted values are listed in table [ tab : para ] .
we parameterize the sb energy feedback with an estimate of the mechanical energy input from core - collapsed sne ( @xmath60 ) .
this estimate depends on the initial mass function ( imf ) of stars formed in the sb . assuming the commonly used salpeter imf over the mass range of 0.08 - 100 m@xmath61 , the number of stars with individual masses greater than 8@xmath62 is @xmath63 .
the total energy deposited into the large - scale surrounding is @xmath64 where @xmath65 ( = @xmath66 ergs ) is the energy released by each sn .
the parameter @xmath67 characterizes the `` efficiency '' of sb feedback , which in principle accounts for the uncertainties in the assumed imf , @xmath65 , possible energy release from other sources ( e.g. , agn ) , and the radiative dissipation within the sb region . the radiative dissipation , for example , depends on the unknown details of the interaction between the sb and the ism ( e.g. , its porosity ; @xcite ) . and
a probably more realistic imf with a low mass turn - over ( e.g. , @xcite ) can give about 10% higher energy input .
following the core - collapsed sne of massive stars is the long - lasting but relatively gentle feedback from stars of lower masses . within the uncertainties in stellar evolution models
the energy input rate from type ia sne can be approximated as ( e.g. , @xcite ) @xmath68 where @xmath36 is the age of stellar population , and @xmath69 is the blue band luminosity of the stellar population at the present age @xmath53 ( table 1 ) , which can be determined from a stellar evolutionary model ( e.g. , @xcite ) .
the type ia sn starts @xmath70gyr @xcite after the starburst .
since the stellar properties in galactic bulges are similar to those in ellipticals , we set @xmath71=0.12 ( snu ) which is the average local sn ia rate for e / s0-type galaxies ( ( * ? ? ?
* ; * ? ? ?
* p467)).depending on the choice of models of snia progenitors @xcite , @xmath72 .
most of the stellar mass loss occurs at early time through the relatively fast evolution of intermediate - mass stars .
we use an updated calculation obtained from ( * ? ? ?
22 ) , which directly relates mass loss to the total initial mass of a stellar population .
the cumulative stellar mass loss of a normalized stellar population with initial mass of 1@xmath62 can be fitted by the following formula : @xmath73 which matches the numerical result to an accuracy of 1% from 0.1gyr to 15gyr .
an alternative formula introduced in @xcite gives the same shape as that of eq .
( [ eq : fitmlr ] ) for @xmath74gyr .
it is worth comparing the relative energetics of the stellar feedback during the sb and the later gradual phase .
there is @xmath75 sne ii per solar mass for a standard salpeter imf . for a general power law decay of the sne ia rate ,
the number of sne ia per solar mass is @xmath76 , where @xmath77 is the mass to light ratio , and @xmath78 with the reasonable ranges of sne ia emerging time @xmath79 ( e.g , 0.1 - 0.5 ) and @xmath80 ( e.g , 1.0 - 1.4 ) , @xmath81 with an uncertainty about a factor of two .
the @xmath82 for a single stellar population with standard salpeter imf is about 7 at 10 gyr @xcite .
thus the number of sne ia per solar mass is @xmath83 .
sb feedback in principle provides much more energy in a short time than the gradual feedback does in the life time of a galactic bulge .
note that the time in eqs .
( 6)-(8 ) is based on `` stellar clock '' and modeled independent of cosmology .
we assume that the `` stellar clock '' is consistent with the concordance cosmology model ( where @xmath84 ) , and scale the time unit in eqs .
( [ eq : snl ] ) and ( [ eq : fitmlr ] ) accordingly in our simulations .
we adopt the 1-d lagrangian hydrodynamical code @xcite to simulate the accretion of the gas and dark matter during the fast assembly regime . as introduced in [ sec : halo ] , we use 1000 and 100,000 shells to represent gas and dark matter , respectively . before the mass collapses into the virial radius
the two components follow each other .
when a gas shell is shocked roughly at the virial radius , the gas is decoupled from the dark matter flow and is heated up to the virial temperature ( see also @xcite ) .
the gas then cools radiatively .
when the temperature falls below @xmath85k , the gas shell is considered to be cold and is dropped out from the hot phase .
we re - distribute the cold gas according to a hernquist profile @xcite with the scale radius of 0.7kpc .
we terminate the 1-d lagrangian simulation at @xmath44 when the fast - accretion regime is over .
this simulation provides both the mass of the dropped - out gas ( @xmath86 ) and the density , temperature , and velocity profiles of the remaining hot gas .
the output of gas properties from the lagrangian code is adopted as the initial condition for our feedback simulations .
after the fast assembly epoch , the bulk of the galactic bulge has been built up and the sb is triggered .
the mass and the sb feedback energy are deposited instantaneously and uniformly in form of thermal energy within the radius @xmath87 kpc , which is about the virial radius of the halo at the assumed sb epoch ( table [ tab : para ] ) .
the included gas is assumed to be well mixed with the ejecta of the sne and is thus considered to be sb - enriched .
we have experimented various methods to deposit the sb energy ( e.g. , kinetic vs. thermal energy as well as radius ) and find that they do nt make significant differences as long as no additional dissipation is allowed before the deposited energy emerges into the large - scale surrounding . with the feedback included ,
we conduct our hydrodynamical simulations using the widely - used flash code @xcite .
this eulerian based code makes it easier to handle the position - wise mass and energy input .
the mass and energy input source terms , specified by eq .
( 6 ) and ( 7 ) , are handled by the operator split method in flash , the gravity force includes the dark matter halo with a time - dependent density profile described by eq .
( [ eq : dmfit ] ) and a fixed hernquist profile for the stellar bulge with a total mass of @xmath88 ( table [ tab : para ] ) .
the gravity is dominated by the bulge stellar mass at small radii and by dark matter at large radii . the gas self - gravity is thus ignored .
the radiative cooling of the gas depends on its metallicity as well as its local density and temperature in the simulations .
we assign representative metallicities of 0.01 , 0.1 , and 1 solar to the igm , the sb - enriched gas , and the type ia sn - enriched stellar material .
a 0.1 solar abundance of the sb - enriched gas corresponds to roughly a yield of 0.02@xmath89 metals per solar mass of the sb .
the metallicity - dependent cooling function is adopted from @xcite .
radiative cooling is switched off for gas of temperature below @xmath90k , which could be maintained by the radiation heating .
the different metallicities are also used to track the individual gas components .
we have chosen a set of models to illustrate how the dynamics and observational properties of the hot gas depend on the sb energy output efficiency ( @xmath67 ) , the gradual stellar feedback evolution , and the treatment of cold gas .
the model parameters are summarized in table [ tab : feedback ] . the reference model ( model ref . )
adopts the canonical values , which sets the reference for comparison with other models .
we demonstrate the effects of @xmath67 in model va , the varied gradual mass loading in model vb , and the varied type ia sn feedback evolution in model vc and vd . the effect of cool and hot gas coupling on the gas dynamics is examined in model ve .
our simulation radius is 2mpc around a model galaxy , including all the region that the feedback can influence . because gas at this radius is still participating in the hubble expansion , an `` outflow '' ( zero - gradient ) boundary condition is adopted , and the standard `` reflection '' boundary condition is used at the center .
in this section , we first focus on the fiducial model and then check the variance due to the changes in the model parameters
. we further give the immediate predictions of our model on the halo gas fraction and x - ray brightness .
[ fig : mwb4a ] illustrates the gas properties of the reference model galaxy at three representative redshifts .
soon after the sb ( e.g. , @xmath27=1.4 ; dot - dashed line ) , a blastwave ( corresponding to the jump at @xmath91 kpc in all the four panels ) is driven into the accretion flow of the galaxy .
this accretion flow is most apparent in the radial velocity panel ( i.e. , the negative velocity just in front of the blastwave ) . at larger radii
, the gas is still trailing the hubble expansion .
later , the blastwave , though weakening , catches up the trailing gas ( i.e. , no negative velocity at @xmath92 and 0.0 ) .
the blastwave heats the gas to a temperature of @xmath93 k , over a region extending well beyond the virial radius at @xmath94 ( @xmath95 kpc ) and even at @xmath96 ( @xmath97 kpc ) .
consequently , _ after this pre - heating by the sb - induced blastwave , no virial shock forms .
_ the sb heating also allows for the transportation of the later stellar feedback away from the galactic bulge . without the sb heating and the resultant rarefaction of the surrounding gas , the energy from the type ia sne
would be radiated away around the bulge and the gradual feedback could never launch a galactic bulge wind . with the low density of the ambient gas due to the expansion
, the gradual stellar feedback develops into a galactic bulge wind , at least in early time when the sn rate is high .
this wind reaches a velocity up to @xmath98 at @xmath99 and @xmath100 at @xmath101 as the rate decreases .
the wind is reverse - shocked at a quite large radius ( the innermost jumps of the profiles in fig .
[ fig : mwb4a ] ) .
the reverse - shock first moves outward and then starts to move back toward the center as the type ia sn rate declines with time .
the reverse - shocked wind material has a temperature of @xmath102 k. this temperature is determined by the specific energy of the feedback , minus the galaxy s gravitational potential that the wind needs to climb through .
the radiative cooling of the wind is negligible ( see table 2 and later discussion in 4.2 ) .
therefore , the sb heating enables the impact of the type ia sn - driven feedback over global scales around the galaxy . in return
, the long - lasting type ia sn - driven feedback helps to maintain and even to enhance the low - density hot environment produced by the sb .
first , the inauguration of the type ia sn - driven feedback about 0.1 gyr after the sb generates a forward shock , which can be seen in figure [ fig : mwb4a ] ( bottom panel ) as a small velocity jump just behind the sb blastwave .
this shock propagates fast in the sb - heated gas with little radiative energy loss @xcite and quickly overtakes the sb blastwave .
secondly , the type ia sn - driven stellar feedback prevents the sb - heated gas from falling back toward the galactic center to form a cooling flow , which would generate too large an x - ray luminosity to be consistent with observations .
sandwiched between the shocked bulge - wind material and the blastwave - heated igm is the sb - enriched gas ( e.g. , marked as the gray region for the @xmath103 density profile ) .
the two contact - discontinuities around the gas are apparent in the density and temperature profiles of fig .
[ fig : mwb4a ] .
the sb - enriched gas has a relatively high density and a low temperature ; and the gas has a cooling time scale of @xmath104gyr even with the assumed moderate metallicity .
the gas can cool to the lowest temperature allowed ( @xmath85k in our simulation ) first , as represented by spike - like features in the density profiles at lower redshifts .
this collapsed dense shell is subject to various hydrodynamical instabilities ( if not prevented by the presence of magnetic field ) ; the fragmented cool gas may be decoupled from the hot gas flow .
lrrrccccc ref .
& 0.12 & 0.5 & 1.4 & 0.25 & 0.45 & & 0.045 & 0.040 + va & 0.12 & 0.25 & 1.4 & 0.25 & 18.6 & & 0.047 & 0.111 + vb@xmath105 & 0.12 & 0.5 & 1.4 & 5.7 & 50.0 & & 0.047 & 0.082 + vc & 0.06@xmath106 & 0.5 & 1.6 & 1.2 & 2.8 & & 0.046 & 0.041 + vd & 0.35@xmath106 & 0.5 & 1.0 & 0.038 & 0.064 & & 0.045 & 0.039 + ve@xmath107 & 0.06 & 0.25 & 1.4 & 3.8 & 1673 & & 0.163 & 0.168 the most uncertain parameter in our reference model is @xmath67 , which determines the magnitude of the energy feedback during the sb phase . in model
va we reduce the sb feedback efficiency to @xmath108 .
the results are compared with those from the reference model in fig .
[ fig : mwb4cmp ] at @xmath21 .
naturally , a smaller @xmath67 value gives a slower and weaker sb blastwave , hence a relatively higher density and faster cooling of the interior gas .
the velocity becomes negative around virial radius and the material blown away earlier is being re - accreted into the halo .
this relatively high density circum - galactic environment affects the dynamics of the type ia sn - driven feedback , which is confined to a smaller region ( fig .
[ fig : mwb4cmp ] ) .
an interesting feature of this model is the higher temperature of the reverse shocked bulge wind material , which has not lost much of its energy in climbing out the gravitational potential of the galaxy .
the reverse shock is moving inward and is only about 20kpc away from the center of the galaxy at @xmath21 .
when it eventually reaches the center , the bulge wind will be quenched .
after that , a cooling flow of previously reverse - shocked bulge wind material will form .
the cooling flow , together with the infall of the already cooled sb - enriched shell may trigger a new sb in the bulge . in general , the smaller @xmath67 is , the earlier the reverse shock front moves back to the center .
a reduced type ia sn energy input rate ( e.g. , due to a smaller @xmath71 and/or @xmath109 than the assumed ) has a similar effect as that of a reduced sb feedback on the galactic gas dynamics : the decrease of energy input can not support a galactic wind or an outflow and even results in an inflow at later time .
we test a model with half of the rate adopted in the reference model , and find that the reverse shock and the cooled sb - enriched dense shell moved back to the center about 2gyr ago .
the resultant flow has an artificially cold core with a high inner density , similar to the figures shown in @xcite and ( * ? ? ?
3 ) , which consumes nearly all the energy feedback locally near the bulge center in the 1-d single phase modeling . such a cold core should undergo a nuclear starburst providing additional feedback .
we do not intend to address such complicated situations here .
a reduction in the mass loss rate has the opposite effect . with the same energy input , but a smaller mass loss rate , the specific heating of stellar ejecta would increase , helping to in reality the mass loss rate is most likely larger than the value calculated from eq .
( [ eq : fitmlr ] ) , which is based on a single burst of star formation scenario . in model
vb we double the stellar mass loss rate given by eq .
( [ eq : fitmlr ] ) . in this model
the inner bulge wind region has shrunk to a radius less than 20kpc at @xmath21 ( dotted grey line in fig .
[ fig : cmpp ] ) . inside the shocked bulge wind region
the density is about twice that of the reference model , and the density profile is leveling off outside .
an even larger mass loss rate can quench the inner bulge wind quickly , inducing an inflow by @xmath21 .
the gas dynamics is also affected by how the type ia sn rate evolves ( e.g , the choice of @xmath80 ) .
the inner bulge wind can be established more easily initially in a model with faster evolutionary trend ( i.e , with a larger @xmath80 ) . to examine this effect
we adopt @xmath110 in model vc , and adjust @xmath71 to ensure that this model has the same total amount of sne ia energy input as the reference model .
the profiles at @xmath21 are shown in fig .
[ fig : cmpp ] ( dashed red lines ) . in this model
more sne ia explode at early time so that the initial blastwave is stronger and can drive the outermost shock front further away than the reference model does .
but at @xmath21 the resultant bulge wind region is smaller and the temperature is lower because less energy is available to maintain the galactic flow at the late stage .
a similar test is done for @xmath111 ( table 2 model vd ) .
though initially the blastwave is weaker and the outermost shock front has a smaller galactocentric radius , this model maintains a much stronger bulge wind at @xmath21 because it has a higher type ia rate at the present time .
the effect of the collapsed dense shell on the hot gas flow depends on the coupling of the cooled and hot gases .
while the physical processes will be studied in our 2-d modeling ( tang & wang 2008 , in preparation ) , here we examine the effect in an extreme case by adding a simple `` dropout '' recipe in our simulation to decouple the cooled gas from the hot gas flow .
when a gas pixel cools below the minimum temperature allowed , we dropout 99% of its mass and increase its temperature accordingly to keep the pressure unchanged . using this recipe ,
the dropout does not generate numerical perturbation into the simulation and conserves the energy in the flow .
simulations with or without the dropout can thus be used to demonstrate the full range of possible dynamic effect .
the dropout in general has little effect on the gas dynamics , if the cooled gas shell is located at a large radius .
this is the case for the reference model ( fig .
[ fig : cmpdo ] ) . with the dropout ,
the flow just becomes a little `` lighter '' .
but because the gravitational potential is shallow at the large radius , the effect is small . on the other hand , if the cold dense shell gets close to or actually falls toward the center , the effect on the gas flow can be substantial .
the shell imposes a weight that the enclosed hot gas needs to hold , which can result in increased radiative cooling and eventually quench the bulge outflow . as mentioned previously , for models with a reduced @xmath71 or @xmath67 the dense shell moves back to the center and forms an artificially cold core before @xmath21 .
when this weight is removed through the dropout , the bulge outflow becomes faster and the radiative cooling becomes slower .
the gas evolution of the model ve with the `` dropout '' enabled is shown in fig .
[ fig : stado ] . in this model ,
the stellar ejecta is confined in a region less than 100kpc , the initial inner bulge wind is quickly quenched and turns into a subsonic outflow .
the outflow is stagnated around half of the virial radius ( i.e. , stagnation is defined as the outflow velocity becomes zero ) , and an accretion flow is built up near the virial radius . thus a density enhanced region is formed between half of the virial radius and the virial shock .
effective cooling first occurs in sb - enriched material because it has a relatively high density and low temperature ( see also fig .
[ fig : mwb4a ] ) .
most of the dropout occurs at radius larger than 100kpc in this simulation . by dropping out the cooler gas
, the galactic flow can stay in a quasi - static state for quite a long time , i.e. , as for the specific case ( model ve ) the density and temperature profiles are hardly changed for half of the hubble time .
the outermost shock front ( actually powered by sne ia which overtakes the blastwave of sb ) is still riding on the turnaround hubble flow at @xmath21 .
note that a global galactic wind would still be maintained by the gradual stellar feedback if it were started from a _ gas free _ initial condition ( fig . [ fig : stado ] ) . the stellar feedback results in a baryonic mass deficit within the virial radius .
the radial variation of the local baryonic fraction ( the baryon to total gravitational mass at each galactocentric radius ) of the reference model is shown in the left panel of fig .
[ fig : mf ] , and the corresponding accumulative baryonic fraction ( the total baryonic mass to the total gravitational mass within that radius ) is shown in the middle panel .
the dash - dot blue line denotes the radial variation at the starburst time @xmath43 .
the variation of the local baryonic fraction at this time follows the fluctuation of the simulated dark matter profile given by the 1-d lagrangian code ( see fig .
[ fig : dmprofile ] ) .
the cumulative fraction shows that the baryonic mass fraction decreases nearly monotonically and reaches the universal value around the virial radius . at @xmath27=0 ,
the baryon fraction drops down to about 5% near the virial radius and reaches the minimum value at the contact - discontinuity between the materials from the two feedback phases .
the baryon fraction then increases outward and finally reaches the universal fraction at the forward blastwave ( located at a distance about 6 times the virial radius ) .
this increase is largely due to the sb - enriched material and the cgm swept up to a large radius by the galactic wind . within the virial radius ,
the hot gas contribution is negligible , as shown by the dash red line ( the inner part has been scaled up by a factor of @xmath112 in the plot for easy visualization ) . the amount of baryonic mass within the virial radius depends on feedback parameters , such as the strength of the sb and the gradual stellar feedback evolution .
the cumulative baryonic mass fraction of model va , for example , is shown in the right panel of figure [ fig : mf ] . with the reduced sb efficiency ,
the baryonic mass fraction at @xmath21 is higher around the virial radius than that of the reference model . in table 2
we list the baryonic mass fractions for all the models within one and two virial radii , respectively . for models with strong sb feedback ( i.e. , models ref . , vc , and vd ) ,
the baryonic fractions within one virial radius are about 5% , and the fractions are even lower within two virial radius , which demonstrates that a large fraction of the baryonic mass is expelled out to a region beyond two virial radius . in model
ve the baryonic fraction is slightly less than the universal value because of its reduced stellar feedbacks .
the x - ray luminosities are listed in table 2 for all the models .
the luminosities within 2kpc characterize the diffuse x - ray emission from the inner bulge wind .
the x - ray luminosity of the reference model in the inner region is @xmath113 , and in model ve the x - ray luminosity is @xmath114 .
the x - ray luminosities within the virial radius vary dramatically with respect to different models . in the reference model
it is only @xmath115 , whereas in model ve the luminosity is as high as @xmath116 .
we now discuss the implications of the above results in connection to the issues mentioned in 1 and the additional work that is needed to further the study of the feedback scenario proposed here . except in model ve ,
the baryonic mass fractions are only about 5% within virial radius , consistent with an inferred low baryonic fraction in the milky way ( roughly about 6% , e.g. , @xcite ) . in our reference model , the radiative cooling after the sb is small because of the low density and high temperature of the halo gas . here
the feedback is treated self - consistently for both the internal heating of the intrahalo gas ( within a galaxy s virial radius ) and the external `` pre - heating '' of the surrounding igm .
the rarefaction of the pre - heated igm substantially reduces and even reverses the accretion .
the degree of the rarification can be characterized by the baryonic deficit around a galaxy ( e.g. , fig .
[ fig : mf ] ) .
the amount of cooling by the gaseous halo at present can , in principle , allow for placing constraints on @xmath67 .
for a too small @xmath67 , the gas would not be sufficiently pre - heated to reduce the accretion into the halo , and the baryonic mass fraction would be much closer to the cosmic universal value , as shown in model ve .
the total cooling rate of the gas around the reference galaxy is only a few of @xmath117 , the bulk of which arises from the accumulated reverse - shocked bulge wind material , located outside of the virial radius ( see fig .
[ fig : lxsb ] left panel ) . for the model with a lower @xmath67 value
, the cooling would be more efficient and occurs in regions closer to the bulge . in general
, no over - cooling problem occurs in our model of galaxy evolution with the proper stellar feedback included .
probably the most thorough study of hot gas outflows around ellipticals is by @xcite .
their 1-d model galaxies evolve through up to three consecutive evolutionary states : an initial wind , a transition to subsonic outflow , and a finial inflow phase .
x - ray faint ellipticals are suggested to be in their wind phase , and brightest ellipticals should mostly be in the inflow regime .
the subsonic outflow is shown to be unstable and short - lived .
our simulations have shown that the sb feedback is able to produce a low - density environment near the galactic bulge , which is similar to the `` gas - free '' condition assumed in @xcite .
but we note that the stellar bulge feedback material does not leave the galaxy freely .
much of the feedback energy may be thermalized well inside the virial radius , forming a hot gaseous halo .
the bulge wind solution applies only within the reverse shock front , shown by dot - dashed blue lines in figure [ fig : mwb4cmp ] , for example .
when the reverse shock moves inward , the wind becomes less and less developed and may eventually turn into a global subsonic outflow .
we find that this subsonic outflow is stable , in contrast to the result of @xcite . the unstable and short life of the subsonic outflow in their model is apparently caused by the use of the `` outflow '' outer boundary condition , which is generally not consistent with the changing properties of the galaxy and its environment .
the `` outflow '' boundary condition has no effect on the internal flow , as long as the outflow is supersonic .
but if subsonic , however , the outflow tends to be suppressed by the artificial force inserted by the leveled off pressure at the boundary .
the same force would amplify an inflow and quickly quench the inner bulge wind / outflow . in our model , no boundary condition needs to be assumed for the bulge wind / outflow , because it occurs well inside the spatial range of the simulations .
the stability of the subsonic outflow state has strong implications for understanding the large scattering in @xmath0 ratios of galactic bulges .
the distribution of the hot gas in the state depends not only on the feedback in the bulge , but also on the condition in outer regions of the galactic halo , which is a manifestation of the large - scale environment and formation history of a galaxy .
this dependence can create a large diversity in the x - ray luminosities of such galaxies . for the bulge outflow to be the solution of the missing feedback problem , a sufficient mechanical energy output from both the sb and type ia sne is needed .
a weak sb ( e.g. , a too small @xmath67 value ) would not create an environment with a low enough density for a lasting galactic bulge outflow .
even with a large sb feedback , a too low sne ia rate would lead to the blowout gas being re - accreted into the halo .
once an inflow is developed near the bulge , the feedback from sne ia would be consumed locally , producing a luminosity comparable to the input , which is disfavored by current observations . while our 1-d model may provide a reasonably good characterization of the large - scale impact of the stellar bulge feedback , the bulge wind / outflow is inefficient in producing diffuse x - ray emission .
in particular , we find that the x - ray luminosity from the 1-d wind / outflow is considerably lower than what is observed from galactic bulges .
observationally the reported `` diffuse '' x - ray emission of our galactic bulge and the one in m31 are about @xmath118 @xcite .
for all the models , the x - ray emission from the inner region is no more than a few of @xmath119 .
although @xmath120 of the energy is radiated within half of the virial radius in model ve , its x - ray emission near the bulge region is only a few times @xmath121 .
note that in model ve about 40% of the total energy input is used to remove the stellar mass ejecta out of the gravitational well at @xmath21 and a less fraction is required at early time .
it is also worth pointing out that the diffuse x - ray surface brightness of a bulge wind is generally steeper than that of stellar mass / light as demonstrated in fig .
[ fig : lxsb ] right panel ( see also @xcite ) .
only model ve , in which the gas flow is now in a subsonic outflow state , has a flatter surface brightness , consistent with observations of x - ray - faint early - type galaxies ( e.g. , @xcite ) .
the x - ray faintness of the bulge wind could be caused by our over - simplified 1-d model
. a careful analysis of the x - ray emission from the m31 bulge indicates the presence of a truly diffuse component with a 0.5 - 2 kev luminosity of @xmath122 ( li & wang 2007 ) .
however , this component has a substantially lower temperature ( @xmath123kev ) and a bipolar morphology in and around the bulge .
these facts indicate that much of this soft x - ray component arises from the interaction of hot gas with cool gas and that the 1-d approximation is too simplistic , at least in the immediate vicinity of the bulge region , which may be expected because of the presence of cool gas in the galactic disk .
a detailed study shows that this interaction may indeed be important , especially for the calculation of x - ray emission , which is sensitive to presence of density structures .
we are carrying out detailed 2-d and 3-d simulations of the m31 bulge region and comparisons with corresponding x - ray observations .
the results will be reported in separate papers .
stellar feedback is an important process in the co - evolution of galactic bulge and host galaxy .
the two - stage stellar feedback model we have introduced in this paper naturally sets up the stellar feedback from a galactic bulge in the context of the assembly history of the host galaxy / halo . in the first stage , violent protogalaxy mergers in the fast - assembly regime trigger a starburst which quickly releases feedback energy through core - collapsed explosions to expel a large fraction of the circum - galactic gas ; in the second stage , the old stellar population in the bulge provides a long - lasting feedback via type ia sne to keep the gaseous halo hot . in this model , the sb feedback and the gradual feedback are roughly synchronized with the early violent merger stage and the later gentle accretion stage of the host galaxy , which is a generic picture of galaxy formation in the cdm model .
a similar co - evolution scenario has been proposed for agn feedback @xcite . in their scenario ,
a central massive black hole builds up its mass during galaxy mergers , and the black hole appears as a `` quasar '' ; when a static hot halo forms in the latter time , the cooling halo gas fuels the central massive black hole to make it active in a so - called low - energy radio mode .
although our stellar feedback model and the agn feedback scenario are aimed at two different processes in galaxy formation , they share some common features .
firstly , the feedback processes happen in two distinct regimes , an early energetic but short - term blow - out and a late low - energy level but long - lasting heating .
secondly , the feedback and the galaxy formation are modeled in a co - evolving fashion . however , our stellar feedback model is proposed for the mw - sized galaxies , whereas the agn scenario is for more massive galaxies , which contribute to the bright end of the galaxy luminosity function .
much of existing work on stellar feedback is focused on whether a sb could drive a galactic superwind ( e.g. , @xcite ) . in the model
we have examined here , the question whether a sb - driven outflow can escape or not is not particularly relevant , because the galaxy is evolving and the galaxy environment is not gas - free . more relevant
is the cooling time scale of the sb - heated material .
this cooling time depends on the detailed physical processes involved in the interplay between massive stars and the ism , which our simple model does not take into account . depending largely on the assumed value of @xmath67
, the sb - heated material may have already fallen back to the center or is still hanging in regions outside the galactic virial radius .
our choice of the initial condition of the sb material ( the distribution scale ) guarantees that the cooling time - scale is long enough for the type ia sn - driven wind to develop an outflow .
it is the rarification of the halo gas by the sb feedback that helps the type ia sn - driven outflow to gently blow out the accreted material and replenish the ism with mass from evolving stars .
we mainly focus on a simultaneous solution to solve the `` over - cooling '' and `` missing energy '' problems in intermediate - mass ( @xmath124 ) galaxies with a substantial stellar bulge component .
the combination of the sb and gradual stellar feedbacks is demonstrated to be a viable way to solve the problems . in the following
we discuss briefly the limitations of our model .
how galactic stellar bulges really form remains a question of debate .
it is likely that there are multiple formation paths .
a popular paradigm is the formation through galaxy merges ( instead of a monolithic collapse ) , which typically happen at high redshifts .
in fact , our halo growth model is based on the average mass accretion history from @xmath1-body simulations ( e.g. , @xcite ) .
after the last major merger , a bulge evolves nearly passively , except for some minor mergers .
if the bulk of bulge stars are formed in a sb during the last major merge , our model still applies .
however , if the bulge stars forms primarily from pre - existing stars by mergers ( the so - called dry mergers ) , a direct application of our model would not be appropriate . we have considered galaxies in `` isolation '' .
this approximation may be reasonable for field galaxies , especially at high redshifts .
but as shown in our simulations , the stellar feedback may have significant effects on scales considerably larger than the virial radius of individual galaxies .
such impact may be sufficient to influence the igm environments of member galaxies in groups .
for example , much of the gas in the local group , which contains two mw - like galaxies with a substantial stellar bulge component , may have been heated to a temperature of @xmath102 k ( e.g. , fig .
[ fig : mwb4a ] ) .
such hot gas with too low a density to be easily detected is consistent with the existing observations .
but the stellar energy feedback in groups containing galaxies with more massive halos ( e.g. , giant ellipticals ) may have relatively minor effects , compared to the gravitational heating . in our numerical model , mass accretion
is assumed to be smooth . in reality , however , the accretion is more complicated .
@xmath1-body simulations of cosmic structure formation have shown that dark matter halos acquire their masses mainly through lumpy mergers rather than smooth accretion .
the sph galaxy formation simulations have further indicated that the gas accretion comes from both a smooth `` hot '' mode and a clumpy `` cold '' mode @xcite .
as discussed in @xcite , lumpy accretions may give rise to higher entropy in the igm than smoothed accretion , and hence may result in less cooling .
in addition , since denser clumps are difficult to be shock - heated ( e.g. @xcite ) , the gas accretion in different modes may lead to a multi - phase medium , which may alter the assembly history of the central galaxy , and hence changes the evolution of the stellar feedback and the thermal properties of the halo gas @xcite .
while the details of such a multi - phase medium are still poorly understood , the clumpy accretion alone can not solve the `` overcooling '' and `` missing energy '' problems , unless additional pre - heating is injected into the gas before it is accreted ( e.g. , @xcite ) .
the expectation is that the multi - phase medium may increase the gas accretion rate onto the central galaxy , hence boosting the predicted x - ray luminosity .
the cold accretion may also be responsible to some of the observed hvcs around our galaxy ( e.g. , @xcite ) , in addition to those that might be produced by the cooled feedback material discussed in 3.1.4 . for spiral galaxies
, additional cool gas may be channeled into the bulge regions via gravitational perturbations because of the presence of a stellar bar .
the heating of the gas ( e.g. , by thermal conduction ) may lead to mass - loading of the hot gas , increasing the x - ray luminosity and reducing the temperature of the gas .
we will study such effects in a separate paper .
we have examined the potential impacts of the galactic stellar bulge feedback on the hot gaseous structure and evolution around intermediate - mass galaxies with substantial bulge component .
these are typically early - type spirals such as the mw , m31 and ellipticals with low @xmath0 ratios , indicating ongoing mass and energy outflows .
the feedback in such a galaxy is approximately divided into two phases : 1 ) a sb - induced blastwave from the formation of the bulge and 2 ) a gradual energy and mass injection from evolved bulge stars .
our 1-d model deals with the feedback in the galaxy evolution context , approximately accounting for the dark matter accretion history . we have carried out various simulations with this model to demonstrate the impacts and their dependences on key assumptions and parameters used .
our major results and conclusions are as follows : * the blastwave initiated during the sb and maintained by subsequent type ia sne in a galactic bulge may heat the igm beyond the virial radius , which may stop further gas accretion .
the heated gas expands and produces a baryon deficit around the galaxy . *
the high temperature and low density of the remaining gas in the galactic halo produces little x - ray emission , consistent with existing observations .
thus the inclusion of both the sb and type ia sn feedbacks may provide a solution to the theoretical over - cooling problem .
* with the low - density circum - galactic gaseous environment , the gradual stellar feedback , powered by type ia sne alone , can form a galactic bulge wind , especially when the bulge is young .
much of the tenuous hot halo is occupied by the thermalized bulge wind material .
the bulge wind region shrinks and likely turns into a subsonic outflow as the sn rate decreases .
this wind / outflow naturally explains the _ missing stellar feedback _ problem of galactic bulges .
* we find that a subsonic outflow from a galactic bulge can be a long - lasting state , the properties of which depend sensitively on the environment and formation history of the galaxy .
this dependency may account for the large dispersion in the x - ray luminosities of galaxies with similar @xmath69 .
* the condensation of the gas injected during the sb may provide an explanation of hvcs observed with moderate metal enrichment .
we thank the referee for useful comments that led to improved presentation of this paper . the software used in this work was in part developed by the doe - supported asc / alliance center for astrophysical thermonuclear flashes at the university of chicago .
this project is supported by nasa through grant sao tm7 - 8005x and nng07ah28 g .
hjm would like to acknowledge the support of nsf ast-0607535 , nasa aisr-126270 and nsf iis-0611948 . | we demonstrate that the feedback from stellar bulges can play an essential role in shaping the halo gas of galaxies with substantial bulge components by conducting 1-d hydrodynamical simulations .
the feedback model we consider consists of two distinct phases : 1 ) an early starburst during the bulge formation and 2 ) a subsequent long - lasting mass and energy injection from stellar winds of low - mass stars and type ia sne .
an energetic outward blastwave is initiated by the starburst and is maintained and enhanced by the long - lasting stellar feedback .
for a milky way - like galactic bulge , this blastwave heats up the circum - galactic medium to a scale much beyond the virial radius , thus the gas accretion into the halo can be completely stopped .
in addition to that , the long - lasting feedback in the later phase powers a galactic bulge wind that is reverse - shocked at a large radius in the presence of circum - galactic medium and hence maintains a hot gaseous halo .
as the mass and energy injection decreases with time , the feedback evolves to a subsonic and quasi - stable outflow , which is enough to prevent halo gas from cooling .
the two phases of the feedback thus re - enforce each - other s impact on the gas dynamics .
the simulation results demonstrate that the stellar bulge feedback may provide a plausible solution to the long - standing problems in understanding the milky - way type galaxies , such as the `` missing stellar feedback '' problem and the `` over - cooling '' problem .
the central point of the present model is that the conspiracy of the two - phase feedback keeps a low density and a high temperature for the circum - galactic medium so that its x - ray emission is significantly lowered and the radiative cooling is largely suppressed .
the simulations also show that the properties of the hot gas in the subsonic outflow state depend sensitively on the environment and the formation history of the bulge .
this dependence and variance may explain the large dispersion in the x - ray to b - band luminosity ratio of the low @xmath0 elliptical galaxies . |
superconductivity in iron arsenides emerges from antiferromagnetic metallic parent compounds in the course of suppressing the magnetic ordering by chemical doping or pressure @xcite , resulting critical temperatures ( @xmath6 ) up to 55 k in sm(o@xmath3f@xmath4)feas @xcite .
relationships between the magneto - structural phase transition and superconductivity in iron arsenides have intensively been studied @xcite . in 2011 the new superconductors ca@xmath0(fe@xmath3pt@xmath4as)@xmath0(pt@xmath1as@xmath2 ) ( 1038 phase , space group @xmath7 ) and polymorphic ca@xmath0(feas)@xmath0(pt@xmath8as@xmath2 ) ( 1048-phases , space groups @xmath9 , @xmath10 , @xmath7 ) with critical temperatures up to 35 k were discovered @xcite .
this new class recently expanded by analogous compounds with iridium ( ir1048 ) @xcite and palladium ( pd1038 ) @xcite instead of platinum . due to the presence of the second metal pnictide layer pt@xmath11as@xmath2 ( @xmath12 = 3 , 4 ) next to feas , as well as the low symmetry of these compounds ( space group @xmath7 ) they were initially considered as rather peculiar representatives of the iron arsenide family .
recent low temperature x - ray structural data together with @xmath13sr spectra @xcite as well as neutron diffraction @xcite revealed a lattice distortion and magnetic phase transition at 120 k , proving that ca@xmath0(feas)@xmath0(pt@xmath1as@xmath2 ) is the parent compound of this branch of the iron arsenide family . thereby closely related superconductors like ca@xmath0(fe@xmath3pt@xmath4as)@xmath0(pt@xmath1as@xmath2 ) ( @xmath14 = 14 k ) , ca@xmath0(feas)@xmath0(pt@xmath8as@xmath2 ) ( @xmath14 = 35 k ) and ( ca@xmath15)@xmath0(feas)@xmath0(pt@xmath1as@xmath2 ) ( @xmath14 = 35 k ) can be derived from this common parent by direct , indirect , and electronic doping , respectively @xcite .
[ fig : structure ] depicts the structure of the 1038 parent compound as well as a section of the pt@xmath1as@xmath2 layer .
crystal structure of the 1038 parent compound ca@xmath0(feas)@xmath0(pt@xmath1as@xmath2).,width=340 ] in this paper we report the crystal structures and superconductivity of transition metal doped ca@xmath0(fe@xmath16as)@xmath0(pt@xmath1as@xmath2 ) with @xmath17 = co , ni , and cu with critical temperatures up to 15 k. our results clearly establish that the 1038/1048 superconductors are more than an exotic exception , but another , even structurally more complex , branch of the iron arsenide family .
polycrystalline samples of co- , ni- , and cu - doped calcium platinum iron - arsenides were synthesized from the elements as described in @xcite , and characterized by x - ray powder diffraction using the rietveld method with topas @xcite .
compositions were determined within errors of 10% by x - ray spectroscopy ( edx ) .
superconducting properties were determined using an @xmath18-susceptometer at 1333 hz in the temperature range of 3.5 k to 300 k at 3 oe .
magnetic measurements were additionally performed on a quantum design mpms xl5 squid magnetometer which allowed for measurements with fields up to 50 koe at temperatures between 1.8 k and 300 k. temperature dependent resistivity measurements between 3.5 k and 300 k were carried out using a standard four - probe method .
[ fig : rietveld ] shows x - ray powder patterns with rietveld fits of ca@xmath0(fe@xmath16as)@xmath0(pt@xmath1as@xmath2 ) samples with @xmath17 = co , ni and @xmath19 = 0.1 . the diffraction patterns are completely described with the 1038-phase structure model @xcite and minor amounts of impurity phases feas and ptas@xmath20 .
the high sample quality implies smooth incorporation of cobalt , nickel , and copper into the 1038 structure .
the amount of impurity phases increase at higher doping levels , which indicates solubility limits of @xmath21 0.25 for co and ni , as well as @xmath22 0.12 for cu , respectively , in the 1038-type structure .
thus , the fully substituted compounds ca@xmath0(coas)@xmath0(pt@xmath1as@xmath2 ) , ca@xmath0(nias)@xmath0(pt@xmath1as@xmath2 ) , ca@xmath0(cuas)@xmath0(pt@xmath1as@xmath2 ) and ca@xmath0(feas)@xmath0(ni@xmath1as@xmath2 ) were not accessible by solid state synthesis , although the pd analogue ca@xmath0(feas)@xmath0(pd@xmath1as@xmath2 ) was recently reported @xcite .
[ fig : lattice ] illustrates the dependency of the lattice parameters on co- or ni - substitution from x - ray powder data refinement , revealing similar behaviour of both co- and ni-1038 to the 122-type compounds ba(fe@xmath3co@xmath4)@xmath20as@xmath20 @xcite and ba(fe@xmath3ni@xmath4)@xmath20as@xmath20 @xcite .
the in - plane axes @xmath23 and @xmath24 , as well as the unit cell angles ( not shown ) remain nearly constant within experimental accuracy , whereas @xmath25 decreases by about 0.8% in the range @xmath26 , accompanied with shrinking of the cell volume by approximately 0.8% . at this point
it should be noted , that , even if small , the analogous effects of co- and ni - doping to the 1038-structure may be indicative for a similar 3@xmath27 electron count localized at co and ni when doped to feas layers .
an interesting aspect arises from the chemical similarity of platinum and nickel .
although the targeted position of co when doped into the 1038 structure is the feas layer , the situation is more complicated in the case of ni . in this system
fe / ni mixing is expected , but ni substitution to the pt sites is as well imaginable as ni occupancy at the pt vacancies in the pt@xmath28a@xmath2 layer .
however , modifications of the pt@xmath28a@xmath2 layers by additional ni incorporation were supposed to produce 1048-type impurity phases , or at least changes of the @xmath23,@xmath24 lattice parameters .
susceptibility measurements would be a sensitive probe to detect even minimal traces of the 1048 phase due to its high @xmath6 well above 30 k. however , none of these phenomena have been observed .
moreover , additional edx measurements gave no indication for pt / ni mixing .
all these results clearly indicate that no modifications of the pt@xmath1as@xmath2 layers occur , tantamount with co- , ni- or cu - doping taking place only in the feas layers .
[ fig : sus ] shows @xmath18-susceptibility data of ca@xmath0(fe@xmath16as)@xmath0(pt@xmath1as@xmath2 ) ( @xmath17 = co , ni ) .
critical temperatures of co-1038 reach 15.3 k at the optimal co concentration @xmath19 = 0.075 , whereas ni-1038 maximum @xmath6 settles at 13.4 k for a doping level of @xmath19 = 0.05 .
superconducting volume fractions indicate bulk superconductivity .
in contrast to this finding no superconductivity was detected in ca@xmath0(fe@xmath3cu@xmath4as)@xmath0(pt@xmath1as@xmath2 ) , except of traces at the lowest doping level sythesized of @xmath19 = 0.02 .
[ fig : res ] shows @xmath29-electrical resistivity and low field @xmath29-susceptibility data .
the left panel ( fig .
[ fig : res]a ) displays resistivity measurements performed with the maximum @xmath6 samples ca@xmath0(fe@xmath30co@xmath31as)@xmath0(pt@xmath1as@xmath2 ) and ca@xmath0(fe@xmath32ni@xmath33as)@xmath0(pt@xmath1as@xmath2 ) .
a steep drop to zero resistance is observed for both samples which coincide with critical temperatures from susceptibility measurements , respectively .
the absolute values of the specific resistivity are in the range of poor metals as typical for iron arsenides in the normal state . remarkably , in our case the normal state resistivity increases with decreasing temperature resembling a semiconductor like behaviour .
this temperature dependence of the specific resistivity is different from 1111- and 122-type iron arsenide superconductors , but was also found for the parent compound ca@xmath0(feas)@xmath0(pt@xmath1as@xmath2 ) .
similar results have also been reported for ca@xmath0(fe@xmath3pt@xmath4as)@xmath0(pt@xmath1as@xmath2 ) @xcite .
the right panel ( fig .
[ fig : res]b ) shows the low - field @xmath29-susceptibility of co- and ni-1038 samples with the highest @xmath6 .
almost 100% shielding at low temperatures proves bulk superconductivity , while the meissner - signal is rather small .
doping dependent critical temperatures for both co- and ni-1038 are compiled in fig .
[ fig : phasedia ] .
superconductivity is induced by small transition metal doping to the iron sites .
ca@xmath0(fe@xmath3co@xmath4as)@xmath0(pt@xmath1as@xmath2 ) reveals a dome like @xmath5 dependency , reaching a maximum of 15.3 k for the optimal doping level @xmath19 = 0.075 .
the critical temperature distinctly drops if @xmath19 exceeds 0.075 .
however , full suppression of superconductivity upon high co concentrations could not be archived within the solubility limit of co. samples with a nominal composition of @xmath19 = 0.2 and @xmath19 = 0.3 show only traces of superconductivity which may come from inhomogeneously distributed cobalt . therewith the superconducting dome of co-1038 is remarkably similar to the co - doped 122-compounds ba(fe@xmath3co@xmath4)@xmath20as@xmath20 @xcite .
the ni-1038 compounds reveal narrower dependency of the critical temperatures from the substitution level , featuring a maximum @xmath6 of 13.4 k at @xmath19 = 0.05 .
this finding is in line with the additional electron of ni with respect to co , giving rise of an increased electron doping at same transition metal substitution @xcite .
@xmath6 rapidly decreases at higher ni concentrations until superconductivity is completely suppressed at 17 % ni .
thus , ca@xmath0(fe@xmath3ni@xmath4as)@xmath0(pt@xmath1as@xmath2 ) reveals very similar properties than its homologue ca@xmath0(fe@xmath3pt@xmath4as)@xmath0(pt@xmath1as@xmath2 ) . in this context ni-1038
is supposed to be a more suitable system to study its properties in the overdoped regime due to the higher solubility of ni .
@xmath5 phase diagrams of ca@xmath0(fe@xmath3co@xmath4as)@xmath0(pt@xmath1as@xmath2 ) and ca@xmath0(fe@xmath3ni@xmath4as)@xmath0(pt@xmath1as@xmath2).,width=340 ]
in conclusion we have demonstrated bulk superconductivity in the 1038 compounds ca@xmath0(fe@xmath16as)@xmath0(pt@xmath1as@xmath2 ) doped by @xmath17 = co or ni , with critical temperatures up to 15.3 k in co-1038 ( @xmath19 = 0.075 ) and 13.4 k in ni-1038 ( @xmath19 = 0.05 ) , respectively .
superconducting properties in both compounds were evidenced by @xmath18- and @xmath29-susceptibility as well as @xmath29-resistivity data .
moreover no superconductivity was evident in cu - doped samples ca@xmath0(fe@xmath3cu@xmath4as)@xmath0(pt@xmath1as@xmath2 ) .
the dependency of @xmath6 of the substitution level @xmath19 reveals similar behaviour than in known directly doped 122-type iron arsenides ba(fe@xmath16)@xmath20as@xmath20 with @xmath17 = co , ni , cu , whereas the comparatively narrow superconducting dome of ni-1038 is indicative for an increased electron doping contribution of ni with respect to co-1038 .
our results clearly show the close resemblance of calcium platinum iron arsenides to other iron arsenide compounds , giving evidence that established doping methods to induce superconductivity are abundantly applicable also to more complex systems like the 1038 and 1048 materials .
this work was financially supported by the german research foundation ( dfg ) within the priority program spp1458 .
mcguire , a.d .
christianson , a.s .
sefat , b.c .
sales , m.d .
lumsden , r.y .
jin , e.a .
payzant , d. mandrus , y.b .
luan , v. keppens , v. varadarajan , j.w .
brill , r.p .
hermann , m.t .
sougrati , f. grandjean , and g.j .
long , phys .
b 78 ( 2008 ) , p. 094517 . | we report the successful substitution of cobalt , nickel , and copper for iron in the 1038 phase parent compound ca@xmath0(feas)@xmath0(pt@xmath1as@xmath2 ) yielding ca@xmath0(fe@xmath3co@xmath4as)@xmath0(pt@xmath1as@xmath2 ) , ca@xmath0(fe@xmath3ni@xmath4as)@xmath0(pt@xmath1as@xmath2 ) , and ca@xmath0(fe@xmath3cu@xmath4as)@xmath0(pt@xmath1as@xmath2 ) , respectively .
superconductivity is induced in co and ni doped compounds reaching critical temperatures up to 15 k , similar to known pt substituted ca@xmath0(fe@xmath3pt@xmath4as)@xmath0(pt@xmath1as@xmath2 ) , whereas no superconductivity was detected in ca@xmath0(fe@xmath3cu@xmath4as)@xmath0(pt@xmath1as@xmath2 ) .
the obtained @xmath5 phase diagrams are very similar to those of other iron arsenide superconductors indicating rather universal behaviour despite the more complex structures of the 1038-type compounds , where the physics is primarily determined by the feas layer .
superconductivity , iron arsenides , doping , crystal structure , platinum , cobalt , nickel , copper |
the study of the electroweak interaction between leptons and hadrons has been a challenging topic ever since the standard model was proposed by glashow , salam and weinberg .
this model predicts the coupling of the electroweak currents to leptons and quarks in terms of the electric charge @xmath6 and the weinberg angle @xmath7 . in particular , the weak neutral current is mediated by the exchange of the @xmath1 gauge boson . at
low and moderate momentum transfer its contribution is suppressed relative to photon exchange by a factor @xmath8 , where @xmath9 is the four - momentum of the exchanged boson and @xmath10 the mass of the @xmath1 . the interference term beween photon and @xmath1 exchange
contains a parity violating ( pv ) effect , which becomes visible as an asymmetry by scattering polarized electrons with helicity along the direction of the beam @xmath11 or opposite to it @xmath12 , @xmath13 the quantity @xmath14 in this equation should represent an inclusive cross section . in the case of a coincidence experiment , e.g. @xmath15 , there also appear parity conserving asymmetries due to the electromagnetic interaction .
specifically , the so - called 5th response function will generate a background of helicity - dependent contributions , which are parity conserving and , therefore , generally larger than the pv effects by several orders of magnitude .
the pioneering experiment to measure pv asymmetries has been performed at slac @xcite by deep inelastic electron scattering on a deuterium target .
this experiment has been followed by investigations of pv quasifree scattering off @xmath16 at mainz @xcite and pv elastic scattering off @xmath17 at mit / bates @xcite . in the latter case
the momentum transfer was only @xmath18 , leading to the tiny asymmetry @xmath19 .
these experiments were originally devised to test the standard model , in particular to measure the weinberg angle .
the value obtained for this angle by high - energy experiments was confirmed within the error bars of about 10% . with the advent of new electron accelerators like cebaf , mami and mit / bates , having high intensity ,
high duty - factor and a polarized beam , the quality of the data can be considerably improved . moreover , with @xmath20 known to 4 decimal places and the standard model firmly established , the strategy of the new experiments will be redirected towards an improved understanding of the structure of the nucleon .
in particular , pv elastic electron scattering will provide information on the strangeness content of the nucleon .
three such experiments are being planned with the 4 gev cebaf beam @xcite and there are also proposals to measure the electric radius and the magnetic moment of strange quark pairs at mami @xcite and mit / bates @xcite in the region of 1 gev . assuming that the present experimental activities will soon yield novel information on the ground state of the nucleon , we deem it appropriate to study the effect of pv interactions for inelastic processes .
it is therefore the aim of this contribution to investigate the pv asymmetries for electroproduction of pions .
previous calculations of nath et al.@xcite and jones et al.@xcite at medium energies and of cahn et al.@xcite at higher energies have been based on the production of stable @xmath3 isobars .
however , there should be non - negligible background contributions interfering coherently with the resonance .
estimates for such contributions have been reported earlier by ishankuliev et al.@xcite and by li et al.@xcite .
in particular , li et al.@xcite have also considered pv effects in the hadronic wave function , i.e. at the @xmath5 vertex .
it turns out , however , that such contributions are of the order of @xmath21 , much smaller than the expected asymmetries due to pv interferences in the electroweak interaction , which are of the order of @xmath22 . in the following section we will briefly review the kinematics , the cross section and the decomposition of both vector and axial currents into the invariant amplitudes .
3 presents our model using effective lagrangians .
it includes a background of born terms with pseudovector @xmath5 coupling and the @xmath3 isobar treated as a rarita - schwinger field with phenomenological @xmath23 transition currents . assuming that the hadronic currents are dominated by @xmath24 and @xmath25 quarks , the weak neutral current may be decomposed in terms of strong isospin @xcite and related to hadronic currents . the matrix elements
are then decomposed into invariant amplitudes @xcite and according to their isospin structure .
the numerical results are presented in sect .
4 . as a test we have first calculated the inclusive electromagnetic cross section .
the experimental data for this process @xcite can be well reproduced by the model .
we present our results for pv asymmetries as function of excitation energy , scattering angle and momentum and compare our asymmetries with previous calculations .
finally , we give a summary and some conclusions in sect .
we consider the reaction shown in fig .
@xmath26 and @xmath27 are the 4-momenta of the nucleon in the initial and the final state , respectively , the produced pion has momentum @xmath28 .
the momentum transfer @xmath29 is the difference of the 4-momenta of the ingoing and outgoing electron @xmath30 , with @xmath31 .
the spins of nucleons and electrons in the initial and final states are denoted by @xmath32 , @xmath33 , @xmath34 , and @xmath35 .
if not stated otherwise , the kinematical variables are evaluated in the laboratory system , which is defined by @xmath36 .
furthermore , the mandelstam variables @xmath37 are equivalent to the following set of lorentz invariant kinematic variables @xcite : @xmath38 with @xmath39 .
the latter set of variables will be used from now on . besides , the coincidence experiment is characterized by the angles shown in fig .
the polar angle @xmath40 is the angle between the pion 3-momentum @xmath41 and the momentum transfer @xmath42 , the azimuthal angle @xmath43 is the angle between the scattering plane , defined by @xmath44 and @xmath45 , and the reaction plane , spanned by @xmath46 and @xmath47 .
furthermore , the scattering angle of the electron is @xmath48 . in the following sections we will be interested in the inclusive cross section ,
i.e. the cross section has to be integrated over the pion angles .
the differential cross section @xcite @xmath49^{\frac{1}{2 } } } \prod^{n_f}_{j = 1 } \frac { d^3 \vec{p}_j } { ( 2 \pi)^3 2 e_j } \mid { \cal m}_{fi } \mid^2 , \ ] ] has to be integrated over the momenta of the pion and the final nucleon . in the one boson exchange approximation and for momentum transfers @xmath50 ,
the invariant matrix element is of the form @xmath51 in this expression , the photon couples to the electromagnetic current of the nucleon , @xmath52 , and the @xmath1 gauge boson to the familiar combination of vector and axial currents @xmath53 .
the corresponding currents of the electron are denoted by @xmath54 and @xmath55 .
these currents may be combined to @xmath56 and @xmath57 , the hadronic and leptonic tensors .
the tensors may be decomposed into symmetric and antisymmetric parts .
if we neglect the electron mass , the leptonic tensor is @xmath58 with @xmath59 and @xmath60 .
a reasonable definition for the hadronic tensor is @xmath61 \ , .
\nonumber\end{aligned}\ ] ] the hadronic matrix elements have the structure @xmath62 where @xmath63 is an abbreviation for the leptonic matrix element .
these matrix elements may be decomposed in isospace according to @xmath64 \chi^{\ } _ { i } } _ { = : i^- } { \cal m}^- \ , , \ ] ] with @xmath65 and @xmath66 the isospinors of the nucleon in the final and initial state , respectively , and @xmath67 the isospin matrix characterizing the produced pion . furthermore
, the hadronic transition currents are decomposed into invariant amplitudes @xcite @xmath68 the amplitudes @xmath69 depend on the three independent variables @xmath70 , and @xmath71 , and the superscript @xmath72 refers to the isospin decomposition ( [ iso ] ) .
a reasonable choice for the vector @xmath73 and axial vector @xmath74 operators is @xcite : @xmath75 & \zeta_4^v = 1 \\ m_\mu^5 = - i \gamma_5 ( q_\mu \ ; k^\pi \cdot q - q^2 k^\pi_\mu ) & \zeta_5^v = -1 \\ m_\mu^6 = i \gamma_5 ( q_\mu { / \!\!\ !
q } - q^2 \gamma_\mu ) & \zeta_6^v = -1 \label{ia_v } \end{array}\ ] ] and @xmath76 the vector current operators are explicitly gauge invariant by construction , @xmath77 .
the constants @xmath78 and @xmath79 specify the behavior of the invariant amplitudes under the crossing transformation @xmath80
@xcite , @xmath81 the vector and axial currents of ( [ ia ] ) may be decomposed into a multipole series following the work of adler @xcite .
the leading @xmath82-wave contributions to the currents are @xmath83 in these equations all variables have to be expressed in the _ cm _ frame , in particular the components of the 4-momentum transfer , @xmath84 , the polarization vector of the virtual photon , @xmath85 and the nucleon spin @xmath86 .
furthermore , @xmath87 and @xmath88 is the polarization vector transverse to the direction of the virtual photon .
the ellipses denote @xmath89-waves and higher multipoles .
since the polarization vector @xmath90 is proportional to the transition current of the electron , @xmath91 or @xmath92 , the four - product @xmath93 vanishes exactly for the ( conserved ) vector current .
however , it can also be safely neglected in the case of the axial vector , because the divergence of the axial current is proportional to the mass of the electron .
the threshold values of the @xmath82-wave multipoles have been predicted by general principles following the arguments of low energy theorems ( let ) .
since these theorems assume lorentz and gauge invariance and the pcac relation , they should be obeyed by our model , too .
however , there have been recently reported large modifications to let due to loop corrections . concerning the vector current these corrections
are particular large for neutral pion photoproduction at threshold @xcite , but do not play an important role for the inclusive cross section , which is dominated by charged pion production .
however , it is interesting that two of the multipoles for the axial vector have contributions containing the @xmath5 @xmath14-term .
according to @xcite these are @xmath94 - \frac{a^+ f_\pi}{m_\pi } + o ( m_{\pi})\ , , \\ { \cal h}_{0+}^{(+ ) } & = & \frac{a^+ f_{\pi}}{q^2 - m_{\pi}^2 } + \frac{\sigma ( 0 ) - \sigma ( q^2 - m_{\pi}^2)}{12 \pi f_{\pi } ( q^2 - m_{\pi}^2 ) } + o ( m_{\pi})\ , , \nonumber\end{aligned}\ ] ] where @xmath95 mev is the pion decay constant and @xmath96 the isospin even @xmath82-wave @xmath5-scattering length . as we see from ( [ mulzer ] ) and the above considerations , the multipole @xmath97 does not contribute in the limit of a vanishing lepton mass .
the longitudinal multipole @xmath98 , however , contributes and its threshold value is dominated by the @xmath14-term .
unfortunately , it appears in combination with the vector coupling of the @xmath1 at the vertex of the electron , i.e. this interesting term is suppressed by a factor @xmath99 . in this subsection
the explicit structure of the lorentz tensors will be given .
the electromagnetic lepton tensor has the familiar form @xmath100 where @xmath101 denotes the helicity of the incoming electron .
the antisymmetric part can be omitted for unpolarized nucleons , because the electromagnetic hadronic tensor is symmetric in this case .
the interference tensor for the lepton is @xmath102 where @xmath103 and @xmath104 denote the weak neutral current couplings of the electron @xmath105 using current conservation , the electromagnetic tensor of the nucleon has the lorentz structure @xmath106 the symmetric part of the hadronic interference tensor @xmath107 has the same structure as @xmath108 .
the corresponding antisymmetric part is @xmath109 as the axial currents are not conserved , there are in principle additional terms proportional to @xmath110 or @xmath111 .
since these terms vanish after contraction with the leptonic tensor ( [ lept ] ) , they have been omitted in ( [ wia ] ) right away .
the structure functions @xmath112 can be expressed in terms of the invariant amplitudes ( [ ia ] ) in a straightforward way .
the result of this calculation is given in appendix [ strf ] .
in this section we will present the phenomenological model that is used to calculate the hadronic currents .
it contains contributions of the born terms and a phenomenological description of the @xmath2 resonance .
the background of the born terms contains both vector current and axial vector current contributions . since we neglect the strangeness of the nucleon , the weak vector current differs from the electromagnetic current only by a coupling constant .
accordingly , we have to calculate the feynman diagrams of fig . [ fig3 ] .
while all diagrams contribute to the vector current , the pion pole diagram does not contribute to the axial current .
for the calculation of the feynman diagrams we use the following interaction lagrangians @xmath113 \psi \ , , \nonumber \\ { \cal l}_{v^\mu \pi \pi } & = & -e\,\xi_v^{i=1 } f_\pi ( \vec{\phi } \times \partial_\mu \vec{\phi } ) _ 3 v^\mu \ , , \nonumber \\
{ \cal l}_{v^\mu n n \pi } & = & e \ , \xi_v^{i=1 } \frac{f_{\pi n n}}{g_a m_\pi } g_a \bar{\psi } ( \vec{\tau } \times \vec{\phi } ) _ 3 \gamma_\mu \gamma_5 \psi \ , v^\mu , \ , \nonumber \\ { \cal l}_{a^\mu n n } & = & - e \ , \xi_a^{i=1 } \bar{\psi } \left [ g_a \gamma_\mu a^\mu + g_p \frac{i \partial_\mu}{2 m } a^\mu \right ] \gamma_5 \frac{\tau_3}{2 } \psi \ , , \nonumber \\
{ \cal l}_{a^\mu n n \pi } & = & e \ , \xi_a^{i=1 } \frac{f_{\pi n
n } } { g_a m_\pi } \bar{\psi } ( \vec{\tau } \times \vec{\phi } ) _ 3 \left ( f_1^v \gamma_\mu a^\mu - \kappa_v f_2^v \frac { \sigma_{\mu\nu}}{2 m } \partial^\nu a^\mu \right ) \psi \ , , \nonumber\end{aligned}\ ] ] with @xmath114 .
the isospin factors @xmath115 follow from the decomposition of the corresponding quark current operators according to strong isospin @xcite . for the electromagnetic current
these factors are equal to unity , for the weak neutral current we have @xmath116 note that there appears no isoscalar contribution to the axial vector current , i.e. @xmath117 , because strange quarks have been neglected .
the form factors @xmath118 , @xmath119 , @xmath120 , @xmath121 and @xmath122 are functions of momentum transfer @xmath71 .
since our calculation will be performed at relatively small momentum transfer , we have used simple dipole forms for the sachs form factors , with the assumption @xmath123 .
this insures gauge invariance without additional gauge terms . with standard methods and neglecting the lepton mass terms
, we obtain the invariant matrix elements @xmath124 u(\vec{p_i})\ , , \\
{ \cal m}_u & = & \xi_v^{i=1 } \frac{f_{\pi n n}}{m_\pi } \bar{u}(\vec{p_f } ) \bigg [ \left ( \tilde i_u^d { / \!\!\ ! \epsilon } + i \tilde i_u^p \frac{\sigma_{\mu\nu}}{2 m } q^\nu \epsilon^\mu \right ) \frac{{/ \!\!\!\!\!\ , p_i } - { / \!\!\ ! k_\pi } + m}{u - m^2 } \gamma_5 { / \!\!\ ! k_\pi } \bigg ] u(\vec{p_i})\ , , \nonumber \\ { \cal m}_t & = & \xi_v^{i=1 } \frac{f_{\pi n n}}{m_\pi } \frac{\epsilon \cdot ( 2 k_\pi - q)}{t - m_\pi^2 } 2 m i_t f_\pi \bar{u}(\vec{p_f } ) \gamma_5 u(\vec{p_i } ) \ , , \nonumber \\ { \cal m}_{c } & = & \xi_v^{i=1 } \frac{f_{\pi n n}}{g_a m_\pi } i_t g_a \bar{u}(\vec{p_f } ) { / \!\!\ !
\epsilon } \gamma_5 u(\vec{p_i } ) \ , , \nonumber
\\ { \cal m}_s^{5 } & = & -\xi_a^{i=1 } \frac{f_{\pi n n}}{m_\pi } \bar{u}(\vec{p_f } ) \left [ { / \!\!\ ! k_\pi } \frac{{/ \!\!\!\!\!\ , p_i } + { / \!\!\ !
q } - m}{s - m^2 } \tilde i_s^a { / \!\!\ ! \epsilon } \right ] u(\vec{p_i})\ , , \nonumber \\ { \cal m}_u^{5 } & = & -\xi_a^{i=1 }
\frac{f_{\pi n n}}{m_\pi } \bar{u}(\vec{p_f } ) \left [ \tilde i_u^a { / \!\!\ ! \epsilon } \frac{{/ \!\!\!\!\!\ , p_i } - { / \!\!\ ! k_\pi } - m}{u - m^2 } { / \!\!\ ! k_\pi } \right ] u(\vec{p_i})\ , , \nonumber \\ { \cal m}_{c}^{5 } & = & \xi_a^{i=1 } \frac{f_{\pi n n}}{g_a m_\pi } i_t \bar{u}(\vec{p_f } ) \left [ \gamma_\mu f_1^v + i
\frac{f_2^v}{2 m } \sigma_{\mu\nu } q^\nu \right ] \epsilon^\mu u(\vec{p_i})\ , . \nonumber\end{aligned}\ ] ] note that there do not appear any induced pseudoscalar terms , because the lepton mass terms have been neglected .
the electromagnetic matrix element is the sum of the first 4 terms in ( [ mat_b ] ) , corresponding to s , u and t channel pole terms and the contact term(c ) .
the matrix element of the weak neutral current is the corresponding sum of the combinations @xmath125 .
the following abbreviations for the isospin matrix elements have been used in ( [ mat_b ] ) : @xmath126 , \nonumber \\
\tilde i_{(s , u)}^p & = & \frac{1}{2 } \left [ \frac{\xi_v^{i=0}}{\xi_v^{i=1 } } f_2^s i^0 + f_2^v ( i^+ \pm
i^- ) \right ] , \nonumber \\
\tilde i_{(s , u)}^a & = & \frac{1}{2 } \left [ g_a ( i^+ \pm i^- ) \right ] \ , , \nonumber\end{aligned}\ ] ] with @xmath127 as defined in ( [ iso ] ) .
the invariant amplitudes obtained from ( [ mat_b ] ) are identical to the results of @xcite .
they can be found in appendix [ iam_b ] . in the kinematical region between threshold and about 400 mev excitation energy ,
the dominant resonant contributions are due to the @xmath2 .
we treat the @xmath3 as a rarita - schwinger field and use the on - shell form of the propagator modified by a phenomenological constant width , @xmath128 @xmath129 is obtained as the real part of the @xmath3 pole @xcite . the width @xmath130 is fitted to the experimental data for pion electroproduction @xcite , which is close to the imaginary part of the pole @xcite , @xmath131 .
the feynman diagrams to be calculated are shown in fig .
we use the phenomenological @xmath4 transition currents @xcite @xmath132 u(\vec{p})\ , , \\ { j}_{\mu}^{nc } & = & \bar{u}^{\lambda}(\vec{p ' } ) \left [ \left ( \frac{c_{3v}^{z}}{m } \gamma^{\nu } + \frac{c_{4v}^{z}}{m^2 } { p'}^{\nu } + \frac{c_{5v}^{z}}{m^2 } p^{\nu } \right ) \vphantom{\frac{1}{2 } } ( g_{\lambda\mu } g_{\rho\nu } - g_{\lambda\rho } g_{\mu\nu } ) q^{\rho } \gamma_5 + c_{6v}^z g_{\lambda\mu } \gamma_5 \right .
\nonumber \\ & & + \left .
\left ( \frac{c_{3a}^{z}}{m } \gamma^{\nu } + \frac{c_{4a}^{z}}{m^2 } { p'}^{\nu } \right ) ( g_{\lambda\mu } g_{\rho\nu } - g_{\lambda\rho } g_{\mu\nu } ) q^{\rho } + c_{5a}^z g_{\lambda\mu } + \frac{c_{6a}^{z}}{m^2 } p_{\lambda } q_{\mu } \right ] u(\vec{p } ) \ , .
\nonumber\end{aligned}\ ] ] the form factors @xmath133 to @xmath134 are taken from nath et al.@xcite , with a @xmath71 dependence as in @xcite .
the @xmath135 vertex has the effective form @xcite @xmath136 with ( [ d_prop])-([pind ] ) we obtain the invariant matrix elements for s channel @xmath3 production , @xmath137 \gamma_5 \epsilon^\mu u(\vec{p_i } ) \ , , \nonumber \\ { \cal m}^{nc}_{s \delta } &
= & \xi_v^{i=1 } { \cal m}^{em}_{s \delta } \vphantom{\frac{1}{2 } } , \nonumber \\ { \cal m}^{5 nc}_{s \delta } & = & -\xi_a^{i=1 } \frac{f_{\pi n \delta}}{m m_{\pi } } \bar{u}(\vec{p_f } ) k^{\pi}_{\sigma } \delta^{\sigma \lambda } \bigg [ c_3^{a } ( q_\lambda \gamma_\mu - g_{\lambda \mu } { / \!\!\ !
q } ) + \frac{c_4^{a}}{m } ( q_\lambda(p_f + k_\pi)_{\mu } \nonumber \\ & & -q \cdot ( p_f + k_\pi ) g_{\lambda \mu } ) - c_5^{a } m g_{\lambda \mu } \bigg ] \epsilon^\mu u(\vec{p_i } ) \ , .
\nonumber\end{aligned}\ ] ] the corresponding u channel amplitudes are constructed by using the crossing properties of the invariant amplitudes ( [ cross ] ) .
these invariant matrix elements can now be decomposed into invariant amplitudes as in ( [ ia ] ) .
the result of this calculation is listed in appendix [ iam ] .
as a first test of the model we have calculated the inclusive electromagnetic cross section . the experimental data for this process @xcite could be reproduced within about 5% over the whole energy region from threshold to about 400 mev excitation energy .
[ fig5 ] shows the asymmetry for the proton and the neutron in various kinematical situations .
the two upper figures compare the asymmetry for proton and neutron as function of the photon equivalent energy , @xmath138 the _ lab _ energy necessary to excite a hadronic system with _
cm _ energy w. at small excitation energies the asymmetry
is essentially given by background contributions , whereas the @xmath3 resonance dominates at the higher energies . while the individual contributions vary strongly , the coherent sum is essentially constant .
in particular , the resonance structure of the inclusive cross sections is essentially wiped out when the asymmetries are calculated . as demonstrated in the two lower figures of fig .
[ fig5 ] , this effect is quite independent of the kinematics except for a trivial dependence on the momentum transfer . for the proton this @xmath139 dependence
is well described by the simple estimate @xmath140 where @xmath141 is the fermi coupling constant . in the case of the neutron ,
the asymmetry is somewhat smaller , at about 80% of the proton case , showing a slight enhancement in the resonance peak . in fig .
[ fig6 ] the asymmetry is shown at an incident electron energy of @xmath142 and at @xmath143 , directly in the resonance peak . in this kinematical situation ,
the asymmetry is strongly dominated by the @xmath3 , the background contributions cancel against the interference terms ( upper panel ) .
therefore , the neglect of background terms as in the work of nath et al.@xcite is well justified . as is shown in the lower panel of this figure
, we are able to reproduce those results except for an overall scaling factor of about 90 % . in the work of cahn et al.@xcite the vector coupling between the @xmath1 and the electron has been neglected .
this corresponds to @xmath144 and produces a somewhat lower curve with a flat distribution in @xmath139 .
we have also compared our calculation to the results of li et al .
@xcite and find a reasonable agreement .
however , the calculation of ishankuliev et al .
@xcite disagrees with all others by a factor @xmath145 seemingly due to a wrong coupling constant .
preliminary studies of the @xmath5 @xmath14-term ( [ mul ] ) have shown only small effects of this physically interesting quantity on the asymmetry . unfortunately , the coupling of the @xmath1 to the axial current of the hadron appears together with the vector coupling to the electron which is strongly suppressed by the value of the weinberg angle .
in addition , the @xmath82-wave multipole @xmath98 contributes to the asymmetry only by interference with the electromagnetic @xmath89-waves . in the usual conventions this contribution is proportional to @xmath146 . as a consequence
, the multipole @xmath98 is further suppressed near threshold .
the aim of this work has been to investigate parity violating ( pv ) contributions to pion electroproduction .
only pv effects due to interference between @xmath0 and @xmath1 exchange have been considered .
pv effects in the strong interaction , as discussed by li et al.@xcite , have been neglected .
the appropriate observable to study these effects is the asymmetry defined in ( [ def_a ] ) .
a simple phenomenological model with effective lagrangian densities has been constructed that allows for the calculation of the asymmetry in the kinematical region from pion threshold to @xmath2 resonance .
the model is fully relativistic , fulfils the crossing symmetry and is gauge invariant due to our simple choice of the form factors .
the invariant amplitudes for the nonresonant contributions respect pcac .
the @xmath3 resonance is treated as a rarita - schwinger field with phenomenological transition currents . for the propagator ,
the on - shell form with the @xmath3 pole of the @xmath5 scattering matrix is used .
the nonresonant contributions are created by the usual pseudovector @xmath5 coupling and phenomenological transition currents .
the weak neutral vector current is traced back to the electromagnetic one by decomposing the quark currents according to strong isospin .
for the axial current , a contact term has been introduced in addition to the @xmath147 and @xmath24 channel nucleon born terms .
this contact term is necessary to fulfil the low energy theorem of adler @xcite for the nonresonant contributions .
the calculated asymmetry is nearly constant as function of excitation energy .
it increases linearly with the square of the four - momentum transfer , @xmath148 for the proton , and has about 80% of that value for the neutron .
previous calculations for the excitation of the @xmath3 isobar could be reproduced reasonably well , while our results are at variance with some of the earlier work on the background contributions . in conclusion
we find the following results : * the asymmetry grows linearly with the momentum transfer @xmath139 and is nearly independent of the excitation energy .
* the expected asymmetries for pion electroproduction are comparable with the asymmetries found for elastic electron scattering . * because of the value of the weinberg angle , the vector coupling of the @xmath1 at the electron vertex is suppressed . as a consequence ,
the asymmetry is dominated by the hadronic vector current , the contributions of the hadronic axial currents being of the order of 10 - 20 % only . *
a precision measurement of the asymmetries is potentially an independent experiment to determine the important @xmath5 @xmath14-term , which appears in the @xmath82-wave multipole @xmath98 of the hadronic axial current .
unfortunately , the @xmath14-term will be difficult to measure , because the hadronic axial current is suppressed and furthermore , the multipole @xmath98 appears in the asymmetry only by its interference with @xmath89-wave multipoles .
finally , we would like to comment on some possible improvements of the calculations : * the simple superposition of born and resonance terms violates the watson theorem .
it would be necessary to perform a multipole decomposition and to unitarize at least the multipoles carrying the phase of the @xmath149 resonance .
* we have neglected the contribution of the strange sea to the hadronic transition currents .
taking account of such effects would require additional , as yet undetermined form factors .
however , it is unlikely that pion production would be strongly modified by strangeness degrees of freedom .
* though we have not found large contributions of the @xmath14-term to the asymmetry , this matter deserves more systematical studies .
clearly , any further independent measurement of this important quantity would be invaluable .
the amplitudes @xmath150 for the electromagnetic current and @xmath151 for the weak neutral current have the following isospin decomposition for born terms and @xmath3 contributions : @xmath152 + * structure functions for the interference tensor * + + * symmetric part * + @xmath153 \nonumber \\ \nonumber \\ w_3^{i } & = & - m^2 \big [ { { \mathrm{re } } } ( v_2^ * \tilde v_2 ) ( \nu_b -\nu)^2 ( 4 m \nu_b + m_\pi^2 + q^2 ) \nonumber\\ & - & { { \mathrm{re } } } ( v_3^ * \tilde v_3 ) ( 4 m^2 ( \nu_b -\nu)^2 + q^2 ( 4 m \nu_b + m_\pi^2 -4 m^2 ) ) \nonumber \\ & + & { { \mathrm{re } } } ( v_4^ * \tilde v_4 ) ( 4 m^2 ( \nu_b -\nu)^2 -q^2 ( 4 m \nu + m_\pi^2 4 m^2 ) ) + { { \mathrm{re } } } ( v_5^ * \tilde v_5 ) q^4 ( 4 m \nu_b + m_\pi^2 + q^2 ) \nonumber \\ & + & ( { { \mathrm{re } } } ( v_1^ * \tilde v_2 ) + { { \mathrm{re } } } ( v_2^ * \tilde v_1 ) ) 2 m ( 2 m ( \nu_b -\nu)^2 + q^2(\nu_b -\nu ) ) \nonumber \\ & - & ( { { \mathrm{re } } } ( v_1^ * \tilde v_3 ) + { { \mathrm{re } } } ( v_3^ * \tilde v_1 ) ) 2 m q^2 + ( { { \mathrm{re } } } ( v_1^ * \tilde v_4 ) + { { \mathrm{re } } } ( v_4^ * \tilde v_1 ) ) 2 m q^2 \nonumber \\ & - & ( { { \mathrm{re } } } ( v_1^ * \tilde v_5 ) + { { \mathrm{re } } } ( v_5^ * \tilde v_1 ) ) q^2 ( 2 m ( \nu_b -\nu ) + q^2 ) -({{\mathrm{re } } } ( v_2^ * \tilde v_3 ) + { { \mathrm{re } } } ( v_3^ * \tilde v_2 ) ) 4 m^2 q^2 ( \nu_b -\nu ) \nonumber \\ & - & ( { { \mathrm{re } } } ( v_2^ * \tilde v_5 ) + { { \mathrm{re } } } ( v_5^ * \tilde v_2 ) ) 2 m q^2 ( \nu_b -\nu ) ( 4 m \nu_b + m_\pi^2 + q^2 ) \nonumber \\ & + & ( { { \mathrm{re } } } ( v_2^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_2 ) ) 4 m^2 q^2 ( \nu_b -\nu ) + ( { { \mathrm{re } } } ( v_3^ * \tilde v_4 ) + { { \mathrm{re } } } ( v_4^ * \tilde v_3 ) ) q^2 ( 2 m(\nu_b + \nu ) + m_\pi^2 ) \nonumber \\ & + & ( { { \mathrm{re } } } ( v_3^ * \tilde v_5 ) + { { \mathrm{re } } } ( v_5^ * \tilde v_3 ) ) 2 m q^4 -({{\mathrm{re } } } ( v_3^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_3 ) ) q^2 ( 2 m(\nu_b -\nu ) + q^2 ) \nonumber \\ & + & ( { { \mathrm{re } } } ( v_4^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_4 ) ) q^2 ( 2 m(\nu_b -\nu ) + q^2 ) -({{\mathrm{re } } } ( v_5^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_5 ) ) 2 m q^4 \big ] \nonumber \\ \nonumber \\ w_4^{i } & = & m^2 \big [ { { \mathrm{re } } } ( v_1^ * \tilde v_1 ) ( 4 m \nu + m_\pi^2 - q^2 ) - { { \mathrm{re } } } ( v_2^ * \tilde v_2 ) 4 m^2 \nu_b^2 ( 4 m \nu_b + m_\pi^2 + q^2 ) \nonumber \\ & + & { { \mathrm{re } } } ( v_4^ * \tilde v_4 ) ( 4 m^2(m_\pi^2 -(\nu + \nu_b)^2 ) + m_\pi^2 ( q^2 - 4 m \nu ) ) - { { \mathrm{re } } } ( v_5^ * \tilde v_5 ) 4 m^2 \nu_b^2 ( 4 m \nu_b + m_\pi^2 + q^2 ) \nonumber \\ & + & { { \mathrm{re } } } ( v_6^ * \tilde v_6 ) ( 4 m^2 ( q^2 + \nu^2 - \nu_b^2 ) - q^2 ( 4 m ( \nu + \nu_b ) + m_\pi^2 ) ) \nonumber \\ & - & ( { { \mathrm{re } } } ( v_1^ * \tilde v_2 ) + { { \mathrm{re } } } ( v_2^ * \tilde v_1 ) ) 2 m \nu_b ( 2 m ( \nu_b - \nu ) + q^2 ) -({{\mathrm{re } } } ( v_1^ * \tilde v_3 ) + { { \mathrm{re } } } ( v_3^ * \tilde v_1 ) ) 4 m^2 \nu_b \nonumber \\ & - & ( { { \mathrm{re } } } ( v_1^ * \tilde v_4 ) + { { \mathrm{re } } } ( v_4^ * \tilde v_1 ) ) 2 m m_\pi^2 -({{\mathrm{re } } } ( v_1^ * \tilde v_5 ) + { { \mathrm{re } } } ( v_5^ * \tilde v_1 ) ) 2 m \nu_b
( 2 m ( \nu_b - \nu ) + q^2 ) \nonumber \\ & - & ( { { \mathrm{re } } } ( v_1^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_1 ) ) 4 m^2 \nu -({{\mathrm{re } } } ( v_2^ * \tilde v_3 ) + { { \mathrm{re } } } ( v_3^ * \tilde v_2 ) ) 8 m^3 \nu_b^2 \nonumber \\ & - & ( { { \mathrm{re } } } ( v_2^ * \tilde v_5 ) + { { \mathrm{re } } } ( v_5^ * \tilde v_2 ) ) 4 m^2 \nu_b^2 ( 4 m \nu_b + m_\pi^2
+ q^2 ) + ( { { \mathrm{re } } } ( v_2^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_2 ) ) 8 m^3 \nu_b^2 \nonumber \\ & + & ( { { \mathrm{re } } } ( v_3^ * \tilde v_4 ) + { { \mathrm{re } } } ( v_4^ * \tilde v_3 ) ) 2 m \nu_b ( 2 m ( \nu_b - \nu ) + m_\pi^2 ) -({{\mathrm{re } } } ( v_3^ * \tilde v_5 ) + { { \mathrm{re } } } ( v_5^ * \tilde v_3 ) ) 8 m^3 \nu_b^2 \nonumber \\
& + & ( { { \mathrm{re } } } ( v_3^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_3 ) ) 2 m
\nu_b ( 4 m^2 -2 m ( \nu_b - \nu ) -m_\pi^2 ) \nonumber \\ & + & ( { { \mathrm{re } } } ( v_4^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_4 ) ) 2 m ( ( 2 m \nu_b + q^2)(\nu+ \nu_b ) + \nu m_\pi^2 ) \nonumber \\ & + & ( { { \mathrm{re } } } ( v_5^ * \tilde v_6 ) + { { \mathrm{re } } } ( v_6^ * \tilde v_5 ) ) 8 m^3 \nu_b^2 \big ] \nonumber\end{aligned}\ ] ] * antisymmetric part * @xmath154 \nonumber \\ \nonumber \\ w_{6}^i & = & 2 m^3 \big [ { { \mathrm{re } } } ( v_1^ * \tilde a_1 ) 2(\nu_b -\nu ) + { { \mathrm{re } } } ( v_3^ * \tilde a_1 ) 4 m \nu_b - { { \mathrm{re } } } ( v_3^ * \tilde a_4 ) 2 m \nu_b - { { \mathrm{re } } } ( v_4^ * \tilde a_1 ) 4 m \nu_b \nonumber \\
& + & { { \mathrm{re } } } ( v_4^ * \tilde a_4 ) 2 m \nu + { { \mathrm{re } } } ( v_6^ * \tilde a_1 ) 2 q^2 - { { \mathrm{re } } } ( v_6^ * \tilde a_4 ) q^2 \big ] \nonumber \\ \nonumber \\ w_{7}^i & = & 2 m^3 \big [ { { \mathrm{re } } } ( v_1^ * \tilde a_4 ) m + { { \mathrm{re } } } ( v_3^ * \tilde a_1 ) 2 m \nu_b - { { \mathrm{re } } } ( v_4^ * \tilde a_1 ) 2 m \nu - { { \mathrm{re } } } ( v_4^ * \tilde a_4 ) 2 m^2 \nonumber \\ & + & { { \mathrm{re } } } ( v_6^ * \tilde a_4 ) q^2 \big ] \nonumber \\ \nonumber \\ w_{8}^i & = & 0 \nonumber \\ \nonumber \\ w_{9}^i & = & 2 m^4 \big [ { { \mathrm{re } } } ( v_1^ * \tilde a_1 ) - { { \mathrm{re } } } ( v_1^ * \tilde a_2 ) + { { \mathrm{re } } } ( v_2^ * \tilde a_1 ) 2 m \nu_b + { { \mathrm{re } } } ( v_3^ * \tilde a_5 ) 2 m \nu_b \nonumber \\ & + & { { \mathrm{re } } } ( v_4^ * \tilde a_2 ) 2 m - { { \mathrm{re } } } ( v_4^ * \tilde a_4 ) m - { { \mathrm{re } } } ( v_4^ * \tilde a_5 ) 2 m \nu + { { \mathrm{re } } } ( v_6^ * \tilde a_5 ) q^2 \big ] \nonumber \\ \nonumber \\ w_{10}^i & = & - m^4 \big [ - { { \mathrm{re } } } ( v_1^ * \tilde a_1 ) - { { \mathrm{re } } } ( v_1^ * \tilde a_2 ) + { { \mathrm{re } } } ( v_1^ * \tilde a_3 ) + { { \mathrm{re } } } ( v_2^ * \tilde a_1 ) 2 m ( \nu_b -\nu ) \nonumber \\ & - & { { \mathrm{re } } } ( v_3^ * \tilde a_1 ) 2 m + { { \mathrm{re } } } ( v_3^ * \tilde a_4 ) m + { { \mathrm{re } } } ( v_3^ * \tilde a_5 ) 2 m \nu_b - { { \mathrm{re } } } ( v_3^ * \tilde a_6 ) 2 m \nu_b \nonumber \\ & + & { { \mathrm{re } } } ( v_4^ * \tilde a_1 ) 4 m + { { \mathrm{re } } } ( v_4^ * \tilde a_2 ) 2 m - { { \mathrm{re } } } ( v_4^ * \tilde a_3 ) 2 m - { { \mathrm{re } } } ( v_4^ * \tilde a_4 ) m \nonumber \\ & - & { { \mathrm{re } } } ( v_4^ * \tilde a_5 ) 2 m \nu + { { \mathrm{re } } } ( v_4^ * \tilde a_6 ) 2 m \nu - { { \mathrm{re } } } ( v_5^ * \tilde a_1 ) q^2 + { { \mathrm{re } } } ( v_6^ * \tilde a_5 ) q^2 \nonumber \\ & - & { { \mathrm{re } } } ( v_6^ * \tilde a_6 ) q^2 \big ] \nonumber\end{aligned}\ ] ] * structure functions for the electromagnetic tensor * + + the structure functions for the electromagnetic tensor follow from ( [ w_sym ] ) via the substitutions @xmath155
* vector currents * @xmath156 the amplitudes @xmath157 for the neutral weak vector current follow from @xmath158 + * axial currents * @xmath159 the amplitudes @xmath160 for the neutral weak axial current follow from @xmath161
in the following , we give the s channel contributions of the resonance .
the corresponding u channel contributions may be obtained by crossing symmetry , see ( [ cross ] ) .
+ + * vector currents * @xmath162 \nonumber \\ v_{2 , \delta}^s & = & i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ c_3^\gamma \bigg ( 1 + { \frac{2}{3m_\delta^2}}\frac{1}{2 } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) { \frac{q^2}{2m\nu_b}}\nonumber \\ & & + \frac{1}{2}{\frac{1}{3m_\delta}}{(m+m_\delta)}{\frac{q^2}{2m\nu_b}}\bigg ) -\frac{c_4^\gamma}{m } \bigg ( \frac{{(m_\delta - m)}}{2 } + \frac{1}{6 } { ( m+m_\delta)}{\frac{q^2}{2m\nu_b}}\nonumber \\ & & -{\frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \frac{1}{2 } { \frac{q^2}{2m\nu_b}}\bigg ) \bigg ] \nonumber \\ v_{3 , \delta}^s & = & i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ c_3^\gamma \bigg ( \frac{{(m+m_\delta)}}{3 } -{\frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \nonumber \\ & & \frac{1}{2 } { ( m+m_\delta)}+ { \frac{1}{3m_\delta } } ( { 2m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \bigg ) -\frac{c_4^\gamma}{m } \bigg ( -(\nu m + \frac{q^2}{2 } ) -\frac{m_\pi^2}{6 } \nonumber \\ & & + { \frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \frac{1}{2 } { ( m+m_\delta)}\bigg ) \bigg ] \nonumber \\ v_{4 , \delta}^s & = & i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ c_3^\gamma \bigg ( -\frac{2}{3 } { ( m+m_\delta)}- { \frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \nonumber \\ & & \frac{1}{2}{(m+m_\delta)}-{\frac{1}{3m_\delta}}\frac{1}{2 } m_\pi^2 \bigg ) -\frac{c_4^\gamma}{m } \bigg ( - m \nu_b -\frac{1}{6 } m_\pi^2 + { \frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \nonumber \\ & & \frac{1}{2 } { ( m+m_\delta)}\bigg ) \bigg ] \nonumber \\ v_{5 , \delta}^s & = & i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ c_3^\gamma \bigg(\frac{1}{2 } ( 1 -\frac{\nu}{\nu_b } ) ( { \frac{2}{3m_\delta^2}}({m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \nonumber \\ & & + { \frac{1}{3m_\delta}}{(m+m_\delta ) } ) \bigg ) -\frac{c_4^\gamma}{m } \bigg ( -\frac{1}{2 } { ( m_\delta - m)}+ \frac{1}{2}(1 -\frac{\nu}{\nu_b } ) \nonumber \\ & & ( \frac{{(m+m_\delta)}}{3 } -{\frac{1}{3m_\delta}}({m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) ) \bigg ) \bigg ] \nonumber \\ v_{6 , \delta}^s & = & i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ c_3^\gamma \bigg ( { \frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2})\frac{1}{2 } { ( m+m_\delta)}\nonumber \\ & & + { \frac{1}{3m_\delta}}\frac{m_\pi^2}{2 } \bigg ) -\frac{c_4^\gamma}{m } \bigg ( m \nu_b + \frac{m_\pi^2}{6 } - { \frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2})\frac{1}{2 } { ( m+m_\delta)}\bigg ) \bigg ] \nonumber\end{aligned}\ ] ] the amplitudes @xmath163 are calculated via @xmath164 and the @xmath165 for the neutral weak vector current follow from @xmath166 + * axial currents * @xmath167 \nonumber \\ a_{2 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( -m \nu_b { ( m+m_\delta)}- \frac{1}{3 } { ( m+m_\delta)}{m(\nu-\nu_b)}\nonumber \\ & & + { \frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) { m(\nu-\nu_b)}\bigg ) - c_5^a m \bigg ( -{\frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \nonumber \\ & & \frac{1}{2 } { ( m+m_\delta)}-\frac{1}{2 } { \frac{1}{3m_\delta}}m_\pi^2 \bigg ) \bigg ] \nonumber \\ a_{3 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( -{(m+m_\delta ) } ( m \nu + \frac{q^2}{2 } ) + \frac{1}{6 } { ( m+m_\delta)}q^2 \nonumber \\ & & -{\frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \frac{q^2}{2 } \bigg ) - c_5^a m \bigg ( \frac{2}{3 } { ( m+m_\delta)}-{\frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+ m_\pi^2 ) \nonumber \\ & & \frac{1}{2 } { ( m+m_\delta)}+ { \frac{1}{3m_\delta}}{m(\nu-\nu_b)}\bigg ) \bigg ] \nonumber \\ a_{4 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( \frac{m_\pi^2}{3 m}({m(\nu-\nu_b)}+\frac{q^2}{2 } ) -{\frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}\nonumber \\ & & + \frac{m_\pi^2}{2 } ) ( { m(\nu-\nu_b)}+\frac{q^2}{2 } ) \frac{{(m+m_\delta)}}{m } \bigg ) - c_5^a m \bigg ( -\frac{m_\pi^2}{3 m } + { \frac{1}{3m_\delta}}\frac{{(m+m_\delta)}}{m } \nonumber \\ & & ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \bigg ) \bigg ] \nonumber \\ a_{5 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( m \nu_b -\frac{1}{6 } ( 2 m { ( m+m_\delta)}- m_\pi^2 ) \nonumber \\ & & -\frac{1}{2 } { \frac{1}{3m_\delta}}{(m_\delta - m ) } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \bigg ) - c_5^a m \bigg ( \frac{1}{2 } { \frac{2}{3m_\delta^2}}({m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \nonumber \\ & & + \frac{1}{2 } { ( m+m_\delta)}{\frac{1}{3m_\delta}}\bigg ) \bigg ] \nonumber \\ a_{6 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( \nu m + \frac{q^2}{2 } - \frac{1}{6 } ( 2 m { ( m+m_\delta)}- m_\pi^2 ) \nonumber \\ & & -\frac{1}{2 } { \frac{1}{3m_\delta}}{(m_\delta - m ) } ( { m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \bigg ) - c_5^a m \bigg ( -1 + \frac{1}{2 } { \frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+ m_\pi^2 ) \nonumber \\ & & + \frac{1}{2 } { ( m+m_\delta)}{\frac{1}{3m_\delta}}\bigg ) \bigg ] \nonumber \\ a_{7 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( -m { ( m+m_\delta)}\nu_b - \frac{1}{6 } { 2m(\nu-\nu_b)}{(m+m_\delta)}\nonumber \\ & & + { \frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) { m(\nu-\nu_b)}\bigg ) - c_5^a m \bigg ( -{\frac{2}{3m_\delta^2 } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \nonumber \\ & & \frac{1}{2 } { ( m+m_\delta)}- { \frac{1}{3m_\delta}}\frac{m_\pi^2}{2 } \bigg ) \bigg ] \nonumber \\ a_{8 , \delta}^s & = & -i \frac{f_{\pi n \delta}}{m m_\pi } \frac{1 } { s - m_\delta^2 + i \gamma m_\delta } \bigg [ \frac{c_4^a}{m } \bigg ( m \nu_b - \frac{1}{6 } ( 2 m { ( m+m_\delta)}-m_\pi^2 ) \nonumber \\ & & -{\frac{1}{3m_\delta } } ( { m(\nu-\nu_b)}+\frac{m_\pi^2}{2 } ) \frac{1}{2 } { ( m_\delta - m)}\bigg ) - c_5^a m \bigg ( \frac{1}{2}{\frac{2}{3m_\delta^2}}({m(\nu-\nu_b)}+ \frac{m_\pi^2}{2 } ) \nonumber \\ & & + \frac{1}{2 } { ( m+m_\delta)}{\frac{1}{3m_\delta}}\bigg ) \bigg ] \nonumber\end{aligned}\ ] ] the amplitudes @xmath168 are calculated via @xmath169 and the @xmath170 for the neutral weak axial current follow from @xmath171 references c.y .
prescott et al .
: phys . lett . * b77*(1978)347 ; ibid . * b84*(1979)524 w. heil et al .
: nucl . phys . * b327*(1989)1 p.a .
souder et al . :
* 65*(1990)694 d. beck et al . :
cebaf * pr-91 - 017*(1991 , revised 1992 ) r. souder et al .
: cebaf * pr-91 - 010*(1991 ) e. beise et al .
: cebaf * pr-91 - 004*(1991 ) d. von harrach et al .
: mainz proposal * a4/1 - 93*(1993 ) r. mckeown et al .
: mit / bates proposal * 89 - 06*(1989 ) l.m .
nath , k. schilcher , m. kretzschmar : phys . rev . * d25*(1982)2300 d.r.t .
jones , s.t .
petcov : phys . lett . * b91*(1980)137 r.n .
cahn , f.j.gilman : phys .
rev . * d17*(1978)1313 d. ishankuliev , m.ya .
safin : sov . j. nucl
* 31*(1980)512 s.p .
henley , w .- y.p .
hwang : ann .
phys.(ny ) * 143*(1982)372 m.j .
musolf et al .
: phys . rep .
* 239*(1994)1 s.l .
adler : ann .
phys.(ny ) * 50*(1968)189 a. cone et al .
: phys . rev .
* 156*(1967)1490 ; ibid . * 163*(1967)1854(e ) h.l .
lynch et al .
: phys . rev .
* 164*(1967)1635 p. renton : electroweak interactions .
cambridge university press 1990 v. bernard , n. kaiser , j. gasser , u .- g .
meiner : phys .
* b268*(1991)291 v. bernard , n. kaiser , u .- g .
meiner : phys . lett . * b331*(1994)137 particle data group , review of particle properties : phys . rev .
* d50*(1994)1173 j.m .
laget : nucl .
* b481*(1988)765 l. tiator , h.j .
weber , d. drechsel : nucl .
a306 * ( 1978)468 | parity violating ( pv ) contributions due to interference between @xmath0 and @xmath1 exchange are calculated for pion electroproduction off the nucleon . a phenomenological model with effective lagrangians
is used to determine the resulting asymmetry for the energy region between threshold and @xmath2 resonance .
the @xmath3 resonance is treated as a rarita - schwinger field with phenomenological @xmath4 transition currents .
the background contributions are given by the usual born terms using the pseudovector @xmath5 lagrangian .
numerical results for the asymmetry are presented .
+ + pacs numbers : 25.30.rw , 13.10.+q |
loop quantum cosmology has provided in past years a useful framework to ask questions about the quantum nature of the early universe @xcite . closely related to loop quantum gravity @xcite ,
the formalism has shown that , for isotropic models , the big bang singularity is replaced by a quantum bounce @xcite .
these results have been obtained for closed and open frw cosmologies @xcite with and without a cosmological constant @xcite for a massless scalar field , and recently extended to the massive case @xcite .
recently , some of these results have been extended to anisotropic @xcite and some inhomogeneous models @xcite . in all these cases , quantum gravitational effects
have been shown to exert a repulsive force and halt the collapse and launch an expanding superinflationary phase before the quantum gravity effects die out and the universe then follows the standard general relativity ( gr ) dynamics . an exactly solvable model for the @xmath0=0 frw with a massless scalar field has allowed to prove analytically the generic nature of the quantum bounce @xcite . in particular , it was shown that the bounce occurs for all physical states , and the energy density possesses a supremum for physical states @xcite . regarding the fluctuations
, it has been shown that they are bounded across the bounce implying that a semiclassical state at late times after the bounce had to come from a semiclassical state before the bounce @xcite .
recently , kaminski and pawlowski have imposed stronger bounds on the asymmetry in relative dispersions before and after the bounce @xcite , thus very strongly ruling out some claims of loss of semiclassicality in @xcite .
these robustness result have support on the fact that , within a large possibility of ` loop quantizations ' , physical considerations select a unique consistent quantization @xcite ( the one introduced in @xcite ) .
an interesting feature of this system is that one can approximate very well the dynamics followed by semiclassical states by an _
effective description _
@xcite . that is , the classical limit of the dynamics is not that of gr + scalar field but rather is defined by an _
effective hamiltonian_. there have been several approaches to reach this result .
for instance , in @xcite the author considered _ kinematical _ coherent states and used them to compute the effective hamiltonian as the expectation value of the quantum hamiltonian constraint of lqc .
more recently , the authors of @xcite showed that , in the path integral description of the model , the paths that contribute the most are those that satisfy the effective equations and not those that follow the classical gr equations . the purpose of this paper is to construct exact physical semiclassical states for the solvable model @xcite .
the strategy will be to define a class of coherent initial states and , using the analytical control we have on the space of solutions to the quantum constraint equations , produce exact physical coherent states , much in the spirit of @xcite .
then , we impose well motivated conditions on the behavior of the fluctuations of physical operators to select from all possible coherent states , those that exhibit a semiclassical behavior .
it turns out that these conditions are sufficient to fix enough parameters of the states and to have states that are ` peaked ' on a given ( physical ) phase space point .
having found semiclassical states , we explore the behavior of these states near the bounce , where one expects the state to be most quantum .
in particular we consider the relative fluctuations of volume and find that , for all states and for all times , the relative fluctuations are bounded by their asymptotic value .
in fact , the minimum value turns out to be precisely at the bounce . the expectation value for volume and density on semiclassical states are very well approximated by the values given by the effective theory .
this leads us to conclude that the effective description is indeed a very good approximation to semiclassical _ physical _ states .
a natural question is whether the class of states considered is the most general one . for that purpose
, we generalize the class of gaussian states to the so called _ squeezed _ states .
the idea is to explore the parameter space near the gaussian states that are well behaved semiclassical states and see whether these squeezed states exhibit better semiclassical properties . of particular interest in this regard
is the asymmetry in volume fluctuations .
claims of a loss of coherence , or ` cosmic forgetfulness ' arise due to results in an approximate framework @xcite ( motivated by , but not within lqc ) that suggest that the fluctuations across the bounce might change significantly and spoil the semiclassical nature of the state @xcite .
even when strong bounds against this possibility within lqc have already appeared @xcite , our _ exact _ formalism is particularly well suited for addressing this question . by imposing bounds on the total asymptotic dispersion in volume ( at late or early times with respect to the bounce )
, we are able to bound how far we can be from gaussian states and still have a semiclassical state .
the range on parameter space allowed turns out to depend on the physical requirements for the state , in such a way that , for ` large volume universes ' , it is severely constrained .
this is turn sets very strong bounds on the allowed asymmetry in volume fluctuations , consistent with @xcite and , in particular , disproving some of the statements of @xcite that claim the loss of coherence across the bounce for semiclassical states .
the structure of the paper is as follows . in sec .
[ sec:2 ] we recall the solvable model in loop quantum cosmology together with the basic operators to be studied . in sec .
[ sec:3 ] we introduce the gaussian states and compute expectation values and fluctuations for the physical operators .
semiclassicality conditions are explored in sec .
[ sec:4 ] , together with the behavior of the states near the bounce .
squeezed states are introduced and compared to the gaussian states in sec .
[ sec:5 ] .
we end with a discussion in sec.[sec:6 ] .
there are four appendices . in appendix [ app : a ] we study the symplectic reduction of the ` effective theory ' to isolate the physical degrees of freedom . in appendix
[ app : b1 ] we recall some useful integrals and formulas . in appendix
[ app : b ] we present an analysis of the errors introduced by considering gaussian states , showing that they are severely suppressed and justify our analysis in the main body . in appendix
[ app : c ] we present in a table a comparison of different quantities for two families of gaussian states .
some of the results reported here were summarized in the short manuscript @xcite , so it is natural that there is some overlap of part of the material .
in this section we shall recall the model we are considering , namely , the solvable @xmath0=0 frw model coupled to a massless scalar field
. our presentation will be self - contained and slightly different from that of @xcite . in terms of the phase space variables used in loop quantum gravity ( lqg )
@xcite , namely a connection @xmath1 and a densitized triad @xmath2 , the homogeneous gravitational sector can be expressed as , a_a^i = cv_0 ^ -(1/3)^o^i_a ; e^a_i = p v_0 ^ -(2/3)^oe_i^a where @xmath3 are a set of orthonormal co - triads and triads compatible with the fiducial ( flat for @xmath4 ) metric @xmath5 and @xmath6 is the volume of the fiducial cell , introduced to define the symplectic structure , with respect to @xmath5 .
the phase space is characterized by conjugate variables @xmath7 satisfying @xmath8 . where @xmath9 @xmath10 is the barbero - immirzi parameter used in lqc .
the triad @xmath11 is related to the physical volume @xmath12 of the fiducial cell , as @xmath13 where @xmath14 is the scale factor and , on the space of classical solutions , @xmath15 .
it is convenient to introduce new variables @xcite , : = c / p^1/2= p^3/2/(2 ^2 ) [ b_eq ] where @xmath16 introduced in @xcite and used thereafter . ] . here
@xmath17 is the orientation of the triad with respect to that of @xmath18 .
the classical constraint , with the choice @xmath19 , then becomes = 38 g ^2 ^ 2 v^2 - p_^22 0 where the matter content consists of a massless scalar field @xmath20 , with canonical momenta @xmath21 .
thus , the kinematical phase space can be though of as a four dimensional space with coordinates @xmath22 .
the strategy for quantization is to define an operator associated to @xmath23 and to look for states that are annihilated by the constraint operator . in the @xmath24 representation ,
the operators basic are represented as ( , ) & = & ( , ) ; ( , ) = -i4 ^ 2 ( , ) ; + _ ( , ) & = & -i ( , ) .
however , the operator @xmath25 is not well defined since @xmath26 is periodic with period @xmath27 , so it has to be replaced by a well defined operator @xmath28 ( see @xcite for details ) .
the quantum constraint of slqc becomes then @xcite , ^2 ^ 2(,)= ^2 ( ) ^2 ( , ) [ const1 ] with @xmath29 , @xmath30 . in the loop quantum cosmology literature ,
the value of @xmath31 is chosen such that @xmath32 corresponds to the minimum eigenvalue of the area operator in loop quantum gravity ( corresponding to an edge of ` spin 1/2 ' ) . with this choice
the free parameter @xmath31 becomes @xmath33 . by introducing x = ^-1 |(/2)| [ x - eq ]
the quantum constraint can be rewritten in a klein - gordon form [ kg - eq ] _
^2 ( x , ) = _ x^2 ( x , ) .
as usual for this system , in its classical evolution the scalar field @xmath20 is a monotonic function and can play the role of internal clock . in the quantum theory
, one can also think of the evolution of the state with respect to @xmath20 . a general solution @xmath34 to eq .
( [ kg - eq ] ) can be decomposed in the left and right moving components : [ chi_eq ] = _ + ( + x ) + _ -(- x ) : = _ + ( x_+ ) + _ -(x_- ) . the physical states that we shall consider are positive frequency solutions of ( @xmath35 ) . since there are no fermions in the model , the orientations of the triad are indistinguishable and @xmath34 satisfy the symmetry requirement @xmath36 .
thus , we can write @xmath37 , where @xmath38 is an arbitrary ` positive frequency solution ' is a positive momentum function , i.e. with a fourier transform that has support on the positive axis .
with such a choice , the solution to the constraint equation become of positive frequency . ] . the physical inner product on solutions is given as @xcite ( _ 1 , _
2)_phy & = & -i_=_0 [ |_1(x,)__2(x , ) -(_|_1(x,))_2(x , ) ] x + & = & i_-^ x .
[ inner - prod ] we can now compute the expectation values and fluctuations of fundamental operator such as @xmath39 and @xmath40 , where @xmath41 is related to the operator @xmath42 . for
_ any _ state on the physical hilbert space the expectation value of the volume operator at ` time @xmath20 ' is given by _ :
= ( , |_)_phy = 2 ^ 2 ( , || ) _ phy where @xmath43 is the absolute value operator obtained from = -2(x)i_x . using the inner product the expectation value of @xmath43
is given by ( , || ) _ phy & = & i_-^ x + & = & 2_-^ [ _ x |f(x_+)(x)_x f(x_+ ) + _ x|f(x_-)(x)_x f(x_- ) ] x + & = & 4_-^ |fx|^2 ( ( x- ) ) x .
is easy to check that @xmath44 . from these expressions
one can find the expectation value of @xmath45 and the dispersion of the operator , @xmath46 with @xmath47 , @xmath48 , @xmath49 , @xmath50 and @xmath51 being real and positive , given by v _ & = & ||^2e^ x x [ const5 ] , + w_0 & = & i 2 ^ 2 ^ 4 ^ 23 g x , + [ w - eq ] w_&=&i^2 ^ 4 ^ 23 g e^2x x , for normalized states . here
@xmath52 and @xmath53 .
note that the expressions for @xmath48 and @xmath54 correct those found in @xcite . from ( [ v - exp ] )
, it follows that the expectation value of the volume @xmath55 is large at both very early and late times and has a non - zero global minimum @xmath56 .
the _ bounce _ ocurrs at time @xmath57 @xcite . around @xmath58 ,
the expectation value of the volume @xmath59 is symmetric .
similarly , @xmath60 is symmetric across the value @xmath61 for the scalar field .
a trivial observation is that if @xmath62 , the difference in the asymptotic values of the relative fluctuation [ d - eq ] d:=_= - vanishes .
it should be noted that this quantity quantifies the change in semiclassicality across the bounce as pointed out in @xcite .
the other observable that one might want to consider as fundamental , is an observable naturally conjugate to @xmath21 . given that we are considering physical states , this does not correspond to the operator @xmath20 as happens in the kinematical setting .
instead , as illustrated in the appendix [ app : a ] , the quantity that is conjugate to @xmath21 , in the reduced physical model , corresponds to the quantity @xmath63 .
thus , it is natural to consider the family of operators @xmath64 , corresponding to the observable ` @xmath65 at time @xmath20 ' .
the expectation value of @xmath41 is defined as , ( _ 1,x _ 2)_phy = 12[(|x|_1,_2)_phy+(_1,|x|_2)_phy ] [ x_sym ] where @xmath66 and @xmath67 .
we have included the two terms in the definition because @xmath68 is not symmetric .
we can now compute the expectation values and fluctuations of @xmath41 for _ any _ state of the physical hilbert space x_&= & _ 1- , [ x - ope ] + x^2_&= & x_2 -2_1 + ^2 , + ( x ) & = & x_2 - _ 1 ^ 2 , with @xmath69 and @xmath70 real ( state dependent ) constants , given by x_1 & = & -2i_-^|f(x)x_x f(x ) x -i_-^|f(x)|^2 x [ x_1 ] + x_2 & = & -2i_-^|f(x ) x^2 _ x f(x ) x - 2i_-^|f(x)|^2 x x , , [ x_2 ] for normalized states .
is important to note that @xmath71 , @xmath72 but the fluctuation @xmath73 is independent of @xmath20 !
this property indeed gives support to the proposal that @xmath41 be the conjugate of @xmath74 .
finally , we compute the expectation values for @xmath21 , a dirac observable as p_=-i _ & & p_= 2 _ -^|fx|^2x + p_^2 = -^2 _ ^2 & & p_^2 = - 2 ^2 i _ -^|fx^2 fx^2x , , for normalized states . in the following sections , it will be convenient to consider a general physical state written as , ( x , ) = _ 0^(k)e^-ik(+x ) k- _ 0^(k)e^-ik(-x ) k where the fourier transform @xmath75 contains all the information of the state .
positive frequency solutions to eq .
( [ kg - eq ] ) means that @xmath75 has support on positive @xmath0 s only .
furthermore , we can consider initial states , for an ` initial time ' @xmath76 , in order to compute some of the quantities that determine the expectation value of relevant operators .
it then suffices to specify @xmath75 and from there to compute all the quantities @xmath77 and @xmath78 .
in this section we shall consider physical states with initial states given _ generalized gaussian states _ , defined by functions @xmath79 of the form : [ coh_state ] f(k)=\ { rl k^n e^-(k - k_0)^2 /^2e^ipk , & k>0 , + 0 , & k0 , . with @xmath80 , @xmath81 .
that is , we are choosing gaussian states centered around the point @xmath82 , with ` dispersion ' given by @xmath83 .
the physical states one has to consider are vanishing in the negative @xmath0 axis , to have positive frequency solutions . in order to gain analytical control over all the quantities at hand
, we shall in what follows consider instead the gaussian states for all values of @xmath0 .
thus , the quantities we shall compute are an approximation to the real quantities but , as we shall argue in detail , this approximation is justified when the values of @xmath82 and @xmath83 are such that the wave function has a negligible contribution from the negative axis . a detailed analysis of the errors introduced by this simplification can be found in the appendix [ app : b ] .
therefore , from now on , in order to have states that approximate positive frequency solutions , we shall impose the condition @xmath84 .
further consistency conditions on @xmath83 , motivated by physical considerations will be derived below .
the norm of the state in @xmath0 space is @xmath85 if we take the initial states in the @xmath0-space as in ( [ coh_state ] ) , the norm is , @xmath86 which can be approximated by @xmath87 ( see appendix [ app : b ] for a justification of this approximation ) . with the change of variables @xmath88 this
last integral takes the form [ norm_int ] _ phy^2=2 ^ 2n+2 _ -^^2n+1 e^-u^2 u . for concreteness
, we shall consider the state corresponding to @xmath89 .
in this case _
phy^2&=&2 ^ 2 _ -^e^-u^2 u = k_0 .
[ norm0 ] the normalized @xmath90=0 states can be used to compute explicitly the expectations values and fluctuations of several physically interesting operators .
the corresponding results for @xmath90=1 are summarized in the appendix [ app : c ] .
let us explore the basic observables for the gaussian states in order to gain a better understanding of the free parameters @xmath91 that characterize the states .
the first observable we shall consider is @xmath21 , that is a dirac observable ( and therefore a ` constant of the motion ' ) . the operator is represented as @xmath92 , and its expectation value is thus given by @xmath93 using parseval s theorem we get p_= 2 ^2 _
_ -^k^2 |f(k)|^2 k .
if we take our initial states @xmath94 then the integral takes the form [ pphi ] p_= 2 ^2 _ _ -^k^2n+2 e^-2(k - k_0)^2 /^2
k .with the change of variables @xmath88 , the expectation value ( [ pphi ] ) takes the form p_= 2 ^2 _ ^2n+3 _ -^^2n+2 e^-u^2 u . in the case of pure gaussian states , namely for @xmath89 , we have p_= = k_0 .this is telling us that , for the approximation we are taking , namely @xmath95 , then @xmath96 .
therefore , as expected , the parameter @xmath82 is giving us a good measure of the expectation value of @xmath21 which can be regarded as the conjugate variable to @xmath65 , on the reduced phase space of the system ( see appendix [ app : a ] ) .
let us now compute the fluctuations of this operator .
let us start by computing p_^2 = - 2 ^2 i^2 _
_ -^|fx^2 fx^2x = 2 ^ 2 ^ 2 _ _ -^k^3
|f(k)|^2 k , . [ parsebal2 ] for our generalized gaussian states the integral takes the form [ pphi2 ] p_^2 = 2 ^2 i^2 _ _ -^k^2n+3 e^-2(k - k_0)^2 /^2
k .which can be rewritten as , @xmath97^{2n+4 } \int_{-\infty}^\infty \left [ u+{\frac}{\sqrt{2}k_0}{\sigma } \right]^{2n+3 } e^{-u^2}\ , \d u\ , .\ ] ] taking @xmath89 , p_^2 = = ^2 k_0 ^ 2 .
[ p2-result ] note that for large @xmath82 , that is for those states satisfying @xmath95 , then @xmath98 .
let us now compute the dispersion of the observable @xmath74 , @xmath99 . for the @xmath89 case we have ( p_)^2&=&^2 k_0 ^ 2-^2 k_0 ^ 2 ^ 2 + & = & ^2 ^ 24which is a constant .
note that for @xmath95 , the dispersion @xmath100 . this is telling us that the parameter @xmath83 has the interpretation one might have expected as the dispersion of the observable associated with the variable @xmath0 , which in this case corresponds to @xmath21 .
recall that the quantity that is conjugate to @xmath21 , in the reduced model , corresponds to @xmath63 .
thus , it is natural to consider the family of operators @xmath64 , corresponding to the observable ` @xmath65 at the time @xmath20 ' .
it is most natural to define the symmetric operator acting on initial states @xmath101 , in the @xmath0-representation the operator defined in eq .
can be written as : f:=(ik + i2k-)f .[x - operator ] the first two terms arise when eq .
is replaced into eq . .
it is now straightforward to compute the expectation value of the @xmath102 operator on the generalized gaussian states .
it is easy to see that , for all values of @xmath90 , [ x_0 ] x = -p - , which confirms the expectation that the parameter @xmath103 represents the point in @xmath65 space , where the gaussian is peaked ( when @xmath76 ) .
this also corresponds to the classical dynamics ( as given by the effective theory ) for the variable @xmath65 ( with our choice @xmath104 ) .
furthermore , one can try to find the expectation value of the operator @xmath105 here it is important to note two important issues that arise when considering the operator defined by ( [ x - operator ] ) .
the first one is that it involves a derivative operator and the second is the factor @xmath106 in the operator .
the derivative term implies that one needs to consider carefully the boundary conditions at @xmath4 , so that the boundary terms ( when integrating by parts ) does not contribute .
in particular , we should have , for the @xmath102 operator , for states defined by functions @xmath107 and @xmath108 , that @xmath109 .
this condition is , of course , satisfied by all of our states @xmath110 .
the term @xmath106 imposes some fall - off conditions at the origin ( @xmath4 ) for the integrals to be finite .
for instance , even when the expectation value of @xmath102 is well defined in the @xmath89 gaussians , these states are no longer in the domain of the operator .
the state that results , when acting with the operator @xmath102 on a @xmath89 state is not normalizable , and therefore , does not belong to the physical hilbert space .
we have to conclude that , in order to have the operator @xmath102 in the set of observables we want to consider , we have to exclude the @xmath89 states and consider only @xmath111 . in the @xmath112 case ,
it is straightforward ( if lengthy ) to compute the expectation value of the operator , [ x2_0 ] x^2 _ = p^2 + 1 ^ 2(1 + 3 ^ 22k_0 ^ 2 ) ( 1 + 3 ^ 24k_0 ^ 2 ) + 2 p+^2 from which the fluctuations of the operator becomes , [ deltax ] ( x)^2 = x^2 _ - x^2_=1 ^ 2(1 + 3 ^ 22k_0 ^ 2 ) ( 1 + 3 ^ 24k_0 ^ 2 ) which also tells us that the parameter @xmath83 has the expected interpretation of providing the inverse of the dispersion for the observable @xmath65 .
this is realized when @xmath113 and therefore the second term in the previous expression is very close to 1 .
we can now examine the uncertainty relations for the observables @xmath114 and @xmath102 ( for @xmath112 ) : ( p_)^2(x)^2 = ^24 ( 1 + 9 ^ 416k_0 ^ 4 + 3 ^
24k_0 ^ 2 - 9 ^ 664k_0 ^ 6 ) ( 1 + 3 ^ 22k_0 ^ 2 ) ( 1 + 3 ^ 24k_0 ^ 2)^3 we note that , if @xmath113 , then the uncertainty relations @xmath115 are very close to being saturated .
as we shall see in the next section , these conditions are necessary for the generalized gaussian states to represent acceptable semiclassical states .
let us start by considering the operator @xmath116 corresponding to the volume at ` time ' @xmath117 .
the expectation value of this operator is given by eq .
( [ v - exp ] ) so it suffices to compute the constants @xmath47 given by v _ & = & ( k ) f(ki)k .
if we take the state as @xmath94 , the integral is @xmath118[k\mp i\alpha]^n e^{-(k\mp i\alpha - k_0)^2 /\sigma^2}e^{ip(k\mp
i\alpha ) } \d k \ , .\ ] ] this can be rewritten as , [ vol1 ] v_= e^p_-^^n+1 e^-(k - k_0)^2 /^2e^-(ki - k_0)^2 /^2
k .it is straightforward to see that , with the change of variable @xmath119 , the integral ( [ vol1 ] ) can be written as @xmath120^{n+1 } e^{-u^2/\sigma^2}\d u\ , .\ ] ] for the @xmath89 case we have , [ vol2 ] v_=(14 ) e^p e^^2/2 ^ 2_-^e^-u^2/^2u .
from which we get @xmath121 if we take now a normalized states using eq .
( [ norm0 ] ) , the result is [ vpm_0 ] v_=v__phy^2 = ^2 2 e^p e^^2/2 ^ 2 . in order to calculate the dispersion of the volume operator , it is necessary to calculate the three integrals @xmath122 .
let us start by computing the quantity @xmath48 .
[ w0_int ] w_0=i_2 ^ 2__-^x , _
2=2 ^ 2 ^4 ^23 g , which can be written as w_0=w_0 ^ 2 _
= 2_2 ^ 2 _ _ -^k^3|f(k)|^2 k .
this is the same integral that the one in eq . .
then , for a normalized gaussian state with @xmath89 we only need to replace @xmath123 by @xmath124 into eq . to obtain that [ w0_0 ]
w_0=_2 k_0 ^ 2 .
let us now compute the quantities @xmath125 , with [ wpm_int ] _
= i_3_-^e^2xx , _
3=^2 ^4 ^23 g , that can be written as , _ = _
3_-^k^2 k .
+ if we take the initial states @xmath94 , the integral is w _ & = & _ 3 e^2 p_-^k^n+2 e^-(k - k_0)^2 /^2 + & & [ ( k2i)^n+1 e^-(k2i - k_0)^2/^2 + ( k2i)^n+1e^-(k2i - k_0)^2/^2 ] k in order to do the integral @xmath126 it is straightforward to show that with the change of variables @xmath127 , we get i()&=&_-^(u+k_0i)^n+2 ( u+k_0i2i)^n+1 e^-u^2 /^2 e^2 ^ 2/^2 u + & = & e^2 ^ 2/^2_-^ ^n+1 ( u+k_0i ) e^-u^2 /^2 u . the real part of @xmath128 is [ real_int ] e[i()]= e^2 ^ 2/^2_-^ ^n+1 ( u+k_0 ) e^-u^2 /^2 u
. then @xmath129={\mathbb{r}}{\rm e}[i(-)]$ ] , and the imaginary part of @xmath128 is @xmath130=\mp\alpha { \frac}{e^{2\alpha^{2}/\sigma^2}}{\sqrt{2}}\int_{-\infty}^\infty \left[\left({\frac}{u}{\sqrt{2}}+k_0\right)^2 + \alpha^2\right]^{n+1 } e^{-u^2 /\sigma^2 } { \d u}\ , , \\\ ] ] therefore , @xmath131=-{\mathbb{i}}{\rm m}[i(-)]$ ] . using the real and imaginary parts
, we have that the integral @xmath132 takes the form w_=_3 e^2p[i(+)+i(- ) ] = 2_3 e^2pe[i(+ ) ] .
this tells us that @xmath132 is real valued ( as expected ) and that it is only necessary to calculate the integral ( [ real_int ] ) .
taking @xmath89 the integral ( [ real_int ] ) takes the form e[i()]&= & e^2 ^ 2/^2_-^ ( u+k_0 ) e^-u^2 /^2 u + & = & e^2 ^ 2/^2 ( u^32 ^ 2/3 + 3u^22k_0 + u(3k_0 ^ 2+^2 ) + ( k_0 ^ 3 + ^2k_0))e^-u^2 /^2 u . using the integrals from appendix [ app : b1 ] , the integral is e[i ( ) ] & = & e^2 ^ 2/^2 ( 34k_0 ^ 3 + k_0 ^ 3+^2k_0 ) , from which we have that , w_= _ 3 e^2pe^2 ^ 2/^2 ( 34k_0 ^ 3+k_0 ^ 3+^2k_0 ) .
including now the normalization of the state we have [ wpm_0 ] w_= ^2 ^4 ^23 g e^2pe^2 ^ 2/^2 .
from the last equation and eq .
( [ vpm_0 ] ) we can observe that the difference in the asymptotic relative fluctuations after and before the bounce , as defined by eq .
, vanishes .
that is , @xmath133 .
it is important to note that this is a general result for any gaussian state of the form of eq .
( [ coh_state ] ) , for all values of @xmath134 , as was expected to happen @xcite .
so far we have considered generalized gaussian states as initial states for physical states in the exactly solvable @xmath0=0 lqc model with a massless scalar field .
we have seen that the parameters that define the states , in the pure gaussian case with @xmath89 , have the expected interpretation : the parameter @xmath82 is related to the expectation value of @xmath114 , @xmath103 to the expectation value of @xmath102 , and @xmath135 to the dispersion of @xmath21 .
the only condition that we have imposed so far is @xmath95 , which guaranties the validity of our approximation .
we have also seen that , for states which satisfy this consistency condition , there are at least two results that follow .
first , it warranties that the relative dispersion of @xmath21 is small .
second , this conditions implies that the uncertainty relations become saturated in the sense that @xmath136 . from this perspective
, it could seem that any value of @xmath82 and @xmath83 , provided they satisfy @xmath137 , might be acceptable to define semiclassical states . the purpose of this section is to explore this issue further and answer the following question : are there more stringent conditions that one must impose in order to have semiclassical states that will further restrict the possible values of @xmath82 and @xmath83 ? as we shall see
, the answer is in the affirmative . in order to answer this question
we shall first consider the volume operator .
since all physical states have the property that the expectation value of the volume operator @xmath138 at time @xmath20 follow the _ same _ functional form and , therefore , follow for large @xmath20 the same dynamics of the classical dynamics ( i.e. einstein s equations ) @xmath139 , we need more criteria to select those states that are semiclassical .
the obvious strategy is to consider the state s relative dispersion @xmath140 of the volume operator at time @xmath20 .
one expects that semiclassical states will have a very small relative dispersion ` at late times ' when the dynamics approaches the classical dynamics .
it is then natural to consider the asymptotic relative dispersions given by _ : = _ ( ) _ ^2= w_v_^2 - 1 .
it is straightforward to find the analytical expression for these quantities in our states ( in the case @xmath89 ) to be _ = w_v_^2 - 1 = e^^2/^2 ( 1 + 3 ^ 24k_0 ^ 2+^2k_0 ^ 2 ) ( 1 + ^2 + ^24k_0 ^ 2 ) ^2 -1 .
[ delta - pm ] it is interesting to note that , in the previous expression , there is some competition between the factor @xmath141 and the second part . the exponential becomes close to one if @xmath83 is large compared to @xmath142 , but can become very large if @xmath83 becomes too small .
on the other hand , the larger @xmath83 becomes , the larger the second term .
thus , that there must be an optimal value @xmath143 for which @xmath144 is the smallest .
the form of @xmath144 can be seen in fig .
[ fig:1.1 ] .
another important aspect to consider is the following .
so far , we have not imposed any condition on @xmath21 .
that is , we have not said that it has to be ` large ' , mainly because in the classical theory there is no dimension - full quantity with respect to which one can compare it . in the quantum theory , with the introduction of @xmath145
we do have such scale . as we can see from eq.([delta - pm ] ) and the expression for @xmath146 , the quantity @xmath147 has the same dimensions of @xmath21 .
furthermore , if we now impose the condition @xmath148 , all the terms in parenthesis in eq.([delta - pm ] ) are small , so we can indeed try to find the optimum value for @xmath83 .
note also that the condition @xmath148 implies @xmath149 .
it is in this sense that we can call the momentum @xmath21 ` large ' .
the condition that the expression ( [ delta - pm ] ) be an extrema is a cubical equation , and the only physically interesting solution can be approximated as , ^2 2 k_0 + 27 ^2 .since we require that @xmath148 , then the second term is very small compared to the first one and we can approximate @xmath150 .
it is important to stress that this condition selects a preferred value for @xmath83 for which the state is most semiclassical .
as can be seen from the figure [ fig:1.1 ] , if we make @xmath83 slightly smaller ( in an attempt to ` make the dispersion small ' ) the asymptotic relative dispersion in volume becomes very large , making the state not a good candidate for a semiclassical state .
if we now introduce this value for @xmath143 into the relative fluctuation function , then this can be approximated by ._|_(k_0 )
k_0 [ opt - dispersion ] this approximation has an error of @xmath151 when @xmath152 ( in planck units ) , and becomes better as @xmath82 grows .
this tells us the precise relation between the parameter @xmath82 that controls the expectation value of @xmath21 and the minimum value of the asymptotic relative fluctuation of the volume operator .
the @xmath83 dependence of the asymptotic relative dispersion of the volume can be seen in figure [ fig:1.1 ] , with @xmath152 .
two features characterize this plot , namely , the first one is that for @xmath153 the function is exponential so it grows very quickly as one approaches zero ; the second feature is that for @xmath154 the function has a polynomial behavior , so one can have values of @xmath155 without @xmath144 changing too much .
is plotted as a function of @xmath156 with @xmath152 in planck units , where @xmath157 .
here we have chosen @xmath89 and @xmath158 .
two features characterize the behavior of the function near the minimum @xmath143 .
the first one is that for @xmath153 the function is exponential ; the second one is that for @xmath154 the function has a polynomial behavior . ] to summarize , we have seen that not any gaussian state will be an admissible semiclassical state . in order to minimize the asymptotic relative fluctuation @xmath144
, one has to choose the parameter @xmath83 carefully as a function of the other relevant parameter ( in this case , the parameter @xmath82 ) .
that is , if we choose @xmath83 to be of the order of @xmath159 , we can then minimize the asymptotic relative dispersion of volume after and before the bounce ( which are the same , as we have seen ) .
gaussian states for which this condition is not satisfied will have , in general , large quantum fluctuations in the regime where we expect them to be small for large times after and before the bounce so they can not be regarded as semiclassical .
let us then assume that we have chosen the value @xmath160 for @xmath83 .
the question we want to ask now is the following : can we reproduce some of the features that are characteristic of the so called `` effective dynamics '' ? as we have discussed before , the ` classical ' hamiltonian constraint one obtains from replacing classical quantities like connection and curvature by a corresponding holonomy functions with a finite parameter @xmath31 , can be seen as either a starting point for ` polymer quantization ' , or as the ` classical limit ' of a loop quantized theory . in any case
, we expect that the semiclassical states of the quantum theory approach this ` classical theory ' in the appropriate regime .
in other words , we want to know when the effective dynamics is a good approximation to the full quantum dynamics .
is there a choice of parameters in the gaussian states for which the effective dynamics is not a good approximation ?
we shall answer these questions in the remainder of this section .
another important issue is the question of how semiclassical the state is at the bounce .
for instance , if the state has a very small relative asymptotic dispersion @xmath144 at late times , one might want to know what the relative dispersion is at the bounce and how it compares to @xmath144 .
a plot of the relative fluctuation , as a function of internal time @xmath20 can be seen in fig .
[ fig:1.2 ] , where it is shown that the relative fluctuation is bounded , it attains its minimum value at the bounce and approaches @xmath144 asymptotically .
thus , somewhat surprisingly , the state does not become very quantum - like " at the planck scale but rather preserves its coherence across the bounce .
let us now try to estimate the value of the relative dispersion in comparison to its asymptotic value .
we know that , at the bounce , @xmath161 , then @xmath162 , and ( v ) _
-4v_+v_- + 2 .
then the relative fluctuation in the bounce is .
( v)^2v ^2 | _ = w_0 + 24v_+v_--1 . for @xmath89 in the gaussian states we have .
( v)^2v ^2 |_-12 .
( v)^2v ^2 | _ = e^-^2/^2(1 + 3 ^ 24k_0 ^ 2 ) 2(1+^2 + ^24k_0 ^ 2 ) ^2 - 12 .
is plotted as a function of internal time @xmath20 with @xmath152 , in planck units where @xmath157 .
note that the minimum of this quantity is at the bounce ( here corresponding to @xmath76 ) .
it is found that the value at the bounce is approximately 1/2 of its asymptotic value for the generalized gaussian states . ] now , if we impose @xmath163 , which is satisfied if we choose @xmath164 , then the last equation has the form , .
( v)^2v ^2 |_-12 .
( v)^2v ^2 | _ & = & o(^2/^2 ) + o(^2/k_0 ^ 2 ) .
this tells us that when @xmath165 , the relative volume fluctuation at the bounce becomes [ eq110 ] .
( v)^2v ^2 |_12 .
( v)^2v ^2 |_= _ .
this approximation , for instance , has an error of @xmath166 when @xmath152 , and it becomes smaller as @xmath82 grows . if we use the eqs . and and choose the value of @xmath83 that minimized the relative fluctuation we get @xmath167 this approximation have an error of @xmath168 when @xmath152 , and becomes smaller as we increase @xmath82 .
we have performed intensive numerical explorations and have seen that , for @xmath111 , the relative fluctuation is again smaller at the bounce and approaches @xmath169 of its asymptotic value .
thus , we can conclude that states that satisfy the conditions @xmath165 , imposed by our requirements on the asymptotic relative dispersion of the volume through the choice @xmath170 , exhibit also semiclassical behavior at the bounce , in the sense that the relative dispersion in volume is bounded and small .
even more , the relative fluctuation at the bounce is approximately @xmath169 of the asymptotic value .
these analytical results give support to the numerical explorations reported in @xcite .
having shown that a semiclassical quantum state behaves also semiclassically at the bounce , it is then natural to ask whether the state yields observables through their expectation values that follow some classical trajectories .
it turns out that it does , but the ` classical theory ' it approaches is not the standard classical theory ( gr+ massless scalar field ) , but rather an _ effective theory _ , as defined by an effective hamiltonian constraint @xmath171 @xcite , which contains the _ quantum geometry scale _ @xmath172 .
the effective dynamics , is generated by the effective hamiltonian constraint @xmath173 and has several important features .
the first one is that all trajectories have a bounce that occurs at the same critical density @xmath174 .
the volume reaches a minimum value at the bounce that depends on the observable @xmath21 : v_min = p _ . in the quantum theory ,
we have seen that every physical states has a bounce where the minimum volume is given by @xmath175 .
it is then natural to find the corresponding volume at the bounce for our family of gaussian states .
it is straightforward to see that the minimal value of the volume , for the case @xmath89 , is [ vmin ] v_bounce=4 ^ 2 e^^2/2 ^ 2 k_0 . which is of the form , @xmath176 if we recall our previous result that @xmath177 , and assume the semiclassically conditions found in previous sections , namely that @xmath178 and @xmath95 ( as is the case when @xmath179 and @xmath148 ) in eq .
, we can compare the last two equations to see that @xmath180 . using the value @xmath181
, we conclude that v_bounce v_min , which is what we wanted to show .
note that the higher the value of @xmath82 , the better the approximation becomes .
in other words , if @xmath82 were not large enough ( compared to @xmath142 ) , not only would the asymptotic relative fluctuation of volume be large , but the effective equations would fail to be a good approximation to the quantum dynamics .
this seems to indicate that quantum gaussian states with a higher value of @xmath82 behave more classically , in contradiction with the classical intuition that tells us that ` rescaling of @xmath82 ' is physically irrelevant=0 frw model has a rescaling symmetry of the equations of motion , and of the underlying spacetime metric given by @xmath182 for @xmath183 a constant .
this can be understood as coming from the freedom of choosing arbitrary fiducial cells for the formulation of the theory @xcite . ] .
thus , we have to conclude that this rescaling symmetry of the classical theory is broken in the quantum theory .
note however , that even when the exact symmetry is broken , one might expect to regain it in the limit of large @xmath82 which , as we have seen , corresponds to @xmath184 . for an in - depth discussion regarding this scaling freedom see @xcite .
another important result for this solvable model is that the energy density @xmath185 is absolutely bounded by the critical density @xmath186 @xcite .
it is then natural to ask what the behavior of the energy density at the bounce is , for semiclassical states .
one might imagine that , for instance , the more semiclassical the state , the higher the density at the bounce .
in this part we shall explore this question by taking several quantities to measure density .
for instance , the simplest one would be @xmath187 , a quantity shown to be bounded by @xmath186 @xcite .
it is straightforward to find this quantity for the @xmath89 states , = _
crite^-^2/^2(1 + ^24 k_0 ^ 2 ) ^2 ( 1 + ^2 + ^24k_0 ^ 2 ) ^2 note that @xmath188 and one approaches it as @xmath189 , assuming again that we have chosen @xmath190 and @xmath148 .
one can also see that , if we were to fix @xmath82 and leave @xmath83 free , the density grows and approaches @xmath186 as @xmath83 grows .
thus , asking for the density to approach the critical density is not a very stringent condition on @xmath83 as the relative fluctuations in volume was .
note also that , if @xmath82 was small enough to be of the order of @xmath142 , the density at the bounce would be far from the critical density giving yet another indication that those states are not semiclassical .
the other quantities representing density that one can build , by taking for instance @xmath191 and @xmath192 , give expressions that have the same qualitative behavior , approaching @xmath186 as @xmath82 grows . of @xmath20 for which one reaches the classical region are plotted as a function of @xmath82 , where @xmath90=0 , @xmath193 and @xmath158 at @xmath76 . in the plot , @xmath194 dotted , @xmath195 dashed , @xmath196 dotdashed , @xmath197 down line , @xmath198 middle line , and @xmath199 the top line .
note that for high values of @xmath82 , the time @xmath200 remains a constant . ] , as function of @xmath82 , where @xmath90=0 , @xmath193 and @xmath158 at @xmath76 .
here , @xmath201 for which @xmath194 dotted , @xmath195 dashed , @xmath196 dotdashed , @xmath197 up line , @xmath198 middle line , and @xmath199 the bottom line . ] in the lqc description of the dynamics as governed by the effective hamiltonian , one can unambiguously ask when one recovers the gr classical dynamics .
one way of measuring this could be in terms of the variable @xmath26 ( when @xmath202 ) , or in terms of the energy density ( @xmath203 ) .
however if one wants to ask a similar question in the quantum realm , one immediately faces the following dilemma .
as we have seen in the previous part , a gaussian state remains semiclassical across and after the bounce .
in fact , the closer to the bounce , the more semiclassical it is .
we can not therefore ask when does the state become semiclassical , based on the behavior of fluctuations of the relevant operators .
still , we would like to have the notion of a transition from a quantum gravity dominated regime to a the regime where the quantum gravity effects are small and one is in the classical regime " .
for this purpose , we want to propose such a criteria , and define the ( percentual ) error between the relative fluctuation and the asymptotic relative fluctuation , ( ) : = .(v)^2v ^2|_-_+ _ + 100 .
we expect that this error will be small when we are approaching the classical region .
let us now motivate this proposal .
while we can not define a transition in terms of a change in the relative fluctuations of the basic operators , we could compare them to those of the standard ` wheeler dewitt ' model .
for we know that the expectation values in the wdw theory follow closely the classical dynamics of general relativity .
there is indeed a canonical way of mapping a lqc state to a wdw state , so the comparison of expectation values is well defined @xcite .
we have done that for our gaussian states and computed the corresponding relative fluctuation for the volume .
it turns out that , for the wdw dynamics , the relative fluctuation is _ constant _ in @xmath20 evolution , and corresponds precisely to the asymptotic relative dispersion of lqc .
thus , the lqc evolution not only approximates the wdw one in terms of expectation values but also in terms of the relative fluctuations .
it is then justified to measure the transition to the ` classical regime ' in terms of how close the lqc evolution is to the corresponding wdw dynamics . in order to study where the classical region begins , we found the value of @xmath20 for which @xmath204 with @xmath205 , taking @xmath76 at the bounce .
the values @xmath206 of @xmath20 that satisfy these conditions are plotted in figure [ fig:1.3 ] .
then we evaluated the densities @xmath207 , @xmath208 , @xmath209 , @xmath210 , at this values of @xmath20 .
the density is plotted in figure [ fig:1.4 ] .
this plot does not depend of the used density ( @xmath211 ) . from figure [ fig:1.4 ]
we can see that ( ) _ crit~()5 , [ dens - delta ] for @xmath212 and @xmath89 .
these results are similar if @xmath213 .
furthermore , if we choose @xmath214 , the only change is in the value of @xmath20 at the bounce .
if we change the value of @xmath83 for a fixed @xmath82 , then for larger ( than @xmath215 ) values , the time @xmath216 of arrival to the classical region is smaller .
in that case , the value of the density is closer to the critical one , and in the limit of large @xmath83 the classical region is ` at the bounce ' , which does not make much sense . in this way
the need to bound @xmath83 to be near the value @xmath215 is manifested .
the relation given by eq .
( [ dens - delta ] ) is telling us the order of the quantum corrections .
thus , if we say that the quantum corrections are negligible when , say , @xmath217 then the density will be of the order @xmath218 . recall that gaussian states are symmetric about the bounce , namely @xmath219 , therefore all the conclusions of this part equally apply before the bounce .
let us now summarize this section .
we have seen that appropriate physical conditions imposed on the behavior of basic physical observables are enough to constraint the parameters that characterize the generalized gaussian states ( [ coh_state ] ) , and obtain truly semiclassical states .
of the three continuous parameters @xmath220 that characterize the states , two of them are fixed by semiclassicality considerations , while the choice of @xmath11 represents a true choice of _ initial condition_. we have seen that asking that the relative fluctuations of the observable @xmath114 be small imposes that @xmath95 .
smallness of the asymptotic relative fluctuation of the volume implies then that @xmath221 .
if these conditions are satisfied , we saw that the semiclassical states at ` late times ' remain semiclassical across the bounce and that the volume and density at the bounce are very well approximated by the effective theory .
we have also seen that the scaling symmetry of the classical theory is not present , within the class of states under consideration , when @xmath82 is of the order of @xmath142 . however , in the limit of large @xmath82 , all the properties of the state remain invariant in terms of being well approximated by the classical effective theory. this leads us to conclude that the scaling symmetry is approximately recovered for large @xmath21 " ( for details regarding this issue , see @xcite ) .
still , the gaussian states we have considered are not the most general states one can consider . in the following section we shall explore the
so called _ squeezed states _ , a generalization of the gaussian states and compare their semiclassicality properties to those of the gaussian states .
in this section we consider states that generalize the gaussian states considered so far
. the strategy will be to consider squeezed states that are nearby the gaussian semiclassical states and compare their properties . in particular , we would like to know if the squeezed states can improve the samiclassicality properties such a better behavior of relative fluctuations of volume or a better approximation to the ` effective theory ' .
the generalized gaussian state we have considered so - far had three free ( real ) parameters @xmath222 and one discrete parameter ( @xmath90 ) .
we saw that , given a point of the physical phase space , we could approximate the internal dynamics of the theory by a choice of @xmath11 , and that semiclassicality imposes conditions on the two other parameters @xmath82 and @xmath83 .
let us now take the initial states as [ sq - state ] f(k)=\ { rl k^n e^-(k-)^2 , & k>0 , , + 0 , & k0 , . depending on two complex parameters @xmath223 .
one can reduce to a gaussian state from a squeezed state by setting m ( ) = : _ i = 0 , e()= : _
r=1 ^ 2 , e ( ) = : _ r = k_0 , 2_r_i = p .
thus , we see that the extra parameter in the definition of the states if given by the imaginary part of @xmath224 . the norm of this state in @xmath0 space is @xmath225 using relation for @xmath226 the last integral can be written as _
-^k^2n+1e^-2_r(k - a)^2 k , where a = _ r-_i_i_r .
if we want the convergence of the integral then @xmath227 and if we want to select the positive frequency then @xmath228 i.e. @xmath229 . with this assumptions the integral
is well defined , and the error is small , as shown in appendix [ app : b ] .
if we take @xmath89 and use the change of variables @xmath230 we have _ phy^2=2e^2_r(a^2-_r^2+_i^2)e^4_i_i_r _
-^(u+a)e^-2_ru^2 k .
then [ norm ] _
phy^2=a e^2_r(a^2-_r^2+_i^2)e^4_i_i_r .
we can compute the expectation value and dispersion of the fundamental observable @xmath231 .
p_= 2_phy^2 _
-^k^2n+2 e^-|(k-|)^2e^-(k-)^2 k .
if we use eq . , with @xmath226 , then we obtain p_= 2_phy^2 e^2_r(a^2-_r^2+_i^2)e^4_i_i_r _
-^k^2n+2e^-2_r(k - a)^2 k .
if we take @xmath89,p_= 2_phy^2 e^2_r(a^2-_r^2+_i^2)e^4_i_i_r .
this becomes , using eq .
( [ norm ] ) , p_=a .
now we want to calculate the expectation value of @xmath232 .
[ p2-int - sq ] p_^2 = 2 ^ 2_phy^2 _
-^k^3 e^-|(k-|)^2e^-(k-)^2 k . if we now use eq .
( [ norm ] ) , the result is for @xmath89 p_^2 = ^2 .[p2-sq ] from this we can find the dispersion and get , ( p_)^2= ^24_r -^216 _ r^2 a^2 . writing this explicitly in the original variables is ( p_)^2= ^24_r -^216_r^2(_r-_i_i_r ) ^-2 .
we shall compute the quantities @xmath233 and @xmath234 from which one can find the expectation value and dispersion of the volume operator .
the relevant coefficients are v_&= & _ -^(-i k ) k^ne^-|(k-|)^2 i [ ki ] [ ki]^n e^-(ki-)^2 k + & = & _ -^k^n+1[ki]^n+1e^-|(k-|)^2 e^-(ki-)^2 k . with @xmath235 . as in previous subsections we will only write down explicitly the @xmath89 case . this expression can be put on the form , [ volume ] v_=awhere @xmath236 . in the gaussian variables it takes the form v_&=&(k_0-_ip^42 ) ^-1 + & & which reduces to the gaussian case when @xmath237 .
the next quantities we need to compute in order to find the dispersion of the volume operator are the coefficients @xmath238 .
the first one is given by w_0=2_2_-^k^3|f(k)|^2 k , _
2=2 ^ 2 ^4 ^23 g .
this is the same integral that the one in eq . .
then , in order to consider the proper normalization for the squeezed state with @xmath89 , we only need to replace @xmath123 by @xmath124 into eq . , the result is w_0=_2 .
the other quantities are given by the expression , w_&=&_3_-^k^2 k , + & = & 2_3e , with @xmath239 and @xmath240 .
this integral can be performed for @xmath89 and we get w_=_3 , where @xmath241 .
it is now straightforward to find @xmath144 , the asymptotic relative dispersion for volume , _ & = & a^24 ^2 - 1 , .
let us now explore those conditions that select semiclassical states .
in this part , we want to see how the semiclassicality conditions are satisfied if we move along the ` squeezing parameter ' @xmath242 , off the gaussian semiclassical states , corresponding to @xmath237 . instead of performing an exhaustive analysis of this question ,
as we have done in previous sections for gaussian states , we have computed the relevant quantities and plotted them . .
in the plot @xmath243 , @xmath244 , @xmath245 , @xmath246 and @xmath89 .
the maximal value of the density correspond to the gaussian states .
is clear that the squeezed states have a maximal density close to @xmath186 only if the state is near to the gaussian states . ]
0.4 cm recall that for gaussian states , asking that the asymptotic relative dispersion in volume be a minimum , selected an optimal ` width ' @xmath247 for the gaussian .
this value had also the property that the density at the bounce was very close to the critical density @xmath248 . in fig .
[ fig:3 ] we have plotted the asymptotic relative dispersion and density at the bounce as functions of @xmath83 and @xmath242 .
the @xmath237 slice corresponds then to the values found for gaussian states .
it can not be appreciated in the figure but , for a fixed @xmath83 , the asymptotic relative dispersion @xmath144 , as a function of @xmath242 does not attain its minimum at zero , but rather at some value @xmath249 for @xmath250 ( and @xmath251 for @xmath252 ) .
thus , if one chooses , for the parameter of the squeezed state the value @xmath253 , one is ` improving ' the asymptotic relative dispersion after the bounce , and , at the same time , increasing the value of the asymptotic dispersion _ before _ the bounce . in this process ,
one is introducing an asymmetry in the fluctuations .
this fact has been used to suggest the possibility , for example , that the asymmetry could be so large to spoil the semiclassicality properties across the bounce . as we shall see below
, this possibility is however , not realized .
what one can indeed see from the figure is that , as we move further away from the gaussian states , that is , as we increase to values @xmath254 , the asymptotic relative dispersion increases steeply , for both signs of @xmath242 .
as can be seen from the figure , the density at the bounce ( in blue ) decreases as we go away from @xmath237 in both directions , reaching densities much less than the critical value very fast . , @xmath255 , @xmath245 , @xmath246 and @xmath89 .
the global minimum is over the gaussian states and has an exponential behavior if we move far away from the gaussian state or if @xmath156 is smaller that @xmath256 . ]
both phenomena suggest that we can indeed consider squeezed states as semiclassical states _
provided we remain very close to the gaussian states_. the question is then how close we can be in order to still have semiclassical behavior as exhibited by the gaussian states . for this purpose ,
we have plotted in fig .
[ fig:2 ] the volume at the bounce as function of @xmath83 and @xmath242 .
we again see that , as was the case for the density at the bounce , the volume at the bounce has indeed a minimum at the gaussian states , namely , for @xmath237 . using both relative volume and density at the bounce
, we can then define an interval , as function of @xmath83 , in which the parameter @xmath242 can take values and the state still be considered as semiclassical . more precisely , we define some tolerance for the asymptotic relative dispersion @xmath257 , and find the maximum value of @xmath258 for which the relative dispersion is below the value @xmath259 .
we then find the value of the density at the bounce for the value @xmath258 .
we have made extensive explorations for different values of @xmath259 and @xmath82 , and have found that , for a fixed value of @xmath259 , the larger the value of @xmath82 , the smaller the allowed interval in @xmath242 .
for example , for @xmath260 , the dependence of @xmath258 on @xmath82 can be approximated ( for large @xmath82 , @xmath170 , @xmath89 and @xmath158 ) as @xmath261 .
the value of @xmath262 at that point is equal to @xmath263 .
( for @xmath264 , @xmath265 and @xmath266 . ) . in the plot @xmath267 , @xmath268 , @xmath245 , @xmath246 and @xmath89 . outside of the vertical one
the squeezed state have an asymptotic relative fluctuations ( @xmath250 and @xmath252 ) bigger that 1 .
one line mark the gaussian states ( @xmath269 ) that is just when the curve is zero and the other line is for states with @xmath256 . ]
let us now return to the issue of the asymmetry across the bounce of the relative dispersion in volume .
as we have mentioned , without any control on the value of the parameters , one could in principle have a large asymmetry that translates into a loss of semiclassicality across the bounce .
this scenario was suggested in @xcite , but where the analysis was made for an effective formalism not directly derived from lqc and with little control on the semiclassicality conditions .
here we are better equipped to make precise statements about this asymmetry based on the control we have on analytical expressions and the possible values the parameters of the states can have . in fig .
[ fig:4 ] we have plotted the relative difference @xmath270 in asymptotic relative dispersion , as a function of @xmath83 and @xmath242 .
we have included , as vertical surfaces , the boundary of the allowed region in parameter space .
two slices in the figure are worth mentioning .
the curve @xmath237 , which is viewed almost vertical in the figure corresponds to the gaussian states . on that curve , @xmath219 so the relative difference vanishes .
the second highlighted curve corresponds to the choice @xmath271 , which is the extension to @xmath272 of the optimal gaussian states .
note that this curve starts at zero , increases ( for positive @xmath242 ) , reaches a maximum and then decreases tending to zero .
the curve is antisymmetric with respect to @xmath242 .
thus , there are two points for which @xmath273 reaches a maximum .
what we have seen , by taking many different values of @xmath82 , is that this maximum value is indeed , extremely small .
for instance , for @xmath274 , the maximum is about @xmath275 .
as we increase @xmath82 this value decreases even further .
these results show that indeed , semiclassical states are very symmetric and that semiclassicality is preserved across the bounce .
one should also note that this quantitative analysis invalidates the claims of @xcite , and @xcite regarding the allowed asymmetry in fluctuations across the bounce and ` cosmic forgetfulness ' , and supports the results of @xcite and @xcite .
for instance , in the first reference of @xcite , it is claimed that the relative change in relative asymptotic dispersion in volume can be as large as 20 for semiclassical states , while we have demonstrated that this quantity is several orders of magnitude smaller for realistic values of the parameter @xmath82 ( of the order of @xmath275 for @xmath274 , and much smaller for the values expected to represent realistic universes ( @xmath276 or larger ) ) .
let us summarize our results .
we have defined coherent gaussian states as candidates for semiclassical states ` peaked ' around points of the classical ( physical ) phase space .
we have imposed consistency conditions on such states by asking that the relative dispersion of the volume be small at late times after the bounce .
this condition implies that one of the parameters , yielding the momentum of the scalar field , be large and fixes one of the other parameters .
we can therefore find a canonical gaussian state for a given point of the classical phase space corresponding to a classical solution at late times .
when exploring the properties of these states near the quantum bounce we found that they behave also semiclassically in the deep quantum region ; the relative fluctuations of volume are of the same order and smaller than in the asymptotic , large volume , region .
furthermore , for those states that have small fluctuations , the expectation value of the volume and energy density at the bounce are very well approximated by the so called effective theory , thus giving support to the claim that this theory is a good approximation to the quantum dynamics , even in the deep planck regime .
next , we introduced squeezed states , a one parameter generalization of gaussian states and studied their properties in a vicinity of the semiclassical gaussian states . as we showed , the range of this extra parameter is severely restricted if we want to maintain a small relative fluctuation for the volume .
furthermore , if one departs from the gaussian states too much , both the volume and density at the bounce differ very rapidly from the value on the gaussian states ( and also the ` effective theory ' ) .
thus , we are lead to conclude that the gaussian states exhibit a better semiclassical behavior than squeezed states .
an important issue that can be studied quantitatively in this solvable model is that of the asymmetry on the volume fluctuations .
on gaussian states the fluctuations are the same before and after the bounce . as one introduces the squeezing parameter , the fluctuations are no longer symmetric .
in fact , depending on the choice of sign of the parameter , either one of the fluctuations becomes smaller and reaches an absolute minimum in the vicinity of the gaussian states .
at the same time , the relative fluctuation on the opposite side of the bounce increases .
the question is then how big can this asymmetry be ?
what we saw is that for a given , somewhat arbitrary choice of parameters , that is still far from those giving rise to a realistic ` large ' universe , the relative difference in relative dispersion is very small ( of the order of @xmath275 ) , and decreases for more realistic values of the parameters .
thus , even if we were to choose the squeezing parameter such that this asymmetry is maximized , the relative change is so small that , for all practical purposes , the state is symmetric . a state that is semiclassical on one side of the bounce not only remains semiclassical on the other side , but maintains its coherence .
this quantitative analysis therefore invalidates claims of ` loss of coherence ' across the bounce @xcite , where it was claimed that the relative change in ( relative ) dispersion can be several orders of magnitude larger .
one particular feature of the classical description of the system under consideration is that it possesses a rescaling symmetry . in the spacetime description , we can rescale the volume by a constant and the physical properties of the spacetime remain invariant . in the phase space description
this is manifested by a symmetry in the equations of motion . this symmetry ,
however , can not be treated as _ gauge _ , in the same sense that hamiltonian symmetries are .
one can understand the origin of this symmetry by recalling that the @xmath0=0 model needs a fiducial volume for its hamiltonian description , and this choice is completely arbitrary . in turn , this means that there is no scale with respect to which one could compare the momentum @xmath21 of the scalar field ; there is no meaning to the statement that ` @xmath21 is large ' .
as we have seen , in the quantum theory planck constant introduces such a scale and this is manifested in the behavior of certain quantum observables .
one of the semiclassicality conditions we have found is that the larger the expectation value of @xmath277 , the more semiclassical the state is .
what we have also seen is that the rescaling symmetry is only approximately recovered , for our class of states , in the ` very large @xmath278 limit ' , which can be taken as the _ classical limit _ of the system . for further details on this issue ,
see @xcite .
we thank a. ashtekar , y. ma , t. pawlowski and p. singh for helpful discussions and comments .
this work was in part supported by dgapa - unam in103610 , by nsf phy0854743 and by the eberly research funds of penn state .
the kinematical phase space of the model is @xmath279 with coordinates @xmath22 . here
@xmath20 is the matter scalar field , @xmath21 its canonical momentum , @xmath12 the volume ( of the fiducial cell ) and @xmath26 its canonical momentum .
the effective hamiltonian constraint of the system is @xcite : @xmath280 which is labeled by a real parameter @xmath31 with dimensions of length .
the symplectic structure is given by : [ forma ] _
0=14gv+p_which induces the following poisson brackets @xmath281 the hamiltonian constraint defines a hypersurface @xmath282 in phase space .
it can be taken as a disjoint union of hypersuperfaces where @xmath283 with @xmath284 . in what follows
we shall only consider the connected component where @xmath285 .
note that the points where @xmath286 , @xmath284 are not in the constrained surface .
these are degenerate points ( hyperplanes ) where the system does not have a well defined dynamics .
we shall see that physically , these points should be excluded .
we shall now consider the pullback @xmath287 of the symplectic structure to the hypersurface @xmath288 .
that is , we find @xmath287 such that @xmath289 , for @xmath290 the embedding of @xmath288 into @xmath291 .
let us write @xmath292 then , on the constrained surface we have [ p - on - surface ] p_v ( ) the pullback of the gradient of @xmath293 to @xmath288 yields , ^2 2v v ( ) ^2 + 2 ^ 2 v^2 ( ) ( ) 2 p_p _
+ p_which can be used for solving @xmath294 , @xmath295 or @xmath296 and inserting back in eq .
( [ forma ] ) .
each of this choices will depend on our choice of coordinates for the constrained surface .
let us study two of such parametrizations . in the first one we solve for @xmath21 , which means that we use @xmath297 as coordinates .
then @xmath298\wedge \d\phi\ ] ] which gives us the expression for the two disconnected hypersuperfaces parametrized by positive and negative values for @xmath21 .
the value @xmath299 is excluded if @xmath300 or @xmath301 , with @xmath284 .
if we take @xmath302 we get @xmath303 if we now solve for @xmath295 , we get the expression for the pre - symplectic structure in the @xmath304 parametrization as , @xmath305 we can again restrict ourselves to @xmath302 , which yields @xmath306 on the corresponding connected component .
regardless of the parametrization chosen , the pre - symplectic form @xmath287 has a degenerate direction corresponding to the hamiltonian vector field @xmath307 of the hamiltonian constraint @xmath293 .
physically the integral curves represent _ gauge _ direction along which points are physically indistinguishable .
we can find the gradient of the constraint , @xmath308 and , using that @xmath309 , and @xmath310}}{\partial v } + { \frac}{\partial^{[a}}{\partial\phi}{\frac}{\partial^{b]}}{\partial p_\phi } \right)\ ] ] we get , @xmath311 as the restriction of the vector field to the constrained hypersurface with coordinates @xmath312 , using eq . .
it is easy to check that this is indeed the degenerate direction of @xmath287 : @xmath313 .
we are now interested in understanding qualitatively the structure of the reduced phase space , for which we would need to take the quotient of @xmath288 by the gauge directions generated by @xmath314 . in this case
, it turns out to be easier to perform a gauge fixing which , as we shall see , is well defined everywhere .
let us define @xmath315 , with @xmath316 .
it is direct to see that it forms , with @xmath293 a second class pair , @xmath317 provided @xmath318 , which is a condition we had asked before .
the gauge fixed phase space @xmath319 is therefore parametrized by @xmath320 ( or @xmath321 ) with @xmath322 , @xmath323 or @xmath324 and @xmath285 .
the corresponding symplectic structures on each of this two different parametrizations are @xmath325 if we introduce the new variable ( as was done in the main text ) , @xmath326 then the symplectic form @xmath327 becomes @xmath328 which indicates that @xmath329 can be seen as conjugate variables in the reduced theory .
here we list some integrals that are useful in the main part of the manuscript .
_ -^e^-x^2x= _ -x_0^x_0 x^2n+1 e^-x^2x=0 , n _
-^x^2n e^-x^2x= ( -1)^n^n ^n , n some error integrals .
_ z^e^-x^2x = 2erfc(z ) , z _ z^x^2e^-x^2x = z2e^-z^2 + 4erfc(z ) , z where @xmath330 is the complementary error function erfc(z)=1-erf(z)=1 - 2_0^z e^-x^2x which can be bounded using its asymptotic expansion @xcite as [ erfc_approx_complex ] @xmath331 .
when @xmath332 is real then [ erfc_approx ] erfc(x)<e^-x^2x -|(k-|)^2-(ki t -)^2 & = & -2_r(k - ait2_r)^2 + & & + ( t^2_r22t_i_r)(1+_i^2_r^2 ) + & & + 2_r(a^2-_r^2+_i^2)+4_i_i_r [ expantion ] [ product ] = u^2 + 2u b_+b_^2+^24
we show in explicit form the errors for the norm and volume and explain why the approximations we take throughout the manuscript are completely justified .
the norm of the gaussian states is given by [ norm_k ] _
phy^2&=&2_0^k k^2ne^-2(k - k_0)^2 /^2 k + [ norm_int2 ] & = & 2 ^ 2n+2 _ -k_0/^^2n+1 e^-u^2 u which for the @xmath89 case becomes , _
phy^2=2 ^ 2 _ -y^e^-u^2 u , y= k_0 . then _
phy^2 & = & ^2 _ -y^u e^-u^2 u+^2_-y^y e^-u^2 u + & = & ^22e^-y^2+^2 y-^2 y2erfc(y ) . using the form of @xmath333 , we get _ phy^2= k_0+^22e^- ( k_0)^2 -k_0 erfc ( k_0 ) . the norm that we used in the manuscript was [ norm0-ap ] _ phy^2 k_0 . then the error in this approximation is just ( _ phy^2)=^22e^- ( k_0)^2 -k_0erfc ( k_0 ) .
this tells us that the error in the norm for @xmath89 is the error in the approximation of the function @xmath334 by the function @xmath335 . if we use eq . to bound the function @xmath336
then the error can be bounded by ( _ phy^2 ) < ^22e^- ( k_0)^2 .
+ since we have seen in the main text that semiclassicallity implies that we take @xmath337 then this error is indeed very small .
these errors were also studied numerically for @xmath213 and present similar behavior .
we can then conclude that the approximation made is very good when computing the norm of the states .
the coefficients in the expectation value for volume are , v_&= & e^p_0^^n+1 e^-(k - k_0)^2 /^2e^-(ki - k_0)^2 /^2
k + [ vol_int ] & = & ( 14)^n+1 e^p e^^2/2 ^ 2_-u_0^^n+1 e^-u^2/^2u , with @xmath338
. then for @xmath89 are v_=14 e^p e^^2/2 ^ 2_-u_0^e^-u^2/^2u .
now we introduce other change of variables @xmath339 , and then the integral takes the form v_=4
e^p e^^2/2 ^ 2_-z^e^-t^2/t , with @xmath340 .
the integral is i_&=&_-^(2 ^ 2 t^2 + 4k_0t+4k_0 ^ 2 + ^2 ) e^-t^2 t - _ -^-z ( 2 ^ 2 t^2 + 4k_0t+4k_0 ^ 2 + ^2 ) e^-t^2 t + & = & k_0 .
then @xmath341 takes the form v_&=&8
e^p e^^2/2 ^ 2 k_0 .
if we normalize using eq . then v_&=&8
e^p e^^2/2 ^ 2 the values that we used in the manuscript was [ vpm_02 ] v_k_02 e^p e^^2/2 ^ 2 .
then the error in the integration is ( v _ ) & = & -8 e^p e^^2/2 ^ 2 using the triangular inequality |error ( v_)| & & 8 e^p e^^2/2 ^ 2 using the relation ( where @xmath342arg@xmath343 because @xmath344 ) |error ( v_)| & < & 8 e^p e^^2/2 ^ 2e^-|z|^2 + & < & 8
22 ^2 ^2 k_0 |z| e^p e^^2/2 ^ 2e^-|z|^2 as all quantities in the brackets are positive and less than one , we can bound the error as @xmath233 can be too small or large then the quantity that really gives a measure of the error is the relative error @xmath345 . using eq .
, this becomes & < & 5 ^ 2 ^ 2 |z| e^-|z|^2 ( 1+^2+^24k_0 ^ 2 ) ^-1 + & < & 5 ^ 2 k_0 e^-2k_0 ^ 2/^2 e^-^2/2 ^ 2 ( 1+^24k_0 ^ 2 )
^1/2(1+^2+^24k_0 ^ 2 ) the relation between the parentheses is less than one , then & < & 5 ^ 2 k_0 e^-2k_0 ^ 2/^2 e^-^2/2 ^ 2 since @xmath346 , then @xmath347 .
furthermore , as @xmath348 , then the negative exponential wins over the polynomial factor , which show that the error is small .
these errors were also studied numerically for @xmath213 and present similar behavior .
in this appendix we summarize the main quantities that were found in the main text , for the generalized gaussian states for @xmath112 and compare them to the @xmath89 case .
it should be noted , that for higher values of @xmath90 , the structure of the terms is similar , with terms of the form @xmath349 and @xmath350 .
m. bojowald , loop quantum cosmology " , _ living rev .
rel . _ * 8 * , 11 ( 2005 ) arxiv : gr - qc/0601085 ; a. ashtekar , m. bojowald and l. lewandowski , mathematical structure of loop quantum cosmology " _ adv .
phys . _ * 7 * 233 ( 2003 ) arxiv : gr - qc/0304074 .
a. ashtekar , `` loop quantum cosmology : an overview , '' gen .
grav . * 41 * , 707 ( 2009 ) arxiv:0812.0177 [ gr - qc ] .
a. ashtekar and j. lewandowski `` background independent quantum gravity : a status report , '' class .
* 21 * ( 2004 ) r53 arxiv : gr - qc/0404018 ; c. rovelli , quantum gravity " , ( cambridge u. press , 2004 ) ; t. thiemann , `` modern canonical quantum general relativity , '' ( cambridge u. press , 2007 ) .
a. ashtekar , t. pawlowski and p. singh , `` quantum nature of the big bang , '' phys .
lett * 96 * ( 2006 ) 141301 arxiv : gr - qc/0602086 .
a. ashtekar , t. pawlowski and p. singh , `` quantum nature of the big bang : improved dynamics , '' phys . rev .
d * 74 * , 084003 ( 2006 ) arxiv : gr - qc/0607039 .
l. szulc , w. kaminski , j. lewandowski , `` closed frw model in loop quantum cosmology , '' class.quant.grav .
* 24 * ( 2007 ) 2621 ; arxiv : gr - qc/0612101 .
a. ashtekar , t. pawlowski , p. singh , k. vandersloot , `` loop quantum cosmology of k=1 frw models , '' phys.rev .
d * 75 * ( 2007 ) 024035 ; arxiv : gr - qc/0612104 .
e. bentivegna and t. pawlowski , `` anti - desitter universe dynamics in lqc , '' arxiv:0803.4446 [ gr - qc ] .
a. ashtekar , t. pawlowski , p. singh , pre - inflationary phase in loop quantum cosmology , " ( in preparation ) .
d. w. chiou and k. vandersloot , `` the behavior of non - linear anisotropies in bouncing bianchi i models of loop quantum cosmology , '' phys .
d * 76 * , 084015 ( 2007 ) arxiv:0707.2548 [ gr - qc ] ; d. w. chiou , `` effective dynamics , big bounces and scaling symmetry in bianchi type i loop quantum cosmology , '' phys .
d * 76 * , 124037 ( 2007 ) .
arxiv:0710.0416 [ gr - qc ] .
a. ashtekar , e. wilson - ewing , `` loop quantum cosmology of bianchi i models , '' phys . rev . *
d79 * , 083535 ( 2009 ) .
arxiv:0903.3397 [ gr - qc ] . a. ashtekar and e. wilson - ewing , `` loop quantum cosmology of bianchi type ii models , '' phys . rev .
d * 80 * , 123532 ( 2009 ) arxiv:0910.1278 [ gr - qc ] .
e. wilson - ewing , `` loop quantum cosmology of bianchi type ix models , '' phys .
d * 82 * , 043508 ( 2010 ) arxiv:1005.5565 [ gr - qc ] . m. martin - benito , g. a. m. marugan and e. wilson - ewing , `` hybrid quantization : from bianchi i to the gowdy model , '' phys .
d * 82 * , 084012 ( 2010 ) arxiv:1006.2369 [ gr - qc ] .
a. ashtekar , a. corichi and p. singh , `` robustness of key features of loop quantum cosmology , '' phys .
d * 77 * , 024046 ( 2008 ) .
arxiv:0710.3565 [ gr - qc ] . a. corichi and p. singh , `` quantum bounce and cosmic recall , '' phys .
* 100 * , 161302 ( 2008 ) arxiv:0710.4543 [ gr - qc ] ; phys .
* 101 * , 209002 ( 2008 ) arxiv:0811.2983 [ gr - qc ] .
w. kaminski and t. pawlowski , `` cosmic recall and the scattering picture of loop quantum cosmology , '' phys .
d * 81 * , 084027 ( 2010 ) arxiv:1001.2663 [ gr - qc ] . m. bojowald , `` what happened before the big bang ? , '' nature phys .
* 3n8 * , 523 ( 2007 ) ; `` dynamical coherent states and physical solutions of quantum cosmological bounces , '' phys . rev .
d * 75 * , 123512 ( 2007 ) arxiv : gr - qc/0703144
. m. bojowald , `` harmonic cosmology : how much can we know about a universe before the big bang ?
, '' arxiv:0710.4919 [ gr - qc ] ; `` quantum nature of cosmological bounces , '' arxiv:0801.4001 [ gr - qc ] .
a. corichi and p. singh , `` is loop quantization in cosmology unique ?
, '' phys .
d * 78 * , 024034 ( 2008 ) arxiv:0805.0136 [ gr - qc ] ; `` a geometric perspective on singularity resolution and uniqueness in loop quantum cosmology , '' phys . rev .
d * 80 * , 044024 ( 2009 ) arxiv:0905.4949 [ gr - qc ] .
v. taveras , corrections to the friedman equations from lqg for a universe with a free scalar field " , phys .
d * 78 * , 064072 ( 2008 ) arxiv:0807.3325 [ gr - qc ] .
a. ashtekar , m. campiglia and a. henderson , `` path integrals and the wkb approximation in loop quantum cosmology , '' phys .
d * 82 * , 124043 ( 2010 ) arxiv:1011.1024 [ gr - qc ] .
a. ashtekar , l. bombelli and a. corichi , `` semiclassical states for constrained systems , '' phys .
d * 72 * , 025008 ( 2005 ) arxiv : gr - qc/0504052 ; b. bolen , l. bombelli and a. corichi , `` semiclassical states in quantum cosmology : bianchi i coherent states , '' class .
* 21 * , 4087 ( 2004 ) arxiv : gr - qc/0404004 .
a. corichi and e. montoya , on the semiclassical limit of loop quantum cosmology " , arxiv:1105.2804 [ gr - qc ] .
m. abramowitz , i.a .
stegun , _ handbook of mathematical functions _
dover publications , inc .
, new york . | the spatially flat friedman - robertson - walker ( frw ) cosmological model with a massless scalar field in loop quantum cosmology admits a description in terms of a completely solvable model . this has been used to prove that : i ) the quantum bounce that replaces the big bang singularity is generic ; ii ) there is an upper bound on the energy density for all states and iii ) semiclassical states at late times had to be semiclassical before the bounce .
here we consider a family of exact solutions to the theory , corresponding to generalized coherent gaussian and squeezed states .
we analyze the behavior of basic physical observables and impose restrictions on the states based on physical considerations .
these turn out to be enough to select , from all the generalized coherent states , those that behave semiclassical at late times .
we study then the properties of such states near the bounce where the most ` quantum behavior ' is expected . as it turns out ,
the states remain sharply peaked and semiclassical _ at the bounce _ and the dynamics is very well approximated by the ` effective theory ' throughout the time evolution .
we compare the semiclassicality properties of squeezed states to those of the gaussian semiclassical states and conclude that the gaussians are better behaved . in particular , the asymmetry in the relative fluctuations before and after the bounce are negligible , thus ruling out claims of so called ` cosmic forgetfulness ' . |
the galactic center ( hereafter gc ) region has been observed at different spatial scales and energy bands by many x ray missions .
the main results in the 0.510 kev range have been obtained with the einstein observatory ( watson et al . , 1981 ) , rosat ( predehl & trumper , 1994 ) and , more recently , with asca ( maeda et al . 1996 , 1998 ; koyama et al .
1996 ) . at higher energies , the telescopes art
p ( pavlinsky , grebenev & sunyaev 1994 ) and sigma ( goldwurm et al . 1994 ) on board the granat satellite performed a monitoring of the gc region leading to the discovery of many new sources .
these observations have shown that the gc x ray emission comes both from point sources and from a diffuse component . within a few degrees from the direction of the gc region
there is a strong concentration of x ray sources .
the zoo of these objects is very rich and comprises both transient and permanent sources , both standard " and very peculiar objects , like e.g. the `` bursting pulsar '' gro j174428 ( lewin et al .
1996 ) , the `` microquasar '' 1e 1740.72942 ( mirabel et al .
1992 ) , and the 2 ms pulsar sax j1808.43658 ( int zand et al . 1998a , wijnands & van der klis 1998 ) .
the region ( @xmath10)@xmath11(@xmath12 ) around the gc was observed with the @xmath2 satellite during 19961998 for a total of @xmath3140 hours of effective exposure time ( sidoli et al .
, 1998a , 1999 ) .
in this paper we report on the observations of the x ray point sources in the gc region obtained with the @xmath2 narrow field instruments ( for the @xmath2 wide field camera results see , e.g. , ubertini et al .
( 1999 ) ) .
although our observations can not equal the spatial resolution obtained with rosat at lower energies , they provide good images with adequate spectral resolution above a few kev .
the sensitivity is similar to that obtained with asca , that thanks to its solid state detectors has a higher spectral resolution . on the other hand ,
the bepposax mirrors provide a narrower and more regular point response function with respect to that of the asca telescopes .
this allows an accurate analysis of the spatial morphology of this complex sky region and reduces the problem of stray light contamination from strong sources outside the field of view .
the detailed results on the x ray diffuse emission from sgra east and west and from the molecular clouds in this region ( sgrb2 , sgrc , sgrd ) will be reported in a separate paper .
the @xmath2 satellite ( boella et al .
1997a ) carries a complement of several imaging and non - imaging x ray detectors , covering a broad energy range from 0.1 kev to 300 kev .
most of the results presented here have been obtained with the imaging instruments placed in the focal planes of four identical x ray mirror units : the medium
energy concentrator spectrometer ( mecs , boella et al .
1997b ) and the low energy concentrator spectrometer ( lecs , parmar et al .
each mirror unit consists of 30 nested , confocal mirrors with a double cone approximation to the wolter i geometry and a focal length of 185 cm .
the mecs instrument is based on three position sensitive gas scintillation proportional counters , providing 1.310 kev images over a field of view with @xmath13 radius .
the mecs is characterized by a total ( three telescopes ) effective area of @xmath3150 @xmath14 at 6 kev , a relatively good angular resolution ( 50% power radius of @xmath15 at 6 kev , on - axis ) and a moderate energy resolution ( fwhm @xmath38.5@xmath16% ) .
one ( m1 ) of the three nearly identical units composing the mecs had a failure in may 1997 ; all the observations performed after that date were carried out only with the remaining two units ( m2 and m3 ) . the fourth mirror unit is associated with the lecs instrument .
this consists of a position sensitive gas scintillation proportional counter utilizing an ultra thin , organic detector entrance window ( 1.25 @xmath17 m ) and a driftless configuration to extend the low energy response down to 0.1 kev .
its spatial and energy resolutions are similar to those of the mecs , but the field of view is slightly smaller ( @xmath18 radius ) . due to
uv contamination leaking through the entrance window during sunlit periods , the lecs is operated only at satellite night time , resulting in effective exposure times smaller than those of the other @xmath2 instruments ( typically @xmath350% of the mecs values ) . for the analysis of 1e 1740.72942 , which has a hard spectrum extending above the mecs energy range we have also used the phoswich detector system instrument ( pds , frontera et al .
this non - imaging instrument points in the same direction of the mecs and lecs telescopes and covers the 15300 kev energy range .
it consists of a square array of four indipendent nai(tl)/csi(na ) phoswich scintillation detectors equipped with collimators defining a field of view of @xmath31.3@xmath19 ( fwhm ) .
there are two separate collimators , one for each pair of detectors , that can be independently rocked off - axis to monitor the background . during our observations
the default rocking mode with offset angles of @xmath20 and dwell time 96 s was used .
the pds provides an energy resolution of @xmath315% at 60 kev ( fwhm ) .
the log of our @xmath2 observations of the gc region performed in 199698 is shown in table 1 .
the pointing directions were chosen to cover a @xmath21 long region at b=0@xmath19 , as well as a few known sources slightly outside the galactic plane .
we chose to partially overlap adjacent pointings , in order to continuously cover the galactic plane with the central part of the mecs field of view ( @xmath22 radius ) , where the instrument sensitivity and angular resolution are best .
as shown in table 1 , a few pointing directions were observed twice , at time intervals of 6 or 12 months .
this was partly planned , to study the long term variability of the sources , and partly resulted from operational constraints .
the exposure time achieved in the different regions is not uniform , ranging from a minimum of 7 ks to @xmath3340 ks at the position of 1e 1743.12843 . an exposure - corrected mosaic of the mecs images in the 210 kev energy range is shown in fig . 1 .
except where noted , the spectral analysis of all the point sources reported here has been carried out as follows : the lecs and mecs counts have been extracted from a circular region with @xmath23 radius centered at the position of the source and rebinned in order to have at least 20 counts per energy bin .
the @xmath23 radius corresponds to an encircled energy of 90% . in several cases we used a smaller extraction radius ( @xmath24 , @xmath360% enclosed energy ) in order to reduce the problems due to source crowding or unfavourable locations in the mecs field of view . in any case
the response matrices appropriate to the adopted extraction radius and source position on the detector have been used . while these matrices include the elements to properly derive the correct spectral shape of the sources observed off - axis , the resulting fluxes can be affected by an uncertainty in their absolute normalization that in some cases could reach @xmath340% . from each spectrum
we have subtracted a local background , extracted from a source free region of the same observation : in fact , the standard @xmath2 background spectra obtained from blank field observations underestimate the actual background present in the gc region , which is dominated by the galactic diffuse emission .
the pds data were reduced using the saxdas software package ( version 1.2.0 ) .
the time intervals before and after earth occultations were excluded and the spurious spikes ( guainazzi & matteuzzi 1997 ) filtered out .
the background subtraction was performed with the standard rocking collimator technique .
the background - subtracted spectra were rebinned in order to have at least 100 counts for each energy channel and then analyzed using the response matrix released on august 31 , 1997 .
the resulting lecs / mecs / pds spectra have been fitted with different models implemented in xspec ( version 10.00 ) .
we took into account the known intercalibration normalizations between the different instruments on
board @xmath2 ( fiore , guainazzi & grandi 1999 ) , using different relative normalizations [email protected] , a@xmath26=1 and [email protected] .
mecs spectra were extracted in the 210 kev energy range , while for the lecs we used only the data below 2 kev ( we verified that using the lecs data above this energy did not significantly improve the results obtained with the mecs ) .
several previously known sources have been observed in our @xmath2 pointings : 1e 1743.12843 ( watson et al . 1981 ) , the persistent black hole candidate 1e 1740.72942 ( sunyaev et al .
1991a ) , the x ray bursters slx 1744299 , slx 1744300 , a 1742294 , ks 1741293 ( skinner et al .
1990 ; pavlinsky , grebenev & sunyaev 1994 ; kawai et al .
1988 ) , and the source(s ) at the gc position ( watson et al .
we also detected two x ray transients discovered very recently : sax j1747.02853 ( int zand et al .
1998b ) and xte j1748288 ( smith , levine & wood 1998 ) , and discovered a new x ray source that we identified as the plerionic emission from the composite supernova remnant g0.9 + 0.1 ( mereghetti , sidoli & israel 1998 ) .
all the sources observed in our survey are listed in table 2 .
note that in table 2 we have reported the best source positions and corresponding uncertainties available in the literature .
the positions derived from our analysis have typical uncertainties of the order of @xmath28 ( 95% confidence level ) and are in agreement with those of table 2 .
we have listed in table 2 also the positions of a few other objects which are relevant in the following discussion of the sources near sgra*. the results for the individual sources are reported in the following sections , while in table 3 we have summarized the parameters of the best fits .
all the luminosities are given for an assumed distance of 8.5 kpc , unless stated differently .
the gc field was observed for 100 ks in august 1997 ( obs .
n.3 ) . strong x ray emission peaked at the gc position is clearly detected with the lecs and mecs instruments .
the central part of the mecs image is shown in fig . 2 , where the positions of the x ray sources previously detected with other satellites have been indicated .
the background subtracted radial profiles of the x ray emission in the soft ( 25 kev ) and hard ( 510 kev ) energy ranges are shown in fig .
the data points represent the surface brightness measured in concentric rings centered at the position of sgra * , while the solid lines show for comparison the profiles expected from a single unresolved point source .
the expected profiles have been normalized to yield the same number of counts within 10@xmath29 as the measured data .
it is clear that the shape of the observed profiles is incompatible with a single point
like source .
the observed radial profile results from the contribution of a diffuse component and one ( or possibly more ) sources close to the sgra * position . the x ray images of this region with the best angular resolution available so far are those obtained with the rosat pspc instrument ( predehl & trumper , 1994 ) which , in addition to the diffuse emission , showed the presence of three different sources . their error regions ( 20@xmath30 radius )
are indicated in fig . 2 with the small circles labelled pt6 , pt7 and pt8 .
one of these sources ( pt7 ) is highly absorbed and located within @xmath31 from sgr a * , while pt6 has been interpreted as a foreground star , due to its smaller column density .
it is evident that the emission detected with @xmath2 is peaked at the position of pt7 , although we can not exclude that also the other rosat sources ( and ax j1745.62901 , see below ) contribute to some of the detected x rays .
though the limited angular resolution hampers a detailed analysis and introduces unavoidable uncertainties , we proceeded as follows to estimate the spectrum and flux to be ascribed to the gc point source(s ) .
a spectrum was extracted from a circular region with @xmath24 radius centered at the position of the x
ray peak flux ( see fig . 2 ) .
the background was taken from an external annular region ( @xmath32 ) , in order to subtract the contribution from the diffuse emission .
a power law does not give an appropriate fit to the resulting spectrum , which clearly shows evidence for emission lines at 6.7 kev and also at lower energies , indicative of a possible thermal origin of the emission . the sulfur line at @xmath33 kev
is particularly bright .
a power law with a gaussian line at 6.7 kev gives a better result , with a photon index @xmath34 and a column density @xmath35(@xmath36 ) .
when the centroid energy and width of the gaussian are let free in the fit , we obtain the values e=6.7 kev , @xmath37=13 ev and an equivalent width ew@xmath38 kev .
the same results were found fixing the @xmath37 of the iron line at 0 .
no evidence for a 6.4 kev line of fluorescent origin is present , contrary to what is found in the surrounding diffuse emission ( a detailed study of which is deferred to a separate paper ) .
the 210 kev flux corrected for the absorption is @xmath39 ergs s@xmath5 and the luminosity is @xmath40 ergs s@xmath5 at a distance of 8.5 kpc .
we also fitted the data with a thermal plasma model ( mekal in xspec , mewe , gronenschild & van den oord 1985 ) .
it properly accounts for the 6.7 kev iron emission line , but leaves some residuals around 2.4 kev ( see fig . 4 ) , possibly indicating the presence of a multitemperature plasma .
in order to properly account for the emission lines at lower energies , we fitted the spectrum with the sum of two mekal components with different temperatures ( kt@xmath41 kev and kt@xmath42 kev ) , obtaining a higher n@xmath43 . also a single mekal model ( kt@xmath44 kev ) plus a power law with a gaussian added at 6.7 kev fit well the data .
all the spectral results are summarized in table 3 . during observations of the gc region performed with the asca satellite in 1993 and 1994 a possibly new source
was discovered @xmath45 sw of sgra * ( maeda et al .
the spectrum of this source , later named ax j1745.62901 ( maeda et al .
1998 ) , was harder than that of the x ray emission coming from the position of sgr a*. ax j1745.62901 varied by a factor @xmath35 in flux between the two oservations separated by one year , and it showed an x ray burst and periodic ( 8.4 hr ) eclipses ( maeda et al .
these authors proposed that ax j1745.62901 could be the quiescent counterpart of the bright soft x ray transient a1742 - 289 observed in outburst in 1975 ( eyles et al .
1975 , branduardi et al .
however , the absence of eclipses at the 8.4 hr period in the ariel v archival data of the a1742 - 289 outburst ( kennea & skinner 1996 ) , makes this association uncertain .
it is therefore likely that ax j1745.62901 is a different neutron star lmxrb . in order to search for the possible presence of ax j1745.62901 in our observations
, we compared the mecs images accumulated in the soft ( 26 kev ) and hard ( 610 kev ) energy ranges .
this was done independently for the two mecs units ( m2 and m3 ) to obtain the best angular resolution by avoiding possible small misalignement errors .
furthermore , we selected only time intervals in which the @xmath2 attitude was determined with two star trackers ( indeed small deviations @xmath46 may be present when no star tracker is active or when only one is in use ) .
the resulting images for the m3 unit are shown in fig .
5 . in the high energy image , an excess emission in the sw direction
is clearly resolved from the softer ( and partly diffuse ) emission due to the 1e 1742.52859 ( gc ) source(s ) .
the position of this excess is consistent with that of ax j1745.62901 .
it is visible only in the higher energy channels of m3 : in fact the higher background due to misplaced events from the iron calibration sources prevents its detection also in m2 ( chiappetti et al .
1998 ) .
the weakness of the source and the presence of the nearby much brighter x ray emission hampers an accurate determimation of its flux and spectrum , which , in any case , is consistent with the lower value observed with asca in 1993 ( maeda et al .
1996 ) .
we obtained two observations ( obs .
n. 1 and 6 ) pointed on the `` microquasar '' source 1e 1740.72942 ( mirabel et al .
our spectral analysis is based on the longer observation performed in september 1997 . as shown by previous observations with coded mask imaging instruments in the hard x ray band ( skinner et al .
1987 , goldwurm et al .
1994 ) , 1e 1740.72942 is the strongest persistent source at energies greater than @xmath330 kev within a few degrees from the gc .
we therefore analyzed also the pds data , in which the source was clearly detected up to @xmath3150 kev . due to the crowding of this region of the sky
, particular care was devoted to the analysis of this non - imaging detector .
the pds field of view includes the positions of the persistent lmxrb a 1742294 and of the transient ks 1741293 .
as shown by our simultaneous mecs data , the latter source was not active at the time of observation .
a 1742294 lies at @xmath47 from 1e 1740.72942 and it has a similar flux in the 210 kev range .
however , it has a much softer spectrum ( see table 3 ) .
using our best fit thermal bremsstrahlung results and taking into account the reduced off - axis response of the pds collimator , we estimate that a 1742294 contributes less than 10% to the observed counts above 40 kev . a further problem arised by the contamination of one of the two offset collimator positions used in the background determination by the bright x ray source gx 3 + 1 .
we therefore estimated the pds background using only one of the two off - axis collimator positions .
a single power law provides an adequate fit to the lecs and mecs spectrum below 10 kev ( photon index @xmath481.5 , @xmath491.5@xmath1110@xmath50 @xmath51 ) .
we searched in the mecs data for the possible presence of iron lines obtaining a negative result : for lines in the range 6.46.7 kev we can give a 90% upper limit to the equivalent width of @xmath52 ev ( with a line width @xmath37 fixed at 0 ) .
this upper limit increases up to @xmath53 ev for @xmath37 = 500 ev .
these results are similar to the findings of sakano et al .
( 1999 ) who recently reanalyzed all the asca data on 1e 1740.72942 .
however , a single power law can not provide an adequate fit over the broad energy range covered by bepposax .
in fact the inclusion of the pds data clearly shows the presence of a spectral turn over at higher energies ( see fig .
good fits could be obtained with a power law with an exponential cut - off or with comptonization models ( see table 3 ) .
1e 1740.7 - 2942 is of particular interest since , beside being one of the few galactic objects with persistent hard x - ray emission extending above @xmath3100 kev , it has a peculiar radio counterpart with double sided relativistic jets ( mirabel et al . 1992 ) .
the observation of a transient broad spectral feature at @xmath3480 kev ( bouchet et al .
1991 ) , suggested that 1e 1740.7 - 2942 might be related to the 511 kev line observed from the gc region .
on the basis of these properties and of its hard x - ray spectrum similar to that of cyg x-1 , 1e 1740.7 - 2942 is generally considered a black hole candidate .
the broad band spectrum measured with bepposax is consistent with this interpretation .
a 1742294 , although sometimes referred to as a transient source , has always been detected in all the gc observations since 1975 ( see references in van paradijs 1995 ) .
it was observed twice in the outer part of the mecs field of view ( obs . n. 8 and 9 ) . in the 1997 observation
a 1742294 was very close to the region of the field of view used for the on board calibration sources .
so we extracted the counts for the spectral analysis from a small circular region with @xmath24 radius . in both observations
acceptable fits were obtained both with a power law and with a thermal bremsstrahlung ( see table 3 ) . in march 1998
the source was softer and brighter ( of roughly a factor of two ) .
moreover , some variability on a time scale of hours was observed during both observations ( see the light curves in fig .
we detected three type i x ray bursts from a 1742294 : one on september 16 ( 20:04:04 ut ) and two on march 31 , 1998 ( 9:44:58 ut and 15:01:20 ut ) .
the poor statistics hampers a detailed study of the neutron star luminosity , radius and temperature variations during the bursts .
slx 1744299 , first discovered with the spartan 1 experiment ( kawai et al .
1988 ) , was subsequently resolved into two distinct objects separated by @xmath54 with the sl2-xrt coded mask telescope ( skinner et al .
1987 , 1990 ) .
we obtained a single observation pointed on these two sources in september 1997 ( obs .
n.7 ) . for the spectral analysis we used extraction radii of only @xmath24 , and
the same background estimated from a circular corona surrounding both sources .
though the results have to be taken with some caution due to the unavoidable cross - contamination of the two spectra , we can safely conclude that slx 1744300 is fainter and has a slightly softer spectrum than slx 1744299 .
however , contrary to predehl et al .
( 1995 ) , we can not claim a significant difference in their absorption column densities ( see results in table 3 ) . though both slx 1744299 and slx 1744300 have been reported to emit x
ray bursts ( skinner et al .
1987 , 1990 ; sunyaev et al .
1991b ) , none was detected during our observation .
this transient source was discovered with the rossixte satellite in june 1998 ( smith , levine & wood 1998 ) .
its hard x ray spectrum and its association with a variable radio source with evidence for the presence of jets establish xte j1748288 as a possible black hole candidate ( strohmayer & marshall 1998 , hjellming et al .
1998 , fender & stappers 1998 ) .
we detected xte j1748288 in the august 1988 observation ( obs .
its spectrum , extracted from a @xmath24 radius region to avoid the problems due to the poorly known additional absorption from the strongback support of the mecs detector entrance window , could be fit equally well by a hard power law ( photon index @xmath31.5 ) or by a thermal bremsstrahlung with kt@xmath340 kev .
the flux was @xmath310@xmath55 erg @xmath51 s@xmath5 ( 210 kev , corrected for the absorption ) .
the light curve of the xte j1748288 outburst obtained with the rossixte asm is shown in fig .
the date of our observation and the measured flux are indicated by the dashed lines , showing that we obtained a positive detection of xte j1748288 when it was well below the asm sensitivity threshold .
it is interesting to note that the flux measured with @xmath2 is consistent with the extrapolation of the decaying light curve observed with the asm until about mid - july .
this suggests that the nearly exponential decay , with e - folding time @xmath320 days , lasted until the date of our observation ( august 26th ) .
the source sax j1747.02853 , positionally consistent with the transient gx 0.20.2 observed in 1976 ( proctor , skinner & willmore 1978 ) , was discovered with the @xmath2 wide field camera ( wfc ) instrument during an outburst in march 1998 ( int zand et al .
1998b , bazzano et al . 1998 ) . during our observation pointed on 1e 1743.12843 ( obs .
n.10 ) , sax j1747.02853 was detected at an off - axis angle of @xmath56 . during this observation ,
performed about 20 days after the wfc ones , a type - i x ray burst from this source was seen ( sidoli et al .
the burst , starting at 1:40 ut of 1998 april 15 , was also visible at energies above 10 kev with the pds instrument .
the burst light curve in different energy ranges is shown in fig . 9 , where we also present the variations of the spectral fit parameters during the burst ( see sidoli et al . 1998b for details ) .
the spectral softening clearly indicates that the burst is of type i , confirming the lmxrb nature of this source . assuming an eddington luminosity at the burst peak , we obtain a distance of @xmath310 kpc , and a blackbody radius consistent with the expectations for a neutron star . during our observation
pointed on the molecular cloud sgrc performed in march 1998 ( obs .
n.9 ) , we detected a source positionally coincident with the transient burster ks1741 - 293 ( int zand et al .
1998c ) .
since the source was located close to the strongback support of the mecs detector entrance window , we used a small extraction radius ( @xmath24 ) for the spectral analysis .
the background was estimated locally from a source free region of the same observation .
the spectrum is relatively soft and could be described equally well by a power law and a thermal bremsstrahlung ( see table 3 ) , giving in both cases a high column density @xmath492@xmath57 @xmath51 , and a luminosity l@xmath58 erg s@xmath5 ( 210 kev , corrected for absorption ) . slightly lower column densities and luminosity were obtained with a blackbody spectrum ( see table 3 ) .
the same region of sky was observed by @xmath2 in september 1997 ( obs .
n.8 ) and ks 1741293 was not detected , with an upper limit of l@xmath59 @xmath60 10@xmath4 erg s@xmath5 .
this bright and highly absorbed source has been repeatedly observed in several of our pointings .
the best fit results reported in table 3 refer to the observation performed in april 1998 ( obs .
n.11 ) , in which 1e 1743.12843 was on - axis .
this source is probably a persistent lmxrb , though no x ray bursts have been detected so far . for a complete discussion of all the @xmath2 and asca observations of 1e 1743.12843 see cremonesi et al .
( 1999 ) .
the discovery of x ray emission from the center of the radio supernova remnant g0.9 + 0.1 has been reported by mereghetti , sidoli & israel ( 1998 ) .
the location of g0.9 + 0.1 was imaged in april and september 1997 , during observations pointed on the sgr b2 molecular cloud . the angular resolution of the mecs at the off axis location of g0.9 + 0.1 ( @xmath61 ) only allowed to establish that the x ray emission is associated with the plerionic radio core of the supernova remnant and not with the surrounding shell ( @xmath22 diameter , helfand & becker 1987 ) .
however , it was not possible to discriminate between a point source ( i.e. a neutron star ) and a synchrotron nebula of a few arcmin extent .
we have performed a new spectral analysis based on the improved calibration results that are now available .
the derived spectral parameters are within the uncertainties of those previously reported ( mereghetti , sidoli & israel 1998 ) , but they are more precise . the best fit power law spectrum has a photon index @xmath62 , @xmath63 , and flux @xmath64 erg @xmath51 s@xmath5 ( 210 kev , corrected for the absorption ) .
equally good fits were also obtained with blackbody and thermal bremsstrahlung spectra ( see table 3 ) , while a raymond - smith thermal plasma model with abundances fixed at the solar values gave a worse result .
the @xmath2 discovery of x ray emission from the central region of g0.9 + 0.1 confirms its plerionic morphology derived from the radio observations .
the presence of a young , energetic pulsar in g0.9 + 0.1 could also explain part of the high energy gamma - ray emission observed with egret from the region of the galactic canter ( mayer hasselwander et al .
1998 ) .
the radio source g359.230.92 , also known as the mouse , belongs to a small class of radio nebulae characterized by an axially symmetric morphology ( yusef - zadeh & bally 1987 ) .
it is believed that these sources are produced by relativistic particles ejected by a compact object ( either a pulsar or a binary system ) moving with high speed along the axis of symmetry ( shaver et al .
1985 , helfand & becker 1985 ) .
predehl & kulkarni ( 1995 ) using the rosat pspc instrument discovered x ray emission associated with the `` head '' of the mouse , located only @xmath65 northeast of slx 1744299 .
since the mecs data of slx 1744299/300 showed some evidence for an elongation in the direction of g359.230.92 , we performed the same procedure described in section 4.2 for ax j1745.62901 .
the resulting 610 kev image , shown in fig .
10 , confirms the mecs detection of g359.230.92 . to estimate the spectral parameters of g359.230.92 we used a small extraction radius ( @xmath24 ) and measured the background ( which is dominated by the counts of the nearby sources ) from a similar region at the same distance from slx 1744299 . a power law fit gave @xmath48 1.92.3 , a column density of @xmath6610@xmath67 @xmath51 and an unabsorbed flux in the 210 kev band of the order of @xmath6810@xmath69 erg @xmath51 s@xmath5 .
this flux is higher than the value expected based on the rosat results at energies below 2.4 kev ( predehl & kulkarni 1995 ) , however this discrepancy might be due to different reasons other than source variability .
first , the flux reported by these authors was based on the assumption of a value of only @xmath70 @xmath51 for the column density : we found that for our n@xmath43 value the mecs and rosat pspc count rate are consistent with the same flux level .
another possibility is that we underestimated the contamination from slx 1744299/300 . in any case , the detection at energies greater than 6 kev is statistically significant and supports a non - thermal origin for the x ray emission from the head of the `` mouse '' .
our @xmath2 survey of the inner 4 degrees of the galactic plane shows that a few persistently bright sources characterize the constellation of x ray sources near the gc .
their properties indicate that they consist of compact accreting objects similar to the classical accretion powered sources found elsewhere in the galaxy .
both systems containing neutron stars ( e.g. slx 1744299/300 , a 1742294 ) and likely black holes ( 1e 1740.72942 ) are present .
however , these persistently bright objects represent only a small fraction of the x ray sources in this region .
three transient sources ( sax j1747.02853 , ks 1741293 and xte j1748 - 288 ) were detected in our survey , but it is known from previous observations that many more are present .
we therefore confirm the clustering of x ray sources towards the gc already noted by several authors .
it is likely that on the large scale ( @xmath3 degrees ) this simply reflect the general mass distribution enhancement , as found by grebenev , pavlinsky & sunyaev ( 1996 ) who considered only the hard x ray sources detected with the art - p coded mask telescope . on the other hand ,
considering the distribution of all the sources reported within several arcminutes from sgra * there is marginal evidence of an additional source population ( skinner 1993 ) .
however , this conclusion might be biased by the different sensitivity and angular resolution of the many observations of this field . though it is obvious that the regions closer to the center have been observed more deeply , the effect of this bias is difficult to quantify .
we finally note that the case of the three sources in the direction of slx j1744299/300 ( at b = 1.5@xmath19 ) is maybe more surprising that the presence of 4 - 5 sources close to sgra*. while g359.230.92 is very likely a foreground object , the two x ray bursters slx j1744299 and slx j1744300 could be at the same distance and therefore somehow related ( e.g. members of a star cluster ) .
observations in x - rays with better angular resolution and deep searches in the infrared are needed to study these objects . in fig .
11 we have plotted the column density and power law photon index ( in the 210 kev range ) for all the sources that we have observed ( to compare the spectra of the different sources , we have used a power law also in a few cases in which other models provided better fits ) .
it is clear that sources containing neutron stars , as indicated by the presence of type i bursts , have softer spectra than the black hole candidates .
based on this spectral dichotomy , it is likely that also 1e 1743.1 - 2843 contain a neutron star .
11 also shows that the column densities of these sources vary over a factor @xmath34 range .
this is probably related to the clumpiness of the molecular clouds in this region rather than to a corresponding spread in the source distances .
at least four point sources ( predehl & trumper 1994 , maeda et al . 1996 ) , maybe five ( kennea & skinner 1996 ) and possibly more could contribute to the x ray flux detected with @xmath2 when pointing at the gc , in addition to the diffuse emission that permeates the region ( maeda & koyama 1996 ) .
however , as shown in fig 2 , the mecs source position is closer to the rosat counterpart of sgra * than to the other two rosat sources ( pt6 and pt8 ) .
the emission lines detected in the mecs spectrum within @xmath24 from the gc clearly indicate a substantial contribution from the thermal diffuse emission associated to the sgra west and sgra east regions ( yusef - zadeh et al .
unfortunately , this contribution is difficult to quantify , and the limited angular resolution does not allow to measure the flux to be ascribed to sgra * itself .
we can place a very conservative upper limit to the luminosity of sgra * by assuming that it is the only point source and that the diffuse emission at its position has the same surface brightness of the annular region ( @xmath32 ) described above . in this case , using the flux reported in table 3 , we obtain a luminosity of @xmath71 ergs s@xmath5 ( 210 kev , corrected for the absorption ) . however , this is probably an overestimate of the true luminosity of sgra * for the fact that the diffuse emission has a surface brightness distribution that increases toward the gc and because there might be a contribution from the other point sources .
a more realistic upper limit to the x ray emission from sgra * can be placed if we consider an x ray background rising towards the gc : we measured the surface brightness from three concentric annular regions ( @xmath72 , @xmath73 and @xmath32 ) and extrapolated it at our source position . in this case we find a new background value , four times higher than the previous one and the new flux for the point source is @xmath74 ergs s@xmath5 , that translates into a luminosity at 8.5 kpc of @xmath75 ergs s@xmath5 .
this upper limit is similar to the value reported by the asca team ( maeda et al . , 1998 ) and is well below the eddington luminosity for a super massive black hole .
several authors tried to explain this high energy underluminosity of sgra * with different models for the accretion , from a spherical accretion ( melia 1992 ) to the advection dominated accretion flow ( narayan , yi & mahadevan 1995 ; narayan et al .
these models have still some problems in the radio and in the @xmath76ray bands ( egret source 2eg 17462852 ) , where substantial excesses are present .
several solution has been proposed : the inclusion of an emission process associated with protons ( mahadevan et al .
1998 ) is able to solve the problem of the radio excess , while the egret source is considered as a gamma
ray upper limit due to the poor spatial resolution .
another possible explanation is that the egret source is not related with sgra * , but with the sgra east supernova remnant shell ( melia et al .
1998 ) .
the xmm and axaf observatories with their superior imaging capabilities will substantially contribute to solve the problem of the high energy emission from sgra * and put more stringent contraints to the accretion models . 1 & 17 43 54.8 29 44 42 & 1996 sep 03 10:56 - 03 16:30 & 7 & 1e 1740.72942 2 & 17 47 20.0 28 24 02 & 1997 apr 05 08:57 - 06 09:33 & 49 & sgr b2 3 & 17 45 40.3 29 00 23 & 1997 aug 24 06:01 - 26 08:41 & 100 & sgr a * 4 & 17 47 20.0 28 24 02 & 1997 sep 03 05:18 - 04 09:36 & 51 & sgr b2 5 & 17 48 39.2 28 05 56 & 1997 sep 04 22:45 - 06 03:22 & 49 & sgr d 6 & 17 43 54.8 29 44 42 & 1997 sep 07 05:45 - 08 06:57 & 43 & 1e 1740.72942 7 & 17 47 25.8 30 01 01 & 1997 sep 10 14:07 - 11 18:36 & 21 & slx 1744299 8 & 17 44 41.3 29 28 14 & 1997 sep 16 11:30 - 16 20:47 & 15 &
sgr c 9 & 17 44 41.3 29 28 14 & 1998 mar 31 06:47 - 31 20:54 & 25 & sgr c 10 & 17 46 19.9 28 44 00 & 1998 apr 13 13:36 - 15 05:51 & 71 & 1e 1743.12843 11 & 17 46 50.2 28
33 04 & 1998 aug 26 16:34 - 28 16:30 & 79 & sgr b1 xte j1748288 & 17 48 08 & -28 28 48 & @xmath77 & transient & 1 slx 1744299 & 17 47 27 & -29 59 55 & @xmath77 & burster & 2 slx 1744300 & 17 47 27 & -30 02 15 & @xmath77 & burster & 2 g0.9 + 0.1 & 17 47 21 & -28 09 22 & & snr & 3 g359.20.8 ( the mouse ) & 17 47 15.2 & -29 58 00 & @xmath78 & & 4 sax j1747.02853 & 17 47 0 & -28 52 12 & @xmath77 & transient / burster & 5 1e 1743.12843 & 17 46 19.5 & -28 53 43 & @xmath77 & & 6 a 1742294 & 17 46 05 & -29 30 54 & @xmath77 & burster & 7 rxj 1745.72858 & 17 45 45.5 & -28 58 17 & @xmath79 & star ?
( pt6 ) & 8 1e
1742.52859 & 17 45 40.7 & -29 00 10 & @xmath77 & sgra * ? & 6 rxj 1745.62900 & 17 45 40.4 & -29 00 23 & @xmath79 & sgra * ? ( pt7 ) & 8 ax j1745.62901 & 17 45 36 & -29 01 34 & @xmath80 & transient / burster ?
& 9 rxj 1745.52859 & 17 45 32 & -28 59 29 & @xmath79 & pt8 & 8 a 1742289 & 17 45 37.3 & -29 01 04.8 & @xmath81 & transient & 10 ks 1741293 & 17 44 49.2 & -29 21 06.3 & @xmath77 & transient / burster & 11 1e 1740.7 - 2942 & 17 43 54.8 & -29 44 42.8 & @xmath82 & black hole candidate & 12 galactic center & pl+line(6.7 ) & @xmath83 & @xmath84 & @xmath85 ( 284 ) & @xmath86 ( mecs ) & & & @xmath87 & & & brems+line(6.7 ) & @xmath88 & @xmath89 & @xmath90 ( 284 ) & @xmath91 & & & @xmath87 & & & m & @xmath92 & @xmath93 & @xmath94 ( 288 ) & @xmath95 & m+m & @xmath96 & @xmath97 & @xmath98 ( 285 ) & @xmath99 & & & @xmath100 & & & m+pl+line(6.7 ) & @xmath101 & @xmath102 & @xmath103 ( 281 ) & @xmath104 & & & @xmath105 & & & & & @xmath106 & & 1e 1740.72942 & pl & @xmath107 & @xmath108 & @xmath109 ( 180 ) & @xmath110 ( mecs ) & brems & @xmath111 & @xmath112 & @xmath85 ( 180 ) & @xmath113 ( lecs+mecs & pl & @xmath114 & @xmath115 & @xmath116 ( 283 ) & @xmath117 + pds ) & brems & @xmath118 & @xmath119 & @xmath120 ( 283 ) & @xmath121 & compst & @xmath122 & @xmath123 & @xmath124 ( 282 ) & @xmath125 & & & @xmath126 & & & pl+e@xmath127 & @xmath128 & @xmath129 & @xmath130 ( 281 ) & @xmath131 & & & e@xmath132 & & & & & e@xmath133 & & a 1742294 & pl & @xmath134 & @xmath135 & @xmath136 ( 238 ) & @xmath137 ( mecs - obs.9 ) & brems & @xmath138 & @xmath139 & @xmath140 ( 238 ) & @xmath141 & pl & @xmath142 & @xmath143 & @xmath144 ( 323 ) & @xmath145 ( mecs - obs.10 ) & brems & @xmath146 & @xmath147 & @xmath148 ( 323 ) & @xmath149 slx 1744299 & pl & @xmath150 & @xmath151 & @xmath152 ( 364 ) & @xmath153 ( mecs ) & brems & @xmath154 & @xmath155 & @xmath144 ( 364 ) & @xmath156 slx 1744300 & pl & @xmath157 & @xmath158 & @xmath159 ( 342 ) & @xmath160 ( mecs ) & brems & @xmath154 & @xmath161 & @xmath162 ( 342 ) & @xmath163 xte j1748288 & pl & @xmath164 & @xmath165 & @xmath166 ( 163 ) & @xmath167 ( mecs ) & brems & @xmath168 & @xmath169 & @xmath170 ( 163 ) & @xmath171 sax j1747.02853&brems & @xmath172 & @xmath173 & @xmath148 ( 138 ) & @xmath174 ( mecs ) & & & & & ks 1741293 & pl & @xmath175 & @xmath176 & @xmath177 ( 213 ) & @xmath178 ( mecs ) & brems & @xmath179 & @xmath180 & @xmath181 ( 213 ) & @xmath182 & bb & @xmath183 & @xmath184 & @xmath185 ( 213 ) & @xmath186 1e 1743.12843 & bb & @xmath187 & @xmath188 & @xmath148 ( 178 ) & @xmath189 ( mecs ) & & & & & g0.9 + 0.1 & pl & @xmath190 & @xmath191 & @xmath192 ( 28 ) & @xmath193 ( mecs ) & brems & @xmath194 & @xmath195 & @xmath192 ( 28 ) & @xmath196 & bb & @xmath197 & @xmath198 & @xmath192 ( 28 ) & @xmath199 g359.23 - 0.92 & pl & @xmath200 & @xmath201 & & @xmath202 ( mecs ) & & & & & goldwurm a. , cordier b. , paul j. , ballet j. , bouchet l. , roques j.p . , vedrenne g. , mandrou , p. , sunyaev , r. , churazov e. , gilfanov m. , finogenov a. , vikhlinin a. , dyachkov a. , khavenson n. & kovtunenko v. 1994 , nature 371 , 589 sidoli l. , israel g.l . , chiappetti l. , treves a. , orlandini m. , kuulkers e. , predehl p. , heise j. & mereghetti s. , 1998a , in the active x ray sky - results from @xmath2 and rxte " , nuclear physics b ( proc.suppl . ) 69/1 - 3 , 88 .
sidoli l. , mereghetti s. , chiappetti l. , heise j. , israel g.l . , kuulkers e. , orlandini m. , predehl p. , tiengo a. & treves a. 1999 , proc .
third integral workshop , the extreme universe " , taormina 1998 , in press sunyaev r.a .
, churazov e. , gilfanov m. , pavlinsky m. , grebenev s. , babalyan g. , dekhanov i. , yamburenko n. , bouchet l. , niel m. , roques j.p . ,
mandrou p. , goldwurm a. , cordier b. , laurent ph .
, paul j. 1991a , a&a 247 , l29 | we report the results of a survey of the galactic center region ( @xmath0 , @xmath1 ) performed with the @xmath2 satellite .
the flux from the center of our galaxy corresponds to a luminosity of @xmath33 10@xmath4 erg s@xmath5 in the 210 kev range .
due to the limited angular resolution ( @xmath6 ) only part of it is supposed to come from sagittarius a * , the non thermal radio source which is believed to mark the dynamical center of the galaxy .
in addition to the diffuse emission , several bright ( l@xmath7 ergs s@xmath5 ) point sources have been observed , both persistent ( a 1742294 , slx 1744299 , slx 1744300 , 1e 1743.12843 , 1e 1740.72942 ) and transient ( xte j1748288 , sax j1747.02853 and ks 1741293 ) . the low mass x
ray binary ax j1745.62901 , discovered with asca at only @xmath8 from sgra * was detected in a low luminosity state in august 1997 .
the 1150 kev spectrum of the hard x ray source 1e 1740.72942 is well described by a comptonization model , typical of black hole candidates in their low / hard state , with no evidence for strong fe lines .
the detection of a type i burst shows that the transient source sax j1747.02853 ( probably the same as the 1976 transient gx 0.20.2 ) is a lmxrb containing a neutron star . the transient black hole candidate xte j1748288 was detected at a luminosity ( @xmath9 ergs s@xmath5 ) consistent with the extrapolation of the exponential decay of the outburst observed with the xte all sky monitor .
two fainter sources are very likely associated with young neutron stars : the ( possibly diffuse ) x ray source at the center of the composite supernova remnant g0.9 + 0.1 , and the head " of the axially symmetric radio source g359.230.92 .
the latter has been detected above @xmath3 6 kev , supporting a non thermal emission mechanism . #
1to 0pt#1 # 1#2d # 1 d # 2 #1_#1 #1 # 2 a&a , # 1 , # 2 # 1 # 2 mnras , # 1 , # 2 # 1 # 2 apj , # 1 , # 2 # 1 # 2 nature , # 1 , # 2 # 1 # 2 pasj , # 1 , # 2 |
heavy fermion materials involve a lattice of localized magnetic moments derived from f - electrons , embedded in a conduction sea formed principally from delocalized d - electrons @xcite .
the physics of these materials can be understood as a consequence of the interplay between two competing physical processes : * the kondo effect , which tends to screen the local moments to produce a band of heavy electrons , and * antiferromagnetism , which locks the local moments together via the rkky interaction , into a state with long range magnetic order .
the characteristic scales for these two processes are @xmath2 where @xmath3 is the strength of the onsite kondo interaction between the localized f - electrons and conduction electrons , @xmath4 is the density of states of the conduction electrons at the fermi energy and @xmath5 the bandwidth .
the kondo temperature , @xmath6 , sets the energy scale for the onset of the kondo effect and consequently the formation of a coherent heavy fermi liquid ( hfl ) while @xmath7 defines the energy scale for the onset of magnetic order .
the family of heavy fermion materials provides an important setting for the study of quantum criticality @xcite which develops when a continuous second - order phase transition is suppressed to absolute zero temperature .
the small characteristic energy scales of these compounds makes them highly tunable , allowing the ready exploration of the phase diagram as a function of pressure , magnetic field or doping .
superconductivity is often found in the vicinity of magnetic quantum critical points ( qcp ) . at temperatures above the quantum critical point non fermi liquid ( nfl ) behavior
is observed , generally characterized by sub - quadratic temperature dependence of the resistivity @xmath8 and a logarithmic temperature dependence of the specific heat coefficient @xmath9 , @xmath10 where @xmath11 is the characteristic scale of the spin fluctuations . for a review of experimental properties of these materials
see stewart@xcite .
one of the central challenges of heavy fermion materials is to understand the mechanism by which magnetism develops within the heavy electron fluid .
traditionally , magnetism and heavy fermion behavior have been regarded as two mutually exclusive states , separated by a single quantum critical point . however , a variety of recent experiments suggest a richer state of affairs , in particular : * ybrh@xmath12si@xmath12 can be driven to a quantum critical point by the application of magnetic field , where both the nel temperature and the kondo energy scale appear to simultaneously vanish . however
, when doped , these two energy scales appear to separate from one - another , indicating that the break - down of fermi liquid behavior and the development of magnetism are not rigidly pinned together @xcite ; * in the 115 superconductor cerhin@xmath13 there is evidence for a microscopic and homogeneous coexistence of local moment magnetism and superconductivity under pressure @xcite ; * neutron scattering experiments observe a partially ordered state in the geometrically frustrated cepdal , in which one third of the ce moments do not participate in the long - range order , suggesting the development of inhomogeneous kondo states @xcite .
various phenomenological frameworks have been proposed for the understanding of heavy fermion systems .
the classical framework proposed in the 70 s by doniach @xcite , involves a competition between @xmath6 and @xmath7 determining the ground state to be a heavy fermi liquid or magnetically ordered .
more recently a new axis was added to this picture , by the inclusion of geometric frustration or reduction of dimensionality @xcite .
these two factors contribute towards the suppression of magnetism in a different way , if compared to the competition with the kondo effect .
also , based on experiments in several families of heavy fermions , a phenomenological two - fluid picture was proposed by nakatsuji and pines , with predictive power on the ground state @xcite .
unfortunately these proposals do not give us information about the character of the transition between the hfl and magnetic phases , and its theoretical description has remained an unsolved challenge for several decades .
theoretical proposals based on a spin density wave description of the qcp @xcite , kondo breakdown @xcite , deconfined quantum criticality @xcite and local quantum criticality @xcite have been suggested , but no one picture is yet able to fully account for experimental observations .
the kondo lattice hamiltonian @xmath14 provides a minimal model for heavy fermion systems .
the first term in @xmath15 describes a band of conduction electrons with dispersion @xmath16 , @xmath17 is the antiferromagnetic kondo coupling between the local moment @xmath18 and the spin of the conduction electron @xmath19 at site @xmath20 .
the local moments are neutral entities uniquely characterized by their spin quantum numbers .
the removal of the charge degrees of freedom from the hilbert space of the localized f - electrons means that spin operators do not follow canonical commutation relations ; consequently , their treatment within a path integral or diagrammatic approach is complicated by the absence of a wick s theorem . to circumvent this difficulty , the spin operator is traditionally factorized in terms of creation and annihilation operators : @xmath21 where @xmath22 , @xmath23 are bosonic or fermionic creation and annihilation operators , respectively , and the indexes @xmath24 for an @xmath25 spin .
there are actually several such spin representations : the holstein - primakoff @xcite , schwinger boson @xcite , abrikosov pseudo - fermion @xcite and the drone or majorana fermion @xcite representations , among others .
the physics that each of these representations describes is profoundly different .
for example , the antiferromagnetic ( afm ) phase at small @xmath17 is very effectively described by a schwinger boson representation of the local moments , with the condensation of the bosons corresponding to the onset of magnetic order @xcite . by contrast , the heavy fermi liquid phase at large @xmath3 is successfully captured by a fermionic representation of the spins @xcite .
we take the view that the success of these two representations in the different limits is not simply one of mathematical convenience ; rather , it reflects the physical transformation of both the spin correlations and the excitations of the local moments : these evolve from collective spin waves to charged heavy fermions .
remarkably , experiment indicates that these two phases connect together _ continuously _ via a quantum critical point , suggesting that at quantum criticality the two representations merge .
in this paper we argue that a full description of heavy fermion materials requires a methodology that can capture the transformation in the character of the ground - state and its spin excitations .
this , in turn , leads us to adopt a _
representation of the spin@xcite @xmath26 here , @xmath27 , @xmath28 and @xmath29 , @xmath30 are respectively , fermionic and bosonic creation and annihilation operators .
the spin is _ supersymmetric _ because it is invariant under transformations that take bosons into fermions and vice versa ; these are generated by fermionic operators which will be introduced in the next section .
one of the challenges of such a factorization , is that it requires a constraint which guarantees that the physics lies within the physical hilbert space @xcite .
for example , an elementary spin @xmath31 kramers doublet requires the constraint @xmath32 . within this
constrained hilbert space , the most general wavefunction is an _ entangled product _ @xmath33 and @xmath34 are the bosonic and fermionic components of the wavefunction , respectively , while @xmath35 is a gutzwiller projection operator .
this operator can be written as : @xmath36 which imposes the constraint @xmath37 at each site @xmath20 .
the unprojected wavefunction @xmath38 describes the formation of long - range magnetic correlations in the form of a bosonic rvb wavefunction , while @xmath39 captures the development of kondo of singlets and the development of a large fermi surface of heavy electrons .
the gutzwiller projection entangles the two components of the wavefunction into a single entity as illustrated in fig .
[ fig : susypsi ] .
this merged wavefunction has , in principle , the potential to capture the two - fluid aspect of the heavy fermion ground - state . and the height @xmath40 of the tableau . ]
the absence of a small parameter in the kondo model effectively rules out the use of conventional perturbation theory .
the alternative approach , followed here , is the use of a large-@xmath41 expansion in which the fundamental representations contain @xmath41 , rather than @xmath42 components . in this approach @xmath43
plays the role of synthetic planck s constant leading to a controlled mean - field ( `` classical '' ) theory in the large-@xmath41 limit , with the possibility of expanding the fluctuations and the constraint condition as a power - series in @xmath44 about the large-@xmath41 limit .
the simplest generalization takes @xmath25 to @xmath45@xcite . written in traceless form
the @xmath46 spin is then @xmath47 where @xmath48 .
however , in this paper we seek to extend the supersymmetric description of spins to the symplectic subgroup @xmath1 of @xmath46@xcite : @xmath49 here @xmath41 must be even , while the range of the elementary spin quantum numbers is @xmath50 . the tilde notation , employed extensively in this article , denotes the sign of the index @xmath51 with the analogous definition for other indexes .
this new spin operator has the symplectic property @xmath52 ( and is thus also traceless ) .
the symplectic group @xmath53 offers many advantages for condensed matter physics , allowing for a consistent extension of the notion of time reversal symmetry to the large-@xmath41 limit , which permits one to form singlet pairs of particles that are absent in the @xmath46 generalization @xcite .
this capability is vital to describe antiferromagnetism and superconductivity .
the concept of a supersymmetric spin was introduced in previous studies of impurity kondo models @xcite . in the work presented here , we follow the lines of coleman et al.@xcite , with the additional generalization to @xmath1 , and discuss the spin representations in terms of young tableaux .
young tableaux provide a precise pictorial rendition of irreducible spin representations : horizontal young tableaux label completely symmetric representations , which are naturally described by bosons , while vertical young tableaux label completely antisymmetric representations , usually described by fermions .
the use of supersymmetric representations lead us to consider the set of representations characterized by l - shaped young tableaux ( fig .
[ lshaped ] ) .
these representations are characterized by two constants : 1 . the total number of elementary spins ( or boxes ) in the representation @xmath54 , where @xmath40 and @xmath55 are height , and width of the young tableau , respectively , and 2 . the asymmetry @xmath56 of the l - shaped young tableau , as discussed in coleman et al.@xcite .
the asymmetry of the representation is absent in a physical @xmath25 spin-1/2 , in which case the young tableau is depicted by a single box , but once we enlarge the symmetry group of the spin in order to develop a large-@xmath41 theory , we find a family of representations that range from a completely symmetric representation , fully described by bosons , to a completely antisymmetric representation , described only by fermions , including a whole plethora of intermediate representations that we refer to as _ mixed representations _ , depicted by l - shaped young tableaux ( see fig .
[ rep ] ) .
the possibility of mean - field solutions described by mixed representations is interesting as it may permit the description of new states of matter , including coexistence of magnetism with superconductivity or with heavy fermi liquid phases .
limit ranging from a fully symmetric ( top ) towards a fully antisymmetric ( bottom ) , passing through a series of l - shape representations . ] for a given value of @xmath41 , one needs to decide which representation to choose in order to proceed with the calculations .
traditionally a purely bosonic representation or a purely fermionic representation is chosen , but the supersymmetric approach provides the possibility of considering an l - shaped representation . to constrain the problem to such a representation one must fix the values of @xmath57 and @xmath58 through the introduction of projection operators into the partition function : @xmath59.\end{aligned}\ ] ] typically , the more negative @xmath60 , the more symmetric the spin representation and the more magnetic the resulting ground - state whereas the more positive @xmath60 , the more antisymmetric the spin representation and the more fermi - liquid like the ground - state . to avoid biasing the physics , we consider a grand - canonical ensemble of representations defined by the partition function with indefinite asymmetry @xmath60 , @xmath61=\sum_{y_0 } tr[p_{q_0 , y_0 } e^{-\beta h}],\end{aligned}\ ] ] where now we can identify @xmath62 with the large-@xmath41 generalization of the gutzwiller projection operator introduced in eq .
[ gutz ] : @xmath63 this procedure will enable the ensemble to explore the lowest energy configurations . another motivation to work with the constraint that fixes only the total number of boxes of the representation is the fact that the asymmetry of the representation appears only in the large-@xmath41 limit , so by letting @xmath60 run free provides an unbiased way to take the limit @xmath64 , as schematically shown in fig .
[ rep ] .
the large-@xmath41 limit we now develop places all l - shaped representations with a given number of boxes on the same footing , and the asymmetry of the representation can be thought of as a variational parameter .
the character of the representation ( bosonic , fermionic or mixed ) will now be decided by the energetics of the problem .
this will permit us to explore the phase diagram of systems as heavy fermions , in which the character of the spin changes from fermionic in the hfl phase towards bosonic in the afm region .
[ spoint ] illustrates schematically the evolution of the energy landscape ( energy as a function of @xmath65 , the number of fermions in the representation ) , for three different values of @xmath66 . for small @xmath66
the energy landscape has a minima for @xmath67 , which means that the system prefers to have a bosonic spin representation and possibly develops magnetic order .
analogously , for large @xmath66 , the energy landscape has a minimum for the maximum value of @xmath68 , indicating a purely fermionic representation , which would possibly lead to the development of a heavy fermi liquid . for intermediate values of @xmath66
we find that the representation is _ mixed _ , with an energy minima developing at an intermediate value of @xmath68 so that the minima is a saddle point as a function of @xmath65 .
we are thus able to identify two classes of solution : 1 . ,
in which the free energy is minimized by a purely bosonic or a purely fermionic representation , indicating that the original supersymmetry of the spin is severely broken . in this case
the results of a purely bosonic or fermionic representations are recovered ; 2 .
, in which mixed representations are energetically favorable .
these kinds of minima are candidate representations for a two - fluid picture of heavy fermions .
since the fermionic and bosonic components of the spin fluid acquire the same chemical potential , this opens up the possibility of a new kind of zero mode : a _ goldstino _ , arising from the zero energy cost of rotations between the fermionic and bosonic spin fluid . for different values of the ratio @xmath66 .
] other new features of this work are : * _ symplectic spins _ : we generalize the rotation group of the spin from @xmath25 to @xmath1 for a large-@xmath41 treatment @xcite .
this generalization guarantees that all the components of the generalized spin properly invert under time reversal which confers various advantages .
in particular , it allows the description of geometrically frustrated magnetism@xcite and it permits the exploration of singlet superconductivity within the large-@xmath41 framework . * _ spatially inhomogeneous representations _ : we explore the possibility of solutions which _ spontaneously _ develop kondo , or magnetic character at different sites . here
we are motivated by the partially ordered phase verified experimentally in cepdal .
[ inhom ] represents this kind of solution schematically in a frustrated triangular geometry : one of the local moments in the triangle develops fermionic character , forming a singlet with electrons in the conduction sea , while the other two local moments have a bosonic representation , forming an antiferromagnetic bond .
this paper is organized as follows : in section [ secss ] we define the supersymmetric - symplectic spin , identifying the gauge group under which it is invariant and discuss the casimir for a given irreducible representation .
details on the derivations are given in appendix [ appss ] . in section [ secfor ]
we introduce the general path integral formalism which leads to a mean field free energy in the large-@xmath41 limit .
we apply this formalism to a two - impurity model in section [ sec2imp ] , and details on the calculations are given in appendices [ appfe ] , [ appbe ] and [ appflu ] . in section [ sec3imp ]
we apply the same formalism to a three - impurity model .
we conclude and discuss the open questions in section [ seccon ] .
we start defining the supersymmetric - symplectic spin : @xmath69 where @xmath27 , @xmath28 and @xmath29 , @xmath30 are respectively , fermionic and bosonic creation and annihilation operators , with indexes @xmath70 , and @xmath71 , @xmath72 . in this form
the invertion of the spin under time reversal is made explicit .
we can write the spin operator more concisely as : @xmath73 by introducing the four component spinor @xmath74 , @xmath75 which carries the explicit spin index @xmath76 , and has an implicit super - index which runs from 1 to 4 related to the supersymmetric and particle - hole character of its entries ; and the matrix @xmath77 we shall follow the convention that super - indices are suppressed and fully contracted with one - another in our formulae , unless otherwise stated . the supersymmetric - symplectic spin defined in eq .
[ ssusy ] commutes with the following operator bilinears and the respective conjugates : @xmath78 these are therefore generators of the symmetry group of the supersymmetric - symplectic spin .
note that @xmath79 and @xmath80 are present in the @xmath1 but not in the @xmath45 generalization of the supersymmetric spin @xcite .
we can rewrite these operators in the form of hubbard operators @xcite as follows : @xmath81 where @xmath82 , @xmath83 and @xmath84 are bosonic hubbard operators , while @xmath85 and @xmath86 are fermionic hubbard operators . in this form
the algebra that these operators follow can be concisely written as : @xmath87_{\pm}=x_{ad}\delta_{bc}\pm x_{cb}\delta_{ad},\end{aligned}\ ] ] where the anticommutator , @xmath88 , is used when both hubbard operators are fermionic , while the commutator , @xmath89 , otherwise .
the hubbard algebra@xcite above defines the @xmath90 supergroup@xcite .
one can explicitly see the @xmath0 subgroup generated by the isospin operators @xmath91 which follow the commutation relation : @xmath92=i\epsilon_{ijk}\psi_k,\end{aligned}\ ] ] and are related to the rotation of the fermionic components of @xmath93 ; while @xmath82 defines the generator for the @xmath94 subgroup associated with the bosons .
note that one can perform a super - rotation @xmath95 taking @xmath96 which leaves the spin invariant : @xmath97 if @xmath95 satisfies @xmath98 the most general transformation @xmath95 can be constructed by exponentiation of the generators of the group listed above . an explicit form for g
is given in appendix [ appa2 ] .
to uniquely characterize an irreducible representation of @xmath1 , in principle one needs to define @xmath99 casimirs , where @xmath100 is the rank of the group ( the dimension of the cartan sub - algebra ) . in terms of young tableaux
, one can understand these parameters as the number of boxes in each row of the tableau ( the maximum number of rows in the young tableau for @xmath1 is @xmath101 ) .
in this work we restrict our attention to l - shaped tableaux , so we have the extra information that all the rows below the first one have no more than a single box .
this reduces the number of parameters required to define the representation to two : @xmath102 , the total number of boxes , and @xmath60 , the asymmetry of the representation , as discussed in coleman et al.@xcite . from the second casimir
we can identify the quantities @xmath102 and @xmath60 . from nwachuku@xcite
, we can deduce that for an l - shaped representation in @xmath1 the second casimir can be written in terms of the width @xmath55 of the first row and the height @xmath40 of the column in the tableau as : @xmath103 and identifying @xmath104 and @xmath56 , we have @xmath105 the details of this derivation are shown in appendix [ appssnwa ] . in terms of operators , the second casimir can be written as the magnitude of the spin : @xmath106 where : @xmath107 this form is similar to that found for the @xmath45 case ( see appendix [ appa4 ] for details of the derivation ) . note that in the lowest weight state @xmath108 , where @xmath109 , ( corresponding to no pairs and to minimizing the number of fermions ) @xmath110 .
we can also write @xmath111 in terms of the @xmath90 casimir of the hubbard operators via the identity : @xmath112- ( x_{00})^{2},\end{aligned}\ ] ] with an implied summation over the repeated indices @xmath113 ( see appendix [ appa5 ] ) .
the main topic of interest in our work is the class of kondo - heisenberg models involving @xmath53 spins , interacting via an additional nearest neighbor antiferromagnetic heisenberg exchange .
these models are written as @xmath114 where @xmath115 describes a conduction band of electrons of dispersion @xmath116 , where @xmath117 creates a conduction electron of momentum @xmath118 , spin index @xmath119 .
the term @xmath120 describes the kondo interaction at site @xmath121 , where @xmath122 defines the local moment operator as in eq .
[ ssusy ] and the conduction electron spin operators are also written in symplectic form : @xmath123 the final term describes the heisenberg interaction between spins at sites @xmath20 and @xmath121 , given by @xmath124 to display the supersymmetric gauge character of the interactions we now rearrange the order of the operators .
first , using the properties of the symplectic spin , @xmath125 , the two parts of the electron spin operator in the kondo interaction are folded into one as follows : @xmath126 next , we super - commute the @xmath127 field to the right - hand side of the interaction , rewriting the interaction in terms of a supertrace , @xmath128\ ] ] where we define the supertrace as @xmath129=a_{11}+a_{22}-a_{33}-a_{44}$ ] and we have used the property that the dot product of two super - spinors @xmath130 and @xmath131 can be rewritten as a supertrace of their outer - product @xmath132 = - { \rm str}[\chi \bar \phi ] $ ] . notice that @xmath133 and @xmath134 are four component column and row spinors that respectively transform like @xmath135 and @xmath136 under super - rotations . in a similar fashion
, we rewrite the heisenberg interaction as @xmath137.\end{aligned}\ ] ] notice that the object @xmath138 is an outer - product of two super - spinors , forming a four - by - four tensor in superspace that transforms as @xmath139 under super - rotations . with these manipulations ,
the kondo - heisenberg model can be written as : @xmath140 -\frac{j_{h}}{n}\sum_{(i , j)\alpha , \beta } { \rm str}\left [ \left(\psi_{i\beta } \bar \psi_{j\beta } \right ) \left(\psi_{j\alpha } \bar \psi_{i\alpha } \right ) \right].\end{aligned}\ ] ] notice that the invariance property of the supertrace @xmath141= { \rm str}[t]$ ] guarantees that these interactions are gauge - invariant under the local super - rotations @xmath142 .
the factorized forms of the interactions are convenient for hubbard - stratonovich transformations .
the constraint fixing the total number of bosons plus fermions at each site @xmath143 can also be written in terms of the spinors @xmath144 : @xmath145 where @xmath146
we can now write the partition function as a functional integral over the constraint field @xmath147 and the spin carrying boson and fermion fields @xmath148 where @xmath149 while @xmath150 , \end{aligned}\ ] ] is the measure of integration over the canonical @xmath151 , @xmath152 and @xmath153 fields and @xmath154 is the measure of integration over the constraint . the components of the action are : @xmath155 the conduction electron part of the action ; @xmath156\end{aligned}\ ] ] describes the berry phase , with the constraint @xmath157 imposed via the introduction of the lagrange multiplier @xmath158 at each site , while @xmath159 are the kondo and heisenberg parts of the action . inside the path integral , we can now carry out a hubbard - stratonovich transformation of the interactions .
the kondo part of the interaction is factorized as follows @xmath160 \rightarrow \cr h_{k } ' ( j)&= & \sum_\alpha { \rm str}\left [ \left(\psi_{j\alpha } c{^{\dagger}}_{j\alpha } \right)\bar v_{j}+ v_{j}\left(c_{j\alpha}\bar \psi_{j\alpha } \right ) + \frac{n}{2j_{k } } \bar { v}_{j}v_{j } \right]\cr & = & \sum_\alpha \left [ \left(\bar v_{j}\psi_{j\alpha } \right ) c{^{\dagger}}_{j\alpha } + c_{j\alpha } \left ( \bar{\psi}_{j\alpha}v_{j}\right ) \right ] + \frac { n}{2 j_k } tr [ v_j \bar v_{j}].\end{aligned}\ ] ] in the last step we have absorbed the minus sign associated with the anticommutation of @xmath161 and @xmath162 , likewise @xmath163 and @xmath164 .
these hubbard - stratonovich fields @xmath165 and @xmath166 are four - component spinors @xmath167 here @xmath168 and @xmath169 are complex fields related to the hybridization between f - fermions and c - electrons and the development of superconductivity by the formation of pairs between f - fermions and c - electrons , respectively . the parameters @xmath170 and @xmath171 are complex grassmann numbers , the first related to the hybridization between b - bosons and c - electrons and the second related to the development os pairs formed between b - bosons and c - fermions .
in a similar fashion , the heisenberg term decouples as : @xmath172 + \frac{n}{j_{h } } { \rm str}\left [ \bar\delta_{ij } \delta_{ij } \right]\cr & = & \sum_{\alpha } \left [ \bar \psi_{i\alpha } \delta_{ij}\psi_{j\alpha } + \bar \psi_{j\alpha } \bar \delta_{ij}\psi_{i\alpha } \right ] + \frac{n}{j_{h}}{\rm str}\left[\bar \delta_{ij}\delta_{ij}\right]\end{aligned}\ ] ] where @xmath173 is a four - by - four matrix and its conjugate is defined as @xmath174 .
the structure of the matrix is as follows @xmath175 where the diagonal block - matrices composed by c - numbers whereas the off - diagonal block matrices @xmath176 and @xmath177 are composed by grassmanians .
the internal structure of these blocks is given by @xmath178 @xmath179 @xmath180 where the remaining matrix is defined through the relation @xmath181 .
the matrix @xmath173 can be thought of as supersymmetric rvb field .
the components fields @xmath182 and @xmath183 promote hopping and pairing amongst the f - fermions in different sites and the complex fields @xmath184 and @xmath185 promote hopping and magnetic bond formation between the b - bosons .
the grassmannian parameters @xmath186 are hopping amplitudes that transmute bosons into fermions and vice versa , while the grassmannian amplitudes @xmath187 describe pairing between bosons and fermions at different sites .
the partition function now reads : @xmath188{\cal d}\mu e^{-s'},\end{aligned}\ ] ] where @xmath189 where the primes denote the actions of the hubbard - stratonovich factorized interactions @xmath190 and the integral over @xmath191 $ ] indicates the integral over all the fluctuating fields introduced by the hubbard - stratonovich transformations . here
c - number fields are represented by the latin letters @xmath192 , while the grassmannian fields are represented by the greek letters @xmath193 .
grassmannian fields are introduced in order to decouple terms with fermionic bilinears .
note that the kondo and heisenberg parts of the action are invariant under the transformation @xmath194 if the fluctuating field matrices transform accordingly as @xmath195 @xmath196 now we move to the discussion of the implementation of the constraint by fixing @xmath57 .
the constraint can be imposed as a projection operator in each site @xmath121 , written as a delta function : @xmath197 in order to treat the bosonic and fermionic components of the spin in the grand canonical ensemble , we split the constraint into two terms as follows : @xmath198 the constraint fixes the total number of bosons and fermions to @xmath199 , leaving the asymmetry of the representation free to adjust according to the energetics of the problem .
this constraint can be implemented in the path integral as a dirac delta function in its integral form : @xmath200 e^{-\sum_{j}s_p ( j ) } , \cr \mathcal{d}[q_{f},\lambda]&= & \prod_{j } \sum_{q_{fj}}d\lambda_{fj } d\lambda_{bj},\end{aligned}\ ] ] where @xmath201,\end{aligned}\ ] ] and the constraint fields , @xmath202 and @xmath203 are integrated along the imaginary axis . from these considerations @xmath204 can be rewritten as : @xmath205 where now @xmath206 and the partition function is now written as : @xmath207 { \cal d}\mu e^{-s'},\end{aligned}\ ] ] where @xmath208 the new feature of this action , is the appearance of the term @xmath209 , which tunes the bosonic / fermionic character of the representation . in the large @xmath41 limit , we will be able to replace the discrete measure of integration over @xmath209 by a continuous measure @xmath210 in the large @xmath41 limit , we anticipate that the functional integral is given by the saddle point value of the effective action , so the large @xmath41 approximation is then given by the exponential of the effective action @xmath211 } \ ] ] where @xmath212 e^{-s'}\ ] ] is the integral over the canonical spin - carrying fermions and bosons in the presence of fixed @xmath209 @xmath213 , @xmath214 , @xmath215 and @xmath216 with the conditions that @xmath217 be stationary with respect to each of its fields .
we may anticipate two classes of mean - field solution 1 .
type i solutions , in which the minimum of the effective action occurs at the extremum of the summation over @xmath218 , i.e @xmath219 or @xmath220 , corresponding to the fully fermionic or bosonic solutions .
2 . type ii solutions , in which the minimum of the effective action occurs at some intermediate value of @xmath221 .
since the action is stationary with respect to variations in @xmath218 , this implies that @xmath222 so that in this phase , the chemical potential of the bosonic and fermionic spinons are equal , @xmath223 this equality of chemical potentials allows us to consider these solutions as _ two fluid _ solutions . at the saddle points
, we can set all fermionic components of @xmath215 and @xmath216 to zero .
these terms only contribute to the fluctuations about the mean field theory , so we have @xmath224 and @xmath225 note that the fermionic and bosonic parts of the action decouple since all the matrices in the action now have the blocks linking the fermionic and bosonic subspaces equal to zero .
now it is possible to solve the fermionic and bosonic problems separately , imposing the constraint @xmath226 to the solution in the end of the calculation .
note that this provides a picture of two asymptotically independent fluids , bosonic and fermionic , in the large-@xmath41 limit and that the introduction of fluctuations will provide interactions between them .
now , as a first exploration of this idea we illustrate the formalism with two simple examples : a two - impurity and a frustrated three - impurity model and show that there are stable mean field solutions with mixed representations , as well as with purely bosonic and purely fermionic representations .
as a first application of the supersymmetric - symplectic spin , we study a minimal model that allows one to make connections to the physics of heavy fermion systems .
the model consists of two local moments interacting among themselves by a heisenberg coupling @xmath227 and interacting with its respective bath of conduction electrons by a kondo coupling @xmath17 .
the hamiltonian is written as : @xmath228 where @xmath229 is the conduction electron hamiltonian , @xmath230 is the lead and local moment index , @xmath231 the momentum and @xmath232 the spin index , which assume values @xmath233 in the large - n limit .
here @xmath234 is the spin density of conduction electrons at the site which is connected to the local moment spin @xmath235 . introducing the supersymmetric - symplectic spin ,
the hamiltonian can be written in the large-@xmath41 limit in terms of fermionic and bosonic operators as : @xmath236\cr & -&\frac{j_{h}}{n}\sum_{\alpha\beta } { \rm str}\left [ \left(\psi_{1\beta } \bar \psi_{2\beta } \right ) \left(\psi_{2\alpha } \bar \psi_{1\alpha } \right ) \right].\end{aligned}\ ] ] where @xmath237 , as defined in eq .
[ phi ] .
we can now apply the formalism introduced in the previous section .
we perform a hubbard - stratonovich transformation to decouple the interacting terms in the hamiltonian by introducing fluctuating fields . within a static mean field solution
the fermionic and bosonic problems decouple and are effectively linked only by the constraint ; we can now factor the partition function as : @xmath238=z_f ( q_{f } ) z_b ( q_{0}-q_{f } ) , \end{aligned}\ ] ] where we have defined @xmath239 , @xmath240 .
we see solutions where the free energy is minimized with respect to @xmath65 .
the fermionic part of the partition function reads : @xmath241\nonumber\end{aligned}\ ] ] @xmath242\nonumber\\ & & + \beta n \sum_a\frac { |v_{a}|^2}{j_k}-2\beta n \lambda_{f } q_f,\end{aligned}\ ] ] where we already dropped the terms in @xmath243 and @xmath244 , since it can be shown that these do not contribute to the saddle point solution .
also , the f - operators can be redefined to eliminate the pairing term between c- and f - operators from the hamiltonian : @xmath245 under these considerations , the fermionic part of the solution reduces to two decoupled impurity problems . taking @xmath246 to be site independent , integrating out the conduction electrons in each lead and transforming from imaginary time to matsubara frequencies ( see details in appendix [ appfe ] ) : @xmath247 where @xmath248 is a constant dos
, @xmath5 is the bandwidth and @xmath249 is a heaviside step function .
summing over matsubara frequencies , in the limit @xmath250 , the free energy has the form : @xmath251,\end{aligned}\ ] ] where we define @xmath252 and the kondo temperature @xmath253 in the large-@xmath41 limit the partition function is dominated by the saddle point .
minimizing the free energy with respect to @xmath254 one finds @xmath255 and substituting back into the fermionic free energy : @xmath256 the bosonic part of the partition function can be concisely written as : @xmath257\nonumber\end{aligned}\ ] ] @xmath258 where @xmath259 @xmath260 here we already dropped the fluctuating field @xmath261 since the saddle point solution results in @xmath262 . integrating out the bosons and summing over matsubara frequencies ( see appendix [ appbe ] ) , in the zero temperature limit , the free energy is given by : @xmath263 minimizing the free energy with respect to @xmath95 and @xmath264 one finds : @xmath265 up to a constant term . making explicit use of the constraint condition , @xmath266
, the total free energy can be written as : @xmath267 where @xmath268 and the free energy is given in units of @xmath227 .
for each value of @xmath269 and @xmath270 the representation was determined by the minimization of the free energy with respect to @xmath68 and the result is plotted in the _ representation diagram _ of fig .
[ pd ] . in case
the free energy is minimized for @xmath271 the phase is purely fermionic , meaning that a completely antisymmetric representation is favored .
analogously , for @xmath67 ( or @xmath272 ) the phase is purely bosonic , and a symmetric representation is more appropriate .
solutions with @xmath273 are solutions in which both bosons and fermions coexist , which we call a _ mixed phase _ and label as @xmath274 in fig .
[ pd ] . note that for a fixed value of @xmath270 , as the ratio @xmath275 is increased the spins tend to develop fermionic character . also , for a fixed value of @xmath276 , increasing @xmath277 ( or reducing @xmath270 , which is equivalent to decreasing the magnitude of the spin ) the spin representation also tends towards a fermionic representation . and @xmath277 .
a purely fermionic representation phase ( red ) is labeled by @xmath278 , a pure bosonic representation phase ( blue ) is labeled as @xmath279 and the mixed representation phase ( intermediate colors ) is labeled by @xmath280 . ] the dashed line between @xmath278 and @xmath280 regions indicates a second order phase transition and can be determined from the condition : @xmath281 which leads to : @xmath282 the continuous line represents a first order phase transition .
the line between the purely fermionic and purely bosonic representations is determined by : @xmath283}{n}=\frac{f[q_f=0]}{n},\end{aligned}\ ] ] which gives the condition : @xmath284 the first order line between the phases @xmath279 and @xmath280 can not be computed analytically and was determined numerically . throughout the mixed phase we have @xmath285 , so both fluids have the same chemical potential .
this is related to the presence of the type ii minima of the free energy ( see discussion in the introduction ) , with a saddle point that allows the coexistence of bosons and fermions and the interchange of one into another at no energy cost .
in particular , at the second order `` phase transition '' line discussed above , when @xmath286 and @xmath287 with @xmath288 , one can check explicitly from eq .
[ lb ] and [ lf ] that the condition @xmath285 gives the same condition that defines the second order line in eq .
[ 2nd ] .
we now analyze the effects of fluctuations of the local fermionic fields . in the previous section we introduced the time dependent fields @xmath289 and @xmath290 , which allow us to decouple the terms @xmath291 and @xmath292 in the action , respectively .
these fields do not acquire an expectation value , but fluctuate around zero .
the partition function can now be written in terms of the saddle point solution determined in the former subsection times @xmath293 and @xmath294 , the new contributions to the partition function due to the presence of the fluctuating fields @xmath295 and @xmath296 , that we take to be site independent .
focusing first in the @xmath297 field : @xmath298}.\end{aligned}\ ] ] expanding to second order in @xmath297 we can identify the propagator for the fluctuating field @xmath297 : @xmath299^{-1}= 2n\left [ \chi_{cb}(i\omega_r ) -\frac{1}{j_k}\right],\end{aligned}\ ] ] with @xmath300{bubbleanew.pdf}}\nonumber\\ & = & -\frac{1}{\beta}\sum_{\bk \bk ' m}g_{b } ( i\omega_{r}+i\omega_m ) g_{\bk\bk'}(i\omega_m),\nonumber\end{aligned}\ ] ] where @xmath301 is the bosonic propagator and @xmath302 is the full c - electron propagator .
we now evaluate @xmath303 ( details of the calculation can be found in appendix [ appflu ] ) .
one interesting region for the analysis of @xmath304 is the second order transition line , where the energy levels of the bosons and fermions are equal . in the infinite bandwidth limit
we find : @xmath305\nonumber\\&-&\frac{\rho_0 \omega ^2}{\delta ^2+\omega ^2}\log ( \lambda -\omega + i\delta).\end{aligned}\ ] ] note that at zero frequency , @xmath306 , so @xmath307^{-1}= 0 $ ] , and the propagator for the fermionic hybridization field @xmath308 diverges at zero frequency at the second order phase transition , indicating the presence of a fermionic zero mode . also , there is a gap of magnitude equal to @xmath309 with a continuum that goes up to the bandwidth .
this gap is always present in the 2-impurity model since @xmath310 is aways finite at the transition . for a kondo - heisenberg model in the lattice
, the bosonic level will acquire a dispersion and when magnetic order sets in it will be gapless at some points in the brillouin zone . in that case the spectrum for the fermionic hybridization field is expected to have a continuum of excitations , which can potentially lead to non fermi liquid behavior .
, the @xmath297 propagator .
the parameters used in this plot were a solution of the mean field theory at a specific point of the second order phase transition : @xmath311 and @xmath312 , which gives @xmath313 and @xmath314 ( in units of @xmath227 ) . ] for the second fermionic mode @xmath296 a similar calculation follows , where now : @xmath315^{-1}= 2n\left [ \bar{\chi}_{cb}(i\omega_r ) -\frac{1}{j_k}\right],\end{aligned}\ ] ] with @xmath316{bubblebnew.pdf}}\nonumber\\ & = & \frac{1}{\beta}\sum_{\bk \bk ' m}g_{b } ( i\omega_{r}-i\omega_m ) g_{\bk\bk'}(i\omega_m),\nonumber\end{aligned}\ ] ] and we can write , at the second order transition line : @xmath317\nonumber\\&-&\frac{\rho_0 ( \omega-2\lambda)^2}{\delta ^2+(\omega-2\lambda)^2}\log ( \lambda -\omega + i\delta).\end{aligned}\ ] ] here we note that there is no zero mode for this fermionic field at the transition , and a continuum starts at a finite @xmath310 .
this is related to the choice we made to implement the constraint , which is not invariant under all transformations that leave the spin invariant .
in particular , it is not invariant under transformations of the form @xmath318 ( see appendix [ appss ] ) , generated by the operators @xmath319 and @xmath320 .
as a second application of the supersymmetric - symplectic spin , we study a minimal model that brings in the issue of geometric frustration into play .
the model consists of three local moments interacting among themselves by an antiferromagnetic heisenberg coupling @xmath227 and interacting with its respective bath of conduction electrons by a kondo coupling @xmath17 , as depicted in fig .
we are motivated to look at this problem by experiments in cepdal in which the equivalent ce sites spontaneously develop a state in which one third are paramagnetic and the other two thirds are magnetically ordered @xcite , exploring the ability of the symplectic representation of the spin to describe frustrated systems .
the hamiltonian is written as : @xmath321 where now @xmath322 is the lead and local moment index with periodic boundary conditions .
as in the previous section , @xmath323 is the conduction electron hamiltonian , @xmath234 is the spin density of conduction electrons at the site that is connected to the local moment spin @xmath235 . introducing the supersymmetric - symplectic spin from eq .
[ s ] into the hamiltonian , this can be written in the large-@xmath41 limit as : @xmath324\cr & -&\frac{j_{h}}{n}\sum_{a,\alpha \beta } { \rm str}\left [ \left(\psi_{a\beta } \bar \psi_{a+1\beta } \right ) \left(\psi_{a+1\alpha } \bar \psi_{a\alpha } \right ) \right].\end{aligned}\ ] ] with @xmath325 defined in eq .
[ phi ] .
proceeding as in the previous section , within a path integral formalism we introduce fluctuating fields in oder to decouple the quartic terms in the action and impose the constraint by the introduction of a delta function in the integral form . within a static saddle point solution
the problem decouples in to a bosonic and a fermionic part , effectively linked by the constraint . in this section
we are going to leave the representation of the spin and the mean field parameters to be determined independently in each site . omitting the details ( similar to the previous section )
, the partition function in the large @xmath41 limit can be written as @xmath326 with the understanding that the partition function is to be stationary with respect to the @xmath327 at the three sites .
the fermionic part of the partition function can be written as : @xmath328\nonumber\end{aligned}\ ] ] @xmath329\nonumber\\&&+\beta n \sum_{a}\frac { |v_{a}|^2}{j_k}-\beta n \sum_a \lambda_{fa } q_{fa},\end{aligned}\ ] ] where as in the previous section , we assume that the only fluctuating field that acquires a finite value at the saddle point solution is @xmath330 .
in this case the fermionic part of the solution reduces to three decoupled impurity problems , with the same solution as the previous section , now for 3 leads : @xmath331 the bosonic part of the partition function reads : @xmath332e^{-s_b } , \end{aligned}\ ] ] where @xmath333\nonumber\\&-&\beta n\sum_a\lambda_{ba } ( q_{ba}+1/2),\end{aligned}\ ] ] where we denote @xmath334 and @xmath335 with @xmath336 @xmath337 and @xmath338 note that the trace term in the action now appears with a minus sign since it is related to the bosonic part of the super - trace introduced in eq .
[ str ] . here
we define @xmath339 as the bosonic part of the original matrix @xmath340 .
the solution of the bosonic part of the partition function is more involved if we allow the representations of the spin to be different in each lead . as a first solution we consider the same representation on every site , and look for an _ homogeneous solution _ , taking @xmath341 , @xmath342 and @xmath343 . integrating out the bosons and summing over matsubara frequencies , in the zero temperature limit
, the free energy can be written as : @xmath344 minimizing the free energy with respect to @xmath261 , @xmath95 and @xmath264 one finds the saddle point free energy : @xmath345 up to a constant term .
the total free energy for the homogeneous solution can be written , already making explicit use of the constraint condition @xmath266 , as : @xmath346 where again @xmath268 and the free energy is given in units of @xmath227 .
note that this is functionally the same as the 2-impurity model up to an overall factor of @xmath347 , and as a consequence the representation diagram determining the most favorable representation for the spin within an homogeneous solution will be identical to the 2-impurity case . and @xmath277 .
the color code is the same as in fig .
the hashed area represents the region of the diagram in which an inhomogeneous solution is energetically favorable . ]
now we move on to investigate solutions which spontaneously develop different representations in each site , which we refer to as _
inhomogeneous representations_. due to frustration , we expect that it is energetically favorable for one of the spins to be in a fermionic representation , essentially disconnected from the other two spins with a bosonic or mixed representation , forming an antiferromagnetic bond .
we assume one of the spins to always have a fermionic representation and let the representation of the two other spins to be selected as the one that minimizes the total energy .
now the problem reduces to a single impurity problem with a purely fermionic representation plus the two - impurity problem solved in the previous section .
the free energy for the inhomogeneous solution reads : @xmath348 in units of @xmath227 , where @xmath268 .
again , the representation diagram will be the same as before , but now we compare the free energies of the homogeneous and inhomogeneous solutions for the 3-impurity problem . the hashed area in fig .
[ pd2 ] is the region of the diagram in which the inhomogeneous solution is more favorable .
this result provides a model for the situation which appears to occur in cepdal @xcite : one third of the local moments in the frustrated kagome lattice ( formed by an assembly of corner - shared triangles ) relieve the frustration by assuming an antisymmetric character and forming a kondo singlet with a conduction electron , while the other two thirds of the local moments assume a bosonic character , developing magnetic order , allowing a partially ordered phase to be formed , as depicted in fig .
[ inhom ] .
in this work we introduced a new supersymmetric - spin representation for large-@xmath41 treatments , based on a symplectic generalization of the spin operator .
we have analyzed the properties of the supersymmetric - symplectic spin and its symmetries , identifying the supergroup @xmath349 of transformations under which the spin is invariant .
we have proposed a new framework in the large-@xmath41 limit , which allows the problem to sample different representations , selecting the one which lowers the energy in a given point in parameter space .
this opens up the possibility of describing the phase diagram of heavy fermions within a single approach that can capture the evolving character of the spin , at the same time that it offers a potential framework for the phenomenological two - fluid picture for heavy fermions . applying this approach to two toy models , the two - impurity model and the frustrated three - impurity model , we have shown two new classes of mean - field theory that may be of interest in developing a unified description of heavy fermion systems . in particular
, we find a _
mixed phase _ solution in which bosons and fermion coexist at each site , which points the way to a description of the coexistence of magnetism and kondo effect .
also , we find stable _ inhomogenous solutions _ , which may provide a basis for describing the partially ordered state in cepdal @xcite .
there are many open questions raised by our work .
firstly , within the mixed phase , we find evidence for a new kind of zero mode , a _ goldstino _ , that results from the partial breaking of supersymmetry .
this can be understood as a consequence of the fact that the hamiltonian is invariant under super - rotations while at the same time , the mean - field solution breaks this rotational symmetry .
given a state @xmath350 with a fermion assigned to the corner box , for example , one can rotate this state as follows : latexmath:[\ ] ] which is the form of @xmath456 discussed in the main text .
one way to avoid the constraint is to use the fedotov - popov trick @xcite , by the introduction of an imaginary chemical potential , but note that this approach makes the hamiltonian non - hermitian and the convexity arguments used in variational approaches to determine the solution as an upper bound to the real ground state energy can not be applied in this case . | we propose a generalization of the supersymmetric representation of spins with symplectic symmetry , generalizing the rotation group of the spin from @xmath0 to @xmath1 . as a test application of this new representation , we consider two toy models involving a competition of the kondo effect and antiferromagnetism : a two - impurity model and a frustrated three - impurity model .
exploring an ensemble of l - shaped representations with a fixed number of boxes in their respective young tableaux , we allow the system to choose which representation is energetically more favorable in each region in parameter space .
we discuss how the features of these preliminary applications can generalize to kondo lattice models . |
note that , in general , @xmath119 .
now , the maximization for @xmath120 in eq . can be obtained by choosing the povms @xmath121 and @xmath122 to be the projections onto @xmath123 , and @xmath1 to be the transpose of the extremal point of @xmath38 , @xmath124 , which gives @xmath125 .
thus , by making use of eq . , we obtain @xmath126 .
it is sufficient to prove that a quantum state @xmath127 if and only if @xmath128 . the proof is as follows . for any state @xmath1 , we can write @xmath129 to prove the necessary part , we note that for every separable state @xmath130 , the right - hand - side of eq .
is just zero by using eq . .
thus , the pay - off can never exceed zero , and hence , the players can at best obtain the same pay - off as @xmath131 .
the sufficient part follows from the fact that , using criteria 1 , for every entangled state @xmath25 there exists an extremal witness @xmath38 such that @xmath132 . since all the summands on the right - hand - side of eq
are nonnegative , it follows that @xmath133 .
the set @xmath63 being convex means that @xmath134 ,
\quad \hat{\eta}_{\rm ab}= p \hat{\varrho}_{\rm ab } + ( 1-p ) \hat{\sigma}_{\rm ab } \in \mathcal{s}_\lambda.\ ] ] it is sufficient to show that @xmath135 we note that @xmath136 where the third inequality is a result of the fact that @xmath137 . the first condition is already proven as criterion 1@xmath48 .
let us prove the condition ( ii ) in a more general context , by allowing the players to make separable operations , @xmath144 , on both their respective parts of shared state and quantum questions .
given that alice and bob receive the ensemble of questions @xmath145 , any normalized separable operation can be written as a convex combination of completely positive trace - preserving separable operations , with kraus decomposition @xmath146 where @xmath147 and @xmath148 .
using the dual operation , we have that @xmath149 , in which @xmath150 is a convex combination of locc effects for alice and bob .
we note that any convex combination of locc effects is also a locc effect .
therefore , it holds true that @xmath151 the inequality is obtained by noticing that a linear function on a convex set achieves its maximum at the extremal points .
we note that the particular requirement of @xmath152 follows as a special case , after maximizing both sides of eq . over all eews ,
if no operation is done on the input questions . finally , a measure of entanglement is universal if it is invariant under all local invertible operations .
this is shown to be the case if the measure is nonincreasing under all separable operations @xcite , which holds true for @xmath66 .
suppose that there are two witnesses @xmath155 and @xmath82 .
the former , corresponding to the game played by charlie , does not detect the npt entangled state @xmath1 shared by alice and bob in the standard witnessing scenario .
the latter , however , detects it .
we know that @xmath76 with schmidt - rank @xmath79 can be transformed into @xmath80 with schmidt - rank @xmath156 via a stochastic locc ( slocc ) map @xmath157 with the success probability @xmath158 @xcite .
that is @xmath159 , where @xmath160 .
slocc operations are a subset of separable ones , and thus linear , with kraus decompositions of the form @xmath161 .
thus , their partial transpose , @xmath162 , is completely positive , and in fact , slocc .
therefore , @xmath163 .
this means that any decomposable eew with schmidt rank @xmath79 can be transformed into an arbitrary decomposable eew with schmidt rank @xmath156 with nonzero probability @xmath164 .
it is clear that the players do not have direct access to the witness as the reward function is fixed by the referee to correspond to @xmath81 alice and bob can not negotiate with the referee on that however , they have the possibility to operate with @xmath165 on their respective questions . then , if alice and bob can win a game @xmath166 corresponding to witness @xmath167 , i.e. , achieve an average reward @xmath168 via measuring some joint locc effect @xmath169 , we have @xmath170 the result is positive because @xmath167 detects the entanglement of @xmath1 upon measurement of @xmath171 , no matter how small the probability @xmath164 is .
equation simply implies that the players can always win a positive pay - off ( even though very small due to the probability @xmath164 being small ) by making use of an appropriate slocc strategy @xmath165 , if their shared state is npt entangled , regardless of the decomposable eew @xmath81 chosen by the referee .
importantly , we may also rearrange eq . as @xmath172 equation represents the possibility of considering slocc as part of the locc measurement strategy , @xmath173 , which incorporates the probability of success of the operation as part of the probability of obtaining the @xmath174 outcome corresponding to the effect @xmath175 .
that is , @xmath176 . as a result of the above discussion
, condition ( i ) follows immediately from the fact that for any npt entangled state in a @xmath78 dimensional hilbert space , there exists a decomposable eew @xmath82 with the schmidt rank of at most @xmath79 which detects it .
thus , a state is ppt if and only if a detection is impossible ( recall that a decomposable witness can only detect the entanglement of npt states ) , which in turn implies @xmath88 .
now , we prove the condition ( ii ) in a more general context , again , by allowing the players to make further separable operations on both their respective parts of the shared state and quantum questions .
given that alice and bob receive the ensemble of questions @xmath145 , any normalized separable operation can be written as a convex combination of completely positive trace - preserving separable operations , with kraus decomposition @xmath146 where @xmath147 and @xmath148 .
suppose that @xmath82 is the witness that detects the shared state @xmath1 , while alice and bob play the game @xmath85 corresponding to the witness @xmath75 with charlie .
the players make a slocc on their respective questions ( or equivalently , on their effects ) as described in eq . .
moreover , suppose that an appropriate measurement denoted by @xmath175 gives the optimal pay - off for the game @xmath85 .
that is , in eq .
, @xmath177 .
we prove that further locc operations @xmath178 on the shared state can not increase the pay - off . using the dual operation
, we have that @xmath179 , in which @xmath180 is a convex combination of locc effects for alice and bob .
we note that any convex combination of locc effects is also a locc effect and thus , it holds true that @xmath181 the inequality is obtained by noticing that a linear function on a convex set achieves its maximum at the extremal points .
we note that the particular requirement of @xmath182 follows as a special case , if no operation is done on the input questions . as before ,
@xmath91 is invariant under all local invertible operations on the state , due to the fact that such operations are a subset of all separable operations .
therefore , it is a universal measure of npt entanglement .
we emphasize that for a different game @xmath183 corresponding to the schmidt rank-@xmath79 decomposable eew @xmath184 , one get another universal measure of npt entanglement @xmath185 .
however , if @xmath186 can be obtained from @xmath81 using locc ( that is , deterministically ) then @xmath187 .
it is known that @xmath76 is locc convertible to @xmath188 if and only if the vector of schmidt coefficients of @xmath76 , @xmath189 , is majorized by the vector of schmidt coefficients of @xmath188 , @xmath190 , denoted by @xmath191 @xcite . in particular , the vector of schmidt coefficients of @xmath192 is majorized by any vector with schmidt rank @xmath79 . as a result
, @xmath193 can be transform into any schmidt rank @xmath79 vector using ( deterministic ) locc . in light of the previous discussion
, charlie can choose the witness to be @xmath194 , that is the swap operator @xmath195 , and make the game easier for alice and bob , while remaining robust against cheating . | within the context of semiquantum nonlocal games , the trust can be removed from the measurement devices in an entanglement - detection procedure . here
we show that a similar approach can be taken to quantify the amount of entanglement . to be specific , first
, we show that in this context a small subset of semiquantum nonlocal games is necessary and sufficient for entanglement detection in the locc paradigm .
second , we prove that the maximum pay - off for these games is a universal measure of entanglement which is convex and continuous .
third , we show that for the quantification of negative - partial - transpose entanglement , this subset can be further reduced down to a single arbitrary element .
importantly , our measure is operationally accessible in a measurement - device - independent way by construction .
finally , our approach is simply extended to quantify the entanglement within any partitioning of multipartite quantum states .
[ [ introduction . ] ] introduction. + + + + + + + + + + + + + + entanglement is a valuable resource for practical as well as fundamental applications of quantum theory , ranging from quantum computation and communication to metrology @xcite .
there are two major challenges in understanding entanglement that stimulates this research .
it is extremely difficult to specify all the nonentangled bipartite or multipartite quantum states .
in fact , the problem is known to be np - hard @xcite .
not surprisingly , the characterization of entangled states is an equally difficult task .
that is , to quantify the amount of entanglement within a quantum state .
the answer to this question is practically very important , because it tells us how well our protocols will perform using a given state @xcite .
recent work by buscemi @xcite has introduced a new way to think about entanglement detection @xcite .
the idea is to map the problem onto a modified class of nonlocal games , called semiquantum nonlocal games ( sqnlgs ) . in any such game , two players ( alice and bob ) share a possibly entangled state .
a referee ( charlie ) starts asking them quantum questions by sending quantum states and receiving classical answers the outcomes of some local measurements ( see fig [ protocol ] ) .
he then evaluates a _ reward _ function from coincidence statistics and pays the players accordingly . confined to not communicate during the game , known as the local operations and shared randomness ( losr ) paradigm , all separable states deliver an equal _ pay - off _ ( maximum average reward ) in any specific sqnlg .
importantly , for every entangled quantum state one can always find a sqnlg which can deliver a larger pay - off than any separable state .
this mapping allows one to merely rely on the coincidence statistics of measurement outcomes , rather than their specific expectation values , to violate an entanglement witnessing inequality @xcite ; a property that was believed to be specific to bell nonlocality scenarios .
consequently , researchers interpreted buscemi s results as a clever way to remove the trust from measurement devices in an entanglement witnessing procedure , since any entanglement witness ( ew ) can be recast as a sqnlg @xcite .
focusing on the second challenge , a first level of hardness is that the quantification of entanglement using almost any entanglement measure , e.g. , entanglement of formation @xcite , negativity @xcite , or random robustness @xcite , requires estimating a large number of density matrix elements ; a task which is difficult to perform on bipartite and multipartite quantum states .
while this difficulty can be partially circumvented by making use of ews when lower bounds on the entanglement are desired @xcite , errors and misalignments of the measurement devices can still lead to the incorrect estimations of the quantities and thus , erroneous conclusions . in this letter , inspired by buscemi s approach , we consider sqnlgs in the paradigm of local operations and classical communication ( locc ) .
we show that a small subset of games which we call extremal semiquantum witnessing games ( esqwgs ) are both necessary and sufficient for the full characterization of entangled states .
we then focus on the entanglement of negative - partial - transpose ( npt ) states , as the necessary ingredient for distillability @xcite .
we present the answer to the practical appeal for a measurement - device - independent ( mdi ) npt - entanglement measure by proving that npt entanglement can be quantified by a referee in _ a single arbitrary _ esqwg .
the main result of the present letter is thus to introduce an operationally accessible measure ( for both general and npt ) entanglement that is convex , universal , and most importantly mdi by construction .
furthermore , we extend our measure to quantify the entanglement in all possible partitions of multipartite quantum states . [ [ from - sqnlgs - to - sqwgs . ] ] from sqnlgs to sqwgs. + + + + + + + + + + + + + + + + + + + + + + let us start by describing sqnlgs more rigorously .
a sqnlg is a collaborative game , denoted here by @xmath0 , in which alice and bob share a quantum state @xmath1 .
charlie prepares a set of quantum questions @xmath2 and @xmath3 with probability @xmath4 and sends them to alice and bob , respectively .
the players respond to each question classically from two sets of labels @xmath5 and @xmath6 . before the game starts ,
given the losr paradigm , they can agree on a best strategy to win the game , however , during the game they are no longer allowed to communicate .
for each question @xmath7 , charlie evaluates a reward corresponding to the answers @xmath8 and @xmath9 according to the function @xmath10 . the average reward of the game is then given by @xmath11 in which @xmath12 , @xmath13 , and the joint probability distribution @xmath14 is @xmath15 where @xmath16 and @xmath17 are local effects ( povm elements ) of the players .
they win or lose some value if the average reward is positive or negative , respectively .
the players goal is of course to maximize the average amount they can obtain in a game .
let us call the maximum average reward the _ pay - off _ value @xcite and denote it by @xmath18 the main result of buscemi @xcite , relevant for entanglement detection , can be recast as follows .
given the set of all sqnlgs , @xmath19 , and the set of all separable states , @xmath20 , for any game @xmath21 and for any two states @xmath22 , one has @xmath23 this simply reads as _ all separable quantum states , at best , are equal in a sqnlg_. [ [ criterion1.-buscemi . ] ] criterion 1 .
( buscemi ) .
+ + + + + + + + + + + + + + + + + + + + + + + a quantum state @xmath1 is entangled if and only if there exists a sqnlg for which @xmath24 .
it is relevant to ask whether one should search within the whole set @xmath19 for a game violating the equality ?
the short answer is in negative @xcite . without loss of generality ,
we assume that the hilbert spaces are finite dimensional , since entanglement can always be verified in finite dimensional subspaces @xcite .
it is well - known that , by hahn - banach theorem , for any entangled state @xmath25 there exist an ew @xmath26 such that @xmath27 note that , here , for the sake of consistency , we have changed the sign of the usual convention .
moreover , we set @xmath28 , with @xmath29 being the minimum dimensionality of alice and bob s subsystems , to compare different ews , and we note that such a normalization is always possible @xcite .
now , every ew can be transformed into a sqnlg as follows .
charlie decomposes the witness in terms of product states as @xmath30 , with @xmath31 denoting the transposition operation and @xmath32 , and defines a sqnlg via @xmath33 it is important to note that this correspondence is not unique .
now , we can rewrite eq .
as @xmath34 we call any such a game a semiquantum witnessing game ( sqwg ) and denote the set of all such games by @xmath35 . we may point out here that , the decomposition of the ew can be generalised to be in a nonseparable basis .
branciard _ et al _
@xcite showed that the set @xmath35 is , indeed , necessary and sufficient for verifying the entanglement of a state @xmath1 shared by the players .
that is , @xmath36 , and for any entangled state @xmath25 there exist a sqwg such that @xmath37 .
[ [ extremal - sqwgs . ] ] extremal sqwgs. + + + + + + + + + + + + + + + + in general , every ew is ( with our sign convention ) a member of the compact convex set of normalized block - negative operators ( defined as operators with negative expectation values in all pure product states ) .
an extremal ew ( eew ) @xmath38 is the one that can not be written as a convex combination of any two other block - negative operators , and hence , there exists a pure product state @xmath39 such that @xmath40 .
we now introduce the set of _ extremal semiquantum witnessing games _ ( esqwgs ) , @xmath41 , which correspond to eews .
this class of games is necessary and sufficient for entanglement detection , since for every entangled state there exists an eew which detects it @xcite .
a very important corollary thus follows . for any @xmath42 ,
we have that @xcite @xmath43 we extend this statement by , first , allowing the local effects to be _ relabeled _ @xcite .
this is the procedure of shuffling the labels of the measurement effects and possibly assigning the same label to multiple outcomes with the help of classical communication .
this leads to locc effects of the form @xmath44 which can be obtained using locc @xcite . here ,
@xmath8 and @xmath9 are labels to be sent to charlie conditioned on @xmath45 and @xmath46 which characterize the outcomes obtained locally and communicated between alice and bob .
note that , any locc povm is necessarily separable , but the converse is not true @xcite .
next , by substituting this into eqs . and , and restricting the games to extremal ones , we define @xmath47 consequently , we have the following entanglement criterion as our first result .
[ [ criterion-1prime . ] ] criterion 1@xmath48 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a quantum state @xmath1 is entangled if and only if @xmath49 . the proof is given in the supplemental material @xcite .
the importance of criterion 1@xmath48 is that , in contrast to criterion 1 , it reduces the entanglement detection down to a much smaller set of games while relaxing to general locc measurements .
[ [ mdi - quantification - of - entanglement . ] ] mdi quantification of entanglement. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + it is interesting that criterion 1@xmath48 also provides an equivalent way to define the set of separable states as the set of all quantum states providing a maximum pay - off of zero : @xmath50 .
this also induces the idea that there exists the following continuous hierarchy of sets .
[ [ definition-1 . ] ] definition 1 .
+ + + + + + + + + + + + + for any @xmath51 , we define @xmath52 . then , for any @xmath53 , @xmath54 with @xmath55 if and only if @xmath56 ; cf . fig .
[ hierarchy ] . .
the color gradient represents the increase of the function @xmath57 . ]
importantly , the set @xmath58 is convex as shown in the supplemental material @xcite .
clearly , a quantum state @xmath1 for which @xmath59 does not belong to the set @xmath58 .
conversely , for any @xmath60 there exists an esqwg @xmath61 and an effect @xmath62 for alice and bob such that they can obtain a pay - off value @xmath59 . to show this
, we note that by hahn - banach theorem there exists a ( nonextremal ) witness @xmath26 for the convex set @xmath63 which detects @xmath1 , and that it can be optimized @xcite .
the resulting optimal witness can be written as a convex combination of extremal points for which at least one of them detects @xmath1 .
[ [ observation-1 . ] ] observation 1 .
+ + + + + + + + + + + + + + @xmath64 if and only if @xmath65 .
observation 1 tells us that esqwgs are also necessary and sufficient for characterizing the continuum of the convex sets @xmath58 via @xmath66 .
moreover , we see that the average reward function provides a lower bound on the amount of entanglement shared by alice and bob . if , for a given quantum state @xmath1 , the reward value that alice and bob obtain in an esqwg is @xmath67 , then @xmath65 for any @xmath68 .
we now state the second result of this letter in the theorem below and point the interested reader to the supplemental material for the proof @xcite . [
[ theorem-1 . ] ] theorem 1 .
+ + + + + + + + + + the pay - off @xmath66 is a universal and faithful measure of entanglement . by universality ,
we mean the value of the measure is invariant under all local invertible operations @xcite .
importantly , not only @xmath66 is a measure of entanglement for the shared state , but also allowing the players to access infinite rounds of locc on input questions will not improve their best achievement .
consequently , we can relax the losr restriction in esqwgs to locc @xcite .
nevertheless , it is clear that this task of measuring @xmath66 is practically challenging in high dimensions .
shortly , we will provide a particularly interesting scenario where the referee is only interested in the amount of npt entanglement which , in turn , eliminates the need for the maximization over all eews , thus removes the aforementioned difficulty . at this point , it is important to mention that the two maximizations in eq .
are , in principle , performed independently by the referee and the players : alice and bob are responsible for the optimization of their measurements , and charlie must choose the optimal game .
the former is less important , because @xmath69 guarantees that the average reward always gives a lower bound on the amount of entanglement of @xmath1 . to charlie
, the average reward @xmath70 can be considered as the _ effective entanglement _ shared by alice and bob .
this is the amount of entanglement contained within their shared state @xmath1 extracted by their locc effect .
we note that , charlie s payment is based on the quantum questions he prepares himself , and the coincidence statistics of the responses from alice and bob . thus , he does not need to make any assumptions about alice and bob s measurements in any form , as long as they are spatially separated .
however , he should hide the indices of the questions by ensuring that his questions can not be unambiguously discriminated and that there are no side channels from his lab to alice and bob @xcite . the players can increase @xmath70 by either sharing a more entangled state or using a better locc strategy .
we also emphasize that there is no need for the referee to trust the players ; if the players do not perform their optimization appropriately they will incur losses .
the latter maximization requires that charlie perform his optimization to choose the best witness ( see , e.g. , ref .
@xcite ) before the game starts , because he must fix his choice of the game and come to an agreement with the players on it .
we also point out that it is still possible for charlie to carry out the optimization after finishing his questions merely based on the outcome statistics . for this ,
alice and bob chose an optimal eew , say @xmath71 , depending on their shared state and take an optimal strategy for that . since the set of games that charlie is optimizing over contains @xmath71 , then he will definitely find the pay - off expected by alice and bob through his optimization .
however , this requires charlie s questions to be informationally complete so that any eew can be expanded in terms of them @xcite . [
[ mdi - quantification - of - npt - entanglement . ] ] mdi quantification of npt entanglement. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + it is a well - known fact that there are two types of entangled states , namely , positive- and negative - partial - transpose ( p- and npt ) entangled states which possess legitimate or unphysical density operators upon partial transposition operation , respectively .
it is also known that the latter is necessary for distillability @xcite , representing the importance of npt entangled states .
similarly , eews are divided into indecomposable and decomposable classes where the latter only detects npt entangled states @xcite . denoting the corresponding games as @xmath72 and @xmath73 , respectively
, we have @xmath74 .
now , we notice that decomposable eews are sufficient for detection of npt entangled states and possess a very simple structure @xcite ; they are of the form @xmath75 , where @xmath76 is a normalized entangled vector and @xmath77 denotes the partial transposition operation with respect to the second party . in a @xmath78 dimensional hilbert space
, we may further restrict the vectors @xmath76 to have a schmidt rank of @xmath79 .
the main point here is that , for any schmidt - rank-@xmath79 pure entangled state @xmath76 there exists a stochastic locc ( slocc ) procedure that converts @xmath76 into an arbitrary pure entangled state @xmath80 with the same schmidt rank @xcite .
the class slocc is a subset of separable operations , implying that the partial transpose of a slocc map is also slocc .
therefore , @xmath81 can be transformed into any other decomposable eew @xmath82 with a nonzero probability @xmath83 via slocc . in the supplemental material @xcite , we discuss and prove in detail how alice and bob can exploit this possibility to win a positive pay - off in an arbitrary schmidt - rank-@xmath79 decomposable esqwg chosen by charlie , if and only if they share a npt entangled state . as a result , the maximization on esqwgs in eq
. becomes unnecessary for charlie , if we restrict the games to schmidt - rank-@xmath79 decomposable extremal ones .
it is also important to note that this restriction replaces the set @xmath20 with the set of ppt states @xmath84 .
[ [ theorem-2 . ] ] theorem 2 .
+ + + + + + + + + + for every schmidt - rank-@xmath79 decomposable esqwg @xmath85 , the pay - off @xmath86 defines a universal measure of npt entanglement . clearly , @xmath87 and @xmath88 if and only if @xmath89 .
we also note that the universality follows from the fact that all completely positive local operations ( in particular , invertible ones ) preserve ppt and npt entanglement . furthermore
, changing the game to a different schmidt - rank-@xmath79 decomposable esqwg provides a different universal measure of npt entanglement .
we emphasize that , the whole procedure described here is mdi and thus , it is not possible for the players to cheat and convince the referee that they have more npt entanglement than that contained in their state .
we also emphasize that there is no ppt entanglement for systems with dimensions up to @xmath90 , and thus their entanglement can be perfectly characterized using @xmath91 .
consider , as an example , the schmidt - rank-@xmath92 decomposable eew @xmath93 , where @xmath94 is a bell state . in a standard witnessing procedure , the bell state @xmath95 is detected by maximally violating the witnessing inequality of eq .
, @xmath96 , while the other bell states can not be detected using @xmath81 and require different witnesses .
in an esqwg corresponding to @xmath81 , on the other hand , by sharing @xmath97 alice and bob will win the pay - off @xmath98 , if they perform the projection onto @xmath99 , just as the witness @xmath81 could not detect their state in the standard witnessing procedure . theorem 2
, however , states the contrary because the shared state is indeed npt entangled .
it can be easily checked that if alice and bob project onto @xmath100 . as a result ,
in accordance with theorem 2 , both @xmath97 and @xmath101 are maximally npt entangled as measured by @xmath91 .
[ [ multipartite - extension . ] ] multipartite extension. + + + + + + + + + + + + + + + + + + + + + + + + it is straightforward to extend our approach to quantify the entanglement within any partitioning of a multipartite quantum state . in such scenarios , there are @xmath102 players denoted by the index set @xmath103 .
a @xmath104-partition of them is uniquely specified by the set @xmath105 such that @xmath106 , and corresponds to the cut @xmath107 .
this time , the referee and the players first agree on a partitioning @xmath108 , meaning that the players within the same party @xmath109 ( @xmath110 ) can perform joint ( global ) measurements on their respective questions , while the group of players in different parties are confined to locc , named as @xmath111-locc here .
the question set for each player is @xmath112 ( @xmath113 ) , chosen at random with joint probability @xmath4 , and the reward function is @xmath114 , where @xmath115 is the set of possible answers returned by the @xmath116th player .
the aim of the players is to maximally win the game . according to refs .
@xcite , in general , multipartite entanglement has a highly complex structure .
however , the subset of witnesses extremal to the set of @xmath108-separable quantum states is necessary and sufficient for detecting entanglement within @xmath108 . depending on the partitioning
, charlie thus performs the optimization over all such games denoted as @xmath117 .
[ [ theorem3 . ] ] theorem 3 .
+ + + + + + + + + + the pay - off @xmath118 is a faithful universal measure of entanglement with respect to the partitioning @xmath108 .
the proof follows from the same line of proof of theorem 1 .
[ [ conclusions . ] ] conclusions. + + + + + + + + + + + + + we showed that entanglement can be quantified operationally in a measurement - device - independent way within the context of extremal semiquantum witnessing games , a subclass of semiquantum nonlocal games , and in the locc paradigm .
thus , we reduced the whole set of games down to a much smaller subset of games .
we proved that the locc does not help the players to increase their maximum reward for a fixed amount of effective shared entanglement . in this way
, the average reward provides a lower bound on the amount of entanglement within the shared state while the pay - off value provides a universal convex measure of entanglement .
we also showed that an arbitrary decomposable member of this class of games is necessary and sufficient for both detection and quantification of npt entanglement , and thus , we reduced the whole set of games down to a single arbitrary game in such scenarios .
we also extended our approach to the multipartite scenario where quantification of entanglement within an arbitrary partitioning of a multipartite quantum state is desired .
[ [ acknowledgements . ] ] acknowledgements. + + + + + + + + + + + + + + + + + + the authors acknowledge useful discussions with maciej lewenstein and fabio costa .
this project was supported by the australian research council centre of excellence for quantum computation and communication technology ( ce110001027 ) .
100 m. a. nielsen and i. l. chunang , _ quantum computation and quantum information _ , ( cambridge university press , cambridge , 2000 ) .
r. horodecki , p. horodecki , m. horodecki , and k. horodecki , rev .
mod .
phys . * 81 * , 865 ( 2009 ) .
the ligo scientific collaboration , nat .
phys .
* 7 * , 962 ( 2011 ) .
l. gurvits , in _ proceedings of the thirty - fifth annual acm symposium on theory of computing , 2003 _ , ( acm , new york , 2003 ) , p. 10 .
s. gharibian , quantum info . comput . * 10 * , 343 ( 2010 ) .
a. furusawa , j. l. srensen , s. l. braunstein , c. a. fuchs , h. j. kimble , and e. s. polzik , science * 282 * , 706 ( 1998 ) .
s. parker , s. bose , and m. b. plenio , phys .
rev .
a * 61 * , 032305 ( 2000 ) .
d. gross , s. t. flammia , and j. eisert , phys .
rev .
lett . * 102 * , 190501 ( 2009 ) . c. gross , t. zibold , e. nicklas , j. estve , and m. k. oberthaler , nature ( london ) * 464 * , 1165 ( 2010 ) .
f. buscemi , phys .
rev .
lett . * 108 * , 200401 ( 2012 ) . c. branciard , d. rosset , y - c .
liang , and n. gisin , phys .
rev .
lett . * 110 * , 060405 ( 2013 ) .
p. xu , x. yuan , l .- k .
chen , h. lu , x .- c .
yao , x. ma , y .- a .
chen , and j .- w .
pan , phys .
rev .
lett . * 112 * , 140506 ( 2014 ) .
m. nawareg , s. muhammad , e. amselem , and m. bourennane , sci . rep . * 5 * , 8048 ( 2015 ) .
e. verbanis , a. martin , d. rosset , c. c. w. lim , r. t. thew , and h. zbinden , phys .
rev .
lett . *
116 * , 190501 ( 2016 ) .
e. cavalcanti , m. j. w. hall , h. m. wiseman , phys .
rev .
a * 87 * , 032306 ( 2013 ) . c. h. bennett , d. p. divincenzo , j. a. smolin , and w. k. wootters , phys .
rev .
a * 54 * , 3824 1996 .
k. zyczkowski , p. horodecki , a. sanpera , and m. lewenstein , phys .
rev .
a * 58 * , 883 ( 1998 ) .
g. vidal and r. f. werner , phys .
rev .
a * 65 * , 032314 ( 2002 ) .
g. vidal and r. tarrach , phys .
rev .
a * 59 * , 141 ( 1999 ) . f. g. s. l. brando , phys
. rev .
a * 72 * , 022310 ( 2005 ) .
j. eisert , f. g. s. l. brando , and k. m. r. audenaert , new j. phys .
* 9 * , 46 ( 2007 ) j. sperling and w. vogel , phys .
rev .
a * 83 * , 042315 ( 2011 ) .
s .-
s .
b. lee , and h .- s .
sim , phys .
rev .
a * 85 * , 022325 ( 2012 ) .
f. shahandeh , j. sperling , and w. vogel , phys .
rev .
a * 88 * , 062323 ( 2013 ) .
f. shahandeh , j. sperling , and w. vogel , phys .
rev .
lett . * 113 * , 260502 ( 2014 ) .
m. horodecki , p. horodecki , and r. horodecki , phys .
rev .
lett . * 80 * , 5239 ( 1998 ) .
note that our terminology here is slightly different from that of buscemi .
as we will see shortly , the pay - off made by charlie is always considered by him to be the best alice and bob could do .
j. sperling and w. vogel , phys .
rev .
a * 79 * , 052313 ( 2009 ) .
m. lewenstein , b. kraus , j. i. cirac , and p. horodecki , phys .
rev .
a * 62 * , 052310 ( 2000 ) .
d. chruciski and g. sarbicki , j. phys .
a : math . theor . * 47 * , 483001 ( 2014 ) .
f. shultz , j. math .
phys . * 57 * , 015218 ( 2016 ) .
see the supplemental material at ... for detailed proofs of eq . , criterion 1@xmath48 , the convexity of the sets @xmath63 , and theorems 1 and 2 .
e. haapasalo , t. heinosaari , and j - p .
pellonp , quantum inf . process .
* 11 * , 1751 ( 2012 ) .
v. vedral , m. b. plenio , m. a. rippin , and p. l. knight , phys .
rev .
lett . * 78 * , 2275 ( 1997 ) . c. h. bennett , d. p. divincenzo , c. a. fuchs , t. mor , e. rains , p. w. shor , j. a. smolin , and w. k. wootters , phys .
rev .
a * 59 * , 1070 ( 1999 ) .
j. sperling and w. vogel , phys .
rev .
a * 79 * , 042337 ( 2009 ) .
j. sperling and w. vogel , phys .
scr . * 83 * , 045002 ( 2011 ) .
d. rosset , c. branciard , n. gisin , and y - c .
liang , new j. phys . * 15 * , 053025 ( 2013 ) .
the authors of ref .
@xcite have shown that the correlations from entangled states in a sqwg can not be simulated classically even if the two spatially separated parties have access to locc . here , however , we have shown in theorem 1 that relaxing the losr to locc within the context of esqwgs allows the pay - off function to induce a faithful measure of entanglement .
x. yuan , q. mei , s. zhou , and x. ma , phys .
rev .
a * 93 * , 042317 ( 2016 ) .
s. l. woronowicz , rev .
mod .
phys .
* 10 * , 165 ( 1976 ) .
m. a. nielsen , phys .
rev .
lett . * 83 * , 436 ( 1999 ) .
g. vidal , phys .
rev .
lett . * 83 * , 1046 ( 1999 ) .
j. sperling , and w. vogel , phys .
rev .
lett . * 111 * , 110503 ( 2013 ) . |
the dielectric response of a collisionless plasma is an ubiquitous quantity in plasma physics and in the many - body theory of coulomb systems in general . for a fermionic system ,
it was first derived by lindhard @xcite and is closely related to the dielectric function @xmath1 in random phase approximation @xcite . in this way ,
the dielectric response is connected to collective effects such as screening , plasmons , and landau damping .
its knowledge is also important for non - ideal plasmas , since the effects of non - ideality are typically parameterized by dynamic local field corrections with respect to the ideal response @xcite .
a detailed study of the ideal dielectric response for fermionic systems has been reported by arista and brandt @xcite . here
, we retain this restriction to fermions , although a treatment for bosonic systems is possible in a similar manner .
in particular , we consider an one - component system with temperature @xmath2 and density @xmath3 interacting by the coulomb potential .
introducing the function @xmath4 as @xmath5 the real part of the dielectric function for fermions is given by @xmath6 \,\,\,,\end{aligned}\ ] ] with @xmath7 , @xmath8 , @xmath9 , and the degeneracy parameter @xmath10 . here , @xmath11 are the fermi energy @xmath12 , fermi velocity @xmath13 , and fermi wave number @xmath14 , respectively .
the bohr radius is denoed by @xmath15 , @xmath16 the chemical potential obtained from @xmath17 with the fermi function @xmath18 .
the imaginary part of the dielectric function is known analytically for all degeneracies @xmath19 } { 1\,+\,\mbox{exp}\left [ \eta-\left(u+z\right)^2/\theta \right ] } \right)\,\,\ , .
\end{aligned}\ ] ] for special cases , see the tables in ref . @xcite .
the dielectric response enters a number of important physical observables such as optical properties @xcite , the single - particle self - energy in @xmath20 approximation @xcite , and the dynamical collision frequency @xcite .
a fast and reliable computation is therefore of crucial importance to obtain these quantities .
however , a direct evaluation of the integral eq .
( [ eq : gx ] ) is often too slow in many applications .
thus , approximative analytical expressions for @xmath4 are desirable
. continued fractions appear in a variety of applications in theoretical physics @xcite .
also , they are of importance in approximation theory and are closely related to pad approximants which in turn enjoy a number of interesting applications in various fields @xcite . in particular , there is a close correspondence between t - fractions @xcite and two point pad approximations , see refs .
since an expansion of @xmath4 at @xmath21 and at @xmath22 is known , the t - fractions representation suggests itself as a powerful approximation . also , due to the compact form of the continued fraction , only a few fitting parameters have to be given .
following arista and brandt @xcite , we can consider the expansion of @xmath4 for small and large values of @xmath23 and obtain @xmath24 where @xmath25 is the fermi integral of order @xmath26 , i.e. @xmath27 and @xmath28 being an integer , @xmath29 an odd number .
note , that @xmath30 are hyper - singular integrals which have to be regularized .
details can be found in appendix [ sec : appendix_small_x ] .
furthermore , the function @xmath4 is analytically known in the non - degenerate limit @xmath31 @xmath32 and in the highly degenerate limit @xmath33 @xmath34 @xmath35 is the so - called plasma dispersion function @xcite @xmath36 and closely related to the complex error function @xcite .
@xmath37 indicates a cauchy principal value integration .
it has been shown by mccabe and murphy @xcite , that there exists a close connection between a two - point pad approximant and a t - fraction @xcite given in general by @xmath38 with @xmath39 being some constants and @xmath40 being a complex variable . here ,
we follow the notation introduced in ref . @xcite . in particular ,
once the expansions for @xmath21 and @xmath22 are known to be of the form @xmath41 and @xmath42 with expansion coefficient @xmath43 and a complex variable @xmath40 , the t - fraction representation can be generated by a q - d algorithm provided that all expansion coefficients are different from zero .
a modified q - d algorithm has been developed by de andrade et al .
@xcite to allow for zero expansion coefficients .
now , the above given expressions for @xmath4 are almost of the form necessary for applying the t - fraction scheme .
we have to take care of the infinitesimal small imaginary contribution to the expansion at @xmath44 @xmath45 by defining @xmath46 and calculating the continued fraction representation for @xmath47 . after generating the representation
, we can go back to the original @xmath4 .
as an illustrative example we study the t - fraction representation of dawson s integral , which is closely related to the plasma dispersion function of a non - degenerate plasma , cf .
( [ eq : dawson ] ) and see also ref . @xcite . for
a detailed discussion of dawson s integral and its continued fraction representation see refs .
the application of t - fractions to this problem has already been considered by j. mccabe , see ref .
we repeat a few of these results for illustration . since dawson s integral does not have the proper expansion at @xmath22 , we consider the modified function @xmath48 for a real variable @xmath23 .
this function does have the expansions @xmath49 using these expansions , one can calculate the t - fraction representation for @xmath50 with the help of the modified q - d algorithm , see app . [
sec : appendix_qd_algorithm ] . in general ,
excellent convergence is found . the relative error as a function of the variable @xmath23 for the @xmath3th approximant is shown in fig .
[ fig : dawson_error ] . for @xmath51 ,
the error is better than @xmath52 for all @xmath23 . as a second example
, we discuss the highly degenerate plasma with @xmath53 .
again , @xmath4 as given by eq . (
[ eq : gx_degenerate ] ) has to be supplemented by an imaginary part to ensure the correct asymptotic behavior @xmath54 with @xmath55 for this function , the expansions read for @xmath56 and @xmath57 @xmath58 respectively .
again , performing the modified q - d algorithm leads to partial numerators and denominators as listed in tab .
[ tab : partial_num_denom ] .
as before , the overall convergence is good , which is verified by inspection of the relative error between the continued fraction representation and the analytic expression eq .
( [ eq : f_0 ] ) , see fig .
[ fig : entart_error ] . only in the vicinity of @xmath59 ,
the agreement drops to about @xmath60 for the 10th approximant of the continued fraction . as a function of the variable @xmath23 for the @xmath3th approximant.,scaledwidth=80.0% ] ,
( [ eq : f_0 ] ) , as a function of the variable @xmath23 for the @xmath3th approximant.,scaledwidth=80.0% ] .[tab : coefficients_h ] expansion coefficients for the small @xmath23 expansion of the function @xmath4 , cf .
( [ eq : gx_small ] ) . [ cols="^,^,^,^,^,^",options="header " , ] next , we turn to the general case of arbitrary degeneracy .
typical expansion coefficients @xmath30 are compiled in tab .
[ tab : coefficients_h ] .
also , the expansion coefficients for large @xmath23 are given essentially by the fermi integrals for order @xmath61 .
they are listed in tab .
[ tab : coefficients_f ] for various values of the degeneracy parameter @xmath0 . finally , the imaginary contribution due to the modified expression in eq .
( [ eq : imaginary_g ] ) has to be determined .
it reads @xmath62 using these coefficients , the t - fraction approximation is determined using the modified q - d algorithm described in app .
[ sec : appendix_qd_algorithm ] .
the resulting partial numerators and denominators are recorded in tab .
[ tab : partial_num_denom ] .
the general form of the t - fraction considered here is @xmath63\,\,\ , .
\end{aligned}\ ] ] with @xmath28 being the imaginary unit , @xmath64 .
it has been shown that t - fractions are a powerful method to obtain reliable and compact approximative expressions for the dielectric response of ideal plasmas at arbitrary degeneracy .
in particular , a set of eight partial numerators and denominators are sufficient to obtain a satisfying accuracy for most practical applications .
fast implementations for the dynamic collision frequency and the self - energy in @xmath20 approximations are possible based on these results due to a notable acceleration compared to a direct evaluation of the integral representation .
an extension to bosonic systems appears to be straightforward and is work in progress .
this work was supported by the deutsche forschungsgemeinschaft within sfb 652 strong correlations and collective effects in radiation fields. the author would like to thank carsten fortmann and mathias winkel for testing the numerical pay - off of the final fit formula . also , he would like to thank john mccabe for pointing out ref .
in order to regularize the hyper - singular integrals in eq .
( [ eq : gx_small ] ) , we consider an expansion of the fermi function around zero @xmath65 the calculation has been done using the computer algebra software mathematica @xcite . in detail , we obtain for the integrands @xmath66 \,\,\,,\\
h_5(y ) & = & \frac { { \rm e}^{\eta } } { 6 \left(1+{\rm e}^{\eta } \right)^4 \left ( { \rm e}^{\eta}\,+\ , { \rm e}^{y^2/\theta } \right)}\ , \left [ -6 \left ( 1 + { \rm e}^{\eta } \right)^3 \left ( -1 + { \rm e}^{y^2/\theta } \right ) \right . \nonumber \\ & & \left .
+ 6 \left ( 1 + { \rm e}^{\eta } \right)^2 \left ( { \rm e}^{\eta}+{\rm e}^{y^2/\theta } \right ) y^2/\theta \,+\ , 3 \left ( -1+{\rm e}^{2 \eta } \right ) \left ( { \rm e}^{\eta}+{\rm e}^{y^2/\theta } \right ) y^4/\theta^2\,+\ , \left ( 1 - 4 { \rm e}^{\eta}+{\rm e}^{2 \eta } \right ) y^6/\theta^3 \right ] \,\,\,,\end{aligned}\ ] ] with @xmath67 . also , for evaluating the integrals involved , it is convenient to determine the expansion of each integrand for small @xmath23 by reading off the corresponding terms in eq .
( [ eq : fermi_function_zero ] ) .
note , that the limits of @xmath30 in the non - degenerate and highly degenerate case are analytically known from the corresponding expansions of eq .
( [ eq : gx_non_degenerate ] ) and eq .
( [ eq : gx_degenerate ] ) , respectively . note also , that the values of @xmath68 and @xmath69 are given by arista and brandt @xcite , while all higher expressions are given here for the first time .
the fermi integrals of order @xmath61 involved in the large @xmath23 expansion have been determined by mathematica @xcite exploiting the relation of the fermi integral to the polylogarithmic function @xcite @xmath70 the fermi integrals @xmath71 can be easily obtained with the interpolation formulas given by antia @xcite .
we summarize the modified q - d algorithm introduced by de andrade et al . @xcite .
we start with the series expansions eq .
( [ eq : expansion_at_zero ] ) and eq .
( [ eq : expansion_at_infinity ] ) and allow for zero coefficients @xmath43 .
next , an auxiliary expansion is introduced by subtracting the expansion of the function @xmath72 with a free parameter @xmath73 from the above given expansion generating a new set of coefficients @xmath74 . for a convenient choice of @xmath73 , all
@xmath74 will be non - zero and the ordinary q - d algorithm can be applied . the continued fraction representation @xmath75 having the original series expansions eq .
( [ eq : expansion_at_zero ] ) and eq .
( [ eq : expansion_at_infinity ] ) can be reconstructed from the coefficients @xmath74 by the following algorithm : * determine @xmath74 from a given set of @xmath43 and a convenient choice of @xmath73 as @xmath76 * perform the q - d algorithm @xmath77 with the initialization @xmath78 and @xmath79 , @xmath80 .
* determine for @xmath81 the following quantities leading to @xmath82 and @xmath83 , @xmath84^{-1}\,\,\,,\nonumber \\
\alpha_{i+1 } & = & \frac{w_{i+1}}{w_i } \,\,\,,\nonumber \\
\beta_{i+1 } & = & -\frac{b_{i+1,0}}{b_{i,0}}\,\,\,,\nonumber \end{aligned}\ ] ] with the initialization @xmath85
there are several algorithms to numerically calculate the value of a continued fraction . here
, we use the backward recurrence algorithm which is known for its numerical stability @xcite . given @xmath3 partial numerators @xmath86 and partial denominators @xmath87 , we determine an approximation @xmath88 to the continued fraction by setting @xmath89 to a small , non - zero number and perform the iteration @xmath90 for @xmath91 .
see e.g. h. mori , prog .
. phys . * 34 * , 399 ( 1965 ) ; m.h .
lee , phys .
rev . lett . * 49 * , 1072 ( 1982 ) ; j. horek and t. sasakawa , phys .
c * 32 * , 70 ( 1985 ) ; d.e .
neuenschwander , am . j. phys .
* 62 * , 871 ( 1994 ) . | we present an approximation scheme for the dielectric response of thermal collisionless plasmas at arbitrary degeneracy .
a t - fraction representation is obtained from the known expansions of the real part of the dielectric function for small and large arguments .
the partial numerators and denominators of the continued fraction are generated by a modified q - d algorithm . for several typical values of the degeneracy parameter @xmath0 , extensive tables for the expansion coefficients and the partial numerators and denominations
are given allowing for an easy implementation of the fitting function . also , an error analysis is performed . |
during the past 20 years a number of experiments with solar , atmospheric , reactor and baseline neutrinos have shown that neutrinos have masses and mixings @xcite .
these experiments provide a framework with @xmath1 that fits remarkably well most of the data . from a model - building point of view the importance of this discovery
can not be overstated .
we should , however , keep in mind that * there are some very basic questions with no answer yet .
we do not know , in particular , if these masses are purely electroweak ( ew ) ( just like the electron mass ) or if neutrinos are majorana fermions and their mass is revealing a new scale in particle physics ( like radiactivity reveals the ew scale ) : @xmath2 * there has been a _ persistent _ anomaly in several experiments with neutrino beams from particle accelerators . namely , lsnd @xcite and miniboone @xcite have observed an excess of @xmath3 events with an electron in the final state per each 1000 @xmath4 charged - current ( cc ) interactions .
the interpretation of these events in terms of @xmath5 oscillations would force the addition of two sterile neutrinos of mass around 1 ev . here
we will go in a completely different direction : we will introduce a sterile mode , but much heavier and unstable . let us very briefly review these experimental anomalies and the heavy neutrino proposal .
in the mid 90 s lsnd @xcite observed an excess of electron - like events that were interpreted as @xmath6 oscillations followed by a cc interaction giving an electron and a free neutron , @xmath7 .
in particular , they could see the cherenkov light of the electron plus a 2.2 mev photon coming from the capture of the neutron to form a deuteron . given the lsnd fluxes @xcite , @xmath8 the initial antineutrinos ( in fig .
1left ) were coming from the decay - at - rest of positive muons , whereas the appearence of oscillations would require the existence of a @xmath9 ev sterile neutrino slightly mixed with the muon flavor . instead , gninenko proposed three years ago a very different interpretation , where the responsible for the signal would be the muon neutrinos from the decay - in - flight of pions ( in fig .
1right ) , with energies of up to 200 mev .
gninenko @xcite showed that these events could be caused by an interaction mediated by a z boson ( see fig .
2 ) that changes the initial @xmath4 into a sterile neutrino @xmath10 of mass around 50 mev plus a free neutron .
the neutrino should then decay through an electromagnetic ( em ) interaction into a light neutrino plus a photon ; the photon would finally convert into an electron pair inside lsnd giving the same cherenkov ring as a single electron .
in other words , the excess of electron - like events at lsnd would actually be due to photon events .
for this to work the mixing of @xmath10 with the muon flavor must be relatively large , @xmath11@xmath12 , and the em dipole transition @xmath13 between @xmath10 and a light neutrino must be such that the lifetime @xmath14 is shorter than @xmath15 s. the advantage of gninenko s interpretation versus the oscillation hypothesis is that it provides an explanation also for karmen @xcite , that in 2002 tried to reproduce lsnd using a similar technique and saw nothing .
the initial proton beam and the distances in both experiments were identical , but the larger angle of the neutrino beam selected at karmen ( @xmath16 , versus just @xmath17 at lsnd ) implied a very different neutrino spectrum : karmen eliminated the neutrinos from in - flight decays , keeping only the less energetic antineutrinos from @xmath18 decays at rest . whereas oscillations should occur in both experiments , the production of a 4080 mev neutrino would be above threshold at karmen .
this mass and mixing of the the sterile neutrino may sound _ unlikely _ ; is there really room for such a particle ?
gninenko @xcite argues that yes , that @xmath10 could appear instead of a regular @xmath19 in up to 1% of all muon and kaon decays .
the bounds from current data depend on whether its lifetime is longer or shorter than @xmath20 s , which corresponds to a decay length of 30 cm .
if the heavy neutrino is longer lived then it tends to be invisible ( _ i.e. _ , it does not decay inside the detector ) and the kinematic changes that it introduces in kaon and muon decays are not significant . if @xmath10 decays faster ( _ e.g. _ , @xmath21 s ) then the @xmath22 or @xmath23 decays will appear with an extra photon .
it turns out , however , that this is also difficult to see due to background processes with photons , @xmath24 although the limits on a 50 mev neutrino @xmath10 with no decay modes into charged particles are weaker than one may think , a recent analysis of the photon distribution in kaon decays by istra+ @xcite disfavors lifetimes @xmath25 s. these shorter lifetimes , however , are necessary in the original gninenko model in order to also explain the low - energy miniboone anomaly ( see below ) .
moreover , it was noticed @xcite that gninenko s model is also in conflict with observations of radiative muon capture at triumf : the shorter - lived heavy neutrino is produced and decays too often there , implying three times more events than observed .
we propose @xcite a variation of gninenko s model with three basic ingredients . *
we consider a longer - lived @xmath10 ( _ e.g. _ , @xmath26 s ) . the target volume at triumf is smaller than @xmath27 , and the number of events with @xmath10 decaying inside the detector goes as @xmath28 when the lifetime grows .
the bounds at istra+ @xcite also weaken with an increased @xmath14 .
* we include the em production of @xmath10 ( see fig . 3 ) .
+ + the same em operator that lets @xmath10 decay into @xmath29 implies , when the light neutrino @xmath30 is a @xmath4 , a production channel at miniboone that was overlooked by gninenko . the extra @xmath10 produced through photon exchange will dominate and compensate the reduction due to the increased lifetime at miniboone , as some @xmath10 will decay outside the detector .
if we write @xmath31 a lifetime @xmath26 s implies @xmath32 , whereas the anomaly at miniboone will require @xmath33 .
this means that only 1% of @xmath10 decays go into a muon neutrino . *
finally , we impose @xcite that the @xmath10 is a dirac particle : neutrino plus antineutrino of both chiralities .
this is important because it will reduce the energy of the events at miniboone .
basically , the initial @xmath4 of negative helicity becomes , after the em transition , a heavy neutrino with the spin pointing ( @xmath34 of the times ) forward .
it turns out that the decay is then not isotropic and the photon is preferably emited backwards , with less energy than if it were emited forward or isotropically . the dirac nature and the increased lifetime that we have assumed
are essential in order to generate an excess only at low energies ( higher - energy neutrinos tend to decay outside the detector ) . just like lsnd , miniboone is unable to distinguish an electron from a photon converted into an @xmath35 pair .
our estimate @xcite of the number of photon events from @xmath10 decays is given in fig . 4 for @xmath36 mev , @xmath37 , @xmath26 s and @xmath33 .
on the left , we plot pot ) @xcite.,title="fig:"]@xmath8 pot ) @xcite.,title="fig : " ] with dashes the distribution of heavy neutrinos produced inside the detector and with long dashes the distribution of neutrinos that decay inside the detector .
notice that these two lines are closer together at low energies and separate as the energy grows .
the dotted line is the energy distribution of the photons ( emitted preferably backwards ) from the decaying @xmath10 , and the final solid line is the energy assigned to an initial @xmath38 that after a quasielastic collision becomes an electron with the same energy as the photon ( _ i.e. _ , the reconstruction of the photon events as @xmath38 events ) . on the right
, we plot the excess observed by miniboone with error bars ( zero would be consistency with the background ) and in dashes what could be expected from @xmath5 oscillations due to the 1 ev sterile neutrino that explains lsnd .
the initial publications by miniboone emphasized that their results disprove the lsnd oscillation hypothesis , although in the later ones they fail to explain the low - energy excess observed in the data @xcite .
our heavy neutrino hypothesis gives a remarkable fit to the data , with similar results for the miniboone excess in the antineutrino mode in fig .
pot ) @xcite .
] three coments are here in order .
* the magnetic dipole transitions that we assume can be generated @xcite in left - right symmetric completions of the model at the tev scale .
in particular , we just need that both the left and the righ - handed components of @xmath10 define @xmath39 doublets together with a charged lepton , and that the breaking of this gauge symmetry gives large mass to this charged lepton while keeping @xmath10 light . *
an analysis of t2k ( where most of the initial muon neutrinos have oscillated into the tau flavor at super - kamiokande ) data @xcite shows that the decay @xmath40 has a @xmath41 branching ratio @xcite .
therefore , there must be a second sterile neutrino @xmath42 lighter , longer lived and less mixed with the standard flavors than @xmath10 that accounts for @xmath43 of the decays : @xmath44 .
* microboone @xcite will investigate whether the low - energy excess at miniboone is caused by electron or by photon events .
it has been argued @xcite that the standard background producing a photon ( in fig .
6 ) may be underestimated in current simulations .
hopefully , there are observables that may distinguish this background from the @xmath45 hypothesis ( in particular , the longitudinal event distribution inside the detector is flat for the background but proportional to @xmath46 in @xmath10 events ) , so the miniboone puzzle should be settled during the next year .
the lsnd and the miniboone anomalies consist of an excess of events with an electron in the final state . the question now would be , could one expect ( at larger energies and distances ) a similar excess in neutrino telescopes
? would it be detectable ? a telescope like icecube observes two types of events : muon tracks and point - like energy depositions ( hadronic and em showers develop in @xmath47 m of ice ) @xcite .
the direction of the first ones can be determined very precisely , while in cascade events the error is around @xmath48 .
upgoing muons crossing the telescope come necesarily from @xmath4 interactions ( most of these tracks do not start inside the detector ) , whereas contained cascades from any direction correspond to @xmath49 cc interactions or @xmath50 neutral current ( nc ) interactions . at energies above 100 gev the flux of atmospheric muons and neutrinos
is dominated by kaon decays @xcite .
since our @xmath10 appears in up to @xmath51 of these decays instead of a muon neutrino , it will be abundant in the atmosphere .
its decay length ( @xmath52 ) at tev energies becomes larger than 10 km , so an atmospheric @xmath10 could reach the center of icecube and decay there .
the resulting photon would produce a shower - like event similar to a cc @xmath38 event or to an inelastic nc collision .
cc interactions , @xmath53 nc interactions , and @xmath54 decays @xcite .
neutrinos from charmed - hadron decays , dominant at @xmath55 gev , have not been included . in dashes , the energy of the parent neutrino ( left ) .
typical @xmath10 event at icecube ( right).,title="fig : " ] @xmath56 cc interactions , @xmath53 nc interactions , and @xmath54 decays @xcite .
neutrinos from charmed - hadron decays , dominant at @xmath55 gev , have not been included . in dashes , the energy of the parent neutrino ( left ) .
typical @xmath10 event at icecube ( right).,title="fig : " ] we have investigated these processes @xcite and have obtained that @xmath10 decays may change substantially the number of contained cascade - like events between 100 gev ( at lower energies the heavy neutrino can not reach the telescope ) and 1 pev ( at very high energies the decay length grows and the probability to decay inside the telescope becomes negligible ) . in fig .
7 we have used the z - moment method to estimate their frequency . generically , the excess of tev contained showers introduced by the heavy neutrino has two basic features : * the excess would only appear in down - going or near - horizontal events ( there are no @xmath10 upgoing events ) * some of these events ( specially the ones from small zenith angles ) would be contaminated with muons . since
the heavy neutrino is produced in the atmosphere together with a very energetic muon , there will be an excess of _ muon plus contained cascade _ events .
we would like to conclude by noticing two recent observations published by icecube .
the first one is the sample of 28 events above 30 tev described in @xcite , which are indeed very interesting data . if we discount 4 muon - like tracks that are consistent with atmospheric muons entering the telescope , we are left with 24 events that should correspond to neutrino interactions .
three of these events are muons and 21 are cascades .
the 3/21 ratio suggests that these events are not atmospheric in origin : below 100 tev the lepton fluxes are still dominated by kaon and pion decays , which produce more muons than electrons ( atmospheric taus are irrelevant ) .
there are just too many cascades relative to muons to be consistent with standard atmospheric - neutrino interactions .
the energy distribution of these events is also much flatter ( @xmath57 ) than the one from atmospheric neutrinos ( @xmath58 ) .
therefore , it is very likely that this is the first observation of cosmic neutrinos . on the other hand , it is _ intriguing _ that all the cascade events of energy below 100 tev are down - going or near - horizontal , since at these energies the earth is not fully opaque ( from zenith angles between @xmath16 and @xmath59 ) to neutrinos , and the events observed do not show preference for the galactic disk neither . in our opinion ,
the data implies an excess of down - going cascade events relative to up - going cascades if one assumes a cosmic origin or relative to muon events if one assumes a standard atmospheric origin , and it could be interesting to run a simulation of the atmospheric @xmath10 hypothesis .
@xmath60 a second interesting observation by icecube @xcite has been obtained in the search for high light density events related to magnetic monopoles and other exotics .
although they did not find any monopole candidate , but their cuts selected two events ( in fig .
8) for an estimated background of 0.14 events .
it seems to us that these two events are not regular contained showers nor muon bundles , but that they could be interpreted as a cascade event _ inside _ a muon bundle .
such topology would be difficult to explain with standard model particles , but it is exactly what the heavy neutrino hypothesis would suggest .
neutrino physics has progressed a great deal during the past 20 years , but a few basic questions and some persistent anomalies should still be clarified .
we have focused on miniboone and icecube , that have provided very interesting although still incomplete results .
hopefully , in the near future microboone and an increased statistics , respectively , will establish the reach of their current results .
we have argued that there is room in neutrino physics for _ serious _ departures from the three - flavor picture , in particular , we have discussed a 50 mev sterile mode that could be a solution to the baseline anomalies .
we find the possible incidence of such neutrino in completely different experiments , like neutrino telescopes , very interesting .
the heavy neutrino could be produced in the atmosphere and decay inside the telescope , introducing an excess of contained showers from downgoing and near horizontal directions , some of them contaminated by atmospheric muons .
there seems to be room in current icecube data for this type of events .
telescopes have the advantage that they are basically background free experients : each individual event must be explained in terms of neutrino interactions or atmospheric muons . in this sense ,
the 50 mev neutrino is a possibility that that may be worth exploring .
this work has been partially supported by micinn of spain ( fpa2010 - 16802 and consolider - ingenio * multidark * csd2009 - 00064 ) , and by junta de andaluca ( fqm 101 and fqm 3048 ) .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * ( 2012 ) 010001 . c. athanassopoulos
_ et al . _
[ lsnd collaboration ] , phys .
* 77 * ( 1996 ) 3082 ; [ arxiv : nucl - ex/9605003 ] ; c. athanassopoulos _ et al . _
[ lsnd collaboration ] , phys .
c * 54 * ( 1996 ) 2685 ; [ arxiv : nucl - ex/9605001 ] ; a. aguilar _ et al . _
[ lsnd collaboration ] , phys .
d * 64 * ( 2001 ) 112007 .
[ arxiv : hep - ex/0104049 ] .
a. a. aguilar - arevalo _ et al . _
[ the miniboone collaboration ] , phys .
* 98 * ( 2007 ) 231801 ; [ arxiv:0704.1500 [ hep - ex ] ] .
a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , phys .
* 102 * ( 2009 ) 101802 .
[ arxiv:0812.2243 [ hep - ex ] ] .
a. a. aguilar - arevalo _ et al .
_ [ miniboone collaboration ] , phys .
* 103 * ( 2009 ) 111801 [ arxiv:0904.1958 [ hep - ex ] ] ; a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , phys . rev .
* 105 * ( 2010 ) 181801 [ arxiv:1007.1150 [ hep - ex ] ] .
s. n. gninenko , phys .
d * 83 * ( 2011 ) 015015 .
[ arxiv:1009.5536 [ hep - ph ] ] .
b. armbruster _ et al .
_ [ karmen collaboration ] , phys .
d * 65 * ( 2002 ) 112001 .
[ arxiv : hep - ex/0203021 ] .
v. a. duk _ et al . _ [ istra+ collaboration ] , phys . lett .
b * 710 * ( 2012 ) 307 [ arxiv:1110.1610 [ hep - ex ] ] .
d. mckeen and m. pospelov , phys .
d * 82 * ( 2010 ) 113018 .
[ arxiv:1011.3046 [ hep - ph ] ] .
m. masip , p. masjuan and d. meloni , jhep * 1301 * ( 2013 ) 106 [ arxiv:1210.1519 [ hep - ph ] ] .
a. radionov , phys .
d * 88 * ( 2013 ) 015016 [ arxiv:1303.4587 [ hep - ph ] ] .
a. bueno , m. masip , p. snchez - lucas and n. setzer , phys
. rev .
d * 88 * ( 2013 ) 073010 [ arxiv:1308.0011 [ hep - ph ] ] .
k. abe _ et al . _
[ t2k collaboration ] , phys .
* 107 * ( 2011 )
041801 [ arxiv:1106.2822 [ hep - ex ] ] . b. j pjones , pos eps * -hep2011 * ( 2011 ) 436 [ j. phys .
* 408 * ( 2013 ) 012028 ] [ arxiv:1110.1678 [ physics.ins-det ] ] .
r. j. hill , phys .
d * 84 * ( 2011 ) 017501 [ arxiv:1002.4215 [ hep - ph ] ] .
f. halzen and s. r. klein , rev .
instrum .
* 81 * ( 2010 ) 081101 .
p. lipari , astropart .
* 1 * ( 1993 ) 195 ; j. i. illana , p. lipari , m. masip and d. meloni , astropart . phys .
* 34 * ( 2011 ) 663 .
m. masip and p. masjuan , phys .
d * 83 * ( 2011 ) 091301 [ arxiv:1103.0689 [ hep - ph ] ] .
m. g. aartsen _ et al .
_ [ icecube collaboration ] , science * 342 * ( 2013 ) 6161 , 947 [ arxiv:1311.5238 [ astro-ph.he ] ]
. j. posselt , _
search for relativistic magnetic monopoles with the icecube 40-string detector _ ,
dissertation , wuppertal univ .
, october 2013 ( http://elpub.bib.uni-wuppertal.de/servlets/documentservlet?id=3814 ) ; r. abbasi _ et al . _
[ icecube collaboration ] , phys .
d * 87 * ( 2013 ) 022001 [ arxiv:1208.4861 [ astro-ph.he ] ] . | we discuss the possibility that a sterile neutrino of mass around @xmath0 mev slightly mixed with the muon flavor may be the origin of the miniboone anomaly .
we show that its production in the atmosphere in a fraction of kaon decays would imply an excess of contained showers at icecube from down - going and near - horizontal directions .
* a sterile neutrino at miniboone and icecube * manuel masip _ cafpe and departamento de fsica terica y del cosmos _ + _ universidad de granada , 18071 granada , spain _ + ` [email protected] ` |
in the linear response theory an electron spin resonance ( esr ) spectrum is given by the fourier transform of the relaxation function @xcite @xmath109 where @xmath110 are transverse magnetization operators and @xmath111 denotes canonical averaging .
line position ( @xmath29-value ) and line width of the spectrum contain important information on local anisotropies @xcite .
the principal axes of the @xmath29-tensor are set by a local symmetry of the crystal field . in bnfso one of the principal axes ( labeled @xmath14 )
is parallel to a two - fold rotational axis passing through the center of each feo@xmath112 tetrahedron , while the other two ( labeled @xmath38 and @xmath37 ) are parallel with the two normals to the o@xmath113-fe - o@xmath113 planes , @xmath114 labeling two pairs of oxygens at a distance of 1.909 and 1.856 from fe , respectively . for site 2 ( see fig . [ fig1s ] and fig . 1 of the main text )
the @xmath14 axis corresponds to the crystallographic @xmath36 direction , while @xmath37 and @xmath38 axes are tilted by @xmath39 and 120@xmath40 from the @xmath115 axis in the crystallographic @xmath42 plane .
the three magnetically non - equivalent fe sites on each triangle yield a single esr line due to exchange narrowing .
therefore , an axially symmetric effective @xmath29-tensor is obtained after averaging over the three sites .
we find tetrahedra in bnfso in ( a ) the @xmath6 and ( b ) the @xmath42 crystallographic plane .
the two crystallographically non - equivalent oxygen sites are labeled as o2 and o3 .
the principal axes of the @xmath29-factor on the fe site 2 are denoted by @xmath37 , @xmath38 and @xmath14 ; the latter is parallel to a local two - fold rotational axis . ]
@xmath116 and @xmath117 , where @xmath118 denotes the polar angle between the crystallographic @xmath10 axis and the applied magnetic field . when the magnetic anisotropy term @xmath48 of the spin hamiltonian [ see eq .
( 1 ) in the main text ] is inferior to the isotropic exchange term @xmath46 and/or the zeeman term @xmath47 it can be treated as a perturbation . in cases of moderate line widths ( @xmath119 )
the well - established kubo - tomita approach applies for @xmath120 .
the esr absorption spectrum then has a lorentzian line shape and a full - width - half - maximum ( fwhm ) line width , given at infinite temperatures as @xmath121 the second and the fourth moment of the resonance spectrum are given by @xcite @xmath122 [ s^{-},\mathcal{h}^{\prime } ] \right\rangle } { \left\langle s^{+}s^{-}\right\rangle } \ ; , \\\label{eq6}m_{4}&=&\frac{\left\langle \left [ \mathcal{h}-\mathcal{h}_{z},\left [ \mathcal{h}^{\prime } , s^{+}\right ] \right ] [ \mathcal{h}-\mathcal{h}_{z},\left [ \mathcal{h}^{\prime } , s^{-}\right ] ] \right\rangle } { \left\langle s^{+}s^{-}\right\rangle } \ ; , \end{aligned}\ ] ] where @xmath123 $ ] denotes a commutator .
the moments given by the spin hamiltonian are always finite , while those of the lorentzian line shape are infinite .
therefore , the true experimental line shape deviates from the lorentzian shape and a modified line shape has to be employed .
typically , this deviation becomes notable only in extreme wings .
the criterion for the selection of a suitable line shape should be the agreement between the theoretical exchange field @xmath124 , determined by the spin hamiltonian , and the one given by a particular line shape @xcite .
we find that in bnfso the approximation where the lorentzian line shape is multiplied with the broad gaussian , @xmath125 $ ] , yields an excellent agreement of the exchange fields .
this situation is usual in systems with strong exchange , where the gaussian time decay of spin correlations can be assumed .
this approximation yields the @xmath126 prefactor in eq ( [ eq4 ] ) .
although tedious , calculations of the two moments lead to analytic expressions of the esr line width .
we calculated it separately for the case of the single - ion ( si ) anisotropy , taking into account spin operators up to the fourth order [ see eq .
( 2 ) in the main text ] , and for the dzyaloshinsky - moriya ( dm ) interaction [ see eq .
( 3 ) in the main text ] . for the spin lattice shown in fig . 1 of the main text we derived the following expressions for the esr line width @xmath127 ^ 3}{20 j^2_{\rm si } \left [ 189 d^2 + 720a^2 + 175 f^2 + 525 af - \frac{7}{5 } \left ( 27d^2 + 25f^2 + 75 af \right ) { \rm cos}\;2 \theta \right]}},\\ \label{eq6}\delta b_\infty^{\rm dm } & = & \sqrt{2\pi } \frac { k_{b}}{g \mu _ { b } } \sqrt{\frac{105\left [ 5d_{ab}^2 + 6d_c^2 + \left(d_{ab}^2 - 2d_c^2 \right ) { \rm cos}\;2 \theta \right]^3}{32\left [ 35 j_{\rm dm}^2 d_{ab}^2 + 6 j'^2_{\rm dm } d_c^2 + \left ( 2 j'^2_{\rm dm } d_c^2 - 7 j_{\rm dm}^2 d_{ab}^2 \right ) { \rm cos}\;2 \theta \right]}}.\end{aligned}\ ] ] here @xmath128 , @xmath129 and @xmath130 .
the ratios of the anisotropy parameters within each model are set by the temperature independent experimental line - width ratio @xmath67 . in case of the si model ,
our calculations predict @xmath131 for the @xmath36 term and @xmath132 for the @xmath65 term . due to its cubic symmetry
the former term does not yield any anisotropy , however in still gives a finite esr line width , because it does not commute with @xmath47 . in case of the dm model
, the @xmath23 anisotropy leads to @xmath132 and the @xmath22 anisotropy to @xmath133 . in both models
, a single dominant parameter thus fails to account for the experiment .
a mean - field approach is used to calculate the antiferromagnetic resonance ( afmr ) modes . in this approximation , sublattices are introduced instead of individually treatment of the spins .
the corresponding magnetizations are given by @xmath134 , where @xmath135 is the number of fe@xmath4 magnetic ions in the @xmath113-th sublattice .
the afmr modes are calculated within the spin hamiltonian given by eqs .
( 1)-(3 ) in the main text . applying the mean - field approximation @xcite we can rewrite this hamiltonian into a magnetic free energy @xmath136 per fe site
, the magnetic ground state is obtained from minimization of the free energy .
this then yields effective magnetic fields acting on individual sublattice magnetization @xmath141 , which is introduced into the equations of motion for all sublattices @xmath142 to make the problem tractable , harmonic oscillations are assumed , @xmath143 , and the equations are linearized .
the problem is solved numerically the parameters @xmath144 , @xmath145 , @xmath146 , @xmath147 and @xmath148 are adjusted to fit the field- and the angular dependence of the antiferromagnetic resonance modes . in this mean - field approach
all information about the interactions between spins within a particular sublattice is lost .
therefore , a meaningful choice of sublattices is of great importance .
the ferro - triangular magnetic order allows a restriction to six sublattices in each triangular plane , in order to account for the lowest lying magnon modes ( see fig . 1 in the main text for the site numbering ) .
thus both the intratriangle exchange @xmath52 and the intertriangle exchange @xmath149 are taken into account .
the effect of the latter is however simplified , since all the 1-3 triangles neighboring the basic 1 - 3 triangle are assumed the same , i.e. , effectively , only one 1 - 3 and one 1-3 triangle with the inter - triangle exchange @xmath149 are needed .
finally , to take into the account also the interlayer interactions ( @xmath150 , @xmath151 and @xmath152 ) , we stacked 14 pairs of 1 - 3 and 1-3 triangles and used periodic boundary conditions , thus allowing for the experimental pitch of 2@xmath153/7 along the @xmath10-axis .
the resulting magnetic ground state is a helix with the ordering vector @xmath154 .
the plane of the helix will , however , be randomly oriented if @xmath48 is set to zero , thus resulting in a sinusoidal tilt ( along @xmath10 ) of spins on a given site from the @xmath10 axis .
this remains the case if a finite in - plane dm component @xmath23 is present .
adding a finite out - of - plane dm component @xmath22 aligns the helix plane with the crystallographic @xmath6 plane .
the triangular chirality is found anti - parallel to the @xmath155 vector , thus minimizing the dm interaction energy .
the helicity is always anti - parallel to the triangular chirality due to the left - handedness of the crystal structure imposing @xmath156 . in the si model , on the other hand , @xmath157 leaves the helix plane randomly oriented , while @xmath158 randomly imposes helix planes with their normals perpendicular to the @xmath10 axis . on the other hand ,
spins are bound to @xmath6 planes for @xmath159 and/or @xmath160 , which , however , always select the state with positive triangular chirality ( and negative helicity ) .
the afmr modes that suit the experimental observations best are plotted in fig . 4 of the main text for both anisotropy models .
in zero field , the three lowest modes define the lowest two branches s@xmath94 and s@xmath70 , which are characterized by @xmath161 and a linear motion of spins .
the crucial difference between the two anisotropy models is that in the dm model the s@xmath94 mode is the gapless goldstone mode and the two s@xmath70 modes occur at ( @xmath162 ghz ) . in the si model ,
on the other hand , all three modes are gapped a two - fold degenerate mode is found at @xmath163 ghz and the third mode at @xmath164 ghz .
the small splitting is due to the fact that the si anisotropy in bnfso is multi - axial ( see fig . 1 of the main text ) .
if the magnetic helix was incommensurate a gapless goldstone mode would exist also in the si model , as there would be no energy cost in rotating the helix around its axis .
the gapless mode is , however , absent for a commensurate helix .
experiments imply the latter .
namely , no changes in the propagation vector , which is a rational fraction of @xmath10 with small numerator and denominator co - primes , were detected in the temperature range of the ordered phase @xcite .
this is not in line with the incommensurability , arising from a delicate balance between the exchange parameters @xmath165 , then subject to strong sensitivity on temperature @xcite .
this then justifies the use of the finite lattice of 14 sites along the @xmath10 axis in our modeling .
the zero - field gap to the lowest gapped mode opens in a square - root fashion ( fig .
[ fig2s ] ) for both the dm and the si model .
we have calculated the dependence of the gap also for symmetric anisotropic exchange ( ae ) @xmath166 where the sum runs over the nearest - neighbor fe pairs , and @xmath167 either corresponds to crystallographic @xmath10 direction or a general local direction with respect to the @xmath140 bond within the @xmath6 plane .
similarly to other two models , in the ae model the square - root dependence of the gap is observed .
moreover , all models predict a very similar size of the anisotropy @xmath168 , which is required to explain the gap .
000 r. kubo and k. tomita , j. phys .
jpn * 9 * , 888 ( 1954 ) .
a. abragam and b. bleaney , _ electron paramagnetic resonance of transition ions _
( clarendon , oxford , 1970 ) .
t. g. castner , jr . and m. s. seehra , phys . rev .
b * 4 * , 38 ( 1971 ) .
h. rohrer and h. thomas , _
_ * 40 * , 1025 ( 1969 ) .
k. marty , phd thesis , university of grenoble ( 2008 ) ; http://tel.archives-ouvertes.fr/tel-00351575/en/ m. pregelj , _ et al . _ ,
b * 82 * , 144438 ( 2010 ) . | we present electron spin resonance ( esr ) investigation of the acentric ba@xmath0nbfe@xmath0si@xmath1o@xmath2 , featuring a unique single - domain double - chiral magnetic ground state .
combining simulations of the esr line - width anisotropy and the antiferromagnetic - resonance modes allows us to single out the dzyaloshinsky - moriya ( dm ) interaction as the leading magnetic anisotropy term .
we demonstrate that the rather minute out - of - plane dm component @xmath3 mk is responsible for selecting a unique ground state , which endures thermal fluctuations up to astonishingly high temperatures .
spin - orbit coupling generally has no relevant effect in ordinary magnets , except to pin the orientation of spins with respect to the background lattice in magnetically ordered phases , singled out by the isotropic heisenberg exchange interactions .
an exception is provided on geometrically frustrated lattices where the isotropic interactions alone are often unable to raise the macroscopic degeneracy of a ground state ( gs ) @xcite , leading to unconventional cooperative electronic states @xcite .
then , the magnetic behavior might be exclusively driven by minute perturbing terms in the form of anisotropic interactions emanating from the spin - orbit coupling .
for instance , the antisymmetric dzyaloshinsky - moriya ( dm ) exchange interaction @xcite can induce order on a kagom lattice in the classical limit @xcite and can lead to quantum criticality in the quantum limit @xcite .
furthermore , dm is known to alter phase diagrams of frustrated ladders @xcite and triangular lattices @xcite .
it is also responsible for spin chirality @xcite that can be long - ranged even in the absence of a classical magnetic order @xcite .
spin chirality is one of the key concepts in the physics of strongly correlated electrons , as it is related to various intriguing phenomena , like the realization of spin liquids @xcite , magnetic - order induced ferroelectricity @xcite , anomalous hall effect @xcite and possibly high - temperature superconductivity @xcite . in this context ,
ba@xmath0nbfe@xmath0si@xmath1o@xmath2 ( bnfso ) is extremely appealing due to its remarkable magnetic properties .
it crystallizes in a non - centrosymmetric trigonal unit cell ( p321 symmetry ) .
the fe@xmath4 ( @xmath5 ) spins reside on vertices of equilateral triangles arranged into a two dimensional ( 2d ) triangular lattice ( crystallographic @xmath6 planes in fig . [ fig1 ] ) @xcite .
the dominant exchange interactions are antiferromagnetic the curie - weiss temperature is @xmath7 k @xcite and thus frustrated .
nevertheless , a long - range - ordered ( lro ) state is realized below @xmath8 k , characterized by a @xmath9 spin arrangement on each triangle .
the moments are bound to the @xmath6 planes and form a magnetic helix along the crystallographic @xmath10 axis , corresponding to the magnetic propagation vector @xmath11 , @xmath12 @xcite . in bnfso .
( a ) , ( c ) exchange interactions @xmath13 and magnetic anisotropies of the dzyaloshinsky - moriya ( * d * ) and single - ion ( * d * , * f * ) type .
the latter vectors are parallel to a local two - fold rotational axis defining the @xmath14 axis .
( b ) two possible double chiral ground states , with @xmath15 and @xmath16 denoting triangular chirality and helicity ( with @xmath17 pitch along @xmath10 axis ) , respectively . ]
the system has been drawing considerable attention because of its magnetoelectric and multiferroic properties @xcite .
however , its uniqueness stems from its distinctive magnetic order and magnetic excitations .
the lro magnetic gs is doubly chiral [ fig .
[ fig1](b ) ] and single - domain a single triangular vector chirality @xmath15 together with a single helicity @xmath18 is chosen from four possible states in a structurally chiral crystal which is believed to be crucial for its multiferroic properties @xcite . moreover , below @xmath19 one of the two magnetic excitation branches emerging from the magnetic satellites is completely chiral over the whole energy spectrum , implying an unprecedented absence of chirality mixing of the spin dynamics @xcite .
the chiral correlations remain present far above @xmath19 @xcite .
the helicity of the gs can be rationalized within the isotropic heisenberg model if three competing interplane interactions [ @xmath20 in fig .
[ fig1](c ) ] are considered .
however , the isotropic exchange still allows @xcite for two @xmath21 pairs [ fig .
[ fig1](b ) ] , and fails to justify the selection of the @xmath6 easy plane .
thus , the fundamental question about the mechanism , responsible for the experimentally observed unique single - domain chiral gs @xcite in bnfso , arises .
it was suggested that the dm anisotropy might be responsible for selecting the chiral gs @xcite and for opening a small gap in magnetic excitations @xcite of bnfso . in another inelastic neutron scattering ( ins )
study the latter was proposed to arise from single - ion ( si ) anisotropy @xcite . a clear and quantitative determination of the source of the magnetic anisotropy and of its impact on the gs is therefore needed . in this letter
, we present our electron spin resonance ( esr ) study , which provides a direct insight to magnetic anisotropy of bnfso . jointly simulating the experimental esr line - width anisotropy in the paramagnetic phase and magnetic excitations in the ordered phase , we determine the dominant magnetic anisotropy , which is of the dm type and has sizable both the out - of - plane ( @xmath22 ) and the in - plane ( @xmath23 ) component ( fig . [ fig1 ] ) . our mean - field based calculations single out @xmath22 as being responsible for selecting the unique gs .
esr is an extremely powerful technique for quantifying magnetic anisotropy @xcite .
the isotropic heisenberg exchange commutes with the @xmath24 spin operator ( @xmath25 denotes the quantization axis set by the applied magnetic field ) and therefore leads to a @xmath26-function resonance .
finite magnetic anisotropy then yields a finite esr line width . in single crystals
the line - width anisotropy in the paramagnetic phase can often unveil the dominant magnetic anisotropy , since different types of anisotropy reflect local symmetries distinctively , as they arise from different microscopic origins .
finally , esr can also detect collective magnon modes in a lro state .
we therefore performed an extensive esr investigation on high - quality bnfso single crystals in the temperature range between 4 k and 500 k. x - band ( at 9.3 ghz ) spectra were recorded on a home - build double - cavity spectrometer , equipped with a helium - flow cooling and a preheated - nitrogen - flow heating system .
measurements at frequencies between 50 ghz and 400 ghz were performed on a couple of custom - made transmission - type spectrometers at nhmfl , florida . in x - band ,
lorentzian - shaped spectra are observed down to @xmath19 ( fig . [ fig2 ] ) , in accord with strong exchange narrowing regularly encountered in dense magnetic systems .
below @xmath19 , the paramagnetic signal rapidly disappears .
our calibration of the esr intensity , which is proportional to spin - only magnetic susceptibility @xmath27 , and the scaling between @xmath27 and the bulk susceptibility @xmath28 [ fig .
[ fig2](b ) ] prove that the esr signal of bnfso is intrinsic .
the observed small deviations of the @xmath29-factor from the free - electron value @xmath30 are typical for fe@xmath4 with a nearly pure @xmath31s@xmath32 orbital singlet gs @xcite .
a uniaxial anisotropy of the @xmath29-tensor and the line width [ fig .
[ fig2](c - d ) ] are justified by a three - fold rotational symmetry of iron triangles .
our fit of the @xmath29-factor anisotropy @xcite yields the three eigenvalues of the @xmath29-tensor : @xmath33 , @xmath34 and @xmath35 , and thus reveals that also locally the symmetry of the @xmath29-tensor is very close to being uniaxial .
the polar axis @xmath14 on each site is set by a local two - fold rotational axis lying within the crystallographic @xmath6 plane ( @xmath36 direction for site 2 in fig .
[ fig1 ] ) . the other two principal axes @xmath37 and @xmath38 for site 2 lie @xmath39 and 120@xmath40 from the @xmath41 axis in the crystallographic @xmath42 plane @xcite .
the anisotropy of the @xmath29-factor reflects mixing of excited orbital states into the ground orbital singlet of the fe@xmath4 ion , which is induced by a spin - orbit coupling .
the anisotropy of the line width arises from magnetic anisotropy present in a spin hamiltonian , which originates from the same perturbation .
we address the issue of the magnetic anisotropy in bnfso in the framework of the spin hamiltonian , and bulk , @xmath43 , susceptibility for @xmath44 .
inset : room - temperature esr spectrum ( blue ) and the corresponding lorentzian fit ( red ) .
( b ) temperature dependence of line widths and the line width ratio ( inset ) .
angular dependence of ( c ) the line width and ( d ) the @xmath29-factor measured at 500 k , with corresponding fits ( solid lines ) explained in the text . ] @xmath45 where , the first sum runs oven the spin pairs connected by one of the five different exchange interactions ( fig .
[ fig1 ] ) and represents the heisenberg term @xmath46 , the second sum gives the zeeman coupling @xmath47 and the third sum the magnetic anisotropy @xmath48 .
the two dominant contributions @xcite to the latter for fe@xmath4 are the single - ion anisotropy @xcite @xmath49 + const.,\end{aligned}\ ] ] and the dzyaloshinsky - moriya anisotropy @xcite @xmath50 the symmetry of the si anisotropy is the same as of the @xmath29-tensor @xcite , therefore we omit the additional @xmath51 term . since @xmath52 is dominant @xcite , we consider the dm interaction only between nearest neighbors . the dm vector may possess a non - zero out - of - plane @xmath22 and in - plane @xmath23 component parallel to the bond [ fig.[fig1](a ) ] .
the third component is forbidden by the two - fold rotational axis passing through the middle of each bond . using the kubo - tomita approach @xcite
, we can analytically calculate the line - width anisotropy @xmath53 at infinite temperature , separately for both types of magnetic anisotropy @xcite . at finite temperatures @xmath54 , where @xmath55 differs from unity due to finite spin correlations . in bnfso
the uncorrelated paramagnetic state is not reached yet even at 500 k ( @xmath56 ) [ fig .
[ fig2](b ) ] , which complements specific heat measurements showing that @xmath57 of magnetic entropy is still missing at 200 k @xcite .
this needs to be contrasted with several cu - based 2d frustrated lattices , where the esr line width was found constant for @xmath58 @xcite .
we find the fits of the experimental line width [ fig .
[ fig2](c ) ] virtually indistinguishable for the si and the dm model . at 500 k
we extract @xmath59 k , @xmath60 , @xmath61 for the si model and @xmath62 k , @xmath63 for the dm model .
it is worth noting that since fe@xmath4 is in the orbital @xmath64 state , in the si model only the @xmath36 term is allowed in purely cubic , octahedral or tetrahedral environments . in bnfso
the local symmetry is lower , therefore , @xmath65 and @xmath36 can have similar magnitudes @xcite . although the absolute size or the sign of the anisotropy terms can not be revealed from these fits due to unknown @xmath66 , the constant ratio @xmath67 , observed in bnfso above @xmath68 [ inset in fig .
[ fig2](b ) ] , unambiguously sets the ratio of the anisotropy terms for both models @xcite .
the temperature - independent line - width ratio reveals that correlations are isotropic in spin space . in our attempt to determine and quantify the dominant magnetic anisotropy in bnfso
, we now turn to high - frequency measurements .
in contrast to x - band , we can follow the esr signal across @xmath19 [ fig .
[ fig3](a ) ] .
the resonance broadens and shifts to lower field by @xmath69 t below @xmath19 .
the temperature dependence of the line shift mimics an order parameter [ fig .
[ fig3](b ) ] , and corresponds to the opening of a zero - field gap in magnetic excitations ( fig . [ fig4 ] ) .
this transformation of the paramagnetic signal into a collective antiferromagnetic resonance ( afmr ) mode s@xmath70 is accompanied by emergence of another afmr mode ( s@xmath1 ) in our frequency window [ inset in fig .
[ fig3](a ) ] .
the frequency - field diagram of the two modes is shown in fig .
[ fig4 ] for both relevant directions of the applied field with respect to the crystallographic @xmath10 axis . .
( d ) angular dependence of the two resonance modes at 4 k ; measured ( symbols ) and simulated for the dm ( solid lines ) and si anisotropy ( dashed lines ) . ]
we have simulated the afmr modes experimentally detected at 4 k in a mean - field approximation @xcite , again for both types of magnetic anisotropy [ fig .
[ fig4](a ) , ( b ) ] .
we note that the zero - field gap to the lowest gapped branch s@xmath70 is opened by finite anisotropy ( @xmath22 or @xmath65 , @xmath36 , @xmath71 ) in a square - root fashion @xcite , while the gap to the next branch s@xmath1 is present already for isotropic @xmath46 and only slightly shifts with anisotropy .
the latter gap is reproduced with the exchange parameters @xmath72 k , @xmath73 k , @xmath74 k , @xmath75 k , and @xmath76 k , which corresponds to a 20% increase @xcite of the parameters obtained in the ins study @xcite .
the si model [ eq . ( [ eq2 ] ) ] , at first sight , yields a reasonable description of the experimental @xmath77 diagram [ fig .
[ fig4](b ) ] for @xmath78 mk , @xmath79 mk and @xmath80 mk @xcite .
however , a closer look at this model unveils its inconsistencies with the experiment .
the simulated angular dependence of the two afmr modes does not match the experiment [ fig .
[ fig4](d ) ] .
the predicted non - monotonic dependence of the s@xmath1 mode is a consequence of a field - induced phase transition , occurring around 5.5 t @xcite for @xmath81 [ see the discontinuity of curves in fig .
[ fig4](c ) ] .
the apparent absence of the transition at exactly @xmath82 is accidental , which can be deduced from fig .
[ fig4](c ) .
the transition corresponds to the field where the lowest branch reaches zero frequency .
such behavior was indeed observed and justified with the si model on another fe - based triangular lattice @xcite .
the absence of any irregularities in low - temperature magnetization isotherms speaks against such transition in bnfso at least up to 23 t @xcite .
we stress that , on the other hand , no such excitation softening exists for the dm model [ fig .
[ fig4](a ) ] . finally , the si model predicts the selection of the @xmath83 gs independent of the structural chirality .
this contradicts with the experimental observation of the @xmath84 gs in a crystal with structural left handedness @xcite , clearly making the si model incompetent of selecting the proper ground state .
this selection is independent of the structural chirality , because it is a single - ion property .
the dm model [ eq . ( [ eq3 ] ) ] , on the other hand , yields excellent agreement with the experiment for @xmath85 mk . both the @xmath86 diagram [ fig .
[ fig4](a ) ] and the finer - detail angular dependence of both afmr modes [ fig .
[ fig4](d ) ] are simulated convincingly .
the dm anisotropy @xmath22 explains the esr line width and the finite gap to the s@xmath70 branch .
its relative size @xmath87 is in good agreement with the order - of - magnitude prediction @xmath88 , as derived by moriya @xcite .
we note that the zero - field gap could also be explained by the presence of symmetric anisotropic exchange ( ae ) of a similar size @xmath89 @xcite . however , as the ae anisotropy results from the second order perturbation in @xmath90 ( while dm results from the first order perturbation ) , it is of the size @xcite @xmath91 and therefore negligible .
we stress that the in - plane dm component @xmath23 does not affect the detected lowest branches and can not be determined from the studies of these excitations .
this makes the above - presented x - band esr investigation , yielding @xmath92 , invaluable in determining also the @xmath93 mk dm component .
there is a very good agreement between the excitations observed in esr and calculated with the dm model on one hand and the modes identified from the ins experiments @xcite on the other hand .
the calculated s@xmath94 branch [ fig . [ fig4](a ) ] is the gapless goldstone mode emerging at the origin of the reciprocal space ( scattering vector @xmath95=0 ) for the correlations of the @xmath10-axis spin components .
the s@xmath70 mode corresponds to the other , gapped excitations for chiral correlations of the @xmath6-plane spin components at @xmath95=0 .
these excitations at @xmath96 involve correlations of the @xmath10-axis spin - component @xcite and cross @xmath95=0 at @xmath97 ghz , yielding the s@xmath1 signal .
the dm anisotropy thus proves to be the origin of the unique magnetic properties of bnfso .
our simulations @xcite further disclose that one of the two possible gs [ fig .
[ fig1](b ) ] is selected depending on the sign of @xmath22 .
the experimentally detected gs @xmath84 is realized for @xmath98 , when the dm energy [ eq .
( [ eq3 ] ) ] is minimized .
formation of a uniform vector chiral state at finite temperature is , however , challenged by formation of domain walls , which prosper on frustrated lattices due to thermal fluctuations @xcite .
it is likely that these are responsible for diffuse neutron scattering coexisting with lro in bnfso far below @xmath19 @xcite .
it is therefore striking that the minute @xmath3 mk term effectively imposes a single - domain gs at two orders of magnitude larger temperatures , as experimentally verified at 1.5 k @xcite , and a chiral unbalance up to t@xmath99 , with chiral correlations persisting even in the paramagnetic state @xcite . in conclusion
, we have shown that the selection of the unique chiral magnetic ground state in bnfso is due to the dzyaloshinsky - moriya interaction as the dominant source of magnetic anisotropy .
although this term is small ( @xmath3 mk , @xmath100 mk ) , it effectively overcomes thermal fluctuations and leads to a fully chiral state in structurally enantiopure crystals , whereas spin - liquid signatures might also be anticipated above @xmath19 , possibly with skyrmion texturation @xcite .
we thank s. de brion for fruitful discussion and acknowledge the financial support of the slovenian research agency ( projects j1 - 2118 and bi - us/09 - 12 - 040 ) .
000 _ introduction to frustrated magnetism _ , edited by c. lacroix , p. mendels , and f. mila ( springer - verlag , berlin , 2011 ) .
l. balents , nature * 464 * , 199 ( 2010 ) . s. t. bramwell and m. j. p. gingras , science * 294 * , 1495 ( 2001 ) .
a. zorko _ et al .
_ , phys .
rev .
lett . * 107 * , 047208 ( 2011 ) .
t. moriya , phys . rev .
* 120 * , 91 ( 1960 ) .
m. elhajal _ et al .
_ , phys .
rev .
b * 66 * , 014422 ( 2002 ) .
r. ballou _ et al .
_ , j. mag . mag . mat . * 262 * , 465 ( 2003 ) .
k. matan _ et al .
_ , phys .
rev .
lett . * 96 * , 247201 ( 2006 ) . o. cpas _ et al .
_ , phys .
rev .
b * 78 * , 140405 ( 2008 ) .
k. penc _ et al .
_ , phys .
rev .
lett .
* 99 * , 117201 ( 2007 ) .
o. a. starykh _
et al .
_ , phys .
rev .
b * 82 * , 014421 ( 2010 ) .
p. bak and m. h. jensen , j. phys .
c 13 , l881 ( 1980 ) .
d. grohol _ et al .
_ , nature mater . * 4 * , 327 ( 2005 ) .
y. machida _ et al .
_ , nature * 463 * , 210 ( 2010 ) .
m. mostovoy , phys .
rev .
lett . * 96 * , 067601 ( 2006 ) .
y. taguchi _ et al .
_ , science * 291 * , 2573 ( 2001 ) .
x. g. wen , f. wilczec , and a. zee , phys .
rev .
b * 39 * , 11413 ( 1989 ) .
k. marty _ et al .
_ , phys .
rev .
lett .
* 101 * , 247201 ( 2008 ) .
k. marty _ et al .
_ , phys .
rev .
b * 81 * , 054416 ( 2010 ) .
h. d. zhou _ et al .
_ , chem . mater .
* 21 * , 156 ( 2009 ) . c. lee _ et al .
_ , chem . mater . * 22 * , 5290 ( 2010 ) .
m. loire _ et al .
_ , phys .
rev .
lett . *
106 * , 207201 ( 2011 ) . c. stock _ et al .
_ , phys .
rev .
b * 83 * , 104426 ( 2011 ) .
a. abragam and b. bleaney , _ electron paramagnetic resonance of transition ions _
( clarendon , oxford , 1970 ) .
see supplementary material for details on the x - band esr @xmath29-factor and line - width fitting as well as for modeling of the afmr modes at higher frequencies .
dipolar interaction between fe@xmath4 moments and their hyperfine coupling to nuclear moments are negligible they yield line widths of @xmath101 mt due to strong exchange narrowing .
symmetric anisotropic exchange being a quadratic correction in @xmath90 is inferior to the antisymmetric dzyaloshinsky - moriya term being linear in @xmath90 @xcite .
a. bencini and d. gatteshi , _ epr of exchange coupled systems _ , ( springer - verlag , berlin , 1990 ) .
r. kubo and k. tomita , j. phys .
soc .
jpn * 9 * , 888 ( 1954 ) .
a. zorko _ et al .
_ , phys .
rev .
b * 69 * , 174420 ( 2004 ) . a. zorko _ et al .
_ , phys .
rev .
lett .
* 101 * , 026405 ( 2008 ) .
the increase of the exchange parameters by @xmath57 compared to those from the ins experiment @xcite is necessary because we use the experimental value of the ordered moments 4@xmath102 instead of the full value 5@xmath102 .
other selections of the sign for the si parameters do not yield afmr modes resembling the experiment .
the critical field is shifted to 7.5 t if exchange parameters from ref .
@xcite , which are significantly different , are used .
a. i. smirnov _ et al .
_ , phys .
rev .
b * 75 * , 134412 ( 2007 ) .
k. marty , phd thesis , university of grenoble ( 2008 ) ; http://tel.archives-ouvertes.fr/tel-00351575/en/ j. jensen , arxiv:1106.1052
. j. t. chalker , p. c. w. holdsworth , and e. f. schender , phys .
rev .
lett . *
68 * , 855 ( 1992 ) .
h. d. zhou _ et al .
_ , phys .
rev .
b * 82 * , 132408 ( 2010 ) . c. pappas ,
_ et al . _ ,
phys .
rev .
lett . * 102 * , 197202 ( 2009 ) . c. pappas , _ et al .
_ , phys .
rev .
b * 83 * , 224405 ( 2011 ) .
+ a. zorko,@xmath103 m. pregelj,@xmath104 a. potonik,@xmath104 j. van tol,@xmath105 a. ozarowski,@xmath105 v. simonet,@xmath106 p. lejay,@xmath106 s. petit,@xmath107 and r. ballou@xmath106 _ @xmath104joef stefan institute , jamova 39 , si-1000 ljubljana , slovenia + @xmath108en fist centre of excellence , dunajska 156 , si-1000 ljubljana , slovenia + @xmath105national high magnetic field laboratory , florida state university , tallahassee , florida 32310 + @xmath106institut nel , cnrs and universit joseph fourier , bp 166 , 38042 grenoble , france + @xmath106laboratoire lon brillouin , cea - cnrs , ce - sacley , f-91191 gif sur yvette , france + _ |
since complex network models @xcite were applied to mathematical epidemiology @xcite in such a way that the host population is modelled as a contact network where nodes represent individuals and links stand for contacts among them the topological structure of the underlying network has shown a strong impact on the spread of infectious diseases @xcite .
for instance , based on mean - field approximations , @xcite suggested that there is no typical epidemic threshold in a scale - free ( sf ) network which follows a power - law degree distribution @xmath0 with the power exponent @xmath1 in the thermodynamic limit .
otherwise , there is a positive critical value @xmath2 of the transmission rate @xmath3 only for @xmath4 .
it is worth remarking that in a more mathematically rigorous analysis , @xcite proved that the critical value @xmath2 is also zero for contact processes on random networks with power - law degree distributions for any power exponent @xmath4 . in other words
, the infection can pervade the whole sf network provided the transmission rate is nonzero . over the past decades
, considerable progress has been made in the study of epidemic models on static networks , where the network topology remains fixed during the process of disease transmission .
many topological features of the network , such as the heterogeneity @xcite , degree correlation @xcite and clustering @xcite , were explicitly addressed , yet most studies ignore the impact of human behavior ( in response to infection ) on the disease @xcite .
such effect is likely to be especially acute when modifications to human behavior causes changes in the structure of the underlying contact network @xcite .
most recently , it has been demonstrated that a better understanding of the interplay between human behavior and epidemic dynamics would help design effective intervention strategies such as vaccination and social distancing policies , and improve the cost effectiveness of disease control @xcite .
it is therefore of great significance to incorporate human behavior in the study of network epidemiology @xcite . in general
, there are two scenarios of human behavior to be considered in the face of a contagious disease .
first , at the macroscopic level authorities enact public containment programs such as quarantine or vaccination ; second , at the microscopic level individuals adopt self - protective measures against the disease based on their perceived risk of infection @xcite .
vaccination is one of the most effective public policies for preventing the transmission of infectious diseases @xcite , and to date there has been a wealth of studies on various vaccination strategies for epidemic models in complex networks @xcite , indicating powerful effects of vaccination on the inhibition of disease transmission , even though some vaccines provide temporary immunity and only reach part of the population .
for instance , targeted vaccination towards high - degree nodes is widely known to have a much higher effectiveness than random vaccination either in sf networks @xcite or in small - world networks @xcite .
however , targeting the highest - degree nodes requires full knowledge of the degree of each node in the network . to overcome this shortcoming ,
an alternative vaccination strategy based on local information has been proposed @xcite . under this strategy vaccination
is implemented on a random acquaintance of a randomly chosen node .
again , this acquaintance vaccination proves to be more effective than random vaccination since it prefers to target high - degree nodes ( this is a consequence of the so - called friendship paradox on average neighbours of a random node have more neighbours than the node itself )
. on the other hand , appropriate and timely response of individuals to a spreading contagion may improve the efficiency of vaccination and save medical resources .
several studies looking at this hypothesis have been framed in network epidemic models through the addition of a rewiring mechanism ( for details , see @xcite and the references therein ) .
a susceptible - infected - susceptible ( sis ) epidemic model on adaptive networks was first proposed by @xcite , where at a given rate susceptible individuals rewire their connections from infected neighbours to the susceptibles .
the basic idea of this adaptive network epidemiology is that individuals tend to change their habits and social contacts to avoid infection in response to a disease threat @xcite .
still further rewiring scenarios have also been considered reflecting different individual reactions to the infectious status of others .
for example , @xcite introduced a _
neighbour exchange _ model to account for the impacts of the stability of social contacts on the spread of infectious diseases @xcite .
alternatively , a random swapping of contact partners between susceptible - infected and infected - infected links was studied in @xcite .
most adaptive epidemic models have largely focused on the standard sis model @xcite , where the rewiring operation isolates infective individuals from the susceptible population in the long run , and therefore influences the epidemic propagation . in turn , the larger the infected fraction in the population , the more likely susceptible individuals will choose to rewire their contacts with infective neighbours .
in other words , the epidemic prevalence can affect individual behavior , while avoiding contacts with infection also exerts a strong impact on epidemic dynamics . typically , in an adaptive network model both the dynamics of and on the network form a feedback loop @xcite , which usually involves many interesting dynamical features such as bistability and oscillation @xcite .
so far , little attention has been paid to the combined effects of public vaccination and individual protection .
only recently have shaw and schwartz adopted a random vaccination strategy in an adaptive sis model to examine the interplay between individual rewiring of social contacts and random vaccination of susceptible individuals @xcite . in their model ,
vaccinated individuals secure complete protection from infection unless they return to the susceptible state .
that is , the vaccinated can not be infected directly from their infective neighbours .
however , there is clear evidence that some vaccines are not completely efficient , namely , they come with no guarantee of perfect protection from infection @xcite rather , they confer only partial immunity and a lower transmission rate than the unvaccinated susceptibles .
moreover , in reality , vaccines rarely cover the entire population and only provide finite - time immunity against infection . in light of this , @xcite proposed a compartmental sis model with imperfect vaccination , where vaccinated individuals may contract infection directly from their infectious contacts , although at a vaccine - reduced transmission rate @xcite .
hereafter , we refer to this as the siv model .
most recently , this imperfect vaccination model has been restudied on static sf networks , revealing surprising epidemic behavior @xcite .
however , the role of individual behavior has been neglected in changing the transmission routes of the disease as well as in reshaping the contact patterns of the underlying network .
this paper aims to provide a systematic framework that couples public vaccination and individual protection . for this purpose
, we develop an adaptive siv model by extending the compartmental siv model @xcite to an adaptive network @xcite .
the adaptive siv model integrates contact rewiring at the microscopic level with vaccination intervention at the macroscopic level in response to disease propagation .
in particular , we consider a population without and with demographic effects , respectively , to account for two different cases as follows : case ( 1)the disease spreads through the population much faster than the births and deaths of individuals , and case ( 2)the disease spreads on a timescale comparable to the speed of individuals births and deaths .
this observation mainly depends on what type of the disease is being modelled .
for instance , the seasonal influenza can be modelled well in a population without demography .
conversely , for the human immunodeficiency virus ( hiv ) where infection span decades , then the demographic impact should be taken into account @xcite .
we establish a mathematical configuration for both cases , which allows for an in - depth analysis of the mutual interaction between disease propagation and network evolution . in sect .
[ sec2 ] we introduce the adaptive siv model , and construct a set of approximating deterministic equations for both a closed system and one with demography .
section [ sec3 ] contains main results and sect .
[ sec4 ] concludes .
-links . at each time step ,
susceptible nodes are infected by their infected neighbours at a transmission rate @xmath5 per @xmath6-link and are vaccinated at a per capita rate @xmath7 . infected nodes
recover and return to being susceptible at a per capita rate @xmath8 .
vaccinated nodes become susceptible at a per capita rate @xmath9 and are infected at a vaccine - reduced transmission rate @xmath10 per @xmath11-link due to imperfect immunity . with the background demographic changes taken into consideration , we assume that each node can give birth to a susceptible node with a birth rate @xmath12 , and each node dies at a natural death rate @xmath13 .
in addition , the infected nodes die with an extra rate @xmath14 because of infection , namely , each infected node will die at a rate @xmath15 . in the adaptive siv model , we assume that only susceptible nodes are permitted to rewire connections away from their infected neighbours at a rate @xmath16 per @xmath6-link . in other words , with rate @xmath16 , for each @xmath6-link , the susceptible node breaks the link to the infected node and forms a new link to a randomly selected susceptible or vaccinated node . ]
we employ individual rewiring of contacts with infective neighbours as an individualised self - protective behavior and the vaccination scheme ( reducing the rate of infection ) as a public intervention in reaction to an infectious disease .
the nodes of the network represent individuals in the population and links are potentially infectious contacts among them . in the adaptive siv model
, each node may have only one of the three possible states : susceptible ( @xmath17 ) , infected ( @xmath18 ) and vaccinated ( @xmath19 ) . as demonstrated in fig .
[ fig1](a ) , the transition probabilities for node states are as follows : a susceptible node becomes infected with the transmission rate @xmath5 per @xmath6-link ( where the parameters in the model definition are listed in table [ table1 ] ) ; 1lx term & meaning + @xmath5 & transmission rate per @xmath6-link + @xmath8 & recovery rate from infection + @xmath7 & vaccination rate of susceptibles + @xmath9 & relapse rate of vaccinated individuals ( i.e. , each vaccinated returns to being susceptible after an average time period of @xmath20 due to temporary immunity ) + @xmath21 & effectivity of the vaccine - induced protection against infection ( i.e. , vaccinated individuals can be infected at a vaccine - reduced transmission rate @xmath10 ) + @xmath12 & birth rate + @xmath13 & natural death rate + @xmath14 & death rate from infection + @xmath16 & rewiring rate per @xmath6-link where the susceptible node breaks away from the infected and then establishes a connection with a randomly chosen noninfected node + an infective node recovers back to the susceptible state with the recovery rate @xmath8 ; according to random vaccination strategy , a susceptible node is vaccinated at the vaccination rate @xmath7 ; a vaccinated node returns to the susceptible state with the resusceptibility rate @xmath9 as the vaccine wears off ; due to imperfect immunity , the vaccinated node will be infected directly from one of its infective neighbours with a reduced transmission rate @xmath10 , where the parameter @xmath22 measures the inefficacy of the vaccine - induced protection against infection ( @xmath23 means the vaccine is completely effective , while @xmath24 means the vaccine is totally ineffective ) . with the background demography taken into consideration , we assume that each node can give birth to a susceptible node with a birth rate @xmath12 , and each node ( irrespective of its state ) dies at a natural death rate @xmath13 .
in addition , the infected nodes die with an extra rate @xmath14 because of infection .
that is to say , each infected node will die at a rate @xmath15 .
meanwhile , the network is rewired adaptively during the disease propagation [ see fig .
[ fig1](b ) as an illustration ] . for simplicity ,
only susceptible nodes are allowed to rewire their connections from infected neighbours to randomly selected noninfected ( either susceptible or vaccinated ) nodes with the rewiring rate @xmath16 .
self links and multiple links are prohibited .
the adaptive siv model therefore consists of two processes ( fig .
[ fig1 ] ) : one is the node dynamics due to the epidemic spreading and the birth and death process , the other is the link dynamics due to the rewiring of @xmath6-links of the underlying network . in what follows , we consider two types of population system a closed system without demography and an open system with demography , respectively , to account for the following two different cases : case ( 1 ) the disease spreads sufficiently fast that the background demographic changes are not significant , and case ( 2 ) the epidemic occurs on a timescale comparable to the speed of the birth and death process .
therefore , some fast diseases like the seasonal influenza can be well described with the case ( 1 ) model , whereas for some slow diseases such as hiv we need to consider the case ( 2 ) model .
1lx term & meaning + @xmath25 & number of nodes in state @xmath26 with degree @xmath27 + @xmath28 & number of nodes in state @xmath26 + @xmath29 & number of nodes with degree @xmath27 + @xmath30 & total number of nodes in the network + @xmath31 & fraction of nodes in state @xmath26 , where @xmath32 corresponding to @xmath33 , respectively + @xmath34 & degree distribution of the network ( i.e. , probability that a randomly selected node has a degree @xmath27 ) + @xmath35 & average degree of the network + @xmath36 & probability for a node in state @xmath26 to have a degree @xmath27 + @xmath37 & probability generating function of @xmath38 + @xmath39 & partial derivative of function @xmath40 with respect to @xmath41 + @xmath42 & second - order partial derivative of function @xmath40 with respect to @xmath41 + @xmath43 & average degree of nodes in state @xmath26 + @xmath44 & degree distribution of newborns ( i.e. , probability for a node entering the network to have @xmath27 links ) + @xmath45 & probability generating function of @xmath44 + @xmath46 & first - order derivative of function @xmath47 with respect to @xmath41 + @xmath48-link & a link connecting a node in state @xmath26 with a node in state @xmath49 + @xmath50 & number of @xmath48-links + @xmath51 & number of links emanating from nodes in state @xmath26 + @xmath52 & total number of links in the network + @xmath53 & fraction of @xmath48-links in the network + @xmath54-type triple & a triple in which a node in state @xmath26 is connected to a node in state @xmath49 who in turn is connected to a node in state @xmath55 + @xmath56 & number of @xmath54-type triples in the network + in this case , obviously the demographic turnover can be ignored by setting the demographic parameters to be zero , i.e. , @xmath57 .
there is neither the addition of new nodes nor the removal of old nodes , and the rewiring mechanism conserves the total number of links among the entire network .
therefore , the numbers @xmath58 ( of nodes ) and @xmath59 ( of links ) are constant during the model evolution .
following @xcite and @xcite , we also develop a mean - field model for the system , which tracks the dynamics of both nodes and links .
the notation summarised in table [ table2 ] allows us to obtain the following set of differential equations for the evolution of the fractions of nodes as follows : @xmath60 on the right hand side ( rhs ) of eq .
( [ eq:1 ] ) , the first and second terms correspond to the conversion of infected and vaccinated nodes into susceptible ones , respectively , while the third term indicates the vaccination process of susceptible nodes , and the fourth term represents the loss of susceptible nodes due to infection , which is proportional to the number @xmath61 of @xmath6-links and the transmission rate @xmath5 per @xmath6-link . on the rhs of eq .
( [ eq:2 ] ) , the first and second terms consider that infection spreads along each @xmath6-link with rate @xmath5 and along each @xmath11-link with rate @xmath10 , respectively , while the third term describes recovery of infected nodes at rate @xmath8 .
similarly , the first term on the rhs of eq .
( [ eq:3 ] ) describes the gain of vaccinated nodes due to vaccination of susceptible nodes , while the second and third terms describe the loss of vaccinated nodes as a result of immunity wearing off and infection along each @xmath11-link , respectively .
the nodal dynamics interact closely with the dynamics of links through the change of nodal states and infection of noninfected nodes along @xmath6- and @xmath11-links .
in addition , rewiring of @xmath6-links also implicitly affects the system since it alters the numbers of @xmath6- , @xmath62- , and @xmath63-links . to close the model system
, we use the low - order moment closure approximation @xmath64 @xcite with the coefficient @xmath65 being omitted , where @xmath66 and @xmath56 represents the number of @xmath54-type triples in which a node in state @xmath26 is connected to a node in state @xmath49 who in turn is connected to a node in state @xmath55 .
this leads to the following equations for the evolution of the fractions of links : @xmath67 where the triple approximations in terms of the number of pairs have been absorbed into this formulation .
the first three terms on the rhs of eq .
( [ eq:4 ] ) describe the gain of @xmath62-links as a result of recovery , loss of immunity , and rewiring , respectively , while the last two terms correspond to the loss of @xmath62-links due to vaccination of susceptible nodes and infection through @xmath68-type triples , respectively . equations ( [ eq:5])([eq:9 ] ) are analogous except that infection of vaccinated nodes at a reduced rate @xmath10 has been taken into account . since the numbers @xmath58 ( of nodes ) and @xmath59 ( of links ) are constant in this model , the whole system is closed under the conservation conditions @xmath69 and @xmath70 .
the steady states of the above ordinary differential equations ( [ eq:1])([eq:9 ] ) can be tracked as a function of parameters via the continuation software @xcite , which helps investigate the dynamical effects of rewiring and vaccination through the bifurcation analysis ( see the results section [ results ] ) . in the previous section
, we assume that the time scale of epidemic spread is much faster than the time scale of population births and deaths .
this is not adequate for slow diseases which last for decades . to overcome this shortcoming ,
we take into account the background demographics renewal in the adaptive siv model , where ( i ) each individual can give birth to a susceptible at a birth rate @xmath12 and the newborn randomly connects to @xmath27 existing individuals according to a preassigned birth degree distribution @xmath44 , which corresponds to the probability generating function @xmath47 , and meanwhile ( ii ) each individual dies from natural causes at a rate @xmath13 and the infective dies at an extra rate @xmath14 ( possibly , one expects @xmath71 due to the disease itself something which is omitted from sis / siv models without demography ) .
we do not consider the possibility of either vaccination or infection being passed directly to offspring .
the low - dimensional mean - field equations in the case ( 1 ) allows us to carry out in - depth bifurcation analysis and predict rich dynamical features for the closed system .
however , the case ( 1 ) model is based on a homogeneous mixing assumption and low - order moment - closure approximations , which fundamentally neglect the effects of the underlying network structure , generally resulting in inaccurate predictions about the time evolution of the system @xcite .
one important network property to be considered is the heterogeneity of the degree distribution among nodes @xcite . to make our model applicable to heterogeneous networks ,
the network nodes are assembled into subgroups @xmath72 according to their degree @xmath27 , where @xmath72 denotes the number of nodes in state @xmath73 with @xmath27 connections . with the notation listed in table [ table2 ] , one can write the evolution equations for the numbers of susceptible , infected and vaccinated nodes with @xmath27 connections @xmath74 , @xmath75 , and @xmath76 as follows : @xmath77 -\mu\frac{m_{si}}{m_{s}}[ks_{k}-(k+1)s_{k+1 } ] , \label{eq:10 } \\
\nonumber & & \\
\frac{\rm d}{{\rm d}t}i_{k } & = & -(\eta_{2}+\mu)i_{k } + \alpha\frac{m_{si}}{m_{s}}ks_{k}+\delta\alpha\frac{m_{iv}}{m_{v}}kv_{k } + \eta_{1}\bar{g}^\prime(1)(i_{k-1}-i_{k } ) -\beta i_{k } \nonumber \\ & & -\omega \frac{m_{si}}{m_{i}}[ki_{k}-(k+1)i_{k+1 } ] -\eta_{2}[ki_{k}-(k+1)i_{k+1 } ] -\mu\frac{2m_{ii}}{m_{i}}[ki_{k}-(k+1)i_{k+1 } ] , \quad\quad\quad \label{eq:11 } \\
\nonumber & & \\ \frac{\rm d}{{\rm d}t}v_{k } & = & -\eta_{2}v_{k } + \varphi s_{k } - \delta\alpha\frac{m_{iv}}{m_{v}}kv_{k } + \eta_{1}\bar{g}^\prime(1)(v_{k-1}-v_{k } ) -\psi v_{k } \nonumber \\ & & + \omega m_{si}\frac{v_{k-1}-v_{k}}{n_{s}+n_{v } } -\eta_{2}[kv_{k}-(k+1)v_{k+1 } ] -\mu\frac{m_{iv}}{m_{v}}[kv_{k}-(k+1)v_{k+1}].\label{eq:12}\end{aligned}\ ] ] on the rhs of eq .
( [ eq:10 ] ) , the first ( second ) term denotes the creation ( loss ) of @xmath78 due to natural births ( deaths ) , where @xmath79 is the increasing rate of newborn susceptibles that are of degree @xmath27 with probability @xmath80 ( according to the degree distribution of newborns ) .
the third term corresponds to the loss of @xmath78 because of infection that is proportional to the number of connections @xmath27 , the transmission rate per @xmath6-link @xmath5 , and the probability @xmath81 for a link originating from the susceptible to point to the infective , which is assumed to have no degree correlation with the starting node .
the fourth term represents the removal of @xmath78 due to vaccination .
the fifth term considers a growth rate @xmath79 of newborns , each of which on average has a number @xmath82 of links . the number @xmath78 increases ( decreases ) when the susceptibles with degree @xmath83 ( degree @xmath27 ) are connected to by one of the @xmath84 links emanating from the newborn nodes .
in addition , the probability for the newborns to randomly link to the susceptibles with @xmath27 contacts is @xmath85 .
note that the establishment of connections of the newborns is simply based on the random attachment mechanism .
the sixth and seventh terms refer to the recovery process from infection and the relapse to susceptibility for vaccinated nodes as the vaccine wears off , respectively .
the eighth term is attributable to the rewiring process with the rewiring rate @xmath16 per @xmath6-link .
the number @xmath78 increases ( decreases ) when the susceptibles with degree @xmath83 ( degree @xmath27 ) is randomly rewired towards by the susceptible endpoint on the @xmath6-link after disconnecting from the infected endpoint . and
, the probability for the susceptible endpoint to randomly rewire to the susceptibles with degree @xmath27 from the pool of noninfected nodes is @xmath86 .
the ninth term depicts the changes of @xmath78 owing to contacts lost from naturally dying nodes .
the number @xmath78 increases ( decreases ) when one of the @xmath87 neighbours ( @xmath27 neighbours ) of the susceptible with degree @xmath87 ( degree @xmath27 ) dies naturally at a rate @xmath13 .
the tenth term describes the changes of @xmath78 due to contacts lost from nodes dying from infection .
the number @xmath78 increases ( decreases ) when one of the infected neighbours of the susceptibles with degree @xmath87 ( degree @xmath27 ) dies from infection at a rate @xmath14 .
there is a fraction @xmath81 of infected nodes among the @xmath88 neighbours of the susceptibles with degree @xmath27 . here
, we assume that the probability for a link emanating from a susceptible node to point to an infected node is uncorrelated to the degree @xmath27 of the starting node . the derivation of eqs .
( [ eq:11 ] ) and ( [ eq:12 ] ) can be explained analogously .
the evolution equations for the numbers of susceptible ( @xmath89 ) , infected ( @xmath90 ) , and vaccinated ( @xmath91 ) nodes , as well as the numbers of links emanating from susceptible ( @xmath92 ) , infected ( @xmath93 ) , and vaccinated ( @xmath94 ) nodes , can be derived according to eqs .
( [ eq:10])([eq:12 ] ) , as given in detail in appendix a. to close the system and to take into account the dynamical correlations in the infection status of connected individuals , additional equations for the evolution of the number @xmath50 of @xmath48-links are required , where @xmath95 . in a similar manner , we obtain @xmath96 on the rhs of eq .
( [ eq:13 ] ) , the first four terms describe the addition of @xmath62-links as a result of newborn nodes ( with an average connectivity @xmath97 ) connecting to old susceptible ones with probability @xmath98 , the recovery of the infected endpoints on @xmath6-links at rate @xmath8 , the resusceptibility of vaccinated endpoints of @xmath63-link due to temporary immunity , and , the rewiring of @xmath6-links away from the infected nodes and towards other susceptible ones with probability @xmath99 , respectively ; while the last three terms correspond to the loss of @xmath62-links due to the natural death ( at a rate @xmath100 ) and the vaccination ( at a rate @xmath101 ) of one of the two endpoints of @xmath62-links , and the infection ( at a rate @xmath5 ) of the intermediate susceptible node along the @xmath68-type triples , respectively .
the appearance of the probability generating functions in eq .
( [ eq:13 ] ) represents an improvement in the approximations of the number of triples in terms of the number of pairs by taking into account the heterogeneity of the central node s degree @xmath27 , rather than simply assuming a homogeneous mixing .
a complete description of this type of formulation is given in appendix b. explanations for the derivation of eqs .
( [ eq:14])([eq:18 ] ) are similar except that the infection of vaccinated nodes at a vaccine - reduced transmission rate @xmath10 has also been considered . equations ( [ eq:13])([eq:18 ] ) characterise the contact behavior among individuals of different states and how they change in response to epidemic transmission and adaptive rewiring of @xmath6-links . in conjunction with the set of eqs .
( [ eq:10])([eq:12 ] ) , they govern the temporal evolution of both the epidemic and the underlying contact network . our analytic approach in eqs .
( [ eq:10])([eq:18 ] ) improves on the usual heterogeneous mean - field approximation by @xcite as we take into account both the connectivity heterogeneity but also the dynamical correlation among the infection states of adjacent nodes , which is generally ignored in the heterogeneous mean - field approximation @xcite .
it is worth noting that @xcite has recently proposed an improved heterogeneous moment - closure approximation , referred to as the active neighbourhood approach , to study the sis epidemic dynamics on adaptive networks .
this active neighbourhood approach categorises all the nodes based on the number of their susceptible and infected neighbours . in the approximation for
the number of triples , our approach only takes the central node s degree into consideration , but the active neighbourhood approach by @xcite , by contrast , takes the starting node s and the central node s degrees ( both the susceptible degree and infectious degree ) into account . in this
regard , our approach may lose some accuracy for the approximation of number of triples .
on the basis of the active neighbourhood approach , @xcite used generating functions with two spatial variables to investigate an adaptive voter model , providing new insights into the connection between network science and the theory of partial differential equations .
moreover , under the formalism framework of the active neighbourhood approximation approach , some improved moment - closure approximations have also been discussed for adaptive networks@xcite . due to the detailed information about the total degree and the infectious degree included in the active neighbourhood approach , the above moment - closure approximations are more accurate but is of much greater computation complexity in comparison with the moment - closure approach as in appendix b , where the approximation of the number of triples only considers the central node s degree . moreover , all these approximation approaches have assumed zero degree correlation among nodes in the network .
[ sec3 ] for the adaptive siv model without population births and deaths , we focus on the effects of rewiring and vaccination on both dynamics of the epidemic and the network .
the steady state of the system of eqs .
( [ eq:1])([eq:9 ] ) can be tracked as a function of parameters with the bifurcation software @xcite . as a function of the transmission rate @xmath5 for different rewiring rates : @xmath102 ( a ) , @xmath103 ( b ) , and @xmath104 ( c ) , respectively .
black solid ( dashed ) lines : stable ( unstable ) branches ; blue ( green ) solid lines : lower ( upper ) turning point of the stable epidemic cycles ; blue ( green ) dashed lines : lower ( upper ) turning point of the unstable epidemic cycles . dots and error bars ( if larger than the dots ) : simulation results .
all the simulation points are obtained by averaging over 50 runs , each of which goes through @xmath105 time steps , starting with an initial fraction @xmath106 of infected seeds and all the others susceptible .
parameters are @xmath107 , @xmath108 , @xmath109 , @xmath110 , @xmath111 , and @xmath112 . ]
we first consider the bifurcation structure of the model .
figure [ fig2 ] shows the bifurcation diagram of the infective fraction @xmath113 as a function of the transmission rate @xmath5 .
the analytic results from the mean - field eqs .
( [ eq:1])([eq:9 ] ) are compared with computational simulations of the full model . in the case of no rewiring ( @xmath102 ) [ fig .
[ fig2](a ) ] , a single transcritical bifurcation occurs at the well - known epidemic threshold @xcite , @xmath114 , which indicates the critical value of transmission rate for invasion of new infections .
hereafter we refer to this threshold as the invasion threshold " . in this
setting , @xmath115 , coincides with the critical behavior @xmath116 obtained in @xcite . if @xmath117 , the disease - free ( healthy ) state loses stability while the endemic state becomes stable . as a low rewiring rate @xmath118
is introduced [ fig .
[ fig2](b ) ] , the epidemic threshold grows rapidly to @xmath119 . with an increment in @xmath5 , a hopf bifurcation takes place at @xmath120 with the appearance of a stable epidemic cycle @xcite ( i.e.
, the limit cycle , which is a periodic solution ) , and then a region of periodic oscillations is observed in the range of @xmath121 .
in addition , two saddle - node bifurcations occur at @xmath122 and @xmath123 , respectively , between which a clear region of bistability is established , as shown in the dark gray region ( where a stable endemic state and a stable epidemic cycle coexist ) and the medium gray region ( with two coexisting stable endemic states ) . as the rewiring rate @xmath16 rises to 0.2 [ fig .
[ fig2](c ) ] , the saddle - node bifurcation on the left induces an endemic threshold , above which an already established endemic persists . for convenience ,
we refer to this threshold as the persistence threshold " . across the turning point of the right saddle - node bifurcation ( a little to the right ) , there is a cycle fold bifurcation ( i.e. , the fold bifurcation of limit cycles ) at @xmath124 , where stable and unstable epidemic cycles converge , and then disappear . between the left saddle - node bifurcation and the cycle fold bifurcation , a wider bistable region is located : in the dark gray region , a stable endemic state and a stable epidemic cycle coexist ; in the medium gray region , there are two stable endemic states ; and in the light gray region , a stable disease - free state coexists with a stable endemic state . again
, an oscillation phenomenon is observed in the dark gray region i. in our simulations searching for the branch points in the bistable regions in fig .
[ fig2 ] , we have first examined which attractor the system goes to by the continuation software @xcite , thereby we obtained distinctive ranges of different stable branches , and then averaged over considerable runs that appeared to be in the same basin of attraction . in a similar manner to @xcite
, we located the stable endemic state branch by sweeping the transmission rate @xmath5 from large to small values , and the steady state was averaged over 50 runs , each of which goes through @xmath105 time steps , starting with an initial fraction @xmath106 of infected seeds .
the simulation results that dropped into the range of high stable endemic branch were regarded as the high endemic branch point , while the simulation results that dropped into the range of low stable endemic branch were regarded as the low endemic branch point . to locate the stable disease - free state branch
, we swept @xmath5 from zero to large .
we considered the disease - free state to be stable if the disease died out in any run of simulation .
however , our method might fail to find stable states with small basins of attraction due to the stochastic nature of computational simulations and because the disease - free state is absorbing .
generally speaking , the simulation and the analytic results in fig .
[ fig2 ] are in good agreement , especially in the upper branches ( endemic state ) and the lower branches ( disease - free state ) . however , in the oscillatory regions where the limit cycles coexist with a stable endemic state , there are a little discrepancies between theory and simulations .
elsewhere it has been found that there is often disagreement between mean - field ( or pairwise ) approximation approaches and network simulations about the parameter ranges for oscillations . in fig .
[ fig2](b ) , while we have observed periodic oscillations ( limit cycles ) in the region , the error bars suggest there is potential for oscillations to appear also in the region .
on the other hand , the error bars in fig .
[ fig2](c ) are so small that it is unclear whether oscillations appear in region as predicted by the mean - field approximation . plane for different rewiring rates : @xmath118 ( a ) and @xmath104 ( b ) .
the yellow ( in panel ( a ) ) , the white and light green regions , respectively , represent a single attractor region , where it is an epidemic cycle in the yellow region ( in panel ( a ) ) , an endemic state in the white ( endemic ) region , and a disease - free state in the light green ( healthy ) region .
the remaining area represents a bistable region , which is divided into three colours in different gray levels : in the light gray region i both the disease - free state and an endemic state are stable ; in the medium gray region ii there are two stable endemic states ; and in the dark gray region iii an endemic state and an epidemic cycle are both stable .
oscillations can be observed both in the yellow region ( in panel ( a ) ) and in the dark gray region iii due to the presence of a stable epidemic cycle .
purple ( solid ) lines : the transcritical bifurcation points ; red ( dash - dotted ) lines : the saddle - node bifurcation points ; black ( dashed ) lines : the hopf bifurcation points ; blue ( dotted ) lines : the cycle fold bifurcation points .
parameters are @xmath107 , @xmath108 , @xmath109 , @xmath111 , and @xmath112 . ] by varying the vaccination rate @xmath7 and tracking each bifurcation curve for the rewiring rate @xmath118 , we obtain the two - parameter bifurcation diagram on the @xmath125 plane in fig .
[ fig3](a ) . in the light green , white , and yellow regions , there is only a single attractor , which is a disease - free state in the light green ( healthy ) region , an endemic state in the white ( endemic ) region , and an epidemic cycle in the yellow region .
the rest of the plane in fig .
[ fig3](a ) represents a bistable region , which is divided into three colours in different gray levels : in the light gray region i both the disease - free equilibrium and an endemic equilibrium are stable ; in the medium gray region ii there are two stable endemic equilibria ; and in the dark gray region iii an endemic equilibrium and an epidemic cycle are both stable .
oscillations can be observed both in the yellow region and the dark gray region iii since there exists a stable epidemic cycle .
analogously , we get in fig .
[ fig3](b ) the two - parametric bifurcation diagram for a larger rewiring rate @xmath126 . in the light green ( healthy ) region
, there is only a stable disease - free equilibrium .
furthermore , there are three different regions of bistability : in the light gray region , both the disease - free and endemic states are stable ; in the medium gray region , there are two stable endemic states ; and in the dark gray region , a stable epidemic cycle coexists with a stable endemic state .
also , oscillatory dynamics can be observed in the dark gray region . on the contrary , in the white ( endemic ) region
, there is only one stable endemic state . from fig .
[ fig3 ] , it can be found that for @xmath127 ( i.e. , without vaccination ) , there is a backward bifurcation and a saddle - node bifurcation .
as the vaccination rate @xmath7 gradually increases , the model exhibits a more complicated bifurcation structure with the occurrence of two saddle - node bifurcations , hopf bifurcations , oscillations , and the cycle fold bifurcation .
finally , for @xmath7 sufficiently large , the epidemic cycles disappear and the dynamical system is characterised by a transcritical bifurcation and two saddle - node bifurcations .
moreover , the bifurcation points of both the transcritical and saddle - node bifurcations shift rightwards with the increase of @xmath7 .
in other words , the public vaccination intervention plays a crucial part in preventing and controlling the spread of infectious diseases by increasing the invasion threshold and also the persistence threshold for epidemics .
furthermore , in comparison with fig .
[ fig3](a ) , the disease - free region is larger in fig .
[ fig3](b ) , which suggests that the larger the adaptive rewiring rate @xmath16 is , the safer the population is from infection . ) plane . in the white , yellow , and light green regions
, there is only a single attractor , which is an endemic state in the white ( endemic ) region , an epidemic cycle in the yellow region , and a disease - free state in the light green ( healthy ) region .
the remaining part represents a bistable region , which is divided into three colors in different gray levels : in the light gray region both of the disease - free and endemic states are stable ; in the medium gray regions and there are two stable endemic states ; and in the dark gray region a stable endemic equilibrium and a stable epidemic cycle coexist .
the oscillatory dynamics can be observed in both the yellow and the dark gray regions since there exists a stable epidemic cycle .
the purple solid , red dash - dotted , black dashed , and blue dotted lines , respectively , correspond to the transcritical , saddle - node , hopf , and cycle fold bifurcation points .
parameters are as in fig .
[ fig2 ] . ]
to understand the effect of individual rewiring on the epidemic dynamics in more detail , a two - parameter bifurcation diagram on the plane of @xmath128 is plotted in fig .
[ fig4 ] . in the case of no rewiring ( i.e. , @xmath102 )
, there is only the transcritical bifurcation ( denoted by a solid line ) , which separates the healthy and endemic regions .
as @xmath16 increases , the system first enters a bistability region ( at @xmath129 ) , i.e. , the medium gray region ( where two stable endemic states coexist ) , and then undergoes oscillatory dynamics ( at @xmath130 ) with the presence of a stable epidemic cycle associated with a supercritical hopf bifurcation ( see the yellow region ) .
as @xmath16 grows further , the system enters another bistable region with one stable endemic state and one stable epidemic cycle ( see the dark gray region ) , where the oscillatory phenomena can still be observed . for large enough @xmath16
, the whole bistable region is divided by transcritical and hopf bifurcations into three parts : in the light gray region , one stable healthy state coexists with one stable endemic state ; in the medium gray region , two stable endemic states coexist ; and in the dark gray region , one stable endemic state coexists with one stable epidemic cycle . while the increase of @xmath16 slightly enlarges the healthy region , it greatly narrows the endemic region and widens the bistable region .
moreover , the significant role of rewiring in hindering epidemics is clear in the enlarged invasion threshold as well as the emergence of a persistence threshold . according to figs .
[ fig3 ] and [ fig4 ] , we conclude that the combination of public vaccination and local rewiring is more effective in preventing and controlling the spread of infectious diseases , although it does cause complicated ( and interesting ) dynamical behavior to occur . with average degrees of susceptible ,
infected , and vaccinated individuals @xmath131 , @xmath132 , and @xmath133 ( a ) ; and a low epidemic prevalence @xmath134 with @xmath135 , @xmath136 , and @xmath137 ( b ) . the black ( solid ) , red ( dashed ) , and blue ( dash - dotted ) lines correspond to the susceptible , infected , and vaccinated individuals , respectively .
the vaccinated individuals have a very broad degree distribution , which extends beyond the domain of the figure .
parameters are @xmath107 , @xmath108 , @xmath138 , @xmath109 , @xmath110 , @xmath111 , @xmath112 , and @xmath118 . ]
we first examine the degree distribution in the endemic steady state .
we choose parameters in the bistable region to compare the degree distributions of nodes in each class at different endemic states .
in particular , the degree distributions at a high epidemic prevalence @xmath139 and a low prevalence @xmath134 are shown in figs .
[ fig5](a ) and [ fig5](b ) , respectively .
the results are obtained by averaging over @xmath140 time steps at the steady state with @xmath141 independent simulations for the model . as shown in fig .
[ fig5 ] , the average degree of infected nodes is smaller than that of susceptible nodes and far smaller than that of vaccinated ones because of rewiring . in the closed population without demographic turnover , the breaking of _
si_-links causes the infective nodes to lose susceptible neighbours , while the susceptible and vaccinated nodes correspondingly gain more connections .
therefore , the average degree of susceptible nodes is larger than the average degree of infected nodes in the steady state . in this regard ,
the random vaccination of the susceptibles favors higher - degree nodes in the network due to rewiring @xcite , and thus is expected to be effective , as in previous studies on targeted vaccination @xcite . since the vaccine - reduced transmission rate is relatively low , vaccinated nodes accumulate connections from rewiring , therefore they have a broad degree distribution with a very large average degree [ see @xmath133 in fig .
[ fig5](a ) and @xmath137 in fig .
[ fig5](b ) ] .
there are more
_ si_-links ( at a lower level of epidemic prevalence but with a higher density of susceptible nodes ) in fig .
[ fig5](b ) than in fig .
[ fig5](a ) , thus adaptive rewiring has stronger effect on the average degree of infected nodes , which in turn reduces the average degree of susceptible nodes due to the recovery process .
it is worth remarking that the presence of the low infection and high infection states is interesting , because the original adaptive network model by @xcite just found bistability between endemic and disease - free states .
in essence , the differences between the low infection and high infection states are rooted in the different selections of target nodes towards which the original susceptible nodes will be rewired . according to the rewiring mechanism in our model ,
after susceptible nodes break away their connections from infected neighbours , the susceptibles establish new connections towards randomly selected noninfected nodes , namely , either susceptible or vaccinated .
note that susceptible nodes can be infected at a much higher rate than vaccinated nodes .
if vaccinated nodes are selected as the rewiring target , then more vaccinated nodes are able to accumulate connection from rewiring .
thus the mean degree of vaccinated nodes are much higher and they become hub nodes in the network . among the neighbours of vaccinated nodes , susceptible nodes are in the majority since they are rewired towards the vaccinated nodes . on the other hand ,
vaccinated nodes are not permitted to rewire away from infected neighbours .
once the vaccinated nodes are infected by one of their neighbours , these hub nodes will induce more infections , leading to a high prevalence in the steady state . on the contrary ,
if more susceptible nodes are selected as the rewiring target , then the susceptible nodes tend to form a cluster and the vaccinated nodes to some extent lose the ability of accumulating new connections from rewiring
. however , susceptible nodes are less likely to become hubs than vaccinated nodes . in this case
, the system will end at a low prevalence at the steady state . as a function of any given node degree @xmath27 in each class of nodes for different epidemic prevalences in the bistable endemic states [ fig .
[ fig2](b ) ] : a high epidemic prevalence @xmath139 ( a ) , and a low epidemic prevalence @xmath134 ( b ) .
the black boxes , red circles , and blue triangles represent susceptible , infected , and vaccinated nodes , respectively .
parameters are @xmath107 , @xmath108 , @xmath138 , @xmath109 , @xmath110 , @xmath111 , @xmath112 , and @xmath118 . ]
another commonly studied topological property of networks is the degree correlation ( or mixing pattern ) , which can be characterised by measuring the average degree of the nearest neighbours of nodes with degree @xmath27 @xcite , @xmath142 here , @xmath143 is the conditional probability that a link belonging to a node with degree @xmath27 points to a node with degree @xmath144 .
there is a positive ( negative ) degree correlation if @xmath145 increases ( decreases ) with @xmath27 , corresponding to the assortative ( disassortative ) mixing pattern , where high - degree nodes preferentially connect to highly ( sparsely ) connected ones @xcite . otherwise , the network is degree uncorrelated as @xmath146 is independent of @xmath27 .
similar to the work by @xcite , the original network is set to be uncorrelated where the initial configuration displays no degree correlations . after we incorporate rewiring into the networked siv model @xcite
, substantial degree correlations appear .
figure [ fig6 ] shows the average degree of the nearest neighbours @xmath145 of degree-@xmath27 nodes as a function of @xmath27 for each class of nodes . in the case of a high prevalence @xmath139 [ fig .
[ fig6](a ) ] , the network shows a positive degree correlation for small degrees . on the contrary , for larger @xmath27
the network is disassortative mixing .
similarly , in the case of a low prevalence @xmath134 [ fig .
[ fig6](b ) ] , the network is generally disassortative mixing as @xmath145 decreases with degree @xmath27 and eventually closes to @xmath147 .
the appearance of the disassortative mixing pattern possibly arises from the fact that @xmath11-links are not allowed for rewiring and that the vaccinated nodes continue to accumulate connections when susceptible nodes rewire towards them , until the vaccinated nodes are infected or turn to being susceptible again . on the other hand
, the rewiring mechanism separates the infected and susceptible nodes , and typically divides the network into an infected cluster and a susceptible cluster . note that the mean degrees of infected nodes and susceptible nodes are both much smaller than the mean degree of vaccinated nodes .
therefore , the original susceptible ( or infected ) nodes that have relatively small degrees are connected to the new susceptible ( or infected ) nodes that are transformed from vaccinated nodes with high degrees , giving rise to a negative degree correlation .
the adaptive siv model with demographic effects is expressed as a set of eqs .
( [ eq:10])([eq:18 ] ) , which can be integrated with a standard numerical integration scheme . in this paper , the fourth - order runge - kutta iteration algorithm is employed to obtain the numerical solution of the model .
the set of ordinary differential equations is of dimension @xmath148 , where @xmath149 is the maximal degree value among all nodes . with eqs .
( [ eq:10])([eq:18 ] ) and notation in tables [ table1 ] and [ table2 ] , we can explore the temporal evolution of both the epidemic and the underlying network structure .
in particular , we derive not only the time series for the number of infectious nodes and the number of _
si_-links , but also the degree distributions for each class of individuals in the steady state .
we confirm the results derived from the above equations by explicitly simulating the dynamics over two different initial networks er @xcite and sf networks .
the er random network with a poisson degree distribution is constructed by linking together each pair of nodes at a certain probability , while an uncorrelated sf random network is generated following the _ uncorrelated configuration model _ proposed by @xcite .
we start the system dynamics by initializing a small fraction @xmath150 ( @xmath151 in this paper ) of infectious seeds .
we also initially vaccinate a fraction @xmath150 of nodes to ensure the denominator of the fraction @xmath152 is not equal to zero at the beginning of the numerical calculation .
hence , the initial conditions for the whole system of eqs .
( [ eq:10])([eq:18 ] ) are given by : @xmath153 and for a small enough @xmath150 ( @xmath154 ) , @xmath155 henceforth , the parameter values are set to @xmath156 , @xmath157 , @xmath158 , @xmath159 , @xmath112 , @xmath160 , and @xmath107 , unless otherwise stated .
the degree distribution of newborns is assumed to be the same as the initial network , i.e. , @xmath161 . for er networks , we use a poisson distribution @xmath162 , with @xmath163 . for random sf networks ,
we choose a power - law distribution @xmath164 with an exponential cutoff @xmath165 , where @xmath55 is the normalisation constant , @xmath166 is the minimal degree and @xmath167 is the maximal degree . in our work
we set them to be @xmath168 , @xmath169 , and @xmath170 to ensure that the initial average degree is @xmath171 . the procedures of the computational simulation at each time step include transmission , rewiring , birth and death events , respectively . of infected nodes for different starting networks : er random network [ on the left column ,
i.e. , ( a ) , ( c ) , and ( e ) ] and random sf network [ on the right column , i.e. , ( b ) , ( d ) , and ( f ) ] , with the same average degree @xmath163 . the black solid ( thick ) lines represent the analytical results obtained from the eqs .
( [ a.1 ] ) , while the gray dashed ( thin ) lines correspond to @xmath141 independent computational simulations .
the results are obtained for different parameter values : without birth and death , @xmath172 [ see ( a ) and ( b ) ] ; with birth and death , @xmath173 [ see ( c ) and ( d ) ] ; with birth and death , @xmath174 [ see ( e ) and ( f ) ] .
the other parameters are common : @xmath157 , @xmath158 , @xmath159 , @xmath22=0.0002 , @xmath160 , and @xmath107 . ] the temporal evolution of the number of infected nodes @xmath175 is presented in fig .
[ fig7 ] , exhibiting a good agreement between the simulation results ( gray dashed lines ) and the analytical solutions ( black solid lines ) obtained from the model equations for different parameter values .
figures [ fig7](a ) and [ fig7](b ) show the time series of the number of infectives , respectively , in the er and sf networks with @xmath163 , where the birth and death processes are not allowed ( @xmath57 ) . due to the rewiring mechanism
, the analytical trajectory shows a very similar trend for both networks ; however , the simulation trajectories exhibit significant variability in terms of the time required to reach the expansion phase and the epidemic prevalence .
this is mainly due to the stochastic effect of early infections .
an initially infected low - degree node , which takes much time to infect its neighbours , will delay the onset of the expansion phase , while the initial infection of a high - degree node may lead to an explosive outbreak @xcite . on the contrary , if we introduce nonzero rates of birth and death @xmath176 , the disease will eventually die out after an initial outbreak in both er and sf networks , as shown in figs .
[ fig7](c ) and [ fig7](d ) .
this happens because the inadequate birth rate fails to offset the natural and disease - induced deaths .
again , the trajectories of infected nodes in both networks are almost the same .
finally , figs .
[ fig7](e ) and [ fig7](f ) show a series of independent simulations for the number of infectives at a larger birth rate @xmath177 and a smaller transmission rate @xmath178 in the er and sf networks , respectively . in this case
, the analytical trajectory traverses through the central region of the dense swarm of simulation trajectories , and the black thick lines in figs .
[ fig7](e ) and [ fig7](f ) demonstrate that on average the epidemic prevalence in the sf network is smaller than that in the er network ( particularly at @xmath179 ) .
the possible reason for this disparity is rooted in the heterogeneity among the node degrees . in the sf network high - degree nodes
are more likely to get infected and hence they die more often ( due to a more rapid death ) , which in turn leads to a greater reduction in the average degree of the entire network and consequently results in a lower infection level . compared with figs .
[ fig7](a ) to [ fig7](d ) , the simulation trajectories in figs .
[ fig7](e ) and [ fig7](f ) exhibit stronger variability in the epidemic prevalence . since the birth rate is greater than the natural death rate , the continuous input of newborn nodes will lead to significant changes in the total network size and the connectivity pattern of the entire network during the epidemic outbreak , and hence the epidemic prevalence . and sf networks [ on the right side : ( b ) , ( d ) , and ( f ) ] , with the same average connectivity @xmath163 .
the black ( thick ) lines indicate the theoretical results obtained by numerical integration of the eq .
( [ eq:14 ] ) , and the gray ( thin ) lines are the simulation results of at least 50 independent realizations .
the results are obtained by varying the vaccination rate @xmath7 and the rewiring rate @xmath16 [ here , we choose @xmath180 for ( a ) and ( b ) ; @xmath181 for ( c ) and ( d ) ; @xmath182 for ( e ) and ( f ) ] , while the other parameters are the same as in figs .
[ fig7](e ) and [ fig7](f ) : @xmath183 , and @xmath107 . ] and the rewiring rate @xmath16 , as indicated in the legend .
the other parameters are : @xmath183 , and @xmath107 . ] ( top ) and @xmath184 ( bottom ) in er networks [ on the left side : ( a ) and ( c ) ] and sf networks [ on the right side : ( b ) and ( d ) ] for the adaptive siv model without birth and death processes . in this case
only the theoretical results are presented .
the red dashed lines indicate the results for @xmath158 and @xmath185 ; the black solid lines for @xmath158 and @xmath160 ; and the blue dotted lines for @xmath186 and @xmath160 .
other parameters are @xmath156 , @xmath157 , @xmath159 , @xmath112 , @xmath57 . ] to understand in more detail how the epidemic spreading is influenced by the adaptive rewiring behavior of susceptible individuals and the random vaccination process , we further investigate the temporal evolution of the number of contacts between susceptible and infected nodes , as shown in fig .
[ fig8 ] . for the parameter values ( @xmath158 and @xmath160 ) given in figs .
[ fig7](e ) and [ fig7](f ) , the number of _
si_-links , @xmath184 , in both er [ fig .
[ fig8](c ) ] and sf [ fig .
[ fig8](d ) ] networks is shown to have the same qualitative behavior as the number of infected nodes , @xmath175 .
si_-links are the major infectious contacts since the probability @xmath10 for vaccinated individuals to be infected is extremely limited ( @xmath112 ) . in this regard
, the number @xmath184 typically determines the number @xmath175 . to show a clear comparison between the effects of public vaccination and adaptive rewiring on the number of @xmath6-links , we present the average simulation results in fig .
[ fig9 ] , in which the number @xmath184 of @xmath6-links in the er network increases as the rewiring rate is raised from @xmath185 to @xmath160 with the vaccination rate remaining the same ( @xmath158 ) .
still , the number @xmath184 is further enlarged as @xmath7 is increased from @xmath158 to @xmath186 , with @xmath16 keeping unchanged ( @xmath160 ) .
this implies that as the parameters @xmath7 and @xmath16 are both increased from @xmath158 , @xmath185 to @xmath186 , @xmath160 , the increment of the number @xmath184 is larger than the increment of @xmath184 caused by increasing either @xmath7 or @xmath16 alone . in this regard , we conclude that public vaccination , in conjunction with adaptive rewiring , facilitates disease spread . in combination with the birth and death processes , lowering the rewiring rate causes @xmath184 to decrease , whereas enhancing the vaccination rate increases @xmath184 .
this case is in sharp contrast to the model without demographic processes , in which both the rewiring and vaccination contribute to decrease @xmath184 and hence @xmath175 , as shown in fig .
[ fig10 ] .
these counterintuitive results imply that in the case of a varying population with demographic effects ( births , deaths and migration ) , susceptible individuals protective behavior and the public vaccination interventions that aim to suppress the epidemic , would make the situation even worse .
these phenomena can be explained as follows .
intuitively , infected nodes die at a more rapid rate since @xmath71 . when the rewiring rate is increased , the infected nodes have lower degree when they die .
therefore , the average degree of the entire network is not reduced as much by node death as it would be in the absence of rewiring .
higher mean degree could lead to more infections .
when the vaccination rate is increased , then there are more vaccinated nodes ( that are able to accumulate connections from rewiring and thus have high degrees ) serving as the hub nodes in the network . on the other hand
, the adaptive rewiring of @xmath6-links , in conjunction with the birth and death processes , contributes to a structural change in the network .
typically , rewiring tends to increase clustering @xmath6-links are broken and then susceptible nodes are forced to select new neighbours from among the more limited subset of noninfected nodes ( which will be in a topologically constrained part of the network ) .
imagine an infected subgraph , the susceptible nodes at the boundary are rewiring their connection to other noninfected nodes elsewhere increasing clustering .
clustering ( particularly with negative correlation between degrees ) means that once infection reaches these regions ( and the hub nodes in particular ) one would expect worse outbreak . as a matter of fact
, infection is facilitated in reaching these regions with the addition of new nodes via the birth and death processes ( since they are connected randomly across the existing structure ) .
moreover , it is shown from fig .
[ fig9 ] that the number @xmath184 in sf networks expands faster but is more restricted than in er networks .
the possible reason for this difference is rooted in the heterogeneity among the node degrees . in the sf network high - degree nodes
are more likely to get infected and hence they die more often ( due to a more rapid death rate ) , which in turn leads to a greater reduction in the average degree of the entire network and consequently results in a lower infection level . .
the results on the left ( right ) column are obtained from an initial er ( sf ) network , with an average degree @xmath163 .
the black boxes , red circles and blue triangles stand for susceptible , infected , and vaccinated individuals , respectively .
the symbols correspond to simulation results averaged over 100 different realizations and the solid lines are numerical solutions obtained from eqs .
( [ eq:10])([eq:12 ] ) and ( [ a.1 ] ) .
the parameters in the panels ( a ) and ( b ) are the same as in figs .
[ fig7](a ) and [ fig7](b ) , while those in the panels ( c ) and ( d ) are the same as in figs . [ fig7](e ) and [ fig7](f ) . ] as a function of any given degree @xmath27 in each class of nodes in the state at @xmath179 .
the results on the left ( right ) column are obtained from an initial er ( sf ) network , with an average degree @xmath163 . the empty boxes ( black ) @xmath187 " , circles ( red ) @xmath188 " and triangles ( blue )
@xmath189 " represent susceptible , infected and vaccinated nodes , respectively , and the crosses ( olive )
@xmath190 " denote the results of the entire network .
all these symbols are simulation results obtained by averaging over 100 different realizations .
the corresponding parameter values in each diagram are the same as in each diagram of fig .
[ fig8 ] . ] furthermore , our model is able to capture the topological structure of the network in the stationary state of the epidemic . in particular ,
the degree distribution in each class of nodes is shown in fig .
[ fig11 ] , where the black , red , and blue colours correspond to the susceptible , infected , and vaccinated nodes , respectively . to be precise
, the symbols represent the simulation results and the solid lines are the analytical solutions to system equations . as shown in figs .
[ fig11](a ) and [ fig11](b ) , without the birth and death processes , the network evolves to the same degree distribution due to the rewiring mechanism . on the contrary ,
the situation is quite different with the introduction of birth and death processes .
for instance , in figs .
[ fig11](c ) and [ fig11](d ) we exhibit the degree distribution of each class of nodes at the final time step @xmath179 , for the settings in figs . [ fig7](e ) and [ fig7](f ) , respectively . in the presence of node birth , different degree distributions of the newborns will lead to different network structures .
it is clear that the final network is narrowly distributed if nodes are born with a poisson degree distribution [ fig .
[ fig11](c ) ] , while the network evolves to a broader degree distribution if the newborns follow a power - law degree distribution [ fig .
[ fig11](d ) ] .
it is worth noting that by the mathematical model of eqs .
( [ eq:10])([eq:18 ] ) , the theoretical results of degree distribution can be obtained not only for the model with demography , but also for the case without demography where the birth and death rates are set to zero , as illustrated in the figs .
[ fig11](a ) and [ fig11](b ) . nevertheless , under the modelling framework of eqs .
( [ eq:1])([eq:9 ] ) where network properties are not included , the theoretical results of degree distribution can not be derived . as mentioned above ,
since the birth process dominates in affecting the degree distribution of nodes , the network structure is degree uncorrelated , which is shown in fig .
[ fig12 ] .
although the individual rewiring changes the local connectivity pattern among nodes , the continuous input of the newborns that randomly link to old ones breaks the degree correlation that emerges in the scenario without birth and death processes ( fig .
[ fig6 ] ) .
note that the average nearest - neighbour degree is close to @xmath191 , @xmath192 , which is larger than the average degree of the network @xmath147 , which is around @xmath193 in our simulations .
this is consistent with the friendship paradox .
similar results are observed for different parameter values among each class of nodes in both er and sf networks ( fig . [ fig12 ] ) .
it is worth noting that although newborn nodes arrive with mean degree @xmath194 , it will not still be the mean degree of the network due to a more rapid death of infectives .
actually , the mean degree of the entire network is parameter dependent .
widespread public vaccination programmes and individual protective behavior are two key factors of disease control . ignoring either component will reduce the accuracy of our models or , ultimately , the effectiveness of containment measures . in this paper
, we developed a mathematical framework to model the combined effects of public vaccination intervention and individual protective behavior on epidemic dynamics by studying an sis model with imperfect vaccination in a dynamic contact network .
the model was analyzed in a closed population without demographic turnover and a varying population with births , deaths and migration . in the case of a closed population ( without demographic effects ) , we focused specifically on the combined effects of macroscopic random vaccination and microscopic contact rewiring on the system dynamics by conducting a bifurcation analysis of the model .
our analysis shows that compared with either individual strategy , the combination of vaccination programme and adaptive rewiring contributes to a more effective control for the spread of infectious diseases by increasing both the invasion threshold and the persistence threshold of the epidemic .
the larger the vaccination and rewiring rates are , the safer the population is from infectious diseases .
the occurrence of stable epidemic cycle at a low prevalence is particularly relevant for the seasonality of influenza epidemics @xcite .
moreover , the conjunction of vaccination and rewiring has significant impacts on the topology of the underlying contact network .
both the degree distributions and degree correlations among different classes of nodes are characteristically distinct . in the case of a varying population with demographic development ( births , deaths and migration ) , we studied the epidemic patterns and the network topology on the basis of the dynamical behavior of nodes and links at a microscopic level , using both the probability generating function and the pairwise approximation .
this modelling approach allows us to incorporate the effects of vaccination and rewiring as well as the demographic changes into network - based epidemiological models with arbitrary degree distributions .
the analytical results of the temporal evolution of both the number of infective individuals and the number of infectious contacts , as well as the degree distribution of nodes in each class , have shown a good agreement with extensive computational simulations . in general , diseases in sf networks spread relatively faster yet end at a lower level of epidemic prevalence , as compared with the er networks .
ironically , the combination of contact rewiring and random vaccination aiming to contain the epidemic would probably facilitate the disease spread . while the network without demography evolves to an identical topology for different initial network configurations , the network with demography develops into different structures depending on the degree distribution of the newborns .
besides , the underlying network structure displays no degree correlation regardless of model parameters and initial settings . in summary
, the present work manifests the significant combined effects of public vaccination intervention and individual protective behavior on both the epidemic dynamics and the underlying network structure .
moreover , the strikingly different results for the two cases of population assumptions indicate the strong impacts of the demographic turnover .
thus , this research provides a more comprehensive insight into disease containment strategies .
the authors would like to thank the two anonymous referees for their valuable comments and suggestions which have greatly improved this paper .
this work was supported by nsfc grants ( 11331009 , 11501340 ) and stcsm grant ( 13zr1416800 ) .
ms was supported by an australian research council future fellowship ( ft 110100896 ) and discovery project ( dp 140100203 ) .
summing eqs .
( [ eq:10])([eq:12 ] ) over all @xmath27 values leads to the differential equations for the total number of susceptibles @xmath89 , infecteds @xmath90 , and vaccinateds @xmath91 as follows : @xmath195 as a consequence , the total network size @xmath196 varies in time such that @xmath197 which reflects the demographic change .
furthermore , the dynamical equations for the total degrees of nodes in each state can be derived as : @xmath198,\\ \frac{\rm d}{{\rm d}t}m_{i}&=&\sum_{k}k(\frac{\rm d}{{\rm d}t}i_{k } ) \\ & = & \eta_{1}\bar{g}^{\prime}(1)n_{i}-2\mu m_{ii}-\omega m_{si } + \alpha\frac{m_{si}}{m_{s}}n_{s}[g^{\prime}_{s}(1,t ) + g^{\prime\prime}_{s}(1,t)]\\ & & + \delta\alpha\frac{m_{iv}}{m_{v}}n_{v}[g^{\prime}_{v}(1,t ) + g^{\prime\prime}_{v}(1,t)]-(2\eta_{2}+\mu+\beta)m_{i},\\ \frac{\rm d}{{\rm d}t}m_{v}&=&\sum_{k}k(\frac{\rm d}{{\rm d}t}v_{k})\\ & = & \eta_{1}\bar{g}^{\prime}(1)n_{v}+\omega m_{si}\frac{n_{v}}{n_{s}+n_{v } } + \varphi m_{s}-(2\eta_{2}+\psi)m_{v}\\ & & -\mu m_{iv}-\delta\alpha\frac{m_{iv}}{m_{v}}n_{v}[g^{\prime}_{v}(1,t)+g^{\prime\prime}_v(1,t ) ] , \end{array}\right\}\ ] ] which yield the evolution equation for the total number @xmath59 of links @xmath199 note that @xmath200 and @xmath201 are the first - order and the second - order derivatives of the generating function @xmath202 , respectively , with regard to variable @xmath41 , where @xmath203 ( see also table [ table2 ] ) . in particular , one gets @xmath204
employing the technique of generating functions used in @xcite , we generalise the standard moment closure approximation in @xcite for the number of triples accountable for new infections to incorporate the heterogeneity in the central node s degree . for example , consider an @xmath6-link connecting an infective with a susceptible , the probability for the susceptible to have degree @xmath27 is @xmath205 , and among the remaining @xmath206 neighbours , there is a fraction @xmath207 of infectious individuals . here
, we assume that the probability for a link starting from a susceptible to point to an infective is independent of the starting node s degree @xmath27 .
therefore , from the mean - field perspective the average number @xmath208 of @xmath209-type triples is @xmath210 analogously , one can get the number @xmath211 of @xmath68-type triples @xmath212 and the number @xmath213 of @xmath214-type , @xmath215 of @xmath216-type , @xmath217 of @xmath218-type , and @xmath219 of @xmath220-type triples @xmath221 where the generating functions are given in ( [ a.5 ] ) .
doedel ej , champneys ar , fairgrieve t , kuznetsov y , oldeman b , paffenroth r , sandstede b , wang x , zhang c ( 2007 ) auto-07p : continuation and bifurcation software for ordinary differential equations .
url : http://www.macs.hw.ac.uk/~gabriel / auto07/auto.html[http://www.macs.hw.ac.uk/.17exgabriel / auto07/auto.html ] fenichel ep , castillo - chavez c , ceddia mg , chowell g , parra pag , hickling gj , holloway g , horan r , morin b , perrings c , springborn m , velazquez l , villalobos c ( 2011 ) adaptive human behavior in epidemiological models .
proc natl acad sci usa 108:6306 - 6311 | in the face of serious infectious diseases , governments endeavour to implement containment measures such as public vaccination at a macroscopic level .
meanwhile , individuals tend to protect themselves by avoiding contacts with infections at a microscopic level . however , a comprehensive understanding of how such combined strategy influences epidemic dynamics is still lacking .
we study a susceptible - infected - susceptible epidemic model with imperfect vaccination on dynamic contact networks , where the macroscopic intervention is represented by random vaccination of the population and the microscopic protection is characterised by susceptible individuals rewiring contacts from infective neighbours . in particular , the model is formulated both in populations without and then with demographic effects ( births , deaths , and migration ) . using the pairwise approximation and the probability generating function approach , we investigate both dynamics of the epidemic and the underlying network . for populations without demography , the emerging degree correlations , bistable states , and oscillations demonstrate the combined effects of the public vaccination program and individual protective behavior . compared to either strategy in isolation , the combination of public vaccination and individual protection is more effective in preventing and controlling the spread of infectious diseases by increasing both the invasion threshold and the persistence threshold . for populations with additional demographic factors ,
we investigate temporal evolution of infected individuals and infectious contacts , as well as degree distributions of nodes in each class .
it is found that the disease spreads faster but is more restricted in scale - free networks than in the erds - rnyi ones .
the integration between vaccination intervention and individual rewiring may promote epidemic spreading due to the birth effect .
moreover , the degree distributions of both networks in the steady state is closely related to the degree distribution of newborns , which leads to uncorrelated connectivity .
all the results demonstrate the importance of both local protection and global intervention , as well as the demographic effects .
our work thus offers a more comprehensive description of disease containment .
= 1 |
_ mancala _ is a family of traditional african games played in many versions and under many names .
it comprises a circular list of _ holes _ containing zero or more _ seeds_. a move consists of selecting a nonempty hole , taking all its seeds and _ sowing _ them in the subsequent holes , one seed per hole .
here we study an idealized version of this called _ open mancala _ ( see @xcite , where the game is called _ open owari _ ) .
we assume there is an infinite sequence of holes .
further we assume that the nonempty holes are consecutive ; such a configuration is described by a sequence of positive integers @xmath0 , @xmath1 , giving the number of seeds in each nonempty hole , starting from the leftmost one .
we assume a move always selects the leftmost nonempty hole , sowing to the right .
thus the leftmost nonempty hole always advances of one position .
it is clear that the sequence of configurations becomes periodic in a finite time .
the periodic configurations have been completely classified ( @xcite ) : see section [ sec : notations ] .
this game is related to a card game called _ bulgarian solitaire _ , discussed by gardner @xcite in a 1983 _ scientific american _ column .
indeed , that game is isomorphic to the case of open mancala where @xmath2 for @xmath3 ( monotone mancala ) , see section [ sec : bulgarian ] .
gardner mentioned a conjecture on the maximum number of moves before the onset of periodicity when the number of cards is of the form @xmath4 .
the conjecture was proved independently a few years later by igusa @xcite and etienne @xcite . in 1998 ,
griggs & ho @xcite revisited the result and proposed a new conjecture in the case of a generic number of cards . due to the isomorphism between monotone mancala and bulgarian solitaire , our result is strongly related to this conjecture .
however , our results apply to open mancala , not to the monotone variant , hence the conjecture about the bulgarian solitaire remains unproven .
the main result of the paper is the optimal bound for the maximum number of moves , before open mancala reaches periodicity , as a function of @xmath5 , the number of seeds . in section
[ sec : lower ] we prove the lower bound by producing , for any @xmath5 , a configuration that requires exactly that number of moves . in section [ sec : smon ] we introduce the main tool in proving the upper bound , namely _
@xmath6-monotonicity_. in section [ sec : upper ] we prove that every configuration reaches periodicity within that number of moves .
we denote by @xmath7 the set of nonnegative integers and by @xmath8 the set of strictly positive integers . by @xmath9 , @xmath10 ,
we denote the @xmath11-th triangular number @xmath12 where in particular @xmath13 . a _ configuration _
@xmath14 is a sequence of nonnegative numbers @xmath0 , @xmath15 , that is @xmath16 .
the _ support _ @xmath17 of a configuration @xmath14 is the set of indices @xmath18 . a _ mancala configuration _
@xmath14 is a configuration having a _ connected _ support of the form @xmath19 for some @xmath20 .
the index @xmath21 is called the _ length _ of the configuration . in the special case of the _ zero _ configuration , we shall conventionally define its length to be zero .
we shall denote by @xmath22 the set of all mancala configurations . if @xmath23 , its _ mass _ is the number @xmath24 we denote by @xmath25 the set of all mancala configurations of mass @xmath5 .
the _ mancala _ game , in our setting , is a discrete dynamical system associated with a function @xmath26 mapping the space of _ mancala _ configurations @xmath22 in itself .
[ def : move ] the _ sowing _ of a mancala configuration @xmath14 is an operation on @xmath14 that results in a new configuration @xmath27 defined as follows : @xmath28 conventionally , @xmath29 maps the empty configuration into itself .
it is clear that @xmath29 preserves the mass of a configuration , so that it can be restricted to @xmath25 .
it is convenient to define the ( right ) _ shift _ operator @xmath30 , that acts on generic sequences simply by a change in the indices . [ def : shift ] if @xmath31 is a sequence , the sequence @xmath32 is defined by @xmath33 clearly , the result of the shift operator is never a mancala configuration ( with the exception of the empty configuration ) .
both @xmath29 and @xmath30 can be iterated , the symbol @xmath34 ( resp .
@xmath35 ) denoting the result of @xmath11 repeated applications of @xmath29 ( resp . of @xmath30 ) .
the shift operator has a left inverse @xmath36 , defined by @xmath37 and satisfying @xmath38 for any configuration @xmath14 .
we say that @xmath39 if @xmath40 for all @xmath41 .
moreover we say that @xmath42 if @xmath43 and @xmath44 .
if @xmath14 , @xmath45 are two integer sequences ( in particular if any of them is a configuration in @xmath22 ) , we define the sum and difference @xmath46 componentwise : @xmath47 .
it is easy to check that if @xmath14 , @xmath48 and @xmath39 , then @xmath49 .
[ defn : monotone ] a configuration is called _ monotone _ if it is weakly decreasing , i.e. if @xmath2 whenever @xmath3 .
if @xmath14 is a monotone configuration , then so it is @xmath50 .
moreover , if @xmath23 has length @xmath51 , then @xmath52 is monotone .
the following special configurations play an important role .
[ def : mg ] for a given @xmath10 we define the special mancala configuration @xmath53 with mass @xmath9 and length @xmath11 , called _ _ marching group _ _ of order @xmath11 , as follows : @xmath54 [ def : amg ] a mancala configuration @xmath14 such that @xmath55 for some @xmath10 is called an _ augmented marching group _ of order @xmath11 .
it satisfies @xmath56 .
notice that a marching group is a particular case of an augmented marching group .
the following important theorem about augmented marching groups is readily proved .
see ( * ? ? ?
* theorems 1 and 2 ) .
[ teo : periodic ] augmented marching groups are the _ only _ periodic configurations for @xmath29 and , if @xmath57 , then the period of @xmath14 is a divisor of @xmath58 .
marching groups are the _ only _ fixed points for @xmath29 .
a remarkable fact about open mancala is that every configuration becomes an augmented marching group , and hence periodic , after a finite number of moves .
the paper is devoted to find an optimal bound on such a number of moves which depends only on the mass of the initial configuration .
[ def : depthdiameter ] the _ depth _ of a configuration @xmath14 is its distance from the periodic configuration , _ i.e. _ the number of moves needed to reach periodicity . for @xmath59
we call the _ depth of @xmath25 _ , denoted by @xmath60 , the maximal depth of all configurations @xmath14 with @xmath61 .
the _ diameter _ of a configuration is the number of moves before the first repetition .
it is equal to its depth plus the length of the period minus one .
the maximal diameter of all configurations of mass @xmath5 is called the _ diameter of @xmath25_. clearly if @xmath62 the diameter of @xmath25 equals its depth @xmath60 , whereas if @xmath63 , @xmath64 the diameter of @xmath25 is bounded above by @xmath65 .
, more precisely a lower bound is given by @xmath66 where @xmath67 is the smallest divisor of @xmath58 larger than @xmath68 . ]
the number @xmath60 can be seen also as the depth of a graph .
indeed , for a given @xmath69 we can construct the directed graph @xmath70 having the configurations in @xmath25 as nodes and an arc from @xmath14 to @xmath71 whenever @xmath72 .
the graph @xmath70 contains exactly @xmath73 nodes .
cycle _ of @xmath70 correspond to sequences of moves that repeat periodically .
a configuration is periodic if it belongs to a cycle . in the case
@xmath74 , if we remove all arcs connecting two periodic configurations ( arcs that belong to a cycle ) , the remaining graph @xmath75 is a disjoint union of trees rooted at a periodic configuration , and in each tree all arcs point toward the root . in this
setting , @xmath60 represents the depth of @xmath70 , _
i.e. _ the maximal depth of the trees of @xmath75 .
we can view the mancala configurations and mancala moves in a way reminiscent of the energy levels for the electrons in an atom .
let us consider the subset @xmath76 given by @xmath77 it is convenient to think of the integers @xmath78 and @xmath79 as a numbering of square cells instead of coordinates of points , in a way similar to the familiar sea - battle game .
each square cell in @xmath80 can either be empty or contain a single seed .
the _ column _ index @xmath78 corresponds to one of the holes in the mancala game , so that @xmath0 , where @xmath81 , is the total number of occupied cells in column @xmath78 of @xmath80 .
we shall also associate an _ energy _ to a seed positioned in cell @xmath82 , given by its second coordinate @xmath79 .
the @xmath11-th level of @xmath80 is the set @xmath83 of cells having energy @xmath11 .
finally we let the seeds free to immediately _ fall _ down without changing their column , but decreasing their energy ( the second coordinate of their current position ) provided the new position ( and all those in between ) are free .
the mancala configuration @xmath23 then corresponds to a positioning of seeds in @xmath80 such that @xmath82 is occupied by a seed if and only if @xmath84 ( see figure [ fig : energyexample ] ) .
( 0,0 ) ; in 0, ... ,4 ( 0,0 ) ( 1,0 )
( 1,1 ) ; in 0, ... ,2 in 0 , ... , ( 0.5,0.5 ) circle ( 0.4 ) ; in 0, ... ,2 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,0 ) ( 0,0 ) node[right , xshift=2cm , yshift=0.25cm]@xmath85 ; ( 0,5 ) ( 0,0 ) ; in 0, ... ,4 ( 0,0 )
( 1,1 ) ; in 0, ... ,2 in 0 , ... , ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,0 ) ( 0,0 ) node[right , xshift=2cm , yshift=0.25cm]@xmath86 ; ( 0,5 ) ( 0,0 ) ; in 0, ... ,4 ( 0,0 ) ( 1,0 ) ( 1,1 ) ; ( 0,0 ) ( 0,0 ) node[right , xshift=0cm , yshift=-0.4cm]@xmath87-th level ; in 0, ... ,2 in 0 , ... , ( 0.5,0.5 ) circle ( 0.4 ) ; in 0, ... ,2 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,0 ) ( 0,0 ) node[right , xshift=2cm , yshift=0.25cm]@xmath88 ; ( 0,5 ) ( 0,0 ) ; in 0, ... ,4 ( 0,0 ) ( 1,0 ) ( 1,1 ) ; in 0, ... ,2 in 0 , ... , ( 0.5,0.5 ) circle ( 0.4 ) ; in 0, ... ,2 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,0 )
( 0,0 ) node[right , xshift=2cm , yshift=0.25cm]@xmath89 ; ( 0,5 )
( 0,0 ) ; in 0, ... ,4 ( 0,0 ) ( 1,0 )
( 1,1 ) ; in 0, ... ,2 in 0 , ... , ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,0 )
( 0,0 ) node[right , xshift=2cm , yshift=0.25cm]@xmath90 ; ( 0,5 )
( 0,0 ) ; in 0, ... ,4 ( 0,0 ) ( 1,0 )
( 1,1 ) ; ( 0,0 )
( 0,0 ) node[right , xshift=0cm , yshift=-0.4cm]@xmath87-th level ; in 0, ... ,2 in 0 , ... , ( 0.5,0.5 ) circle ( 0.4 ) ; in 0, ... ,2 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 3.5,4.5 ) ( 3.5,3.5 ) ; ( 0,0 ) ( 0,0 ) node[right , xshift=2cm , yshift=0.25cm]@xmath91 ; a _ mancala move _ can now be viewed in this setting as follows .
firstly , each seed is moved one position to the left without changing its energy level , while seeds in column @xmath68 are _ rotated _ to the last column of that energy level .
namely , a seed at position @xmath82 is moved in the new position @xmath92 if @xmath93 , @xmath94 if @xmath95 .
this is just a left rotation of the seeds of each level , so that cells will receive at most one seed .
secondly we let gravity pack all the seeds in each column at the lowest energy possible .
we notice that if all energy levels up to @xmath11 are completely filled , then a mancala move does not change those levels ( up to a rearrangement of the seeds ) , and all the action takes place from the lowest level that is not completely filled , say @xmath58 , up to the highest level that is not completely empty .
lowest active level _ is the lowest level that is not completely filled .
the overall energy of a configuration ( the sum of the individual energy of every seed ) does not change after the first part of the move ( the rotation within each energy level ) whereas it can decrease during the second part ( vertical downward shift ) .
[ def : gap ] we shall call _ gap _ an empty cell at the lowest energy level , say @xmath11 , that is not completely filled
. we shall think of a gap as an _ absent seed _ , and in this respect we can _ follow _ a gap through mancala moves as it rotates left along its energy level .
a gap can be filled when it receives a seed that falls from above after a move , and this can happen only when the gap is in column @xmath11 .
the filling of a gap at the lowest energy level that is not completely filled is permanent , since there is no possibility of further energy decrease for seeds at that energy level .
all gaps will be eventually filled , provided there are sufficiently many seeds at energy levels larger than @xmath11 .
in order to prove a good lower bound @xmath96 for the depth @xmath60 we need to find appropriate configurations and to compute the number of moves required to reach a periodic configuration ( which is an augmented marching group ) .
let @xmath63 with @xmath97 .
@xmath98-biaugmented marching group _ is a mancala configuration @xmath99 of the form @xmath100 where @xmath53 is the marching group of order @xmath11 and @xmath101 is an increment of the form @xmath102 for some @xmath103 .
the second ( resp .
fourth ) row in the definition is void if @xmath104 ( resp . if @xmath105 ) .
clearly such a configuration is not an augmented marching group .
suitable choices of @xmath98 and @xmath106 ( compatible with the definition above ) allow to obtain all the values of @xmath107 such that @xmath108 .
[ prop : biaugmented1 ] a @xmath98-biaugmented marching group has depth @xmath109 , _
i.e. _ it becomes an augmented marching group exactly after @xmath109 moves . if @xmath104 , a direct check shows that after two moves we obtain an augmented marching group .
if @xmath110 , let us _ color _ the seedthe idea of
the colored seed is due to bouchet ( * ? ? ?
* section 3 ) ] in the first hole which is at the energy level @xmath111 ( the highest energy level ) ; we agree that the colored seed is the last to be sown in the mancala move , so that it moves at the energy level @xmath111 until it falls on the level @xmath58 , when the configuration becomes an augmented marching group .
notice that the lowest active level is @xmath58 .
the colored seed is again at the leftmost position after exactly @xmath111 moves , while @xmath58 is the period of the seeds moving in the lowest active level .
hence the first `` hole '' at level @xmath58 , which were in position @xmath112 at the beginning , gets one step closer to the colored seed every @xmath111 moves . at the end ,
the configuration becomes a @xmath113-biaugmented marching group in @xmath114 moves , and in two further moves it becomes periodic .
as an example , figure [ fig : energyexample ] shows the energy - level interpretation of a @xmath115-biaugmented marching group with @xmath116 and its evolution .
the gray circle in the pictures is the colored seed of the proof .
let @xmath117 with @xmath118 .
@xmath98-biaugmented marching group of the second kind _ is a configuration @xmath99 of the form @xmath119 , where @xmath120 is the marching group of order @xmath121 and @xmath101 is an increment of the form @xmath122 for some @xmath123 .
again , such a configuration is not an augmented marching group .
the depth of a @xmath98-biaugmented marching group of the second kind is given by @xmath124 , _
i.e. _ it becomes an augmented marching group precisely after @xmath125 moves .
let us color red the seeds which are at the beginning in the energy level @xmath58 , and blue the ones in the energy level @xmath11 , which is the lowest active level .
as in the proof of the previous proposition , we agree that the colored seeds are the last to be sown in every mancala move , the red seed ( if present ) being sown after the blue one . in that way , the blue seeds remain forever at the level @xmath11 , while one of the red seeds will eventually lose one energy level , when the configuration becomes an augmented marching group .
we track the evolution of the colored seeds , which move in the two highest energy levels .
the pattern of red seeds rotates with a period @xmath58 , while the pattern of blue seeds rotate with a period of @xmath11 ; hence the pattern of red seeds slowly slides one position to the right with respect to the pattern of blue seeds every @xmath11 moves , and the value of @xmath98 decreases by one .
when @xmath104 , it is easy to see that after @xmath11 moves a red seed reaches the lower level .
hence the initial configuration takes exactly @xmath125 moves in order to reach periodicity .
we define the _ heaviside configuration _ @xmath126 by @xmath127 we shall also study the special configuration @xmath128 with a single seed in all holes @xmath15 : although @xmath128 is not in @xmath22 since it has unbounded support , the sowing @xmath29 is still well - defined . [
theorem : ones ] let @xmath129 and @xmath130 with @xmath131 .
then the configuration after @xmath132 moves starting from @xmath128 is given by @xmath133 it is easy to check that exactly at every move @xmath134 the configuration becomes greater than the marching group @xmath135 , hence we have the result for @xmath136 . in particular , for any @xmath137 the leftmost element of the sequence at time @xmath138 is given by @xmath139 .
hence in the subsequent move we have to add one seed in the holes with indices @xmath140 and then shift all piles of one position to the left ( dropping the leftmost ) .
if @xmath141 we get @xmath142 and the result follows by induction . if @xmath143 , then @xmath144 and we have to add one seed to the piles with indices @xmath140 , which contain the values @xmath145 , to begin with .
again the result follows immediately .
[ rem : truncatedheaviside ] notice that at time @xmath146 , @xmath131 we moved at most @xmath11 seeds ( @xmath147 , so that we do not touch the piles at position @xmath148 , which came from @xmath149 shifts to the left .
this means that proposition [ theorem : ones ] holds for a starting configuration @xmath150 with @xmath63 ( we denote it by ` 1^n ` ) until time @xmath149 , with truncation to zero for elements with index @xmath148 .
moreover , the number of seeds in the leftmost pile ( index @xmath151 ) is exactly @xmath11 at times between @xmath152 and @xmath153 and is @xmath58 at time @xmath154 if @xmath155 .
[ rem : truncatedheaviside2 ] in the special case @xmath156 ( @xmath157 ) we can exactly describe the resulting configuration at time @xmath158 .
this is done by observing that the number of seeds in the first pile ( remark [ rem : truncatedheaviside ] ) is the same as for the full @xmath128 sequence up to time @xmath9 , meaning that the sowing process is exactly the same .
we can thus recover the resulting sequence at time @xmath9 by subtracting from the missing seeds in their expected position ( @xmath159 ) obtaining @xmath160 or equivalently @xmath161 where @xmath162 is the sequence with one in position 1 and zero elsewhere .
now we want to analyze the evolution of the augmented sequence obtained from @xmath150 by adding a seed at the position with index @xmath67 .
we can write the sequence as @xmath163 where @xmath164 denotes the sequence with a @xmath68 at the position @xmath67 and zero elsewhere . the analysis can be conveniently done by using again bouchet s trick , see footnote [ fn : bouchet ] . in this way
the evolution of the augmented configuration coincides exactly with the evolution of the non - augmented configuration with the addition of the colored seed , to be positioned appropriately . at each sowing the colored seed moves one position to the left ; when it reaches the leftmost pile ( index @xmath151 ) , at the next move its new position depends on how large the pile is .
we however know precisely how the heaviside sequence evolves , so that we can explicitly predict the position of the colored seed at each time .
we are particularly interested in the case where @xmath156 and @xmath165 with @xmath166 .
[ teo : augmentedheaviside ] the _ colored _ seed in the evolution of the augmented heaviside sequence @xmath167 with @xmath156 and @xmath168 , @xmath166 is located at position @xmath169 at time @xmath158 .
after @xmath170 moves the colored seed reaches the leftmost pile which , according to proposition [ theorem : ones ] , will contain also @xmath171 noncolored seeds .
this is true even in the special case @xmath172 , where the colored seed is added at the rightmost nonempty pile of the sequence @xmath150 , see remark [ rem : truncatedheaviside ] .
this covers the case @xmath172 .
if @xmath173 , at time @xmath174 the colored seed will go into the hole with index @xmath175 and will be again into the leftmost pile at time @xmath176 . if @xmath177 we can again resort to proposition [ theorem : ones ] and conclude that the leftmost pile will contain the colored seed and @xmath175 noncolored seeds at that time .
this argument can be repeated for @xmath98 cycles as long as @xmath178 with the leftmost pile containing @xmath179 noncolored seeds at time @xmath180 . by taking @xmath181 , the largest admissible value , we have the colored seed in the leftmost pile at time @xmath182 together with @xmath11 noncolored seed . at the next time
@xmath183 the colored seed will be sown in position @xmath58 and will move left of one position in the next @xmath184 moves so that at time @xmath185 it will be in the hole at position @xmath186 , which concludes the proof .
the method of coloring seeds can be also used with more than one seed , provided that we do not place two or more colored seeds in the same hole and as long as the colored seeds do not interact ( staying all in different piles ) . thanks to lemma [ teo : augmentedheaviside ]
we indeed have noninteracting colored seeds for @xmath9 moves if we place them in positions with indices of the form @xmath174 for a set of values @xmath166 : if two of them would land in the same pile during the evolution , then the expected position given by lemma [ teo : augmentedheaviside ] would be the same , which is not the case .
we can thus exactly predict their position at time @xmath158 , there will be a colored seed in position @xmath6 , if and only if there was a colored seed in position @xmath187 .
if this is true for @xmath188 ( with the addition of @xmath189 colored seeds ) , recalling remark [ rem : truncatedheaviside2 ] we obtain a @xmath98-biaugmented marching group . by proposition [ prop : biaugmented1 ] ,
the biaugmented marching group will become periodic after exactly @xmath109 further moves .
we want to construct , for a given @xmath63 , a configuration that takes as long as possible to become periodic . in view of the previous discussion to achieve this goal it is mostly convenient to place the colored @xmath107 seeds in the rightmost positions with index of the form @xmath174 .
we call _ augmented heaviside sequence _ the configuration of mass @xmath63 , @xmath190 , and length @xmath191 , which is obtained from @xmath192 by adding a seed in the rightmost @xmath193 positions having index of the form @xmath174 ( in particular there will be no added seeds if @xmath194 ) . as an example , the augmented heaviside sequence with mass @xmath195 is given by 1 1 1 1 1 1 2 1 1 1 2 1 1 1 1 2 .
the preceding arguments give a proof of the following [ teo : lower1 ] let @xmath63 with @xmath97 . then the corresponding augmented heaviside sequence becomes an @xmath196-biaugmented marching group after @xmath9 moves .
in particular , its evolution becomes periodic after exactly @xmath197 moves with @xmath198 the choice @xmath199 is a limit case corresponding to @xmath200 , and we have @xmath201 .
moreover if we write @xmath202 with @xmath203 , then @xmath204 and @xmath197 in can be equivalently written as @xmath205 .
direct inspection allows to show that for @xmath206 the particular biaugmented marching group at which we arrive at time @xmath9 evolves into an augmented marching group having period of length exactly @xmath58 ( and not a proper divisor ) .
hence the diameter of the evolution is @xmath207 .
we now consider the augmented heaviside sequence corresponding to @xmath208 : it has length @xmath58 , contains an increment at all the positions of the form @xmath174 , @xmath166 and the total number @xmath5 of seeds is itself a triangular number . in the rightmost positions
, this sequence has a @xmath115 followed by @xmath121 ones and a final @xmath115 .
given @xmath209 , let us remove from this sequence the rightmost @xmath210 piles , obtaining a configuration with mass @xmath211 .
equivalently , we can obtain the same configuration starting from the augmented heaviside sequence corresponding to @xmath212 and joining to the right a sequence of @xmath107 piles with one seed each .
let us call _ truncated augmented heaviside sequence _ such a configuration .
as an example , the truncated augmented heaviside sequence with mass @xmath213 is given by 1 2 1 2 1 1 2 1 1 1 2 1 1 .
[ prop : truncated ] the truncated augmented heaviside sequence with @xmath63 , @xmath118 , becomes a @xmath214-biaugmented marching group of the second kind after @xmath215 moves .
simply color the rightmost sequence of ones and track their position during the first @xmath215 moves of the augmented heaviside sequence of proposition [ teo : lower1 ] with @xmath121 in place of @xmath11 and @xmath143 .
[ teo : lower2 ] a truncated augmented heaviside sequence with @xmath63 , @xmath118 , becomes periodic after exactly @xmath216 moves . if @xmath217 or @xmath218 , the function @xmath219 is not defined by the previous corollary , and we shall conventionally set it to @xmath220 . as for the augmented heaviside sequences of section [ sec :
augheaviside ] , the special biaugmented marching group of the second kind at which we arrive if @xmath155 evolves to an augmented marching group with period exactly @xmath58 , and again we have a diameter given by @xmath221 . in the special case @xmath222 , @xmath223 , we need a further type of configuration . for @xmath222 , @xmath223 , we define the _ biaugmented heaviside sequence _ in the following way : take the truncated augmented heaviside sequence for @xmath224 , and then add a third seed in the pile with two seeds at position @xmath225 . as an example
, @xmath226 produces the sequence 1 2 1 2 1 1 2 1 1 1 2 1 1 1 1 3 1 1 .
[ teo : lower3 ] the biaugmented heaviside sequence with @xmath227 becomes periodic after exactly @xmath228 moves .
let us color the extra added third seed in the pile at position @xmath225 . by construction , without the colored seed the biaugmented heaviside sequence becomes a truncated augmented heaviside sequence with mass @xmath229 , hence by proposition [ prop : truncated ] , after @xmath230 moves we have an @xmath231-biaugmented marching group of the second kind with an extra seed ( corresponding to an increment of @xmath232 with respect to the marching group ) in the leftmost pile .
the argument we used to study the biaugmented marching groups can be adapted to the present case , proving that the periodic configuration is reached exactly after the same number of moves as for the biaugmented marching group of the second kind : @xmath125 with @xmath233 and @xmath234 .
hence the initial configuration becomes periodic after @xmath235 also in this case a direct inspection of the resulting augmented marching group allows to show that its period is exactly @xmath58 , and again the diameter of the special plateau sequence is @xmath236 .
the lower bound @xmath237 is only defined for special values of @xmath5 , of the form @xmath238 , and we conventionally set it to @xmath220 in all the other cases .
we can now gather the three lower bounds and define @xmath239 so that the previous results can be summarized as the following : @xmath240 simply gather the results obtained in propositions [ teo : lower1 ] , [ teo : lower3 ] and corollary [ teo : lower2 ] .
a plot of the function @xmath96 for @xmath241 can be seen in figure [ fig : ldepth ] . for @xmath241 .
we have @xmath242 . ]
there is an explicit way to express the lower bound @xmath96 . indeed , writing @xmath243 in the form @xmath244 with @xmath245 and @xmath246 , one can prove that @xmath247 notice that the middle case occurs only when @xmath11 is even .
in this section we introduce one of the main notions of the paper , namely _ @xmath6-monotonicity_. firstly we define a sequence which is strictly related to the energy levels interpretation of the mancala game . [ def : energyseq ] given a mancala configuration @xmath23 we define the _ energy sequence _ @xmath248 as follows : setting @xmath249 , we define @xmath250 and extend it on @xmath251 as a periodic function of period @xmath252 .
notice that if @xmath253 we actually use values of @xmath14 outside its support , so that we have @xmath254 for @xmath255 .
we have that @xmath256 if and only if @xmath257 for all @xmath258 .
[ def : smon ] given @xmath259 and @xmath23 , we say that @xmath14 is _
@xmath6-monotone _ if : * @xmath260 , and * its energy sequence @xmath261 satisfies the inequality @xmath262 ( notice that the first requirement @xmath260 is redundant if @xmath263 ) .
it is convenient to introduce also the ( void ) notion of @xmath113-monotonicity , satisfied by all mancala configurations .
[ def : increasingplateau ] let @xmath264 .
an _ increasing plateau of length @xmath6 _ is a subset of @xmath265 consecutive values @xmath266 of an energy sequence such that @xmath267 it is readily proved that a configuration is @xmath6-monotone if and only if each increasing plateau has length at least @xmath6 .
let us give a few examples of @xmath6-monotone configurations : * 1 2is @xmath113-monotone but not @xmath68-monotone * 1 1 1is @xmath68-monotone but not @xmath115-monotone * 4 1 1andare @xmath115-monotone but not @xmath268-monotone * 5 2 2 1andare @xmath268-monotone but not @xmath87-monotone * 6 3 3 2 1andare @xmath87-monotone but not @xmath269-monotone .
if a mancala configuration is @xmath270-monotone , then it is also @xmath271-monotone for any @xmath272 .
moreover , a configuration is an augmented marching group if and only if it is @xmath6-monotone for all @xmath273 .
[ rem : weakstrong ] the notions of monotonicity , in the sense of definition [ defn : monotone ] , and of @xmath68-monotonicity , in the sense of the previous definition , are slightly different . for instance , the sequence 4 2 is monotone but not @xmath68-monotone .
however , if a configuration @xmath14 is monotone , then @xmath50 is @xmath68-monotone .
hence , in view also of next theorem , during the evolution of a configuration there is at most one time when it is monotone but not @xmath68-monotone . in passing , we observe that any @xmath68-monotone sequence has a predecessor that is merely monotone , _
i.e. _ a @xmath68-monotone configuration can be obtained by applying a move to some monotone configurations .
hence the maximal depth / diameter in the family of monotone configurations is exactly one plus the maximal depth / diameter in the family of @xmath68-monotone configurations .
the following theorem shows that the class of @xmath6-monotone configurations is closed under mancala moves .
[ teo : smonclosed ] if @xmath14 is @xmath6-monotone , then @xmath50 is @xmath6-monotone .
let us set @xmath274 , @xmath275 , @xmath261 and @xmath276 .
clearly any mancala configuration is at least @xmath113-monotone , hence the set of @xmath113-monotone configurations is closed under mancala moves .
a direct check shows that the property @xmath260 is preserved under a mancala move whenever @xmath14 is a monotone configuration , indeed we have @xmath277 , whence @xmath278 .
consequently the set of @xmath68-monotone configurations is closed under mancala moves .
now assume that @xmath279 , hence we have to check only the values of the energy sequence .
we divide the proof in five cases . * case @xmath280 .
* we have @xmath281 and the new period is @xmath282 . since @xmath279 , then @xmath14 is monotone and @xmath283 , hence we have @xmath284 and the new energy sequence @xmath285 is obtained by a left - shift of @xmath286 possibly followed by the insertion of a new value . indeed @xmath287 for @xmath288 and @xmath289 , since @xmath290 . if @xmath291 , then @xmath292 and the new energy sequence @xmath285 is exactly a left - shift of @xmath286 , so that @xmath71 is itself @xmath6-monotone . if @xmath293 , then @xmath294 and the new value to be inserted is @xmath295 ( being @xmath296 ) and it is equal to the previous value @xmath297 .
since inserting a value in the sequence can not destroy the @xmath6-monotonicity property if it is equal to the previous ( or the next ) value , we can conclude that @xmath71 is @xmath6-monotone .
* case @xmath298 .
* we have @xmath299 and @xmath300 .
since @xmath283 we have @xmath301 , so that @xmath302 and @xmath303 .
moreover , @xmath304 implies that @xmath305 , so that @xmath306 .
using the inequalities above we conclude @xmath307 the energy sequence @xmath285 is obtained from the energy sequence @xmath286 with the following operations : 1 .
translate the values of @xmath286 to the left of one position @xmath308 .
this has no impact on the @xmath6-monotonicity ; 2 .
decrease by one the values with indices in the range @xmath309 ; 3 .
remove the value at index @xmath252 and extend to @xmath251 with period @xmath310 .
the last two operations can only impact @xmath6-monotonicity in the case of indices @xmath311 with @xmath312 ( up to addition of multiples of @xmath310 ) . however using we have @xmath313 .
* case @xmath314 . *
the argument is similar to the previous , but now becomes @xmath315 moreover the set of indices ( within the period @xmath316 ) where we decrease the energy by one ( step 2 above ) reduces to @xmath317 and is further reduced to the single index @xmath318 after removal of the energy value with index @xmath252 ( step 3 above ) .
now suppose by contradiction that @xmath319 with @xmath311 , @xmath312 .
this is only possible if both @xmath320 and @xmath321 .
the latter equation implies @xmath322 and , in view of the discussion above , it can be rewritten as @xmath323 which contradicts the @xmath6-monotonicity of @xmath14 .
* case @xmath324 and @xmath291 .
* it turns out that @xmath325 .
moreover @xmath326 whereas @xmath327 , so that @xmath328 .
the resulting energy @xmath285 can be obtained by a left shift of @xmath286 followed by removal of element at index @xmath252 , which is the same as first removing the element at index @xmath68 of @xmath286 and then performing a left shift . the fact that @xmath329 allows to ensure that the removal of the value @xmath330 does not impact the @xmath6-monotonicity , which is then unaffected by the final left shift . * case @xmath324 and @xmath293 .
* it turns out that in this case @xmath331 and that the energy sequence of @xmath71 is just a left - shift of the energy sequence of @xmath14 : @xmath332 .
these five cases cover all possible situations so that we conclude the proof . for particular configurations it is possible to prove that repeated mancala moves actually increase the order of monotonicity .
the following is our first important result for which the distinction between monotonicity and @xmath68-monotonicity is crucial .
for instance , the result is false for the sequence ` 4 2 ` , which is monotone but not @xmath68-monotone .
[ teo : incrmon ] let @xmath23 , @xmath273 and @xmath333 such that 1 .
@xmath14 is @xmath6-monotone ; 2 .
@xmath334 ; 3 .
@xmath335 , ( i.e. @xmath336 ) . then @xmath337 ( the configuration after @xmath98 moves ) is @xmath338-monotone .
denote by @xmath71 the configuration obtained after @xmath98 moves . since we apply a number of moves that is larger than the length of the sequence
, it follows that @xmath71 is at least @xmath68-monotone ( it is already monotone after @xmath339 moves , hence @xmath68-monotone after @xmath340 moves by remark [ rem : weakstrong ] ) .
this covers the case @xmath341 , and we can suppose @xmath259 .
in particular we can assume that @xmath342 where the first inequality follows from requirement ( ii ) and the second inequality comes from ( iii ) and constraint ( i ) of definition [ def : smon ] .
using the energy level interpretation of the game , we want to track the position after @xmath98 moves of the seeds with energy larger than @xmath98 . first observe that requirement ( ii ) implies that all energy levels up to the @xmath343-th are completely filled and that seeds at level @xmath98 ( which is not completely filled ) are subjected to @xmath98 rotations to the left and end up at the same initial position .
a seed in column @xmath78 ( notice that requirement ( iii ) implies that @xmath344 ) at level @xmath345 is also rotated @xmath98 times to the left , however the @xmath346-th energy level has length @xmath347 , so that it will end up in column @xmath348 and can possibly decrease its energy ( that is , its height ) once or more than once . requirements ( i ) and ( iii ) imply that there is no seed at height @xmath349 in the first @xmath6 columns , whereas seeds in column @xmath350 and height at least @xmath349 get moved to the right after the set of @xmath98 moves .
this allows to conclude that @xmath351 for all @xmath311 , @xmath352 , where @xmath276 is the energy sequence of @xmath71 given by definition [ def : energyseq ] .
if there were no energy decrease , then all increasing plateaus would increase their length by @xmath68 and the @xmath338-monotonicity would follow . the decrease in energy after each move involves all seeds in the final columns and it can be easily checked that it maintains the @xmath338-monotonicity .
[ def : qcanonical ] let @xmath333 , @xmath353 and @xmath99 .
we say that @xmath14 is @xmath98-canonical if 1 .
@xmath334 ; 2 .
@xmath335 , in particular @xmath14 * is not * @xmath354 ; is the lowest that is not completely filled . ]
the gap ( see definition [ def : gap ] ) in the rightmost position ( column @xmath98 ) in the energy level at height @xmath98 is the last to be filled after repeated moves .
notice that requirement @xmath355 implies that the @xmath98-th energy level will eventually fill up completely . due to the central importance of @xmath98-canonical configurations a few remarks
are in order .
the value @xmath98 of a @xmath98-canonical configuration is necessarily the height of the lowest active energy level ( the lowest level that is not completely filled ) .
this follows from requirements ( 1 ) and ( 2 ) of the above definition .
hence a configuration can not be @xmath98-canonical with two different values of @xmath98 . in view of the comments following definition [ def : gap ] , in particular the fact that _ filling a gap _ is permanent
, we can indeed define a _ filling order _ on the gaps at level @xmath98 .
all the gaps ( at level @xmath98 ) will be eventually filled due to requirement @xmath355 , and we can identify the gap that will be filled last .
an augmented marching group can not be @xmath98-canonical , since if @xmath99 is an augmented marching group with @xmath355 then @xmath356 .
this is incompatible with requirement ( 2 ) of definition [ def : qcanonical ] .
[ rem : sameq ] since the gap in the rightmost position of a @xmath98-canonical configuration @xmath14 returns in the same position after every sequence of exactly @xmath98 moves , it follows that the @xmath98-th energy level fills up after exactly @xmath357 moves , for some @xmath358 .
in particular @xmath359 .
moreover @xmath360 is itself a @xmath98-canonical configuration for all @xmath361 and there are no other @xmath98-canonical configurations in between . as an example , let us consider the configuration @xmath362 , it satisfies requirements ( 1 ) and ( 2 ) of definition [ def : qcanonical ] with @xmath363 ( see figure [ fig : ex55362 ] for an energy - level interpretation ) , it has three gaps at the energy level @xmath364 with three seeds at higher level , enough to completely fill the sixth energy level ( @xmath365 ) .
however this configuration is not a @xmath364-canonical configuration because the rightmost gap in column @xmath364 is not the last filled .
this can only be seen by letting the configuration evolve and annotating the filling order of the three gaps . in this example
only seven mancala moves are sufficient to see that the gap in column @xmath364 is filled first , whereas the gap that is filled last is the one originally in column @xmath268 .
the configuration @xmath366 ( obtained from the previous one with three mancala moves ) is indeed a @xmath364-canonical configuration .
as noted in remark [ rem : sameq ] , after six mancala moves we have either another @xmath364-canonical configuration or the sixth level becomes completely filled .
it turns out that we have a total of four other @xmath364-canonical configurations , one after each sequence of six moves starting from the @xmath364-canonical configuration ` 9 5 3 2 2 ` .
( 0,3.8 ) ; ( 4 , 3.8 )
( 4 , 4 ) ; in 4, ... ,5 ( 0,0 ) ( 1,0 ) ( 1,1 ) ; ( 0,0 )
( 0,0 ) node[right , xshift=0cm , yshift=-0.3cm]@xmath269-th level ; in 0, ... ,4 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) node[right , xshift=0.0cm , xshift=-0.5mm]@xmath115 ; ( 0.5,0.5 ) circle ( 0.1 ) node[right , xshift=0.0cm , xshift=-0.5mm]@xmath268 ; ( 0.5,0.5 ) circle ( 0.1 ) node[right , xshift=0.0cm , xshift=-0.5mm]@xmath68 ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,6 ) ( 0,3.8 ) ; ( 4 , 3.8 )
( 4 , 4 ) ; in 4, ... ,5 ( 0,0 ) ( 1,0 )
( 1,1 ) ; in 0, ... ,4 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) node[right , xshift=0.0cm , xshift=-0.5mm]@xmath68 ; ( 0.5,0.5 ) circle ( 0.1 ) node[right , xshift=0.0cm , xshift=-0.5mm]@xmath115 ; ( 0.5,0.5 ) circle ( 0.1 ) node[right , xshift=0.0cm , xshift=-0.5mm]@xmath268 ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0,6 ) ( 0,3.8 ) ; ( 4 , 3.8 ) ( 4 , 4 ) ; in 4, ... ,5 ( 0,0 )
( 1,1 ) ; in 0, ... ,4 ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; ( 0.5,0.5 ) circle ( 0.1 ) ; ( 0.5,0.5 ) circle ( 0.4 ) ; [ teo : maxmoncanon ] a @xmath98-canonical configuration can not be @xmath343-monotone .
in other words it is at most @xmath367-monotone .
let @xmath14 be a @xmath98-canonical configuration , @xmath368 ; hence @xmath369 .
now suppose by contradiction that @xmath14 is also @xmath343-monotone ; this implies that @xmath370 by ( i ) of definition [ def : smon ] .
set @xmath261 ( the energy sequence ) , with period @xmath371 , hence @xmath372 either @xmath373 or @xmath374 .
energy levels strictly larger than @xmath98 must also be present , otherwise @xmath14 would be an augmented marching group , incompatible with @xmath98-canonicity . hence there must exist an increasing plateau , and its length can not be longer than @xmath375 due to the periodicity of @xmath286 , which is at most @xmath339 . on the other hand
, @xmath343-monotonicity implies that the length of every increasing plateau is at least @xmath376 , a contradiction .
in this section we shall fix @xmath377 for some @xmath10 and consider a mancala configuration @xmath99 .
we let @xmath14 evolve ( applying mancala moves ) until it reaches an augmented marching group ; we denote by @xmath378 the result of @xmath379 mancala moves , so that in particular @xmath380 .
let @xmath381 be the smallest time @xmath132 for which @xmath382 is an augmented marching group , in particular @xmath383 .
we also denote by @xmath384 the smallest time @xmath132 such that @xmath385 . during the evolution @xmath382
will occasionally become a @xmath98-canonical configuration .
specifically we indicate with @xmath386 the times before @xmath387 when we have a @xmath98-canonical configuration for some @xmath388 , @xmath389 can be itself @xmath98-canonical , however @xmath390 would imply @xmath391 , hence @xmath392 . in other words ,
if we restrict @xmath393 , then @xmath389 can not be @xmath98-canonical .
we call such times _ critical times_. specifically we have a sequence @xmath394 such that @xmath395 is a @xmath396-canonical configuration
. consider two consecutive critical times @xmath397 and @xmath398 .
remark [ rem : sameq ] applied to @xmath395 tells us that after @xmath396 moves , we have either another @xmath396-canonical configuration , with no other @xmath98-canonical configurations in between , in which case @xmath399 and @xmath400 , or we have just filled the @xmath396-th energy level so that we have @xmath401 and @xmath402 . in both cases we have @xmath403 .
we can now apply lemma [ teo : incrmon ] to @xmath395 with @xmath404 and conclude that the @xmath6-monotonicity of @xmath405 increases strictly . since @xmath406 is @xmath113-monotone , we obtain by induction that @xmath405 is @xmath78-monotone and that @xmath389 is @xmath6-monotone .
it is possible for the set of critical times to be empty . in this case
we shall see that the filling up of the energy levels is very fast .
[ teo : stepcanon ] if @xmath334 and @xmath407 , then there exists @xmath408 such that either @xmath409 or @xmath410 is a @xmath98-canonical configuration . if @xmath14 is itself @xmath98-canonical , the result holds true with @xmath411 ( besides the trivial value @xmath412 ) .
if the @xmath98-th energy level is full , then there is nothing to prove , otherwise locate the gap that will be filled last in column @xmath79 .
if @xmath413 we already have a @xmath98-canonical sequence , otherwise we apply @xmath79 moves in order to rotate the gap position into column @xmath98 , now we either have a @xmath98-canonical sequence or we filled the @xmath98-th energy level .
the second part of the lemma easily follows by first performing a single move and then reasoning as in the first part .
[ teo : emptycrit ] if the set of critical times is empty ( @xmath414 ) , then one has @xmath415 ,
i.e. @xmath416 it can be readily seen by repeated applications of the previous lemma .
similarly , the @xmath417-canonical sequence @xmath406 is reached after at most @xmath418 moves , i.e. @xmath419 .
now let @xmath420 ( notice that @xmath421 ) , and decompose @xmath422 if @xmath423
we first apply one move to @xmath395 and then invoke lemma [ teo : stepcanon ] @xmath424 times , concluding that @xmath425 on the other hand , if @xmath426 then @xmath427 ( remark [ rem : sameq ] ) . in the end
we have the estimate @xmath428 at each critical time we gain one level of monotonicity , so that @xmath429 is @xmath430-monotone .
the maximal possible value of @xmath431 is @xmath432 , which gives an upper bound @xmath433 for the @xmath6-monotonicity of @xmath429 ( theorem [ teo : maxmoncanon ] ) and hence we have @xmath434 .
we get @xmath435 to obtain the configuration @xmath390 we need to apply a move to @xmath429 and again invoke lemma [ teo : stepcanon ] ( more than once if @xmath436 ) for a total of at most @xmath437 moves .
we notice that the last term in the previous sum is also present in the sum that defines @xmath438 , so that using we arrive at @xmath439 we summarize the previous analysis in the following lemma .
[ teo : balance ] let @xmath377 and @xmath99 .
then there exists @xmath440 such that after at most @xmath441 moves the resulting configuration @xmath382 satisfies : 1 .
@xmath385 ; 2 .
@xmath382 is @xmath6-monotone .
if @xmath62 or @xmath156 then the resulting configuration is an augmented marching group . [ def : functionw ] given @xmath442 two integers ,
we define @xmath443 by @xmath444 the value of @xmath445 can be explicitly computed by first expressing @xmath446 with @xmath447 and @xmath448 ( euclidean division ) and then using the formula for the sum of the first natural numbers .
we end up with @xmath449 .
we have the special values * @xmath450 , @xmath451 if @xmath452 ( only one term in the above sum ) , * @xmath453 if @xmath454 ( two terms in the above sum ) , * @xmath455 if @xmath456 ( more than two terms in the above sum ) . moreover ,
if @xmath457 then @xmath445 is strictly increasing with respect to @xmath458 and strictly decreasing with respect to @xmath459 .
we can then `` invert '' @xmath445 with respect to @xmath459 and define the function @xmath460 which is ( weakly ) decreasing with respect to @xmath461 and ( weakly ) increasing with respect to @xmath458 .
we have the special values * @xmath462 if @xmath463 , * @xmath464 , if @xmath465 .
referring to figure [ fig : vmap ] the value of @xmath443 can be graphically interpreted as the number of unit squares completely included in the right triangle having base ( long cathetus ) of length @xmath458 and height ( short cathetus ) of length @xmath466 .
( 0,2.333 )
( 0,0 ) ; in 0, ... ,3 ( 0,0 )
( 0,1 ) ; ( 0,0 )
( 0,0 ) ( 1,1 )
( 0,1 ) ; with the same graphical interpretation in mind function @xmath467 can be viewed as the maximal value of @xmath459 such that the right triangle of sides @xmath458 and @xmath466 contains at least @xmath461 unit squares . for @xmath468 , let us consider a @xmath11-canonical configuration . by lemma [ teo : stepcanon ] , after a cycle of @xmath11 moves we have that the resulting configuration is either itself @xmath11-canonical or it has just become larger than @xmath53 .
iterating this procedure , we want to relate the number of cycles needed to become larger than @xmath53 with the degree of monotonicity of the configuration .
[ teo : cyclenumbound ] let @xmath469 with @xmath63 , @xmath223 and suppose that 1 .
@xmath14 is a @xmath11-canonical configuration ; 2 .
@xmath14 is @xmath6-monotone for some @xmath273 .
let @xmath470 denote the number of cycles of @xmath11 moves that are required to become larger than @xmath53 and suppose that @xmath471 .
then @xmath472 equivalently , reasoning by contradiction , we have the upper bound @xmath473 first we analyze the case @xmath474 , so that @xmath259 . since the configuration is @xmath6-monotone and @xmath11-canonical , in this case @xmath475 , hence @xmath476 , @xmath477 @xmath478 and so on , where @xmath286 is the energy sequence @xmath261 .
then the total number of seeds with energy above @xmath11 can not be larger than @xmath479 .
this is obtained by observing that we have a sequence of @xmath98 plateaus of length @xmath6 and increasing height from @xmath113 to @xmath340 and a final top plateau of length @xmath184 and height @xmath98 .
this implies that @xmath480 where the `` @xmath220 '' is a consequence of the fact that the @xmath11-th energy level is not completely full . for a generic value of @xmath481
we simply perform @xmath482 cycles of @xmath11 moves obtaining a configuration with level of monotonicity @xmath483 with still @xmath484 and we can resort to the special case discussed above
. the following lemma will be fundamental in proving the optimal upper bound .
[ teo : magico ] let @xmath99 with @xmath63 , @xmath485 , and denote by @xmath486 the number of @xmath98-canonical configurations encountered during the evolution of @xmath14 , with @xmath487 .
then the total number @xmath488 of canonical configurations encountered , and hence the least level of monotonicity of the first configuration that becomes larger than ( or equal to ) @xmath53 , is bounded by @xmath489 in particular , for @xmath490 we have @xmath491 and for @xmath492 we have @xmath493 where we used the special values of function @xmath494 computed in definition [ def : functionw ] .
if @xmath495 for all @xmath98 , then we can resort to corollary [ teo : emptycrit ] and conclude immediately
. otherwise , set @xmath496 and let @xmath497 be the largest value of @xmath98 such that @xmath498 .
we apply lemma [ teo : cyclenumbound ] with @xmath499 , @xmath500 and @xmath501 and obtain an upper bound for @xmath502 that , since @xmath503 can be written as @xmath504 or equivalently as @xmath505 where the last inequality follows from the monotonicity of @xmath494 with respect to its arguments . the evolution of a configuration @xmath469 , with @xmath63 and @xmath64 ,
will be decomposed into two stages : 1 .
the evolution from the beginning up to the first time @xmath132 when @xmath385 , where as usual @xmath506 ; 2 .
the subsequent evolution until we reach an augmented marching group .
the duration of each of these stages will be estimated from above with bounds that depend upon a few parameters and we shall find a uniform bound valid for all possible feasible choices of such parameters finally obtaining an upper bound for @xmath60 . we need some further lemmas . [ teo : augmentedifrmonotone ] let @xmath469 and let @xmath11 be the largest integer such that @xmath256 .
write @xmath63 and suppose that @xmath14 is @xmath107-monotone
. then @xmath14 is an augmented marching group . by contradiction
suppose that @xmath14 is not an augmented marching group , then the energy sequence has an element with value @xmath11 ( otherwise @xmath507 ) and an element with value @xmath111 ( otherwise @xmath14 would be an augmented marching group ) .
this implies the existence of an increasing plateau at level @xmath58 , which must have length at least @xmath107 .
an energy level of @xmath58 at position @xmath78 is possible without using one of the @xmath107 exceeding seeds only if @xmath508 , however if that position is part of the increasing plateau , then the seed at level @xmath111 must necessarily stay on top of a seed at level @xmath58 . in all cases we need at least @xmath509 seeds to build such an increasing plateau , which contradicts the assumptions .
[ teo : nolongersuboptimal ] let @xmath99 with @xmath63 , @xmath131 , @xmath256 and @xmath6-monotone .
if @xmath510 becomes an augmented marching group , then @xmath511 using the energy level interpretation we identify the gap at level @xmath58 that will be filled last , and call @xmath512 its column index .
we consider two cases . *
case @xmath513 or @xmath514 .
* the first term @xmath193 in the estimate is then the maximal number of moves that are required to move that gap into the rightmost @xmath515-th column .
we observe that starting from that configuration the augmented marching group will be obtained exactly after a number of moves multiple of @xmath58 after each cycle of @xmath58 moves the level of monotonicity increases at least of one unit ( lemma [ teo : incrmon ] ) , so that after @xmath516 cycles of @xmath58 moves we obtain a configuration @xmath71 that is at least @xmath231-monotone .
now we want to prove that @xmath231-monotonicity ( and not higher ) is actually impossible under the circumstances .
this would imply at least @xmath107-monotonicity that in turn implies that we have an augmented marching group thanks to lemma [ teo : augmentedifrmonotone ] .
suppose then by contradiction that @xmath71 is @xmath231-monotone but not @xmath107-monotone and hence it is not an augmented marching group .
this means that the gap at level @xmath58 , column @xmath58 is still empty .
the energy sequence of @xmath71 has then both elements at level @xmath11 and elements at level @xmath111 , so that there exists at least an increasing plateau at level @xmath58 ( i.e. a sequence of energy elements as contiguous positions at level @xmath58 following an element at level @xmath11 and preceding an element at level @xmath111 ) .
such plateau has length at least @xmath193 due to the @xmath231-monotonicity .
now recall that we have an excess of @xmath107 seeds with energy larger or equal to @xmath58 ( levels @xmath11 and below are completely filled thanks to the assumptions ) .
column @xmath111 is the only one having energy @xmath58 with no need to use one of the excess seeds , however we have barely enough seeds to form an increasing plateau of length @xmath193 which requires @xmath509 seeds that reduce to @xmath107 only if we take advantage of the energy step at column @xmath111 . recalling that column @xmath58 has energy level @xmath11 ( the gap is still empty ) , the only possible disposition of the @xmath107 excess seeds is the one with @xmath193 of them in columns @xmath68 to @xmath193 ( energy @xmath58 ) and the last excess seed itself in column @xmath193 ( then with energy @xmath111 ) .
this configuration becomes an augmented marching group after exactly @xmath107 moves , however we know that an augmented marching group is reached after a multiple of @xmath58 moves and hence @xmath517 , a contradiction . since @xmath231-monotonicity is impossible for @xmath71 ,
then we have at least @xmath107-monotonicity and hence we have an augmented marching group . * case @xmath518 . * let us consider the result of the evolution as soon as it becomes an augmented marching group and call it @xmath71 , the position of the `` marked '' gap is then column @xmath58 and it has just been filled with the last of the @xmath107 seeds the were originally at energy levels larger then @xmath58 . immediately to the left of this seed
there is a contiguous sequence of , say , @xmath519 seeds , @xmath520 at level @xmath58 , preceded by a gap in column @xmath521 .
clearly this latter gap ( let us call it `` gap number @xmath115 '' ) can be traced back to its corresponding position in the initial configuration @xmath14 , and will originally be empty in @xmath14 and will continue to be empty during the whole evolution .
it is located in column @xmath522 of @xmath14 , to the left of the marked gap ( in column @xmath512 ) .
no seed at energy larger then @xmath58 ( they are the `` active '' seeds , the only ones that `` move '' with respect to the @xmath58 level ) can ever cross or reach the column containing gap number @xmath115 , so that the active seeds originally to the left of gap number @xmath115 or to the right of the marked gap ( call this region the `` outer region '' ) can ever interact with the region between gap number @xmath115 and the active gap . without loss of generality
we can then suppose that the outer region has reached its final configuration already in @xmath14 .
this actually means that there are no active seeds to the right of the marked gap .
the final step is then to observe that in this situation @xmath14 can be interpreted as the result of the evolution ( with @xmath523 moves ) from a configuration @xmath524 that itself satisfies the hypotheses of this lemma and the marked gap already in the @xmath58 position ( in other words we can `` steal '' a few moves ) and we can resort to the first case of the lemma .
the proof is thus complete .
we now analyze the evolution of @xmath469 , @xmath63 , @xmath131 .
we denote with @xmath273 the number of @xmath98-canonical configurations encountered during the evolution with @xmath525 . using lemma [ teo : magico ]
we obtain the constraint @xmath526 , which in particular gives @xmath527 the third estimate is actually just a special case of the second one ; the first and second estimates coincide for @xmath107 in the common interval @xmath528 . as usual
we denote by @xmath378 and we set * @xmath529 the smallest time such that @xmath530 * @xmath531 the smallest time such that @xmath532 is an augmented marching group .
[ teo : timesestimates ] we have the following estimates : 1 .
@xmath533 ; 2 .
@xmath534 if @xmath535 ; 3 .
@xmath536 if @xmath537 .
the first point is exactly lemma [ teo : balance ] .
point ( 2 ) is simply lemma [ teo : augmentedifrmonotone ] .
point ( 3 ) follows from lemma [ teo : nolongersuboptimal ] we then have a total estimate of @xmath538 as a function of @xmath6 that reads as follows : @xmath539 clearly in the case @xmath537 it is most convenient to take @xmath414 , whereas if @xmath535 the maximum value is achieved by selecting the largest possible value of @xmath6 , which is dictated by the constraints .
we recall from section [ sec : lower ] the definition of @xmath96 as @xmath540 [ teo : mainresult ] let @xmath63 with @xmath541 and @xmath131
. then @xmath542 .
the case @xmath543 corresponds to @xmath544 , easily studied manually . in view of the results of section
[ sec : lower ] it suffices to prove the upper bound @xmath545 .
let @xmath469 .
we invoke lemma [ teo : balance ] so that there exists @xmath6 satisfying the constraints such that after at most @xmath546 moves the resulting configuration is @xmath6-monotone and @xmath547 .
now we enforce lemma [ teo : timesestimates ] and distinguish three cases . * case @xmath490 . *
if @xmath535 , we use the first inequality of in to conclude @xmath548 if @xmath537 , using @xmath549 , and @xmath550 , i.e. @xmath551 we have @xmath552 so that in any case @xmath553 . * case @xmath554 .
* again we distinguish the case @xmath535 , together with the third inequality of to obtain @xmath555 whereas if @xmath537 , again using @xmath549 we have @xmath556 * case @xmath557 . *
we start with the case @xmath537 and again use @xmath549 to get @xmath558 finally , if @xmath535 , enforcing the second inequality in , using @xmath559 we obtain @xmath560 we have covered all the possible cases and the proof is complete . the previous theorem [ teo : mainresult ] , and in particular the upper bound @xmath545 also allows to compute the maximal diameter of @xmath25 : the diameter of @xmath25 ( see definition [ def : depthdiameter ] ) is given by @xmath561 where @xmath562 is the maximal period of an augmented marching group , i.e. @xmath563 if @xmath62 is a triangular number , otherwise @xmath564 if @xmath63 , with @xmath565 , is not a triangular number . that the diameter can not exceed @xmath561 immediately follows from theorem [ teo : mainresult ] . the reverse inequality follows from the constructions of section [ sec : lower ] , in particular we make use of proposition [ teo : lower1 ] , corollary [ teo : lower2 ] , proposition [ teo : lower3 ] and the discussion about the period of the resulting augmented marching groups following each of these results .
the notion of @xmath6-monotonicity naturally introduces a sort of graduation on the graph of moves . indeed ,
theorem [ teo : smonclosed ] asserts that the subgraph obtained by considering only the @xmath6-monotone configurations for some @xmath273 is closed under mancala moves ( no arc leaves this subset ) .
it is then natural to ask what is the depth of such subgraphs .
more precisely , for @xmath566 let us denote by @xmath567 and @xmath568 the subsets of @xmath25 given by @xmath569 similarly , with @xmath570 and @xmath571 we denote the subgraphs of @xmath70 having nodes in @xmath567 and @xmath568 respectively .
then the _ depth _ of @xmath570 , denoted by @xmath572 , is the maximal distance from a node of @xmath570 to a periodic configuration . in this context , recalling that @xmath113-monotonicity is a void notion , our main theorem [ teo : mainresult ] gives a value for @xmath573 . it would be desirable to have a result about the depth @xmath572 for any value of @xmath6 ; unfortunately , at the moment we are not able to extend our result to nontrivial values of @xmath6 .
especially the value @xmath574 is interesting : indeed , remark [ rem : weakstrong ] relates @xmath575 to @xmath571 and in particular their respective depth .
adding one to @xmath576 we obtain the depth of the graph of monotone configurations .
this is of special interest in view of the following the monotone mancala graph is isomorphic to the graph of the bulgarian solitaire @xcite,@xcite,@xcite . the bulgarian solitaire , already mentioned in the introduction , is played as follows .
the starting position consists of a number of piles , each with a ( possibly different ) number of cards .
each move consists in removing one card from each pile and collecting the removed cards in a new pile .
piles that become empty are simply neglected and the order of the piles ( and of the cards in a pile ) is inessential .
we can associate a mancala configuration to a bulgarian solitaire configuration by defining @xmath0 equal to the number of piles with at least @xmath78 cards . in this way we obtain a valid mancala configuration which is clearly monotone .
it can be readily shown that this correspondence is one to one and that it is consistent with the respective game rules .
hence , the depth of the bulgarian solitaire subgraph is given by @xmath577 .
the following conjecture is stated in @xcite : let @xmath63 with @xmath541 and @xmath131 .
the depth of @xmath575 is given by @xmath578 in @xcite the authors proved this result for some special values of @xmath5 , while the above expression is proved to be a lower bound .
our computer simulations validates the conjecture for values up to @xmath579 , improving the previous limit of @xmath580 .
the authors are grateful to the anonymous reviewers for their precious suggestions .
a. bouchet , _
owari i. marching groups and periodical queues _ , 2005 , preprint ( http://isgem.rpi.edu/files/1545 ) a. bouchet , _ owari ii .
marching groups and bulgarian solitaire _ , 2005 , preprint ( http://isgem.rpi.edu/files/1546 ) h. bruhn , _ periodical states and marching groups in a closed owari _ , discrete math .
* 308 * ( 2008 ) , no . 16 , 36943698 . | we determine the optimal bound for the maximum number of moves required to reach a periodic configuration of _ open mancala _
( also called open owari ) , inspired by a popular african game .
a mancala move can be interpreted as a map from the set of compositions of a given integer in itself , thus relating our result to the study of the corresponding finite dynamical system . |
one of the key problems to overcome when conducting multi - wavelength surveys is determining which sources are associated with one another in different wave - bands , and which are unrelated .
when multiple observations have been conducted at similar wavelengths and with similar resolution and sensitivity , this problem can be reliably addressed by using a simple nearest neighbour match . however , in the situation where the two distinct sets of observations to be matched have considerably different resolution for example matching far - infrared or sub - millimetre survey data to an optical catalogue ( e.g. sutherland et al . , 1991 , clements et al . , 1996 ,
serjeant et al . , 2003 , clements et al . , 2004 , ivison et al .
2005 , 2007 , wang & rowan - robinson , 2009 , biggs et al .
2010 ) the large positional uncertainties in the longer - wavelength data can make it much more difficult to find reliable associations between sub millimetre sources and their optical / near infrared counterparts .
one method which can be used to identify the most likely counterpart to a low - resolution source , is the likelihood ratio technique ( hereafter lr ) , first suggested by richter ( 1975 ) , and expanded by sutherland & saunders ( 1992 ) and ciliegi et al .
the crucial advantage of the lr technique over other methods is that it not only uses the positional information contained within the two catalogues , but also includes brightness information ( both of the individual potential counterparts , and of the higher resolution catalogue as a whole ) to identify the most reliable counterpart to a low - resolution source .
the _ herschel _ astrophysical terahertz large area survey ( _ herschel_atlas , eales et al . , 2010 ) is the largest open - time key project that will be carried out with the _ herschel space observatory _ ( pilbratt et al . , 2010 ) .
the _ herschel_atlas will survey in excess of 550 deg@xmath7 in five channels centred on 100 , 160 , 250 , 350 and 500@xmath0 m , using the pacs ( poglitsch et al .
, 2010 ) and spire instruments ( griffin et al . , 2010 ) .
this makes _
herschel_atlas currently the largest area extragalactic _ herschel _ survey .
herschel_atlas observations consist of two scans in parallel mode reaching 5@xmath8 point source sensitivities of 132 , 126 , 32 , 36 and 45 mjy in the 100@xmath0 m , 160@xmath0 m , 250@xmath0 m , 350@xmath0 m and 500@xmath0 m channels respectively , with beam sizes of approximately 9 , 13 , 18 , 25 and 35 arcsec in the same five bands .
the spire and pacs map - making procedures are described in the papers by pascale et al .
( 2010 ) and ibar et al .
( 2010 ) , while the catalogues are described in rigby et al .
one of the primary aims of the _
herschel_atlas was to obtain the first unbiased survey of the local universe at sub - mm wavelengths , and as a result the survey was designed to overlap with existing large optical and infrared surveys . in this paper , we present a discussion of our implementation of the lr technique to identify the most reliable counterparts to 250@xmath0m selected sources in the _ herschel_atlas science demonstration phase ( sdp ) data field ( eales et al . , 2010 ) .
this field was chosen in order to take advantage of multi wavelength data from the sloan digital sky survey ( sdss york et al . , 2000 ) , and the uk infrared deep sky survey large area survey ( ukidss - las lawrence et al . , 2007 ) .
this field also overlaps with the 9 hour field of the galaxy and mass assembly survey ( gama driver et al . , 2010 ) .
the gama catalogue ( hill et al . , 2011 ) , comprises not only thousands of redshifts ( for galaxies selected as described in baldry et al . , 2010 , and observed with the maximum possible tiling efficiency
robotham et al . , 2010 ) , but also @xmath9band defined aperture matched photometry in the @xmath10 bands .
in addition , the gama fields are being systematically observed using the _ galaxy evolution explorer ( galex ) _ satellite ( martin et al . , 2005 ) at medium
imaging survey depth to provide aperture matched fuv and nuv counterparts to the catalogued gama sources ( the _ galex_gama survey ; seibert et al .
, in prep ) .
these counterparts will potentially be of great scientific value once the most reliable optical counterpart can be established for each _
herschel_atlas source .
in section [ lrcalc ] we present the specific lr method that we have used to identify counterparts to 250@xmath0m selected sources from the _ herschel_atlas sdp catalogue in an @xmath9band catalogue of model magnitudes derived from the sdss dr7 . in section
[ sec : redshifts ] we present the redshift properties of our catalogue , which covers @xmath11deg@xmath7 over the gama 9 hour field .
section [ results ] contains some basic results based on our reliable catalogue , and in section [ conclusions ] we present some concluding remarks about the likelihood ratio technique and the resulting catalogue .
the likelihood ratio , i.e. the ratio between the probability that the source is the correct identification and the corresponding probability for an unrelated background source , is calculated as in sutherland & saunders ( 1992 ) : in which @xmath13 , and @xmath14 correspond to the sdss @xmath9band magnitude probability distributions of the full @xmath9band catalogue and of the true counterparts to the sub
millimetre sources , respectively , while @xmath15 represents the radial probability distribution of offsets between the 250@xmath0 m positions and the sdss @xmath9band centroids .
we will now describe how we calculate each component of this relationship in turn . for _
herschel_atlas sdp observations , it was necessary to determine the spire positional uncertainties .
since this information was not available _ a priori _ , we empirically estimated @xmath17 using the sdss dr7 @xmath9band catalogue positions , assuming that the sdss positional errors were negligible in comparison to the spire errors . to determine @xmath17 , we derived histograms of the separations between the positions in the mad - x spire catalogue ( rigby et al . , 2010 ) of the @xmath18 250@xmath0 m sources , and all of those objects in the @xmath9band sdss dr7 catalogue within 50 arcsec , doing this for both the north south and east west directions ( figure [ fig : clust ] ) .
these histograms can be well described as the sum of the gaussian positional errors plus the clustering signal for sdss sources convolved with gaussian errors , @xmath19 , with @xmath20 : where @xmath23 , with @xmath24 being measured in degrees for the purposes of comparison with the literature .
we determined the values of @xmath25 and @xmath26 empirically based solely on galaxies in the sdss catalogue over @xmath27 deg@xmath7 centred on the _ herschel_atlas sdp field ( limited to @xmath28 ) , with the best fit parameters @xmath29 and @xmath30 , in reasonable agreement with the values of connolly et al .
the effects of clustering ( i.e. @xmath31 ) are shown in the top panel of figure [ fig : clust ] . in order to determine the 1@xmath8 positional error of the 250@xmath0 m selected catalogue
, we conducted a simple @xmath32 fit of our model ( equations [ xhist ] & [ yhist ] ) to the histograms .
the results are shown in figure [ fig : clust ] for the summations in the east - west and north - south directions in the middle and bottom panels respectively .
the clustering signal is shown in the bottom two panels by the dotted lines , with the histograms and their poisson error bars overlaid with the best fit model ( solid lines ) .
the 1@xmath8 positional errors were found to be @xmath33 arcsec and @xmath34 arcsec in the two directions , consistent with one another within the errors .
the advantages of this method are two fold ; firstly , it is not necessary to identify the counterparts to the 250@xmath0 m sources _ a priori _ , and secondly , the centroids of the best fit gaussians may be used to determine astrometric corrections in the spire maps ( e.g. pascale et al . , 2010 ) .
the value for @xmath17 that we adopted was the weighted mean @xmath35 arcsec .
positional errors of the spire 250@xmath0m selected sources , we produced histograms of the total number of sdss sources within a box 50 arcsec on a side around the spire 250@xmath0 m centres .
after accounting for the clustering of sdss sources ( the top panel shows the signal expected for the clustering of sdss sources in the ra and dec directions convolved with gaussian positional errors the results are shown as solid and dashed lines for ra and dec , respectively , and these results appear as the dotted lines in the bottom two panels ) , we can add in an appropriate gaussian distribution of centres to account for the actual positions of the spire sources ( equations [ xhist ] & [ yhist ] ) .
performing a @xmath32 minimisation allows us to then empirically determine the 1@xmath8 positional uncertainty for these sources , which are shown as @xmath36 and @xmath37 . ] theoretically
, the positional uncertainty should depend on the signal - to - noise ratio ( snr ) of the detection and on the full - width at half maximum ( fwhm ) of the spire 250 @xmath0 m beam ( 18.1 arcsec , pascale et al . , 2010 ) , following the results derived in ivison et al .
( 2007 ; @xmath38 ) and assuming the case of uncorrelated noise .
we use our empirical results in figure [ fig : clust ] to calibrate the theoretical relation presented in ivison et al .
( 2007 ) to our data , and assume that our results are symmetric in ra and dec .
this leads us to introduce a factor of 1.09 to give equation [ ivison2007 ] : * whilst it is acceptable to neglect the sdss dr7 positional errors for the purposes of determining @xmath17 ( section [ sec : f_r ] ) , the astrometric precision for sources in the sdss dr7 catalogue is non - zero ( @xmath40 arcsec
abazajian et al . ,
* large sources , especially those without gaussian surface brightness profiles ( e.g. bright spiral galaxies ) , have considerably larger positional uncertainties associated with them .
* confusion provides a lower limit to the positional errors of the spire catalogue , although the snr in equation [ ivison2007 ] does include confusion noise as described in rigby et al .
( 2010 ) and pascale et al ( 2010 ) . to account for these effects
, we do not allow the positional uncertainty to fall below 1 arcsec , and we also include a term which adds 5 percent of the sdss @xmath9band isophotal major axis in quadrature to the value determined by equation [ ivison2007 ] , for those sources with @xmath9band model magnitudes @xmath41 .
finally , @xmath15 must be renormalised so that calculating the lr requires two further pieces of magnitude information , @xmath13 and @xmath14 .
the quantity @xmath13 is simply the probability that a background source is observed with magnitude @xmath43 . to estimate this
, we calculate the distribution of sdss dr7 @xmath9band model magnitudes for all of the primary photometry sources in the catalogue , normalised to the total area of the catalogue ( which is approximately 36.0 deg@xmath7 for the sdss catalogue that we use for this purpose ) .
the non - triviality lies in the calculation of @xmath14 the probability that a true counterpart to a 250@xmath0 m source has a magnitude @xmath43 . to estimate this
we calculate the @xmath9band magnitude distribution of the counterparts to the 250@xmath0 m sources using the method of ciliegi et al .
this method involves counting all objects in the optical catalogue within some fixed maximum search radius ( @xmath44 ) of the spire positions . to avoid influencing the results of this analysis with erroneous deblends in the sdss dr7 catalogue ( which artificially alter the number counts ) , we eyeballed the sdss @xmath9band images of each of the 5@xmath8 250@xmath0 m sources , removing 370 sdss sources from the input catalogue .
the magnitude distribution of the remaining objects is referred to as total@xmath45 .
here we have adopted @xmath46arcsec , which encloses @xmath4799.996% of the real counterparts to the 250@xmath0 m sources based on our derived value for @xmath17 .
the distribution total@xmath45 is then background subtracted to leave the magnitude distribution of excess sources around the 250@xmath0 m centres , real@xmath45 : where @xmath49 is the number of 250@xmath0 m sources in the catalogue .
this enables us to empirically estimate @xmath14 from the sources in our optical catalogue rather than modelling the @xmath9band magnitude distribution of 250@xmath0m selected _
herschel_atlas sources .
the distribution @xmath14 is given by equation [ eq : q_m ] : here @xmath53 represents the number of possible ids within 10.0 arcsec of the spire positions , and @xmath49 is defined as above . since the value of @xmath51 will be different for galaxies and unresolved sources in our catalogue , we must calculate @xmath14 , @xmath13 and @xmath51 separately for each population .
we separate resolved and unresolved sources using a slightly modified version of the gama colour colour relation from baldry et al .
( 2010 , modified such that @xmath54 rather than 0.20 to avoid adding an unphysical sharp edge to the stellar locus in figure [ fig : stargalsep ] ) .
having separated the two populations , we corrected the positions of the unresolved sources for known proper motions in the usno / sdss dr7 catalogue ( munn et al .
, 2004 ) , precessing their co ordinates to the epoch of the _ herschel_atlas sdp observations .
only those unresolved sources with proper motions detected at a snr @xmath55 were updated .
sources in our catalogue classified as unresolved , we find that three satisfy the gama colour selection criteria for being stellar , and so are potentially evolved stars , dust - obscured qsos or debris disk candidates possibly indicative of a proto - planetary system ( e.g. thompson et al . , 2010 ) .
the dashed line describes the first order star
galaxy separation locus ( for more details see baldry et al .
galaxy separation locus has been modified slightly from the baldry et al .
value due to the fainter magnitudes considered in our survey . ] for our sdss dr7 @xmath9band catalogue , @xmath56 , i.e. 58.3 percent of the galaxy counterparts are brighter than our magnitude limit . for the unresolved sources the value is @xmath57 , indicating that only 1 percent of the unresolved sources in the catalogue are detected at @xmath58 in our 250@xmath0 m data ( although see section [ sec : randcat ] ) .
thus we determine that overall @xmath59 .
the distributions of @xmath14 , and @xmath13 ( as well as the magnitude dependence of the lr
@xmath14/@xmath13 ) are shown in figure [ fig : nm ] , in which the left and right columns show the values for the resolved and unresolved sources , respectively . while the @xmath14 distribution for galaxies is well
sampled at @xmath60 mag , we assume that @xmath61 is constant for all sources brighter than this , enabling us to use our well - defined @xmath13 to estimate @xmath14 for the brightest galaxies .
since the fraction of _
sources associated with unresolved counterparts is low ( reflected in @xmath62 ) , the method used to determine @xmath14 for these sources differs . in order to ensure that the lr results for stars / qsos are not dominated by small number statistics , we assume a flat prior on @xmath14 , normalised to retain @xmath63 ( figure [ fig : nm ] )
. we can correct our value for @xmath51 for the clustering of sdss sources by simply dividing @xmath51 by @xmath64 ( remembering that @xmath24 is measured in degrees ) , giving a clustering corrected value of @xmath65 .
this value is broadly consistent with the recent results of dunlop et al .
( 2010 ) , who recover optical counterparts to 8 out of 20 250@xmath0 m sources brighter than 36@xmath66 in data from the blast observations of the goods - south field to a comparable @xmath67band magnitude ( albeit with lower angular resolution at 250@xmath0 m and much more sensitive optical , infrared and radio data ) , while dye et al .
( 2009 ) found 80 counterparts to the 175 blast 250@xmath0 m sources brighter than 55@xmath66 down to similar magnitude limits in @xmath9 or @xmath68band data ( s. dye , private communication ) . to account for the fact that an _ herschel_atlas source may have more than one possible counterpart
, we also define a reliability @xmath69 for each object @xmath70 being the correct counterpart out of all those counterparts within @xmath44 , again following sutherland & saunders ( 1992 ) : where the lr values have been determined for the resolved and unresolved counterparts separately ( see figure [ fig : lrhist ] ) .
the reliability is a key statistic ; we recommend using only those counterparts with reliability @xmath72 for analysis , since this ensures not only that the contamination rate is low ( see below ) , but also that only one @xmath9band source dominates the far infrared emission ( as required for e.g. deriving spectral energy distributions for 250@xmath0m selected galaxies in the _ herschel_atlas catalogue , smith et al .
_ in prep _ )
this is more conservative than other works in the literature ( e.g. chapin et al . ,
2010 ) , where the chosen lr limit was defined based on a 10 percent sample contamination rate .
sources in the spire sdp catalogue .
the lr values for the resolved sources ( i.e. galaxies ) are shown as the solid histogram , with unresolved sources shown as the dotted histogram .
middle : reliabilities for each counterpart .
once more , the solid histogram represents the resolved sources , while the dotted histogram represents the unresolved sources .
there are a total of 2423 sources which have a reliability @xmath73 , of which five are unresolved using the star / galaxy separation criteria of baldry et al .
bottom : the variation of the reliability as a function of the likelihood ratio .
this is not a linear relation since some sources have more than one counterpart with a high likelihood ratio .
there are 263 sdss @xmath9band sources with reliability @xmath74 but @xmath75 ( the value above which @xmath4 for a single counterpart within the 10.0 arcsec maximum search radius )
. these may be interacting systems , as discussed in section [ mergers ] .
these sources also demonstrate a possible limitation of the lr method , since the method implicitly assumes that there is only one true counterpart to a given 250@xmath0 m source . ] as a result we expect 103 false ids in our sample , which corresponds to a contamination rate of 4.2% . for those investigations
in which it is desirable only to determine whether an optical source is associated with an _
herschel_atlas object ( with additional caveats about lensed sources and the de - blending efficiency in the optical catalogue ) , it is sufficient to use a likelihood ratio cut ( e.g. @xmath77 , i.e. the source is 5 times more likely to be associated with the sub
millimetre object than it is to be a chance superposition of sources ) .
this aspect of the likelihood ratio technique is discussed in more detail in section [ mergers ] . in table
[ tab : nmatch ] , we present the number of possible optical counterparts within 10.0 arcsec of the 250@xmath0 m sample , including the relative fractions of reliable associations . only half of the 250@xmath0
m sources with a single optical counterpart within the search radius are deemed reliable .
.the distribution of the number of sdss @xmath9band sources within 10.0 arcsec of the 250@xmath0 m positions , and the fraction of reliable counterparts .
there are 2869 sources with only one possible match within 10.0 arcsec , and yet only 1389 of these are determined to be reliable ; the vast superiority of the lr technique over a simple nearest - neighbour algorithm is evident . [ cols="^,^,^,^",options="header " , ] we also compared the results of our likelihood ratio analysis to data from the faint images of the radio sky at twenty centimetres ( first ) survey ( becker , white & helfand , 1995 ) .
the first survey covers 9,000 square degrees of sky with a resolution of 5 arcsec , with a source density of approximately 90 per square degree brighter than the detection threshold of 1@xmath66 . at these relatively bright flux limits ,
the source population is dominated by active galactic nuclei ( agn ) rather than star forming sources ( e.g. wilman et al . , 2008 ) ; as a result the overlapping population of sources between the _ herschel_atlas and first catalogues is not expected to dominate the number counts . to make the comparison between our lr analysis and first sources
, we used the frequentist identification procedure of downes et al .
( 1986 ) , commonly used to quantify the formal significance of possible counterparts to sub - millimetre galaxies in radio survey data ( e.g. lilly et al . , 1999 , ivison et al .
2007 ) . in this procedure , the statistic used to assess the probability that a nearby radio source is _ not _ associated with the spire source is @xmath78 , where @xmath9 is the angular distance between the spire source and the radio source , @xmath79 is the flux density of the radio source , and @xmath80 is the surface density of radio sources with flux densities greater than this . for each spire source
, we looked for radio sources in the first catalogue within 10.0 arcsec , and treated the radio source with the lowest value of @xmath81 as the one most likely to be associated with the spire source .
we used a monte - carlo simulation ( e.g. eales et al . 2009 ) to determine the probability distribution of @xmath82 on the null hypothesis that there are no genuine associations between radio sources and spire sources .
we then used this probability distribution to determine the probability that each measured value of @xmath82 would have occured by chance .
we call this probability @xmath83 . of the 6621 @xmath18 250@xmath0 m spire sources , 105 have radio counterparts within 10.0 arcsec , all with values of @xmath84 . however , this does not take account of the fact that with such a large sample of spire sources one expects to find some low values of @xmath83 even if there were no genuine associations between the spire sources and first objects .
we used a monte - carlo simulation to determine that 15 of the 105 associations are likely to be spurious . to correct for this , we calculated a new probability for each association , @xmath85 , where @xmath86 is a constant that we calculated using @xmath87 .
we took the conservative decision to treat associations with @xmath88 as counterparts which are likely to be genuine , which rejected 29 of the original 105 associations .
there were a total of 76 spire sources with @xmath89 counterparts , and each of these was scrutinised using the first and sdss images displayed side by side with the downes et al . and lr analysis overlaid . in this manner , we compared the results of the two independant identification methods . in forty - two cases ,
the @xmath89 radio counterpart is also identified as having @xmath90 in the @xmath9band data , and the two methods choose the same counterpart .
there are thirty spire sources with high quality ( @xmath89 ) first counterparts which we do not recover in our lr analysis , including twenty three spire sources which do not have any @xmath9band counterparts in our sdss dr7 data ( presumably distant , optically
faint radio sources ) . of the remaining seven sources with @xmath89 first counterparts : * four counterparts are detected in the optical data but have low reliabilities due to their faint magnitudes , or large separations in comparison to the value of @xmath17 derived based on the 250@xmath0 m source snr .
* two sources have multiple , possibly interacting components with @xmath91 but @xmath92 , only one of which is a radio source ( these sources are discussed in section [ mergers ] ) . * in one further instance , the radio source has a double - lobed structure ( a so - called fr - ii , following fanaroff & riley , 1974 ) , not coincident with either the dust emission or the starlight in the plane of the sky .
the lobes of this frii are extremely bright ; as a result , the @xmath93 statistic suggests that there is a low probability of a chance association , even though the separation between the spire position and the first centroid is large .
the lr technique identifies the apparent host galaxy aligned at the centre , between the two luminous radio jets as having @xmath94 due to its large separation ( @xmath95 arcsec ) from the spire centroid ; this is an example of the limitations of the downes et al . method .
however , these possibilities do not contaminate the 250@xmath0 m selected sample with incorrect associations .
there are however , four instances where distinct counterparts have @xmath96 and @xmath97 ; here , the opposite is potentially true and the two methods conflict .
these sources have derived reliabilities of 0.87 , 0.98 , 0.81 and 0.93 as compared with distinct downes et al .
counterparts with @xmath93 statistics of 0.08 , 0.07 , 0.02 and 0.19 , respectively .
these sources are shown in figure [ fig : downes_lr_dis ] , in which the 10.0 arcsec search radius centred on the 250@xmath0 m position is shown in red , any unreliable optical counterparts in black , the reliable optical i d in light blue , and the radio contours overlaid in royal blue . in two of the four cases , the additional sources implied by the radio data are visible in @xmath98band observations from viking ( sutherland , 2009 ) , indicating that these sources are not merely effects of the larger positional uncertainty in first as compared with sdss .
futhermore , three of the four sources have spire colours @xmath99 , suggesting high redshifts ( @xmath6 ) or cold dust temperatures , with the former being at odds with the photometric redshifts of their most reliable counterparts ( @xmath100 ) .
sources with similar spire colours and low - redshift counterparts are discussed in more detail in section [ submmcolours ] .
finally , we note that probabilistic arguments such as those discussed here will inevitably present apparent disagreements for a small number of sources within large samples . in the remaining 101 out of 105 cases however , the results of our lr analysis are consistent with those using the first catalogue and the @xmath93 statistic , and
crucially we recover an additonal 2,348 counterparts , compared with 31 extra counterparts to the 250@xmath0 m sources obtained by using only the radio data . an additional check on the identification process
was conducted by searching for mid infrared data from the _ spitzer space telescope _ heritage archive , in order to compare the reliabilities from our @xmath9band catalogue with near- and mid
infrared images between 3.6 and 160@xmath0 m .
four sets of observations were found which overlapped with the _ herschel_atlas sdp observations these data can be used to examine the regions surrounding the spire ids for additional sources which may not be present in the @xmath9band catalogue used for the identification process , as a visual check on the effectiveness of the lr technique .
there are a total of 49 sources that have _ spitzer _ data , and although these data vary in sensitivity , there is no evidence that would suggest a mis - identification from the @xmath9band catalogue .
such indications of wrong ids would include reliable ( @xmath72 ) @xmath9band counterparts indicated for spire sources which have previously unrevealed bright _ spitzer _
sources nearer to the centre of the spire centroid .
indeed , in one case in particular ( h atlas j090913.2 + 012111 ) , the sensitive irac data reveal the power of the lr technique .
although there are three potential counterparts in the sdss dr7 @xmath9band catalogue all within 6 arcsec of the spire centroid , they have all been given low reliability ( @xmath101 , and also @xmath102 ) .
the irac 3.6@xmath0 m data reveal a fourth candidate counterpart within 1 arcsec of the spire position , which is presumably the true counterpart .
the @xmath9band and irac 3.6@xmath0 m data are presented in figure [ fig : rband_irac1 ] , with the various source positions overlaid to demonstrate the robustness of the lr method for this particular source , but also the need for longer - wavelength observations in order to be able to reliably identify the counterparts to higher redshift sources .
the forthcoming data from the vista kilo - degree infrared galaxy ( viking ) survey and from the wide - field infrared survey explorer ( _ wise _ duval et al . , 2004 )
satellite will enable this .
this example also highlights one crucial advantage of using the lr technique for _ herschel _ surveys rather than opting simply for the downes method ; the lr method takes into account the fact that not every source has a counterpart that is brighter than the detection limit in ancillary survey data . | we present a technique to identify optical counterparts of 250@xmath0m - selected sources from the _ herschel_-atlas survey . of the 6621 250@xmath0 m @xmath1mjy sources in our science demonstration catalogue
we find that @xmath2 percent have counterparts brighter than @xmath3mag in the sloan digital sky survey . applying a likelihood ratio technique
we are able to identify 2423 of the counterparts with a reliability @xmath4 .
this is approximately 37 percent of the full 250@xmath0 m catalogue .
we have estimated photometric redshifts for each of these 2423 reliable counterparts , while 1099 also have spectroscopic redshifts collated from several different sources , including the gama survey .
we estimate the completeness of identifying counterparts as a function of redshift , and present evidence that 250@xmath0m - selected _ herschel_-atlas galaxies have a bimodal redshift distribution .
those with reliable optical identifications have a redshift distribution peaking at @xmath5 , while sub - mm colours suggest that a significant fraction with no counterpart above the r - band limit have @xmath6 .
we also suggest a method for selecting populations of strongly - lensed high redshift galaxies .
our identifications are matched to uv
nir photometry from the gama survey , and these data are available as part of the _ herschel_-atlas public data release .
[ firstpage ] galaxies : local , galaxies : infrared , galaxies : star - forming , methods : statistical , submillimetre : galaxies |
it is widely believed that the most natural and appealing explanation of the recent neutrino oscillation results is provided by the seesaw mechanism @xcite incorporated into extensions of the standard model that include a local @xmath0 symmetry .
the simplest models with local @xmath0 symmetry are the left - right symmetric models @xcite based on the gauge group @xmath6 .
these models have the additional virtue that they explain the origin of parity violation in weak interactions as a consequence of spontaneous symmetry breaking in very much the same way as one explains the strength of the weak interaction in the standard model .
stability of the higgs sector under radiative corrections calls for weak scale supersymmetry as in the minimal supersymmetric standard model ( mssm ) .
it has recently been shown that if the mssm is embedded into a left
right symmetric framework at a high scale @xmath7 gev , as suggested by neutrino oscillation data and by gauge coupling unification , it helps solve some important problems faced by the mssm , viz . , the susy cp problem @xcite , the strong cp problem @xcite and the @xmath8 problem .
supersymmetric models with such a high scale embedding are therefore attractive candidates for physics beyond the standard model .
it was noted many years ago @xcite that the electric charge formula of the left
right symmetric models , @xmath9 , allows one to conclude from pure group theoretic arguments that parity symmetry breaking implies a breakdown of @xmath0 symmetry as well with the constraint that @xmath10
. this simple relation is profoundly revealing .
it says that the neutrinos must be majorana particles since the lepton number breaking terms in the theory must obey @xmath11 selection rule .
this conclusion follows directly if higgs triplets are used to break @xmath12 symmetry since @xmath13 for triplets , it also holds when higgs doublets are used for this purpose , since gauge invariance requires the presence of two such doublets in the mass term for the neutrinos .
secondly , for purely hadronic baryon number violating processes , baryon number must change by at least two units , @xmath14 .
this means that models based on left
right symmetric gauge structure can lead to the process where a neutron transforms itself into an antineutron ( @xmath2 oscillation@xcite ) , while they may forbid the decay of the proton , which is a @xmath15 process .
while the above group theory argument predicts the existence of @xmath2 oscillation in left
right symmetric models , its strength will depend on the details of the model . using simple dimensional analysis
it is easy to find that the lowest dimensional operators that contributes to @xmath2 oscillation are six quark operators , a typical one being @xmath16 .
this operator has dimension 9 and therefore the coupling strength scales as @xmath17 , where @xmath18 is the scale of new physics .
it is natural to identify @xmath18 with the scale of @xmath0 ( or parity ) breaking . the current lower limit on @xmath2 oscillation time , @xmath19 sec @xcite , sec @xcite .
] implies an upper limit @xmath20 gev@xmath21 . for @xmath2 oscillations to be observable
then , the scale @xmath18 should be rather low , @xmath22 gev .
one class of models where @xmath23 transition manifests itself through higgs boson exchange has been discussed in ref .
there it was shown that if the @xmath6 model is embedded into the @xmath24 gauge group , then @xmath2 oscillations can arise at an observable level if the @xmath25 breaking scale is in the 100 tev range . in these models , @xmath2 oscillation amplitude
is intimately tied to an understanding of small neutrino masses via the seesaw mechanism as well as the breaking of quark lepton degeneracy implied by @xmath25 symmetry .
the same higgs field that breaks @xmath25 and generates heavy majorana masses for the right handed neutrinos also mediate @xmath2 oscillations here . with the scale of @xmath25 breaking in the 100 tev range
, these models would appear to be incompatible with gauge coupling unification .
furthermore , such a low scale of parity breaking would not yield naturally neutrino masses in the range suggested by current experiments .
if we raise the scale of parity/@xmath25 breaking to values above @xmath26 gev , so that small neutrino masses in the right range are generated naturally , then @xmath2 transition amplitude becomes unobservably small in these models .
oscillation was noted in the context of a susy @xmath27 model in ref .
these models possess accidental symmetries that lead to light ( @xmath28 gev ) diquark higgs bosons even though the scale of parity violation is high . as a result
, the @xmath2 oscillation operator can have observable strength .
unification of gauge couplings is however difficult to achieve in these models . ]
does the above arguments mean that @xmath2 oscillations are beyond experimental reach based on current neutrino oscillation phenomenology ?
in this letter we will show that this is not the case in a class of attractive seesaw models with local @xmath0 symmetry .
we will see that in these models a new class of @xmath29 operators is induced as a consequence of parity breaking .
these operators lead to observable @xmath2 oscillation despite the scale @xmath30 of parity breaking being close to the conventional gut scale of @xmath31 gev .
in fact , @xmath32 increases with @xmath30 and therefore one has the inverse phenomenon that increasing @xmath30 leads to stronger @xmath2 oscillation amplitude .
interestingly , the scale @xmath30 implied by neutrino masses is such that @xmath2 oscillation should be accessible experimentally with a modest improvement in the current limit .
we obtain an _ upper _ limit of @xmath33 sec in this class of models . this prediction becomes sharper in a concrete model where flavor symmetries reduce considerably the uncertainties in the estimate of @xmath4 .
we emphasize that our upper limit is derived in the context of conventional seesaw models of neutrino mass without using any special ingredients to enhance @xmath2 oscillation amplitude .
this should provide new impetus for an improved experimental search for @xmath2 oscillations .
the basic framework of our model involves the embedding of the mssm into a minimal susy left
right gauge structure at a scale @xmath30 close to the gut scale .
the electroweak gauge group of the model , as already mentioned , is @xmath34 with the standard assignment of quarks and leptons left handed quarks and leptons ( @xmath35 ) transform as doublets of @xmath36 , while the right handed conjugate ones ( @xmath37 ) are doublets of @xmath12 .
the quarks @xmath38 transform under the gauge group as @xmath39 and @xmath40 as @xmath41 , while the lepton fields @xmath42 and @xmath43 transform as @xmath44 and @xmath45 respectively .
the dirac masses of fermions arise through their yukawa couplings to two higgs bidoublet @xmath46 , @xmath47 .
the @xmath48 symmetry is broken down to @xmath49 in the supersymmetric limit by @xmath50 doublet scalar fields , the right handed doublet denoted by @xmath51 accompanied by its left
handed partner @xmath52 .
anomaly cancellation requires the presence of their charge conjugate fields as well , denoted as @xmath53 and @xmath54 .
the vacuum expectation values ( vevs ) @xmath55 break the left
right symmetry group down to the mssm gauge symmetry .
a singlet @xmath56 is also used to facilitate symmetry breaking in the susy limit .
it has recently been shown that if there exists a @xmath57 r symmetry , the minimal model just described will solve the strong cp problem and the susy phase problem based on parity symmetry @xcite .
furthermore , the @xmath8 term will have a natural origin .
one possible @xmath57 assignment was given in ref .
@xcite . here
we present a slight variant , which yields the same superpotential at the renormalizable level as in ref .
@xcite and thus preserves all its success . under this @xmath57 , the superpotential @xmath58 changes sign , as do @xmath59 and @xmath60 .
the quark fields @xmath61 are even , while @xmath62 transform as @xmath63 .
the fields @xmath64 are all odd under @xmath57 .
the gauge invariant superpotential consistent with this @xmath57 r symmetry at the renormalizable level is @xmath65 this superpotential breaks the gauge symmetry to that of the standard model in the susy limit without leaving any unwanted goldstone bosons and induces realistic quark masses and mxings .
the baryon number violating processes as well as neutrino masses arise in this model from higher dimensional operators induced by planck scale physics
. they will be the main focus of the rest of the paper .
we shall pay special attention to the relation between the neutrino mass and the @xmath2 oscillation time .
the relevant dimension four operators in the superpotential which are scaled by @xmath66 and are allowed by the @xmath57 symmetry are : @xmath67~,\nonumber\\ { \cal o}_2 & = & f'\left [ q^cq^cq^c\bar{\chi^c } + qqq\bar{\chi } \right]~.\end{aligned}\ ] ] operator @xmath68 gives rise to majorana masses for @xmath69 of order @xmath70 . combining this with the dirac neutrino masses arising from eq .
( 1 ) , small neutrino masses will be generated by the seesaw mechanism .
for @xmath71 gev , the magnitude of the light neutrino masses are in the right range to explain the atmospheric and the solar neutrino oscillation data .
operator @xmath72 , which is also invariant under the @xmath57 , leads to baryon number violation .
while @xmath73 could have their origin in quantum gravity , they may also be induced by integrating out vector states that have @xmath57inviariant masses of order the planck scale .
note that operators such as @xmath74 are not allowed by the @xmath57 symmetry .
if they were present along with @xmath72 , they would lead to rapid proton decay .
note also that the well known proton decay operator @xmath75 is not allowed by the @xmath57 symmetry . in any case its presence would not have been a problem since it is scaled by the planck mass and therefore can lead to a proton lifetime consistent with the present lower limit . to see the connection between neutrino masses and the @xmath2 oscillation time @xmath76 qualitatively
, first we note that the operator @xmath68 leads to the majorana mass for the right handed neutrino @xmath77 .
the seesaw formula then leads to the relation @xmath78 . on the other hand ,
the operator @xmath72 leads to a @xmath29 operator with strength @xmath79 . leaving aside the details of the flavor structure of @xmath72 and how actually @xmath80 arises
, it is clear that we have a simple linear relation between the neutrino masses and the @xmath2 oscillation time : @xmath81 where @xmath82 is a dimensional constant which depends only on the details of weak scale physics and does not involve the high scale @xmath30 .
we will evaluate @xmath82 in the next section .
this simple relation makes it clear that our present knowledge of the neutrino masses allows a direct prediction of the @xmath2 oscillation time in the context of the supersymmetric left - right models broken by doublet higgs fields .
let us now proceed to examine the expected @xmath2 oscillation time resulting from the @xmath29 operator @xmath72 of eq .
an important point to note here is that since @xmath72 is a superpotential term with antisymmetric color contraction it must have antisymmetric flavor contraction as well .
the flavor structure of this operator is then of the type @xmath83 , @xmath84 or @xmath85 in terms of the superfields .
we must then use flavor mixings to obtain the fermionic operator of the type @xmath86 and then the six quark @xmath2 operator . the dominant contribution to
this process comes from the feynman diagram shown in fig . 1 which proceeds through the exchange of a gluino and squarks @xcite and involves two @xmath87 mixings .
the strength of the @xmath23 operator resulting from fig .
1 can be estimated to be @xmath88 ^ 2f'^2 \over { m_{\tilde{g } } m_{\tilde{q}}^4}}~,\end{aligned}\ ] ] where @xmath89 is the gluino mass , @xmath90 is the squark mass and @xmath91 is the @xmath87 mixing angle .
the effective baryon number violating @xmath84 yukawa coupling in the superpotential is parametrized here as @xmath92 ( see eq . ( 2 ) ) .
( fig1nnbar.epsf width 8 cm ) let us first discuss the origin of the flavor mixing that changes a @xmath84 operator to the required @xmath86 operator .
the dominant source for this in the present context turns out to be the mixing of @xmath93 with @xmath94 .
such mixings occur in the left
right supersymmetric model since the right handed quark mixings are physical above the scale @xmath30 .
the renormalization group evolution of the soft susy breaking mass parameters between @xmath95 and @xmath30 will then induce mixings in the right handed down squark sector proportional to the top
quark yukawa coupling and the right handed ckm mixings .
this is analogous to the rge evolution in the mssm inducing squark mixing in the left handed down squark sector proportional to the left handed ckm angles and the top
quark yukawa coupling .
we estimate this right handed @xmath87 mixing to be @xmath96 this estimate is obtained by integrating out the rge between @xmath95 and @xmath30 assuming universality of masses at @xmath95 @xcite . since above @xmath30 , both @xmath97 and @xmath98 are part of the same @xmath12 multiplet , unlike in the mssm , @xmath98 yukawa coupling is of order one . in this momentum range ,
quark yukawa coupling reduces the mass of @xmath93 . in going to the physical basis of the quarks
, this effect will induce the squark mixing quoted in eq .
( 5 ) . for the numerical estimate we took @xmath99 and @xmath100 and @xmath101 gev for illustration .
there is a second source of flavor violation that induces @xmath102 mixing in general susy models .
that is the baryon number violating yukawa couplings themselves . if we write in standard notation , the effective @xmath103violating superpotential arising from eq .
( 2 ) as @xmath104 , the rge evolution from planck scale to the weak scale will induce @xmath87 mixing proportional to @xmath105 .
for example , if we keep only the couplings involving the @xmath106 quark , viz . ,
@xmath107 and @xmath108 , we can estimate the induced @xmath91 by integrating the relevant rge @xcite to be @xmath109 recalling that @xmath110 , we see that while this source of flavor mixing may not be negligible , it would be typically smaller than the ones from the right handed quark mixings of eq .
a third source of flavor violation relevant for @xmath2 oscillations has been identified in ref .
@xcite involving the exchange of the wino .
such diagrams will have an electroweak loop suppression and a chirality suppression necessary to convert the left handed squark to the right handed one .
we find that this contribution to @xmath80 has a suppression factor given approximately by @xmath111 ^ 2 \sim 1 \times 10^{-9}$ ] ( valid for small @xmath112 ) which is about two orders of magnitude smaller in this class of models compared to the gluino exchange diagram of fig . 1 .
one has to calculate the hadronic matrix element of the six quark operator in order to obtain the @xmath4 .
this has been discussed in several places in the literature@xcite .
the calculations of this `` conversion '' factor can be done using crude physical arguments , according to which one has to multiply the @xmath32 by @xmath113 to obtain @xmath80 where @xmath114 is the baryonic wave function for three quarks inside a nucleon . on dimensional grounds
, one can deduce that @xmath115 , which implies that @xmath116 gev .
more detailed bag model calculations have been carried out .
rao and shrock in ref .
@xcite quote this conversion factor to be @xmath117 .
we shall use this number for our numerical illustrations . combining this matrix element with eq .
( 4)-(5 ) we obtain @xmath118 we can rewrite eq . ( 7 ) in a form that makes the connection with the neutrino mass more transparent .
the mass of @xmath119 can be expressed through the seesaw formula from eq . ( 2 ) as @xmath120 where @xmath121 denotes the dirac mass of @xmath119 . eliminating the high scale @xmath30 from this
, we have from eq . ( 7 ) , @xmath122 since the value of @xmath123 can be determined from the atmospheric neutrino data under certain assumptions , we conclude that within the seesaw framework , measurement of @xmath2 oscillation will be a measure of the dirac mass of the tau neutrino .
this can then be used as a way to discriminate between models of neutrino masses . to see the specific prediction for @xmath2 oscillations within the context of the class of models under consideration ,
we need to know the @xmath119 dirac mass . we can estimate it from the following relations for the dirac masses of the third generation quarks and leptons in the susy left
right model : @xmath124 here @xmath125 are the higgs mixing parameters obtained from eq .
( 1 ) ( eg : @xmath126 ) and @xmath127 are the vevs of the mssm doublets . from eq . ( 9 ) it follows that in the limit of @xmath128 and @xmath129 , we get @xmath130 .
since at such high scales @xmath131 , this predicts @xmath132 .
in fact , we find that unless the two terms in eq .
( 9 ) for @xmath121 are precisely canceled , the dirac mass of @xmath119 will be approximately equal to @xmath133 . using @xmath134 and @xmath135 , @xmath136
, we get a value for the @xmath2 oscillation time which is tantalizingly close to the present experimental lower limit @xcite . for values of @xmath137 10 times smaller than @xmath133 , and taking the supersymmetric particle masses as large as 1 tev
, we see that @xmath4 is less than @xmath138 seconds , case should be about 2 , which would reduce the estimate of @xmath4 by a factor of 2 . ]
which is in the range accessible to a recently proposed experiment @xcite .
it would thus appear that a search for neutron - antineutron oscillation will provide an enormously useful window into neutrino mass models and as such a powerful constraint on the nature of new physics beyond the standard model .
the prediction for @xmath2 oscillations can be sharpened if we make use of flavor symmetries to determine the coefficients @xmath139 and @xmath140 in eq .
we illustrate this with a specific choice of flavor symmetry @xcite taken to be @xmath141 .
the first two families of fermions form doublets of @xmath142 and have a @xmath143 charge of @xmath144 while the third family fermions are singlets under both groups .
this flavor symmetry is broken by a pair of doublets @xmath145 and singlets @xmath146 . allowing for effective operators suppressed by a scale @xmath18 larger than the vevs of these fields provides a natural explanation of the fermion mass and mixing angle hierarchy .
if we choose @xmath147 and @xmath148 , a resonable fit to all quark and lepton masses is obtained , including neutrinos , for @xmath149 and @xmath150 and all dimensionless couplings being order one @xcite . in this model
, we can estimate the couplings @xmath139 and @xmath140 from the horizontal quantum numbers .
they are @xmath151 and @xmath152 , so that could have been of order one but in the horizontal model or ref .
@xcite , due to large @xmath153 mixing , it is the @xmath154 flavor entry that dominates the atmospheric neutrino mass difference and hence the horizontal suppression factor @xmath155 .
the @xmath156 factor is due the fact that the operator must be invariant under @xmath143 . ]
@xmath157 this estimate leads to @xmath158 sec from eq .
( 8) . allowing for uncertainties of order 1 in this estimate , we expect that @xmath4 not to exceed about @xmath159 sec .
before we conclude a few comments are in order : 1 .
the model becomes unacceptable as soon as @xmath160 is embedded into a higher symmetry such as @xmath25 or so(10 ) group because in that case , the @xmath29 operator described in eq .
( 2 ) is accompanied by other r - parity violating operators coming from the same higher dimensional operator @xmath73 due to the higher symmetry .
together , they would lead to unacceptable proton decay rate .
thus observation of @xmath2 oscillation would be a signal of an explicit @xmath160 symmetry all the way upto the planck ( or string ) scale .
2 . baryogenesis has to proceed through a weak scale scenario since the @xmath29 interactions in the model are in equilibrium down to the tev scale and will wash out any primordial baryon or lepton asymmetry .
we note that the baryon number violating interactions contained in @xmath72 themselves can potentially be the source of weak scale baryogenesis @xcite .
the lightest neutralino in this model is unstable and will decay via @xmath161 modes due to the presence of the effective @xmath162 operator @xmath72 .
this prediction is directly testable at colliders .
an alternative candidate for dark matter must be sought .
in conclusion , we have found that in a large class of seesaw models for neutrino masses , despite the high scale of sessaw dictated by the current neutrino oscillation data , neutron
antineutron oscillation is in the observable range .
in fact , unless the dirac masses of neutrinos are far below those deduced under simple and reasonable assumptions , we predict an upper bound on the neutron - antineutron oscillation time in the range of @xmath163 sec .
this is very close to the present experimental lower limit on @xmath2 oscillations . in the most conservative theoretical scenario ,
the measurement of @xmath2 oscillation time would be a measure of the dirac mass for the tau neutrino , given the values of squark masses .
this in itself would be an extremely interesting result , since it would discriminate among theoretical models of neutrino masses .
this is apart from the fundamental importance that any observation of baryon number violation will carry .
we therefore strongly urge a new experimental search for neutron antineutron oscillation .
the work of ksb has been supported in part by doe grant # de - fg03 - 98er-41076 , a grant from the research corporation , doe grant # de - fg02 - 01er4864 and by the osu environmental institute .
rnm is supported by the national science foundation grant no .
ksb is grateful to the elementary particle theory group at the university of maryland for the warm hospitality extended to him during a visit when this work was completed .
one of the authors ( r. n. m. ) would like to thank y. kamyshkov and r. shrock for discussions .
m. gell - mann , p. ramond and r. slansky , in _
supergravity _ ,
p. van niewenhuizen and d.z .
freedman ( north holland 1979 ) ; t. yanagida , in proceedings of _ workshop on unified theory and baryon number in the universe _ , eds . o. sawada and a. sugamoto ( kek 1979 ) ; r. n. mohapatra and g. senjanovi , phys .
lett . * 44 * , 912 ( 1980 ) .
imb collaboration , t.w .
al . , phys .
lett . * 52 * , 720 ( 1984 ) ; kamiokande collaboration , m. takita et .
al . , phys . rev .
* d34 * , 902 ( 1986 ) ; frejus collaboration , ch .
berger et .
al . , phys . lett . *
b240 * , 237 ( 1990 ) . s. rao and r. shrock , phys
116b * , 238 ( 1982 ) ; j. pasupathy , phys . lett . * 114b * 172 ( 1982 ) ; riazzuddin , phys . rev . *
d25 * , 885 ( 1982 ) ; s. misra and u. sarkar , phys . rev . *
d28 * , 249 ( 1983 ) ; s. fajfer and r. j. oakes , phys . lett . *
b132 * , 432 ( 1983 ) . | we show that in a large class of supersymmetric models with spontaneously broken @xmath0 symmetry , neutron
antineutron oscillations occur at an observable level even though the scale of @xmath0 breaking is very high , @xmath1 gev , as suggested by gauge coupling unification and neutrino masses .
we illustrate this phenomenon in the context of a recently proposed class of seesaw models that solves the strong cp problem and the susy phase problem using parity symmetry .
we obtain an _ upper _ limit on @xmath2 oscillation time in these models , @xmath3 sec .
this suggests that a modest improvement in the current limit on @xmath4 of @xmath5 sec will either lead to the discovery of @xmath2 oscillations , or will considerably restrict the allowed parameter space of an interesting class of neutrino mass models .
epsf.tex ( # 1 width # 2)=#2 |
the gravitational wave ( gw ) is the propagation of ripples of the space - time curvature with speed of light .
although @xcite predicted the existence of gravitational radiation ( gr ) , the gws have never been directly detected yet . from 30 years observations , @xcite found that the binary pulsar psr 1913 + 16 , discovered by @xcite , exhibited the decrease of the orbital period and the amount of decrease exactly coincide with the prediction of general relativity .
there are several types of astronomical objects that can produce significant amount of gravitational waves : core - collapse supernovae , spinning neutron stars ( e.g. , * ? ? ?
* ) , compact binary coalescences ( e.g. , * ? ? ?
* ) , supermassive black holes ( smbh ) ( e.g. , binary smbh merger , * ? ? ?
* ) ( extreme mass ratio inspirals , emri , * ? ? ?
* ; * ? ? ?
* ) and cosmological density fluctuations ( e.g. , * ? ? ?
* ) . among those ,
the compact binary coalescence ( cbc ) involving neutron stars ( ns ) or stellar mass bh is the primary targets for the second generation of ground based detectors such as advanced ligo and virgo . up to now
, only ten ns - ns binary pulsars have been discovered in our galaxy including psr 1913 + 16 , and a half of them will merge within a hubble time @xcite .
based on these binaries and theoretical studies , the rate of ns - ns merger within the horizon of the advanced detectors is estimated to lie between 0.4 and 100 per year @xcite .
these ns - ns binaries are thought to have been evolved from primordial binaries .
obviously other types such as bh - ns and bh - bh binaries can form by the evolution of binaries of high mass stars , but their rates are highly uncertain especially because of the lack of observed systems . since the detection of the gws coming from the cbc event will provide us with the information regarding the masses of the binaries , we should be able to distinguish different types of binaries through the observations of gws in the future .
compact binaries can also be formed dynamically in stellar systems : three - body processes and dissipative two - body processes by tidal effect @xcite or gw emission ( hereafter gr capture , * ? ? ?
* ; * ? ? ?
stellar systems such as globular clusters ( gc ) and nuclear star clusters ( nc ) at the galactic nuclei provide good environments for the dynamical formation of compact binaries . in gcs ,
when the core is dense enough , compact binaries can be formed by three - body encounters .
these binaries become more compact through close encounters with other stars and is eventually kicked out from gcs when their orbital separations become very small .
some of the ejected binaries merge within a hubble time in galactic field @xcite . on the other hand , binary formation by three - body processes
is suppressed by the existence of massive bh ( mbh ) in galactic nuclei @xcite .
instead , coalescences of primordial compact binaries can be driven by the orbit coupling with central mbh , as known as the kozai effect @xcite , and could occur within hubble time @xcite .
compact binaries , especially bh - bh binaries , also can be formed by gr capture in ncs due to the high stellar density and velocity dispersion at the vicinity of mbh @xcite .
the black holes are expected to be highly concentrated in the central parts through the dynamical friction against low mass stars , providing a favorable environment for the close encounters between black holes .
such captured compact binaries usually have large eccentricities with small pericentre distance , and thus , their waveforms will be different from the ones produced by the circular binaries @xcite .
numerous authors made estimates of the detection rates of gws from cbcs with present- and planned- gw detectors using various methods : population synthesis models for primordial binaries @xcite , monte - carlo simulations for gcs @xcite and fokker - planck simulations for ncs @xcite , and estimated that a few tens of events will be detected by new generation gw detectors every year ( see , for example , * ? ? ?
. however , most of these studies are based on simplified models and assumptions on the evolution of the stellar systems and the binaries .
there remain substantial uncertainties because of difficulties in accurately modeling of the systems with the large number of stars .
direct @xmath4-body approach is impossible for realistic systems , and therefore , statistical approaches such as fokker - planck models and monte - carlo simulations have been used , so far . in this paper , we focus on the binary formation of bhs by gr capture in ncs by direct @xmath4-body simulations .
although , we can not use realistic number of stars , we try to deduce important information regarding the binary formation and evolution based on scaled - down and simplified version of @xmath4-body simulations .
our study should provide useful comparison with the statistical models .
this paper is organized as follows . in
2 and 3 , we introduce the numerical method and the model for star clusters in galactic nuclei with a central mbh .
the dynamical evolution of our model is presented in 4 . in 5 ,
we describe binary formation in ncs and estimate the merger rate per galaxy . in 6 , the expected detection rate for new generation gw detectors is estimated . to give the interpretation for binary coalescences and their waveforms
, we implement post - newtonian approximations on the two - body motions .
an example waveform of a binary bh coalescence in milky - way - like galaxies is provided in 7 .
finally , we summarize in 8 .
ncs are very dense stellar systems located at the nuclei of galaxies , regardless of the type ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
their typical mass is @xmath5 @xcite , the size of ncs is comparable to that of galactic gcs @xcite , thus average stellar density of ncs is much higher than gcs .
it is also well known that most of galaxies host mbhs at the centre .
the coexistence and correlation of mbhs and ncs at the central region of galaxies have been studied by @xcite .
observational and theoretical ( e.g. , * ? ? ?
* ) studies for star clusters with the central mbh showed that the density and velocity dispersion increases steeply toward the mbh , following kepler s profile ( i.e. , @xmath6 ) . the modeling of stellar systems with a central black hole has been done by numerous authors ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ?
* ; * ? ? ?
in order to generate @xmath4-body realizations for ncs with mbh , we adopt ` adiabatic growth ' of the mbh as suggested by @xcite .
the mbh is assumed to grow with time as @xmath7 & t \le t_{\rm mbh}\\ m_{\rm mbh } & t > t_{\rm mbh } \end{cases } \end{aligned}\ ] ] where @xmath8 and @xmath9 are the final mass of mbh and the black hole growth time scale , respectively . during the growth of the mbh
, the stellar system is adjusted against the potential of the mbh .
the mbh is fully grown after @xmath9 , and the gravitational potential of the mbh is assumed to follow that of the plummer model @xmath10 where @xmath11 is the softening parameter for avoiding unexpected effect at the singularity .
@xcite noted that @xmath9 should be larger than the half - mass dynamical time to ensure the adiabatic growth of the mbh .
while they used the hernquist model as the initial density distribution , we used plummer model with half - mass dynamical time @xmath12 for standard @xmath4-body scaling ( i.e. , @xmath13 ) .
according to recent observations of the centre of the milky way ( i.e. , the vicinity of sgr a@xmath14 ) , the velocity dispersion of stars at about 1 parsec scale is nearly flat .
this implies that the nc is almost in the isothermal state .
however , it is not possible for the isolated stellar systems to become fully isothermal . in order to realize nearly isothermal sphere
we need a very large and massive system .
instead we model the nc as a isothermal sphere surrounded by external potential which is assumed to be fixed .
such an external potential could be provided by the galactic bulge .
@xcite have investigated a self - gravitating stellar systems embedded in an external potential well .
they considered a plummer external potential , @xmath15 where @xmath16 and @xmath17 are the mass and scale length of the plummer potential , respectively .
they revealed that if the external potential well is deep enough compared to the potential of embedded stellar system , the external potential well behaves like a heat bath .
the velocity dispersion of the embedded stellar system , therefore , becomes isothermal and there exists a quasi - equilibrium solution of a potential - density pair for the nearly isothermal stellar system ( see model 2 of fig .
10 in * ? ? ?
the bulge is one of galactic components , extending over few kpc scales .
nc and bulge are independent components of galaxies with different surface brightness profiles ( e.g. , * ? ? ?
in the milky way , the effective radius ( i.e. , half - light radius ) of the bulge is about 0.1 kpc , and the mass is estimated to be roughly @xmath18 . according to a dynamical model of galactic bulge suggested by @xcite , the kinematics of the bulge is affected by the mbh , sgr a@xmath14 , at inner parsec scale , and the velocity dispersion is nearly flat from 1 to 10 parsec and increases gradually at the large radii .
we assume that the the bulge is a sphere in this study for the simplicity .
this may not cause a serious effects on the dynamics of the nuclear cluster we are considering since the role of the bulge in our model is to confine the nuclear cluster within a few parsec .
in this study , we used the gpu accelerated version of nbody6 @xcite .
this code includes many efficient and accurate algorithms such as the fourth - order hermite integrator , the individual and block time steps , the ahmad - cohen neighbor scheme , the kustaaheimo - stiefel ( ks ) and chain regularization scheme @xcite .
recently , by using numerous stream processors of gpu devices , calculations of gravitational interactions among stars have been significantly accelerated through massive parallelism . force calculation in nbody6 code
is divided into _
regular _ force from distant particles whose gravitational potential change slowly and _ irregular _ force from neighboring particles that give strong gravitational fluctuations . in the gpu accelerated version , the regular forces are calculated by gpu machines with single precision while irregular force calculation that needs better accuracy is still done by cpus .
as we mentioned above , the external potential is composed of two parts , @xmath19 where @xmath20 is the keplerian potential due to the central mbh . in our implementation
, these external forces are added to the sum of regular forces .
all the calculations in nbody6 code use dimensionless time , length and mass units . from given unit of length @xmath21 in parsec and mean stellar mass @xmath22 in @xmath23 , the physical unit of velocity and time can be expressed as @xcite @xmath24 @xmath25 where @xmath4 is total number of stars .
table [ tblg1 ] shows model parameters of our simulations .
although galactic nuclei contain @xmath26 stars in a cubic parsec , it is hard to treat such a large number of particles for nbody6 code even with gpu machine .
we , therefore , used several different number of particles ranging from 10,000 to 100,000 in order to build a scaling relation .
we also made several simulations with different random seeds and took average in order to reduce the statistical noises .
the masses of mbh are chosen to be 10 and 20 % of the entire mass of cluster stars @xmath27 ( but excluding the mass of the plummer potential ) . because these masses are relatively larger than those of @xcite and @xcite , we also used longer black hole growth time scale @xmath28 .
the softening parameter of the mbh is fixed to @xmath29 , which is much smaller than the radius of influence of the mbh ( c.f .
, @xmath30 for @xmath31 ) .
we assumed that all stars have same masses ( i.e. , @xmath32 ) of stellar mass bhs with the mass of @xmath33 .
thus , the total mass of the cluster in physical unit becomes @xmath34 .
it is well known that there is a correlation between the mass of mbhs and the kinematics of bulges .
recently , @xcite updated the @xmath35 relation and found that the mass of mbh is roughly 0.2% of the bulge mass for @xmath36 ( also see * ? ? ? * ) .
for the external potential of a bulge , we fixed the mass and scale length of the plummer potential to 100 and 5 , respectively , which corresponds to @xmath37 and 0.002 for our models . under this potential
well , the embedded stellar system is expected to become nearly isothermal in a few half - mass relaxation time @xcite .
we also considered a model without the external plummer potential well ( model 0 ) for the purpose of comparison . [ cols="<,^,^,^,^,^,^,^,^ " , ] @xmath38mean mass fraction of bhs from @xcite .
+ @xmath39low , realistic and high detection rates depending on uncertainties .
+ @xmath40@xmath41 with snr 8 form @xcite divided by 2.26 , the correction factor for sky + location and orientation of sources .
the effect of redshift is not included .
+ @xmath42@xmath41 with snr 10 form fig . 16 in @xcite .
+ @xmath43@xmath41 with snr 8 form fig . 3 in @xcite .
+ table [ tblg2 ] shows the detection rates expected for advanced ligo . because we mainly considered the equal - mass models for only bhs ,
the detection rates are corrected by the factor of @xmath44 .
we estimated detection rates by using @xmath41 from different studies @xcite .
all @xmath41s are corrected for the orientation of sources .
note that @xmath41 from @xcite does not include the cosmological effect of the redshift .
@xmath41 from @xcite is for spinning bhs but independent of the spin of bhs for 10@xmath23 bhs .
we list three detection rates @xmath45 , @xmath46 and @xmath47 in table [ tblg2 ] for different models and different estimates of detection horizon ( @xmath41 ) .
we first calculated the ` reasonable ' expected detection rate @xmath46 using the merger rate given in equation ( [ eqns36 ] ) .
given several uncertainties including the variance of the number density , dynamical evolution and the lower limits of the mbh , we assume that the total range of uncertainty is a factor of 60 . thus ` low ' expected rate @xmath45 is simply obtained by dividing @xmath46 by @xmath48 while the ` high ' expected rate @xmath47 is calculated by multiplying the same factor .
the whole range of the expected detection rate lies between 0.02 @xmath49 14 yr@xmath3 depending on the maximum horizon distance and uncertainties .
these estimates are able to cover those of @xcite ( 5 - 20 yr@xmath3 ) for bhs around @xmath33 although the lowest value is quite smaller than that .
our estimations have some limitations ; ( 1 ) we ignore the initial mass function .
this mass function may affect not only the evolution of systems by the relaxation between mass components but also the merger rates for bhs with different masses .
( 2 ) we need to consider various range of @xmath50 ( e.g. , * ? ? ?
* suggested that there is a rough relation between the mass of mbh and nc . )
( 3 ) the exact calculation for snr is necessary in order to obtain more reasonable detection rates .
we considered the mass segregation and its effect only approximately .
these limitations would be considered in future works .
solving einstein field equation exactly is very difficult and has only been done numerically . fortunately , during inspiral phase of compact binary coalescences , the einstein equation can be simplified with the pn expansion .
when a compact binary is formed , the orbit decays with time due to the gr .
the orbit - averaged change of the semi - major axis and the eccentricity by gr is first derived by @xcite as @xmath51 @xmath52 in the 2.5pn order ( @xmath53 , the first order gr term ) . however , the orbital evolution is also affected by other pn order terms such as 1pn ( relativistic precession ) , 1.5pn ( spin - orbit coupling ) , 2pn ( spin - spin coupling , high order relativistic precession ) and higher orders . with full consideration of pn terms up to 2.5 order
, @xcite noted that the decay of the binary orbit is much faster than that with 2.5pn only .
although the effect of the spin is quite important for the motions and waveforms of bh - bh binary coalescences , we only consider non - spinning bhs and take 1 , 2 and 2.5pn order terms in this study .
the equation of motion with pn correction up to 2.5pn order can be simply written in the centre of mass frame as @xcite @xmath54 where @xmath55 are pn coefficients depending on their masses , the relative position @xmath56 and relative velocity @xmath57 ( see appendix a , for more details ) .
many authors have incorporated the pn corrected force in direct @xmath4-body simulations with different pn orders @xcite .
similarly , we implemented the pn equation of motion to the ks regularization process in nbody6 code . in the ks regularization process , a two - body motion is sometimes perturbed by other neighboring stars , and these perturbation should be corrected .
thus , we can consider the pn force as a perturbing force in the code by adding the pn force and its time derivative .
we designed that the binary will merge when the separation is smaller than four schwarzschild radii @xmath58 because the pn approximation is not valid in this regime any more .
for instance , fig .
[ figg_orb0 ] shows the time evolution of semi - major axis and eccentricity for a 10@xmath23 bh - bh binary with pn approximation .
the initial semi major axis and eccentricity are @xmath29 au and 0.9 , respectively .
the solid line represents the time integration of peters formula , equations ( [ eqns38 ] ) and ( [ eqns39 ] ) .
there is a good agreement between the results of simulation with only 2.5pn term ( open circle ) and integration of peters formula .
on the other hand , the merging time of simulation with all terms up to 2.5pn ( filled circle ) is significantly smaller than that of peters formula as reported in @xcite . in case of this binary
, it takes less then a half hour for merging ( i.e. , @xmath59 ) . in our simulations ,
most of pairs of close encounters have not been perturbed by other nearby stars .
thus , it is possible to separate the two - body motion with pn correction from the main loop of simulations .
we , here , use a toysverre / web / pages / nbody.htm ] code only for a ks two - body motion written by s. j. aarseth for convenience of exploration of the evolution of merging binaries .
the pn implementation mentioned earlier is also adopted in the toy code . for given semi - major axis and eccentricity ,
we simulate the orbital evolution of binaries .
( see [ subsec5.1 ] for details ) bh - bh binaries with different velocity dispersion of embedded star clusters .
the merging times are obtained from two - body simulations with all corrections up to 2.5pn .
the range of velocity dispersion is 50 to 400 km / s , which is correspond to the range of the mass of smbh from @xmath60 to @xmath61 according to the @xmath62 relation from @xcite .
star symbol represents the binary merging time in the milky - way - like galaxies . in all cases ,
the merging times are smaller than a year.,width=317 ] fig .
[ figg_tmr ] shows the merging time of typical bh - bh binaries in equations ( [ eqns19 ] ) and ( [ eqns20 ] ) with different velocity dispersion of systems . the merging time is estimated by two - body simulations with all pn corrections .
a star symbol is showing the result of a galactic nucleus in milky - way - like galaxy .
in all cases , the merging times are less than a year .
binaries formed by gr capture , therefore , will merge immediately .
the gw waveforms of coalescing binaries have been studied by many authors ( e.g. , * ? ? ?
* ; * ? ? ?
as the perturbation of flat - space metric , @xmath63 in can be expressed by ( for more details , see equations 3.21 and 3.22 in * ? ? ?
* ) @xmath64_{\rm tt}\end{aligned}\ ] ] where @xmath65 is the reduced mass , @xmath66 is the distance from the source to the detector , @xmath67 is the time derivative of quadrupole moment tensor , @xmath68 is the pn corrections with order of @xmath69 , and so , ss and tt denote spin - orbit coupling , spin - spin coupling and transverse - traceless gauge , respectively . here @xmath70 is used . since we are interested in the aspects of gw rather than the exact waveforms , we take the leading order of @xmath63 @xmath71\ ] ] with @xmath72\ ] ] where @xmath73 and @xmath74 are the relative velocity and the normal vector of the relative position , respectively .
it is well known that gws have two polarization @xmath75 and @xmath76 and waveforms are the linear combination of these two polarizations .
if we assume that the orbital plane lies on the _ xy _ plane initially in the source coordinate , and the angle between the direction to the detector @xmath77 and the angular momentum @xmath78 is @xmath79 , the polarizations @xmath80 and @xmath81 are given by @xcite @xmath82 @xmath83 now we provide a waveform of a typical 10@xmath23 bh - bh binary coalescence in a milky - way - like galaxy for an example .
the semi - major axis and eccentricity after gr capture are 0.153 au and 0.99989 from the equations ( [ eqns19 ] ) and ( [ eqns20 ] ) , respectively . in fig .
[ figg_xy ] , the relative motion of bhs on _ xy _ plane is shown .
due to the 1pn and 2pn terms , the position of perihelion is shifted counterclockwise .
in addition , by emitting gws , the orbit shrinks more and more with time .
[ figg_wav ] shows the waveforms for this bh - bh binary coalescence . for simplicity
, we assume that the axis of angular momentum is aligned with the direction to the detector ( i.e. , face - on view , @xmath84 ) . in fig .
[ figg_wav](a ) , the waveform of @xmath75 polarization during whole evolution is presented .
the merging time is about 8 days .
interestingly , the waveform is burst - like at the beginning , and it takes more than 2 days for the first burst after capture .
the detailed waveforms @xmath80 and @xmath81 at this moment are shown in [ figg_wav](c ) and ( d ) .
these waveforms are similar to those of eccentric orbits in @xcite . fig .
[ figg_wav](b ) , ( e ) and ( f ) show the waveform in the last minute , the detailed view of @xmath80 and @xmath81 a half minute before merging , respectively . in this stage , the orbit is much circularized compared to the beginning , and the orbital frequency is about 10 hz . at the moment of coalescence ,
the orbital frequency becomes few hundreds hz which is the detectable frequency by ground - base gw detectors .
one of the important questions with regards to the gr captured binaries is the typical value of eccentricity when they enter the detectable band of the ground - based detectors .
we computed the eccentricity when gw frequency becomes 30 hz as a function of initial orbital parameters as shown in fig .
[ figg_ecc ] using the post - newtonian approximation with 2.5pn term .
the diagonal line is the locus of the typical @xmath85 plane for different velocity dispersion , and we computed the eccentricities at 30 hz for up to 10% of peak in probability distribution function along the initial eccentricity for a given velocity dispersion .
different colors represent different eccentricity at 30 hz .
as can be seen from this figure , the gr captured binaries from typical nc with @xmath86 km / sec become nearly circular when they enter the ligo / virgo band . only those from very massive nc with high velocity dispersion are likely to remain eccentric up to the high frequency band .
note that we have computed the orbital evolution only up to the pericentre distance of 4 @xmath87 and some extreme binaries will not emit gw signal higher than 30 hz .
this could be an artifact of using post - newtonian approximation .
eventually the binaries in this regime should emit gravitational waves at higher frequencies than 30 hz .
we simply refer to @xcite for the derivation of more realistic derivation of the waveforms for eccentric binaries and their detectability with the ground - based gw detectors . ) plane .
the eccentricity at 30 hz is expressed by the color and the diagonal solid line represents the typical values of @xmath88 for different @xmath89 .
we limited the range of calculation down to 10% of the probability peak of the initial probability distribution function of the initial eccentricity for a given velocity dispersion marked at the central solid line .
also we show the locus of 50 % of the peak as broken lines.,width=317 ]
we have generated @xmath4-body realizations for nuclear star clusters ( ncs ) located at the centre of galactic bulges and hosting a massive black hole ( mbh ) . in our simulations , the surrounding bulge is considered as the external potential well which makes the velocity dispersion of the embedded star cluster isothermal since a deep potential well behaves like a heat bath @xcite .
the mbh , in the same manner , is also modeled as a point - mass potential but growing with time to ensure the adiabatic adjustment of the stellar system .
consequently , our @xmath4-body realizations have a stellar density cusp ( @xmath90 ; * ? ? ?
* ) and keplerian velocity dispersion within the radius of influence .
in addition , the overall velocity structure is similar to observations of the star cluster at the centre of milky way . strictly speaking , however , these star clusters are not in equilibrium but expand continuously since the mbh can generate kinetic energies by the interaction with stars in the cusp .
the slopes of density and velocity dispersion profiles of our @xmath4-body realizations are slightly shallower than those of theoretical expectations .
this environment of ncs is a good laboratory for gravitational wave ( gw ) sources . in order to investigate gw event rates in ncs ,
we have collected the orbital information of close encounters ( i.e. , semi - major axis and eccentricity ) in our @xmath4-body simulations .
while most of binaries are disrupted by the strong tidal field from mbh , there can be many hyperbolic encounters of stellar mass black holes ( bhs ) whose pericentre distances are sufficiently small to radiate gws efficiently due to the high density and velocity dispersion at the vicinity of the mbh and the high number fraction of bhs due to the mass segregation @xcite .
when the energy loss by gravitational radiation ( gr ) is greater than the orbital energy , two bhs make a binary and merge quickly because of the small separation and large eccentricity after capture .
thus , the capture event rate corresponds to the merger rate .
the capture happens most frequently near the half mass radius rather than within the radius of influence .
thus , our investigation of gr capture event rates is still valid although our models can not precisely realize the cluster inside of the radius of influence . by counting the number of gr capture events , we have built scaling relations of merger rates for a nc as a function of the mass of the mbh and the velocity dispersion of the star cluster . as the result ,
the merger rate for a milky - way - like galaxy is @xmath91 proportional to the mass ratio of mbh to the star cluster . from the @xmath62 relation ( e.g. , * ? ? ?
* ) , the merger rate becomes a function of the mass of mbh only . by using realistic mass function of mbhs ( e.g. , * ? ? ?
* ) , we have determined the merger rate density per unit volume .
then , the detection rates can be expressed with the merger rate density and the size of cosmological volume covered by gw detectors if we simply assume the uniformness of merger events over cosmic time and volume .
we have obtained the expected detection rates @xmath92 for advanced ligo depending on the maximum horizon distances from different studies @xcite and the mass ratio of mbh to the star cluster 0.1 and 0.2 .
this estimate can cover those of @xcite who suggested the detection rate of @xmath93 of 5 to 15 @xmath23 bh - bh binary coalescences in galactic nuclei for advanced ligo .
however , this study still have some limitations ; ( 1 ) it is necessary to consider realistic mass function for bhs instead of assuming the mean mass fraction of bhs . ( 2 ) there is a relation between the mass of mbhs and the ncs @xcite . however , we fixed the mass ratio of mbh to the star cluster to 0.1 and 0.2 .
( 3 ) in order to determine the maximum horizon distance for bh - bh binary mergers , the precise signal - to - noise ratio calculation is needed .
we have investigated the statistics of coalescing bh - bh binaries and found that the typical semi - major axis and eccentricity of these binaries are related to the velocity dispersion of the system .
we also have implemented the post - newtonian ( pn ) approximation on the two body motions up to 2.5pn orders . with a given set of semi - major axis and eccentricity
, we calculated the two - body motion under the pn approximation and the waveform of gw emission .
the merging time is about a few hours for a typical bh - bh binaries in a milky - way - like galaxy .
we also estimated the eccentricities of gr captured binaries when they enter the ligo / virgo band ( assumed to be 30hz ) .
most of the binaries formed in relatively low velocity environment ( @xmath94 100 km / s ) would become nearly circular while those formed in very high velocity environment are likely to keep significant eccentricity when they cross the 30 hz . considering the fact that the low velocity galaxies contribute more to the total merger rate ( cf .
equation [ eqns35 ] ) , the gr captured binaries are likely to be indistinguishable from those of different origin in the sense the eccentricity effect is nearly negligible .
this work was supported by the national research foundation of korea ( nrf ) grant nrf-2006 - 0093852 funded by the korean government .
hml acknowledges alexander von humboldt foundation for the research award in 2013 .
here , we present the post newtonian ( pn ) equation of motion of binary system in the centre of mass frame up to 2.5pn order by following @xcite ( also see appendix in * ? ? ? * ) .
because we assume non - spinning bhs , we do not consider spin terms in this paper .
for the beginning , we borrow notations from @xcite as @xmath95 where @xmath96 and @xmath97 are the mass of stars , @xmath98 and @xmath99 are the 3-dimensional positions and velocities , and @xmath100 is the symmetric mass ratio . as mentioned in the text , the pn acceleration can be considered as a perturbation and added to the gravitational acceleration as @xmath101 where @xmath102 and @xmath103 are the pn coefficients related to the relative position and velocity , respectively , where @xmath104 and @xmath106 @xmath107 @xmath108 where @xmath109 is the first time derivative of radial distance defined as @xmath110
. then the coefficients @xmath102 and @xmath103 are given by the summations of coefficients for different pn order divided by the speed of light @xmath111 , @xmath112 .
note that the sign of @xmath103 is opposite of that in @xcite .
since the nbody code uses 4th - order hermite integrator , we have to have the first time derivatives of accelerations as similar to the accelerations @xmath113 the derivative of pn acceleration @xmath114 can be expressed as @xmath115 where @xmath116 and @xmath117 are the time derivatives of the coefficients @xmath102 and @xmath103 , which are the summations of aarseth , s. j. 1999 , pasp , 111 , 1333 aarseth , s. j. 2007 , mnras , 378 , 285 aarseth , s. j. 2010 , gravitational @xmath4-body simulations , by sverre j. aarseth , cambridge , uk : cambridge university press , 2010 , abadie , j. , abbott , b. p. , abbott , r. , et al .
2010 , classical and quantum gravity , 27 , 173001 abramovici , a. , althouse , w. e. , drever , r. w. p. , et al .
1992 , science , 256 , 325 alexander , t. 2005 , phys . rep . , 419 , 65 aller , m. c. , & richstone , d. 2002 , aj , 124 , 3035 ananda , k. n. , clarkson , c. , & wands , d. 2007 , phys .
d , 75 , 123518 andersson , n. , ferrari , v. , jones , d. i. , et al . 2011 ,
general relativity and gravitation , 43 , 409 antonini , f. , & perets , h. b. 2012 , apj , 757 , 27 bae , y. , lee , h. m. , kim , c. 2014 , mnras , 440 , 2714 bahcall , j. n. , & wolf , r. a. 1976 , apj , 209 , 214 baker , j. g. , mcwilliams , s. t. , van meter , j. r. , et al .
2007 , phys .
d , 75 , 124024 balcells , m. , graham , a. w. , domnguez - palmero , l. , & peletier , r. f. 2003 , apj , 582 , l79 barth , a. j. , greene , j. e. , & ho , l. c. 2005 , apj , 619 , l151 banerjee , s. , baumgardt , h. , & kroupa , p. 2010 , mnras , 402 , 371 baumgardt , h. , makino , j. , & ebisuzaki , t. 2004a , apj , 613 , 1133 baumgardt , h. , makino , j. , & ebisuzaki , t. 2004b , apj , 613 , 1143 belczynski , k. , taam , r. e. , kalogera , v. , rasio , f. a. , & bulik , t. 2007 , apj , 662 , 504 berentzen , i. , preto , m. , berczik , p. , merritt , d. , & spurzem , r. 2009 , apj , 695 , 455 binney , j. , & tremaine , s. 2008 , galactic dynamics : second edition , princeton university press , princeton , nj blanchet , l. , & iyer , b. r. 2003 , classical and quantum gravity , 20 , 755 bker , t. , laine , s. , van der marel , r. p. , et al .
2002 , aj , 123 , 1389 bker , t. , sarzi , m. , mclaughlin , d. e. , et al .
2004 , aj , 127 , 105 brem , p. , amaro - seoane , p. , & sopuerta , c. f. 2014 , mnras , 437 , 1259 carollo , c. m. , stiavelli , m. , de zeeuw , p. t. , & mack , j. 1997 , aj , 114 , 2366 ct , p. , blakeslee , j. p. , ferrarese , l. , et al .
2004 , apjs , 153 , 223 ct , p. , piatek , s. , ferrarese , l. , et al .
2006 , apjs , 165 , 57 dehnen , w. 1993 , mnras , 265 , 250 downing , j. m. b. , benacquista , m. j. , giersz , m. , & spurzem , r. 2011 , mnras , 416 , 133 east , w. e. , mcwilliams , s. t. , levin , ja . & pretorius , f. , 2013 , phys .
d , 87 , 043004 einstein , a. 1916 , sitzungsberichte der kniglich preuischen akademie der wissenschaften ( berlin ) , 688 ferrarese , l. , & ford , h. 2005 , space sci .
, 116 , 523 fiestas , j. , porth , o. , berczik , p. , & spurzem , r. 2012 , mnras , 419 , 57 freitag , m. , amaro - seoane , p. , & kalogera , v. 2006 , apj , 649 , 91 giersz , m. , & heggie , d. c. 1996 , mnras , 279 , 1037 goodman , j. , & binney , j. 1984 , mnras , 207 , 511 graham , a. w. , & driver , s. p. 2007 ,
mnras , 380 , l15 graham , a. w. , & scott , n. 2015 , apj , 798 , 54 graham , a. w. , & spitler , l. r. 2009 , mnras , 397 , 2148 hansen , r. o. 1972 , phys .
d , 5 , 1021 holley - bockelmann , k. , mihos , j. c. , sigurdsson , s. , hernquist , l. , & norman , c. 2002 , apj , 567 , 817 hopman , c. , & alexander , t. 2006 , apj , 645 , l133 hulse , r. a. , & taylor , j. h. 1974 , apj , 191 , l59 kalogera , v. , kim , c. , lorimer , d. r. , et al .
2004 , apj , 601 , l179 kang , w .- r . , woo , j .- h . ,
schulze , a. , et al .
2013 , apj , 767 , 26 kent , s. m. 1992 , apj , 387 , 181 kidder , l. e. 1995 , phys .
d , 52 , 821 kormendy , j. , & richstone , d. 1995 , ara&a , 33 , 581 kozai , y. 1962 , aj , 67 , 591 lee , h. m. , & ostriker , j. p. 1986
, apj , 310 , 176 lee , m. h. 1993 , apj , 418 , 147 lin , d. n. c. , & tremaine , s. 1980 , apj , 242 , 789 lincoln , c. w. , & will , c. m. 1990 , phys .
d , 42 , 1123 mcconnell , n. j. , & ma , c .- p .
2013 , apj , 764 , 184 merritt , d. , alexander , t. , mikkola , s. , & will , c. m. 2011 , phys .
d , 84 , 044024 merritt , d. , mikkola , s. , & szell , a. 2007 , apj , 671 , 53 mora , t. , & will , c. m. 2004 , phys .
d , 69 , 104021 mueller , e. , & janka , h .-
1997 , a&a , 317 , 140 nitadori , k. , & aarseth , s. j. 2012 , mnras , 424 , 545 oleary , r. m. , kocsis , b. , & loeb , a. 2009 , mnras , 395 , 2127 oleary , r. m. , rasio , f. a. , fregeau , j. m. , ivanova , n. , & oshaughnessy , r. 2006 , apj , 637 , 937 oshaughnessy , r. , & kim , c. , 2010 , apj , 715 , 230 peters , p. c. 1964 , physical review , 136 , 1224 peters , p. c. , & mathews , j. 1963 , physical review , 131 , 435 quinlan , g. d. , & shapiro , s. l. 1987 , apj , 321 , 199 quinlan , g. d. , & shapiro , s. l. 1989 , apj , 343 , 725 quinlan , g. d. , hernquist , l. , & sigurdsson , s. 1995 , apj , 440 , 554 reisswig , c. , husa , s. , rezzolla , l. , et al .
2009 , phys .
d , 80 , 124026 schdel , r. , merritt , d. , & eckart , a. 2009 , a&a , 502 , 91 shapiro , s. l. 1977 , apj , 217 , 281 sigurdsson , s. , hernquist , l. , & quinlan , g. d. 1995 , apj , 446 , 75 spitzer , l. 1987 , princeton , nj , princeton university press , 1987 , 191 p. , takahashi , k. , lee , h. m. , & inagaki , s. 1997 , mnras , 292 , 331 tremaine , s. , gebhardt , k. , bender , r. , et al . 2002 , apj , 574 , 740 tsang , d. , 2013 , apj , 777 , 203 walcher , c. j. , van der marel , r. p. , mclaughlin , d. , et al .
2005 , apj , 618 , 237 weisberg , j. m. , & taylor , j. h. 2005 , binary radio pulsars , 328 , 25 yakunin , k. n. , marronetti , p. , mezzacappa , a. , et al .
2010 , classical and quantum gravity , 27 , 194005 yoon , i. , lee , h. m. , & hong , j. 2011 , mnras , 414 , 2728 young , p. 1980 , apj , 242 , 1232 | stellar black hole ( bh ) binaries are one of the most promising gravitational wave ( gw ) sources for gw detection by the ground - based detectors .
nuclear star clusters ( ncs ) located at the centre of galaxies are known to harbour massive black holes ( mbhs ) and to be bounded by a gravitational potential by other galactic components such as the galactic bulge .
such an environment of ncs provides a favourable conditions for the bh - bh binary formation by the gravitational radiation capture due to the high bh number density and velocity dispersion .
we carried out detailed numerical study of the formation of bh binaries in the ncs using a series of n - body simulations for equal - mass cases .
there is no mass segregation introduced .
we have derived scaling relations of the binary formation rate with the velocity dispersion of the stellar system beyond the radius of influence and made estimates of the rate of formation of black hole binaries per unit comoving volume and thus expected detection rate by integrating the binary formation rate over galaxy population within the detection distance of the advanced detectors .
we find that the overall formation rates for bh - bh binaries per nc is @xmath0 for the milky - way - like galaxies and weakly dependent on the mass of mbh as @xmath1 @xmath2 .
we estimate the detection rate of 0.02 - 14 yr@xmath3 for advanced ligo / virgo considering several factors such as the dynamical evolution of ncs , the variance of the number density of stars and the mass range of mbh giving uncertainties .
[ firstpage ] gravitational waves methods : numerical galaxies : nuclei . |
the high - acceptance dielectron spectrometer ( hades ) experiment at gsi pursues a comprehensive program of dielectron emission studies in few - gevnucleon - nucleon @xcite , proton - nucleus @xcite , and nucleus - nucleus collisions @xcite .
dilepton spectroscopy allows to investigate the properties of hadrons produced , propagated , and decayed in a strongly interacting medium .
this is because leptons ( electrons and muons ) do not themselves interact strongly when traveling through finite - sized hadronic matter , that is , their kinematics remain basically undistorted .
lepton - pair measurements are hence ideally suited to search for medium modifications of hadrons in nuclear matter @xcite .
the observed dilepton spectra consist , however , of a complex superposition of various mesonic and baryonic contributions , and their interpretation requires a detailed knowledge of all sources .
indeed , early interpretations of dilepton spectra from relativistic heavy - ion collisions commonly introduced a schematic distinction of ( i ) hard initial contributions related to drell - yan type processes , ( ii ) the thermal radiation off the fireball , and ( iii ) the hadronic cocktail from late decays following its disassembly ( cf .
transport models supersede this artificial separation as they describe all phases of the collision on an equal footing by following continuously virtual and real photon emission over time .
presently , they are commonly employed in the few - gev bombarding energy regime to describe particle production and propagation through the medium , in particular , when dealing with the complex dynamics of nucleus - nucleus reactions @xcite .
comprehensive information on meson production is thereby mandatory to benchmark and constrain those calculations . in this context
the neutral pion and eta mesons are of particular interest as they contribute largely to the dilepton spectrum via their dalitz decays , @xmath5@xmath6and @xmath7@xmath6 , respectively .
although in the few - gev energy regime a large body of systematic data on pion and , to a lesser extent , on eta production in nucleus - nucleus collisions has been gathered over the last decades , mostly at the bevalac , the ags , and sis18 accelerators , there is much less information available from proton - nucleus reactions .
the latter ones are , however , important as an intermediate step between nucleon - nucleon and nucleus - nucleus collisions .
charged pions from p+a have been measured at the bevalac with proton beams of kinetic energies up to 2.1 gev @xcite , at triumf up to 0.5 gev @xcite , and recently by the harp experiment at the cern ps with proton energies between 2 and 12 gev @xcite .
information on eta production in p+a reactions is even more scarce .
only the pinot experiment at the saturne accelerator in saclay provided data for proton energies in the range of 0.8 1.5 gev @xcite .
in this paper we supplement the available body of experimental results on pion ( @xmath8 ) and eta production in p+a collisions with data obtained with hades in the p+nb reaction at 3.5 gev .
negative pions have been identified via their characteristic energy loss vs. momentum signature in the hades time - of - flight ( tof ) system .
the neutral mesons , @xmath3 and @xmath4 , were reconstructed with the photon - conversion technique in which meson decay photons are detected via their external conversion into an @xmath6 pair via a bethe - heitler process , preferentially in high-@xmath9 materials .
this method has been developed foremost in high - energy physics experiments @xcite to study the radiative decays of the quarkonium states @xmath10 into @xmath11 and @xmath12 into @xmath13 .
it has also been used in high - energy heavy - ion reactions , namely by the experiment at rhic , studying au+au collisions at a nucleon - nucleon center - of - mass energy of @xmath14 gev @xcite and by the experiment at the lhc in @xmath15 tev p+p collisions @xcite . making use of the good momentum resolution of charged - particle trackers , in particular at low energies , the conversion technique offers typically better energy resolution than a photon calorimeter .
we demonstrate here the applicability of the method with hades in few - gev reactions . our paper is organized as follows .
section [ experiment ] describes the experiment and the employed particle identification procedures . in sec .
[ method ] the photon - conversion method is introduced .
pion and eta spectra , as well as meson multiplicities are presented in sec .
[ yields ] . in sec .
[ transport ] we compare the data with transport - model calculations and , finally , in sec .
[ summary ] we summarize our findings .
a preliminary version of the @xmath16 data shown here has already been presented elsewhere @xcite .
the six - sector high - acceptance spectrometer hades operates at the gsi helmholtzzentrum fr schwerionenforschung in darmstadt where it takes beams from the heavy - ion synchrotron sis18 .
although its setup was originally optimized for dielectron spectroscopy , hades is in fact a versatile charged - particle detector with both good efficiency and momentum resolution .
its main component serving for electron and positron selection is a hadron - blind ring - imaging cherenkov detector ( rich ) .
further particle identification power is provided by the time of flight measured in a plastic scintillator tof wall , the electromagnetic shower characteristics observed in a pre - shower detector , and the energy - loss signals from the scintillators of the tof wall as well as from the four planes of drift chambers serving as tracking stations .
charged particles are tracked through a toroidal magnetic field provided by a six - coil iron - less superconducting magnet .
all technical aspects of the detector are described in @xcite . in the experiment discussed here a proton beam with a kinetic energy of @xmath17 = 3.5 gev and an average intensity of about @xmath18 particles per second impinged onto a 12-fold segmented niobium ( @xmath0nb ) target with a total thickness of 5.4 mm and corresponding to 2.8% nuclear interaction probability .
the online event selection was done in two steps : a 1@xmath19-level trigger ( lvl1 ) selected events with at least three charged - particle hits in the tof wall ( @xmath20 ) and a 2@xmath21-level trigger ( lvl2 ) fired if an electron or positron candidate was recognized .
while all lvl2 events were recorded , the lvl1-only events were downscaled by a factor three before being written to data storage .
this trigger scheme was in fact primarily optimized for studying inclusive @xmath6 production @xcite . to allow for trigger bias studies also
lvl1 events requiring only two charged particles were recorded during part of the experiment .
the lvl1 triggered on p+nb reactions with 56% efficiency in the @xmath20 mode and 72% efficiency in the @xmath22 mode .
these efficiencies correspond to the fraction of reactions that actually fired the lvl1 trigger . in total @xmath23 events
downscaled lvl1 or lvl2 were recorded , corresponding to @xmath24 inspected lvl1 events and @xmath25 reactions in the target . to study neutral - meson production the off - line data analysis searched for events containing four lepton tracks from which the 4-momentum of @xmath3 and @xmath4 mesons was fully reconstructed .
indeed , the electromagnetic decays of the latter , that is mostly @xmath26e@xmath1e@xmath2 ( dalitz ) and @xmath27 , combined with the external conversion of the decay photon(s ) on material in the inner region of the hades detector lead to events with a characteristic 4-lepton signature , namely @xmath28 e@xmath1e@xmath2e@xmath1e@xmath2 .
simulations show that the dalitz decay contributes 25% ( 30% ) of the detectable @xmath3 ( @xmath4 ) yield .
the direct branching into e@xmath1e@xmath2e@xmath1e@xmath2 due to the decay into two virtual photons ( @xmath29 ) is however very small @xcite and contributes in our case only about 2 2.5% of the total yield .
electron and positron tracks were identified following the procedures described in detail in @xcite .
as no dedicated start detector was present in this experimental run , the start time for the time - of - flight measurement was reconstructed event - by - event from the most optimal assignment of different particle hypotheses ( e@xmath2 or @xmath16 ) to tracks of negatively charged fast particles . in the extraction of the inclusive negative pion yields
no direct use was however made of the time - of - flight information , and the pion identification was solely based on energy - loss vs. momentum cuts ( for details see @xcite ) .
positive pions were identified likewise but , their spectra being partially contaminated by the much more abundant protons , they were not analyzed further .
the detection of high - energy photons via a full reconstruction of conversion pairs has been applied with success by the phenix @xcite and alice @xcite collaborations in ultrarelativistic collisions .
although those experiments comprise also electromagnetic calorimeters for photon detection , this alternative method is considered to offer valuable supplementary information on thermal photon as well as neutral - meson production . as hades
is presently not equipped with such a calorimeter , the conversion method opens a unique approach to @xmath3 and @xmath4 detection .
meson reconstruction was realized with identified lepton tracks by joining opposite - sign leptons into @xmath6 pairs and by further combining those dileptons pairwise into e@xmath1e@xmath2e@xmath1e@xmath2 multiplets .
calculating the four - lepton invariant mass @xmath30 and setting appropriate mass cuts allows to select the @xmath3 and @xmath4 mesons , respectively .
various opening angle cuts , optimized in extensive monte - carlo simulations , were applied to suppress combinations of uncorrelated leptons , namely @xmath31 on the dilepton with the smaller opening angle and @xmath32 on the second one ( all angles given in the laboratory system ) .
the @xmath33 selection is optimal for conversion pairs while the @xmath34 cut accepts also the more massive and hence wider dalitz pairs .
in addition , a cut was applied on the relative angle between the two dileptons in a multiplet , namely @xmath35 , to suppress spurious counts at low @xmath30 .
these cuts were adjusted on the data in order to maximize the meson yield while keeping the background of uncorrelated combinations low .
[ meeee ] .characteristics of the reconstructed meson peaks : raw signal counts above cb ( integrated in the range 0.10 0.16 gev/@xmath36 for @xmath3 and 0.46 0.60 gev/@xmath36 for @xmath4 ) , signal / cb ratio in those mass ranges , position of the peak maximum , and @xmath37 width of the peak ( @xmath38fwhm/2.35 ) .
all errors are statistical . [
cols="<,^,^",options="header " , ] [ efftable ] we have performed extensive detector simulations to study the reconstruction efficiency of the conversion technique . to do this
we have generated meson distributions with the pluto event generator @xcite , tracked the resulting particles through a realistic model of the hades setup with the geant3 physics simulation tool @xcite , embedded those tracks into real events from the p+nb experiment , and reconstructed the overlayed events with the full hades lepton analysis .
the purpose of the embedding was to include realistic detector noise into the procedure .
as the lepton identification in hades relies primarily on the rich detector , only conversion pairs produced in the inner parts of the setup can contribute to the e@xmath1e@xmath2e@xmath1e@xmath2 signal .
these are the niobium target segments , the carbon - composite target holder , the carbon - composite beam pipe , and the rich radiator gas c@xmath39f@xmath40 .
conversion pairs produced in the rich mirror or in any of the following materials are not detectable .
average conversion probabilities @xmath41 of @xmath3 and @xmath4 decay photons obtained from the simulations are listed in table [ convtable ] .
more than half of the conversion takes place in the niobium targets and the rest in the target holder , the beam pipe , and the rich converter gas , with a the cumulated probability of 3.54.1% per photon . the difference in probabilities for @xmath3 and @xmath4 ,
respectively , is related to the energy dependence of the conversion process .
as the hades pair vertex resolution , of the order of 23 cm , is not good enough to cleanly isolate the various converter parts in the event reconstruction we have refrained from applying specific vertex cuts and have exploited the cumulated conversion effect .
systematic errors on the total conversion probability result mainly from uncertainties on the material budget and the alignment of the relevant inner detector parts .
the thickness of the target foils is known with an error of 2% and their misalignment in the beam pipe adds another 4% error .
inhomogeneities of the carbon - composite material lead to an estimated uncertainty of 10% on the holder and pipe contributions . finally , for the radiator contribution we assume a 5% error .
this leads then to an overall systematic error on the total material budget of 5% and hence an error of about 10% on the effciency for detecting double conversion events . as we do not apply a secondary vertex selection to the @xmath6 pairs
also dalitz decays are included in our pair signal . in the dalitz case , of course , only the one real photon is required to convert .
the combined branching into all contributing final states ( @xmath42 , @xmath43 , and @xmath44 ) amounts to @xmath45% for the @xmath3 and to @xmath46% for the @xmath4 @xcite .
the meson sources used in the simulation were modeled according to a fireball @xcite characterized by a temperature in the range @xmath47= 8090 mev and a central laboratory rapidity in the range @xmath48= 0.920.96 .
as discussed below , the choice of these values is motivated by the data itself ( cf . also figs .
[ mult_pt ] and [ mult_y ] ) .
the monte carlo shows that the total detection efficiency _
eff_@xmath49 depends only weakly on the meson source properties and mostly on the photon conversion probability , on the geometric detector acceptance with respect to @xmath50 _ acc _ , and on the pair reconstruction efficiency _ eff_. in this definition , _ acc _ includes the lepton low - momentum cutoff at around 50 mevdue to the track bending in the hades magnet field and _ eff _ accounts for all detection and reconstruction losses within the hades acceptance . as all of these quantities are averaged over the two - photon , the dalitz , and the small direct decay channels it is useful to introduce an effective branching ratio _
br_@xmath51 into the e@xmath1e@xmath2e@xmath1e@xmath2 final state which includes the photon conversion probability .
table [ efftable ] summarizes the result of our @xmath52 e@xmath1e@xmath2e@xmath1e@xmath2 and @xmath53 e@xmath1e@xmath2e@xmath1e@xmath2 event simulations showing in particular that the total detection efficiencies are of order @xmath54 @xmath55 .
systematic errors on the simulated efficiencies result from the uncertainties on the conversion probabilities ( 5% on @xmath56 , 7.5% on @xmath57 ) , on the branching ratios of the contributing decays ( 1% on @xmath57 for @xmath3 and 2% for @xmath4 , from @xcite ) , and on the detector and reconstruction efficiencies ( 10% , from a comparison of various simulated and measured observables in the hades detector @xcite ) .
combining all of these contributions we assign to the total efficiency a conservative systematic error of 15% .
the transverse momentum ( @xmath58 ) dependence of the meson reconstruction efficiency is depicted in fig .
[ effpt ] , with and without the photon conversion probability included .
the cutoff of the @xmath3 efficiency at @xmath59 gev/@xmath60 is caused mostly by the strong bending of low - momentum tracks , i.e. those with @xmath61
gev/@xmath60 , in the hades magnetic field . because of the large mass of the eta meson , its efficiency is much less afflicted by low - momentum tracks and
consequently reconstruction is possible down to zero @xmath58 .
[ hades_pim ] as pointed out above , besides measuring e@xmath1 and e@xmath2 , the hades detector also provides high - quality data on charged hadrons . here
we use the concurrently measured negative pions @xcite to validate the reconstructed @xmath3 yields and extrapolate them to @xmath62 gev/@xmath60 .
figure [ hades_pim ] shows efficiency - corrected yields are corrected as well for the gaps caused in the azimuthal acceptance by the hades magnet coils . ]
double differential @xmath16 yields @xmath63 adapted from @xcite .
the charged pion acceptance is constrained geometrically to polar angles of @xmath64 @xmath65 as well as by a momentum cut of @xmath66 gev/@xmath60 .
this results in an acceptance in laboratory rapidity of @xmath67 0.2 1.8 and in transverse momentum of @xmath68 gev/@xmath60 .
because of this quite large coverage one can expect that the extrapolation to full phase space will lead to moderate systematic uncertainties only . at a bombarding energy of 3.5 gevthe rapidity of the nucleon - nucleon center - of - mass system
is at @xmath69 . from the fig .
[ hades_pim ] it is however apparent that in the p+nb reaction the pion yield is not peaked at a mid - rapidity @xmath70 but at a lower value with , in addition , a marked @xmath58 dependence .
fits of a gaussian function to @xmath71 projections done for various @xmath58 slices give the loci @xmath48 of the maximum yield vs. @xmath58 , as shown by stars in fig .
[ hades_pim ] .
in particular , low-@xmath58 pions seem to be radiated mostly from a target - like source , near @xmath72 , pointing to a high degree of stopping of the incoming projectile .
obviously not only first - chance nucleon - nucleon collisions contribute to pion emission in the p+nb reaction , but proton elastic and inelastic rescattering followed by secondary production processes add a soft target - like component .
this is also corroborated by various transport - model calculations @xcite .
integrating @xmath63 within the hades rapidity coverage we find an accepted @xmath16 yield of 0.50 per lvl1 event . from simulations
we know that the lvl1 trigger leads to a 42% enhancement of the average detected charged - pion yield per event ( see @xcite ) .
correcting for this trigger bias we obtain an accepted yield of 0.35 per p+nb reaction .
the gauss fits done to the @xmath71 projections provide furthermore a means to extrapolate the measured @xmath16 yield outside of the hades rapidity coverage .
alternatively , transport models , e.g. hsd @xcite , urqmd @xcite , or gibuu @xcite ( see sec . [ transport ] below ) , can be used to perform the extrapolation in @xmath73 and @xmath58 to full solid angle .
in fact , integrating the yield within the geometric acceptance limits , extrapolating it either way via gauss fits to @xmath71 projections in @xmath58 slices or , better , with the help of transport calculations done for p+nb and correcting for the lvl1 bias we obtain on average a minimum - bias inclusive @xmath16 multiplicity of @xmath74 . the error on this corrected @xmath16 multiplicity is dominated by systematic effects introduced mostly by the correction of the lvl1 trigger bias ( @xmath75% ) and the spread in the model - dependent extrapolation in phase space ( -10% , + 15% ) ; statistical errors are however negligible .
table [ multtable ] lists the extracted multiplicity values and their associated uncertainties .
l c c particle & @xmath76 & @xmath77 + ' '' '' @xmath16 & @xmath78 & @xmath79 + ' '' '' @xmath3 & @xmath80 & @xmath81 + ' '' '' @xmath4 & @xmath82 & @xmath83 + [ multtable ] we come now to the presentation of our differential @xmath3 and @xmath4 yields .
first notice that the lvl1 trigger ( meaning at least three charged hits in the tof wall ) does not introduce a bias on events with an e@xmath1e@xmath2e@xmath1e@xmath2 signature .
likewise , the lvl2 trigger efficiency is found to be for such events .
hence no explicit corrections for trigger effects are needed .
the systematic uncertainties on the e@xmath1e@xmath2e@xmath1e@xmath2 observables to be taken into account are those introduced by the efficiency correction ( @xmath84% ) , the cb subtraction ( @xmath3 : @xmath85% , @xmath4 : @xmath86% ) , and the model - dependent extrapolation to full solid angle ( @xmath3 : @xmath84% , @xmath4 : @xmath86% ) .
[ mult_pt ] [ mult_y ] as is visible from the invariant - mass spectrum in fig .
[ meeee ] , both neutral mesons can be selected with appropriate mass cuts .
we have used @xmath87 for the @xmath3 and @xmath88 for the @xmath4 .
subtracting the corresponding combinatorial background and applying corrections for photon conversion , as well as for lepton track identification and reconstruction efficiencies the double - differential yields @xmath89 are obtained as a function of rapidity and transverse momentum . as stated in sec .
[ method ] , our efficiency corrections are based on the reconstruction of simulated meson decays embedded into real events . to do a first correction , we started out with relativistic boltzmann distributions of an assumed temperature of 100 mevand then refined this value in a second pass .
the same holds for the central rapidities of the @xmath3 and @xmath4 sources which , like in case of the @xmath16 , are observed to be substantially below the @xmath90 value .
all corrections were furthermore done concurrently in two dimensions , @xmath73 and @xmath58 , in order to alleviate any remaining dependence on our assumptions about the meson source characteristics .
the resulting final @xmath91 and @xmath71 distributions , normalized per minimum - bias event , are displayed in figs .
[ mult_pt ] and [ mult_y ] . for comparison ,
the negative pion distributions are also shown .
[ pion_mt ] due to the efficiency cutoff discussed in sec
. [ method ] , no @xmath3 yield is detected at low @xmath58 .
this is directly visible in fig .
[ mult_pt ] and it is also the reason why in fig .
[ mult_y ] we show both pion rapidity distributions ( i.e. @xmath3 and @xmath16 ) for @xmath58 larger than 0.35 gev/@xmath60 only . for the @xmath4 meson , however , the @xmath58 coverage is complete .
the rapidity coverage of both neutral mesons is restricted to @xmath92 = 0.2 1.8 by the detector geometry . in a first attempt to characterize the observed meson yields we confront them with the isotropic fireball model @xcite . adjusting a boltzmann type distribution @xmath93 to the meson transverse momentum distribution and a gaussian @xmath94 to the rapidity distribution we find @xmath95 mev ( @xmath96 ) , @xmath97 , and @xmath98 ( @xmath99 ) for the @xmath3 , respectively @xmath100 mev ( @xmath101 ) , @xmath102 , and @xmath103 ( @xmath104 ) for the @xmath4 . here
@xmath105 is the modified bessel function , @xmath106 is the transverse mass , @xmath47 is the fitted temperature parameter , @xmath48 is the average source mid - rapidity ( actually the peak position thereof ) , and @xmath107 is the width of the accepted rapidity distribution .
the fit results show in particular that , within error bars , the rapidity distributions of both pion species agree in shape . to characterize the pion source further , fig .
[ pion_mt ] shows the pion transverse - mass distributions @xmath108 projected for various rapidity selections .
it is apparent from this figure , and also fig .
[ mult_pt ] , that a boltzmann source does not describe very well the low-@xmath58 behavior .
on the other hand , both pion species display in general a very similar behavior as function of @xmath109 and @xmath73 . only in the first rapidity bin
( @xmath92=0.2 0.4 ) the @xmath16 show a somewhat harder spectrum than the @xmath3 , which we attribute to a contamination of the @xmath16 spectrum at small polar angles with fake tracks .
we prefer , still , to use directly the shape of the measured negative pion distribution spectrum was first corrected for its coulomb shift @xmath110 mev , where @xmath111 and @xmath112 fm . ] for extrapolating the @xmath3 yield below 0.35 gev/@xmath60 .
doing this , we get a yield per minimum - bias event of @xmath113 within the accepted rapidity range of @xmath114 . in a second step
the extrapolation to full solid angle can be done , based on transport - model calculations as discussed above , giving a minimum - bias inclusive multiplicity of @xmath115 .
statistical and systematic error bars on those results are given in table [ multtable ] .
for the eta meson , being four times heavier than the pion , the hades detector provides complete transverse - momentum coverage .
the rapidity coverage , although restricted to @xmath92 = 0.2 1.8 , is very large too .
figures [ mult_pt ] and [ mult_y ] show that the @xmath4 phase space distribution is well described by a boltzmann fit in transverse momentum and by a gaussian in rapidity .
the latter fit yields a width @xmath107 , which can be related @xcite to the longitudinal temperature parameter @xmath116 of the eta source via the relation @xmath117 . from @xmath118 and @xmath119 one
obtains @xmath120 mevwhich is within error bars still consistent with the transverse temperature parameter obtained from the above boltzmann fit , namely @xmath121 mev . finally , fig .
[ eta_mt ] shows boltzmann fits to the @xmath4 transverse mass distributions @xmath108 for various rapidity selections , as well as the evolution with rapidity of the fitted slope parameter , @xmath122 .
all of those are compatible with the assumption of an isotropic fireball : the @xmath109 distributions are thermal with their slope varying like @xmath123 where @xmath124 is taken from the previous boltzmann fit to @xmath91 and @xmath48 is the central rapidity obtained in the above gauss fit to @xmath71 . [ eta_mt ] integrating either of the @xmath71 or @xmath91 distributions we obtain a yield of 0.031 accepted eta per reaction and , extrapolating with the help of our fireball fits to full solid angle , an inclusive multiplicity of @xmath125 .
extrapolations based on transport models yield slightly larger values ( see sec . [ transport ] ) .
this is taken into account in the systematic uncertainties listed in table [ multtable ] .
having available differential yields of both pions and etas from the same reaction , we can compare their scaling with transverse mass .
so - called @xmath109-scaling has indeed been found previously for @xmath3 and @xmath4 production in 1 and 1.5 gev/@xmath126 ar+ca collisions @xcite .
the observation was that the production cross sections at mid - rapidity of different mesons are identical at a given @xmath109 value .
model calculations have been able to reproduce this phenomenon @xcite . according to @xcite ,
in particular , proton - nucleus collisions should display @xmath109-scaling as well . in fig .
[ mtscaling](a ) we show , therefore , an overlay of our pion and eta @xmath109 distributions for a few broad rapidity bins .
we have chosen here the @xmath127 representation , ( i ) because this form represents a boltzmann source as approximate exponential function , and ( ii ) because it is particularly well suited for visualizing @xmath109-scaling . despite slight differences in slope
consistent with the temperature parameters obtained from the boltzmann fits discussed above an overall good agreement of the @xmath3 and @xmath4 yields at @xmath128 is apparent .
the @xmath16 yields follow the @xmath109-scaling as well , except at low rapidities . as we discussed already in the context of fig .
[ pion_mt ] , we attribute this deviation to a contamination of fake tracks in the @xmath16 spectrum .
the same trends are visible in panel ( b ) of fig .
[ mtscaling ] which shows ratios of the meson yields , namely @xmath129 and @xmath130 . except for part of the low - rapidity bin ,
the ratios are compatible with unity at transverse masses above @xmath131 .
we conclude that , albeit the pion spectra do not follow boltzmann distributions at low transverse momentum , @xmath109-scaling seems to hold .
this finding suggests that the meson yields are determined mostly by phase space : although we can assume that meson production is mediated mostly by baryon resonance excitation , it does not matter whether one produces an eta meson at low momentum or a pion at high momentum , as long as their transverse mass is the same .
[ mtscaling ] finally , we want to point out that the extrapolated meson multiplicities can be transformed into a production cross section by multiplication with the total reaction cross section , @xmath132 .
parameterizations of the proton - nucleus absorption cross section as function of bombarding energy do exist @xcite and they suggest for p(3.5 gev)+nb values of @xmath132 ranging from 990 mb ( ref .
@xcite ) to 1060 mb ( ref .
the comparison of @xmath16 multiplicities measured with hades and interpolated harp @xmath16 cross sections @xcite yields a compatible value of @xmath133 mb @xcite .
[ transport_pty ] our observation that the rapidity distributions measured in the asymmetric p+nb system are centered at values @xmath134 strongly suggests that , beyond first - chance nucleon - nucleon collisions , secondary reactions , i.e. processes involving multiple successive interactions of baryons and/or mesons contribute sizably to particle production .
a similar behavior had already been noticed for kaon production in a previous study of p+au collisions at comparable bombarding energies @xcite . while the observed phase space population of the eta agrees quite well with a fireball description , this is questionable for the pion .
complete thermalization is apparently not reached in the p+nb reaction and a transport - theoretical approach is required to model the complex interplay between reaction dynamics and particle production .
transport models typically handle meson and baryon production at energies up to a few gev by resonance excitation and at higher energies through string fragmentation . in that respect ,
our beam energy is particularly challenging because it is situated in the transition region between these two regimes . in the following we compare our results with three transport calculations done with either urqmd ,
the ultrarelativistic quantum molecular dynamics model ( version v3.3p1 , see @xcite ) , gibuu , the giessen boltzmann - uehling - uhlenbeck model ( version 1.5 , see @xcite ) , or hsd , the hadron string dynamics model ( version 2.7 , see @xcite ) . at 3.5 gev bombarding energy , corresponding to @xmath135 = 3.18 gev , urqmd runs in the resonance regime only whereas hsd switches over to string fragmentation mode at @xmath135 = 2.6 gev . for gibuu , on the other hand
, we present calculations done with two different realizations of this model : ( i ) the original implementation with a smooth transition to string fragmentation at @xmath135 = 2.6 gev and ( ii ) a version ( denoted hereafter by `` ext . res . '' ) where the resonance region has been extended up to about @xmath135 = 3.5 gev @xcite .
l c c c c ' '' '' & & + ' '' '' model & @xmath136 & @xmath50 & @xmath136 & @xmath50 + urqmd v3.3p1 & 0.38 & 0.66 & 0.013 & 0.016 + hsd v2.7 & 0.38 & 0.69 & 0.028 & 0.038 + gibuu v1.5 & 0.39 & 0.64 & 0.039 & 0.046 + gibuu ext res & 0.32 & 0.49 & 0.031 & 0.034 + [ modeltable ] figure [ transport_pty ] , panels ( a ) and ( b ) , shows that @xmath3 production is fairly well described by all models , and this in both observables @xmath58 and @xmath73 , within the hades rapidity and transverse momentum acceptance .
the pion yields reconstructed in the rapidity range @xmath92 = 0.2 1.8 are in fact reproduced to within 10 - 25% ( cf .
table [ multtable ] vs. table [ modeltable ] ) .
the rapidity distributions of all models are very similar with a slight tendency of being shifted towards target rapidity .
differences between the various calculations are mostly visible for @xmath137 gev/@xmath60and for @xmath138 gev/@xmath60 .
notice also that at low @xmath58 the standard implementation of gibuu behaves more in line with urqmd than the one with an extended resonance region .
this is somewhat surprising as one would rather expect that this modification of gibuu increases the similarity of the two models .
comparing next in fig .
[ transport_pty ] , panels ( c ) and ( d ) , the calculated and measured @xmath4 distributions , larger discrepancies between the models do appear . while urqmd reproduces quite well the @xmath91 shape , it underestimates the accepted yield by a factor of 23 and also misses the @xmath71 shape
. both versions of gibuu , on the other hand , do fairly well in describing the @xmath4 rapidity distribution and integrated yield ( see table [ modeltable ] ) , and its extended - resonance implementation also possesses the correct transverse - momentum behavior .
the hsd pion and eta yields , finally , do agree fairly well with the data , albeit their @xmath73 and @xmath58 distributions deviate substantially .
the complete lack of data in the energy range discussed here probably explains why the models tend to perform worse for @xmath4 production than they do for pions
. we are confident that with the help of our new results a more detailed theoretical investigation of the relevant production processes will now be possible .
to summarize , we have presented data on inclusive pion and eta production in the reaction p+nb at 3.5 gev kinetic beam energy . in this study
we have used the photon - conversion method to detect and reconstruct neutral mesons from 4-lepton final states .
we have demonstrated that with hades quantitative results on differential @xmath3 and @xmath4 yields can be obtained over a large range of transverse momentum and rapidity .
our data provide valuable new input for the theoretical description of proton - nucleus and nucleus - nucleus collisions in the few - gev energy regime with respect to both meson dynamics and dilepton emission .
this is exemplified in our comparison with a selection of available transport models revealing an overall fair to good agreement in various observables . together with our previous studies of p+p reactions @xcite ,
the present results provide the required baseline for measurements with heavy - ion beams at the future fair facility . indeed , as its central component
the sis100 accelerator is designed to provide intense beams of even the heaviest ions up to 8 gev/@xmath126 , we will be in the position to isolate unambiguously those effects induced by the hot and dense baryonic medium . | data on inclusive pion and eta production measured with the dielectron spectrometer hades in the reaction p+@xmath0nb at a kinetic beam energy of 3.5 gev are presented .
our results , obtained with the photon - conversion method , supplement the rather sparse information on neutral - meson production in proton - nucleus reactions existing for this bombarding energy regime .
the reconstructed e@xmath1e@xmath2e@xmath1e@xmath2 transverse - momentum and rapidity distributions are confronted with transport - model calculations , which account fairly well for both @xmath3 and @xmath4 production . |
the research about the cosmological scale factor has attracted intense attention during the last years as well as the investigation about the connection between the cosmological constant and the dark energy issue . as the vacuum has a non - trivial role in the early universe ,
a @xmath0-term in the einstein field equations is generated .
this @xmath0-term leads to the inflationary phase @xcite .
based on the inflationary cosmology we can say that during an early exponential phase , the vacuum energy was a large cosmological constant .
however , the current observed small value of the cosmological constant makes us to assume that @xmath0 , representing the energy density of the vacuum , is a variable dynamic degree of freedom which being initially very large went down to its small present value in an expanding universe @xcite .
the last one can be measured through the discrepancies between the infinitesimal value that the cosmological constant has for the present universe ( it is very small in planck units ) and the values expected by the standard model @xcite .
there is an extensive literature that show examples of phenomenological @xmath0-decay laws . to mention some ideas ,
the studies comprise a @xmath0 depending on temperature , time , hubble parameter and scale factor @xcite .
a dynamically decaying cosmological constant with cosmic expansion has been considered by several authors @xcite .
the importance of the subject resides in the fact that any nonzero value of @xmath0 introduces a length scale and a time scale into the theory of general relativity @xcite .
the cosmological constant perturbs spacetime dynamics , although the general relativity works on scales much larger than the planck scales . however , vacuum quantum fluctuations of the standard model rekindle the theory of the cosmological constant problem .
moreover , the cosmological constant is one of the candidates to be a de model , namely , with the equation of state @xmath1 .
a plaudible de model is very important to explain the observations introduced by supernovae ( sne ) type ia @xcite .
other candidates are scalar fields models such as quintessence , k - essence , tachion , phantom and etc .. for a review see @xcite .
albeit there are many references about a scale factor inverse - square law for the cosmological constant there is a lack of information about the explicit form of the function of the scale factor that obeys this requirement . in reference
@xcite there are some analytical solutions . however , those were obtained for a perfect - fluid matter energy - momentum tensor . besides , it is our opinion that there is not an appropriate analysis in the literature that includes the scale factor analytical solutions in a de approach .
the underlying motivation to carry out these solutions is the fact that the time variation of @xmath0 can lead to a creation of matter ( i.e. , @xmath0(t)cdm models ) and/or radiation such as to help us in the investigation of the age of the universe . in this work
we will use the requirement of the energy - momentum tensor conservation in order to introduce some new analytical solutions for the scale factor different from the published and unpublished @xcite literature .
while the current observations clearly pinpoint to a flat universe , in the name of generalization , we will find solutions that have different curvatures , @xmath2 as parameters , i.e. , for flat , open and closed universes respectively .
the obvious interest in a flat @xmath3 cosmological constant model has its motivation rooted in the fact that a @xmath0-term helps to connect inflation with observations .
moreover , with @xmath0 , it is possible to obtain , for flat universes , a theoretical age in the observed range , even for a high value of the hubble parameter @xcite , but it is not the objective of this work . with these parameters
we analyze some de models in the background of these analytical calculations . in this paper
we also assume , as will be said below , that the parameter of the de equation of state , @xmath4 , is constant .
this parame4ter measures the ratio between the dark energy pressure and its energyh density and is its precse value is one of the main tarets in general relativistic cosmology . in the next section
we will depict the field equations and the differential form of the scale factor equation . in section iii
we will present some motivations about the reasons to study this specific form for the cosmological constant .
we present some possible analytical solutions for the scale equation in section iv . in section
v we carry out a very brief review about de and we analyze its connection with the obtained results .
the final discussions are left for the final section .
the observations showed that the universe was not only in expansion but that this expansion is not uniform , as believed by the scientific community , but it is accelerated , one evidence of this acceleration is through the hubble - sandage diagram that shows the observed brightness of the sne type ia as a function of the redshift .
it suggest that the bulk of the energy density in the universe is repulsive and appears like an exotic form of energy with negative pressure .
this exotic energy was dubbed as dak energy and it is considered as an additional component to the ordinary dark energy .
there are many good reviews in the literature @xcite . the analysis of dark energy phenomena begun with the observations of supernovae ( sne ) type ia published independently by two different groups @xcite in the late 90 s . as we said
before the equation of state paramete @xmath4 , is one of the most wanted number in relativistic cosmology today .
one motivation for this quest is that if one could fix @xmath4 to be precisely @xmath5 , then , arguably , we could identify the de with the vacuum state of all existing fields in the universe , namely , the cosmological constant . on the other hand ,
if the precise value obtained is @xmath6 , then we could not rule out the cosmological constant , but we can work on a model of the de pressure responsible for the current cosmic acceleration as the potential energu density associated with a dynamical scalar field .
in other words , we have the so - called quintessence field if @xmath7 @xcite or a phantom component for @xmath8 @xcite . in both cases
we have a violation of the strong energy condition ( @xmath9 ) . in the case of the phantom component ,
moreover , we have the violation of the null energy condition ( @xmath10 , for a homogeneous and isotropic universe ) .
bounds on the value of @xmath4 can be obtained from observation such as the mentioned magnetic measurements of sne ia .
there are other evidences that came with the anisotropies of the cosmic microwave background ( cmb ) observations @xcite , the large - scale structure ( lss ) data @xcite , the age of high - z objects @xcite , measurements of the angular size of radio sources @xcite , observation of the x - ray luminosity of galaxy clusters @xcite , gravitational lensing @xcite and there are others that the interested reader can find in @xcite showing that a standard cold matter scenario is not enough .
recently , in @xcite the authors showed , using the lookback time techique , how to provide bounds on the cosmological parameters with high accuracy . in spite of the number of de models , the cosmological constant and the quintessence model are the most probables candidates . for the quintessence model , the equation of state is determined dynamically by the scalar field or tachyon .
it is important to notice that the concept of an accelerated universe and consequently of the existence of a de component can be considered as an introduction of a new era in general relativity ( gr ) .
it is possibloe that we have to review all the gr concepts since repulsive gravity is possible in gr . through this section
we will present briefly some possible alternative explanations for the acceleration of the universe and connect these models with the results obtained in the last section , which is our actual objective . to accomplish this we have to analyze the main dark energy ingredient which is its equation of state given by equation @xmath11 .
we will see that the alternative models of dark energy have their origin through different values of @xmath4 .
it is worthwhile to stress that it is very plausible to consider this parameter constant because observational data can hardly distinguish between a varying and a constant equation of state @xcite .
but there are many studies that consider this parameter as a function of the redshift @xmath12 or the scale factor @xmath13 . here
we will consider this parameter as constant . from the conservation of the energy momentum tensor through the bianchi identities we have a continuity equation given by @xmath14 and substituting the equation of state above and solving for @xmath15 we have that @xmath16 where @xmath17 is the density of non - relativistic matter today .
for @xmath18 we have that @xmath19 and this scenario is characterized by a dust dominated universe , i.e. , @xmath20 .
for @xmath21 we have @xmath22 , i.e. , a radiation dominated universe .
however , for both cases we have a deceleration expansion of the universe scenario .
notice that it is the same result provided by the newton gravitational theory . as we claimed in the first sections ,
the scale factor been proportional to the universe radius , as said above , describes the growth of the universe and it is defined so that at the present time @xmath23 , @xmath24 .
it can be proved both classically and relativistically that the time evolution of the expansion obeys @xmath25 where @xmath26 measures de acceleration of the universe . the experiments described above
confirm that at the present time , @xmath27 , characterizing an accelerating universe .
this acceleration implies that , either : the universe is dominated by some particles ( dark matter ) or by a field ( dark energy ) that has a negative pressure . in particular , substituting the equation of state in ( [ 21 ] ) we have the condition @xmath28 . and
again we have two situations : or there is in fact a non - zero cosmological constant ; or the gr or the standard model , is incorrect . for @xmath29
we have an acceleration / deceleration boundary .
the de is so enigmatic that we can have two different scenarios : if de is a cosmological constant the universe will be acceleration indefinitely and will become empty ; or if it is a dynamical scalar field , this last may decay and refill the universe with matter and energy .
another scenario that appears in the literature is the one of a x - matter plus cold dark matter ( the well known xcdm parametrization ) where both fluid components are separately conserved .
this cosmological model is characterized by an equation of state @xmath30 .
the model dominated by the cosmological constant @xmath31 , i.e. , @xmath1 is a limiting case of the x - matter parametrization . in terms of energy density
we have that , using ( [ 23 ] ) , @xmath32 , and the energy density is constant in any epoch .
hence , we can say that the cosmological constant corresponds to a fluid with a constant equation of state @xmath1 .
let us begin reviewing some basic concepts .
the einstein field equations are @xmath33 where @xmath0 has often been treated as a constant of nature and where @xmath34 is the einstein tensor given by @xmath35 and @xmath36 is the energy - momentum tensor .
we are using relativistic units ( i.e. , @xmath37 ) and we are assuming that the universe is homogeneous and isotropic . in this case
we will use the robertson - walker line element .
let us consider also a perfect - fluid - like ordinary matter with pressure @xmath38 and energy density @xmath15 @xcite .
taking the covariant divergence of ( [ 1 ] ) and using bianchi identities to guarantee the vanishing of the covariant divergence of the einstein tensor , it follows ( assuming the energy - momentum conservation law ) that the covariant divergence of @xmath39 must vanish also and therefore @xmath0 is constant constituting a geometrical interpretation of @xmath0 . as usual in recent works , we can move the cosmological term to the right side of ( [ 1 ] ) in order to reconstruct the einstein equation with an alternative form such as , @xmath40 where @xmath41 and we can interpret the cosmological term as part of the matter content of the universe , rather than a geometrical term . with this effective momentum - tensor @xmath42 satisfying the energy conservation as @xmath43 there are no reasons in order to refuse a varying cosmological terms @xcite .
at this point we have to observe that since we use @xmath37 , we will not consider here cosmological models with a time varying speed of light @xcite .
the incorporation of the cosmological term in the definition written in ( [ 4 ] ) implies that the effective energy - momentum tensor has the perfect fluid form , with an effective pressure given by @xmath44 and an energy density as @xmath45 @xcite . using the field equations ( [ 3 ] ) and
( [ 4 ] ) and the law of energy - momentum conservation , equation ( [ 5 ] ) , we can write @xmath46 and @xmath47\,=\,-3\left(p\,-\,\frac{\lambda}{8\pi g}\right)a^2\,\,.\ ] ] for the equation of state we can write that @xmath48 with @xmath49 constant . note from ( [ 8 ] ) that for @xmath50 we have matter - dominated epoch ( @xmath19 ) and for @xmath51 we have radiation - dominated epoch .
notice that , in order to obtain a convenient form for athe nest equations we constructed the equation of state in ( [ 8 ] ) . in a de scenario , as said in the last section , we are clearly making @xmath52 .
the reason for this will be clarified in the next lines . substituting ( [ 8 ] ) in ( [ 7 ] ) we have that @xmath53 now , using ( [ 6 ] ) and ( [ 9 ] ) we obtain that @xmath54 using ( [ 6 ] ) to eliminate @xmath15 in ( [ 10 ] ) ( or vice - versa ) we can write the equation @xmath55 which is a differential equation for the scale factor and has a cosmological term .
we clearly see that this equation is independent of the fact that @xmath0 is constant or not @xcite .
our objective from now on is to present analytical solutions for the scale factor in equation ( [ 11 ] ) .
but to accomplish this we have to choose a relevant expression for the cosmological term .
as said above , there is a whole literature that approaches this issue ( see @xcite and the references therein ) . here we choose a specific one ( well known in the literature ) but , we produce a variation in the curvature also .
we can summarize the forms of the cosmological constant depending on the scale factor in a general expression as @xcite @xmath56 where @xmath57 , @xmath58 and @xmath59 are constants .
ozer and taha @xcite and chen and wu @xcite concluded , although in different contexts , that the time dependence of @xmath0 should have @xmath60 and @xmath61 @xcite .
the interested reader can see the different variations concerning these parameters in @xcite . hence , let us choose the cosmological term as @xmath62 where @xmath63 ( we use here the same parameter used in @xcite ) is a pure number of order @xmath64 which should be calculable in a model for the time variation of @xmath0 .
we can consider that the parameter @xmath63 is a new cosmological parameter to be computed from observations , substituting @xmath0 . if @xmath65 we have , obviously , an ansatz of @xmath66 .
with a time variation like the exposed in ( [ 11.1 ] ) the values of @xmath0 in the early universe could be much bigger than the present value of @xmath0 .
however , even that order of small magnitude it was arguably large enough to drive various symmetry breakings that occurred in the early universe . with a @xmath66 ansatz we can not have such scenario .
otherwise , if @xmath67 , the equation ( [ 11.1 ] ) can be identified as a medium " time variation @xcite . for a positive @xmath63
our universe is flat .
in fact , considerations based on the second law of thermodynamics lead to the restriction @xmath68 @xcite . also , if @xmath69 the universe will expand forever ; if @xmath70 the universe will collapse in the future . for @xmath71 in equation (
[ 11 ] ) , even a closed universe will expand forever and the @xmath72 case is no longer critical for the collapse of our universe @xcite .
it can be shown that if our universe is flat , observationally @xmath63 should be positive @xcite .
if we write equation ( [ 12 ] ) in a general form , i.e. , as a function of the scale factor @xmath73 , like in the first term in equation ( [ 11.1 ] ) , the value @xmath74 has dimensional reasons . with @xmath74 , both @xmath75 and @xmath76
disappear and we have , with @xmath37 only , the equation ( [ 12 ] ) @xcite
. the presence of @xmath75 on the einstein equations would be rather disturbing .
nevertheless , equation ( [ 12 ] ) can be obtained theoretically based on general assumptions on quantum gravity .
it can be considered also that this formulation has no conflicts with current data and can furnish some explanation about the inflationary scenario @xcite . in other words
we can say that this model is singular and preserves the standard picture of the early universe .
notice that the condition @xmath77 requires that @xmath78 independently of the curvature index @xmath79 and hence @xmath80 for all @xmath81 @xcite .
we will not make inflation considerations for the time being . in another point of view ,
the dependence on @xmath82 used in ( [ 12 ] ) was first introduced by gasperini @xcite in a thermal approach .
an alternative motivation for this kind of dependence is that it also appear in some string dominated cosmologies @xcite .
we can also obtain an equation ( [ 12 ] ) form based on semiclassical lorentzian analysis of quantum tunneling @xcite .
although this kind of dependence is only phenomenological and does not come from particle physics first principles @xcite it can be considered as a laboratory in order to investigate the models it generalizes . as mathematically simple , we can obtain some analytical solutions .
after this brief historical review , let us now substitute ( [ 12 ] ) in ( [ 11 ] ) to have that @xmath83 which can be written as in @xcite as @xmath84 where @xmath85 and @xmath86 , @xmath79 and @xmath87 constitute a group of free parameters which have the respective physical considerations mentioned above . in our analysis , the parameter @xmath79 ,
as said above , will have the values @xmath88 , @xmath5 and @xmath89 .
just to remember , if we observe ( [ 12 ] ) we can see clearly that if @xmath90 , then the expression covers the @xmath66 case . if @xmath91 @xcite we have two options , @xmath92 where the universe will expand forever or @xmath93 and then the universe will collapse in the future . if @xmath94 , even a closed universe ( with @xmath95 ) will expand forever and the @xmath72 case is no longer critical for the collapse of our universe @xcite . for the case @xmath96 and substituting in ( [ 14 ] ) we have that @xmath97 after some algebraic work we can reduce the solution to a simple form as @xmath98 where @xmath99 and @xmath100 are constants to be determined . observe that we have written the equation above in order to maintain the parameter @xmath101 in evidence for the following analysis below . for this case of @xmath96 , from equations ( [ 15 ] )
we have that @xmath102 and consequently that @xmath101 has a dependence on @xmath79 and @xmath87 that can be written as @xmath103 and we have that for the values of @xmath104 , @xmath105 and the analysis about the expansion or collapse of the universe can be accomplished as above .
since the value @xmath90 can be obviously discarded , we can stress that the general behavior of the solution ( [ 18 ] ) does not change . as a final observation we can say for @xmath96 the equation of state
can be written as @xmath106 , where we have an inflation scenario . as usual we measured in units of hubble times ( @xmath107 ) and we have imposed the following boundary conditions at present epoch :
@xmath108 after have calculated the expression for the hubble parameter we used these boundary conditions to determine the free parameters values in the scale factor since we have two equations ( scale factor and hubble parameter ) and two incognites c1 and c2 . to plot the figures of this work we have set @xmath109 the @xmath110 that is inside of the current accept value of @xmath109 ( @xmath111 @xmath112 @xmath113 @xmath114 @xmath115 @xmath116 @xmath117 ) @xcite . in figure 1 , after the change of variable @xmath118 , we present the plot of the scale factor versus @xmath119 with @xmath120 and @xmath121 varying from @xmath122 to @xmath123 . [ fig1 ] for @xmath124 we have two general solutions for the equation ( [ 13 ] ) , @xmath125 \,\,,\nonumber\\ a_{2}(t)\,=\,{\frac{1}{4c_{1 } } } \left [ e^{-c_{1}(t+c_{2})}\,-4\,\lambda _ { 2}\,e^{c_{1}(t+c_{2})}\right ] \,\,,\end{aligned}\ ] ] where the difference is the signs of the exponentials .
for @xmath124 we have a little different behavior . from ( [ 15 ] ) we have that @xmath126 and consequently that @xmath127 , so we can write that @xmath128 , @xmath129 and @xmath130
. for @xmath131 the solutions ( [ 19 ] ) have different behaviors altogether , namely , @xmath132 and the equation of state now is @xmath31 , again an inflation scenario .
for @xmath133 we have that , @xmath134 \,\,,\nonumber\\ a_{2}^{k=\pm1}(t ) & = { \frac{1}{4c_{1 } } } \left [ e^{-c_{1 } ( t+c_{2})}% \,\mp\,4\,e^{c_{1}(t+c_{2})}\right ] \,\,,\end{aligned}\ ] ] and we can see again completely different formulations for the scale factor in an inflation scenario . in figure 2 , we present the plot of the scale factor versus @xmath135 first equation ( 23 ) , with @xmath136 and @xmath137 @xmath138 [ fig2 ] in figure 3 , we present the plot of the scale factor versus @xmath135 second equation 23 , with @xmath136 and @xmath137 @xmath139 [ fig3 ] in the figure 4 , we present the plot of the scale factor versus @xmath135 first equation ( 23 ) , with @xmath136 and @xmath140 @xmath141 with singularity in @xmath119=@xmath142 . [ fig4 ] for @xmath143 we have the following solution , @xmath144 and from equations ( [ 15 ] ) we have that @xmath145 and @xmath146 and the general behavior of the solutions ( [ 20 ] ) do not change since we have the condition @xmath147
. however , we have to notice that the square root demands attention in order to not obtain imaginary scale factors for @xmath148 .
but this is a trivial task to substitute in ( [ 20 ] ) the values of @xmath101 obtained in ( [ 20.1 ] ) with different @xmath79 .
we will not have any new physical feature in this specific mathematical analysis .
the equation of state is @xmath149 , a radiation - dominated epoch as said above .
next we show a solution that does not have constraints in @xmath150 and @xmath101 besides those given by the equations ( 16 ) @xmath151 if we set @xmath136 and @xmath140 the solution has a singularity in @xmath119 @xmath152 figure 5 show this result .
the equation ( [ 12 ] ) was studied in @xcite and it can also be obtained theoretically from some assumptions from quantum cosmology . it is not in conflict with experimental observations and with the inflationary scenario . from quantum cosmology arguments
it is more convenient to use the scale factor instead of the age @xmath153 of the universe @xcite . in this paper
we have computed some analytical solutions for the scale factor since the effective cosmological constant @xmath0 is time varying according to an inverse square law in the scale factor .
this inverse square law can be justified based on the principles of quantum cosmology .
since such time variation of @xmath0 leads to creation of matter and/or radiation , we believe that the formulation of a scale factor analysis is of current interest both theoretical and experimental . another motivation can be found in the strict relation between the cosmological constant problem and the problem of the age of the universe treated in the light of quantum cosmology . there is a great literature about different functions for the cosmological constant . specifically concerning a function like in the equation ( [ 12 ] )
the existing published papers do not bring any analytical expressions for the scale factor itself as a building term in @xmath0 .
as we are interested in a general analytical behavior of @xmath0 , numerical solutions of ( [ 14 ] ) are out of the scope of this work .
therefore we found three analytical solutions .
other alternatives varying the parameter @xmath154 in ( [ 14 ] ) brought singularities which demands numerical efforts .
our main motivation is to show simply that there is another possible and different approach concerning the cosmological constant in order to obtain some progress in some questions about the universe .
although the model proposed by chen and wu in equation ( [ 12 ] ) are compatible with the astrophysical data , it alters the predictions of the standard model for the matter - dominated epoch . as a future perspective
it can be investigated , with the solutions obtained here , some aspects of inflation like the current value of the deceleration parameter and the parameter @xmath155 , namely , to obtain its value and test if they are compatible with the standard model demands .
note that the knowledge of the form of the scale factor , as well known , can describe if an universe is an expanding one or not through the line element .
an investigation about the age of the universe in different scenarios as well as the equations of state , with the results obtained in this paper , is a work in progress .
the authors would like to thank prof .
felipe t. falciano for helpfull discussions ; fundao de amparo pesquisa do estado do rio de janeiro ( faperj ) and conselho nacional de desenvolvimento cient ' ifico e tecnolgico ( cnpq ) ( brazilian agencies ) for financial support .
if @xmath165 we have the following solutions @xmath166 @xmath167 @xmath168 @xmath169 @xmath170 @xmath171 @xmath172 @xmath173 @xmath174 @xmath175 @xmath176 @xmath177 @xmath178 @xmath179 @xmath180 @xmath181 @xmath182 @xmath183 s. weinberg , rev .
61 ( 1989 ) 1 ; y. b. zeldovich , sov .
jetp 14 ( 1968 ) 1143 ; usp 11 ( 1968)384 ; a. d. sakarov , sov .
dokl . 12 ( 1968 ) 1040 ; l. h. ford , phys . rev .
d 11 ( 1975 ) 2270 ; e. steeruwiotz , phys .
d 11 ( 1975 ) 3378 ; p. j. e. peebles and b. ratra , astrophys .
j. 325 ( 1988 ) l17 .
v. sahni and a. a. starobinsky , int . j. mod
d 9 ( 2000 ) 373 ; p.
j. e. peebles and b. ratra , rev .
phys . 75 ( 2003 ) 559 ; t. padmanabhan , phys . rept . 380
( 2003 ) 235 ; j. a. s. lima , braz .
34 ( 2004 ) 194 ; d. h. weinberg , new astron .
49 ( 2005 ) 337 ; e. j. copeland _
et al _ , hep - th/0603057 ; j. s. alcaniz , braz . j. phys .
36 ( 2006 ) 1109
. p. j. e. peebles and b. ratra , astrophys .
325 , l17 ( 1988 ) ; c. wetterich , nucl .
b302 , 668 ( 1988 ) ; j. a. frieman , c. t. hill , a. stebbins and i. waga , phys .
lett . 75 , 2077 ( 1995 ) ; r. r. caldwell , r. dave and p. j. steinhardt , phys . rev
80 , 1582 ( 1998 ) ; p.g .
ferreira and m. joyce , phys .
d58 , 023503 ( 1998 ) ; t. padmanabhan , phys . rev .
d 66 , 021301 ( 2002 ) ; f. c. carvalho , j. s. alcaniz , j. a. s. lima and r. silva , phys .
97 , 081301 ( 2006 ) .
astro - ph/0608439 .
r. r. caldwell , phys .
b545 , 23 ( 2002 ) ; v. faraoni , int . j. mod
d 11 , 471 ( 2002 ) ; r. r. caldwell , m. kamionkowski , and n. n. weinberg , phys .
91 , 071301 ( 2003 ) ; p. f. gonzalez - diaz , phys .
d 68 , 021303r ( 2003 ) ; s. nojiri and s.d .
odintsov , phys .
b 562 , 147 ( 2003 ) ; j. s. alcaniz , phys . rev .
d 69 , 083521 ( 2004 ) .
astro - ph/0312424 ; s. nesseris and l. perivolaropoulos , phys .
d 70 , 123529 ( 2004 ) ; j. santos and j. s. alcaniz , phys .
b 619 , 11 ( 2005 ) .
astro - ph/0502031 ; s. nesseris and l. perivolaropoulos , jcap 0701 , 018 ( 2007 ) . j. s. dunlop _
et al . _ , nature 381 , 581 ( 1996 ) ; j.s .
dunlop , 1999 , in the most distant radio galaxies " , eds . h. j. a. rttgering , p. n. best , and m. d. lehnert , p. 71 ; h. spinrad _
et al . _ ,
j. 484 , 581 ( 1997 ) ; l. a. nolan _ et al .
_ , mon . not .
323 , 385 ( 2001 ) ; l. m. krauss , astrophys .
j. 480 , 486 ( 1997 ) ; j. s. alcaniz and j. a. s. lima , astrophys . j. 521 , l87 ( 1999).astro - ph/9902298 ; j. a. s. lima and j. s. alcaniz , mon . not .
317 , 893 ( 2000).astro - ph/0005441 ; j. s. alcaniz and j. a. s. lima , astrophys .
j. 550 , l133 ( 2001 ) .
astro - ph/0101544 ; g. hasinger , n. schartel and s. komossa , astrophys . j. 573 , l77 ( 2002 ) ; j. s. alcaniz , j. a. s. lima and j. v. cunha , mon . not .
soc . 340 , l39 ( 2003 ) .
astro - ph/0301226 ; d. jain and a. dev , phys .
b 633 , 436 ( 2006 ) .
a. buchalter , d. j. helfand , r. h. becker , and r. l. white , astrophys .
j. 494 , 503 ( 1998 ) ; l. i. gurvits , k. i. kellermann and s. frey , astron . astrophys .
342 , 378 ( 1999 ) ; j. a. s. lima and j. s. alcaniz , astron .
astrophys .
357 , 393 ( 2000 ) .
astro - ph/0003189 ; j. s. alcaniz , phys .
d 65 , 123514 ( 2002 ) .
astro - ph/0202492 ; r. g. vishwakarma , class .
18 , 1159 ( 2001 ) ; j. a. s. lima and j. s. alcaniz , astrophys . j. 566 , 15 ( 2002 ) .
astro - ph/0109047 ; z. h. zhu and m. k. fujimoto , astrophys . j. 581 , 1 ( 2002 ) ; g. chen and b. ratra , astrophys .
j. 582 , 586 ( 2003 ) ; d. jain , a. dev and j. s. alcaniz , class .
20 , 4485 ( 2003 ) .
astro - ph/0302025 ; s. podariu , r. a. daly , m. p. mory and b. ratra , astrophys .
j. 584 , 577 ( 2003 ) ; l. i. gurvits , new astron .
48 , 1211 ( 2004 ) ; z. h. zhu , m. k. fujimoto and x. t. he , astron .
astrophys .
417 , 833 ( 2004 ) ; j. c. jackson , jcap 0411 , 007 ( 2004 ) .
s. w. allen , s. ettori , and a.c .
fabian , mon . not .
roy . astron .
324 , 877 ( 2002 ) ; s.w .
allen , r.w .
schmidt , and a.c .
fabian , mon . not .
astron . soc .
334 , l11 ( 2002 ) ; j. a. s. lima , j. v. cunha and j. s. alcaniz , phys .
d 68 , 023510 ( 2003 ) .
astro - ph/0303388 ; s. w. allen _ et al .
_ , mon . not .
353 , 457 ( 2004 ) ; g. chen and b. ratra , astrophys .
j. 612 , l1 ( 2004 ) ; j. v. cunha , j. s. alcaniz and j. a. s. lima , phys .
d 69 , 083501 ( 2004 ) .
astro - ph/0306319 ; s. w. allen _ et al .
_ , mon . not .
383 , 879 ( 2008 ) ; z. h. zhu , m. hu , j. s. alcaniz and y. x. liu , astro.astrophys . 483 , 15 ( 2008 ) .
arxiv:0712.3602 [ astro - ph ] ; l. samushia and b. ratra , astrophys . j. 680 , l1 ( 2008 ) .
c. s. kochaneck , astrophys .
j. 466 , 638 ( 1996 ) ; a.r .
cooray and d. huterer , astrophys .
513 , l95 ( 1999 ) ; i. waga and a. p.
m. miceli , phys .
d 59 , 103507 ( 1999 ) ; d. jain , a. dev and j. s. alcaniz , phys .
d 66 , 083511 ( 2002 ) .
astro - ph/0206224 ; a. dev , d. jain and j. s. alcaniz , phys .
d 67 , 023515 ( 2003 ) .
[ arxiv : astro - ph/0209379 ] ; d. jain , a. dev , n. panchapakesan , s. mahajan and v. b. bhatia , int .
j. mod .
d 12 , 953 ( 2003 ) ; k .- h .
et al . _ ,
89 , 151301 ( 2002 ) ; k. h. chae , g. chen , b. ratra and d. w. lee , astrophys . j. 607 , l71 ( 2004 ) . | the accelerated expansion of the universe has now been confirmed by several independent observations including those of high redshift type ia supernovae , and the cosmic microwave background combined with the large scale structure of the universe .
another way of presenting this kinematic property of the universe is to postulate the existence of a new and exotic entity , with negative pressure , the dark energy ( de ) . in spite of observationally
well established , no single theoretical model provides an entirely compelling framework within which cosmic acceleration or de can be understood . at present
all existing observational data are in agreement with the simplest possibility of the cosmological constant to be a candidate for de .
this case is internally self - consistent and non - contradictory .
the extreme smallness of the cosmological constant expressed in either planck , or even atomic units means only that its origin is not related to strong , electromagnetic and weak interactions .
although in this case de reduces to only a single fundamental constant we still have no derivation from any underlying quantum field theory for its small value . from the principles of quantum cosmologies , it is possible to obtain the reason for an inverse - square law for the cosmological constant with no conflict with observations .
despite the fact that this general expression is well known , in this work we introduce families of analytical solutions for the scale factor different from the current literature .
the knowledge of the scale factor behavior might shed some light on these questions since the entire evolution of a homogeneous isotropic universe is contained in the scale factor .
we use different parameters for these solutions and with these parameters we stablish a connection with the equation of state for different dark energy scenarios .
.2 in |
cold molecules currently attract great attention due to their potential application for precision measurements @xcite , quantum information schemes @xcite , degenerate quantum gases with complex interactions @xcite , and for cold chemistry @xcite . besides the approach of forming cold molecules out of a sample of ultracold trapped atoms @xcite , recent experimental efforts mainly focus on developing techniques of slowing or filtering , cooling , and trapping molecules produced in effusive or nozzle beams .
well established techniques are buffer gas cooling @xcite , the deceleration of beams of polar molecules using time - varying electric fields @xcite , as well as filtering of polar molecules out of an effusive source using static or time - varying electric fields @xcite .
recently , the method of decelerating supersonic molecular beams has been extended to magnetic @xcite and optical fields @xcite .
alternative routes to producing cold molecules have been demonstrated utilizing the kinematics in elastically or reactively colliding molecular beams @xcite .
furthermore , producing beams of slow and cold atoms and molecules by mechanical means has been demonstrated @xcite . in particular the technique of translating a supersonic jet to low longitudinal velocities by means of a rapidly counter - rotating nozzle , pioneered by gupta and herschbach ,
holds the promise of producing cold , slow and intense beams of nearly any molecular species available as a gas at ambient temperatures @xcite .
attempts to improve the original arrangement have been made by the groups of h .- j .
loesch in bielefeld and of m. dekieviet in heidelberg @xcite . by controlling the sense and speed of rotation of the nozzle as well as by using the seeded beams technique , tunable beam velocities ranging from thousands down to a few tens of meters per second are feasible .
typically , the translational and internal temperatures in the moving frame of the slowed beam are in the range of a few kelvin and a few tens of kelvin , respectively . due to the extremely simple concept
, this technique can be realized by a relatively simple mechanical device .
therefore this versatile source of cold molecules may be particularly useful for novel studies of molecular reaction dynamics at low collision energies in the gas phase or with surfaces @xcite .
the principle of operation relies on the galilei transformation of particle velocities in the moving frame of the rotating nozzle into the resting laboratory frame . as the nozzle moves backwards with respect to the molecular beam motion
the beam velocity @xmath0 is reduced to @xmath1 . in this way ,
gupta and herschbach succeeded in slowing beams of krypton ( kr ) and xenon down to 42 and 59 m / s , respectively , at beam temperatures down to 1.5 and 2.6 k , respectively @xcite . in their prototype setup
they used a tapered aluminium barrel with a hollow bore as a rotor .
the special aluminum alloy was chosen for its favorable ratio of tensile strength to mass density , allowing high rotational speeds . however , machining such a rotor barrel requires considerable efforts .
the gas injection was realized by inserting a polymer needle into the rotor .
however , erosion of this type of sealing and the resulting gas leakage was found to be a serious limitation .
the nozzle orifice was fabricated by gluing a stainless steel disk with a 0.1 mm aperture close to the tip of the rotor . in this case , swatting of slowly emerging molecules at high rotational speeds by the rotor arm as it comes back after one full revolution may be an issue . furthermore , a general drawback of producing slow beams using a moving source is the fact that the flux of molecules @xmath2 on the beam axis drops down sharply with decreasing laboratory beam velocity @xmath3 and with increasing distance @xmath4 of the detector from the nozzle , @xmath5 .
this is a consequence of the beam spreading in transverse directions which becomes more and more important as @xmath3 decreases and hence the flight time goes up .
the aim of the improved setup presented in this paper is to remedy the drawbacks mentioned above .
schematic overview over the experimental arrangement used for characterizing the rotating nozzle molecular beam source.,width=226 ] the experimental arrangement used for characterizing the new rotating nozzle setup is schematically depicted in fig .
[ fig : schematic ] .
the upper part represents the moving molecular beam source including a hollow rotor and a skimmer with 1 mm diameter .
the nozzle chamber is pumped by an 8000l / s oil diffusion pump backed by roots and rotary vane pumps .
typical background gas pressure during operation with argon ( ar ) at @xmath6bar nozzle pressure is @xmath7 mbar .
the molecular beam follows a trajectory through the following elements : at a distance of about 20 mm behind the nozzle it enters an intermediate chamber through a skimmer with 1 mm opening diameter .
the intermediate chamber is pumped by a 600l / s turbo pump providing a background gas pressure @xmath8 mbar during operation .
the intermediate chamber is connected to a detector chamber pumped by a 150l / s turbo pump to maintain a background pressure @xmath9 mbar . for the purpose of guiding beams of polar molecules , a set of four parallel steel rods is installed around the beam axis between the skimmer and a commercial quadrupole mass spectrometer ( qms ) ( pfeiffer qms200 ) which we use as a detector ( see sec .
[ sec : guide ] ) .
the mounting discs used for holding the rods also serve as apertures for differential pumping between the intermediate chamber and the detector chamber . the total fight distance from the nozzle to the detector
amounts to @xmath10 cm .
since the positions of the quadrupole guides and the qms detector can not be adjusted , the beam axis is initially aligned by sight by adjusting the position of the skimmer . for additional characterization of the beam ,
a pitot tube is installed about @xmath11 mm behind the ionizer of the qms detector .
it consists of a cylindrical chamber 4 cm in diameter and 10 cm in length having a conical aperture on the side close to the qms and a cold cathode ion gauge on the opposite side as an end cap .
when a continuous molecular beam ( nozzle at rest ) is directed into the detector chamber , the residual gas pressure inside the pitot tube rises until equilibrium between incoming flux and outgoing flux of molecules effusing out of the pitot tube aperture according to its molecular flow conductance is reached .
thus , from the difference between pressures inside the pitot tube and in the qms chamber , @xmath12 , the flux of molecules on the beam axis , @xmath2 , can be inferred according to @xmath13 .
here @xmath14 , @xmath15 , and @xmath16 denote the area of the entrance channel into the pitot tube , a scaling parameter to account for the channel length , and the mean velocity of the molecules at ambient temperature @xmath17 .
the on - axis beam density , which is obtained from the molecule flux using @xmath18 , is compared to the one determined by the qms signals in sec .
[ sec : results ] .
overview ( a ) and close - up views ( b , c ) of the experimental setup . ( a ) the nozzle orifice is mounted on one tip of the hollow rotor and produces molecule pulses along the beam axis each time it passes nearby the skimmer .
the gas is injected into the rotor through a rotary feedthrough from below .
the motor is mounted to a massive brass mount on top of the rotor , to which it is connected by a flexible coupler ( silicone tube ) .
( b ) sectional view of the central part of the setup .
( c ) close - up view of the rotor tip , which contains a bored thin - walled titanium ferrule as a nozzle.,width=264 ] the experimental realization of our rotating source of cold atoms and molecules is depicted in fig .
[ fig : setup ] ( a ) .
the following main modifications with respect to the work of gupta and herschbach have been implemented : the rotor is made of a thin carbon fiber ( cfk ) tube with 4 mm outer diameter and 38 cm in length .
cfk is superior to any metallic material regarding the ratio of tensile strength to density by at least a factor of @xmath19 .
moreover , the small mass of the cfk tube makes it simpler to suppress deterioral mechanical vibrations that may occur due to slight imbalance of the rotor at high speeds of rotation .
the cfk tube is sealed on both ends by gluing thin - walled ( 0.25 mm ) titanium grade 5 ferrules as end - caps into the inner surface of the cfk tube using special epoxy for cfk ( fig .
[ fig : setup ] ( c ) ) .
one of the ferrules has a borehole @xmath20 mm in diameter at a distance of 0.5 mm from the closed end , thus forming a well defined nozzle aperture . a simple alternative that we also tried is capping the cfk tube by gluing circular pieces of capton foil to front and back sides of the tube .
capping the cfk tube this way has proven to be stable enough to withstand the high centrifugal forces acting at rotational speeds up to @xmath21hz ( 21000 rpm ) .
the nozzle orifice was fabricated by inserting a @xmath22 m wire half way in between the cfk front side and the capton foil prior to gluing .
after the epoxy has hardened the wire is pulled out and an approximately 1 mm long channel remains .
this method allows to easily realize small nozzle channels of diameters @xmath23 m , which limits the gas load and thus the required pumping speed .
moreover , using this method of fabrication the distance of the nozzle from the tip of the rotor is kept minimum , thereby suppressing swatting of slow molecules by the rotor .
however , the expansion out of a long nozzle channel is unfavorable compared to a thin - walled orifice as far as adiabatic cooling conditions are concerned . besides
, nozzles fabricated using this procedure vary considerably with respect to length and shape of the nozzle channels . for all measurements presented in this work
, bored thin - walled titanium ferrules are used .
the cfk tube is inserted and glued into the core of the rotor consisting of an aluminium or brass cylinder ( fig .
[ fig : setup ] ( b ) ) .
the latter has a borehole along its axis for inserting the bored shaft of a rotary feedthrough from below and the shaft of a flexible coupler attached to a motor ( faulhaber-group.com ) from above . in order to avoid any imbalance it is important to perfectly align the motor axis and the feedthrough axis
this is ensured by inserting the two axes into the same through - hole through the rotor core .
the rotor including the metal core is first balanced statically by suspending the rotor core between two needles in the center of gravity and by abrading cfk material from the heavier side of the rotor until it is perfectly balanced . in a second step ,
the rotor is dynamically balanced by adding screws to the rotor core at different places and by monitoring the vibration spectrum using acceleration sensors attached to the setup .
using this procedure , vibrations at high rotation speeds resulting from the rotor are small compared to vibrations induced by the motor itself .
therefore , additional masses are attached to the motor ( fig .
[ fig : setup ] ( b ) ) , which reduce vibrations at frequencies @xmath24hz .
gas injection into the rotor is realized using a commercial rotary feedthrough sealed by ferrofluidic metal sealings .
the bored shaft of the feedthrough is inserted into the rotor core and o - ring sealed .
metal - sealed feedthroughs ( ferrotec.com ) features very high leak tightness ( @xmath25mbarl / s ) at low friction torque .
the rotor axis is aligned with the motor axis by connecting the rotary feedthrough and the motor to a u - shaped brass construction .
this brass mount is broken in two parts to introduce a rubber layer in between the two parts as a damping element for suppressing vibrations .
the brass mount is fixed to a triangular massive metal plate supported by neoprene dampers . in our new setup ,
the neoprene damper positioned underneath the skimmer resides on a vertical translation stage and the remaining two dampers reside on two horizontal translation stages .
the translation stages are actuated by turning knobs outside of the vacuum chamber which allows full adjustment of the nozzle position with respect to the beam axis even during operation ( cf .
[ fig : twopeaks ] ) . in our present setup
the maximum speed of rotation is limited to about 350hz ( 21000rpm ) .
the limiting factor so far is mechanical vibrations , which rapidly increase in amplitude at frequencies @xmath26hz .
the origin of these vibrations is presumably a slight unbalance of the rotor as well as vibrations caused by the motor .
these vibrations are found to reduce the lifetime of the ferrofluidic sealings of the feedthrough , which is specified to rotation frequencies @xmath27hz ( 10000rpm ) .
the motor as well as the cfk rotor , however , are able to withstand much higher frequencies up to 970 and 800hz , respectively .
illustration of the dependence of measured signal ( right side ) on the alignment of the rotating nozzle setup with respect to the beam axis ( dotted line ) .
( a ) proper nozzle alignment ; when the rotor is set to a perpendicular position with respect to the beam axis , nozzle and skimmer are in line with the beam axis up to the detector .
( b ) nozzle shifted above the beam axis ; two pulses are produced each time the nozzle intersects the beam axis at two slightly tilted rotor positions .
( c ) nozzle shifted below the beam axis ; one pulse is measured with reduced amplitude and width.,width=302 ] although the expansion out of the rotating nozzle is continuous , the beam passing through the skimmer and entering the detector chamber is chopped .
each time the nozzle comes close to the skimmer part of the ejected particles travel along trajectories that match the beam axis , determined by the positions of skimmer , quadrupole guides , and qms detector . using the vertical and horizontal translation stages
the rotor position can be aligned during rotation of the nozzle to optimize the detector signal .
a peculiarity of the rotating - nozzle arrangement is the fact that the signal traces as a function of time are changed in an asymmetric manner as the rotor is shifted horizontally across to the beam axis , as illustrated in fig .
[ fig : twopeaks ] .
the signal traces are recorded using ar gas expanding out of the nozzle at a nozzle pressure @xmath28 mbar and at a frequency of rotation of 10hz . at this rotation frequency , the flight time of ar atoms to the detector is short compared to the nozzle motion in front of the skimmer .
thus , the recorded signal traces mainly reflect the effective slit opening function given by the skimmer with respect to the moving nozzle . in fig .
[ fig : twopeaks ] ( a ) , the nozzle is properly aligned such that the signal trace features a slightly asymmetric trapezoidal shape . as the nozzle is moved across the beam axis by 2 mm , the measured signal splits into two maxima , as shown in fig .
[ fig : twopeaks ] ( b ) . in this geometry
, particle trajectories matching the beam axis occur at two different angles of inclination of the rotor with respect to the beam axis .
misalignment by 2 mm in the other direction leads to a reduction of the signal width and amplitude , as shown in fig .
[ fig : twopeaks ] ( c ) . the inclination signal envelopes in fig .
[ fig : twopeaks ] ( a ) and ( b ) results from the decreasing solid angle defined by nozzle and skimmer aperture as the nozzle backs off away from the skimmer .
slit opening function measured at low rotation speeds of the nozzle ( 10hz ) ( solid line ) .
the dashed line represents a model curve fitted to the data , consisting of the sum of 3 gaussian functions ( light solid lines).,width=340 ] in order to extract the exact values of the beam velocity , temperature , and density from the measured signal traces , the effective slit opening function has to be characterized quantitatively and included into a model for the time - dependent density distribution of the molecular beam . the effective slit opening function for optimum horizontal nozzle alignment measured at low rotor frequency ( 10hz )
is plotted in fig .
[ fig : slit ] ( solid line ) . at such low rotor frequency longitudinal dispersion of the beam pulse during flight time is negligible .
this is verified measuring the signal intensity at rest as a function of the rotor position using a micrometer driven manipulator .
the resulting curve has identical shape compared to the time of flight curves at low rotor velocities . the dashed red line in fig .
[ fig : slit ] represents a fitted model function consisting of the sum of 3 gaussian functions , @xmath29 where amplitude factors @xmath30 , center positions @xmath31 , and widths @xmath32 are determined by the fit to the experimental data . the density distribution in phase - space of a gas pulse produced by the rotating nozzle
is modeled by the expression @xmath33 which is normalized to the total number of molecules per bunch , @xmath34 .
the spatial distribution @xmath35 is assumed to be anisotropic according to @xmath36 the distribution in transverse direction , @xmath37 is normalized to @xmath38 .
the distribution along the beam axis ( @xmath39-axis ) is obtained from the slit opening function ( eq . ( [ eq : slit ] ) ) by rescaling the z - coordinate according to @xmath40 , where @xmath41 denotes the ratio between beam velocity @xmath42 and the velocity of the rotor tip , @xmath43 . here ,
@xmath44 for monoatomic gases , @xmath45 is the most probable velocity of the molecules inside the nozzle at a temperature @xmath46 , and @xmath47 cm stands for the rotor radius .
thus , @xmath48 where the amplitude factor @xmath49 is given by @xmath50 . the total number of molecules per pulse ( integral over space of @xmath51 ) is a function of speed of rotation , @xmath52 , where @xmath53 @xcite is the flux of molecules out of the nozzle orifice of area @xmath54 at pressure @xmath55 and @xmath56@xmath57 . in other words ,
as the nozzle velocity @xmath58 falls below the beam velocity @xmath0 , the molecule bunches become longer than the spatial width of the slit opening function @xmath59 and the number of molecules per bunch increases accordingly .
the velocity distribution is chosen according to the most frequently used ellipsoidal drifting maxwellian model @xmath60 , where @xmath61 with normalization constants
@xmath62 and @xmath63 @xcite .
the widths of the velocity distributions , @xmath64 and @xmath65 are related to the transverse and longitudinal beam temperatures @xmath66 and @xmath67 by @xmath68 . while @xmath67 results from adiabatic cooling of the gas in the jet expansion
, @xmath66 is an approximate measure of the transverse beam divergence .
typical values of @xmath67 attainable in free jet expansions range between @xmath69k and @xmath70k , while the transverse beam divergence corresponds to @xmath71k . during free expansion , in the absence of collisions and external forces
, the distribution function remains unchanged according to liouville s theorem .
the time evolution is given by @xmath72 the spatial density as a function of time is then obtained by integrating over velocities , @xmath73 the analytic solution of this expression is given by @xmath74 where @xmath75 and @xmath76 simulated density distributions on the beam axis as a function of time of flight at various speeds of rotation of the nozzle .
solid lines are calculated according to eq .
( [ eq : densitytime ] ) based on the measured slit - opening function , dashed lines reflect the longitudinal velocity distribution rescaled to flight times.,width=302 ] fig .
[ fig : faltung ] displays simulated ar density distributions on the beam axis , @xmath77 , according to eq .
( [ eq : densitytime ] ) for typical beam temperatures @xmath78k and @xmath79k at various speeds of rotation ( solid lines ) . here , @xmath80 m denotes the total flight distance from the nozzle to the qms detector .
the nozzle speed is @xmath81 m/s ( @xmath82hz ) in fig .
[ fig : faltung ] ( a ) , 24 m/s ( 20hz ) ( b ) , 179 m/s ( 150hz ) ( c ) , and 358 m/s ( 300hz ) ( d ) .
besides shifting linearly to longer flight times at higher rotational speeds , the distribution is first broadened at rotations around 0hz due to the effect of the slit opening function . at higher speeds ,
the distribution first narrows down as the nozzle moves faster past the skimmer ( @xmath83hz ) , and then becomes broader again as a consequence of longitudinal dispersion , _
i.e. _ due to the fact that longer flight times allow the molecule bunches to longitudinally broaden due to the finite beam temperature @xmath67 .
the dashed lines represent the neat velocity distributions horizontally rescaled to flight times according to @xmath84 .
thus , at negative speeds of rotation ( molecular beam is accelerated ) the measured signal is predominantly determined by the shape of the velocity distribution , while at speeds around 0hz the measured trace is strongly distorted by the slit - opening function . at higher rotation frequencies , the effect of the slit function diminishes and the two distributions eventually converge . therefore , at rotations @xmath85hz , eq .
( [ eq : zdensitytime ] ) may be replaced by the simpler expression @xmath86 in the procedure of fitting the measured data .
( a ) measured time of flight density distributions of decelerated argon beams at various speeds of rotation of the nozzle .
( b ) same data as in ( a ) , where the horizontal axis is rescaled to velocities.,width=302 ] typical time of flight distributions measured with ar at a nozzle pressure @xmath87 mbar are displayed in fig .
[ fig : tofs ] ( a ) for various speeds of rotation . in these measurements ,
the ar gas pressure inside the gas line was carefully adjusted such that the effective pressure at the nozzle remains constant . due to the centrifugal force
the nozzle pressure @xmath55 rises sharply at increasing speed of rotation @xmath58 while holding the gas - line pressure @xmath88 constant according to @xmath89 @xcite .
signal broadening due to the convolution with the slit opening function around 0hz as well as broadening due to longitudinal dispersion at high rotation frequencies are clearly visible ( _ cf .
_ inset of fig .
[ fig : tofs ] ( a ) ) .
the time of flight distributions shown in fig .
[ fig : tofs ] ( a ) are horizontally rescaled to velocity distributions according to @xmath90 and replotted in fig .
[ fig : tofs ] ( b ) . in this representation of the data
the signal broadening due to dispersion at high rotation frequencies cancels out .
however , broadening due to the slit opening function still remains and masks the real velocity distribution .
the absolute densities are estimated by scaling the qms detector signal to the output of the pressure gauge in the detector chamber . besides , densities can be estimated from the beam intensity measurement using the pitot as discussed in sec .
[ sec : setup ] .
however , the density values obtained in this way turn out to be about a factor of 10 higher than the ones from the qms measurement
. this may be due to some feedback effect of the cold cathode ion gauge onto the vacuum inside the small volume of the pitot tube , thus causing a nonlinear response of the vacuum gauge .
peak values of the time of flight argon density distributions as a function of the rotation frequency .
symbols are the measured values , the solid line represents the result of fitting ( eq . ( [ eq : densitytime ] ) ) to the data.,width=302 ] strikingly , the ar signal amplitude rapidly drops by about a factor of 30 when increasing the speed of rotation from 0hz to 300hz . in contrast , when the beam is accelerated by spinning the nozzle in the opposite direction up to 250hz , the signal only slightly increases by a factor of 2 . the dependence of the measured peak density from the speed of rotation is shown in fig .
[ fig : peakdens ] as filled symbols .
the solid line represents maximum values of the model function , @xmath91 , according to eq .
( [ eq : densitytime ] ) . here
, the amplitude factor of the model function , @xmath34 , is scaled by a factor of 0.5 to fit the experimental data .
this strong amplitude modulation as a function of sense and speed of rotation is caused by both longitudinal as well as transverse beam dispersion . since the rate of the signal drop is determined by the beam temperatures @xmath67 and
@xmath66 , it is evident that beam temperatures have to be kept low by ensuring optimum free jet expansion conditions , _
i.e. _ high effective nozzle pressure @xmath92 mbar at sufficiently low background pressure @xmath93 mbar as well as a high quality nozzle aperture and skimmer @xcite . measured time of flight density distributions of krypton at various speeds of rotation of the nozzle.,width=340 ] since the terminal beam velocity of jet beams is reduced with increasing mass of the beam particles , deceleration of beams of the heavier nobel gas atoms krypton ( kr ) and xenon to small velocities is achieved at lower speeds of rotation .
beams of kr , _
e.g. _ , have terminal velocities around @xmath94 m/s , such that deceleration down to velocities below 100 m/s are easily achieved at rotation frequencies @xmath95hz , as shown in fig .
[ fig : krtof ] .
however , due to longitudinal dispersion , consecutive beam packets start to overlap at frequencies @xmath95hz ( `` wrap around effect '' ) , which impedes quantitative analysis of the time of flight distributions
. one way of passing around this effect could be the implementation of an additional mechanical beam chopper to pick out individual beam packets . in comparison with other methods of producing beams of slow molecules ,
our results are quite similar regarding beam densities at low velocities . using electrostatic quadrupole filters , _
e.g. _ , typical peak densities around @xmath96@xmath97 at velocities around 60 m/s are achieved @xcite .
however , the rotating nozzle setup offers the possibility of tuning beam velocities in a wide range . due to free jet expansion conditions ,
the internal degrees of freedom of the molecules are expected to be cooled to the kelvin range , which can be reached using quadrupole filters only with substantial experimental efforts @xcite . besides , the rotating nozzle is suitable for any atomic of molecular substance available in the gas - phase at ambient temperatures , in contrast to filtering or deceleration techniques , which rely on particularly suitable stark of zeeman effects of molecules in external electrostatic or magnetic fields .
argon and krypton beam parameters as a function of nozzle pressure.,width=302 ] in order to characterize the beam parameters and to test the analytic model , flight distributions of decelerated ar and kr beams are measured at various expansion pressures and speeds of rotation and are fitted by the model function , eq .
( [ eq : densitytime ] ) .
free fit parameters are the total number of atoms per bunch , @xmath34 , the reduced beam velocity @xmath3 , and the longitudinal beam temperature @xmath67 .
the resulting values for ar and kr beams are depicted in fig .
[ fig : tempvel ] as symbols . clearly , the results do not depend on the speed of rotation of the nozzle within the experimental uncertainty .
as expected , the peak densities of ar and kr beams measured at rotation frequencies 150 and 100hz , respectively , evolve roughly proportionally to the nozzle pressure @xmath55 ( fig .
[ fig : tempvel ] ( a ) ) .
the kr densities fall behind the ones reached with ar by about a factor of 3 due to the smaller flux of kr out of the nozzle and due to the smaller longitudinal beam velocity of kr .
the fitted beam velocities @xmath0 shown in fig .
[ fig : tempvel ] ( b ) are found to perfectly match the theoretical values for supersonic beams , @xmath0 , provided the nozzle pressure exceeds about 100 mbar .
the latter are given by @xmath98 , where @xmath46 and @xmath55 are the nozzle temperature and pressure , @xmath99 is the background gas pressure and @xmath100 is the adiabatic exponent for a gas of particles with @xmath101 degrees of freedom @xcite . for mono - atomic gas ( @xmath102 ) and low background pressure @xmath103 we obtain the simpler formula @xmath104 , which yields @xmath105 m/s for ar and @xmath106 m/s for kr at @xmath107k . at lower nozzle pressure @xmath55 , the fitted values of @xmath3 slightly deviate from the model curve , indicating expansion conditions at the transition from the effusive to the free jet expansion regime .
fit values of the longitudinal temperature @xmath67 of ar beams are shown in fig .
[ fig : tempvel ] ( c ) as symbols .
the solid lines represent model curves according to two different collision models .
assuming hard - spheres ( hs ) collisions , the terminal speed ratio @xmath108 is given by @xmath109 , while atom - atom interactions according to the lennard - jones ( lj ) potential yield @xmath110 @xcite . here , @xmath111[@xmath112@xmath113 represents the hs - cross section , @xmath114[@xmath112k@xmath115 and @xmath116[@xmath117@xmath118 stand for the potential well depth and the equilibrium distance of the lj - potential for argon [ kr ] , respectively . in comparison with the experimental data ,
the model assuming interactions according to the lj - potential appears to yield more precise values than the one assuming hs - collisions .
however , the lj - model systematically slightly under - estimates longitudinal temperatures , in contrast to the hs - model , which gives too large values .
the measured temperature continuously decreases as a nozzle pressure up to about 1bar is applied , reaching values @xmath119k with both ar and kr .
evidently , high nozzle pressures are desirable if low temperatures are to be achieved .
attenuation of the krypton peak beam intensity as a function of nozzle pressure for different speeds of rotation of the nozzle.,width=302 ] in order to work at high nozzle pressures while keeping the gas load low it may be favorable to use smaller apertures , which goes at the expense of beam intensity , though .
the maximum applicable gas pressure , however , is limited by the pumping speed available in the rotor chamber . in our setup , the rotor - chamber pressure increases roughly linearly from @xmath120 mbar to @xmath121 mbar , as the ar pressure at the nozzle rises from 100 mbar up to 1000 mbar . at these vacuum conditions , collisions of beam molecules with the ones from the residual gas may seriously deplete the beam .
this effect is studied by measuring the peak beam density of kr beams as a function of nozzle pressure up to @xmath122bar at different rotation frequencies , @xmath123hz ( @xmath124 m/s ) , @xmath125hz ( @xmath126 m/s ) , @xmath127hz ( @xmath128 m/s ) . as shown in fig .
[ fig : kratt ] , the kr beam density is affected by rest gas collisions as the nozzle pressure exceeds about 600 mbar .
nozzle pressures higher than about 900 mbar even lead to the destruction of the beam and beam densities drop down again , irrespective of the rotor frequency . only at high frequencies leading to beam velocities below about @xmath129
m/s the position of the maximum of the dependence @xmath130 shifts down to lower values of @xmath55 due to the fact that slow molecules spend more time inside the region of high rest - gas pressure and are therefore more sensitive to rest - gas collisions .
the solid lines in fig .
[ fig : kratt ] are fit curves according to the model of the peak density ( eq . ( [ eq : densitytime ] ) ) , taking into account a density reduction due to collisions , @xmath131 . here
, the relative velocity between colliding atoms is taken as @xmath132 , @xmath133 cm is the average flight distance from nozzle to skimmer , and @xmath134 and @xmath135 are the residual gas density and mean thermal velocity , respectively .
the kr - kr collision cross section @xmath136 is varied as a free fit parameter , yielding an average value @xmath137 .
this value is in reasonable agreement with the landau - lifshitz approximation for the maximum scattering cross section , @xmath138 , with @xmath139ev@xmath140 for kr @xcite .
comparison of beam densities of the continuous nozzle beam ( rotor at rest ) and the effusive beam produced by the residual gas in the rotor chamber .
circles depict the data measured using the quadrupole mass spectrometer scaled to densities obtained from the detector chamber pressure measured using a cold cathode pressure gauge .
squares represent the density values obtained from the pitot tube measurement .
, width=340 ] high nozzle pressures , which lead to increased rest - gas pressure in the nozzle chamber , cause another unwanted effect to occur the formation of a continuous effusive beam .
the latter adds up to the pulsed nozzle beam and may have comparable beam intensities as the nozzle beam is decelerated to low speeds and low densities .
we therefore compare the densities of nozzle beams and effusive beams by recording the detector signal with the nozzle standing properly aligned in front of the skimmer , on the one hand , and with the rotor being tilted completely out of the beam axis , on the other .
the resulting density values are plotted in fig .
[ fig : effusive ] .
while the measurement using the pitot tube yields higher nozzle beam densities by a factor of 10 , equal densities are found for the effusive beam .
thus , the nozzle beam is higher in density by 3 - 4 orders of magnitude than the effusive beam . for decelerated beams of ar and kr ,
however , equal densities of nozzle and effusive beams are reached at beam velocities @xmath141 and @xmath142 m/s , respectively . provided the expansion is operated in the supersonic regime , the ratio of nozzle and effusive beam densities is independent of the nozzle pressure @xmath55 .
this ratio can be improved either by increasing the pumping speed in the nozzle chamber , which is characteristic for such an arrangement based on a continuous gas expansion .
the effusive background can be further suppressed by implementing a mechanical velocity selector synchronized with the rotating nozzle .
therefore , in the near future a chopper wheel will be placed behind the skimmer .
typical chf@xmath143 density distributions as a function of flight time for two different speeds of rotation of the nozzle .
the data recorded at 250hz are scaled up by a factor of 400 .
signals shown as back lines are recorded with no voltage applied to the guide electrodes , light red lines are obtained at 3.5kv electrode voltage.,width=340 ] the issue of transverse divergence of decelerated beams can be relieved for beams of polar molecules by implementing additional guiding elements using inhomogeneous electrostatic fields .
this technique has been used in the past for focussing and even state selecting molecular beams for reactions dynamics experiments @xcite .
more recently , the technique has been used for guiding and trapping of cold polar molecules produced by stark deceleration or stark filtering @xcite . in our setup
we have implemented an electrostatic quadrupole guide between the skimmer and the ionizer of the qms .
it consists of four 259 mm long stainless steel rods 2 mm in diameter with a gap of 2 mm between two diagonally opposing rods . in a first experiment ,
slow beams of fluoroform ( chf@xmath143 ) molecules have been produced by expanding chf@xmath143 either as a neat gas or seeded in kr out of the rotating nozzle .
chf@xmath143 is a polar molecule with a permanent dipole moment of 1.65debye . since its mass ( @xmath144amu ) lies close to the mass of kr
( @xmath145amu ) it can be efficiently cooled and decelerated by coexpansion with kr .
when applying a high voltage @xmath146kv to a pair of opposing electrodes the qms detector signal clearly rises , as depicted in fig .
[ fig : enhancementtof ] . in these measurements ,
chf@xmath143 is expanded at a nozzle pressure @xmath147 mbar at rotor frequencies 75 and 250hz , leading to most probable beam velocities @xmath148 and @xmath149 m/s , respectively .
clearly , the enhancement of the qms signal at an electrode voltage @xmath146kv in proportion to the signal with no high voltage is higher at low beam velocities , reaching roughly a factor of 5 for @xmath149 m/s .
great care was taken to suppress any direct influence of the electrode voltage on the ionization efficiency of the qms ionizer .
similar experiments have also been performed with deuterated ammonia ( nd@xmath143 ) .
in addition to increasing the beam intensity , the quadrupole fields act as quantum state selectors which filter out those states that are attracted towards the beam axis where the electric field amplitude is minimal ( low - field seeking states ) . provided
the rotational temperature of the nd@xmath143 molecules is comparable to the translational one , _
i.e. _ in the range of a few kelvin , we may expect to produce nd@xmath143 molecules in the state @xmath150 , @xmath151 with very high purity @xcite .
relative efficiency of guiding chf@xmath143 molecules at various beam velocities as a function of electrode voltage.,width=340 ] interestingly , the dependence of the enhancement of peak density on the electrode voltage @xmath152 features a non - monotonic increase , as shown in fig .
[ fig : enhancementvoltage ] .
a first local signal maximum is observed at electrode voltages between 200 and 500v , depending on beam velocity .
we interpret this behavior in terms of two competing mechanisms being active : guiding of molecules in low - field seeking states enhances the detector signal whereas deviation of high - field seeking molecules away from the beam axis reduces the measured beam density . at low voltages @xmath153v , the guiding effect
outreaches the deviation efficiency , leading to a local density maximum . as the electrode voltage rises ,
all high - field seeking molecules are expelled out of the beam and only low - field seekers are guided to the detector .
this guiding effect partly compensates the transverse blowing up of the beam , which is most relevant for low longitudinal beam velocities .
consequently , the relative guiding efficiency increases up to a factor of 5 as the beam velocity is reduced down to @xmath149 m/s .
the fact that the same qualitative behavior is observed with beams of nd@xmath143 supports our interpretation , which does not consider specific stark shifts of the molecules . in future efforts
the transmission of our setup will be further studied both experimentally and using molecular trajectory simulations . besides
, guiding of decelerated molecules around bent electrodes will be implemented to provide state - purified beams of cold molecules .
this has the additional advantage of efficiently suppressing the hot effusive background .
enhancement of the chf@xmath143 peak density as a function of beam velocity for fixed electrode voltage @xmath154kv.,width=340 ] the dependence of the enhancement factor on the beam velocity is illustrated in fig .
[ fig : enhancementrot ] . in this measurement ,
the electrode voltage is set to the maximum value @xmath146kv and the beam velocity is tuned from @xmath155 down to @xmath156 m/s by changing the frequency of rotation of the nozzle .
further deceleration down to @xmath157 m/s is obtained by seeded expansion of chf@xmath143 in kr at a mixing ratio 1:2 .
the enhancement factor continuously rises as the beam velocity is reduced , which highlights the potential of combining decelerated nozzle beams with electrostatic guiding fields .
in conclusion , we have presented a versatile apparatus that produces cold beams of accelerated or decelerated molecules based on a rapidly rotating nozzle .
various technical improvements with respect to the original demonstration by gupta and herschbach @xcite are introduced . in particular , gas injection into the rotor is now realized using a fluid metal - sealed rotary feedthrough which eliminates gas leakage into the vacuum chamber . using this setup , beam velocities well below @xmath129 m/s and longitudinal beam temperatures down to about
1k are achieved .
a detailed analytic model to simulate the measured density distributions as functions of flight time is presented .
the fundamental drawback of this technique , the sharp drop of beam intensity at slow beam velocities , is relieved by combining the setup with electrostatic guiding elements , provided the molecules feature a suitable stark effect .
moreover , stark guiding goes along with internal - state selectivity .
thus , a significant increase of beam intensity as well as an expected high state - purity is achieved with beams of fluoroform and ammonia molecules .
thanks to the technical improvements the rotating nozzle setup has evolved into a simple and reliable source of cold molecular beams with tunable velocity . in particular ,
slow beams of molecules that are not amenable to efficient filtering or deceleration by means of the stark interaction with external electric fields can now be produced with high intensity . in the near future , we will utilize this beam source for studying reactive collisions between slow molecules and cold atoms produced by other means , _
e.g. _ using a magneto - optical trap . at collision energies down to @xmath158mev
, interesting new quantum effects may be expected in the reaction dynamics , _
e.g. _ strong modulations of the scattering cross - section in reactions of the type li+hf@xmath159lif+h @xcite . furthermore
, intense cold molecular beams with tunable velocity and internal - state purity are of current interest for surface scattering experiments @xcite .
we thank h.j .
loesch for ceding to us the main part of the experimental setup .
generous advice by d. weber , k. zrost and m. dekieviet as well as assistance in setting up the apparatus by j. humburg , ch .
glck and o. bnermann is gratefully acknowledged .
we are grateful for support by the landesstiftung baden - wrttemberg as well as by dfg . | intense beams of cold and slow molecules are produced by supersonic expansion out of a rapidly rotating nozzle , as first demonstrated by gupta and herschbach @xcite .
an improved setup is presented that allows to accelerate or decelerate cold atomic and molecular beams by up to 500 m/s .
technical improvements are discussed and beam parameters are characterized by detailed analysis of time of flight density distributions . the possibility of combining this beam source with electrostatic fields for guiding polar molecules is demonstrated . |
one of the most intriguing features of qcd for massless quarks is its underlying conformal invariance , invariance under both scale ( dilatation ) and special conformal transformations @xcite .
for example , in the case of perturbative qcd , the running coupling @xmath0 becomes constant in the limit of zero @xmath1-function and zero quark mass , and conformal symmetry becomes manifest .
in fact , the renormalization scale uncertainty in pqcd predictions can be eliminated by using the principle of maximum conformality ( pmc ) @xcite . using the pmc / blm procedure @xcite , all non -
conformal contributions in the perturbative expansion series are summed into the running coupling by shifting the renormalization scale in @xmath2 from its initial value , and one obtains unique , scale - fixed , scheme - independent predictions at any finite order .
one can also introduce a generalization of conventional dimensional regularization which illuminates the renormalization scheme and scale ambiguities of pqcd predictions , exposes the general pattern of nonconformal terms , and allows one to systematically determine the argument of the running coupling order by order in pqcd in a form which can be readily automatized @xcite .
the resulting pmc scales and finite - order pmc predictions are both to high accuracy independent of the choice of initial renormalization scale . for example
, pmc scale - setting leads to a scheme - independent pqcd prediction @xcite for the top - quark forward - backward asymmetry which is within one @xmath3 of the tevatron measurements .
the pmc procedure also provides scale - fixed , scheme - independent commensurate scale relations @xcite , relations between observables which are based on the underlying conformal behavior of qcd such as the generalized crewther relation @xcite .
the pmc satisfies all of the principles of the renormalization group : reflectivity , symmetry , and transitivity , and it thus eliminates an unnecessary source of systematic error in pqcd predictions @xcite .
anti - de sitter space in five dimensions ( ads@xmath4 ) provides a geometric representation of the conformal group .
one can modify ads space by using a dilaton factor in the ads metric @xmath5 to introduce the qcd confinement scale .
although such a mass scale explicitly breaks the dilatation invariance of the equations of motion , the action can still be conformally invariant , as was first shown by v. de alfaro , s. fubini and g. furlan ( daff ) @xcite in the context of one - dimensional quantum field theory ; i.e , the change in the mass scale of the potential can be compensated by the time scale in the action . however , this manifestation of conformal symmetry can only occur if the dilaton profile @xmath6 is constrained to have the specific power @xmath7 , a remarkable result which follows from the daff construction of conformally invariant quantum mechanics @xcite .
the quadratic form @xmath8 leads to linear regge trajectories @xcite in the hadron mass squared .
a remarkable holographic feature of dynamics in ads space in five dimensions is that it is dual to hamiltonian theory in physical space - time , quantized at fixed light - front ( lf ) time @xcite .
for example , the equation of motion for mesons on the light - front has exactly the same single - variable form as the ads / qcd equation of motion ; one can then interpret the ads fifth dimension variable @xmath9 in terms of the physical variable @xmath10 , representing the invariant separation of the @xmath11 and @xmath12 at fixed light - front time .
as discussed in the next sections , this light - front holographic principle provides a precise relation between the bound - state amplitudes in ads space and the boost - invariant lf wavefunctions describing the internal structure of hadrons in physical space - time .
the resulting valence fock - state wavefunction eigensolutions of the lf qcd hamiltonian satisfy a single - variable relativistic equation of motion analogous to the nonrelativistic radial schrdinger equation . the quadratic dependence in the quark - antiquark potential @xmath13 in the light - front schrdinger equation ( lfse )
is determined uniquely from conformal invariance , whereas the constant term @xmath14 is fixed by the duality between ads and lf quantization , a correspondence which follows specifically from the separation of kinematics and dynamics on the light - front @xcite .
the lf potential thus has a specific power dependence in effect , it is a light - front harmonic oscillator potential .
it is confining and reproduces the observed linear regge behavior of the light - quark hadron spectrum in both the orbital angular momentum @xmath15 and the radial node number @xmath16 .
the pion is predicted to be massless in the chiral limit @xcite - the positive contributions to @xmath17 from the lf potential and kinetic energy is cancelled by the constant term in @xmath18 for @xmath19 the derived running qcd coupling displays an infrared fixed point @xcite .
the construction of daff retains conformal invariance of the action despite the presence of a fundamental mass scale .
the ads approach , however , goes beyond the purely group - theoretical considerations of daff , since features such as the masslessness of the pion and the separate dependence on @xmath20 and @xmath15 are consequences of the potential derived from the holographic lf duality with ads for general @xmath20 and @xmath15 @xcite .
the quantization of qcd at fixed light - front time @xcite ( dirac s front form ) provides a first - principles hamiltonian method for solving nonperturbative qcd .
it is rigorous , has no fermion - doubling , is formulated in minkowski space , and it is frame - independent . given the boost - invariant light - front wavefunctions @xmath21 ( lfwfs ) , one can compute a large range of hadron observables , starting with structure functions , generalized parton distributions , and form factors .
it is also possible to compute jet hadronization at the amplitude level from first principles from the lfwfs @xcite .
a similar method has been used to predict the production of antihydrogen from the off - shell coalescence of relativistic antiprotons and positrons @xcite .
the lfwfs of hadrons thus provide a direct connection between observables and the qcd lagrangian .
solving nonperturbative qcd is thus equivalent to solving the light - front heisenberg matrix eigenvalue problem .
angular momentum @xmath22 is conserved at every vertex .
the lf vacuum is defined as the state of lowest invariant mass and is trivial up to zero modes .
there are thus no quark or gluon vacuum condensates in the lf vacuum the corresponding physics is contained within the lfwfs themselves @xcite , thus eliminating a major contribution to the cosmological constant .
the simplicity of the front form contrasts with the usual instant - form formalism .
current matrix elements defined at ordinary time @xmath23 must include the coupling of photons and vector bosons fields to connected vacuum - induced currents ; otherwise , the result is not lorentz - invariant .
thus the knowledge of the hadronic eigensolutions of the instant - form hamiltonian are insufficient for determining form factors or other observables .
in addition , the boost of an instant form wavefunction from @xmath24 to @xmath25 changes particle number and is an extraordinarily complicated dynamical problem .
it is remarkable fact that ads / qcd , which was originally motivated by the ads / cft correspondence between gravity on a higher - dimensional space and conformal field theories in physical space - time @xcite , has a direct holographic mapping to light - front hamiltonian theory @xcite .
the ads mass parameter @xmath26 maps to the lf orbital angular momentum .
the formulae for electromagnetic @xcite and gravitational @xcite form factors in ads space map to the exact drell - yan - west formulae in light - front qcd @xcite .
thus the light - front holographic approach provides an analytic frame - independent first approximation to the color - confining dynamics , spectroscopy , and excitation spectra of the relativistic light - quark bound states of qcd .
it is systematically improvable in full qcd using the basis light - front quantization ( blfq ) method @xcite and other methods .
in the limit of zero quark masses the longitudinal modes decouple from the invariant lf hamiltonian equation @xmath27 with @xmath28 .
the generators @xmath29 , @xmath30 , are constructed canonically from the qcd lagrangian by quantizing the system on the light - front at fixed lf time @xmath31 , @xmath32 @xcite .
the lf hamiltonian @xmath33 generates the lf time evolution with respect to @xmath31 , whereas the lf longitudinal @xmath34 and transverse momentum @xmath35 are kinematical generators .
it is advantageous to reduce the full multiparticle eigenvalue problem of the lf hamiltonian to an effective light - front schrdinger equation which acts on the valence sector lf wavefunction and determines each eigensolution separately @xcite .
in contrast , diagonalizing the lf hamiltonian yields all eigensolutions simultaneously , a complex task . the central problem for deriving the lfse becomes the derivation of the effective interaction @xmath36 which acts only on the valence sector of the theory and has , by definition , the same eigenvalue spectrum as the initial hamiltonian problem . in order to carry out this program
one must systematically express the higher fock components as functionals of the lower ones .
this method has the advantage that the fock space is not truncated , and the symmetries of the lagrangian are preserved @xcite .
the light - front hamiltonian for qcd can be derived directly from the qcd lagrangian @xcite .
the result is relativistic and frame - independent .
the @xmath37 lf fock state wavefunction for a meson can be written as @xmath38 thus factoring the angular dependence @xmath39 and the longitudinal , @xmath40 , and transverse mode @xmath41 . in the limit of zero quark
masses the longitudinal mode decouples and the lf eigenvalue equation @xmath42 takes the form of a light - front wave equation for @xmath43 @xcite @xmath44 \phi_{j , l , n}(\zeta^2 ) = m^2 \phi_{j , l , n}(\zeta^2),\ ] ] a relativistic _ single - variable _ lf schrdinger equation .
this equation describes the spectrum of mesons as a function of @xmath16 , the number of nodes in @xmath10 , the total angular momentum @xmath20 , which represent the maximum value of @xmath45 , @xmath46 , and the internal orbital angular momentum of the constituents @xmath47 .
the variable @xmath9 of ads space is identified with the lf boost - invariant transverse - impact variable @xmath10 @xcite , thus giving the holographic variable a precise definition in lf qcd @xcite .
for a two - parton bound state @xmath48 , where @xmath49 is the longitudinal momentum fraction and @xmath50 is the transverse - impact distance between the quark and antiquark . in the exact qcd theory @xmath36 is related to the two - particle irreducible @xmath37 green s function .
the potential in the lfse is determined from the two - particle irreducible ( 2pi ) @xmath51 greens function .
in particular , the higher fock states in intermediate states leads to an effective interaction @xmath52 for the valence @xmath53 fock state @xcite .
a related approach for determining the valence light - front wavefunction and studying the effects of higher fock states without truncation has been given in ref .
@xcite . unlike ordinary instant - time quantization ,
the light - front hamiltonian equations of motion are frame independent ; remarkably , they have a structure which matches exactly the eigenmode equations in ads space .
this makes a direct connection of qcd with ads methods possible .
in fact , one can derive the light - front holographic duality of ads by starting from the light - front hamiltonian equations of motion for a relativistic bound - state system in physical space - time @xcite .
recently we have derived wave equations for hadrons with arbitrary spin starting from an effective action in ads space @xcite .
an essential element is the mapping of the higher - dimensional equations to the lf hamiltonian equation found in ref .
this procedure allows a clear distinction between the kinematical and dynamical aspects of the lf holographic approach to hadron physics .
accordingly , the non - trivial geometry of pure ads space encodes the kinematics , and the additional deformations of ads encode the dynamics , including confinement @xcite . a spin-@xmath20 field in ads@xmath54
is represented by a rank @xmath20 tensor field @xmath55 , which is totally symmetric in all its indices . in presence of a dilaton background field
@xmath56 the effective action is @xcite @xmath57 where the indices @xmath58 , @xmath59 and @xmath60 is the covariant derivative which includes parallel transport .
the coordinates of ads are the minkowski coordinates @xmath61 and the holographic variable @xmath9 , @xmath62 .
the effective mass @xmath63 , which encodes kinematical aspects of the problem , is an _ a priori _ unknown function , but the additional symmetry breaking due to its @xmath9-dependence allows a clear separation of kinematical and dynamical effects @xcite . the dilaton background field @xmath56 in ( [ seff ] ) introduces an energy scale in the five - dimensional ads action , thus breaking conformal invariance .
it vanishes in the conformal ultraviolet limit @xmath64 .
a physical hadron has plane - wave solutions and polarization indices along the 3 + 1 physical coordinates @xmath65 , with four - momentum @xmath66 and invariant hadronic mass @xmath67 .
all other components vanish identically .
the wave equations for hadronic modes follow from the euler - lagrange equation for tensors orthogonal to the holographic coordinate @xmath9 , @xmath68 .
terms in the action which are linear in tensor fields , with one or more indices along the holographic direction , @xmath69 , give us the kinematical constraints required to eliminate the lower - spin states @xcite . upon variation with respect to @xmath70
, we find the equation of motion @xcite @xmath71 \phi_j = m^2 \phi_j,\ ] ] with @xmath72 , which is the result found in refs .
@xcite by rescaling the wave equation for a scalar field .
similar results were found in ref .
@xcite . upon variation with respect to @xmath73
we find the kinematical constraints which eliminate lower spin states from the symmetric field tensor @xcite @xmath74 upon the substitution of the holographic variable @xmath9 by the lf invariant variable @xmath10 and replacing @xmath75 in ( [ phijm ] ) , we find for @xmath76 the lfse ( [ lfwe ] ) with effective potential @xcite @xmath77 provided that the ads mass @xmath78 in ( [ phijm ] ) is related to the internal orbital angular momentum @xmath79 and the total angular momentum @xmath80 according to @xmath81 .
the critical value @xmath82 corresponds to the lowest possible stable solution , the ground state of the lf hamiltonian .
for @xmath83 the five dimensional mass @xmath78 is related to the orbital momentum of the hadronic bound state by @xmath84 and thus @xmath85 .
the quantum mechanical stability condition @xmath86 is thus equivalent to the breitenlohner - freedman stability bound in ads @xcite .
a particularly interesting example is a dilaton profile @xmath87 of either sign , since it leads to linear regge trajectories @xcite and avoids the ambiguities in the choice of boundary conditions at the infrared wall . for the confining solution @xmath88
the effective potential is @xmath89 and eq .
( [ lfwe ] ) has eigenvalues @xmath90 , with a string regge form @xmath91 . a discussion of the light meson and baryon spectrum , as well as the elastic and transition form factors of the light hadrons using lf holographic methods , is given in ref .
@xcite . as an example
the spectral predictions for the @xmath92 light pseudoscalar and vector meson states are compared with experimental data in fig .
[ pionspec ] for the positive sign dilaton model .
parent and daughter regge trajectories for the @xmath93-meson family ( left ) with @xmath94 gev ; and the @xmath95-meson family ( right ) with @xmath96 gev.,title="fig:",width=232 ] parent and daughter regge trajectories for the @xmath93-meson family ( left ) with @xmath94 gev ; and the @xmath95-meson family ( right ) with @xmath96 gev.,title="fig:",width=232 ] the effective interaction @xmath97 is instantaneous in lf time and acts on the lowest state of the lf hamiltonian .
this equation describes the spectrum of mesons as a function of @xmath16 , the number of nodes in @xmath98 , the internal orbital angular momentum @xmath99 , and the total angular momentum @xmath100 , with @xmath80 the sum of the orbital angular momentum of the constituents and their internal spins .
[ the @xmath101 casimir @xmath102 corresponds to the group of rotations in the transverse lf plane . ]
it is the relativistic frame - independent front - form analog of the non - relativistic radial schrdinger equation for muonium and other hydrogenic atoms in presence of an instantaneous coulomb potential .
the ads / qcd harmonic oscillator potential could in fact emerge from the exact qcd formulation when one includes contributions from the lfse potential @xmath36 which are due to the exchange of two connected gluons ; _ i.e. _ , `` h '' diagrams @xcite .
we notice that @xmath36 becomes complex for an excited state since a denominator can vanish ; this gives a complex eigenvalue and the decay width .
the correspondence between the lf and ads equations thus determines the effective confining interaction @xmath36 in terms of the infrared behavior of ads space and gives the holographic variable @xmath9 a kinematical interpretation .
the identification of the orbital angular momentum is also a key element of our description of the internal structure of hadrons using holographic principles .
if one starts with a dilaton profile @xmath103 with @xmath104 the existence of a massless pion in the limit of massless quarks determines uniquely the value @xmath7 . to show this , one can use the stationarity of bound - state energies with respect to variation of parameters @xcite .
the quadratic dilaton profile also follows from the algebraic construction of hamiltonian operators by ( daff ) @xcite .
the action @xmath105 is invariant under conformal transformations in the variable @xmath23 , and there are in addition to the hamiltonian @xmath106 two more invariants of motion for this field theory , namely the dilation operator @xmath107 and @xmath108 , corresponding to the special conformal transformations in @xmath23 .
specifically , if one introduces the the new variable @xmath109 defined through @xmath110 and the rescaled fields @xmath111 , it then follows that the the operator @xmath112 generates evolution in @xmath109 @xcite .
the hamiltonian corresponding to the operator @xmath113 which introduces the mass scale is a linear combination of the old hamiltonian @xmath106 , @xmath107 , the generator of dilations , and @xmath108 , the generator of special conformal transformations .
it contains the confining potential @xmath114 , that is the confining term in ( [ u ] ) for a quadratic dilaton profile and thus @xmath115 .
the construction of new hamiltonians from the generators of the conformal group has been used by daff to construct algebraically the spectra and the eigenfunctions of these operators .
the conformal group in one dimension is locally isomorphic to the group of pseudo - rotations @xmath116 .
the compact operator @xmath117 generates rotations in the euclidean 1 - 2 plane , whereas the non - compact operators @xmath118 and @xmath119 generate pseudo - rotations ( boosts ) in the non - euclidean 2 - 3 and 1 - 3 plane respectively . as in the familiar case of angular momentum , one can introduce raising and lowering operators @xmath120 and construct the spectrum and the eigenfunctions analogously to the angular momentum operators .
interestingly , this is just the method employed in ref .
@xcite to obtain an integrable lf confining hamiltonian by following infeld s observation that integrability follows immediately if the equation of motion can be factorized as a product of linear operators @xcite .
the method can be extended to describe baryons in ads while preserving the algebraic structure @xcite .
our approach has elements in common with those of ref .
@xcite , where the scale of the confinement potential arises from a boundary condition when solving gauss equation .
the triple complementary connection of ( a ) ads space , ( b ) its lf holographic dual , and ( c ) the relation to the algebra of the conformal group in one dimension , is characterized by a quadratic confinement lf potential , and thus a dilaton profile with the power @xmath121 , with the unique power @xmath7 .
in fact , for @xmath122 the mass of the @xmath123 pion is automatically zero in the chiral limit , and the separate dependence on @xmath20 and @xmath15 leads to a mass ratio of the @xmath95 and the @xmath124 mesons which coincides with the result of the weinberg sum rules @xcite .
one predicts linear regge trajectories with the same slope in the relative orbital angular momentum @xmath15 and the principal quantum humber @xmath16 .
the ads approach , however , goes beyond the purely group theoretical considerations of daff , since features such as the masslessness of the pion and the separate dependence on @xmath20 and @xmath15 are a consequence of the potential ( [ u ] ) derived from the duality with ads for general @xmath20 and @xmath15 .
the constant term in the potential , which is not determined by the group theoretical arguments , is fixed by the holographic duality to lf quantized qcd @xcite .
the resulting lagrangian , constrained by the conformal invariance of the action , has the same form as the ads lagrangian with a quadratic dilaton profile . in their discussion of the evolution operator @xmath125 as a model for confinement ,
daff mention a critical point , namely that `` the time evolution is quite different from a stationary one '' . by this statement
they refer to the fact that the variable @xmath109 is related to the variable @xmath23 for the case @xmath126 by @xmath127 _
i.e. _ , @xmath109 has only a limited range .
the finite range of invariant lf time @xmath128 can be interpreted as a feature of the internal frame - independent lf time difference between the confined constituents in a bound state .
for example , in the collision of two mesons , it would allow us to compute the lf time difference between the two possible quark - quark collisions .
the treatment of the chiral limit in the lf holographic approach to strongly coupled qcd is substantially different from the standard approach based on chiral perturbation theory . in the conventional approach
@xcite , spontaneous symmetry breaking by a non - vanishing chiral quark condensate @xmath129 plays the crucial role . in qcd sum rules
@xcite @xmath130 brings in non - perturbative elements into the perturbatively calculated spectral sum rules .
it should be noted , however , that the definition of the condensate , even in lattice qcd necessitates a renormalization procedure for the operator product , and it is not a directly observable quantity . in bethe - salpeter
@xcite and light - front analyses @xcite , the gell mann - oakes - renner relation @xcite for @xmath131 involves the decay matrix element @xmath132 instead of @xmath133 . in the color - confining ads / qcd light - front model discussed here ,
the vanishing of the pion mass in the chiral limit , a phenomenon usually ascribed to spontaneous symmetry breaking of the chiral symmetry , is obtained specifically from the precise cancellation of the lf kinetic energy and lf potential energy terms for the quadratic confinement potential .
this mechanism provides a viable alternative to the conventional description of nonperturbative qcd based on vacuum condensates , and it eliminates a major conflict of hadron physics with the empirical value for the cosmological constant @xcite .
invited talk , presented by sjb at the third workshop on the qcd structure of the nucleon ( qcd - n12 ) , bilbao , spain , october 22 - 26 , 2012 .
this work was supported by the department of energy contract de | we show that ( a ) the conformal properties of anti - de sitter ( ads ) space , ( b ) the properties of a field theory in one dimension under the full conformal group found by de alfaro , fubini and furlan , and ( c ) the frame - independent single - variable light - front schrdinger equation for bound states all lead to the same result : a relativistic nonperturbative model which successfully incorporates salient features of hadronic physics , including confinement , linear regge trajectories , and results which are conventionally attributed to spontaneous chiral symmetry breaking . |
relaying is receiving attention for wireless cellular applications because it improves rates and reliabilities .
there are two simple geometric scenarios that give insight into relaying strategies , and that show how relaying achieves distributed multi - input , multi - output ( mimo ) gains @xcite .
first , relays that are close to a source node can achieve multi - input " gains by using a decode - forward ( df ) strategy .
second , relays that are close to a destination node can achieve multi - output " gains by using a compress - forward ( cf ) strategy .
both the df and cf strategies appeared for abstract channels in the work of cover and el gamal @xcite .
this document focuses on the cf strategy whose usefulness for network communication has been demonstrated . , e.g. , in @xcite and follow - up works .
recently , a method called _ noisy network coding _ was developed @xcite that is a quantize - forward ( qf ) variant of the cf strategy .
the qf strategy uses simple relays and achieves a remarkably simple - to - describe rate region that is sometimes close to a cut - set upper bound .
our main goal is to outline why the short - message cf method of ( * ? ? ?
v ) can achieve better rates and reliabilities with less complexity than the long - message qf methods of @xcite .
this extends results by wu and xie @xcite who found that short - message encoding achieves the same rates as long - message encoding for a single source .
finally , we describe a hashing scheme that combines some of the advantages of long- and short - message encoding .
we first address terminology . variations of the compression strategy of cover and el gamal @xcite are known by names such as estimate - forward " ( ef ) , compress - forward " ( cf ) , quantize - forward " ( qf ) , quantize - map " , hash - forward " ( hf ) , and so forth .
we make the following observations . *
the word compress " is a generic name that refers to both lossless and lossy source coding , the latter including quantization " and hashing " ( or binning " ) . * without hashing one obtains a qf strategy @xcite . * without quantization one obtains a hf strategy @xcite . *
the name of a relay function should not depend on the choice of operations at other nodes .
in particular , it should not depend on whether other nodes perform optimal or suboptimal processing .
the last bullet point sometimes causes confusion .
some literature takes cf to mean that the next " relay along a route _ must decode _ certain indices , perhaps even with a suboptimal decoder .
however , if one accepts this bullet point , such terminology makes little sense .
we therefore advocate to use the ( generic ) name cf for the general strategy , hf for a strategy without quantization , and qf for a strategy without hashing .
of course , this makes hf and qf ( and noisy network coding ) special cases of cf .
we review the recent qf strategy , see @xcite .
however , rather than using long - message repetition encoding we use short "- message encoding ( see ( * ? ? ?
6 ) , ( * ? ? ?
v ) , @xcite ) and pipelined decoding via a sliding - window method ( see ( * ? ? ?
i.a ) , @xcite , @xcite , @xcite , ) . as usual
, we use independent random codebooks for each block ( see @xcite and @xcite ) .
we use the notation @xmath0 and @xmath1 for @xmath2-typical sets . _ code construction _ : encoding is performed in @xmath3 blocks , and we generate a different code book for each block ( see [ fig : relay1-compressforward ] where @xmath4 ) . for block @xmath5 , @xmath6 , generate @xmath7 codewords @xmath8 , @xmath9 , by choosing the symbols @xmath10 , @xmath11 , independently using @xmath12 .
similarly , generate @xmath13 codewords @xmath14 , @xmath15 , by choosing the @xmath16 independently using @xmath17 .
finally , introduce an auxiliary random variable @xmath18 that represents a quantized version of @xmath19 , and consider a distribution @xmath20 . for each @xmath14 ,
generate @xmath13 codewords @xmath21 , @xmath22 , by choosing the @xmath23 independently using @xmath24 .
_ source _ : the message @xmath25 of @xmath26 bits is split into @xmath27 blocks @xmath28 of @xmath7 bits each . in block
@xmath5 , @xmath6 , the source transmits @xmath29 , where @xmath30 .
_ relay _ : in block @xmath31 , the relay transmits @xmath32 .
after block @xmath5 , the relay has seen @xmath33 .
the relay tries to find a @xmath34 such that @xmath35 if one or more such @xmath34 are found , then the relay chooses one of them , sets @xmath36 , and transmits @xmath37 . if no such pair is found , the relay sets @xmath38 and transmits @xmath39
. _ sink terminal _ :
after block @xmath5 , @xmath40 , the receiver has seen @xmath41 and @xmath42 , and tries to find a pair @xmath43 such that @xmath44 and we assume that @xmath45 . if one or more such @xmath43 are found , then the sink chooses one of them , and puts out this choice as @xmath46 .
if no such @xmath43 is found , the sink puts out @xmath47 .
_ analysis _ : the analysis follows familiar steps , see ( * ? ? ?
15.2 ) and we summarize the results . 1 .
the relay quantization requires @xmath48 2 .
the sink s decoder can be viewed as a multi - access channel ( mac ) decoder for two messages @xmath49 and @xmath50 and therefore we have the bounds @xmath51 observe that we _ can not _ ignore the bound , as might be expected , because we require that @xmath45 in .
the sums in and are due to the intersection of independent events and . the joint distribution of the random variables factors as @xmath52 for all @xmath53 .
performing a fourier - motzkin elimination of @xmath54 , and manipulating the mutual information expressions , the bounds - become @xmath55 but suppose that is not satisfied so that and give @xmath56 which is a stronger bound than .
the rates satisfying are achievable with qf , e.g. , by choosing @xmath18 independent of @xmath57 and @xmath19 ( and thus @xmath58 also ) .
hence we may ignore the constraint .
the resulting qf rates are as close as desired to the known cf rate @xmath59 \label{eq : cfrate}\end{aligned}\ ] ] where the maximization is over all distributions factoring as in .
the rate is the same as the more commonly used expression ( see ( * ? ? ?
* thm . 3 and eq . ( 6 ) ) )
one may decode short - message encoded packets by using backward decoding ( see , e.g. , @xcite ) as long as the final transmission block is sufficiently long to be able to decode @xmath61 .
the bound is replaced with the weaker constraint @xmath62 which means that is replaced with the weaker constraint @xmath63 we thus find that backward decoding should outperform pipelined decoding for slow - fading channels with outage .
another possibility is for the receiver to jointly decode all indices after transmission is completed .
yet another possibility is to use a pipelined decoder with a longer and variable window length , either in the forward or backward directions .
for example , the window length may be @xmath5 for block @xmath5 .
one advantage of long - message encoding is that the quantization constraint disappears ( which means that disappears ) .
thus , long - message encoding should outperform short - message encoding for slow - fading channels with outage . as mentioned in the introduction , relaying achieves distributed mimo gains if relays close to a source use df and relays far from a source use cf / qf @xcite .
unfortunately , long - message encoding inhibits df because the message is usually too long to decode after receiving one block of channel outputs .
in contrast , short - message encoding lets relays close to a source decode messages early .
these relays can form a distributed transmit array with this source .
a subtle advantage of short messages is that one can map a small number of bits onto the modulation , i.e. , the modulation set can be kept small .
when using long - message encoding , one would either have to use a ( very ) large modulation set , or one must first hash the long message to a shorter message ( see below ) . either way , long - message decoding is complex and will suffer implementation and synchronization losses if a large modulation alphabet is used .
short - message encoding has a considerably - reduced _ encoding _ delay as compared to long - message encoding .
similarly , pipelined decoding enjoys a considerably - reduced _ decoding _ delay .
the combination of these two approaches might support streaming applications .
we may combine the reliability advantage of long - message encoding with the early decoding and modulation complexity advantages of short messages by hashing the long message to many short messages before transmission .
relays may now decode short hash messages if they can , thereby enabling df and distributed mimo .
furthermore , the quantization constraints such as disappear like for long - message encoding .
the price paid is long encoding and decoding delays .
as shown in @xcite , short - message encoding and backward decoding achieves the same bounds on @xmath64 as in @xcite for multiple relays .
this statement is also valid for multiple relays and multiple sources .
the knowledgable reader may wonder why hashing ( or binning ) is not needed to achieve the cf rates , in seeming contradiction to results in , e.g. , @xcite and @xcite .
of course , one obvious explanation is that we are using a better ( joint ) decoder rather than a step - by - step decoder .
however , the model of @xcite deserves closer inspection .
the relay channel in @xcite does not have the standard " form with a memoryless channel @xmath65 ; there is instead a rate constraint @xmath66 on the relay - destination link .
but we can bring such a channel into standard form by introducing a random variable @xmath57 that represents the relay s transmit symbols and choose its alphabet size @xmath67 as @xmath68 ( if @xmath68 is not an integer we may choose a model with memory on the relay - destinaton link and again appropriately limit the size of the input alphabet ) .
now suppose that @xmath69 in which case qf necessarily assigns the same codeword @xmath70 to ( exponentially in @xmath71 ) many indices @xmath50 .
in other words , qf implicitly performs hashing .
the same consideration shows that wyner - ziv coding may be considered to be using qf only ( without a binning step ) .
a similar claim can be made for slepian - wolf coding .
of course , this statement lacks depth since whether we call implicit binning qf or hf is not important .
the reader may now wonder whether qf always performs hashing implicitly .
we emphasize that this is generally _ not _ the case when @xmath72 .
for example , for real - input channels such as gaussian channels we have @xmath73 and qf will generally assign a unique @xmath70 to every index @xmath50 .
for the single - relay channel , short - message qf with pipelined decoding achieves the same rates as long - message qf . for the multi - relay ,
multi - source channel , short - message qf with backward decoding recovers the rates of long - message qf .
several advantages of short - message encoding are pointed out , e.g. , substantial reduction in delay , reduced modulation complexity , and added flexibility in letting relays choose df or cf ( or qf ) .
g. kramer was supported by an alexander von humboldt professorship endowed by the german federal ministry of education and research and by nsf grant ccf-09 - 05235 .
jie hou was supported by newcom++ . | recent work for single - relay channels shows that quantize - forward ( qf ) with long - message encoding achieves the same reliable rates as compress - forward ( cf ) with short - message encoding .
it is shown that short - message qf with backward or pipelined ( sliding - window ) decoding also achieves the same rates .
similarly , for many relays and sources , short - message qf with backward decoding achieves the same rates as long - message qf .
several practical advantages of short - message encoding are pointed out , e.g. , reduced delay and simpler modulation .
furthermore , short - message encoding lets relays use decode - forward ( df ) if their channel quality is good , thereby enabling multi - input , multi - output ( mimo ) gains that are not possible with long - message encoding . finally , one may combine the advantages of long- and short - message encoding by hashing a long message to short messages . |
gamma - ray bursts ( grbs ) have been proposed as a potential origin of ultra - high - energy cosmic rays ( uhecrs ) ( e.g. , waxman 1995 ; vietri 1995 ; milgrom and usov 1995 ) .
high - energy neutrinos are thought to be a useful messenger to probe cr acceleration in grbs , as they are predicted to be produced in the dissipative fireballs , where cosmic - ray protons interact with fireball gamma - ray photons through photopion process ( e.g. , waxman & bahcall 1997 ; guetta et al . 2004 ; dermer & atoyan 2006 ; murase et al . 2006 ;
wang & dai 2009 ) .
however , search for high - energy neutrinos in coincident with grbs using icecube has failed to find any associated neutrinos so far ( abbasi et al .
2012 ; aartsen et al . 2015 ) .
the non - detection has put stringent constraints on the neutrino production efficiency and fireball properties of grbs ( he et al .
2012 ; zhang & kumar 2013 ; gao et al . 2013 ; yacobi et al .
2014 ) . because the low neutrino flux could be simply due to a low efficiency in converting crs to neutrinos , as expected in some grb models with large dissipation radius
, one can not rule out cr acceleration in grbs with the non - detection of neutrinos .
crs accelerated by grbs will finally escape from the source and enter the interstellar space of the host galaxy .
the proton - proton collisions between crs and nuclei in the interstellar medium will also produce high - energy neutrinos . recently , icecube collaboration reported the detection of extraterrestrial tev - pev neutrinos with a best - fit flux of @xmath3 per flavor ( aartsen et al .
2014 ) . noting the coincidence between the icecube neutrino flux and the waxman - bahcall bound ( @xmath4 ) derived from the flux of uhecrs above @xmath5
, some authors suggest that icecube neutrinos may be produced by the same source responsible these uhecrs ( waxman 2013 ; katz et al .
as grbs are one candidate source of uhecrs , one may wonder whether neutrinos resulted from the cr collisions with the ism in the host galaxy can explain the icecube observations ( waxman 2013 , wang et al .
2014 ) . in this paper
, we study this interesting possibility by perform a calculation of the expected neutrino flux with the knowledge about the properties of grbs and their host galaxies . one key unknown factor that decides the neutrino flux from grb host galaxies is the pion production efficiency , i.e. the energy loss of cr protons into pions due to collisions with the ism of the host galaxy .
crs interact with the ism along their path before escape , so this efficiency relates with the gas column density that crs traversed . as crs
are transported outward by galactic winds , the gas surface density @xmath6 represents an averaged column density of matter that crs traversed .
however , for most grbs , we do not have measurements of @xmath6 at the grb explosion site as grb hosts are hardly spatially resolved .
we note that the x - ray absorption column density @xmath0 , inferred from the x - ray observations of the prompt and afterglow emission , has a similar role , but @xmath0 reflects the column density along the direction of the line of sight ( rather than the averaged gas column density ) .
moreover , the value of @xmath0 is derived assuming solar metallicity for the absorbing gas , while grbs seem to occur preferentially in low metallicity galaxies ( stanek et al . 2006 ; prieto et al . 2008 ) . the x - ray absorption , as measured with swift / xrt observations , mainly takes place through inner shell electrons of metals , thus it is linked to the metallicity of the ism in the host galaxies .
if grb host galaxies have lower metallicity , the true absorbing gas column density should be higher than the quoted values obtained from fitting x - ray afterglows , considering this metallicity difference .
we thus first study how well @xmath0 can trace the gas surface density @xmath6 in 2 .
we find that @xmath0 is well correlated with the surface star - formation rate ( @xmath7 ) of the host galaxies .
then taking into account the kennicutt - schmidt law that relates the surface gas density and the surface star - formation rate ( kennicutt 1998 ) , we obtain a relation between @xmath0 and @xmath6 .
once @xmath6 is known for each grb , we can calculate the neutrino production efficiency and further calculate the accumulated neutrino flux from all grb host galaxies ( 3 ) .
because we have very few measurements of @xmath8 , but have measurement of @xmath0 for many grbs , we study whether @xmath0 are correlated with @xmath8 and act as an indicator of the gas surface density .
we collect the grb hosts that have spatially resolved observations from literatures ( svensson et al .
2010 ; fruchter et al .
2006 ; savaglio et al . 2009 ) , which are given in table 1 .
the total sfr and @xmath9 light radius of these hosts @xmath10 are obtained in these observations . by comparing the @xmath9 light radius and half light radius of some grb hosts that are spatially resolved ( fruchter et al .
2006 ; kelly et al .
2014 ) , we find that @xmath11 . converting @xmath9 light radius to the half - light radius @xmath12 with @xmath13 , we calculate the surface star - formation rates @xmath8 . with a sample of 18 grbs that have both measurements of @xmath0 from the xrt data of swift observations and @xmath8 of their host galaxies , we then study the relation between @xmath0 and @xmath14 . through the ordinary least - squares ( ols )
bisector fitting and @xmath14 , and it is not clear which variable is causative .
therefore , the ols bisector fitting which treat the variables symmetrically should be used , according to isobe et al .
( 1990 ) . ]
( isobe et al . 1990 ) , we find that @xmath15 with a correlation coefficient of @xmath16 and a null hypothesis probability of @xmath17 , indicating a good correlation .
the result of the fit is shown in the left panel of fig.1 . considering the kennicutt - schmidt law between the surface sfr and the surface density of molecular plus atomic gas ( kennicutt 1998 ) , i.e. , @xmath18 we get @xmath19 where @xmath20 is the mass of the hydrogen atom .
the roughly linear relation between @xmath21 and @xmath0 suggests that @xmath0 is a good indicator of the gas surface density @xmath6 .
note that @xmath0 is derived assuming solar metallicity for the absorbing gas .
since grb host galaxies usually have lower metallicity , the true absorbing gas column density should be corrected by the metallicity effect .
the coefficient @xmath22 may reflect such a correction .
we also study whether @xmath0 are correlated with the total sfrs of the host galaxies .
we collect 35 grbs in total that have measurements of @xmath0 and total sfrs , which are also listed in table 1 . through the ols bisector
fitting of the data of @xmath0 and sfr , we find that @xmath23 with a correlation coefficient of @xmath24 and a null hypothesis probability @xmath25 , which indicates a tight positive correlation .
the result of the fitting is shown in the right panel of fig.1 . with this relation
, we can infer the sfr of grb hosts with @xmath0 for each grb in our sample , which will be used to compute the galactic wind velocity in 3.1 .
the grb accelerated cr protons travel through the host galaxy and produce high - energy neutrinos via proton proton(_pp_)-collisions with its interstellar medium ( ism ) .
the collisions produce charged pions , which decay to neutrinos ( @xmath26 ) .
meanwhile , crs can escape out of the galaxy through diffusion or galactic wind advection .
these two competing processes ( i.e. collision and escape ) regulate the efficiency of the pion - production of crs , which can be described by @xmath27 , where @xmath28 is the energy - loss time of crs via ( @xmath29 ) collisions and @xmath30 is the escape time of crs .
pp_-collision energy loss time is @xmath31 , where @xmath32 is the inelasticity , @xmath33 is the gas number density , @xmath34 is the scale height of the gaseous disc of the galaxy .
@xmath35 is the _ pp_-collision inelastic cross - section , which slightly increases with the proton energy , given by @xmath36 ^ 2{\rm mb}$ ] where @xmath37 ( kelner et al .
thus , the collision energy loss time is @xmath38 . the crs escape from the grb host galaxy in two ways .
one is advective transportation by the galactic wind .
galactic - scale gaseous winds are ubiquitous in star - forming galaxies at all cosmic epochs ( heckman et al . 1990 ; pettini et al .
2001 ; shapley et al .
such winds can be driven by stellar winds , supernova explosions or other processes .
it is found that , for luminous and ultraluminous infrared galaxies at low redshifts , the winds from more luminous starbursts have higher speeds , roughly as @xmath39 sfr@xmath40 with @xmath41 being in the range of 0.25 - 0.35 ( martin 2005 ; rupke et al . 2005 ;
arribas et al . 2014 ) .
a similar trend is found for star - forming galaxies at @xmath42 ( weiner et al .
2009 ) , i.e. @xmath43 with the error in @xmath44 being @xmath45 .
such type of relation between the velocity of the outflowing material and sfr is expected if the mechanical energy is supplied by stellar winds and supernova explosions .
thus , one can calculate @xmath44 for each grb hosts with measured @xmath0 under the help of eq . 4 and eq . 5 .
taking @xmath46 as the reference value for grb host galaxies , the advective escape time is @xmath47 . the other way that crs escape is diffusion , i.e. , crs are scattered off small - scale magnetic field inhomogeneities randomly and diffuse out of the host galaxy . the diffusion time is estimated to be @xmath48 , where @xmath49 is the diffusion coefficient , and @xmath50 and @xmath51 are normalization factors .
since little is known about the diffusion coefficient in grb host galaxies , in the calculation we allow lower diffusion coefficients than that of our galaxy , which is @xmath52 .
the energy dependence of the diffusion coefficient is also unknown and @xmath53 depending on the spectrum of interstellar magnetic turbulence .
we assume two cases , one is the commonly - used value @xmath54 , based on the measurement of the cr confinement time in our galaxy ( engelmann et al .
1990 ; webber et al . 2003 ) .
another choice is @xmath55 , assuming the kolmogorov - type turbulence .
then the diffusive escape time is @xmath56 , with @xmath57 . combining the advective and diffusive escape timescales , the total escape time is @xmath58 when the difference between @xmath59 and @xmath60 is large , @xmath61 .
note that when @xmath62 , which holds for low - energy crs and small values of @xmath50 , the pion production efficiency is independent of @xmath34 and is given by @xmath63 the dependence of @xmath64 on cr energy has a break at the energy where @xmath65 .
below the break energy , @xmath66 slightly increases with the energy due to the _ pp_-collision inelastic cross - section , while above the break @xmath66 decreases as @xmath67 , leading to a steeper neutrino spectrum .
the fermi / gbm detection rate of grbs is about 250 per year ( paciesas et al . 2012 ) , and the total gamma - ray ( 10 - 1000 kev ) fluence of grbs detected in one year is @xmath68 .
the gbm field of view ( fov ) is roughly @xmath69 steradians , so the total all - sky flux of gamma - rays in 10 - 1000 kev is @xmath70 . then the energy production rate of gamma - rays in the energy range of 10 - 1000 kev , is estimated to be @xmath71 , where @xmath72 is the hubble constant and @xmath73 is a factor accounting for the source density evolution with redshift ( eichler et al .
according to the estimate of katz et al .
( 2009 ) , the present - day _ differential _ energy production rate of uhecrs above 30 eev is @xmath74 , where @xmath75 is the power - law index of the cr spectrum .
we assume @xmath76 as suggested by fermi acceleration .
define @xmath77 as the ratio between the _ differential _ energy production rate of uhecrs and the energy production rate of gamma - rays in 10 - 1000 kev . if grbs are the source of uhecrs above 30 eev , @xmath78 , assuming @xmath79 ( eichler et al .
2011 ) .
we collect a sample of grbs that have both measurements of @xmath80 and gamma - ray fluence . as the energy coverage of swift bat is small ,
we choose grbs that are detected by fermi / gbm to get more accurate values of the gamma - ray fluence .
there are 45 grbs in total that have measurements of both @xmath0 and gamma - ray fluence , as shown in table 2 . for each grb ,
the ( single - flavor ) neutrino fluence is estimated to be @xmath81 where @xmath82 is the differential energy flux of cr protons and @xmath83 is the gamma - ray fluence of this grb .
the accumulated neutrino flux produced by all grbs in the universe is @xmath84 where @xmath85 represents the @xmath85th grb in our sample , @xmath86 is the total number of grbs in our sample , and @xmath87 is the all - sky rate of grbs that would be detected by fermi / gbm if the field of view of gbm is @xmath88 steradians .
the result of the accumulated neutrino flux is shown in fig.2 .
we find that the accumulated neutrino flux at pev is @xmath89 per flavor for @xmath55 and @xmath90 .
the 1 sigma error in the accumulated flux is obtained by using formulas for the propagation of error .
it results mainly from the uncertainties in the neutrino production efficiency @xmath64 , in the values of @xmath0 , and in the cr - to - gamma - ray ratio @xmath77 . since the cr diffusion coefficient is not well - understood , we study its effect on the neutrino flux .
the results for different values of @xmath91 and @xmath50 are shown in fig .
3 . one can see that , when @xmath50 decreases , the neutrino flux increases at pev energies . for @xmath55 ,
if @xmath50 is as low as @xmath92 , the neutrino spectrum becomes too hard , although the flux increases . for @xmath54 , lower values of @xmath50
is needed to produce the same flux compared to the case of @xmath55 . however , the confinement of 100 pev crs producing pev neutrinos requires that @xmath50 be larger than @xmath93 ( murase et al .
thus , even under favorable conditions about the cr diffusion coefficient , the neutrino flux produced by grb crs alone can not explain the icecube observations if other assumptions used in our calculation are correct . in the calculation of the accumulated neutrino flux , we only considered those grbs that triggered fermi / gbm detector ( i.e. adopting @xmath94 ) .
there are dim grbs that do not trigger the detector and their total gamma - ray fluence may even be larger than that of the triggered ones according to the simulation results in liu & wang ( 2013 ) .
these untriggered grbs may also produce diffuse neutrinos as triggered grbs .
however , one should note that considering the contribution by untriggered grbs would not change our result about the accumulated neutrino flux , because the the cr - to - gamma - ray ratio @xmath77 needs to be re - calibrated correspondingly and thus the accumulated neutrino flux remains unchanged
. there could be an exceptional case , i.e. , if some dim grbs do not accelerate protons to energy @xmath95 ev ( although nuclei can still be accelerated to ultra high energies ) , but they can still contribute to pev neutrinos with crs of @xmath96 .
note also that this does not exclude the possibility that low - luminosity grbs themselves are responsible for their origin ( e.g. , murase & ioka 2013 , bhattacharya et al .
the above calculation has some uncertainties in the following aspects .
first , the estimate of the energy production rate of gamma - rays @xmath97 has uncertainty ( e.g. , dermer 2012 ) .
note that the factor @xmath79 is obtained by assuming that grb rate follows the cosmic sfr of porciani & madau ( 2001 ) or that of hopkins & beacom ( 2008 ) .
if grb density evolves faster than the cosmic sfr , @xmath73 is smaller and then @xmath77 is larger .
second , the estimate of the uhecr energy budget has uncertainty . the value can be a bit larger if one uses the telescope array or hires data .
third , the power - law index @xmath75 could be softer than 2 , then the energy budget of crs at 100 pev could be larger .
thus , given these uncertainties , the neutrino flux in the optimistic case could reach the observed value of icecube .
on the other hand , if one consider that grb host galaxies that are not detected by optical observations are possibly smaller galaxies , such as dwarf galaxies , the pion production efficiency may be smaller and thus the neutrino flux contributed by these galaxies would be smaller .
also , as shown by eq .
6 , the pion production efficiency depends on the speeds of the galactic winds .
the properties of the galactic winds in grb host galaxies are not well - explored .
x - ray observations of starburst galactic winds , such as m82 , usually give a higher speeds than that inferred from the optical observations ( strickland & heckman 2009 ) .
if the galactic wind speeds of grb hosts are proven to be higher , the neutrino flux produced by grb hosts would decrease .
the pion production efficiency also depends on the diffusive coefficient .
if @xmath98 is larger than @xmath99 ( for example , @xmath100 for our galaxy ) , the accumulated neutrino flex would be lower than @xmath101 at @xmath102 , as shown in fig .
3 . in summary , we calculated the neutrino flux produced by crs accelerated by grbs while they are propagating in the host galaxies based on our current knowledge about grb and their host galaxies .
these crs collide with nuclei of ism and produce neutrinos before they escape out of the galaxy . when the flux of crs produced by grbs is normalized with the observed flux of uhecrs above @xmath1 , the accumulated neutrino flux is @xmath103 per flavor under the usual assumptions about the grb properties and favorable assumptions about the cr diffusion coefficient .
the estimate , however , has uncertainty due to uncertainty in our current knowledge of grb and their host galaxies and the accumulated neutrino flux could reach the observed value by icecube in the optimistic case , so we can not rule out this scenario at present .
we thank zhuo li and ruoyu liu for useful discussions .
this work is supported by the 973 program under grant 2014cb845800 , the nsfc under grants 11273016 and 11033002 , and the excellent youth foundation of jiangsu province ( bk2012011 ) . | gamma - ray bursts ( grbs ) are proposed as candidate sources of ultra - high energy cosmic rays ( uhecrs ) .
we study the possibility that the pev neutrinos recently observed by icecube are produced by grb cosmic rays interacting with the interstellar gas in the host galaxies . by studying the relation between the x - ray absorption column density @xmath0 and the surface star - formation rate of grb host galaxies , we find that @xmath0 is a good indicator of the surface gas density of the host galaxies
. then we are able to calculate the neutrino production efficiency of crs for grbs with known @xmath0 .
we collect a sample of grbs that have both measurements of @xmath0 and accurate gamma - ray fluence , and attempt to calculate the accumulated neutrino flux based on the current knowledge about grbs and their host galaxies .
when the cr intensity produced by grbs is normalized with the observed uhecr flux above @xmath1 , the accumulated neutrino flux at pev energies is estimated to be about @xmath2 ( per flavor ) under the assumption that grb energy production rate follows the cosmic star - formation rate and the favorable assumption about the cr diffusion coefficient .
this flux is insufficient to account for the icecube observations , but the estimate suffers from some assumptions in the calculation and thus we can not rule out this scenario at present . |
in context of ads / cft correspondence , it was proposed that variation of cosmological constant corresponds to varying the number of colors in the boundary field theory of yang - mills with chemical potential interpretation @xcite . on the other hand , according to teitelboim and henneaux s mechanism , coupling four - dimensional gravity with antisymmetric gauge field without cosmological constant results into appearance of the cosmological constant as a constant of motion teit , and therefore , the cosmological constant will be a variable .
recently , the cosmological constant was considered as a state - dependant parameter in two dimensional dilaton gravity @xcite .
it was shown that treating the cosmological constant as a @xmath4 charge with non - minimal coupling leads to confining electrostatic potential .
larranaga showed that considering the cosmological constant as thermodynamical variable can be extended to use the smarr formula for inner and outer horizons of btz black hole @xcite .
recently , there has been an increasing interest in thermodynamical behavior of black holes in asymptotically ads spacetime .
this growing interest comes from the fact that using ads / cft correspondence , one can find answers regarding a conformal field theory of @xmath5-dimensions by solving problems that a gravitational field presents in ( @xmath6)-dimensional anti de sitter spacetime @xcite . hawking and page , in their pioneering work , showed that the similarity of phase transition between the stable large black hole and thermal gas in ads space can be interpreted as the confinement / deconfinement phase transition in the dual strongly coupled gauge theory @xcite .
later , in an interesting article , witten showed that using ads / cft correspondence , one can study the thermal phase transition and the interpretation of confinement in gauge theories @xcite . on the other hand
, one may consider the cosmological constant as a thermodynamical pressure ( in order to investigate the phase transition of black holes ) to extend the phase space ( see @xcite for more details ) and modify the first law of black holes thermodynamics @xcite .
another contribution of this consideration is a renewed interpretation for the mass of black holes which from internal energy becomes enthalpy @xcite .
this interpretation indicates that the mass of black holes plays more important role in thermodynamical structure of black holes and contains more information regarding phase structure of black holes @xcite . for a canonical ensemble with a fixed charge
, it was found that there exists a phase transition between small and large black holes .
this phase transition behaves very like the gas / liquid phase transition in a van der waals system @xcite . on the other hand ,
phase transition of small / large black holes in ads / cft correspondence may be interpreted as conductor / superconductor regions of the condescend matter systems @xcite .
considering the fact that maxwell theory contains some fundamental problems and nonlinear electromagnetic fields solve some of these shortcomings @xcite , one is motivated to study different models of nonlinear electrodynamics ( ned ) . one interesting class of these models is born - infeld ( bi ) type which is acquired in the low energy limit of heterotic string theory @xcite .
therefore , one is motivated to study these theories ( which in this paper we have considered logarithmic @xcite and exponential forms @xcite ) and the nonlinearity effects of electromagnetic field on critical values representing phase transition of black holes . on the other hand ,
einstein gravity is not flawless and has some fundamental problems @xcite .
generalization of the einstein gravity to higher orders of lovelock gravity is one way to solve some of these problems lovelock . besides , the lagrangian of lovelock gravity is obtainable through the use of the low energy effective action of string theory bi - st .
one can take this fact into account that modification of einstein gravity may change the conserved quantities of black holes and therefore it is inevitable to see that critical values and phase transition may depend on the choice of gravity model . in literature
, there have been various studies regarding higher orders of lovelock gravity in presence of different ned @xcite .
also , the phase transition and stability conditions of black holes in various gravity models have been studied intensively @xcite .
these investigations lead to interesting consequences and phenomenologies @xcite . in this paper , we consider higher orders of lovelock gravity in presence of two classes of ned and study their phase structure .
we investigate the effects of both nonlinearity of the electrodynamic models and lovelock parameters on the phase diagrams and the critical values .
considering the fact that we are treating black holes as thermodynamical system and interpret the first law of black holes mechanics as the first law of thermodynamics , we expect to see the similar thermodynamical behavior for the black holes and usual thermodynamical systems .
therefore , it is crucial to investigate phase transition and critical values of black holes .
moreover , lovelock gravity is a generalization of the einstein gravity and it is a theory that solves some of the einstein gravity problems .
hence , this generalization also gives a correction to calculated critical values of the einstein gravity .
the thermodynamical behavior of this modification should be reasonable and consistent with thermodynamical concepts .
furthermore , as it was mentioned before , nonlinear electromagnetic fields are introduced to solve shortcomings of maxwell theory .
so it is reasonable to investigate nonlinear effects of ned on the phase transition of black holes .
also , modifications of electrodynamics like gravitational parts must have consistent thermodynamical behavior .
this paper is constructed in the following form .
next section is devoted to introduction to higher orders of lovelock gravity and conserved quantities .
then , we extend the phase space by considering cosmological constant as thermodynamical pressure and study the smarr formula of these black holes . after that , we calculate critical values and plot related diagrams for different cases .
we give a detailed discussion regarding diagrams , their physical interpretations , and the effects of both ned and gravitational parameters .
we finish our paper with closing remarks .
here , we restrict ourselves to bi types ned models that were introduced by soleng @xcite and hendi @xcite with the following lagrangians @xmath7where enef and lnef denote _ exponential form of nonlinear electromagnetic field _ and _ logarithmic form of nonlinear electromagnetic field _ , respectively , and @xmath8 is nonlinearity parameter . considering the fact that we are interested in studying the spherically symmetric spacetime , we employ the following metric @xmath9 where @xmath10 denotes the standard metric of @xmath5-dimensional sphere , @xmath11 , with the volume @xmath12 .
it is known that the field equations of einstein gravity in the presence of ned are in the following forms @xcite @xmath13 where @xmath14 and @xmath15 are einstein tensor and cosmological constant , respectively .
einstein gravity in the presence of mentioned ned has been studied in ref .
regardless of gravitational sector , one may consider eqs .
( [ lned ] ) and ( [ maxwelleq ] ) with the metric ( [ metric ] ) to obtain the nonzero components of electromagnetic fields with the following explicit forms @xcite @xmath16 where @xmath17 is an integration constant which is proportional to total electric charge of the black hole solutions ( with regarding @xmath18 as proportionality constant ) extended phase space thermodynamics and phase diagrams of the einstein gravity in the presence of ned were investigated in ref .
so , we consider gauss - bonnet ( gb ) and third order of lovelock ( tol ) gravities and investigate extended phase space thermodynamics and critical behavior of these gravities . at first , we take into account the gb gravity . in order to obtain the field equation of gb gravity , one should add the @xmath19 tensor to the left hand side of eq .
( [ einsteineq ] ) , in which @xmath20 is @xmath21 , \label{gbterm}\]]where @xmath22 and @xmath23 is gb parameter .
considering eq .
( [ einsteineq ] ) with the extra term ( [ gbterm ] ) , one can obtain the following solutions @xcite @xmath24where @xmath25 and@xmath26 @xmath27 , & \;\;\text{enef } \\
\frac{8(d-2)}{\left ( d-1\right ) } \left [ \frac{\left ( 2d-3\right ) ( \gamma -1)% } { d-2}-\frac{\left ( d-1\right ) \ln \left ( \frac{1+\gamma } { 2}\right ) } { d-2}+% \frac{\left ( d-2\right ) \left ( 1-\gamma ^{2}\right ) \mathcal{h}}{d-3}\right ] , & \;\;\text{lnef } \end{array}% \right.\]]in which @xmath28 and @xmath29 is an integration constant which is related to the total mass @xcite @xmath30 in addition , @xmath31 and @xmath32 are in the following forms @xmath33 , % \left [ \frac{3d-7}{2d-4}\right ] , 1-\gamma ^{2 } \right ) , \nonumber \\ \gamma & = & \sqrt{1+\frac{q^{2}}{\beta ^{2}r^{2d-4}}}. \nonumber\end{aligned}\ ] ] calculating the kretschmann scalar , one finds that it diverges at @xmath34 , so the metric function ( [ ggb ] ) has an essential singularity at @xmath34 @xcite .
we should note that these solutions may be interpreted as asymptotically ads black holes as those of einstein case @xcite .
now , we take into account the surface gravity interpretation to obtain the hawking temperature of the mentioned black hole solutions , yielding @xcite @xmath35where@xmath36 e^{\frac{-2e^{2}}{\beta ^{2}}}-1\right ) , \;\;\text{enef } \\
\;8r^{3}\beta ^{2}\ln \left [ 1-\left ( \frac{e}{2\beta } \right ) ^{2}\right ] + % \frac{4r^{3}e^{2}}{1-\left ( \frac{e}{2\beta } \right ) ^{2}},\;\;\text{lnef}% \end{array}% \right .
, \]]and @xmath37 .
since obtained solutions are asymptotically ads , one may obtain the entropy of the black hole solutions by the use of the gibbs - duhem relation .
after some calculations , one can obtain @xcite @xmath38which confirms that the area law is recovered for @xmath39 .
now , we insert the following tol term , @xmath40 , to the field equation of gb gravity to obtain the solutions of tol gravity
. the tensor @xmath41 may be written as @xmath42 , \label{tolterm}\end{aligned}\]]where @xmath43 and @xmath44 are , respectively , the coefficient and lagrangian of tol gravity @xmath45 hereafter , we consider the special case @xmath46 to simplify the calculations .
it has been shown that the metric function of tol gravity in the presence of ned can be written as hendidehghani@xmath47 @xmath48 the geometric and thermodynamic properties of the asymptotically ads black holes have been studied before @xcite .
the finite mass of these solutions is the same as that of einstein gravity , where @xmath29 can be obtained as a function of @xmath49 from the metric function of tol gravity .
the hawking temperature and the entropy of the tol solutions can be calculated as @xcite @xmath50and @xmath51
as we mentioned in introduction , there are some motivations to view the cosmological constant as a variable . in addition
, there are various theories in which some physical constants ( such as gauge coupling constants , newton constant , lovelock coefficients and bi parameter ) are not fixed but dynamical ones . in that case , it is logical to consider the variation of these parameters into the first law of black hole thermodynamics ctevariable . in order to investigate the phase structure of these classes of gravities ,
we employ the approach in which the cosmological constant is a thermodynamical variable ( pressure ) with the following relation @xmath52 this consideration could be justified due to the fact that in quantum context , fundamental fixed parameters could vary .
as one can see the conjugating thermodynamical variable to this assumption ( cosmological constant as thermodynamical pressure ) will be volume where in literature the derived volume for different types of black holes are in agreement with the topology of the spacetime @xcite . in order to calculate the volume of these thermodynamical systems we use the following relation @xmath53 also
, we should consider the effects of cosmological constant in the first law of thermodynamics and extend our phase space .
with doing so the total finite mass of the black hole will play the role of enthalpy and hence the corresponding gibbs free energy will be in form of @xmath54 using the mentioned comments , one can obtain the volume with the following form @xmath55which is consistent with topological structure of spherical symmetric spacetime .
equation ( [ g ] ) was obtained in einstein gravity @xcite which indicates that although considering lovelock gravity modifies the metric function and some conserved quantities of the black hole , it does not change the volume of the black hole .
in addition , it was shown that the smarr formula may be extended to lovelock gravity as well as nonlinear theories of electrodynamics @xcite .
geometrical techniques ( scaling argument ) were used to derive an extension of the first law and its related modified smarr formula .
the result includes variations in the cosmological constant , lovelock coefficients and also nonlinearity parameter . in our case , lovelock gravity in the presence of ned
, @xmath56 should be the function of entropy , pressure , charge , lovelock parameters and bi coupling coefficient @xcite .
regarding the previous section , we find that those thermodynamic quantities satisfy the following differential form @xmath57 where we have achieved @xmath58 , and one can obtain @xmath59 using the redefinition of @xmath60 and @xmath61 with respect to the single parameter , @xmath62 , we can rewrite @xmath63 as a single differential form @xmath64 moreover , by scaling argument , we can obtain the generalized smarr relation for our asymptotically ads solutions in the extended phase space @xmath65where @xmath66 , & \text{ened } \\
-\frac{{2{\beta ^{2}}r_{+}^{d-1}}}{{(d-1)q}}\left ( { { \eta _ { + } } -1}\right ) , & \text{lned}% \end{array}% \right . ,
\end{aligned}\ ] ] @xmath67 , \end{aligned}\]]in which @xmath68 , \,\left [ { \frac{{d-3}}{{2d-4}}}\right ] , \,1-}\gamma _ { + } ^{2}\right ) , \\ \mathcal{d}_{+}&=&_{1}{f}_{1 } \left ( [ 1],\left[\frac{5d-11}{2d-4}\right],\frac{% l_{w+}}{2d-4}\right ) , \\ \mathcal{h}_{+}&=&_{2}{f}_{1}\left ( \left [ \frac{1}{2},\frac{d-3}{2d-4}\right ] , \left [ \frac{3d-7}{2d-4}\right ] , 1-\gamma _ { + } ^{2 } \right).\end{aligned}\ ] ] next step will be calculating critical values .
due to relation between volume and radius of the black hole , we use horizon radius ( specific volume ) in order to investigate the critical behavior of these systems phase - higher .
in order to do so , we use the method in which critical values are obtained through the use of @xmath69 diagrams . at first , we use the following relations to obtain the proper equations for critical radius @xmath70 for the economical reasons we will not bring obtained relations for calculating critical horizon radius . we employ numerical method for calculating critical values which result into following diagrams for different classes of lovelock gravity . we present various tables in order to plot @xmath69 , @xmath71 and @xmath2 and study the effects of gravitational parameter which is presented by @xmath72 and ned parameter which is presented by @xmath73 . in this paper , we have considered two classes of ned .
studying different phase diagrams for these ned shows that they have similar behavior .
therefore , for economical reason , we will regard only lnef branch and calculate related critical values of this ned model
. it will be constructive to give a short description regarding to different phase diagrams and the information they contain before presenting tables and phase diagrams .
@xmath2 diagrams are representing energy level of different states that phase transition takes place between them and shows the changes in energy level of before and after phase transition states .
the characteristic swallowtail that is seen in these diagrams shows the process that we know as phase transition .
it also gives interesting information regarding temperature of critical points .
for @xmath71 plot , it contains information regarding critical temperature and horizon radius in which phase transition takes place .
also , it gives some insight about single state regions which in our case is small / large black holes .
finally , studying @xmath74 plot gives us information regarding the behavior of pressure as function of horizon radius , critical pressure and critical horizon radius ( volume ) of phase transition .
one of the reasons for studying these diagrams is the similarity between phase structure of black holes and the van der waals thermodynamical systems . in what follows ,
we present various tables to investigate the effects of electrodynamics and gravity models on the critical values of phase transition .
we also plot @xmath69 , @xmath71 and @xmath2 diagrams for gb and tol gravities and interpret them .
it is notable to mention that considering the metric function of gb gravity , one finds that there is an upper limit for gb parameter to have a real solution . [ cols="^,^,^,^,^",options="header " , ] + table ( @xmath75 ) : tol gravity with @xmath76 , @xmath77 and @xmath78 .
@xmath79 @xmath80 @xmath81 @xmath82 @xmath83 @xmath84
as one can confirm , higher orders of lovelock gravity modify the phase diagrams and critical values of volume , pressure and temperature ( see tables @xmath85 and related figs .
of [ fig1 ] - [ fig7 ] ) .
it is clear that considering higher orders of lovelock gravity leads to different kinds of thermodynamical systems which was also evident through calculated conserved quantities . in this paper
, we consider thermodynamical effects of lovelock gravity up to gb and tol , separately .
it is evident that critical temperature and pressure are decreasing functions of @xmath72 and also orders of lovelock gravity whereas the critical volume is an increasing function of these two factors ( see figs .
[ fig8 ] , [ fig9 ] , [ fig11 ] and [ fig12 ] ) . on the other hand ,
the energy gap between two phases increases drastically by transforming from lower order of lovelock gravity to higher one ( see figs .
[ fig8 ] , [ fig9 ] , [ fig11 ] and [ fig12 ] right panels ) or by increasing @xmath72 for each order ( see figs .
[ fig8 ] and [ fig9 ] right panels ) .
in addition , the length of subcritical isobars increases which means that the single phase region of small / large black holes decreases ( see figs . fig9 and [ fig12 ] middle panels ) .
also , phase transitions region is an increasing function of @xmath72 and order of lovelock gravity ( see figs .
[ fig9 ] and [ fig12 ] left panels ) . to conclude ,
plotted figures show that the highest critical temperature and pressure , and the lowest critical volume and energy gap belong to einstein gravity . on the other hand
, system needs higher temperature to have phase transition in einstein gravity . whereas the critical temperature decreases by considering higher orders of lovelock gravity or increasing lovelock coefficients .
considering the fact that increasing lovelock parameters and/or adding higher orders of lovelock gravity increase the power of gravity , one may say that in stronger gravitational regimes , phase transitions take place in lower temperature . as for the effects of ned , we find the following results .
it is evident that the critical temperature in which swallow tail is formed ( see figs . [ fig5 ] , [ fig11 ] and [ fig12 ] right panels ) and critical pressure ( see figs .
[ fig5 ] , [ fig11 ] and [ fig12 ] left panels ) are decreasing functions of nonlinearity parameter whereas the critical volume is an increasing function .
in addition , the length of subcritical isobars is an increasing function of nonlinearity parameter .
this means that single phase region of small / large black holes is a decreasing function of @xmath73 ( see figs .
[ fig5 ] , [ fig11 ] and [ fig12 ] middle panels ) .
therefore , for higher values of nonlinearity parameter ( weak nonlinearity strength ) , black holes need to absorb less mass in order to have phase transition .
another issue that must be taken into account is the fact that as @xmath73 increases , the gap between isobars decreases whereas for small @xmath73 this gap is greater . due to the fact that we take into account bi type models , for large values of nonlinearity parameter
, they will lead to maxwell theory . obtained results
show that the lowest critical temperature and pressure and highest critical volume belong to maxwell theory . on the other hand
, one can conclude that the power of the nonlinearity causes the system to need higher critical temperature to have a phase transition .
this effect is opposite of what was observed for gravity .
in addition , the energy gap between two phases , critical temperature and gibbs free energy are increasing functions of dimensions ( fig . [ fig10 ] ) .
in other words , for higher dimensions , system needs to have more energy for having phase transition .
it is worthwhile to mention that subcritical isobars ( also critical region ) are increasing functions of dimensions ( see figs .
[ fig10 ] middle panels ) . whereas critical volume is a decreasing function of dimensions .
finally , the ratio @xmath3 is a decreasing ( an increasing ) function of @xmath72 ( @xmath73 ) .
@xmath86 @xmath87 @xmath88 @xmath89 @xmath90 @xmath91
in this paper , we have considered both gb and tol gravities in presence of two classes of bi type ned and studied their phase diagrams .
we have considered cosmological constant as pressure and its conjugating quantity as thermodynamical volume .
the obtained volume for these cases was consistent with topological structure of black holes and what was obtained previously @xcite .
it was shown that although both higher orders of lovelock gravity and ned modify thermodynamical quantities but the volume of the black holes in these cases is independent of these two modifications and only depends on the topology of the solutions . by employing numerical method ,
we have calculated critical thermodynamical values for different cases and studied the effects of gravitational and nonlinear electromagnetic field parameters on these critical values .
it was shown that black holes under consideration have similar behavior as van der waals liquid - gas system .
we found that critical temperature and pressure were decreasing functions of orders and/or coefficients of lovelock gravity and critical volume and energy gap were increasing functions of them .
in other words , black holes with higher orders of lovelock gravity go under phase transition and acquire stable state with lower temperature comparing to einstein case . on the other hand ,
the length of the subcritical isobars and region of the phase transition were increasing functions of the orders and/or coefficients of lovelock gravity .
orders of the lovelock gravity are denoting different powers of the curvature scalar . from what we have obtained
, one can argue that the critical temperature , pressure and the region of the small / large black holes are decreasing functions of the power of the curvature scalar .
while , the critical volume , subcritical isobars and region of the phase transition are increasing functions of it . therefore , the power of the curvature scalar indeed has a crucial role in variation of the critical values .
it is also regarded that considering the higher orders of lovelock theory , increases the power of gravity .
it was shown that critical temperature and pressure were decreasing functions of @xmath73 whereas the critical volume is an increasing function of it . in comparison between bi type and maxwell electrodynamics
, it was found that the lowest critical pressure and temperature and the largest critical volume belong to linear ( maxwell ) theory .
the gravitational and electromagnetic fields have opposite effects on critical pressure and also phase transition . according to these results
, one can say that phase transition is fundamentally related to both gravity and electrodynamics and their powers . as the power of gravitational field ( electromagnetic field ) increases ( decreases ) the critical temperature decreases ( increases ) .
interaction between the gravitational and electrodynamic sectors of charged black hole may be found through the metric function .
in addition , investigations of the phase diagrams confirm that they are weakening each others effects . as for dimensionality
, we found that as it increases , the energy gap , critical temperature and critical pressure increase .
the universal ratio of @xmath3 was a decreasing function of @xmath73 and increasing function of dimensions and lovelock parameter . for higher orders of lovelock gravity we have higher value of entropy ( @xmath92 ) which indicates that the thermodynamical system that the gravity describe contains higher value of disorder .
in addition , considering higher orders of lovelock gravity will cause the black holes to have higher degree of complexity in their geometrical structure . if one considers the complexity of the geometrical structure as a disorder measurement of the system , it is logical to expect to see higher value of entropy for higher order of lovelock gravity . on the other hand
, higher value of entropy means that our systems will have phase transition in lower critical temperature .
that is the result that we have derived through our numerical calculations .
it was shown that @xcite gb gravity in the presence of maxwell field has no phase transition for arbitrary electric charge in higher than @xmath93-dimensions . while we found that adjusting the nonlinearity parameter , @xmath73 , there is phase transition for various values of @xmath17 and @xmath94 .
in other words , the nonlinearity parameter of electrodynamics has modified both electrodynamics and thermodynamical behavior of a black hole system .
due to the opposite effects of gravitational power in lovelock gravity and the power of nonlinearity in electrodynamics , it is constructive to find the dominant effect for various domains of thermodynamical systems .
also , one can study whether these two different fields for certain values of @xmath72 and @xmath73 , cancel each other effects or not .
considering the effects of hawking radiation , we expect to see different behavior for higher orders of lovelock gravity .
this indicates that in order to investigate black holes evaporation and phase transition of black holes , one could take both effects simultaneously .
another interesting issue is studying the connection between complexity of the spacetime ( topological structure of black hole ) and entropy of the system and the interpretation of entropy as geometrical property .
we leave these problems for the future work . | in this paper , we consider lovelock gravity in presence of two born - infeld types of nonlinear electrodynamics and study their thermodynamical behavior .
we extend the phase space by considering cosmological constant as a thermodynamical pressure . we obtain critical values of pressure , volume and temperature and investigate the effects of both the lovelock gravity and the nonlinear electrodynamics on these values .
we plot @xmath0 , @xmath1 and @xmath2 diagrams to study the phase transition of these thermodynamical systems .
we show that power of the nonlinearity and gravity have opposite effects .
we also show how considering cosmological constant , nonlinearity and lovelock parameters as thermodynamical variables will modify smarr formula and first law of thermodynamics .
in addition , we study the behavior of universal ratio of @xmath3 for different values of nonlinearity power of electrodynamics as well as the lovelock coefficients . |
this paper is the second part of a general study about the feasibility of the forecast of meteorological parameters and optical turbulence at eso sites ( cerro paranal and cerro armazones ) in the framework of the mose project ( modeling eso sites ) .
the mose project , and the first results obtained for the vertical stratification of different atmospheric parameters , are presented in a joint paper @xcite .
the reader can refer to this joint paper to have more details about mose .
we only recall here that the mose project aims at proving the feasibility of the forecast of the most relevant classical atmospherical parameters for astronomical applications ( wind speed intensity and direction , temperature , relative humidity ) and the optical turbulence ot ( @xmath0 profiles ) with the integrated astro - climatic parameters derived from the @xmath0 i.e. the seeing ( @xmath10 ) , the isoplanatic angle ( @xmath2 ) , the wavefront coherence time ( @xmath3 ) above the two eso sites of cerro paranal ( site of the very large telescope - vlt ) and cerro armazones ( site selected for the european extremely large telescope - e - elt ) .
+ the final outcome of the project is to investigate the opportunity to implement an automatic system for the forecast of these parameters at the vlt observatory at cerro paranal and at the e - elt observatory at cerro armazones .
+ + the meso - nh model has already successfully been used to investigate some atmospherical parameters above sites of interest for the astronomy : at roque de los muchachos ( near ground temperature , @xcite ) ; at san pedro martir ( wind speed profiles , @xcite ) ; at cerro paranal in chile and maidanak in uzbekistan ( near ground wind speed , @xcite ) ; in antarctica ( wind speed and temperature profiles at dome c , @xcite ) ; at mount graham , arizona ( wind speed vertical distribution , @xcite and use of the meso - nh wind speed for wavefront coherence time reconstruction , @xcite ) . in the present study , we focus our analysis on the model ability in reconstructing the meteorological surface parameters ( temperature , wind speed and direction ) , from the ground up to 30 m a. g. l , at cerro paranal and cerro armazones . in section 2
, we describe the numerical set - up used for the mesoscale simulations . in section 3
, we present the data - set ( instruments and measurements ) and the statistical parameters used to evaluate the performances of the model . in section 4 , we evaluate the overall statistical performances of the model . in section 5 ,
we look at the single nights temporal evolutions of the meteorological surface forecasted parameters . in section 6 , we look more closely at the performances of the model for the individual nights . in section 7 , we discuss the usefulness of the correlation coefficient in our context .
conclusions are drawn in section 8 .
& & & & ground altitude ( m ) + & & & @xmath11x=500 m & @xmath11x=100 m & + cerro paranal @xmath12 & [email protected]``s & [email protected]''w & 2478 ( @xmath11h=156 m ) & 2545 ( @xmath11h=89 m ) &
2634 + cerro armazones@xmath13 & 24@xmath43521``s & 70@xmath41130''w & 2901 ( @xmath11h=164 m ) & 3010 ( @xmath11h=55 m ) & 3065 + + + all the numerical simulations of the nights presented in this study have been performed with the mesoscale numerical weather model meso - nh:@xmath14-@xmath15 @xcite .
the model has been developed by the centre national des recherches mtorologiques ( cnrm ) and laboratoire dareologie ( la ) from universit paul sabatier ( toulouse ) .
the meso - nh model can simulate the temporal evolution of three - dimensional meteorological parameters over a selected finite area of the globe .
we refer the reader to masciadri et al .
( 2013 ) , sec . 3.3 , for the general model configuration and the physical packages .
we just recall here the use of the grid - nesting technique @xcite , that consists in using different imbricated domains of the digital elevation models ( dem i.e orography ) extended on smaller and smaller surfaces , with increasing horizontal resolution but with the same vertical grid . in this study
we use two different configurations .
the first grid - nesting configuration employed three domains ( fig .
[ fig : oro ] and table [ tab : config ] ) and the innermost resolution is @xmath16 = 500 m. the second configuration is made of five imbricated domains , the first same three as the previous configuration , and other two centered at both paranal and armazones sites , with a horizontal resolution of @xmath16 = 100 m ( all domains of fig .
[ fig : oro ] and table [ tab : config ] ) .
one can notice that , using these configurations , we are able to do the forecast at both sites simultaneously .
the orographic dems we used for this project are the gtopo:@xmath17-@xmath18 with an intrinsic horizontal resolution of 1 km ( used for the domains 1 and 2 ) and the istar with an intrinsic horizontal resolution of 0.5 km ( used for the domain 3 ) . along the z - axis
we have 62 levels distributed as follows : a first vertical grid point equal to 5 m , a logarithmic stretching of 20 @xmath19 up to 3.5 km above the ground , and an almost constant vertical grid size of @xmath5600 m up to 23.8 km .
+ all the simulations done for the analyses discussed in this paper , were initialized the day before at 18 ut and forced every 6 hours with the analyses from the ecmwf , and finished at 09 ut of the simulated day ( for a total duration of 15 hours ) .
the statistics is computed only during night local time , from 00 ut to 09 ut . | this article is the second of a series of articles aiming at proving the feasibility of the forecast of all the most relevant classical atmospherical parameters for astronomical applications ( wind speed and direction , temperature , relative humidity ) and the optical turbulence ( @xmath0 and the derived astro - climatic parameters like seeing @xmath1 , isoplanatic angle @xmath2 , wavefront coherence time @xmath3 ... ) .
this study is done in the framework of the mose project , and focused above the two eso ground - bases sites of cerro paranal and cerro armazones . in this paper
we present the results related to the meso - nh model ability in reconstructing the surface layer atmospherical parameters ( wind speed intensity , wind direction and absolute temperature , [ 0 - 30 ] m a. g. l. ) .
the model reconstruction of all the atmospherical parameters in the surface layer is very satisfactory .
for the temperature , at all levels , the rmse ( root mean square error ) is inferior to 1@xmath4c . for the wind speed ,
it is @xmath52 m@xmath6s@xmath7 , and for the wind direction , it is in the range [ 38 - 46@xmath4 ] , at all levels , that corresponds to a rmse@xmath9 in a range [ 21 - 26% ] . if a filter is applied for the wind direction ( the winds inferior to 3 m@xmath6s@xmath7 are discarded from the computations ) , the wind direction rmse is in the range [ 30 - 41@xmath4 ] , i.e. a rmse@xmath9 in the range [ 17 - 23% ] .
the model operational forecast of the surface layer atmospherical parameters is suitable for different applications , among others : thermalization of the dome using the reconstructed temperature , hours in advance , of the beginning the night ; knowing in advance the main direction which the strong winds will come from during the night could allow the astronomer to anticipate the occurrence of a good / bad seeing night , and plan the observations accordingly ; preventing adaptive secondary mirrors shake generated by the wind speed .
[ firstpage ] turbulence - site testing - atmospheric effects - methods : data analysis - methods : numerical |
magnetic fields in differentially rotating disks play essential roles in the angular momentum transport which enable the accretion and various activities such as x - ray flares and jet formation . motivated by the skylab observations of the solar corona , galeev , rosner , & vaiana ( 1979 ) proposed a model of magnetically structured corona in accretion disks consisting of x - ray emitting magnetic loops .
the magnetic loops can be created due to the buoyant rise of magnetic flux tubes ( or flux sheet ) from the interior of the accretion disk .
matsumoto et al . ( 1988 ) carried out two - dimensional magnetohydrodynamic ( mhd ) simulations of the parker instability ( parker 1966 ) in nonuniform gravitational fields which mimic those in accretion disks .
they showed that when @xmath5 , a plane - parallel disk deforms itself into evacuated undulating magnetic loops and dense blobs accumulated in the valley of magnetic field lines .
the effects of rotation and shear flow , however , were not included in their simulations . shibata , tajima , & matsumoto ( 1990 ) carried out two - dimensional mhd simulations of the parker instability including the effects of shear flow and suggested that magnetic accretion disks fall into two types ; gas pressure dominated ( high-@xmath4 ) solar type disks , and magnetic pressure dominated ( low-@xmath4 ) cataclysmic disks . in high-@xmath4 ( @xmath6 ) disks , magnetic flux escapes from the disk more efficiently as @xmath4 decreases . several authors ( hawley , gammie & balbus 1995 ; matsumoto & tajima 1995 ; brandenburg et al . 1995 ; stone et al .
1996 ) have reported the results of three - dimensional local mhd simulations of magnetized accretion disks by adopting a shearing box approximation ( hawley et al .
1995 ) . in differentially rotating disks , magnetorotational instability ( balbus & hawley 1991 ) couples with the parker instability ( foglizzo & tagger 1995 ) .
since parker instability grows for long wave length perturbations along magnetic field lines , non - local effects may affect the stability and nonlinear evolution .
results of global 3d mhd simulations including vertical gravity were reported by matsumoto ( 1999 ) and hawley ( 1999 ) . by adopting an initially gas pressure dominated torus threaded by toroidal magnetic fields ,
matsumoto ( 1999 ) showed that magnetic energy is amplified exponentially owing to the growth of the balbus & hawley instability and that the system approaches a quasi - steady state with @xmath7 .
matsumoto ( 1999 ) assumed @xmath8 in the torus at the initial state .
when @xmath9 , the deviation from magneto - rotational equilibrium introduces large amplitude perturbations . in this paper , we present the results of 3d mhd simulations starting from an equilibrium mhd torus threaded by initially equipartition strength ( @xmath10 ) toroidal magnetic fields .
the basic equations we use are ideal mhd equations in cylindrical coordinate system @xmath11 .
we assume that the gas is inviscid and evolves adiabatically .
since we neglect radiative cooling , our numerical simulations postulates that the disk is advection - dominated ( see kato , fukue , & mineshige 1998 for a review ) .
the initial condition is an equilibrium model of an axisymmetric mhd torus threaded by toroidal magnetic fields ( okada et al .
we assume that the torus is embedded in hot , isothermal , non - rotating , spherical coronal gas . for gravity
, we use the newtonian potential .
we neglect the self - gravity of the gas . at the initial state
the torus is assumed to have a constant specific angular momentum @xmath12 .
we assume polytropic equation of state @xmath13 where @xmath14 is a constant , and @xmath15 is the specific heat ratio . according to okada et al .
( 1989 ) , we assume @xmath16 where @xmath17 is the alfvn speed and @xmath18 is a constant . for normalization
we take the radius @xmath19 where the rotation speed @xmath20 equals the keplerian velocity @xmath21 as unit radius and set @xmath22 at @xmath23 . using these normalizations , we can integrate the equation of motion into the potential form ; @xmath24 where @xmath25 is the square of the sound speed , @xmath26 and @xmath27 . by using equation ( 2 )
, we obtain the density distribution .
@xmath28 } { k[\gamma/(\gamma-1)][1+\beta_0^{-1 } r^{2(\gamma-1 ) } ] } \right\ } ^{1/(\gamma-1)}\ ] ] where @xmath29 is the plasma @xmath4 at @xmath30 .
the parameters describing the structure of the mhd torus are @xmath15 , @xmath31 , @xmath12 and @xmath14 . in this paper
we report the results of simulations for parameters @xmath32 , @xmath33 , @xmath34 , and @xmath35 .
the density of the halo at @xmath36 is taken to be @xmath37 .
the unit field strength is @xmath38 .
we solve the ideal mhd equations in a cylindrical coordinate system by using a modified lax - wendroff scheme with artificial viscosity .
we simulated only the upper half space ( @xmath39 ) and assumed that at the equatorial plane , @xmath40 , @xmath41 , @xmath42 , @xmath43 , @xmath44 , and @xmath45 are symmetric and @xmath46 and @xmath47 are antisymmetric .
the outer boundaries at @xmath48 and at @xmath49 are free boundaries where waves can transmit .
the singularity at @xmath50 is treated by softening the gravitational potential near the gravitating center ( @xmath51 ) .
the number of grid points is @xmath52 .
to initiate non - axisymmetric evolution , small amplitude , random perturbations are imposed at @xmath53 for azimuthal velocity .
figure 1a shows the initial condition .
color scale denotes the density distribution and red curves depict magnetic field lines .
figure 1b shows density distribution and velocity vectors at @xmath54 where @xmath55 is the orbit time @xmath56 . as the matter which lost angular momentum accretes to the central region , the mhd torus becomes disk - like .
velocity vectors indicate that matter flows out from the disk .
after the non - axisymmetric balbus & hawley instability grows , the inner region of the torus becomes turbulent .
the turbulent motions tangle magnetic field lines in small scale and create numerous current sheets ( or current filaments ) .
figure 1c shows the magnetic field lines projected onto the equatorial plane and the density distribution at @xmath57 .
note that figure 1c shows only the inner region where @xmath58 and @xmath59 .
in large scales , magnetic field lines and density distribution show low azimuthal wave number spiral structure .
figure 1d shows that magnetic loop structures similar to those in the solar corona are created .
the yellow surfaces show strongly magnetized regions where @xmath60 and red curves show magnetic field lines at @xmath61 .
the magnetic loops buoyantly rise from the disk due to the parker instability .
numerical results indicate that small loops which newly appeared above the photosphere ( emerging magnetic loops ) develop into expanding coronal loops .
we can observe magnetic loops elongated in the azimuthal direction and loops twisted by the rotation of the disk .
figure 2 shows the isosurfaces of plasma-@xmath4 at @xmath62 .
orange surfaces show strongly magnetized regions where @xmath63 .
yellow and green surfaces show the region where @xmath64 and @xmath65 , respectively . inside the disk , strongly magnetized , low-@xmath4 filaments are created . as we shall show below , low-@xmath4 region occupies a small fraction of the total volume . intermittent magnetic structures ( filamentary strongly magnetized regions ) similar to those in the solar photosphere develop in the disk .
figure 3a shows the spatial average of the magnetic energy , @xmath66 ( dashed curve ) , @xmath67 ( dash - dotted curve ) and @xmath68 ( solid curve ) averaged in @xmath69 and @xmath70 .
this figure shows that magnetic energy in the equatorial region decreases within a few rotation period due to the buoyant escape of magnetic flux , and that averaged plasma @xmath4 oscillates quasi - periodically around @xmath71 .
the ratio of averaged maxwell stress to initial equatorial pressure , @xmath72 is @xmath73 when @xmath74 .
figure 3b shows the time evolution of the accretion rate @xmath75 at @xmath76 , and of the outflow rate @xmath77 at @xmath78 .
accretion rate increases and fluctuates around a mean value .
numerical results also indicate quasi - periodic ejection of the disk material .
figure 3c shows the poynting flux @xmath79 which passes through the plane at @xmath80 .
after a few rotation period , magnetic flux is convected from the equatorial region to the disk surface .
the volume filling factor of the region where @xmath3 is shown in figure 3d .
strongly magnetized region where @xmath3 occupies about 8% of the total volume after a few rotation period .
later , the filling factor decreases to 2% .
after about 10 rotation period , magnetic flux is regenerated in the disk and filling factor shows the second peak when the averaged @xmath81 is maximum . following this second peak , poynting flux at @xmath82 increases , which suggest that magnetic flux escapes from the disk .
we showed by 3d global mhd simulations that when a differentially rotating torus is threaded by equipartition strength toroidal magnetic fields , magnetic loops emerging from the disk continue to rise and form coronal magnetic loops similar to those in the solar corona .
inside the disk , magnetic turbulence drives dynamo action which maintains magnetic fields and keeps the disk in a quasi - steady state with @xmath71 .
we successfully simulated the formation process of the solar type disk by 3d direct mhd simulations .
numerical results indicate that in differentially rotating disks magnetic field lines globally show spiral structure with low azimuthal wave number but locally fluctuating components create numerous current sheets .
we expect that magnetic reconnection taking place in current sheets generate hot plasmas which emit hard x - rays . when the disk is optically thin , such reconnection events
may be observed as large amplitude sporadic x - ray time variations characteristic of low - states in black hole candidates ( kawaguchi et al .
1999 ) .
we found that inside the torus , filamentary shaped , locally strongly magnetized , low-@xmath4 regions appear .
even when @xmath71 in average , low-@xmath4 regions where @xmath3 occupy 2 - 10% of the total volume .
the low-@xmath4 filaments are embedded in high-@xmath4 plasma .
such an intermittent structure is common in magnetized astrophysical plasmas .
numerical results indicate that low-@xmath4 filaments are re - generated after they once disappear .
although we assumed point gravity suitable for accretion disks , numerical results can qualitatively be applied to galactic gas disks .
our numerical results suggest that although @xmath83 in average , low-@xmath4 filaments exist in galactic gas disks .
magnetic reconnection taking place in such low-@xmath4 regions may heat the interstellar gas and create hot , x - ray emitting plasmas .
tanuma et al .
( 1999 ) proposed a model that magnetic reconnection in strongly magnetized ( @xmath84 ) regions in our galaxy creates hot plasma which emit 7kev component of galactic ridge x - ray emission ( grxe ) .
the low-@xmath4 filaments can also confine the hot plasma and prevent it from escaping from the galaxy .
we thank drs .
k. shibata , t. tajima , s. mineshige , t. kawaguchi and k. makishima for discussion .
numerical computations were carried out by vpp300/16r at naoj .
this work is supported in part by the grant of japan science and technology corporation and the grant - in - aid of the ministry of education , science , sports , and culture , japan(07640348 , 10147105 ) .
balbus , s. a. , & hawley , j. f. 1991 , , 376 , 214 brandenburg , a. , nordlund , a. , stein , r. , & torkelsson , u. 1995 , , 446 , 741 foglizzo , t. , & tagger , m. 1995 , , 301 , 293 galeev , a. a. , rosner , r. , & vaiana , g. s. 1979 , , 229 , 318 hawley , j. f. , gammie , c. f. , & balbus , s. a. 1995 , , 440 , 742 hawley , j. f. 1999 , , submitted kato , s. , fukue , j. , & mineshige , s. 1998 , black - hole accretion disks ( kyoto univ .
press ) kawaguchi , t. , mineshige , s. , machida , m. , matsumoto , r. , & shibata , k. 1999 , submitted to matsumoto , r. , horiuchi , t. , shibata , k. , & hanawa , t. 1988 , , 40 , 171 matsumoto , r. , & tajima , t. 1995 , , 445 , 767 matsumoto , r. 1999 , numerical astrophysics , ed .
miyama , s. , tomisaka , k. , and hanawa , t. , ( kluwer academic publishers ) , p.195 okada , r. , fukue , j. , & matsumoto , r. 1989 , , 41 , 133 parker , e. n. 1966 , , 145 , 811 shibata , k. , tajima , t. , & matsumoto , r. 1990 , 350 , 295 stone , j. m. , hawley , j. f. , gammie , c. f. , & balbus , s. a. 1996 , 463 , 656 tamura , s. , yokoyama , t , kudoh , t. , matsumoto , r. , shibata , k. , & makishima , k. , 1999 , , 51 , 161 | we present the results of three - dimensional global magnetohydrodynamic ( mhd ) simulations of the parker - shearing instability in a differentially rotating torus initially threaded by toroidal magnetic fields .
an equilibrium model of magnetized torus is adopted as an initial condition .
when @xmath0 at the initial state , magnetic flux buoyantly escapes from the disk and creates loop - like structures similar to those in the solar corona . inside the torus , growth of non - axisymmetric magneto - rotational ( or balbus & hawley ) instability generates magnetic turbulence .
magnetic field lines are tangled in small scale but in large scale they show low azimuthal wave number spiral structure .
after several rotation period , the system oscillates around a state with @xmath1 .
we found that magnetic pressure dominated ( @xmath2 ) filaments are created in the torus . the volume filling factor of the region where @xmath3 is 2 - 10% .
magnetic energy release in such low-@xmath4 regions may lead to violent flaring activities in accretion disks and in galactic gas disks . |
in the past few decades several proposals of modification of quantum mechanics ( qm ) have appeared in the literature , aimed at unifying its internal fundamental dichotomy between unitary deterministic quantum dynamics and non - linear irreversible state collapse following a measurement process @xcite . on the other hand big efforts
have been devoted towards an attempt to reconcile einstein gravity with quantum theory . in this context
, some approaches have focused on the possible role of gravity in state function collapse as a result of the incompatibility of general relativity and the unitary time evolution of qm @xcite .
it has been shown , in fact , that the existence of linear superpositions of states with macroscopic mass - distribution differences would entail a breakdown of classical space - time making the traditional quantum dynamics somehow troubling @xcite .
as distinct from penrose proposal , some other detailed collapse models have been proposed , which are based on a spontaneous stochastic state vector reduction : the ghirardi - rimini - weber ( grw ) model @xcite , the quantum mechanics with universal position location model ( qmupl ) of wave function collapse @xcite and the continuous spontaneous localization ( csl ) model @xcite .
recently the mechanism of spontaneous symmetry breakdown of time translation symmetry has been invoked as well in order to give rise to the quantum state reduction @xcite . in a different proposal ,
de filippo introduced a nonunitary model of newtonian gravity ( nng , from now on ) , which can be seen as the non - relativistic limit of a classically stable version of higher derivative gravity ( see e.g. ref.@xcite , references therein ) .
this model presents several appealing features to become a natural candidate for an effective low - energy model of gravity .
for example , while reproducing at a macroscopic level the ordinary newtonian interaction , it presents a mass threshold for gravitational localization , which for ordinary matter densities is around @xmath0 proton masses @xcite . the model can be seen as a realistic version of the nonunitary toy models @xcite inspired by the emergence of the information loss paradox @xcite from black hole physics . on the other hand the violation of unitarity when matching quantum mechanics and gravity
was argued also outside black hole physics , on general consistency grounds @xcite .
the model affords a mechanism for the evolution of macroscopic coherent superpositions of states into ensembles of pure states , each one of them corresponding within a future consistent general covariant theory to an unambiguous space - time .
its features include its ability to produce an evolution of the density matrix compatible with the expectations leading to the phenomenological spontaneous localization models , as it was argued that they should be both nonlinear and nonunitary @xcite . while sharing with the other proposals the non - linear non - unitary character , at variance with them , however , it does not present obstructions consistent with its special - relativistic extension @xcite .
another success of the model is the emergence of a unified picture for ordinary and black hole entropy as entanglement entropy with hidden degrees of freedom @xcite , in agreement with bekenstein - hawking entropy @xcite and hawking evaporation temperature ; that arises from the smoothed singularity of the black hole introduced by the model and paves the way for the quantum foundations of the second law of thermodynamics .
it is important to realize that the subject of a fundamental non - unitarity , and the various detailed mechanisms proposed @xcite , is not just a matter of philosophy , but could be , in principle , experimentally proved or disproved .
current technological progresses which have being achieved in isolating , manipulating and controlling an higher and higher number of degrees of freedom indicate a not far possibility of detecting fundamental decoherence , which would manifest in a clean way only once the system has been sufficiently protected against the sources of environmental noise @xcite .
indeed an experiment designed to detect fundamental deviations from unitary quantum evolution would be of considerable importance .
some technologies and devices have , at present , been recognized to be particularly suitable to create quantum state superpositions which are _ macroscopically _ distinct @xcite . among them
, there are diffraction of complex molecules up to @xmath1 proton masses @xcite , current cat states in squid devices @xcite and superpositions of atomic matter waves in bose einstein condensates exp3 .
recent progress in optomechanical systems may soon allow one to make superpositions of even larger objects , such as micro - sized mirrors or cantilevers @xcite , and to test quantum phenomena at larger scales . in this context
an appealing quite recent proposal for the practical realization of the penrose `` gedanken experiment '' @xcite considers the relatively small _
cm_-displacement of a lump of @xmath2 proton masses in a interferometric device in which two high - finesse optical cavities are inserted into its arms @xcite . the cavity in arm * a * has a very small end mirror mounted on a micro - mechanical oscillator ( cantilever ) , which suffers the radiation pressure of the photon confined inside it and as a consequence can be excited into a distinguishable quantum state .
a single photon incident on a 50 - 50 beam splitter will realize a superposition of being in either of the two arms ; then , the coupling between the photon and the cantilever will lead to an entangled state putting the cantilever into a superposition of distinct positions . after a full mechanical period of the cantilever , it recovers its original position ; if the photon leaks out of the cavity at this stage , a revival of the interference ( visibility ) is observed , provided that the quantum superposition state of the system survives at the intermediate times .
conversely , if the state of the system collapses due to some decoherence mechanism , visibility will not revive . summarizing
, a measurement of the magnitude of the revival of visibility gives a measurement of decoherence occurred in the time interval under consideration .
our work is devoted to calculate explicitly the output of this experiment @xcite according to nonunitary newtonian gravity model @xcite .
we would point out that detailed calculations for the expectations of some other collapse models , gravitational or not , have already been done , demonstrating a far reaching possibility to confirm the theory 31,32,33 .
the plan of the paper is as follows .
section ii contains a brief description of the basic nng model . in section
iii we give a qualitative discussion to argue , in the relevant parameter space of the experiment by marshall et al .
@xcite , the subspaces where gravitational effect could , in principle , be visible with both homogeneous and granular assumption on mirror mass distribution .
section iv is devoted to the application of nng model to the experiment , and its general solution .
then , in section v , we use the wigner function to monitor the mirror s state , and verify its behavior in time , after a measurement of photons state . finally , in sec.vi , we draw conclusions and outline perspectives of this work .
calculational details are devoted to appendices .
the aim of this section is to briefly recall the key features of the nng model which we will use for calculations . on the basis of a number of considerations ( among which consistency with basic formal relations of qm , approximate energy conservation , classical and quantum behavior of matter , requirement that non - unitary terms have a gravitational origin , etc ) , it is possible to isolate a two - parameter class of non - unitary gravity models , as discussed in detail in refs .
we will comment later in the section on these parameters , while we give here a concise definition of the model in its simplest form , which will allow us to carry out our calculations .
let @xmath3 $ ] be the non - relativistic hamiltonian of a finite number of particle species , like electrons , nuclei , ions , atoms and/or molecules , where @xmath4 denote the whole set @xmath5 of creation - annihilation operators , _
i.e. _ one couple per particle species and spin component .
@xmath6$]includes the usual electromagnetic interactions accounted for in atomic , molecular and condensed - matter physics . to incorporate that part of gravitational interactions responsible for non - unitarity
, one has to introduce complementary creation - annihilation operators @xmath7 and the overall ( meta-)hamiltonian : @xmath8+h[\widetilde{\psi } ^{\dagger } , \widetilde{\psi } ] + \\ & -{g}\sum_{j , k}m_{j}m_{k}\int d\textbf{x}d\textbf{y}\frac{\psi _ { j}^{\dagger } ( \textbf{x})\psi _ { j}(\textbf{x})\widetilde{\psi } _ { k}^{\dagger } ( \textbf{y})\widetilde{\psi}_{k}(\textbf{y } ) } { \end{split } \label{meta - hamiltonian}\ ] ] acting on the product @xmath9 of fock spaces of the @xmath10 and @xmath11 operators , where @xmath12 is the mass of the @xmath13-th particle species and @xmath14 is the gravitational constant .
the @xmath11 operators obey the same statistics as the corresponding operators @xmath10 , while @xmath15_{-}=[\psi , \widetilde{\psi } ^{\dagger } ] _ { -}=0 $ ] .
the meta - particle state space @xmath16 is the subspace of @xmath17 , including the meta - states obtained from the vacuum @xmath18 by applying operators built in terms of the products @xmath19 and symmetrical with respect to the interchange @xmath20 ; as a consequence they have the same number of @xmath10 ( physical ) and @xmath11 ( hidden ) meta - particles of each species .
since constrained meta - states can not distinguish between physical and hidden operators , the observable algebra is identified with the physical operator algebra . in view of this
, expectation values can be evaluated by preliminarily tracing out the @xmath11 operators . in particular ,
the most general meta - state corresponding to one particle states is represented by @xmath21 @xmath22 this is a consistent definition since @xmath23 generates a group of ( unitary ) endomorphisms of @xmath16 .
if we prepare a pure @xmath24-particle state , represented in the original setting , excluding gravitational interactions , by@xmath25 its representation in @xmath16 is given by the metastate @xmath26 a comment is in order on the possible extensions of the model outlined in refs.@xcite .
as said in the beginning of the section , the phenomenological general model depends on the two parameters @xmath27 .
the first refers to the number @xmath28 of copies in interaction , which on thermodynamical grounds can be inferred to be @xmath29 ( as in the model presented above ) .
it is however interesting to note that the limit @xmath30 ( with @xmath31 ) reproduces the famous non - linear newton - schrdinger equation , sometimes considered in the literature as a possible candidate equation for the inclusion of the self - gravity in qm , relevant to the quantum - classical transition @xcite .
the second , @xmath32 modulates the degree of nununitarity encoded in the gravitational interaction .
the above model definition corresponds to @xmath31 , for which all newtonian interaction is of nonunitary type .
this choice has been made in order to maximize the effect of nonunitarity , while calculating the prediction on the experiment by marshall et al .
@xcite in the best model - setting which gives the largest possible deviations from unitarity .
before considering the detailed application to the mirror experiment of ref .
@xcite , in this section we give a semi - quantitative argument for a gross identification of nng effects .
when considering self - interaction gravitational energy , the threshold mass of localization , @xmath33 @xmath34protonmasses @xmath35 , can be identified in the following way . if @xmath36 the metasystem behaves like an hydrogen - like system , while in the case @xmath37 the hidden mass - copy is quite well superposed to the physical one . as a consequence , the interaction potential can be approximated , within the lowest energy state of the meta - system , by the harmonic oscillator ground state with gaussian wave function width @xmath38 where @xmath39 @xmath40 is silicon density .
for nonunitary gravity to be effective in localizing the mirror , this length scale must be at least comparable with the wave packets separation : @xmath41where @xmath42 is the mirror s frequency and @xmath43 is the optomechanical coupling constant ; then the condition @xmath44 amounts to @xmath45 for the experimental parameters , _
i.e. _ @xmath46 and @xmath47 , we get @xmath48 , @xmath49 .
as pointed out in ref @xcite , an enhancement in the possibility of observing gravitational decoherence effects is provided by taking into account the real distribution of mass inside a crystal , which is very concentrated within nuclei ( see appendix a.2 ) . in that case , one should consider instead of @xmath50 a matter density @xmath51 , given by a silicon nucleus mass divided by its effective radius , which can be estimated as the typical spread of the wavefunction inside a crystal .
this leads to @xmath52 , which is still much greater than unity .
it should be stressed that the choice between homogeneous or granular matter distribution is not arbitrary , but is dictated by the experimental situation . as a matter of fact , when the relative displacement of meta - masses @xmath53 is of the order of the nucleus effective radius , it seems appropriate to take into account granularity
; when @xmath53 is made much greater than interatomic separation then homogeneity assumption appears to be the most suitable one ; finally , for @xmath53 of the order of interatomic separation , if imperfections like dislocations are present in the sample ( as it usually happens , even when very accurate preparation methods are used ) , then meta - masses are likely to ` feel ' an effective homogeneous masses potential ( this is because in the presence of a sufficient number of dislocations , as two meta - nuclei get nearer and nearer in one place , two nuclei in another place in the crystal can equally well go farther and farther from each other ) ; otherwise , for a really perfect crystal granularity should come again into play .
on the other hand , ( fundamental ) decoherence rate must be at least comparable with ( or lower than ) a period of natural oscillation of the mirror , which , in its turn , must be comparable with ( or lower than ) environmental decoherence rate for the experiment to be feasible .
then the following chain of relations must be satisfied : @xmath54 where @xmath55 is the gravitational interaction energy of the meta - masses ( see appendix a ) , @xmath56 is the environmental decoherence rate of the mirror @xcite . with the parameters of the experiment the value
@xmath57 has been proposed .
an exploration of parameter space within the exact solution has confirmed that first inequality ( [ eqn : ineq ] ) and approximate equality in eq .
( ineq2 ) must hold in order to see some relevant deviation from unitarity .
it should be stressed that , in spite of the improvement of mass size in the mirror experiment with respect to double - slit type experiments , the degree of _ macroscopicity _ of superposition is controlled not only by mass but also by space separation of the superposed wave packets .
a comment is in order on the apparent independence of the above result on the mass .
the main point is that we have chosen to measure cm state separation in units of coherent states size , which means that a larger mass is associated with a smaller unit of length and then , fixing @xmath58 , to a narrower peak separation .
one could also choose to fix the absolute separation @xmath59 , and express @xmath60 in the above formulas .
in this section we carry out interference visibility calculations within the mirror experiment @xcite in the framework of the nng model @xcite .
+ to this aim , let us start by defining the gravity - free hamiltonian @xmath61=\hbar\omega _ { ph}\left(\mathcal{n}_{a}+\mathcal{n}_{b}\right)+\hbar\omega _ { m}b^{\dagger } b-\hbar g\mathcal{n}_{a}\left ( b+b^{\dagger } \right),\ ] ] where @xmath62 @xmath63 are the number operators for the photon in the interferometer arms @xmath64 and @xmath65 respectively , while @xmath66 and @xmath67 are the phonon destruction and creation operators associated with the motion of the mirror s cm . in this way
our ( meta-)hamiltonian can be written as @xmath68 + h_{free}\left [ \tilde{b},\tilde{b}^{\dagger } , \widetilde{\mathcal{n } } _ { a};\omega _ { m}^{\ast } \right ] -k_{g}\left ( b+b^{\dagger } \right ) ( \widetilde{b}+\widetilde{b}^{\dagger}),\ ] ] with @xmath69 ( see appendix a for the calculation of the gravitational interaction strength @xmath70 in both homogeneous and granular case ) . in practice ,
the ( relevant degrees of freedom ) meta - system is formed by two gravitationally coupled harmonic oscillators and two by two photonic modes , each couple of modes interacting with its own mirror .
then , schrdinger state at time _
t _ is given by : @xmath71 where @xmath72 with @xmath73 } \\
\left\vert \beta _ { c}\right\rangle & = & \left\vert \beta e^{-i\omega _ { m}^*t}\right\rangle ; \left\vert \beta _ { l}\right\rangle = \left\vert \beta e^{-i\omega_{m}^*t}+\kappa \left ( 1-e^{-i\omega _ { m}^*t}\right ) \right\rangle.\end{aligned}\ ] ] here @xmath74 denotes the imaginary part of @xmath75 .
computational details on kernel @xmath76 are devoted to appendix b. since the only experimentally accessible quantity is the _ visibility _ , defined as twice the off - diagonal term ( in the absolute value ) of the physical photon s state @xmath77 , we are going to calculate this quantity as @xmath78 where @xmath79 is defined as @xmath80 it can be shown that visibility has a one - to - one correspondence with von neumann entropy , which represents a good measure of entanglement for a pure bipartite state @xcite .
we write the kernel as : @xmath81 + \mathcal{f}_{1}^{\gamma _ { + } , \gamma _ { + } } + \mathcal{f}_{2}^{\gamma _ { -},\gamma _ { -}}\biggr]}\;e^{-|\alpha|^2}}}\times\\ & \times & e^{-\frac{|\beta|^2}{2}-\frac{|\widetilde{\beta}|^2}{2}+\beta^*\alpha + \widetilde{\beta}^*\alpha } \times\\ & \times & e^{-\frac{k_{g}}{2\hbar ^{2 } } \biggl [ \underset{f_{p}}{\underbrace{\left ( \mathcal{f}_{1}^{\gamma , \gamma } + \mathcal{f}_{2}^{\gamma , \gamma } \right ) } } \beta ^{\ast 2}+2\underset { f_{\gamma p}}{\underbrace { ( \mathcal{f}_{1}^{\gamma , \gamma _ { + } } + \mathcal{f}_{2}^{\gamma , \gamma _ { -}}+2\alpha \mathcal{f}_{1}^{\gamma ^{\ast } , \gamma } ) } } \beta ^{\ast } \biggr]}\times\\ & \times & e^ { -\frac{k_{g}}{2\hbar ^{2}}\biggl [ \underset{f_{p}}{\underbrace { \left ( \mathcal{f}_{1}^{\gamma , \gamma } + \mathcal{f}_{2}^{\gamma , \gamma } \right ) } } \widetilde{\beta } ^{\ast 2}+2\underset{f_{m}}{\underbrace{\left ( \mathcal{f}_{1}^{\gamma , \gamma } -\mathcal{f}_{2}^{\gamma , \gamma } \right ) } } \beta ^{\ast } \widetilde{\beta } ^{\ast } + 2\underset{f_{\gamma m } } { \underbrace { ( \mathcal{f}_{1}^{\gamma , \gamma _ { + } } -\mathcal{f } _ { 2}^{\gamma , \gamma _ { -}}+2\alpha \mathcal{f}_{1}^{\gamma ^{\ast } , \gamma } ) } } \widetilde{\beta } ^{\ast } \biggr]}\end{aligned}\ ] ] where the functions @xmath82 are defined in appendix b , eq . ( [ eqn : f1 ] ) .
then we proceed to write the products @xmath83 in the form ( @xmath84 ) : @xmath85 } } } \
\underset{{\large k } _ { 2}\left ( \beta , \widetilde{\beta } \right ) } { \underbrace{e^{- k^{\prime } \left [ f_p\widetilde{\beta } ^{\ast 2}+2f_{m}\beta ^{\ast } \widetilde{\beta } ^{\ast } + 2f_{\gamma m}\widetilde{\beta } ^{\ast } \right ] } } } \underset{{\large k}_{3}\left ( \beta ^{\prime } \right ) } { \underbrace{e^{- k^{\prime } \left [ f^*_p\beta^{\prime 2}+2f_{\gamma p}^*\beta ^{\prime } \right ] } } } \ \underset{{\large k}_{4}\left ( \beta ^{\prime } , \widetilde{\beta } ^{\prime } \right ) } { \underbrace{e^{- k^{\prime } \left [ f^*_p\widetilde{\beta } ^{\prime 2}+2f_{m}^*\beta ^{\prime } \widetilde { \beta } ^{\prime } + 2f_{\gamma m}^*\widetilde{\beta } ^{\prime } \right ] } } } \times\\ & \times & \underbrace{e^{-\frac{|\beta|^2}{2}-\frac{|\beta'|^2}{2}+\beta^*\alpha+\beta'\alpha^ * } } _ { h(\beta,\beta')}\underbrace{e^{-\frac{|\widetilde{\beta}|^2}{2}-\frac{|\widetilde{\beta'}|^2}{2}+\widetilde{\beta}^*\alpha + \widetilde{\beta}'\alpha^*}}_{g(\widetilde{\beta},\widetilde{\beta}')}.\end{aligned}\ ] ] let s calculate the traces : @xmath86 } \underset{{\large l } \left ( \beta , \beta ^{\prime } \right ) } { \underbrace{e^{\beta _ { l}\beta _ { c}^{\prime \ast } -\frac{1}{2}\left\vert \beta ' _ { c}\right\vert ^{2}-\frac{1 } { 2}\left\vert \beta _ { l}\right\vert ^{2}}}}\biggl [ { \large k}^{10}(\beta , \widetilde{\beta } ) { \large k}^{00\ast } ( \beta ^{\prime } , \widetilde{\beta } ^{\prime } ) \underset{{\large h } _ { c}\left ( \widetilde{\beta } , \widetilde{\beta } ^{\prime } \right ) } { \underbrace{e^{\widetilde{\beta } _ { c}\widetilde{\beta } _ { c}^{\prime \ast } -\frac{1}{2}\left\vert \widetilde{\beta } _ { c}\right\vert ^{2}-\frac{1}{2}\left\vert \widetilde{\beta } _ { c}^{\prime } \right\vert ^{2}}}}\;+\\ & + & { \large k}^{11}(\beta , \widetilde{\beta } ) { \large k}^{01\ast } ( \beta ^{\prime } , \widetilde{\beta } ^{\prime } ) \underbrace{e^{i\kappa\im[\widetilde{\beta}(1-e^{-i\omega_m^*t})]}}_{\varepsilon(\widetilde{\beta } ) } \underbrace{e^{-i\kappa\im[\widetilde{\beta}'(1-e^{-i\omega_m^*t})]}}_{\delta(\widetilde{\beta } ' ) } \underset{{\large h } _ { l}\left ( \widetilde{\beta } , \widetilde{\beta } ^{\prime } \right ) } { \underbrace{e^{\widetilde{\beta } _ { l}\widetilde{\beta } _ { l}^{\prime \ast } - \frac{1}{2}\left\vert \widetilde{\beta } _ { l}\right\vert ^{2}-\frac{1}{2 } \left\vert \widetilde{\beta } _ { l}^{\prime } \right\vert ^{2 } } } } \biggr].\end{aligned}\ ] ] visibility
is then given by two contributions : @xmath87 here we do not write the explicit integrals for @xmath88 and @xmath89 , which can be found in appendix c ; it is worth noting that , if we discard their photons number dependence , they are formally similar . in the free case of no nng interaction , _
i.e. _ by putting @xmath90 , and for @xmath91 , we get @xmath92 as found in @xcite .
@xmath93 for the mirror size @xmath94 and several values of the frequency : @xmath95 ( dot - dashed ) , @xmath96 dashed , @xmath97 ( dotted ) and @xmath98 ( continuous black line ) . visibility and time
are expressed in dimensionless units.,height=151 ] the behavior of visibility for @xmath99 and four different values of @xmath42 is depicted in figure [ fig : visibility ] .
the first case shows no difference with the free case , and the remaining three , in decreasing order , progressively give a more and more reduction of revival effect at the end of the cycle .
a direct comparison can be made with the predictions of other phenomenological collapse models , like grw , qmupl and cls @xcite .
let us focus , for example , on the grw model @xcite .
in such a case the explicit formula for visibility is @xmath100 where @xmath101 is the visibility in the free case , _
i.e. _ in the absence of any mechanism of decoherence ( of fundamental or environmental nature ) , @xmath102 and @xmath103 is nucleon mass .
notice that also in this case no explicit dependence on mass appears .
it turns out that in the `` worst '' nng case ( @xmath104 ) , in which visibility goes practically to zero , grw predicts a visibility lowering at the end of the cycle of about @xmath105 with respect to the free case .
it is interesting to note that for smaller @xmath106 an enhancement of the effect s observability comes from the consideration of granularity . as discussed in the previous section , in fact , for @xmath107 or less matter granularity may come into play ; as a consequence , we obtain that already for @xmath108 visibility behaves in a similar way to the homogeneous case with @xmath109 ( dashed curve of figure [ fig : visibility ] ) .
in the previous section we have considered a measure of photons interference , and shown that with a proper choice of parameters one gets a reduced revival effect at the end of one oscillation period . here
we want to elucidate about the mirror s state soon after a photon measurement process . for monitoring mirror s state , and to get a physical insight of what s going on , we use the wigner function .
as it is well known , this quasi - distribution have both positive and negative parts , the latter being a signature of quantum coherence survival .
it is expected that the action of nng - induced decoherence would reduce , after some time , the interference patterns in the wigner distribution .
we calculate the wigner function starting from the expression : @xmath110 , \label{wigner}\ ] ] with @xmath111 where @xmath112 is the reduced density matrix of the mirror after a photon detection @xcite .
following @xcite , we take this measurement as the process projecting ( physical ) photons state onto the state @xmath113 where @xmath114 is a phase constant ( it can be shown that the corresponding results are quasi - independent of @xmath114 and then we put @xmath115 ) .
correspondingly , mirror s density matrix is ( @xmath116 ) @xmath117 .\ ] ] after some calculations , we get the wigner function as @xmath118 \right ] = \notag \\ & = & \frac{1}{4\hbar \pi ^{5}}\int d^{2}(\beta , \widetilde{\beta } , \beta ^{\prime } , \widetilde{\beta ^{\prime } } ) e^{-2\eta \eta ^{\ast } } \biggl[\alpha _ { 1}e^{-\beta _ { c}\beta _ { c}^{\prime ^{\ast } } -\frac{1}{2}\left\vert \beta _ { c}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { c}^{\prime } \right\vert ^{2}+2\beta _ { c}\eta ^{\ast } + 2\beta _ { c}^{\prime ^{\ast } } \eta } + \alpha _ { 2}e^{-\beta _ { l}\beta _ { l}^{\prime ^{\ast } } -\frac{1}{2}\left\vert \beta _ { l}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { l}^{\prime } \right\vert
^{2}+2\beta _ { l}\eta ^{\ast } + 2\beta _ { l}^{\prime ^{\ast } } \eta } \notag \\ & + & \alpha _ { 3}e^{-\beta _ { c}\beta _ { l}^{\prime ^{\ast } } -\frac{1}{2}\left\vert \beta _ { c}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { l}^{\prime } \right\vert ^{2}+2\beta _ { c}\eta ^{\ast } + 2\beta _ { l}^{\prime ^{\ast } } \eta } + \alpha _ { 4}e^{-\beta _ { l}\beta _ { c}^{\prime ^{\ast } } -\frac{1}{2}\left\vert \beta _ { l}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { c}^{\prime } \right\vert ^{2}+2\beta _ { l}\eta ^{\ast } + 2\beta _ { c}^{\prime ^{\ast } } \eta } \biggr ] .
\label{wigner_int}\end{aligned}\ ] ] explicit expressions for @xmath119 and @xmath120 , together with calculational details , are reported in appendix d. results for @xmath121 and @xmath122 and @xmath123 , are shown in figures [ fig : wig_func_1 ] and [ fig : wig_func_2 ] respectively for the homogeneous case . as seen for visibility , in the case of ( proposed ) experimental values , wigner function after measurement is undistinguishable from the free case .
for the second ( presently unattainable ) much smaller value of @xmath124 , after a certain time a diminution of interference fringes is observed together with a contextual lowering of the first rest peak .
in this paper the output of the mirror experiment proposed by marshall et al . @xcite has been calculated within the framework of nng , assuming both homogeneous and granular mass distributions . by varying the experimental parameters in a wide range beyond the proposed values @xcite
, a window of sensible parameters has been found in which the nng induced decoherence effect is manifest . in conclusion , even if the experimental test of nonunitary gravity has been proved to lie beyond current technology yet , requiring an unprecedented control of decoherence , its peculiar form of self - gravitational interaction has been shown to be in principle distinguishable from the action of other collapse models .
an exploration of the relevant parameter space could in fact , in a feasible experiment , lead to a clear distinction of the most appropriate model .
the signature of nng model is ultimately connected with the fact that fundamental interaction occurs with a simple system ( the mirror s copy in this case ) rather than with a fundamental thermal bath random field leading to a visibility output somehow indistinguishable from the effect of temperature .
this is essentially due to the laboratory artificially created superposition state , while in naturally occurring circumstances it is expected that fundamental environment , being as complex as the system itself , could eventually lead to auto - thermalization effects .
+ in spite of the huge technical challenges , however , we believe that due to the rapid progress in developing high - quality micro - optomechanical devices , a prototypal experiment of this type could be soon realized .
a remark is finally in order concerning a finite temperature inclusion into the model .
it should be clear that , when our initial knowledge of the system state is characterized by a density matrix like a thermal state , there is no unique prescription to associate it with a pure meta - state .
in such a case one has to consider the possibility of using mixed meta - states to encode our incomplete knowledge .
this more general case , independently of the specific experiment treated here , will pave the way towards a generalized model of gravity induced thermalization .
such an interesting issue will be addressed in a future publication @xcite .
the authors would like to thank mario salerno for discussions and useful comments on the manuscript .
anonymous referee s critical comment , pointing out an important issue of the work , is also kindly acknowledged .
in this appendix the gravitational interaction potential between meta - mirrors is computed within the two assumption of homogeneous and granular mass distributions .
let s consider the newtonian potential energy for two particles with masses @xmath125 and @xmath126 @xmath127 starting from the initial condition where the two mirrors of length @xmath128 are overlapped , we consider the shifting of the first mirror of @xmath129 along @xmath130 positive axis and the second one along @xmath131 negative axis .
because of the very small relative displacement @xmath132 between the meta - mirrors ( which is at maximum of the order of the size of the wave - function describing , in the ordinary setting , the cm coordinate of the mirror ) , it is enough to calculate the quadratic term of the expansion in the distance of the total gravitational interaction energy @xmath133 d^{2}+o\left ( d^{3}\right),\end{aligned}\ ] ] where @xmath134 is the interaction energy of two square infinitesimally tiny mirror sheets parallel to the @xmath135 plane , @xmath136 here we have used the assumption that , for example , @xmath137 in this way the terms linear in @xmath132 vanish . defining the non - dimensional coordinates
@xmath138and so on , we get : @xmath139=\\ & = & \frac{g\,m^{2}}{l^{3 } } \iint\limits_{-1/2}^{1/2}dx_{1}^{\prime } dy_{1}^{\prime } \iint\limits_{-1/2}^{1/2}dx_{2}^{\prime } dy_{2}^{\prime } \left ( \frac{1 } { \left [ \left ( x_{1}^{\prime } -x_{2}^{\prime } \right ) ^{2}+\left ( y_{1}^{\prime } -y_{2}^{\prime } \right ) ^{2}\right ] ^{1/2}}-\frac{1}{\left [ \left ( x_{1}^{\prime } -x_{2}^{\prime } \right ) ^{2}+\left ( y_{1}^{\prime } -y_{2}^{\prime } \right ) ^{2}+1\right ] ^{1/2}}\right)=\\ & = & 4\frac{g\,m^{2}}{l^{3}}\int_{0}^{1}d\xi _ { -}\int_{0}^{1-\xi _ { -}}d\xi _ { + } \int_{0}^{1}d\eta _ { -}\int_{0}^{1-\eta _ { -}}d\eta _ { + } \left ( \frac{1 } { \left [ \xi _ { -}^{2}+\eta _ { -}^{2}\right ] ^{1/2}}-\frac{1}{\left [ \xi _ { -}^{2}+\eta _ { -}^{2}+1\right ] ^{1/2}}\right ) \\ & = & 4\frac{g\,m^{2}}{l^{3}}\int_{0}^{1}d\xi _ { -}\int_{0}^{1}d\eta _ { -}\left ( 1-\xi _ { -}\right ) \left ( 1-\eta _ { -}\right ) \left ( \frac{1}{\left [ \xi _ { -}^{2}+\eta _ { -}^{2}\right ] ^{1/2}}-\frac{1}{\left [ \xi _ { -}^{2}+\eta _ { -}^{2}+1\right ] ^{1/2}}\right ) = \frac{2\pi } { 3}\frac{g\,m^{2}}{l^{3}},\end{aligned}\ ] ] where we have introduced the new variables @xmath140 and @xmath141
. then we obtain @xmath142 from which , writing the interaction term in the meta - hamiltonian as @xmath143 , we get @xmath144 looking at the meta - mirrors as aggregates of atoms disposed in a lattice @xmath145 , we are led to consider the gravitational potential between the meta - crystals described by the state ( to be symmetrized ) @xmath146 where superscripts parameters @xmath147 refer to the center of mass displacements from the origin of coordinate system along , say , @xmath148axis .
the problem is similar to that of potential between two atoms or molecules , where ( within the born - oppenheimer approximation ) nuclei positions are treated as parameters of the atoms / molecules .
+ the interaction potential is then given by ( @xmath149 ) @xmath150}}{\left\vert \mathbf{x}-\mathbf{y}\right\vert } \\ & \simeq -gm_{nuc}^{2}\left ( \frac{m_{nuc}\omega _ { crystal}}{\hbar \pi } \right ) n_{nuc}\int \int d\mathbf{x}d\mathbf{y}\frac{e^{-\frac { m_{nuc}\omega _ { crystal}}{\hbar } \left [ \left ( \mathbf{x}+\mathbf{d}/2\right ) ^{2}+\left ( \mathbf{y}-\mathbf{d}/2\right ) ^{2}\right ] } } { \left\vert \mathbf{x - y}\right\vert } , \end{split}\end{aligned}\ ] ] where , for simplicity , einstein model for the crystal has been used , with @xmath151 einstein frequency @xmath152 , @xmath153 is nucleus mass , @xmath154 the number of nuclei . in the last line of the above formula , we have made the assumption that the nuclei wavefunction spreads are much lower than interatomic distance , and that the formers are greater than @xmath132 . to get a simple estimate
, we can consider the interaction between interpenetrating spheres of radius @xmath155 separated by a distance @xmath132 , so that @xmath156 giving the ( enhanced ) gravitational coupling @xmath157
in the following we study in detail the dynamical evolution of the meta - system . as initial meta - state we take @xmath158 \otimes \left [ \left ( \left\vert 0_{\tilde{a}}1_{\tilde{b}}\right\rangle + \left\vert 1_{\tilde{a}}0_{\tilde{b}}\right\rangle \right ) \
\left\vert \widetilde{\alpha}\right\rangle \right]\label{state}\equiv \left\vert \psi \left ( 0\right ) \right\rangle \otimes \left\vert \psi \left ( 0\right ) \right\rangle\ ] ] where general coherent states @xmath159 and @xmath160 are considered , although acceptable meta - states must be symmetrized with respect to the physical and hidden parts .
let s start by calculating the meta - state at time @xmath161 .
we will use the interaction picture , defining @xmath162 + h_{free}\left [ \tilde{b},\tilde{b}^{\dag},\mathcal{n}_{a}^{\sim } , \mathcal{n}_{b}^{\sim } ; \omega _ { m}^{\ast } \right ] \right ) t/\hbar } \left\vert \left\vert \psi(t)\right\rangle \right\rangle _ { sch}\,\ ] ] where @xmath163 and @xmath164 } \left\vert 1_{a}0_{b}\right\rangle \otimes \left\vert \alpha e^{-i\omega _ { m}^*t}+\kappa \left ( 1-e^{-i\omega _ { m}^*t}\right ) \right\rangle\right).\ ] ] the interaction hamiltonian @xmath165 evolves as @xmath166 + h_{free}\left [ \tilde{b},\tilde{b}^{\dagger } , \mathcal{n}_{a}^{\sim } , \mathcal{n}_{b}^{\sim } ; \omega _ { m}^{\ast } \right ] \right ) t/\hbar } \left ( b+b^{\dagger } \right)\times\\ & \times & ( \widetilde{b}+\widetilde{b}^{\dagger } ) e^{-i\left(h_{free}\left[b , b^{\dagger } , \mathcal{n}_{a},\mathcal{n } _ { b};\omega _ { m}^{\ast } \right ] + h_{free}\left [ \tilde{b},\tilde{b}^{\dagger},\mathcal{n}_{a}^{\sim } , \mathcal{n}_{b}^{\sim } ; \omega _ { m}^{\ast } \right]\right ) t/\hbar}=\\ & = & -k_{g}e^{ih_{free}\left [ b , b^{\dagger } , \mathcal{n}_{a},\mathcal{n } _ { b};\omega _ { m}^{\ast } \right ] t/\hbar } \left ( b+b^{\dagger } \right ) e^{-ih_{free}\left [ b , b^{\dagger } , \mathcal{n}_{a},\mathcal{n}_{b};\omega
_ { m}^{\ast } \right ] t/\hbar } \times\\ & \times & e^{ih_{free}\left [ \tilde{b},\tilde{b}^{\dagger } , \mathcal{n}_{a}^{\sim } , \mathcal{n}_{b}^{\sim } ; \omega _ { m}^{\ast } \right ] t/\hbar } ( \widetilde{b}+\widetilde{b}^{\dagger } ) e^{-ih_{free}\left[\tilde{b},\tilde{b}^{\dagger } , \mathcal{n}_{a}^{\sim } , \mathcal{n}_{b}^{\sim } ; \omega _ { m}^{\ast } \right ] t/\hbar } .\end{aligned}\ ] ] the above expression is the product of two specular terms , so it suffices to calculate the first , say .
we make use of the backer - hausdorff lemma : @xmath167_{n},\ ] ] with @xmath168_{0}=g,\quad\left[f , g\right ] _ { n}=\left [ f,\left[f , g\right ]
_ { n-1}\right ] , \ ] ] and @xmath169 t,\qquad g = b+b^{\dagger } , \ ] ] ( @xmath170 is now defined as @xmath171 ) by which , noting that @xmath172_{0 } & = & g = b+b^{\dagger } , \\
\left[f , g\right]_{1 } & = & -i\omega _ { m}^*t\left[b^{\dagger } b , b+b^{\dagger } \right ] = i\omega_{m}^*t\left(b - b^{\dagger}\right ) , \\
\left[f , g\right]_{2 } & = & \left [ f,\left[f , g\right]_{1}\right ] = -(\omega_{m}^*t)^{2}\left[\left(b+b^{\dagger } \right)-2\left ( g/\omega _ { m}^*\right)\mathcal{n}_{a}\right ] , \\
\left[f , g\right ] _ { 3 } & = & -i(\omega _ { m}^*t)^{3}\left ( b - b^{\dagger } \right ) = -(\omega _ { m}^*t)^{2}\left [ f , g\right ] _ { 1 } , \\ \left [ f , g\right]_{4 } & = & -(\omega
_ { m}^*t)^{2}\left [ f , g\right ] _ { 2 } \\ & & ... \\ & & ... \\ \left[f , g\right]_{n_{even}\neq 0}&=&(-1)^{n_{even}/2}\left ( \omega_{m}^*t\right)^{n_{even}}\left[\left(b+b^{\dagger } \right ) -2\left(g/\omega _ { m}^*\right ) \mathcal{n}_{a}\right ] \\ \left [ f , g\right]_{n_{odd } } & = & \left ( i\omega _ { m}^*t\right ) ^{n_{odd}}\left ( b - b^{\dagger } \right ) , \end{aligned}\ ] ] we obtain : @xmath173 t/\hbar } \left ( b+b^{\dagger } \right ) e^{-ih_{free}\left [ a , a^{\dagger } , \mathcal{n}_{a},\mathcal{n}_{b};\omega _ { m}^{\ast } \right ] t/\hbar } \\ & = & b^{\dagger } ( \cos ( \omega _ { m}^ * t ) + i \sin ( \omega _ { m}^ * t ) ) + b ( \cos ( \omega _ { m}^ * t ) -i \sin ( \omega _ { m}^ * t ) ) -\frac{2 g \mathcal{n}_a \cos ( \omega _ { m}^ * t ) } { \omega _ { m}^ * } + \frac{2 g \mathcal{n}_a}{\omega _ { m}^ * } , \end{aligned}\ ] ] and a similar result for @xmath174 @xmath175 t/\hbar } \left ( \widetilde{b}+\widetilde{b}^{\dagger}\right ) e^{-ih_{free}\left [ \widetilde{a},\widetilde{a}^{\dagger } , \widetilde{\mathcal{n}}_{a},\widetilde{\mathcal{n}}_{b};\omega _ { m}^{\ast } \right ] t/\hbar } \\ & = & \widetilde{b}^{\dagger } ( \cos ( \omega _ { m}^ * t)+i \sin ( \omega _ { m}^ * t ) ) + \widetilde{b } ( \cos ( \omega _ { m}^ * t ) -i \sin ( \omega _ { m}^ * t ) ) -\frac{2 g \mathcal{\widetilde{n}}_a \cos ( \omega _ { m}^*t ) } { \omega _ { m}^ * } + \frac{2 g \mathcal{\widetilde{n}}_a}{\omega _ { m}^ * } , \end{aligned}\ ] ] or @xmath176 where @xmath177 the evolution equation is @xmath178 i.e. , using the hubbard - stratonovich transformation @xcite , @xmath179e^{-\frac{ic^{2}}{\hbar } \int_{0}^{t}dt^{\prime}(\varphi_{1}^{2}-\varphi_{2}^{2 } ) } \widehat{t}e^ { \frac{ic\sqrt{k_{g}}}{\hbar } \int_{0}^{t}dt^{\prime}(\varphi_{1}+\varphi_{2 } ) \widehat{\mathcal{o}}(t ) } \widehat{t}e^{\frac{ic\sqrt{k_{g}}}{\hbar } \int_{0}^{t}dt^{\prime } ( \varphi_{1}-\varphi_{2})\widehat{\widetilde { \mathcal{o}}}(t)}\|\psi ( 0)\rangle\rangle \\ & = & \frac{1}{\pi^2}\iint d^{2}\beta\,d^{2}\widetilde{\beta } \int d[\varphi_{1}(t ) , \varphi_{2}(t)]e^{-ic^{2}/\hbar\int_{0}^{t}dt^{\prime}(\varphi_{1}^{2}- \varphi _ { 2}^{2})}\langle\beta |\widehat{t}e^{ic\sqrt{k_{g}}/\hbar \int_{0}^{t}dt^{\prime}(\varphi_{1}+\varphi_{2})\widehat { \mathcal{o } } ( t)}|\alpha\rangle\times \\ & \times & \langle\widetilde{\beta}| \widehat{t } e^{ic\sqrt{k_{g}}/\hbar \int_{0}^{t}dt^{\prime } ( \varphi_{1}-\varphi_{2})\widehat { \widetilde { \mathcal{o}}}(t)}|\widetilde{\alpha}\rangle|\beta\rangle| ent\rangle\otimes| \widetilde{\beta } \rangle|\widetilde{ent}\rangle=\frac{1}{\pi^2}\iint d^2\beta d^2\widetilde{\beta}\;k^{\mathcal{n}_{a } \widetilde { \mathcal{n}}_{a}}(\beta , \widetilde{\beta } ; \alpha;t)|\beta\rangle| ent\rangle\otimes|\widetilde{\beta } \rangle| \widetilde{ent}\rangle,\end{aligned}\ ] ] where @xmath180 is a constant . before dealing with the kernel @xmath181 , we calculate the amplitude with the help of the baker - campbell - hausdorff formula : @xmath182 } e^{\frac{t^3}{6}(2[\widehat{b},[\widehat{a } , \widehat{b}]]+[\widehat{a},[\widehat{a},\widehat{b}]])}\dots,\ ] ] @xmath183 @xmath184 a similar result holds on for the other amplitude by making the following substitutions : @xmath185 we get then @xmath186}\times \\ & \times&\lim_{\substack { n\rightarrow\infty \\ \delta t\rightarrow 0}}\iint\dots\int \prod_{i=1}^nd\varphi_{2,i}\exp\biggl\{\sum_{i=1}^n\biggl[\frac{ic^2}{\hbar } \delta t\delta_{i , j}+\frac{f^2}{2}(\delta t)^2\gamma_{i}\gamma^{\ast}_{j}\biggr]\varphi_{2,i}\varphi_{2,j}+ f\sum_{i=1}^n\delta t\bigl(\beta^{\ast}\gamma_{i}+\gamma_i-\widetilde{\beta}^{\ast}\gamma_{i}-\widetilde{\gamma}_i\bigr)\varphi_{2,i}\biggr\}\times \\ & \times&\iint\dots\int\prod_{i=1}^nd\varphi_{1,i}\exp\biggl\{\sum_{i , j=1}^n\biggl[-\frac{ic^2}{\hbar}\delta t\delta_{i , j}+\frac{f^2}{2}(\delta t)^2\gamma_{i}\gamma^{\ast}_{j}\biggr]\varphi_{1,i}\varphi_{1,j}+f\sum_{i=1}^n\delta t(\beta^{\ast}\gamma_{i}+2\alpha\gamma^{\ast}_{i}+\gamma_i+\widetilde{\beta}^{\ast}\gamma_{i } + \widetilde{\gamma}_i)\varphi_{1,i } \biggl\}= \\ & = & e^{-\frac{|\beta|^2}{2}-\frac{|\alpha|^2}{2}+\beta^*\alpha } e^{-\frac{|\widetilde{\beta}|^2}{2}-\frac{|\alpha|^2}{2}+\widetilde{\beta}^*\alpha}\times\lim_{\substack { n\rightarrow\infty \\
\delta t\rightarrow 0 } } \sqrt{\frac{(2\pi)^{2n}}{\det\mathbf{a}^{(n)}}}\;e^{-\frac{1}{2}\mathbf{j}^t(\mathbf{a}^{(n)})^{-1}\mathbf{j}}= \\ & = & e^{-\frac{|\beta|^2}{2}-\frac{|\alpha|^2}{2}+\beta^*\alpha } e^{-\frac{|\widetilde{\beta}|^2}{2}-\frac{|\alpha|^2}{2}+\widetilde{\beta}^*\alpha}\times\lim_{\substack { n\rightarrow\infty \\
\delta t\rightarrow 0 } } \frac{(2\pi)^n}{\sqrt{\det\mathbf{a}_1^{(n)}\det\mathbf{a}_2^{(n)}}}\ ; e^{-\frac{1}{2}\mathbf{j}_1^t(\mathbf{a}_1^{(n)})^{-1}\mathbf{j}_1-\frac{1}{2}\mathbf{j}_2^t(\mathbf{a}_2^{(n)})^{-1}\mathbf{j}_2 } \\ & = & \lambda(t)\;e^{-\frac{|\beta|^2}{2}-\frac{|\alpha|^2}{2}+\beta^*\alpha } e^{-\frac{|\widetilde{\beta}|^2}{2}-\frac{|\alpha|^2}{2}+\widetilde{\beta}^*\alpha}\times \\ & \times&e^{-\frac{k_g}{2\hbar^2}\int_0^t\!\!\int_0^tdt'dt''[\beta^{\ast}\gamma(t')+ 2\alpha\gamma^{\ast}(t')+\gamma(t')+\widetilde{\beta}^{\ast}\gamma(t')+\widetilde{\gamma}(t')]\textbf{a}_1^{-1}(t',t '' ) [ \beta^{\ast}\gamma(t'')+2\alpha\gamma^{\ast}(t'')+\gamma(t'')+\widetilde{\beta}^{\ast}\gamma(t'')+\widetilde{\gamma}(t'')]}\times\\ & \times & e^{-\frac{k_g}{2\hbar^2}\int_0^t\!\!\int_0^tdt'dt''[\beta^{\ast}\gamma(t ' ) + \gamma(t')-\widetilde{\beta}^{\ast}\gamma(t')-\widetilde{\gamma}(t')]\textbf{a}_2^{-1}(t',t '' ) [ \beta^{\ast}\gamma(t'')+\gamma(t'')-\widetilde{\beta}^{\ast}\gamma(t'')-\widetilde{\gamma}(t'')]}.\end{aligned}\ ] ] where @xmath187 and @xmath188 @xmath189 _ { ij}=\omega _ { m}^{\ast -2}\left ( \delta \tau \right ) ^{2}\left[\frac{2i\omega _ { m}^{\ast \ } } { \hbar } \frac{\delta _ { i , j}}{\delta \tau } + \frac{k_{g}}{\hbar ^{2}}\re { [ \gamma _ { i}\cdot\gamma _ { j}^{\ast } ] } \right],\qquad \left [ \mathbf{a}_{2}^{\left ( n\right ) } \right ] _ { ij}\ = \omega _ { m}^{\ast -2}\left ( \delta \tau \right ) ^{2}\ \left[-\frac{2i\omega
_ { m}^{\ast \ } } { \hbar } \frac
{ \delta _ { i , j}}{\delta \tau } + \frac{k_{g}}{\hbar ^{2}}\re{[\gamma _ { i}\cdot\gamma _ { j}^{\ast } ] } \right],\ ] ] @xmath190 _ { i}=\frac{1}{\hbar } \sqrt { k_{g}}\;\frac{\delta\tau}{\omega _ { m}^{\ast}}\left [ \left ( \beta ^{\ast } + \widetilde{\beta } ^{\ast } \right ) \gamma _ { i}+2\alpha \gamma _ { i}^{\ast } + \gamma^{\mathcal{n}_{a}\widetilde{\mathcal{n}}_{a } } _ { + } \right],\qquad \left [ \mathbf{j}_{2}^{\left ( n\right ) } \right ] _ { i}=\frac{1}{\hbar } \sqrt{k_{g}}\;\frac{\delta\tau}{\omega _ { m}^{\ast}}\left [ \left ( \beta ^{\ast } -\widetilde{\beta } ^{\ast}\right ) \gamma_{i}+\gamma_{-}^{\mathcal{n}_{a}\widetilde{\mathcal{n}}_{a}}\right];\ ] ] here with @xmath191}$ ] we indicate the real part of @xmath75 and @xmath192 we take now the continuous limit as @xmath193}^{\left ( n\right ) } & \underset{_{\substack { n\rightarrow \infty \\
\delta \tau \rightarrow 0 } } } \longrightarrow & \qquad \mathbf{a}_{1[2]}\left ( \tau ^{\prime } , \tau ^{\prime \prime } \right ) , \\ \frac{\mathbf{j}_{1}^{\left ( n\right)}}{\delta \tau}&\underset{_{\substack{n\rightarrow \infty \\
\delta \tau\rightarrow 0}}}\longrightarrow & \qquad\mathbf { j}_{1}(\tau ) = \qquad \frac{1}{\hbar\omega _ { m}^{\ast}}\sqrt{k_{g } } \left [ ( \beta^{\ast}+\widetilde{\beta } ^{\ast } ) \gamma(\tau)+2\alpha \gamma^{\ast}(\tau)+\gamma_{+}^{\mathcal{n}_{a}\widetilde{\mathcal{n } } _ { a}}(\tau)\right ] , \\ \frac{\mathbf{j}_{2}^{(n)}}{\delta \tau } & \underset{_{\substack { n\rightarrow \infty \\
\delta \tau\rightarrow 0}}}\longrightarrow & \qquad\mathbf{j } _ { 2}(\tau ) = \qquad \frac{1}{\hbar\omega
_ { m}^{\ast}}\sqrt{k_{g}}\left [ ( \beta ^{\ast } -\widetilde{\beta } ^{\ast } ) \gamma(\tau)+\gamma_{-}^ { \mathcal{n}_{a}\widetilde{\mathcal{n}}_{a}}(\tau ) \right],\end{aligned}\ ] ] @xmath194 alternatively , the function @xmath195 can also be obtained from the normalization condition @xmath196 defining now , for general functions @xmath197 and @xmath198 , @xmath199}^{f , g}\left ( \tau \right ) = \int\limits_{0}^{\tau } \int\limits_{0}^{\tau } d\tau ^{\prime } d\tau ^{\prime \prime } f\left ( \tau ^{\prime } \right ) \mathbf{a}_{1[2]}^{-1}\left ( \tau ^{\prime } , \tau ^{\prime \prime } \right ) g\left ( \tau ^{\prime \prime } \right ) , \ ] ] we get @xmath200 \mathbf{a}_{1}^{-1}\left ( \tau ^{\prime } , \tau ^{\prime \prime } \right ) \left [ \left ( \beta ^{\ast } + \widetilde{\beta } ^{\ast } \right ) \gamma \left ( \tau ^{\prime \prime } \right ) + 2\alpha \gamma ^{\ast } \left ( \tau ^{\prime \prime } \right ) + \gamma _ { + } ^{\mathcal{n}_{a}\widetilde{\mathcal{n}}_{a}}\left ( \tau ^{\prime \prime } \right ) \right ] } \times \\ & \times & e^{-\frac{1}{2\hbar ^{2}}k_{g}\int\limits_{0}^{\omega _ { m}^{\ast }
t}\int\limits_{0}^{\omega _ { m}^{\ast }
t}d\tau ^{\prime } d\tau ^{\prime \prime } \left [ \left ( \beta ^{\ast } -\widetilde{\beta } ^{\ast } \right ) \gamma \left ( \tau ^{\prime } \right ) + \gamma _ { -}^{\mathcal{n}_{a}\widetilde{\mathcal{n}}_{a}}\left ( \tau ^{\prime } \right ) \right ] \mathbf{a}_{2}^{-1}\left ( \tau ^{\prime } , \tau ^{\prime \prime } \right ) \left [ \left ( \beta ^{\ast } -\widetilde{\beta } ^{\ast } \right ) \gamma \left ( \tau ^{\prime \prime } \right ) + \gamma _ { -}^{\mathcal{n}_{a}\widetilde{\mathcal{n}}_{a}}\left ( \tau ^{\prime \prime } \right ) \right ] } = \end{aligned}\ ] ] @xmath201 finally @xmath202 + \mathcal{f}_{1}^{\gamma _ { + } , \gamma _ { + } } + \mathcal{f}_{2}^{\gamma _ { -},\gamma _ { -}}\right\ } } \times \\ & \times e^{-\frac{1}{2\hbar ^{2}}k_{g}\left\ { \mathcal{f}_{1}^{\gamma , \gamma } \left ( \beta ^{\ast } + \widetilde{\beta } ^{\ast } \right ) ^{2}+ \mathcal{f}_{2}^{\gamma , \gamma } \left ( \beta ^{\ast } -\widetilde{\beta } ^{\ast } \right ) ^{2}+2\left ( \beta ^{\ast } + \widetilde{\beta } ^{\ast } \right ) \left [ 2\alpha \mathcal{f}_{1}^{\gamma ^{\ast } , \gamma } \ + \mathcal { f}_{1}^{\gamma , \gamma _ { + } } \right ] + 2\mathcal{f}_{2}^{\gamma , \gamma _ { -}}\left ( \beta ^{\ast } -\widetilde{\beta } ^{\ast } \right ) \ \right\ } } . \end{split } \label{kappa}\ ] ] inverse operators @xmath203 have been evaluated numerically .
note that the dependence on @xmath204 is present only in the second factor ( second and third row ) , while the dependence on photon states is hidden in @xmath205 .
schrdinger state at time @xmath161 is then given by equation ( [ eqn : prsch ] ) .
integrals ( i ) and ( ii ) appearing in equation ( [ visibbility ] ) are given by @xmath206}l\left ( \beta , \beta ^{\prime } \right)h(\beta,\beta^{\prime } ) k_{1}^{10}\left ( \beta \right ) k_{3}^{00}\left ( \beta ^{\prime } \right ) \iint d^{2}\widetilde{\beta } d^{2}\widetilde{\beta } ^{\prime } k_{2}^{10 } ( \widetilde{\beta } ) k_{4}^{00 } ( \widetilde{\beta } ^{\prime } ) g(\widetilde{\beta},\widetilde{\beta } ^{\prime } ) h_{c } ( \widetilde{\beta } , \widetilde{\beta } ^{\prime } ) , \\ & & \left ( ii\right ) = \frac{1}{4\pi^4}e^{i\kappa ^{2}\left ( \omega^ * _ { m}t-\sin \omega^ * _ { m}t\right ) } \mathbb{k}^{11}\left ( t\right ) \mathbb{k } ^{01\,*}\left ( t\right)\times \\ & \times & \iint d^{2}\beta d^{2}\beta ^{\prime } e^{i\kappa\im [ \beta(1-e^{-i\omega_m^*t})]}l\left ( \beta , \beta ^{\prime } \right)h(\beta,\beta^{\prime } ) k_{1}^{11}\left ( \beta \right ) k_{3}^{01}\left ( \beta ^{\prime } \right ) \iint d^{2}\widetilde{\beta } d^{2 } \widetilde{\beta } ^{\prime } k_{2}^{11 } ( \widetilde{\beta } ) k_{4}^{01 } ( \widetilde{\beta } ^{\prime } ) \varepsilon(\widetilde{\beta } ) \delta(\widetilde{\beta}^{\prime } ) g(\widetilde{\beta},\widetilde{\beta } ^{\prime } ) h_{l } ( \widetilde { \beta } , \widetilde{\beta } ^{\prime } ) . \ ] ]
in the following we present the details of the calculation of the wigner function . let s start by computing @xmath207 defining @xmath215 & = & \frac{1}{\pi } \int d^{2}\chi \left\langle \chi
\right\vert \rho _ { m}\left ( t\right ) e^{\lambda b^{\dagger } -\lambda ^{\ast } b}\left\vert \chi \right\rangle = \\ & = & \frac{1}{4\pi ^{5}}\int d^{2}(\beta , \widetilde{\beta } , \beta ^{\prime } , \widetilde{\beta ^{\prime } } ) \int d^{2}\chi e^{-\left\vert \chi \right\vert ^{2}-\lambda ^{\ast } \chi -\frac{1}{2}\left\vert \lambda \right\vert ^{2}}\biggl(\alpha _ { 1}e^{\beta _ { c}\chi ^{\ast } -\frac{1}{2}\left\vert \beta _ { c}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { c}^{\prime } \right\vert ^{2}+\beta _ { c}^{\prime ^{\ast } } \chi + \beta _ { c}^{\prime ^{\ast } } \lambda } + \\ & + & \alpha _ { 2}e^{\beta _ { l}\chi ^{\ast } -\frac{1}{2}\left\vert \beta _ { l}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { l}^{\prime } \right\vert ^{2}+\beta _ { l}^{\prime ^{\ast } } \chi + \beta _ { l}^{\prime ^{\ast } } \lambda } + \alpha _ { 3}e^{\beta _ { c}\chi ^{\ast } -\frac{1}{2}\left\vert \beta _ { c}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { l}^{\prime } \right\vert ^{2}+\beta _ { l}^{\prime ^{\ast } } \chi + \beta _ { l}^{\prime ^{\ast } } \lambda } + \\ & + & \alpha _ { 4}e^{\beta _ { l}\chi ^{\ast } -\frac{1}{2}\left\vert \beta _ { l}\right\vert ^{2}-\frac{1}{2}\left\vert \beta _ { c}^{\prime } \right\vert ^{2}+\beta _ { c}^{\prime ^{\ast } } \chi + \beta _ { c}^{\prime ^{\ast } } \lambda } \biggr)\end{aligned}\ ] ] w. shollkopf , j. p. toennies , science * 266 * , 1345 ( 1994 ) ; m. arndt , o. nairz , j. vos - andreae , c. keller , g. van der zouw , a. zeilinger , nature * 401 * , 680 ( 1999 ) ; m. arndt , l. hackermuller , s. uttenthaler , k. hornberger , e. reiger , b. brezger , a. zeilinger , phys .
91 * , 090408 ( 2003 ) .
j. r. friedman , v. patel , w. chen , s. k. tolpygo , j. e. lukens , nature * 406 * , 43 ( 2000 ) ; c. wan der wal , a. ter haar , f. wilhelm , r. schouten , c. harmans , t. orlando , s. lloyd , j. mooij , science * 290 * , 773 ( 2000 ) ; v. e. manucharyan , j. koch , m. brink , l. i. glazman , m. h. devoret , arxiv:0910.3039v1 .
j. i. cirac , m. lewenstein , k. molmer , p. zoller , phys .
a * 57 * , 1208 ( 1998 ) ; c. gross , t. zibold , e. nicklas , j. esteve , m. k. oberthaler , nature * 464 * , 1165 ( 2010 ) ; n. bar - gill , c. gross , g. kurizski , i. mazets , m. k. oberthaler , _ phys . rev .
lett . _ * 106 * , 120404 ( 2011 ) ; q. y. he , m. d. reid , c. gross , m. k. oberthaler , p. d. drummond , _ phys . rev
. lett . _ * 106 * , 120405 ( 2011 ) ; y. v. bludov , v. v. konotop , m. salerno , _ europhys .
lett . _ * 93 * , 30003 ( 2011 ) . | aim of this work is to calculate explicitly the result of the experiment of superposition of a mirror in the michelson photon cavities interferometric device proposed by marshall , simon , penrose and bownmeester , as expected within a recently proposed model of non - unitary self - gravity inducing localization . as for other proposals of modifications of quantum mechanics in a non - unitary sense , aimed to account for both unitary evolution and irreversible collapse , like in the famous ghirardi - rimini - weber and pearle s models
, it turns out that , for the experimental parameters proposed , no effect is detectable at all .
it is pointed out that the enhancing properties of matter granularity does not substantially change this conclusion .
parameters have also been exploratively varied in a certain range beyond the proposed values .
it is shown that within ` sensible ' parameters , that are not yet attainable within current technology , the model exhibits a peculiar signature with respect to other collapse models as far as parameters space is explored .
besides , the calculation offers a way to see non - unitary gravity at work in a quasi - realistic setting . |
the light shifts calculated in ref .
@xcite are based upon eq .
( 2 ) in ref .
the text states that the basis for eq .
( 2 ) is ` spherical ' ( i.e. , irreducible spherical tensors ) .
in fact , the basis for eq .
( 2 ) is actually a cartesian basis ( i.e. , @xmath1 ) , which we used to perform all calculations in ref .
@xcite . in eq .
( 2 ) of ref .
@xcite , @xmath2,@xmath3 and @xmath4 are the scalar , vector and tensor atomic dynamic polarizabilities defined in ref .
the dipole matrix element in ref .
@xcite is @xmath5 , as defined in ref .
however , as ref .
@xcite only considered light shifts of the ground states , the definition for @xmath6 in ref .
@xcite becomes problematic when defining @xmath6 for the excited states .
@xmath7 and @xmath8 are defined with respect to states of lower to higher energy , respectively , rather than with respect to initial and final states .
this notation is therefore ambiguous if the initial state is an excited state .
also , the counter - rotating term was not taken into account in ref .
therefore , we follow the formalism in refs .
@xcite , in which case eq .
( 2 ) of ref .
@xcite should be rewritten as follows : @xmath9 , \end{array } \label{hamiltonian}\ ] ] where the dipole matrix element in refs .
@xcite is @xmath10 , where @xmath11 are for ( lower , upper ) levels , respectively .
@xmath2 , @xmath3 , and @xmath4 of eq . include counter - rotating terms . however , in ref .
@xcite , we made an error in the definition of the vector polarizability , @xmath3 , by neglecting the rank dependence of the counter - rotating terms .
in this errata , we correct the definition of @xmath3 to incorporate the rank dependence of the counter - rotating terms in our calculations of light shifts @xcite .
the dynamic polarizabilities are then given by @xmath12 @xmath13 where @xmath14 is the rank-@xmath15 propagator defined as @xmath16 we note that the counter - rotating term gains an overall minus sign for @xmath17 in the expression for the vector propagator @xmath18 .
we have confirmed that our corrected formula for the dynamic polarizabilities ( i.e. , eqs .
( 1 - 5 ) ) is now consistent with the expressions in ref .
in addition to corrections to our formalism discussed above , we correct two numerical errors for the atomic data used to calculate the light shifts in refs . @xcite .
these errors stem from a mistake for a partial lifetime @xmath19 and from an error for a transition wavelength @xmath20 .
all values for @xmath19 and @xmath20 of ref .
@xcite were taken from tables 7.17.3 of j. mckeever s ph.d thesis @xcite .
the specific errors are as follows : 1 . the wavelength for the transition @xmath21 in table 7.2 of ref .
@xcite is listed as @xmath22 nm , whereas the correct value is @xmath23 nm .
2 . the partial lifetime for the transition @xmath24 in table 7.3 of ref .
@xcite is listed as @xmath25 @xmath26s , but instead should be @xmath27 @xmath26s . the corrected values for tables 7.17.3 of ref .
@xcite are highlighted with green color in tables [ comparisontable1 ] and [ comparisontable2 ] of this errata .
the partial lifetimes and wavelengths from table 7.17.3 of ref . @xcite were obtained from refs . @xcite with the exception of the values for @xmath28 from ref .
@xcite and @xmath29 from ref .
@xcite . to further confirm the accuracy of the corrected values from ref .
@xcite , we refer to tables v and vi of arora _ et al . _
@xcite ( heretofore , labelled as ` clark ' )
. tables [ comparisontable1 ] and [ comparisontable2 ] in this errata compare the corrected tables from ref .
@xcite to tables v vi of ref .
our values agree well with those from ref .
@xcite .
we note that all previous publications prior to ref .
@xcite from our group , including ref .
@xcite , have used the correct values in tables [ comparisontable1][comparisontable2 ] and not those from mckeever s thesis @xcite .
furthermore , these earlier results are for linearly polarized trapping fields for which the contribution of the vector term @xmath3 vanishes
. therefore , these earlier results from our group stand without revision . using the revised formalism from section i and based on the corrected numerical values in tables [ comparisontable1][comparisontable2 ] from section ii , we present corrected figures to replace those in refs .
@xcite as described in the following sections .
the figure numbering in this errata mirrors that of ref.@xcite with corrected figures given here in roman numerals . fig .
[ fig6 ] relates to magic wavelengths for the cs @xmath0 line and replaces fig .
@xcite . the magic wavelength in this note
is defined by the weighted average of the zeeman sublevels @xmath30 ( i.e. , not by @xmath2 alone ) . as in ref .
@xcite , we include a surface interaction potential of an atom with the dielectric nanofiber in our calculation of the total atomic trap potential .
the surface potential of the ground state cs atom near a planar dielectric surface can be approximated by the van der waals potential @xmath31 , where @xmath32 and @xmath33 @xcite . in figs .
[ fig7 ] , [ fig8 ] , and [ fig9 ] , the two - color evanescent trap from ref .
@xcite is constructed from a pair of counter - propagating @xmath34-polarized ( @xmath35 ) red - detuned beams ( @xmath36 mw ) at @xmath37 nm , forming an optical lattice , and a single repulsive @xmath38-polarized ( @xmath39 ) blue - detuned beam ( @xmath40 mw ) at @xmath41 nm . the sio@xmath42 tapered optical fiber has radius @xmath43 nm in the trapping region
[ fig7],[fig8 ] , [ fig9 ] replace figures of fig . 7 , 8 , 9 of ref . @xcite . for the magic , compensated trap in figs
. [ fig10 ] , [ fig11 ] , [ fig12 ] , we use a pair of counter - propagating @xmath34-polarized ( @xmath44 ) red - detuned beams ( @xmath45 mw ) at the magic wavelength @xmath46 nm .
counter - propagating , @xmath34-polarized blue - detuned beams at a second magic wavelength @xmath47 nm are used with a power @xmath48 mw . the resulting interference is averaged out by detuning the beams to @xmath49 ghz . figs .
[ fig10 ] , [ fig11 ] , [ fig12 ] replace figs .
10 , 11 , 12 of ref .
@xcite .
because of the errors described in sections i and ii , our experiment in ref .
@xcite used red- and blue - detuned beams at wavelengths @xmath50 nm and @xmath51 nm instead of the correct values of @xmath52 nm and @xmath53 nm calculated in the same fashion as fig .
[ fig6 ] but now for @xmath54 of @xmath55 to @xmath56 of @xmath57 .
[ fig13 ] shows the trapping potentials for the ground and excited states for the correct magic wavelengths of @xmath52 nm and @xmath53 nm for this transition . for the actual wavelengths @xmath50 nm and @xmath51 nm used in our experiment @xcite , fig.1 and fig .
sm5 of ref .
@xcite are here replaced by figs .
[ fig14 ] and [ fig15 ] , respectively , which incorporate the revisions described in sections i and ii .
our emphasis has been to correct the formalism ( section i ) and atomic data ( section ii and tables i , ii ) that are the basis for our calculations in refs .
recently a more extensive set of atomic data than in tables i and ii has become available @xcite .
we have confirmed that these data with our formalism in eqs .
( 1 - 5 ) reproduce figs .
( 4 , 5 ) from ref .
@xcite . however , the expanded set of atomic levels and lifetimes in ref .
@xcite lead to small differences between figs .
( 4 , 5 ) @xcite and corresponding figures computed from our tables i , ii . these differences are most pronounced around @xmath58 nm ( e.g. , our fig .
vi(a ) ) principally due to excited - state contributions up to @xmath59 , which are not included in tables i , ii . we therefore recommend that the data set from ref .
@xcite be employed for the calculation of ac stark shifts for the @xmath0 line in atomic cesium rather than the less extensive data in our tables i , ii .
30 c. lacroute _
_ new j. phys .
_ , * 14 * , 023056 ( 2012 ) .
_ , * 109 * , 033603 ( 2012 ) . i. h. deutsch and p. s. jessen _ opt . comm .
_ , * 283 * , 681 ( 2010 )
. d. a. steck _ quantum and atom optics textbook_. + http://atomoptics.uoregon.edu/ dsteck / teaching / quantum - optics/ , 2011 .
p. rosenbusch _
_ phys . rev .
a _ , * 79 * , 013404 ( 2009 ) . k. beloy , _ phd thesis _ ,
theory of the ac stark effect on the atomic hyper fine structure and applications to microwave atomic clocks , ( 2009 ) .
work / beloy _ dissertation _
2009.pdf fam le kien , p. schneeweiss , and a. rauschenbeutel arxiv:1211.2673v1 ( 2012 ) j. mckeever _ phd thesis _ ( 2004 ) + http://thesis.library.caltech.edu/2421 , 2004 .
+ in tables [ comparisontable1 ] and [ comparisontable2 ] of this errata , two errors from mckeever s thesis have been corrected and are highlighted in green , as described in section [ comp ] .
g. laplanche _
_ phys . rev .
a _ , * 30 * , 2881 ( 1984 ) .
m. fabry and j. r. cussenot _ can .
_ , * 54 * , 836 ( 1976 ) .
d. a. steck _ cesium d line data _
+ http://steck.us/alkalidata/ , 2003 .
c. j. hood _ phd thesis _ ( 2000 ) .
+ http://thesis.library.caltech.edu/3688 .
b. arora , m. s. safronova , c. w. clark _ phys .
a _ , * 76 * , 052509 ( 2007 ) .
j. mckeever _
_ phys . rev .
_ , * 90 * , 133602 ( 2003 ) .
stern _ et al . _
_ new j. phys .
_ , * 13 * , 085004(2011 ) .
e. vetsch _
_ , * 104 * , 203603 ( 2010 ) . | several errors in refs .
@xcite are corrected related to the optical trapping potentials for a state - insensitive , compensated nanofiber trap for the @xmath0 transition of atomic cesium .
section i corrects our basic formalism in ref .
@xcite for calculating dipole - force potentials .
section ii corrects erroneous values for a partial lifetime and a transition wavelength in ref .
@xcite .
sections iii and iv present corrected figures for various trapping configurations considered in refs .
@xcite and @xcite , respectively . |
this paper is concerned with the bianchi groups , i.e. the class of groups defined by psl@xmath0 of the integers of an imaginary quadratic number field @xmath3 $ ] where @xmath4 is a product of different primes .
any such group is a group of orientation - preserving isometries of hyperbolic three space @xmath5 by means of the right - coset identification @xmath6 .
we aim at computing the left hand side of the baum - connes - conjecture for these groups .
our main result is theorem [ statement ] that accomplishes the calculation in the example case @xmath7 .
+ there is vast literature on the bianchi groups @xcite and the conjecture @xcite .
the latter is concerned with the difficult problem to parametrize k - theory classes of projective modules over the reduced @xmath2-algebra of @xmath1 .
this @xmath2-algebra is defined as the closure in the operator norm of the left regular representation of @xmath1 on @xmath8 .
the conjecture asserts that the k - theory of @xmath9 is isomorphic to equivariant k - homology with @xmath1-compact supports .
we shall use the `` official '' definition @xmath10 of equivariant k - theory with @xmath1-compact supports as in @xcite , where @xmath11 is a classifying space for proper actions of @xmath1 , and @xmath12 is equivariant bivariant k - theory as defined in @xcite . in the present context of a closed subgroup @xmath1 of a semisimple lie group @xmath13 , the symmetric space
@xmath14 is a typical universal space @xmath11 @xcite , where @xmath15 is a maximal compact subgroup of @xmath13 .
+ the baum - connes conjecture applies equally to lie groups and has , in fact , been proved for all connected lie groups @xcite .
however , most often the plancherel formula allows to understand the structure of the left regular representation of lie groups , whereas the @xmath2-algebras of arithmetic groups have turned out to be very hard to understand .
one of the reasons is that for many arithmetic groups , the @xmath2-algebra is not type i , i.e. there are irreducible representations not contained in the operator ideal of compact operators
. + hence , the knowledge of the k - theory of the @xmath2-algebra yields valuable information which seems not to be accessible by more elementary or more explicit approaches .
furthermore , the baum - connes conjecture is , so far , the method of choice for computing the k - theory whenever the structure of the @xmath2-algebra is unknown .
+ however , it often turns out that the left - hand side is not immediately calculable either .
the present paper is in spirit quite close to @xcite which computes the left - hand - side for the group @xmath16 ; however , it is unknown if the assembly map is surjective in this case .
in contrast , julg and kasparov have verified the baum - connes conjecture for all discrete subgroups of @xmath17 and @xmath18 @xcite . since @xmath19
, this readily implies that the assembly map is an isomorphism for all bianchi groups .
therefore , the bianchi groups are arguably the most interesting arithmetic groups for which complete knowledge of ( the isomorphism type of ) @xmath20 is available .
+ the canonical way to understand arithmetic groups is via actions on retracts of the symmetric space , as computed by reduction theory . here
, we shall use ( a suitable subset of ) flge s cw - complex @xcite , a union of a 2-dimensional retract of hyperbolic 3-space with a countable subset of the satake ( spherical ) boundary of @xmath5 .
these boundary points of the complex are the @xmath1-orbits of so - called singular vertices of the flge complex , and such orbits are in bijection with the non - trivial elements of the class group @xcite .
the singular points are clearly visible in the visualizations of several exemplary fundamental domains in @xcite and @xcite .
contractibility of the flge complex has been shown by rahm and the author in @xcite .
+ the equivariant @xmath15-homology of bianchi groups has been calculated by rahm in @xcite .
however , he considers only the cases of trivial class number @xmath21 .
the reason for this restriction is that in these cases there are no singular points .
however , for higher class numbers the flge complex is not proper anymore , as the stabilizers of the singular points are parabolic .
an algorithmic approach for computation of the complex has been introduced , described and exploited for the computation of the integral homology by a. rahm and the author in @xcite .
furthermore , a program for the calculation of fundamental domains in the pari / gp language has been published as part of rahm s phd thesis @xcite .
+ since the presence of singular points in the flge complex leads to the action being non - proper , in these cases the flge complex is not a model for the universal classifying space @xmath11 of proper actions , and therefore not immediately suitable for calculation of equivariant @xmath15-homology for higher class numbers . moreover , the flge complex is not locally finite since there are infinitely many edges emanating from the singular points .
so , there is no obvious locally compact topology on it , leading to severe difficulties in defining an associated commutative @xmath2-algebra .
+ there is a general construction for turning arbitrary , possibly non - proper , @xmath1-cw - complexes into proper actions @xcite which could in principle be applied to the flge complex ; however , the construction is not in general cofinite and therefore difficult to use for computational purposes .
+ in the present paper , we show how to overcome these difficulties .
the approach pursued in this paper is more akin of classical @xmath2-algebraic techniques , and hence closer to the original object under study , the @xmath2-algebra @xmath9 .
the four main ingredients are : reduction theory in the form of the flge complex , the existence of the borel - serre compactification , a certain amount of kasparov theory , and topology in the form of an atiyah - hirzebruch spectral sequence .
+ let us give a short summary of the steps and results .
section [ fl ] introduces the topological objects under study .
we identify the universal cover of the borel - serre compactification of the locally symmetric space associated to @xmath1 as a universal @xmath1-space ( lemma [ isup ] ) .
one may note that for torsion - free subgroups of bianchi groups , borel - serre compactifications of the quotient @xmath22 ( which is then a manifold ) have already been mentioned in @xcite .
there is an extension [ 55 ] of @xmath2-algebras naturally associated to the inclusion of the boundary component into this universal space .
we use this to show that the equivariant @xmath15-homology of @xmath1 forms a 6-term exact sequence ( sequence [ hy2lhs ] ) with topological @xmath15-homology of a disjoint sum of tori and the kasparov k - homology of the crossed product @xmath23 .
the latter , in turn , is determined ( lemma [ iszero ] ) by the k - homology of the subset @xmath24 of the flge complex @xmath25 .
it is necessary to work with @xmath26 instead of @xmath25 in order to overcome the problem that @xmath25 is not locally finite .
the space @xmath26 is naturally endowed with a locally compact topology , so there is no problem in associating a well - behaved commutative @xmath2-algebra to it .
the goal of section [ ahs ] is to introduce the spectral sequence used to compute the k - homology of @xmath26 . in principle , the atiyah - hirzebruch spectral sequence starting from bredon homology associated to a group action on a complex is the classical tool to compute its equivariant k - homology .
that spectral sequence is explained in @xcite and was used in @xcite .
here , we have to consider a slightly more general form of the spectral sequence ( lemma [ bigone ] ) in order to be able to treat the space @xmath26 instead of @xmath25 .
this generalization is done by identifying the atiyah - hirzebruch spectral sequence for the equivariant k - homology of a complex as the schochet spectral sequence associated to a filtration by closed ideals of a @xmath2-algebra . in this case , these ideals come from the intersections of the skeleta of @xmath25 with @xmath5 . at that point ,
all technical difficulties are resolved , and we can illustrate the computation by means of the concrete example @xmath7 in section [ concrete ] .
this leads to the main theorem [ statement ] . +
the bianchi groups , although far from being amenable , are k - amenable @xcite , as they are closed subgroups of a k - amenable group , see @xcite for the proof for @xmath27 . as a consequence
, there is a kk - equivalence between the reduced and full group @xmath2-algebra .
therefore , by calculating the k - theory of the former we simultaneously calculate that of the latter . hence , the bianchi groups may also be the most interesting lattices for which complete knowledge of the k - theory of the full algebra is now available .
+ i would like to thank prof .
ralf meyer very much for several inspiring discussions on the subject , which alone made this paper possible .
i would also like to thank prof .
edgar wingender for the opportunity to complete the present work under his hood at the medical center gttingen .
briefly , the flge complex @xmath28 is defined as follows , see @xcite for details and references .
let now @xmath29 be square - free , and let @xmath30}$ ] and @xmath31 . denote the hyperbolic three - space by @xmath32 .
the space @xmath5 is a homogeneous space under the lie group @xmath33 .
the satake ( or spherical ) boundary @xmath34 of @xmath5 is @xmath35 , and the action of @xmath13 , and hence that of @xmath1 , extend continuously to actions on the boundary by mbius transformations @xcite*section 12 .
a point @xmath36 is called a _ singular point _ if for all @xmath37 , we have @xmath38 .
+ we shall denote the union of all orbits of singular points by @xmath39 .
note that the point @xmath40 is never in the orbit of a singular point .
the set @xmath39 can either be viewed as a subset of @xmath41 , or of @xmath42 , or of @xmath43 . in the first case
, @xmath39 is identified with the number field @xmath44 $ ] @xcite .
+ the set @xmath39 is a subset of the set of cusps of @xmath1 as defined in @xcite .
serre considers the set @xmath45 of rational boundary points defined as @xmath3\cup \{\infty\}$ ] as a subset of the boundary viewed as the projective space @xmath46 .
the action of @xmath1 falls into orbits which are in natural bijection with the class group , and the orbit corresponding to the trivial element is the orbit of @xmath47 , whereas the set @xmath39 is the union of all orbits corresponding to non - trivial elements of the class group .
+ furthermore , one considers the union of all hemispheres @xmath48 for any two @xmath49 with @xmath50 , as well as the `` space above the hemispheres '' @xmath51 let @xmath52 denote the modulus in @xmath53 . as a set
, @xmath54 is the union @xmath55 ( where we omit , for simplicity , the standard identification @xmath56 . )
the topology of @xmath57 is generated by the topology of @xmath5 together with the following neighborhoods of the translates @xmath58 of singular points : @xmath59 this definition makes @xmath39 a closed subset of @xmath57 .
+ there is a retraction @xmath60 from @xmath57 onto the set @xmath61 of the union @xmath39 with all @xmath1-translates of @xmath62 , i. e. there is a continuous map @xmath63 such that @xmath64 for all @xmath65 .
the set @xmath25 admits a natural structure as a cellular complex @xmath28 , such that @xmath1 acts cellularly on @xmath28 .
we shall refer to the complex thus defined , as well as to its structure as a @xmath1-subset of @xmath57 , as to the `` flge complex '' .
we shall not make use of any topology on @xmath25 , only of the structure as a subset of @xmath57 and of the combinatorial structure . since @xmath66 comprises bianchi s fundamental polyhedron ( which is called @xmath67 in @xcite ) , we have @xmath68 and hence this definition of @xmath25 coincides with flge s original definition as the @xmath1-closure of @xmath69 .
note that the topology of @xmath57 is not the one inherited from the satake ( spherical ) compactification of @xmath5 .
however , @xmath57 coincides on @xmath5 with the usual topology ; furthermore , it is path - connected , locally path - connected and simply connected @xcite*satz 1 , and contractible @xcite*lemma 8 , as is the cellular complex @xmath28 @xcite*corollary 7 . + since @xmath25 is not locally finite at the points of @xmath39 , we shall not work with @xmath25 directly , but only with the pruned intersection @xmath70 of @xmath25 with @xmath5 . on @xmath71 ,
the cellular topology coincides with the topology inherited from @xmath5 , and is locally compact ( unlike that of @xmath57 ) , whence there are no difficulties in associating a @xmath2-algebra @xmath72 .
although @xmath71 is not a complex anymore , the @xmath2-algebra still possesses a filtration by closed ideals , defined by functions that vanish on the intersections of the skeleta of @xmath25 with @xmath5 .
the associated schochet - spectral sequence will be one of the keys to the calculation of the equivariant k - homology of @xmath1 . in the following , the pruned cellular complex @xmath73 endowed with the subset topology of @xmath5
shall be denoted by @xmath26 , and its pruned skeleta by @xmath74.[defofcirc ] + it is important to note that any two edges adjacent to a common singular point in @xmath25 define half - open intervals that are _ disjoint _ from each other as subsets of @xmath26 .
furthermore , the space @xmath26 is , as subset of @xmath5 , locally compact .
this contrasts the failure of local finiteness of @xmath25 . as a warning
, one should not expect to have @xmath26 the same equivariant k - homology as the subcomplex of @xmath25 that consists of all cells that do not touch a singular point although these spaces are equivariantly homotopic .
the reason is that these homotopies are not proper since the singular points are at infinity , for the geometry of @xmath26 .
( recall that the `` compactly supported '' version @xmath75 of k - homology is only invariant under proper homotopies . ) besides the flge construction @xmath57 , we shall need another enlargment of @xmath5 , the borel - serre construction @xmath76 for the set @xmath77 of rational boundary points .
very roughly speaking , the space @xmath76 is obtained from @xmath5 by gluing copies of @xmath78 onto each @xmath79 ( whereas @xmath57 was obtained from @xmath5 by merely gluing copies of a point onto each @xmath80 ) .
let us recall the setup and notation of serre s introduction of the borel - serre boundary for linear algebraic groups of @xmath81-rank one @xcite*appendix 1 .
+ serre s notation translates as follows .
the space @xmath5 is hyperbolic space , @xmath34 is the ordinary spherical ( satake ) boundary of @xmath5 , @xmath67 is a translate of a singular point or of the point @xmath47 , the subgroup @xmath82 is its stabilizer inside @xmath13 .
we are going to view @xmath5 as a space acted upon by @xmath13 from the right ; thus , there is the natural identification @xmath83 .
the group @xmath82 is a minimal parabolic subgroup , and the singular points defined by flge naturally embed into the satake boundary .
for instance , consider the simplest choice where @xmath67 is the origin of the poincar plane .
then @xmath84 is the set of upper triangular matrices in @xmath1 , hence isomorphic to @xmath30}$ ] .
even though there are several orbits of singular points , all stabilizer groups @xmath84 are free abelian of rank two @xcite .
+ the group @xmath85 is conjugated inside @xmath86 to the group @xmath87 whence @xmath88 is isomorphic to @xmath89 , the multiplicative group of non - zero complex numbers . the space @xmath90 is defined as the union of hyperbolic space with a boundary component @xmath91 , defined as the space of rank one - tori of @xmath82 . in our cases ,
all @xmath91 are diffeomorphic to @xmath78 .
more specifically , we obtain for @xmath92 , the regular cusp at infinity of the riemann sphere : @xmath93 furthermore , there is a unique fixed point @xmath94 of the cartan involution on @xmath91 defined by @xmath15 .
this defines an iwasawa decomposition @xmath95 associated to the subgroup @xmath15 , where @xmath96 is the neutral connected component of @xmath94 , and @xmath97 denotes the ordinary multiplication in the group @xmath13 .
the group @xmath96 identifies with @xmath98 in a canonical way ( induced by the positive root of @xmath94 with respect to the order relation coming from @xmath85 ) , and for @xmath99 , the associated element in @xmath96 is denoted by @xmath100 .
these observations permit to topologize the space @xmath90 by means of the family of bijective maps @xmath101 , defined by @xmath102 where @xmath15 is a maximal compact subgroup of @xmath13 .
these maps are compatible for different choices of @xmath15 and therefore define a unique structure of manifold with boundary on @xmath90 , independent of @xmath15 . + let now @xmath45 be the set of cusps of @xmath1 , i.e. the set of orbits of singular points together with the orbits of the trivial cusp at @xmath47
. in serre s terminology , the space @xmath76 is the union @xmath103 defined by gluing the spaces @xmath90 together along their common intersection @xmath5 . hence , as a set , @xmath76 is the disjoint union @xmath104 .
serre shows that @xmath105 is compact , and called the borel - serre compactification of @xmath106 .
the space @xmath76 is its universal cover .
+ of course , it is also possible to consider a slightly smaller @xmath1-space , defined by considering only the subset @xmath39 of @xmath45 , @xmath107 in order to obtain a space that surjects onto @xmath57 .
in fact , one has there is a natural map @xmath108 defined by collapsing each boundary component to the corresponding singular point .
this map is continuous and surjective .
this is proved by checking directly that preimages of the open sets that define the topology on @xmath57 are open in @xmath109 .
the relation between the spaces @xmath110 and @xmath111 can be summarized by the equivariant diagram @xmath112&\partial{\mathcal{h}}(s)\ar[l]\ar[d]\\ { \mathcal{h}}\ar[r]\ar[u ] & \widehat{{\mathcal{h}}}. } \ ] ] the following lemma uses the notion of amenable transformation group , discussed in detail in @xcite .
[ biglemma ] let @xmath113 with @xmath114 .
then the transformation group @xmath115 is amenable .
in particular , the @xmath2-algebraic crossed product @xmath116 is nuclear and unique , i.e. the quotient map from the maximal to the reduced crossed product is an isomorphism .
+ the set @xmath45 can be replaced by any @xmath1-closed subset of @xmath45 , but we shall not need that fact . by @xcite*theorem 5.3 , it suffices to show amenability of the transformation group . + for @xmath117 , write the symmetric space as a coset space @xmath118 where @xmath15 is a maximal compact subgroup of @xmath13 .
for two closed subgroups @xmath1 and @xmath15 of a locally compact group @xmath13 , the associated transformation group @xmath119 is an amenable transformation group @xcite*example 2.7(5 ) . + for @xmath120 ,
observe that the action of @xmath1 on @xmath121 is proper in the sense that the map @xmath122 defined by @xmath123 is topologically proper because the action is free and cocompact .
any proper action defines an amenable transformation group .
in fact , the functions @xmath124 satisfy the conditions of @xcite*propositions 2.2(2 ) where @xmath125 is a continuous non - negative function on @xmath121 such that @xmath126 for all @xmath127 . + form the associated extension of maximal crossed products associated to the invariant ideal @xmath128 in @xmath129 ( for maximal crossed products , exactness is automatic ) .
we have shown that the ideal and the quotient are nuclear ( and therefore coincide with minimal crossed product ) .
nuclearity for @xmath130 then follows from the fact that nuclearity is stable under extensions @xcite .
moreover , since @xmath1 is discrete , it has property @xmath131 @xcite*example 4.4(3 ) , and therefore also the transformation groups @xmath132 and @xmath133 are amenable @xcite*theorem 5.8 .
+ for @xmath134 , write @xmath45 as the disjoint union of orbits @xmath135 ( this is a decomposition in finitely , namely @xmath4 components , where @xmath4 is the class number ) .
any crossed product splits into a direct sum over these components , so it is enough to prove the statement separately for each . on such a component
, we can apply @xcite*example 2.7(5 ) again , this time to the ambient group @xmath1 ( which is locally compact ) and the two closed subgroups @xmath1 itself and the stabilizer @xmath136 , since the orbit @xmath137 is then equal to the @xmath1-space @xmath138 , and @xmath139 is abelian , hence amenable .
lemma [ isup ] below uses the following simple criterion , well - known in the literature , for a space to be universal , i. e. to serve as a model for the classifying space for proper actions for @xmath1 . note that whereas @xmath5 is both a universal @xmath1- and @xmath13-space , the space @xmath76 is not acted upon by @xmath13 .
any free and proper @xmath1-space is universal if and only if it is @xmath140-equivariantly contractible for any finite subgroup @xmath141 .
[ isup ] for any @xmath1-closed subset @xmath142 , the universal cover @xmath76 , as constructed in @xcite*appendix 1 , of the borel - serre compactification of @xmath143 is a universal proper @xmath1-space . by construction
, the action of @xmath1 on @xmath144 is free .
moreover , for each boundary component we can choose an isomorphism of its ( non - pointwise ) stabilizer with @xmath145 such that its action on the boundary component is equivariantly homeomorphic to the action of @xmath145 on @xmath78 , hence proper .
thus , the action on @xmath121 is proper .
it ensues that @xmath76 is proper .
let @xmath140 be a finite subgroup of @xmath1 .
the quotient @xmath146 is an orbifold with boundary which has the same homotopy type as its interior @xmath147 .
the latter is contractible because @xmath5 is a universal proper @xmath1-space @xcite*section 2 .
so @xmath146 is contractible .
moreover , since @xmath76 is contractible , the contraction of @xmath146 thus obtained lifts to an @xmath140-equivariant contraction of @xmath76 whence the assertion . in fact
, the space @xmath5 itself is the typical example of a universal proper @xmath1-space , so @xmath5 and @xmath76 can both be used as models for @xmath11 and for computation of equivariant k - homology .
an isomorphism @xmath148 is induced by the inclusion map @xmath149 .
however , @xmath76 has , by construction , the advantage that it is cocompact unlike @xmath5 .
therefore , we can pass from @xmath150 to @xmath151 as follows .
@xmath152 the last isomorphism is the dual green - julg theorem @xcite*theorem 20.2.7(b ) .
since the crossed product is nuclear , every ideal is semisplit @xcite*theorem 15.8.3 .
+ throughout the paper , the reader should bear in mind that the notations @xmath153 and @xmath154 refer to the `` original '' kasparov kk - groups instead of the compact - support group @xmath150 .
( we shall prefer to work with the former . ) for a commutative @xmath2-algebra @xmath155 , the group @xmath156 is k - homology with locally finite support , rather than the usual group with compact support .
locally finite k - homology is called k - homology with compact support in @xcite .
we can now make use of 6-term sequences which are available in @xmath151 , in contrast to @xmath150 .
the space @xmath26 was defined in subsection [ defofcirc ] .
the sequences @xmath157 defined by the respective evaluation maps , are exact and of course equivariant .
note that in the latter sequence we choose to work with @xmath76 instead of @xmath109 because only @xmath76 is cocompact , whence leading to equivariant k - homology .
+ note that none of these algebras have a unit .
we can take reduced crossed products by @xmath1 .
( we have shown that they coincide with the maximal ones . )
it is well - known @xcite*rem .
14 that a @xmath1-invariant ideal @xmath158 in a coefficient @xmath1-algebra @xmath159 induces a short exact sequence of reduced crossed products @xmath160 when @xmath1 is discrete .
therefore , we arrive at a short exact sequence of reduced crossed products with the sequences [ 55 ] . these induce exact 6-term sequences in k - homology @xmath161 for this , note that all algebras occurring in extensions throughout this paper are nuclear , and therefore all extensions are semisplit , so excision in kk - theory holds .
+ for a countable discrete group @xmath1 , there is an identification of @xmath162 with @xmath163 @xcite .
+ the k - homology of the quotient @xmath2-algebras is easy to compute . to start with the second extension ,
the action of @xmath1 on @xmath144 is free and proper , so there is a strong morita equivalence @xmath164 .
let @xmath165 denote the class number of the underlying number field .
then the number of orbits of singular points is @xmath165 .
hence , the quotient space is a disjoint union of @xmath165 compact 2-tori @xcite , so we have @xmath166 hence , the second ses . of [ 55 ] allows to compute the equivariant k - homology of to compute that of @xmath76 from @xmath5 .
+ the former ses . of [ 55 ] , in turn , allows to compute that of @xmath5 from that of @xmath26 .
as we have shown in lemma [ isup ] , the space @xmath76 is a classifying space , so this will achieve the computation . + the interior of bianchi s fundamental polyhedron has trivial stabilizer @xcite .
therefore , the action on @xmath167 is free and proper so @xmath168 is morita equivalent to @xmath169 which is isomorphic to @xmath170 of an open 3-cell , i.e. @xmath171 .
therefore , @xmath172 thus , we can compute the left hand side @xmath173 via successive computation of the groups @xmath174 and @xmath175 without support condition , by the exact 6-term sequences @xmath176 & { \operatorname{k}}^0(\gamma{\ltimes}\mathscr{c}_0({\mathcal{h } } ) ) \ar[r ] & 0 \ar[d]\\ { { \mathbb{z}}}\ar[u ] & { \operatorname{k}}^1(\gamma{\ltimes}\mathscr{c}_0({\mathcal{h } } ) ) \ar[l ] & { \operatorname{k}}^1(\gamma{\ltimes}\mathscr{c}_0(x_\circ ) ) \ar[l ] } \ ] ] and , using [ tori ] , @xmath177 & { r{\operatorname{k}}^\gamma_0(\underline{e}\gamma)}\ar[r ] & k^0(\gamma{\ltimes}\mathscr{c}_0({\mathcal{h } } ) ) \ar[d]\\ k^1(\gamma{\ltimes}\mathscr{c}_0({\mathcal{h}}))\ar[u ] & { r{\operatorname{k}}^\gamma_1(\underline{e}\gamma)}\ar[l ] & \mathbb{z}^{2k } \ar[l ] } \ ] ] which paves the way to reduce the computation of the equivariant k - homology of @xmath1 to that of the @xmath15-homology of the reduced crossed product and that of the boundary tori .
the invertible in @xmath178 ( namely , kasparov s dual - dirac element ) leads to an isomorphism @xmath179 .
therefore , rewriting the 6-term sequence [ hy2lhs ] and using the fact that the assembly map is an isomorphism , we arrive at the following exact 6-term - sequence .
@xmath180 & { \operatorname{k}}_0({c^\ast_{\textit{red}}}\gamma)\ar[r ] & r_1(\gamma ) \ar[d]\\
r_0(\gamma)\ar[u ] & { \operatorname{k}}_1({c^\ast_{\textit{red}}}\gamma ) \ar[l ] & \mathbb{z}^{2k } \ar[l ] } \ ] ] where @xmath181 is the kasparov representation ring .
this is remarkable since there are rarely exact sequences available that connect k - theory and k - homology of the same algebra ( in view of @xmath182 by k - amenability . )
[ iszero ] the connecting homomorphism of the 6-term - sequence [ x2hy ] is zero , so there is an isomorphism @xmath183 and a short exact sequence @xmath184 let @xmath185 denote the interior of the bianchi fundamental polyhedron which is called @xmath67 in @xcite ( in the present manuscript , the letter @xmath67 is already occupied ) ; hence , there are homeomorphisms @xmath186 and @xmath187 because the interior of @xmath188 is trivially stabilized .
it is well known that the kk - equivalence coming from the strong morita - rieffel equivalence @xmath189 is induced by the inclusion of @xmath2-algebras @xmath190 where the first algebra is viewed as a crossed product with the trivial group , and a function on @xmath188 is viewed as a function on @xmath167 by setting it to zero outside @xmath188 .
+ consider the map of short exact sequences @xmath191 & \mathscr{c}_0(\mathcal{d})\ar[d ] \ar[r]^{= } & \mathscr{c}_0(\mathcal{d } ) \ar[r]\ar[d ] & 0 \ar[r]\ar[d ] & 0\\ 0 \ar[r ] & \gamma{\ltimes}\mathscr{c}_0({\mathcal{h}}- x_\circ ) \ar[r ] & \gamma{\ltimes}\mathscr{c}_0({\mathcal{h } } ) \ar[r ] & \gamma{\ltimes}\mathscr{c}_0(x_\circ ) \ar[r ] & 0 } \ ] ] we claim that the center vertical arrow induces a surjective map @xmath192 in k - homology ( recall that we are dealing with locally finite homology ) .
the assertion then follows .
+ consider the following diagram of @xmath193-theory groups , @xmath194\\ { \operatorname{k}}^0(\text{pt } ) \ar[ur]\ar[r ] \ar[d ] & { \operatorname{k}}^1({\mathcal{h}})\ar[d]\\ { \operatorname{k}}_0({c^\ast_{\textit{red}}}\gamma)\ar[r ] & { \operatorname{k}}_1(\gamma{\ltimes}{\mathcal{h } } ) } \ ] ] where all vertical arrows are induced by @xmath195-inclusions , the diagonal arrow is the bott isomorphism associated to the standard orientation of @xmath196 , the upper horizontal arrow is bott isomorphism , and the lower horizontal arrow is multiplication with kasparov s dual - dirac element @xmath197 .
the commutativity of the triangle is inherent in the definition of the bott element , and that of the square follows from the fact that the forgetful homomorphism @xmath198 sends @xmath199 to the bott element . +
all arrows except the two lower vertical arrows are isomorphisms ( recall that @xmath199 is invertible ) .
it follows that the dual map @xmath200 is injective , since the class of the unit in @xmath201 is non - zero and non - torsion .
it is classical that this follows from the existence of the trace map @xmath202 which sends @xmath203 $ ] to @xmath204 .
this completes the proof .
bredon homology is the main tool for computation of equivariant k - homology in similar contexts @xcite , so we refer to these references for a more thorough introduction . here
, we shall only fix the notation . + recall the usual terminology for bredon homology :
the orbit category @xmath205 of @xmath1 has an object @xmath206 for each subgroup @xmath140 of @xmath1 , and all @xmath1-maps between any two objects as morphisms .
a bredon module is a functor @xmath207 from the orbit category to abelian groups . here
, we are going to consider the bredon module @xmath208 that associates to @xmath206 the k - homology group @xmath209 .
this reduces to even degree where it is @xmath210 canonically .
let @xmath211 be a homomorphism of discrete countable groups .
the differentials @xmath212 of the atiyah - hirzebruch spectral sequence are given by left multiplication by an abstractly defined element @xmath213 .
bredon homology of a proper complex @xmath214 with respect to the family @xmath215 of finite subgroups with coefficients in the bredon module @xmath216 is defined as the homology of the cellular complex with coefficients in @xmath216 @xmath217 where the sum extending over orbit representatives . in case @xmath140 and @xmath218
are finite , @xmath208 is a functor @xmath219 , and @xmath220 defines a map @xmath221 which reduces to the usual bredon differential as described in @xcite , and in even more detail in @xcite .
specifically , the latter reference gives the information about the morphisms induced on the representation rings by the finite group inclusions occurring among stabilizers of flge simplices , as well as details about how to simplify the computations in the pari / gp language @xcite .
in fact , the complex representation ring of a finite group as the free @xmath222-module the basis of which are the irreducible characters of the group .
these irreducible characters are given by the character tables for the finite subgroups of the bianchi groups .
one has to consider all possible inclusions , and identify the said morphism ; then , this information is fed into the program bianchi.gp in order to obtain the bredon chain complex , from which we shall easily deduce the information on the modified bredon chain complex , introduced below .
lemma [ iszero ] motivates to calculate the k - homology of @xmath223 . in the usual
setting of a proper cellular complex , there is the atiyah - hirzebruch spectral sequence computing equivariant k - homology .
its @xmath224-term is the bredon complex , and its @xmath225-term is bredon homology . for the present purpose ,
we modify this setting in the following way .
note that @xmath26 is not a cellular complex .
however , there is still a natural notion of @xmath226-cell , with the difference that there are one - cells that are adjacent to only one vertex . in the cellular complex @xmath25
, there are one- and two - cells @xmath227 adjacent to singular points .
these are compact subsets of @xmath25 .
however , their intersection @xmath228 is , of course , not compact in @xmath26 .
however , this makes no difference with regard to the combinatorial structure : they are still adjacent to the same edges .
since the definition of bredon homology only takes into account the combinatorial structure , we may still write down a combinatorial bredon complex for @xmath26 , and this complex still satisfies @xmath229
. we shall refer to this complex as the _ modified bredon complex _ and accordingly to its homology as the _ modified bredon homology _ , denoted by @xmath230 with coefficients in the bredon module @xmath208 .
the following lemma says , roughly speaking , that this modification of bredon homology is not only a modification of the @xmath224 term , but that in fact this modified @xmath224-term fits into an entire modified spectral sequence whose @xmath224-term is the modified @xmath224-term .
furthermore , this modified spectral sequence computes `` the right thing '' , namely the equivariant k - homology of @xmath26 .
+ let @xmath208 denote the bredon module that associates to @xmath206 the k - homology group @xmath231 as above .
[ bigone ] there is a homological spectral sequence computing equivariant k - homology , @xmath232 where the sum extends over representatives @xmath233 of orbits of @xmath226-cells in @xmath26 , and @xmath234 is the respective stabilizer isomorphism type . it is concentrated in the first and fourth quadrants , and 2-periodic in the index @xmath235 .
its differential @xmath236 has bidegree @xmath237 .
+ the summand of the @xmath224-term belonging to the cells of @xmath26 that do not touch a singular point is equal to the usual bredon complex of these cells .
the summand of the @xmath224-term belonging to the one - cells of @xmath26 that touch a singular point has only a differential to one vertex orbit .
the summand of the @xmath224-term belonging to the two - cells of @xmath26 that touch a singular point is equal to the usual bredon complex associated to the combinatorial structure .
therefore , the @xmath225-term is @xmath230 .
+ this is a slight generalization of the atiyah - hirzebruch spectral sequence .
it is constructed as the spectral sequence associated as in @xcite to the three - step filtration by closed ideals @xmath238 of functions vanishing on the pruned skeleta . + the sub - quotient @xmath239 is a direct sum over the orbits of @xmath235-cells , and a summand corresponding to the orbit of a cell @xmath240 stabilized by the ( necessarily finite ) subgroup @xmath241 is isomorphic to @xmath242 which is strongly morita equivalent to the @xmath235-fold suspension of the complex group ring @xmath243 whose k - homology is isomorphic to @xmath244 .
this yields the atiyah - hirzebruch spectral sequence for equivariant k - homology , upon implementing the @xmath235-fold dimension shifts pertaining to the cells dimensions .
its @xmath224 computes ( and , therefore , its @xmath225-term is ) bredon homology with respect to finite subgroups @xcite , as it is considered in @xcite for computation of equivariant k - homology .
the fact that we deal with pruned skeleta introduces no peculiarities since on the level of sub - quotients it only means to pay attention to the fact that for each edge @xmath245 touching a singular point , there is only one inclusion of vertex stabilizers instead of two ( recall that each edge touches at most one singular point ) .
it is easy to visualize the modified atiyah - hirzebruch spectral sequence thus defined for a very simple toy case of a pruned cellular complex , for instance the case of a single edge acted upon by the trivial group .
one can then compare with the schochet spectral sequence associated to a half - open interval with the filtration @xmath246 , where the endpoint is the single simplex .
the latter s @xmath224-term takes the form @xmath247 in even rows , and zero in odd rows , according to a single edge and a single point .
the homology of this vanishes , in accordance with @xmath248 .
bearing this example in mind might help in understanding the slight generalization of the atiyah - hirzebruch spectral sequence to pruned cellular complexes , considered here .
roughly speaking , `` it is enough to omit the missing points from the spectral sequence '' .
let us write down how to pass , in our case where the spectral sequence is applied to the algebra @xmath223 , from @xmath249 to the k - homology using the edge homomorphisms .
the odd rows of the @xmath250-term vanish since @xmath251 of a finite - dimensional @xmath2-algebra such as a finite group s ring necessarily vanishes .
since @xmath252 , the spectral sequence is concentrated in the columns @xmath253 .
therefore , @xmath254 , and there is an isomorphism @xmath255 and a short exact sequence @xmath256
let us go through all computations for @xmath257 .
details , including a picture , of a simple fundamental domain for the action of @xmath25 are given in @xcite and @xcite*section 3.2 .
we shall stick to the notations therein .
+ recall that we set out to calculate the spectral sequence , associated to a commutative filtrated algebra , that almost comes from the skeleton filtration of a cellular complex algebra . however , there is `` one vertex orbit missing '' in @xmath26 .
so , there are only four orbits of vertices , namely those of @xmath258 , but the same numbers of edges and faces as in @xcite*section 3.2 , namely seven orbits of edges , those of @xmath259 , and three orbits of faces .
stabilizers are given in explicit form in the same references .
there is only one orbit of singular points , since the class number of @xmath260 $ ] is two . + all information on the representation rings @xmath261 of finite stabilizers of vertices @xmath227 is given in @xcite and used as explained in subsection [ bredon ] .
+ the @xmath224-term is a complex @xmath262 in even rows , and zero in odd rows .
the @xmath263-matrices are displayed in tables [ d_one ] and [ d_two ] .
their elementary divisors are easily determined ( for instance by the pari / gp software @xcite ) by calculating the smith normal forms , yielding @xmath264 and @xmath265 therefore , the modified bredon homology is @xmath266 where we write @xmath267 .
( a quick check of a necessary criterion is the correct rational euler characteristic @xmath268 . )
using [ k1 ] and [ extension ] , we obtain @xmath269 using lemma [ iszero ] , we have @xmath270 we can now state the main theorem .
it only remains to solve the 6-term extension problem [ hy2lhs ] .
recall that the class number @xmath165 is two .
the strategy for solving this 6-term problem is similar to the one used in @xcite where the analogous long term sequence in homology was computed .
as stated in the introduction , this simultaneously computes @xmath273 and @xmath274 . + analogous computations for other @xmath4 should present no fundamentally new difficulties .
however , they are beyond the scope of the present paper . + another possibly interesting question asks for the least matrix size over @xmath9 containing an idempotent that realizes any given k - theory class associated to a list of generators of the left hand side thus computed . | we compute the equivariant k - homology of the groups psl@xmath0 of imaginary quadratic integers with trivial and non - trivial class - group .
this was done before only for cases of trivial class number .
+ we rely on reduction theory in the form of the @xmath1-cw - complex defined by flge .
we show that the difficulty arising from the non - proper action of @xmath1 on this complex can be overcome by considering a natural short exact sequence of @xmath2-algebras associated to the universal cover of the borel - serre compactification of the locally symmetric space associated to @xmath1 .
we use rather elementary @xmath2-algebraic techniques including a slightly modified atiyah - hirzebruch spectral sequence as well as several 6-term sequences . +
this computes the k - theory of the reduced and full group @xmath2-algebras of the bianchi groups . |
high - frequency quasi - periodic oscillations ( qpo ) in neutron - star x - ray binaries often appear in pairs , with frequencies ranging from a few hundred hz to more than 1 khz ( see van der klis 2006 for a review ) .
these so - called khz qpos , which were discovered with the rossi x - ray timing explorer ( rxte ) , provide a probe into the accretion flow very close to the compact objects and can possibly serve as a tool to observe effects of general relativity .
their frequencies are strongly correlated with other timing and spectral features ( see ford & van der klis 1998 ; psaltis , belloni & van der klis 1999 ; belloni , psaltis & van der klis 2002 ; van straaten et al .
in particular , the frequencies of the lower ( @xmath1 ) and upper ( @xmath2 ) khz qpo are strongly correlated with each other ( see belloni , mndez & homan 2005 , hereafter bmh05 ) .
the correlation is roughly linear ( but see zhang et al . 2006 ) and similar for all sources ( belloni , mndez & homan 2007 ) .
a number of theoretical models have been proposed for the interpretation of these oscillations , based on the identification of their frequencies with various characteristic frequencies in the inner accretion flow ( stella & vietri 1999 ; osherovich & titarchuk 1999 ; zhang 2004 ; lamb & miller 2003 ) .
however , there is still no consensus as to the origin of the qpos . in an attempt to determine the possible presence of preferred frequencies , as predicted by certain resonance models ( see e.g. abramowicz et al .
2003 ) , bmh05 investigated the distribution of the khz qpo frequencies for separate sources .
they analyzed large samples of rxte data of five systems and found indications for preferred frequencies in each of the sources .
however , it is known that the time evolution of khz qpo frequencies is similar to a random walk , i.e. the current frequency can not jump arbitrarily away from the one from a few dozen seconds before ( see van der klis 2006 ) .
bmh05 showed that a random walk in frequency , if not sampled for a long time compared to the random walk time scale , also produces apparent preferred frequencies . to obtain a data set with an improved sampling of the khz qpo frequency evolution and to determine an unbiased frequency distribution we started a monitoring campaign of one of the five sources studied by bmh05 , 4u 163653 .
this source is a member of the atoll class ( see hasinger & van der klis 1989 ) .
it contains a neutron star accreting matter from a companion star of mass @xmath3 @xmath4 with an orbital period of 3.8 hr ( see , e.g. , van paradijs et al .
type - i x - ray bursts were observed several times from 4u 163653 , sometimes with unusual shapes ( e.g. turner & breedon 1984 ; van paradijs et al .
1986 ) , or unusually long durations ( e.g. wijnands 2001 ) .
new results on the x - ray bursts have recently been presented by bhattacharyya & strohmayer ( 2005,2006 ) .
a long - term quasi - periodic variability has been reported from all - sky monitor ( asm ) data ( shih et al . 2005 ) .
two simultaneous khz qpos have been observed in 4u 163653 , with rms amplitudes increasing with photon energy ( zhang et al .
1996 ; wijnands et al .
the frequency of the lower khz qpo increases , and its amplitude decreases , with inferred @xmath5 ( di salvo , mndez & van der klis 2003 ; their data are the same as those used by bmh05 ) , and at high @xmath5 the frequency difference between the two qpos is in the range @xmath6 hz , significantly lower than half the frequency of the oscillations detected during type - i bursts ( mndez et al .
it has been shown that at low inferred mass accretion rate the khz qpo peak separation in 4u 163653 is @xmath7 hz , which exceeds half the neutron star spin frequency as inferred from burst oscillations ( jonker et al .
third , and possibly fourth , weaker khz qpos have been discovered simultaneously with the previously known khz qpo pair ( jonker et al .
2000 ) ; the new khz qpos are at frequencies of @xmath8 hz above and below the frequency of the lower khz qpo respectively , suggesting that these are sidebands to the lower khz qpo .
barret et al .
( 2005 ) reported the results of a systematic archival study of the proportional counter array ( pca ) observations of 4u 163653 concentrating on the drop in coherence of the lower khz qpo at high frequencies , as previously found by di salvo et al .
after extension of their analysis to a large sample of sources , barret et al .
( 2006 ) interpreted this as evidence for the presence of the innermost stable orbit around the neutron star .
this conclusion is under discussion ( see mndez 2006 and barret et al .
2007 ) .
4u 163653 , together with 4u 160852 and aql x-1 , also shows qpos at mhz frequencies ( revnivtsev et al .
this feature seems to occur only in a rather narrow range of mass accretion rates , corresponding to x - ray luminosities of @xmath9 ergs / s , and they disappear after x - ray bursts .
contrary to the general behavior of qpos , the rms amplitude of the mhz qpos strongly decreases with energy .
these oscillations are thought to be related to nuclear burning on the surface of the neutron star ( see revnivtsev et al .
2001 ) .
in this paper we present the results of the first 1.5 years of our ongoing monitoring campaign of 4u 163653 .
this paper is devoted to the frequency distribution of the khz qpos , while following articles will present the results on the x - ray bursts and on a more detailed analysis of the overall evolution .
in march 2005,we started to observe 4u 163653 with rxte for about 2 ks every two days . here
we include all observations from 1 march 2005 to 14 september 2006 , a total of 305 pointings . during the period 2006 march 3 through april 10 ,
daily observations were performed . for each dataset , we extracted a background - subtracted average count rate in the full pca channel band ( corresponding to @xmath0260 kev ) using standard ftools .
we only used data form proportional counter unit # 2 ( pcu2 ) as it was the only one consistently operational during all observations .
the resulting long - term light curve is shown in the top panel of fig .
an oscillation with a period of 30 - 40 days ( see shih et al .
2005 ) is evident .
a total of 49 x - ray bursts was detected ( marked with crosses in fig .
[ licu ] ) . on two occasions , 2005
april 20 and 2005 july 31 , two x - ray bursts within the same observation were detected .
a study of the x - ray bursts will be presented in a forthcoming paper .
in order to follow the x - ray color evolution of the source along the count rate oscillation ( see top panel of fig . [ licu ] , we extracted background - subtracted count rates from the _ standard2 _ channel bands a=310 and b=2040 ( numbering from 1 to 129 ) , nominally 2.55.7 kev and 7.816.4 kev respectively .
we then constructed a hardness ratio @xmath10 .
the time evolution of the hardness ratio can be seen in the bottom panel of fig .
the count - rate oscillations correspond to strong spectral changes , which appear to be out of phase with changes in hardness . in order to study this behaviour in more detail
, we constructed the hardness - intensity diagram ( hid ) , shown in fig .
[ hid ] , using the count rate and hardness ratio from fig .
the source followed a very repetitive path in this diagram , corresponding to the long - period oscillation visible in fig .
two main regions can be identified in the hid : the hardest points are all at low count rate and correspond to the island state of atoll sources .
the soft points span a rather large range of count rate while keeping a nearly constant hardness : they correspond to the banana branch . in between ,
transitional points are observed . in fig .
[ hid ] we connected consecutive observations with a line : a dotted line if the source moved from hard to soft ( leftward ) and a solid line if the source moved from soft to hard ( rightward ) .
it is clear that 4u 163653 traveled the diagram describing a counter - clockwise rotation ( although the vertical soft branch in the hid is not actually followed monotonically from top to bottom ) .
also , the density of transitional points indicate that the transition from hard to soft is faster than the reverse transition .
the 47 observations with x - ray bursts are marked with crosses in fig .
bursts are observed all over the diagram , although there are none corresponding to the hardest points of the island state ( hr @xmath11 1.25 ) and around hr=0.8 .
for each observation , we divided the light curve from the full pca energy band in intervals 16 seconds long and produced a power density spectrum ( pds ) for each stretch of data .
the time resolution used was 256@xmath12s , corresponding to a nyquist frequency of 2048 hz .
we normalized the pds according to leahy et al .
( 1983 ) , obtaining a time - frequency image .
we also averaged all pds corresponding to a single observation in order to increase statistics .
the average pds were then normalized to fractional rms ( belloni & hasinger 1990 ) . the frequency region 3001300 hz was rebinned into 128 bins ( frequency resolution @xmath08 hz ) and the resulting power spectrum was fitted using xspec 11.3 with a model consisting of a constant and one or two lorentzians . in a few cases ,
an additional zero - centered lorentzian was added to take into account a broad component . in this way we detected a significant qpo peak in 95 observations from our sample of 305 . in three cases
also a second significant qpo was detected .
since khz qpo frequencies drift on short time scales , we rebinned each time - frequency image by a factor of 4 in time ( resolution 64s ) and inspected all images for peaks showing changes in their centroid frequency .
we found such a variable - frequency peak in 53 cases out of 95 , while for the other 42 observations , no significant peak was detected at 64 s resolution .
for the observations with a visible variable peak , we fitted each 64s pds separately with a constant plus lorentzian model and recovered the frequency evolution versus time .
the fits were performed with a lorentzian with a fwhm fixed at 4 hz in order to constrain the automatic fits to narrow components . in the following ,
we refer to these as variable - peak observations . in the three cases where two peaks were detected
, we obviously have a lower and an upper khz qpo .
when only one peak is detected , the identification of the qpo has to be made in a different way . in fig .
[ q - hardness ] , for each observation we plot the average qpo centroid frequency as a function of spectral hardness .
two populations of points immediately emerge : above a hardness of 0.75 , the qpo frequency is anti - correlated with hardness , while below that threshold there is no correlation . such a segregation is known ( see mndez & van der klis 1999 , mndez et al .
1999 ) and allows us to identify the two types of qpo : lower - khz qpo at low hardness , upper - khz qpo at high hardness . with this classification of peaks
, we can examine the distribution of centroid frequencies .
the bottom panel of fig .
[ distribution ] shows the distribution of variable - peak frequencies , together with the corresponding histogram from bmh05 .
the new data are distributed in a very different way and over a wider range of frequencies .
there are broad peaks visible , but none corresponding to the main peak observed by bmh05 .
a kolmogorov - smirnov test , applied over the same frequency range gives a probability of 10@xmath13 that the two sets were drawn from the same parent population .
the dashed lines mark the region corresponding , within errors , to frequencies that would correspond to a 3:2 ratio of upper and lower khz qpo frequency , calculated with the frequency correlations in bmh05 . while no bmh05 detections fell in that region , a few do so in the new sample . while the data in fig .
[ distribution ] represent only the variable - peak frequencies , i.e. lower khz qpos which were sufficiently narrow and strong to be followed in time with a 64s resolution , it is interesting to examine the histograms of the remaining frequencies identified with a lower khz qpo , consisting of one frequency per observation , and of the frequencies identified with the upper khz qpo .
these can be seen in fig .
[ dist_broad ] . in the top panel
are all frequencies identified with as lower - khz qpo , one point per observation : in gray the ones which could also be followed on 64s time scale , in black the others . in the bottom panel are all frequencies identified with as upper - khz qpo , also one point per observation .
the upper khz frequencies are evenly distributed over the full range , while the lower khz qpos show a concentration at high frequencies .
the distributions described above derive from a set of positive detections , while a large number of observations did not yield a significant qpo .
this raises a problem of completeness that needs to be addressed .
we ran simulations to determine the percentage of lower khz qpo peaks that we could have missed as a function of centroid frequency .
for each frequency in our range , we produced 100 pds corresponding to an observation of 2ks with 3 pcus on and a source count rate equal to the minimum observed value ( see fig .
[ licu ] ) .
the rms and @xmath14 value for the qpo were derived from the curves reported by barret et al .
the result is that we are near 100% detection efficiency across the frequency range 400 - 1000 hz , with a decrease at the boundaries , reaching 30% at 250 hz and 1150 hz .
this means that our search is complete within the observed range and the frequency distributions are not biased for missing additional peaks .
using a series of short observations at regular intervals with rxte , we found that the 30 40 day quasi - periodicity seen in the rxte / asm light curve of 4u 163653 ( shih et al .
2005 ) is due to regular state transitions between the hard ( island ) and soft ( banana ) states .
these cyclic state transitions in 4u 163653 have been repeting for at least 5 years now , and they follow a hysteresis cycle similar to the one seen in black - hole binaries in outburst .
this is the first time ever that this hysteresis pattern has been observed in a persistent neutron star system .
we found that the distribution of frequencies of the lower khz qpo in 4u 163653 changed significantly compared to the distribution in the observations used by bmh05 .
this is consistent with the proposal by bmh05 that the peaks in the distribution of the khz qpo frequencies are due to the combination of a random walk of the qpo frequencies and the observational sampling .
this definitely shows that in 4u 163653 there are no preferred frequencies ( or frequency ratios ) in the distribution of khz qpo frequencies . in fig .
[ distribution ] , we compare the distribution of lower khz qpo frequencies in 4u 163653 from this campaign to the distribution of lower khz qpo frequencies for the same source taken from bmh05 , which was based on data taken before 1998 .
this comparison shows that in the new data the peaks in the distribution are less pronounced than they were in the data of bmh05 , and are not located at the same frequencies .
this shows that , at least for this source , peaks in the distribution of qpo frequencies have no special significance but , as suggested by bmh05 , they are due to a combination of the random walk of the qpo frequencies and the observational sampling .
however , there could be other reasons for the difference in frequency distributions . as mentioned above ,
the simplest interpretation is in the framework of a random walk .
sampling a random walk over a limited time span yields peaks in the distribution of frequencies even though there are no preferred frequencies .
this has already been shown by bmh05 . even assuming that the frequency variations follow a simple random walk , although we know that in reality this can not be strictly a random walk , we need to consider the fact that the sampling strategies of the earlier bmh05 data and our new data were very different . in fig .
[ time_history ] , we show the long term frequency evolution ( one point every 16 seconds ) of the data used in bmh05 compared to those of this campaign , using the same time scale on the x axis .
the bmh05 data consisted of a few long pointings separated by long intervals in time , while our regular coverage samples the evolution much better . the khz qpos in neutron - star lmxbs , both z and atoll sources , are known to show large ( @xmath15 hz ) frequency changes on time scales longer than our typical 1/2 hour exposure .
therefore , long observations or clusters of observations such as those implemented in the past are not optimal to obtain an unbiased view of the qpo evolution .
our new campaign yields a set of observations sufficiently spaced so that any correlation between consecutive measurements introduced by the random walk of the frequency is less important .
finally , there is another possible concurrent cause for the difference in distribution of frequencies between the two datasets .
the data used by bmh05 correspond to observations in the period 19961998 , when the rxte / asm data indicate that 4u 163653 was at a rather stable level of @xmath020 asm counts s@xmath16 ( shih et al . 2005 ) .
after 2000 , the overall rate decreased and the long - term oscillation started ( shih et al .
this means that the bmh05 frequencies are associated to a high stable flux , while the data from our campaign are associated to a much more variable flux level .
it is known that the frequency of the khz qpos in these systems show a positive correlation with source count rate on short ( hours ) time scales . on long time scales ,
this correlation shows shifts in count rate , producing what is known as parallel lines " ( mndez et al . 1999 ) .
a possible explanation for this phenomenon was proposed by van der klis ( 2001 ) : the frequency would depend on the current mass accretion rate normalized by its long - term average .
this would mean that in the bmh05 data , since the long - term average did not vary much , the frequencies would be observed to vary less than in our more recent data . the distribution of lower khz qpo frequencies shown in fig .
[ distribution ] gives a different view than what was found by bmh05 , but confirms their conclusions .
our new results confirm the absence of one or more preferred frequencies .
in particular , there is nothing special in the distribution of frequencies at around 650 hz , which corresponds to an upper to lower khz qpo frequency ratio of 1.5 .
our data set also provides a unique view of the long term evolution of a persistent neutron star lmxbs .
we found that the @xmath17 day periodicity seen in the rxte / asm light curve of 4u 163653 up to december 2004 ( shih et al .
2005 ) is still present in our observations , up to sepember 2006 .
[ licu ] and [ hid ] show that 4u 163653 moves regularly along the hid track , following this @xmath03040 days periodicity .
in other words , the periodicity in the light curve is caused by a regular alternation of soft ( island ) and hard ( banana ) states . in combination with rxte / asm light curves ( see shih et al .
2005 ) , our observations suggest that the state transitions have been repeating regularly for at least 5 years .
the hid of 4u 163653 shares some similarities with the corresponding diagrams for neutron star and black - hole transients ( maccarone & coppi 2003 ; belloni et al . 2005 ; homan & belloni 2005 ) .
in particular , during outbursts , these systems follow a counter - clockwise path in their hid as they switch between their hard and soft states .
for these systems , the outbursts start with a hard peak , which means that the hysteresis can also be interpreted as the soft flux lagging the hard one ( see e.g. yu et al .
the case of 4u 163653 is the same .
the hard - soft ( island - banana ) transition always takes place at a higher flux value than the returning transition to the hard ( island ) state , similar to what is observed in black - hole transients , where the hard - soft transition always takes place at a higher flux than the reverse transition .
4u 163653 is the first persistent system for which such a hysteresis cycle has been observed repeatedly .
a comparison of the asm light curve of 4u 170544 with the color analysis of olive , barret & gierliski ( 2003 ) indicates that this source might behave in a similar way .
shih et al .
( 2005 ) show a similar quasi - periodic modulation in the asm flux of ks 1731260 , but no dense sampling with the pca is available to check for state transitions .
a more detailed analysis of the hysteresis behavior will presented in a future paper .
we reported the results of a one and a half year campaign to observe 4u 1635 - 53 .
the observational strategy of a regular monitoring has the advantage of yielding an unbiased coverage of the evolution of the system along its quasi - regular 30-day oscillation .
we found that this oscillation corresponds to a regular alternation of island and banana states .
we traced the evolution of these oscillations in a hardness - intensity diagram and found that the source follows a hysteresis cycle similar to that observed from transient black - hole and neutron - star lmxbs . in many observations , we detected khz oscillations , most of which we could classify as lower khz qpo .
their frequency distribution is rather flat and it is significantly different from the frequency distribution found by bmh05 , based on previous rxte observations .
the conclusion is that there is no preferred frequency or frequency ratio in 4u 163653 .
tb acknowledges financial contribution from contract asi - inaf i/023/05/0 . | we have monitored the atoll - type neutron star low - mass x - ray binary 4u 163653 with the rossi x - ray timing explorer ( rxte ) for more than 1.5 years .
our campaign consisted of short ( @xmath02 ks ) pointings separated by two days , regularly monitoring the spectral and timing properties of the source . during the campaign we observed a clear long - term oscillation with a period of @xmath030 - 40 days , already seen in the light curves from the rxte all - sky monitor , which corresponded to regular transitions between the hard ( island ) and soft ( banana ) states .
we detected khz qpos in about a third of the observations , most of which were in the soft ( banana ) state .
the distribution of the frequencies of the peak identified as the lower khz qpo is found to be different from that previously observed in an independent data set .
this suggests that the khz qpos in the system shows no intrinsically preferred frequency .
[ firstpage ] x - rays : binaries accretion : accretion discs stars : neutron |
topological defects in crystals @xcite exhibit a rich variety of kinematic properties @xcite .
defects such as interstitials / vacancies , dislocations , and disclinations @xcite perturb the ideal perfectly periodic crystal . to date
, numerous works extensively studied the dynamics and coarsening of these defects , e.g. @xcite . in normal solids ,
dislocations are present at low concentrations and their peculiar ` glide motions ' are an important factor in determining the plastic properties of the medium .
it has long been recognized that the energy - entropy balance of topological defects is responsible ( via deconfinement ) for melting transitions @xcite . in glasses , an extensive configurational entropy of these defects @xcite ( from which ensuing restrictive slow dynamics might follow @xcite )
may be sparked .
investigations of frank kasper phases @xcite of defect lines and of related systems @xcite have flourished . in some solids , such as shape memory alloys @xcite , reducing the appearance of defects is a matter of pertinence .
the study of topological defects in liquid crystals @xcite , compounded by the very crisp imaging techniques of these systems , has evolved into a broad field .
topological defects often display sharp dynamical imprints .
perhaps the best studied such dynamical effect is the glide principle which forms the focus of the current article . throughout the years
, much work has been carried out on the classical glide motion of dislocations .
many of these works entail detailed studies of the peach - kohler forces between dislocations and peierls potentials in systems ranging from classical solids to vortex arrays to liquid crystals , e.g. @xcite .
although the discussion of our current publication is aimed towards quantum systems , our formalism may be applied with no change to many of these classical systems to rigorously derive , in the continuum limit , the well known classical glide motion ( see fig.([glidefig ] ) ) occurring in the absence of interstitials / vacancies and high order effects .
this rigorous result concerning the linear regime of continuum limit elasticity complements older works addressing detailed climb diffusion in various systems , e.g. @xcite .
we further report on _ new generalized glide equations _ in the presence of both dislocations and disclinations . as we illustrate in detail here , the basic physical ingredient leading to these generalized glide equations in solids is mass conservation which strictly restrains the dynamics .
recent years saw the extension of investigations of defect dynamics in classical media to quantum systems @xcite addressing electronic liquid crystal @xcite and other phases @xcite in which the electronic constituents favor , in a certain parameter regime , the formation of an ideal crystal like stripe pattern which may then be perturbed , through a cascade of transitions , to produce a rich variety of phases .
such stripe patterns are observed in the high temperature superconductors and other oxides @xcite and in quantum hall systems @xcite .
defects naturally alter the local electronic density of states allowing for spatial ( and temporal ) inhomogeneities of electronic properties @xcite . following the general notion that melting occurs by the condensation of topological defects , e.g. @xcite
, we may naturally anticipate that the study of topological defects is pertinent to the understanding of quantum phase transitions between various zero - temperature states .
much unlike classical physics where statics and dynamics are decoupled , in quantum systems , time and space are deeply entangled . as a consequence
, we expect the zero point _
kinematics _ of defects to be of paramount importance both to the character of the zero - temperature states and to the nature of the phase transitions . only quite recently ,
along with mukhin , two of us @xcite presented a thorough analysis of the problem of topological quantum melting of a solid in 2 + 1 space - time dimensions . at first sight , this follows the pattern of the famous nelson - halperin - young @xcite theory of classical melting in 2d . however , the quantum problem is far richer . leaning heavily on the formalism developed by kleinert @xcite ,
it was shown in @xcite that the transition from elastic solid to nematic ( or ` hexatic ' ) liquid crystal is closely related to the vortex ( or ` abelian higgs ' ) duality in 2 + 1d .
the nematic quantum fluid can be viewed as a bose - condensate of dislocations subjected to a ` dual ' higgs mechanism . in this formalism , the rigidity of the elastic medium is parameterized in terms of gauge fields ( ` stress photons ' ) . in the quantum fluid ,
the shear components acquire a higgs mass due to the presence of the dislocation superfluid .
as it turns out , such a substance is at the same time a conventional superfluid , which may now be viewed as an elastic medium which has lost its capacity to sustain shear stress .
the glide principle as known from metallurgy amounts to the observation that dislocations only move easily in the direction of their ( vectorial ) topological charge . in the nematically ordered state , the burgers vectors of the dislocations forming the condensate are oriented along the macroscopic director leading to an anisotropic screening of the shear stress @xcite . here , the dislocations coast freely only in the direction of their director .
accordingly , the elastic propagator can only be fully screened in the direction perpendicular to the director .
even more striking , it turns out @xcite that in the quantum field - theoretic setting the glide principle acquires a meaning which goes well beyond the conventional understanding of this phenomenon .
in fact , the main goal of this paper will be to further generalize these notions beyond the linearized continuum limit of @xcite , including a study of theory in higher dimensions and the incorporation of far richer defect configurations . in many standard texts , e.g.@xcite ,
the explanation of glide takes little effort .
a dislocation corresponds with a row of particles ( atoms ) ` coming to an end ' in the middle of a solid .
one way to move this entity is to cut the neighboring row at the ` altitude ' of the dislocation and consequently move over one tail to cure the cut ( fig .
[ glidefig ] ) .
the net effect is that the dislocation is displaced .
this easy mode of motion is termed ` glide ' . moving in the orthogonal direction
is not as easy .
let us try to move the dislocation ` upward ' .
this requires loose particles to make the row of particles longer or shorter ( ` interstitials ' or ` vacancies ' ) and since loose particles are energetically very costly while they move very slowly ( by diffusion ) this ` climb ' motion is strongly hindered .
estimates on ` climb ' diffusion rates are provided in e.g. @xcite .
climb is hindered to such an extent in real life situations ( e.g. pieces of steel at room temperature ) that it may be ignored altogether .
the lower dimensional glide motion of dislocations is reminiscent of dynamics in the heavily studied sliding phases @xcite in which an effective reduction of dimensionality occurs .
although this simple pictorial explanation suffices for practical matters in metallurgy , it is immediately obvious that there is more to it . dislocations are completely specified by their topological charge , the burgers vector . expressed in this language , glide is a prescription of the dynamics of defects as dictated by their own topology : dislocations can only move in the direction of their burgers vector !
how can this be ? as we review in section ii , this kinematic role of topology arises most naturally in the context of field theoretic description of plasticity .
linear elasticity does not capture interstitial / vacancy defects .
the only degrees of freedom which are admitted by continuum limit linear elasticity are phonons ( or , in dualized form , the stress photons ) . given the absence of interstitials , the secret behind glide is that normal crystals are ` non - relativistic'- solids can not displace along the time direction .
as the field theory lacks knowledge of the interstitials , this condition translates ( as we will illustrate in section iv ) directly into the glide constraint for the dislocation matter ! from the glide constraint
it immediately follows that gliding dislocations do not occupy volume .
volume is exclusively associated with interstitial matter .
accordingly , dislocations do not communicate with compressional stress and in the dislocation condensate pressure stays massless because of the glide constraint !
it better be so , because the dislocation condensate corresponds with a conventional superfluid , and conventional non - relativistic fluids carry sound .
amusingly , glide needs a preferred time direction and in a relativistic medium there is no such thing as a preferred direction .
as kleinert and one of us pointed out @xcite , since glide can not be defined in a truly relativistic medium , dislocations have to couple to compressional stress with the effect that in the relativistic quantum nematic crystal sound also acquires a mass : this incompressible state turns out to be nothing else than the space time of 2 + 1d general relativity .
the main limitation of @xcite is that they focus is exclusively on the special case of dislocation glide in 2 + 1 dimensions within the realm of the linearized theory .
our main result , presented in section iv , amounts to expressions for _ new glide - type constraints in arbitrary dimensions_. we further investigate what these constraints imply for both dislocations and disclinations via an analogous study of the more fundamental double curl defect densities .
in addition , we show in section v how these wisdoms come to an end in the full non - linear theory where dislocations can in principle climb , acting as sources and sinks of interstitials .
this paper is organized as follows : after reviewing the basics of elasticity and introducing some notations in section ii , we derive expressions for the dynamical currents for an arbitrarily dimensional medium in section iii . in section iv
we deduce a general relation describing the glide constraint as it acts on dislocation , disclination , and defect currents according to the linear theory , which is applicable in arbitrary dimensions .
this allows us to analyze the particulars of dislocation glide in 3 + 1d ( as well as in arbitrary higher dimensions ) . in high dimensions ,
dislocations are no longer particles but instead extended higher dimensional p - branes with p=1 in 3 + 1d and p=2 in 4 + 1d .
we will show that the glide constraint influences the _ center of mass motion _ of the branes while the relative ( transversal ) motions of these extended manifolds remain unconstrained .
this amounts to a precise mathematical description of the ` restricted climb ' known from metallurgy .
we subsequently turn to the disclination and defect currents and presents proofs for deep- but also disappointing traits of these currents : the mass conservation underlying dislocation glide turns into simple conservation laws .
another important property of the constraint is represented in the symmetry of the constrained current . by identifying this constraint as applying only to the unique ( rotational symmetry invariant ) ` singlet ' component of the dynamical dislocation current
, the relation of the glide constraint to compression is firmly established . in the final section ( section ( v ) )
, we lift this to the full non - linear level incorporating the presence of a lattice cut - off and interstitial matter in the field theory formalism .
deformations in ideal crystals are parameterized by displacement fields @xmath0 , where @xmath1 and @xmath2 refer to the actual position and the position in the ideal crystal of the constituent , respectively .
the action of an elastic medium contains strain , kinetic , and external potential components .
these are , in general , functionals of the displacements @xmath3 and their derivatives .
the kinetic energy density term is simply @xmath4 .
the incorporation of the kinetic energy density @xmath5 in the total lagrangian density leads to the variational equations of motion for the dynamics of @xmath3 in both quantum systems ( with bosonic displacements @xmath3 ) and simple classical ones .
much throughout , we will perform a wick rotation ( wherein the real time @xmath6 replaced by @xmath7 ) to euclidean space - time . in the resulting imaginary - time
euclidean action , the lagrangian density now becomes an `` energy '' density given by the _ sum _ of kinetic and potential parts .
the elastic portion of the lagrangian density follows from the gradient expansion in terms of the displacements . to leading order ( ` first order elasticity ' )
, the elastic energy density is @xmath8 with @xmath9 the elastic tensor whose general form is dictated by the symmetries of the medium . in the remainder of the article , we will employ latin indices to refer to spatial components , while greek indices will be reserved for space - time components .
einstein summation convention will further be assumed for repeated indices . in full generality ,
the action will further contain higher order derivatives and anharmonic strain couplings : @xmath10 corresponding to the partition function @xmath11 in the absence of singularities , the action above corresponds with the simple problem of acoustic phonons .
however , when defects are present , the theory becomes far richer . in its full non - linear form
, elasticity is closely related to einstein gravity@xcite . in the 1980 s ,
kleinert @xcite achieved a considerable progress by his recognition of the underlying gauge field - theoretic structures , employing the mathematical machinery of gauge theory to penetrate deeper into this subject than ever before ; much of this is found in his book .
this work focused primarily on the ` plasticity ' of 3d classical media , and only very recently it was extended to the problem of quantum melting in 2 + 1d@xcite .
the key notion introduced by kleinert is the dualization of the action eq.([s0 ] ) into stress variables to recognize subsequently that stress can be expressed in terms of stress gauge fields or ` stress photons ' .
topological defects then take the form of _ sources _ for the stress photons .
this follows the same pattern as the vortex ( or abelian higgs ) duality where the phase modes of the superfluid are dualized in @xmath12 gauge fields , mediating the interactions between vortex sources .
this analogy is quite close when only dislocations are in the game .
the quantum nematics addressed by two of us earlier@xcite are dual superconductors in the same sense that the quantum disordered superfluids ( mott - insulators ) can be regarded as dual superconductors , the difference being that in the dislocation condensate shear acquires a higgs mass while in the vortex condensate super - currents acquire the higgs mass .
when disclinations come into play , we need to employ rank two tensorial gauge fields ( ` double curl gauge fields ' ) . on this level ,
the correspondence with gravity become manifest@xcite .
as this construct is not widely known , we summarize the basic dualization steps in appendix a. .
the discontinuity occurs at the volterra cut ( rastered plane).,width=264 ] there is a famous theorem due to weingarten @xcite which states that the singular displacement recorded by encircling a defect line in a three - dimensional elastic medium may always be expressed as a sum of a constant vector and an antisymmetric operator acting on the radial vector @xmath13 ( see fig .
[ weingartenfig ] ) @xmath14 the reference point @xmath15 can be arbitrary , but when fixed on the defect line it defines the burgers vector @xmath16 or ` dislocation charge ' of the defect .
the three - dimensional antisymmetric operator @xmath17 can be represented as the cross product of the radial vector @xmath13 with the pseudo - vector @xmath18 : @xmath19 .
the pseudo - vector @xmath18 is the frank vector or the ` disclination charge ' of the defect .
this theorem makes immediately clear that dislocation and disclination charges are related .
for instance , a dislocation can be viewed as a bound state of a disclination and an anti - disclination , while the disclination corresponds with a ` stack ' of an infinite number of dislocations .
when addressing linear elasticity , the proper topological current turns out to be a particular combination of dislocation- and disclination currents ( the ` topological defect current'@xcite ) . employing the dualization recipe of appendix a in three classical dimensions , we can straightforwardly derive differential expressions for the dislocation and disclination densities respectively , @xmath20 where @xmath21 denotes the 3d antisymmetric ( levi - civita ) tensor . for
a defect line in a given spatial direction @xmath22 , the charges in eqs.([staticdislocation ] , [ staticdisclination ] ) , derived in appendix a via a dualization procedure , coincide with the burgers and frank vectors , respectively . to see this , consider , for instance , the edge dislocation of fig.([glidefig ] ) . here , in going around the dislocation point there is a net change of the displacement fields .
couched in standard mathematical terms , the burgers vector signaling the total variation in the displacement field is , as well known , @xmath23 in eq.([burgers ] ) , @xmath24 denotes a unit vector along the horizontal direction and @xmath25 is a planar area element within a region @xmath26 containing the dislocation point .
we recognize the argument of the last integrand in eq.([burgers ] ) as the dislocation density of eq.([staticdislocation])- @xmath27 .
similar considerations lead to the identification of eq.([staticdisclination ] ) as the angular mismatch ( the disclination ) density .
a nice construct for its visualization is the volterra cut @xcite .
needless to say , the derivation of the defect densities via integrals such as that in eq.([burgers ] ) is _ topological _ ; these forms are valid for any crystal regardless of its fundamental constituents ( whether they are classical or quantum matter of one particular statistics or another ) . in what will briefly follow , within the arena of dynamical defects ( section iii )
, we will elevate the densities of eqs.([staticdislocation ] , [ staticdisclination ] ) into the zero ( temporal ) component of euclidean space - time defect currents . as a curiosity , an old famous anecdote concerning time like defects
is provided in @xcite .
as stated earlier , dislocations and disclinations are not independent entities : a dislocation may be viewed as a disclination- antidisclination pair while a disclination is an infinite stack of dislocation lines .
thus , the densities of eqs.([staticdislocation ] , [ staticdisclination ] ) harbor redundant information .
this redundancy may be removed by fusing the dislocation and disclination densities into a more fundamental ( double curl ) topological defect density , @xmath28 within this publication , most of our studies of these densities will be within the confines of the linearized theory of elasticity .
when higher derivatives of the displacement field @xmath3 appear in the action eq.([s0 ] ) , the defect densities discussed above will fail to capture all of the relevant singularities .
the role of higher order corrections will be briefly touched upon in section v. depending on the alignment of the charge vector and the three dimensional defect line , dislocations are classified as edge ( perpendicular , @xmath29 ) or screw ( parallel , @xmath30 ) , while the disclinations may be of the wedge ( parallel , @xmath30 ) , sway ( perpendicular , @xmath29 ) or twist variants .
in this section , we will generalize the gauge theoretical description of defect currents to higher dimensions limiting ourself to the linearized level ( first order elasticity ) . as a defect drifts through the medium
, the defect charge transforms into dynamical defect current . for the particular case of the 2 + 1-dimensional medium ,
such issues were addressed for the first time by two of us @xcite . as a first extension ,
let us now consider how these currents look like in 3 + 1d .
one has now to dualize the full action including the kinetic term .
this is a straightforward extension of the 2 + 1d case , @xmath31 with the local rotation given by @xmath32 the currents are now two forms in the ` lower ' space time labels , actually keeping track of the fact that the defect lines spread out in world sheets ( or 2-branes ) in 3 + 1d space - time .
obviously , by taking one of these labels to correspond with the time direction one immediately recovers the static densities , @xmath33 and @xmath34 .
the currents with two lower spatial indices correspond with the dynamical currents : the lower indices @xmath35 refer to a current in the @xmath22 direction , of a defect line extending into the @xmath36 direction . ) .
depending on the orientation of the volterra cut with respect to the ` velocity burgers vector ' it represents slip of surfaces ( parallel , full line ) or adding / removing material ( perpendicular , dashed line ) .
for simplicity , we use the word ` slip ' for both motions.,width=264 ] in solids ( and other condensed matter systems ) , the medium can not displace in the temporal direction @xmath37 : @xmath38 . consequently , in the definition of eq.([currents3j ] ) , the`upper ' burgers - labels @xmath39 of the dislocation currents @xmath40 are purely spatial .
formally , these spatial directions can be regarded as ` flavor ' degrees of freedom @xcite .
we derive eq.([currents3 t ] ) by employing a generalization the static current of eq.([staticdisclination ] ) along with an application of the weingarten theorem ( eq.([weingartenstrain ] ) ) .
the local rotation operator eq.([omegaab ] ) used in the definition has also been generalized to deal with the in - equivalence of space and time .
when one of the indices @xmath41 is temporal , eq.([omegaab ] ) reduces to the usual@xcite spatial rotation @xmath42 .
the corresponding disclination current then carries the frank charge of the defect .
what is the physical meaning of the operator eq.([omegaab ] ) when both indices @xmath41 are spatial ? due to the constraint on temporal displacements ( @xmath43 ) , the only label that can take the temporal value @xmath37 is @xmath44 and the generalized rotation represents the velocity field @xmath45 the corresponding dynamical disclination current @xmath46 records discontinuity , a ` dislocation ' , in the velocity field .
physically , this discontinuity represents slip of two surfaces relative to each other , acting as a ` defect factory ' in the crystal ( fig .
[ figslip ] ) .
however , the weingarten theorem , which is the underneath the definitions eq.([currents3j ] - [ currents3 t ] ) , assumes that all the symmetrized strains @xmath47 are smooth everywhere .
choosing one of the indices @xmath44 or @xmath48 to correspond with time and the other to space , this strain becomes a local rotation , @xmath49 a nonzero value of the disclination current @xmath46 would mean discontinuity in the velocity field , which in turn violates the weingarten theorem . to avoid confusion between real disclination densities and these surface slips
we redefine them as follows , @xmath50 so that only the frank vector is represented by the upper label ( disclination ` flavor ' ) .
when generalized to 3 + 1d , the dualization procedure summarized in appendix a indicates that the surface slips which we left out will eventually play no role , because they do not couple to stress photons .
let us now derive the form of the topological currents in arbitrary dimensions .
the key lies , once again , in the weingarten theorem of eq.([weingartenstrain ] ) which is valid in any dimension . in @xmath51
dimensional space , the burgers vector is a @xmath51-dimensional while the disclination charge is a tensor of rank @xmath51 - 2 .
the rank of the disclination charge may be ascertained from the fact that the antisymmetric tensor @xmath17 can be written as a contraction of levi - civita symbol and another , @xmath51 - 2 dimensional antisymmetric tensor : @xmath52 .
thus , two , three and four dimensional disclinations are characterized by a frank scalar , vector and antisymmetric rank 2 tensor respectively .
the @xmath51 + 1-dimensional extension of the definitions of the currents eq.([currents3j],[currents3ta ] ) are @xmath53 both of the currents of eqs.([currentjd],[currenttd ] ) are antisymmetric in the lower indices , and represent oriented @xmath54-branes .
the disclination currents are antisymmetric also in their upper indices where the frank charge is displayed .
we reiterate that within linear elasticity the true ( double curl ) topological defect density of eq.([etad ] ) removes the redundant information given by the inter - related dislocation and disclination currents . extending the duality procedure of linearized elasticity ( appendix a ) to arbitrary dimensions
, we easily deduce that the fundamental double curl defect current within a @xmath51 + 1 dimensional medium is @xmath55 this current is antisymmetric under exchange of any two lower or any two upper indices .
this defect current is related to the dislocation and disclination currents eq.([currentjd ] - [ currenttd ] ) via @xmath56 to avoid surface slips , one of the upper indices in the definition eq.([etad ] ) must be temporal when addressing the disclination currents @xmath57 .
when all the upper indices are spatial , such components record only the derivatives of the dislocation currents of eq.([currentjd ] ) .
let us now switch gear , to derive mathematical expressions of the glide constraint in terms of the currents introduced above .
we consider an electrically charged medium ( ` bosonic wigner crystal ' ) , both because it is interesting on its own right @xcite , and also because it provides us with a convenient vehicle for the derivation of general expressions for the glide constraint .
later on , we will independently derive these glide constraints by direct _
mass conservation _ without resorting to local gauge invariance to implement it . as
mass conservation pertains to a scalar quantity , such a conservation law translates into a condition on a linear combination of the topological defect currents which is necessarily invariant under spatial rotations .
conservation laws ( equivalent to gauge invariance ) may greatly restrict the system dynamics leading to an effective reduction in the dimensionality .
we will now illustrate how this indeed transpires in solids : in linear elasticity , mass ( ` charge ' ) conservation allows only a glide motion of a dislocation .
we proceed with the treatment of a charged uniformly charged medium by employing the standard electromagnetic ( em ) gauge field formulation . as usual , in euclidean space , the electrical currents defined in terms of the constituent particles are minimally coupled to the electromagnetic potentials via @xmath58 .
the spatial current components relate to the velocity of the charged particles , so that every particle with charge @xmath59 contributes to the current as @xmath60 . for a medium with @xmath61 of such particles per unit volume
, the corresponding current density will be @xmath62 .
the time components describe the coupling between the coulomb potential @xmath63 and the charge density @xmath64 . for sufficiently small strains ,
the density is @xmath65 $ ] , subtracting the constant contribution ( compensating electrical background ) . on this linearized level
, we should add the following term to the lagrangian of quantum elasticity to describe the coupling of the em field to the electrically charged medium , @xmath66 the electromagnetic fields further have their own dynamics , described by the maxwell term @xmath67 with field strengths @xmath68 .
eq.([lem ] ) must be invariant under em gauge transformations @xmath69 , with @xmath70 being an arbitrary , non - singular scalar function .
the maxwell term is automatically gauge invariant but the demand of gauge - invariance of the minimal coupling term eq.([lem ] ) implies that the currents are locally conserved .
after partial integration , @xmath71 as @xmath72 is arbitrary , it follows that the current is conserved , @xmath73 the standard result that gauge invariance implies the conservation of the gauge charge .
let us now see how this works out considering instead the displacement . performing a gauge transformation on eq.([lem ] ) one finds immediately a gauge non invariant part , @xmath74 and gauge invariance implies that @xmath75 although this derivation is very simple , this equation is a key result of this paper . combining it with definition of the dislocation current in arbitrary dimensions eq.([currentjd ] ) and using the contraction identity , @xmath76
it follows that the requirement of conservation of electrical charge has transformed itself into a constraint acting on the dislocation current , @xmath77 this is none other than the glide constraint acting on the dislocation current in arbitrary dimensions ! at first sight this might appear as magic but it is easy to see what is behind this derivation . in order to derive eq.([lem ] ) from eq.([emmin ] ) one has to assume that the gradient expansion is well behaved , i.e. the displacements should be finite .
this is not the case when interstitials are present , because an interstitial is by definition an object which can dwell away an infinite distance from its lattice position .
hence , in the starting point eq.([lem ] ) it is implicitly assumed that the interstitial density is identical zero .
the gauge argument then shows that gauge invariance exclusively communicates with the _ non - integrability _ of the displacement fields , eq.([keyresult ] ) .
these non - integrabilities are of course nothing else than the topological currents the glide constraint is a constraint on the dislocation current . if the glide constraint was not satisfied , electrical conservation would be violated locally , i.e. electrical charges would ( dis)appear spontaneously , as if the dislocation is capable to create or destroy crystalline matter . in the absence of interstitials
this is not possible and , henceforth , dislocations can only glide .
the key is , of course , that , by default , dislocation currents are decoupled from compressional stress in the linear non - relativistic theory .
we may , indeed , equivalently derive eq.([glidecurrent ] ) without explicitly invoking em gauge invariance to arrive at eq.([keyresult ] ) . instead
, we may directly rely from the very start on _ mass conservation_- the continuity equation of the mass currents ( which , as alluded to above , is equivalent to local gauge invariance ) , @xcite @xmath78 to see how this is done directly , we compute the various mass current components @xmath79 . by simple geometrical considerations , within the linear elastic regime , the mass density @xmath80 , \label{j0.}\end{aligned}\ ] ] with @xmath81 the uniform background value : the divergence of @xmath82 ( signaling the local volume increase ) yields the negative net mass ( ` charge ' ) density variation at any point .
similarly , the spatial current density @xmath83 compounding the mass continuity equation of eq.([gliderho ] ) with the physical identification of the current ( eqs.([j0 .
] , [ ji . ] ) ) , we obtain eq.([keyresult ] ) from which eq.([glidecurrent ] ) follows . in section ( [ gclca ] ) , we will return to such a physical interpretation of the glide constraint from this perspective in order to determine corrections to the glide principle which follow from anharmonic terms .
we emphasize that the mass conservation law leading to eq.([keyresult ] ) trivially holds in any medium regardless of the underlying statistics @xcite of potential quantum systems or their dimension .
furthermore , within the linear elastic regime , such a discussion highlights the validity of eq.([keyresult ] ) and the ensuing glide equation of eq.([glidecurrent ] ) ( when interpreted as density matrix averages ) in crystals at any temperature in which strict linear order mass conservation condition is imposed on all configurations .
needless to say , as temperature is elevated , a departure occurs from such an imposed linear order condition through the enhanced appearance and diffusion of interstitials and vacancies leading to climb motions .
the restriction on the dynamics in this regime is captured by a higher order variant of eq.([glidecurrent ] ) ( eq.([glidecubic ] ) ) which will be derived later on .
let us now pause to consider what eq.([glidecurrent ] ) means physically .
in two spatial dimensions , eq.([glidecurrent ] ) implies that the dislocation currents have to be symmetric @xcite : @xmath84 .
now , consider fig .
( 1 ) . here
, the burgers vector is pointing in the horizontal x - direction implying that @xmath85 , while the glide constraint reduces to @xmath86 .
this current ( @xmath87 ) is , by its very definition , the climb current perpendicular to the burgers vector . in three and higher spatial dimensions ,
the story is less easy , the reason being that the constraint on the motion is less absolute .
this is of course known in the classic theory @xcite , but the reader might convince him / herself that making use of eq .
( [ glidecurrent ] ) the analysis is much helped as compared to the rather pain staking effort based on the ` intuitive ' arguments . in 3 + 1d , the constraint of eq.([glidecurrent ] ) becomes @xmath88 let us first consider a screw dislocation .
these correspond with dislocation currents of the form @xmath89 ( i.e. the static @xmath90 component corresponds with the orientation of the dislocation loop being parallel to the burgers vector ) .
it follows immediately that the constraint eq.([glide3d ] ) is not acting on screw dislocations and , henceforth , screw dislocations can move freely in all directions .
edge dislocations are the other extreme , corresponding with dynamical currents of the form @xmath91 where @xmath92 and @xmath93 are orthogonal .
the condition @xmath94 corresponds with glide : an edge dislocation with its loop oriented in a direction ( @xmath93 ) perpendicular to the burgers vector ( @xmath92 ) can still move freely in the direction of the burgers vector ( @xmath92 ) .
the displacement of a dislocation along the line is not a topological object and the current with two identical lower indices ( @xmath95 ) vanishes .
the glide constraint only strikes when all three labels are different .
let us consider a dislocation line extending in @xmath96 direction with burgers vector in the @xmath97 direction ( see fig .
[ fig3dglide ] ) .
the only nonzero components of the current are @xmath98 and the glide constraint becomes @xmath99 .
this automatically forbids any motion in the @xmath100 direction , that is perpendicular both to the dislocation line and its burgers vector .
the constraint is ` leaky ' due to the extended nature of the defect .
the material needed for the climb of one segment of the dislocation can be supplied by an adjacent segment .
the glide constraint has therefore only a real meaning through its integral form , @xmath101 as illustrated in fig .
[ fig3dglide ] , only the dislocation s ` center of mass ' is prohibited to move in the climb direction .
local segments of the line may still move at expense of their neighbors , effectively transporting matter along the defect line .
the ` leaky ' , locally defined , constraint of eq.([glide3dintegral ] ) corresponds with the intuitive idea of ` restricted climb ' found in the elasticity literature .
having a mathematical definition of the glide constraint at our fingertips enables us to address its incarnations in even higher dimensional systems where direct visualization is of little use . notwithstanding that such crystals are , of course , not to be found in standard condensed matter
, higher dimensional glide constraints might have implications for ` emergence theories ' of fundamental phenomena resting on elasticity theory @xcite .
let us for example look how the constraint eq.([glidecurrent ] ) acts in a 4 + 1d crystal .
a dislocation is now a 2-brane , say , a plane extending in @xmath100 and @xmath96 directions .
when its burgers vector lies in this plane , this brane is analogous to a screw dislocation in the sense that its motions are not affected by the glide constraint . the other extreme is the ` edge dislocation brane ' with a burgers vector perpendicular to the brane , say in the @xmath97 direction . in this case
the nontrivial currents correspond with @xmath102 .
as noted earlier for three - dimensional defects , topological currents do not record any motion taking place within the defect - brane but only in directions perpendicular to it .
the bottom line is that besides the static current ( density ) @xmath103 , the topological dynamical currents are @xmath104 and @xmath105 , representing dislocation glide and climb respectively . the glide constraint eq .
( [ glidecurrent ] ) forbids the latter ` in integral form ' , allowing climb of a certain brane element only at the expense of the brane volume taken by a neighboring element . up to this point we focused on dislocations , the traditional scenery for glide motions
however , as alluded to in section ii , dislocation and disclination currents are not independent .
the truly fundamental topological currents are the ( double curl ) defect currents of eq.([etad ] ) . a natural question to ask is what the form of the glide constraint is on these fundamental entities .
this task is made easy by eq.([keyresult ] ) . in what follows
, we will derive new generalized glide equations for a medium in which both dislocations and disclinations are present .
let us first consider the most trivial case , the 2 + 1d crystal . excluding surface slips , eq.([generaljt2eta ] )
becomes @xmath106 in two spatial dimensions , we immediately note from eq.([eta2d1 ] ) that the glide constraint implies that @xmath107 .
the extension to arbitrary dimensions is straightforward .
we start with the definition eq.([etad ] ) and then employ the levi - civita contraction identity to obtain @xmath108 the terms containing @xmath109 vanish .
the no - slip condition ( @xmath110 ) needs to be imposed supplanted by the glide constraint in its strain form of eq.([glidestrain ] ) .
the outcome is that the scalar spherical tensor component of the general defect current tensor of eq .
( [ generaljt2eta ] ) vanishes @xmath111 we next investigate what the glide constraint implies for disclinations .
the constraint of eq.([glideeta ] ) does not shed much light on this issue .
linear elasticity directly addresses dislocations : as we already argued , one of the upper indices in the defect current of eq.([etad ] ) needs to be temporal in order to record disclination currents . on the other hand , it is clear that disclinations somehow do know about glide because they can be viewed as an infinite stack of dislocations . as an example , consider a wedge disclination : the material added by the volterra cut is proportional to the frank charge and this should not change over time .
it turns out that the matter conservation associated with a general defect distribution is related to the volterra cutting procedure .
after the cut has been applied , no additional material should be introduced , which is the same as the requirement that all symmetrized strains and derivatives thereof are smooth everywhere , including the locus of the volterra cut .
this condition is hard wired into the proof of the generalized weingarten theorem eq.([weingartenstrain ] ) .
as we will now show , the ramification of this principle for disclinations is a conservation law . to make headway
, it is convenient to first consider euclidean lorentz - invariant @xmath112 space time .
the non - relativistic case will turn out to be a special case which directly follows from imposing the condition of the absence of time like displacements ( @xmath43 ) .
what follows rests heavily on elastic analogs of identities in differential geometry which are discussed in detail in the last part of kleinert s book@xcite .
first , we introduce the tensor @xmath113 representing the riemann - christoffel curvature tensor in the geometrical formulation of the theory of elasticity ( the @xmath114 s are crystal displacements ) .
the smoothness assumptions underlying the weingarten theorem may be expressed as , @xmath115 which in turn imply smoothness of the displacement second derivative @xmath116 cycling through the indices @xmath117 , @xmath118 , and @xmath119 of the smoothness equation eq.([verysmooth ] ) , we obtain the bianchi identity for the displacement fields , @xmath120 starting from the other end , let us analyze a candidate for a disclination conservation law , corresponding with @xmath121 .
this can be expressed in terms of the riemann tensor where @xmath122 refers to a string of indices labeled by @xmath122 , @xmath123 the first term is zero as a consequence of the contraction of the bianchi identity eq.([bianchi ] ) and the levi - civita symbol @xmath124 .
this implies that relativistic disclination currents are conserved .
the non - relativistic case is just a special case : the vanishing of time like displacements means that all upper labels are space - like ( @xmath125 ) and it follows , @xmath126 in the 2 + 1d medium , only wedge dislocations exist and the message of the conservation law eq.([conservationt ] ) is clear : the frank charge- the angle of an inserted wedge in the volterra construction- behaves as a trivially conserved scalar component of a tensor . in higher dimensions ,
the disclination current similarly behaves as a conserved tensorial current .
the defect density has information regarding both dislocations and disclinations .
the disclination conservation law has separate ramifications for the defect density .
the fundamental condition eq.([verysmooth ] ) can be directly rewritten into a conservation law for the defect density of a similar form as for the disclinations , @xmath127 this is not surprising as defect currents are proportional to to disclination currents .
this completes the picture : the conservation of matter mandates that the ` proper ' disclinations currents are also conserved .
however , the ` handicapped ' dislocation currents are not conserved a - priori ( disclinations form their sources ) but they have to pay the price that they can only glide . the static dislocations and disclinations of higher dimensional media are geometrically complex entities such as lines , sheets or @xmath128-branes with burgers vectors and frank tensors attached .
when in motion , these branes sweep the additional time dimension .
for instance , defects in two space dimensions are point - like particles turning into world - lines in space - time , in three space dimensions they form loops spreading out in strings etc . nevertheless ,
regardless of the embedding dimensionality , all of these ` branes ' share an universal property : there is a unique direction perpendicular to the brane .
this direction can be related to the dynamical defect currents by contracting it with the @xmath129 dimensional antisymmetric tensor having one index set equal to time , @xmath130 isolating the perpendicular direction @xmath22 . instead of the dislocation current @xmath131
, we could have used here also the disclination current @xmath57 , but our interest in this subsection will be in the former .
the comparison of the tensor eq.([uia ] ) with the glide constraint eq.([glidestrain ] ) illustrates that the glide condition turns into a constraint on the trace @xmath132 . for rank
2 tensors , the trace is the only invariant tensorial component of the @xmath51-dimensional orthogonal group ( @xmath133 ) , corresponding with all rotations and the inversion in the @xmath51-dimensional medium .
it follows that the constrained current is the only ` singlet ' ( scalar ) under the point group symmetries of the crystal ( point group symmetries constitute a subgroup of @xmath133 ) .
the conjugate degree of freedom must have the same symmetry and this can only be compression , the only physical entity being a singlet under @xmath133 .
we have identified the fundamental reason that glide implies the decoupling of dislocations and compressional stress@xcite .
apart from rotations , the glide constraint is also invariant under galilean space - time translations .
needless to say , the constraint does not obey euclidean lorentz ( space - time ) invariance as the time direction has a special status both in eq.([glidestrain ] ) and eq.([glidecurrent ] ) .
the origin of this lies , of course , in the definition of the crystalline displacement @xmath3 and its role in the minimal coupling eq.([lem ] ) .
the displacements are defined under the assumption that every crystalline site has an equilibrium position which implicitly wires in that their world lines extend exclusively in the temporal direction .
the crystal defect currents ( eq.([glideeta ] ) ) and the disclination conservation law of eq.([conservationt ] ) are invariant under galilean transformations while they do not respect invariance under lorentz boosts .
contrary to our rigorous ( ` glide only ' ) result concerning the linear regime of continuum elasticity , in real crystals dislocations do climb ( albeit at small rates ) .
as discussed earlier by two of us@xcite , interstitial matter will be exchanged when dislocations collide and this process releases climb motions . in what follows ,
we will rederive the glide constraint , yet now do so within a fully general framework which will enable us to address the implications of both ( higher order ) _ non - linear elasticity _ and the presence of a lattice cut - off .
higher order corrections to linear elasticity modify the original glide constraints of eq.([glidecurrent ] ) giving rise dislocation climb .
we will leave the detailed analysis of the non - linear problem to a later publication @xcite . in
what follows , we outline how the full non - linear theory ( including finite lattice size effects ) captures higher order effects such as climb . to achieve this aim , we return to the mass continuity equation invoked earlier ( in unison with eqs.([j0 .
] , [ ji . ] ) ) yet now , by examining contributions of higher order derivatives of the displacement field , we exercise far greater care in examining its ramifications .
the continuity equation of eq.([gliderho ] ) implicitly assumes that the density and current fields are functionals of local eulerian ( distorted lattice ) coordinates . on the other hand ,
the displacement , stress , and other elastic fields are functionals of substantial coordinates ( i.e. the coordinates defined relative to the undistorted lattice coordinates)- the lagrangian coordinate frame . as was briefly done earlier ( eqs.([j0 . ] , [ ji . ] ) ) , we express the local density and currents of eq.([gliderho ] ) in terms of volume and velocities as @xmath134 and @xmath135 , with @xmath81 the mass of the ideal uniform medium in a unit volume in the undistorted original medium ( i.e. the ideal background mass density ) . following a distortion , a unit volume element of the original medium now occupies a region of volume @xmath136 . with these relations in tow , eq.([gliderho ] ) reads @xmath137 this equation can be interpreted as a law governing the change of volume of the elastic medium : the change in volume ( the derivative on the left hand side ) is exclusively dictated by the motion of the boundaries- the gradient on the right hand side corresponds with a surface integral of the velocity field .
this is just a reformulation of the same basic constraint : the conservation of mass ( or electrical charge ) . throughout this paper ,
mass conservation was the primary ingredient leading to the glide constraint . to invoke the mass continuity equation in the form of eq.([glideeuler ] ) , let us express the actual atomic coordinates @xmath138 in terms of the eulerian coordinates ( henceforth denoted by @xmath139 ) by employing the identity , @xmath140 where @xmath92 is a ( lattice ) cut - off scale . simplifying , we find that eq .
( [ glideeuler ] ) may be recast as @xmath141 this is an exact expression ( entailing corrections to all order in gradients of @xmath3 ) detailing the glide constraint . retaining the leading order contributions and employing @xmath142
, we recover the ` familiar ' linearized glide constraint of eq.([glidestrain ] ) .
the exact glide constraint eq.([glidelagr ] ) can also be derived using the coupling to electromagnetism as generating functional as used in section iv .
how may we generalize this analysis so that it deals with the non - linearities up to all orders ?
a key is provided by eq.([lem ] ) . due to changes in the volume of the medium arbitrarily large lattice deformations
translate into current densities @xmath79 .
in addition , we should use substantial- instead of local coordinates such that the spatial arguments of the em fields and strains no longer coincide .
a crystalline constituent that was originally at position @xmath1 feels the em field at point @xmath143 .
hence , the full non - linear generalization of eq .
( [ lem ] ) is @xmath144 using the procedure of section iv , the exact glide constraint is straightforwardly derived from this expression .
what can we learn from this exact form of the glide constraint ?
let us specialize to the simple 2 + 1d medium .
assuming a vanishing lattice constant @xmath92 , the volume of an elementary cell is simply given as @xmath145 . inserting this in the exact glide expression eq.([glidelagr ] )
we find that it simplifies , @xmath146 the displacement derivative @xmath147 includes both regular and singular components .
when @xmath147 is small compared to unity , the linearized glide constraint of eq.([glidecurrent ] ) is recovered .
let us now analyze the conditions required for a dislocation in climbing motion to satisfy the generalized constraint of eq.([glide2d1exact ] ) .
as before , we set the burgers vector of a dislocation to @xmath148 implying @xmath149 .
a climbing dislocation then moves in the @xmath100 direction resulting in the dynamical current ( @xmath150 is the position of the dislocation ) @xmath151 assuming that the dislocation does not glide simultaneously , the current eq.([jclimb ] ) is the only nonzero dynamical dislocation current .
the exact glide constraint is then satisfied , if and only if @xmath152 clearly , so long as we keep the displacements small compared than the lattice size ( effectively prohibiting interstitials and vacancies ) , the condition of eq.([climbstrain ] ) is impossible to satisfy and climb motion is strictly forbidden .
however , as soon as particles are allowed to leave their initial positions in the crystal ( interstitial density develops ) , @xmath153 becomes finite and climb motion is liberated !
more precisely , the right hand side of the condition eq.([climbstrain ] ) represents the derivative of euler to lagrangian coordinates @xmath154 and if vanishing , it means that two neighboring sites of the original crystal are , upon deformation , occupying the same place .
this is none other than an interstitial ` event ' .
we can thus interpret the 2 + 1-dimensional exact glide constraint of eq.([glide2d1exact ] ) as formally proving the old metallurgical maxim that `` climb motion is permitted if and only if interstitials / vacancies are present '' .
this climb will have the effect that the interstitial is absorbed by the dislocation .
we reach the conclusion that higher orders in the displacements ( non - linearities , anharmonicities ) have the effect of restoring the physics of interstitials .
another natural question to ask regarding the continuum limit is whether there may be a hidden dependence on the lattice cut - off @xmath92 . to investigate this issue ,
let us see what happens to eq.([glidelagr ] ) when we only keep terms which are harmonic in the displacements , while imposing no conditions on the number of derivatives .
these have to do with the volume of the unit cell .
let us specialize to a simple hypercubic lattice with sides defined by the vectors ( @xmath155 ) , @xmath156 the corresponding volume is a determinant of the matrix eq.([uijcubic ] ) and in the harmonic approximation only the diagonal elements remain , @xmath157 using this expression for the volume , the exact equation of constraint ( eq .
( [ glidelagr ] ) ) reads @xmath158 the remarkable fact is that it is possible to collect all higher order derivatives in a simple exponent .
eq.([glidecubic ] ) amounts to a double check on the non - singular nature of the continuum limit .
it corresponds with a resummation to all orders and if there would be a non - perturbative surprise it should show up in this equation .
one sees immediately that this is not the case : in the long wavelength limit ( @xmath159 ) , the spatial derivatives on the right hand side vanish recovering the linearized equation of constraint ( eq.([glidecurrent ] ) ) .
what has happened in this paper ?
this pursuit was driven by a bewilderment with the glide phenomenon .
surely , it is very easy to explain using the intuitive arguments found in the elasticity textbook ( fig.1 ) .
however , we got puzzled when we discovered that it plays a rather central and outstanding role in the field theoretic formulation @xcite .
it is just not something that one adds by hand , but instead it is deeply wired into the structure of the theory , and it is responsible for the theory to have sensible outcomes like the ` prediction ' that non relativistic fluids carry sound . in this earlier work @xcite
, it was assumed that glide was at work as a prerequisite to find sensible outcomes . in this paper
we nailed down its origin : assuming that mass is conserved in the non - relativistic elastic medium , dislocations have to to glide as long as the ` interstitial ' ( non - topological ) displacements are small compared to the lattice constant . in the linearized theory
the glide constraint becomes exact .
we learned in section v that this is to an extent an artifact of the field theory since there is room for interstitials in the fully non - linear theory .
however , one might want to view the field theory as a platonic form , a mathematical ideal , and it is just remarkable that topology , absence of symmetry ( broken lorentz invariance ) and non - topological conservation laws ( mass conservation ) conspire to produce a kinematic law determined by a topological invariant . our key results in eqs.([glidestrain],[glidelagr ] ) translate mass conservation into the mathematical language of the topological defects .
these are quite useful because they make possible to address the workings of the constraint in circumstances where the visualization ` method ' of the elasticity books is of little use .
we illustrated this by analyzing dislocation motions in arbitrary dimensions , demonstrating that this is about restricted climb ( ` mass transfer is free inside the brane ' ) .
this is surely of some practical value : we imagine that it might be useful in the complicated , real life situations encountered by metallurgists .
it will be surely of importance in the generalization of the duality by zaanen , nussinov and mukhin@xcite from 2 + 1d to 3 + 1d , a task which is still to be accomplished and not unimportant given the fact that quantum nematics are more likely be found in a three dimensional world .
a related issue is that it still remains to be demonstrated that the lorentz - invariant version of such a 3 + 1d quantum nematic has to do with 4d einstein gravity@xcite . why are disclinations not subjected to glide constraints , while dislocations are ? using our formalism we could address this question in detail ( section iv ) . in a way ,
the outcomes are surprising .
although we find that indeed the mass conservation does not impose kinematic constraints on the disclinations , we do find that they imply a disclination conservation law .
although at first sight this might appear as obvious , it is actually quite confusing .
usually topological conservation laws are rooted in the presence of an order parameter and are thereby associated with some generalized rigidity . in this way , dislocations are about shear rigidity while disclinations are associated with curvature rigidity , one of these beautiful facts becoming clear in the geometrical formulation .
volume and mass are associated with compressional rigidity .
how can it be that this seemingly non - topological entity volume can still give rise to conservation laws pertaining to topological currents ?
we conclude with a speculation .
as emphasized throughout , the basic physical ingredient leading to the derived constrained glide dynamics in solids was mass conservation . in a formal setting , similar albeit more restrictive volume conserving diffeomorphisms ( parameterized by time ) may be directly examined via the @xmath160 algebra in two spatial dimensions and its extensions , e.g. @xcite .
we suspect that there might be a more fundamental way of casting our relations by relying on the intricacies of such symmetries of space and time .
this work was supported by the netherlands foundation for fundamental research of matter ( fom ) .
jz acknowledges the support by the fulbright foundation in the form of a senior fellowship .
support ( zn ) from the us doe under ldrd x1wx at lanl is acknowledged .
in this appendix , we sketch the dualization procedure for an action describing a three - dimensional elastic medium . we will then elaborate on the relation between the dual fields and the defect densities .
the extension of our treatment to other dimensions ( as well as to imaginary time ) is straightforward . the dislocation / disclination densities of eqs.([staticdislocation],[staticdisclination ] ) harbor a redundancy which may be removed by considering a single symmetric tensor encapsulating information about both dislocations and disclinations- the double curl topological defect density @xmath161 of eq.([etad ] ) .
linear elasticity is valid only when the displacements are minute as compared to the lattice size . within this regime , the lagrangian density of eq.([s0 ] ) is @xmath162 this is a functional of displacements @xmath163 and their first derivatives @xmath147 .
the elasticity tensor @xmath164 must adhere to symmetries inherited from the underlying lattice .
thus , as dictated by the triclinical point symmetry group in a three dimensional medium , it may contain up to 21 independent constants .
[ in general dimension @xmath128 , the elastic tensor may have up to @xmath165 independent components . ] the displacements @xmath3 have generalized momenta conjugate to them the stresses given by @xmath166 from the aforementioned symmetry of the elastic tensor @xmath164 , an ehrenfest constraint encapsulating a symmetry of the stress tensor ( @xmath167 ) rigorously follows . by a legendre transformation ,
we may obtain the hamiltonian as functional of displacements and stresses @xmath168 the dualization of the linear elastic action proceeds by a complete removal of the displacement degree of freedom .
this is achieved by splitting the displacement field into the smooth and singular components @xmath171 .
this prescription generalizes the standard abelian higgs duality , e.g. @xcite , to the elastic arena .
the singular displacement field @xmath172 harbors the defect density while the smooth field @xmath173 is everywhere regular .
much as in other dualities @xcite , the smooth part can be integrated out , leading a constraint on stresses ( momentum conservation ) @xmath174 by virtue of the ehrenfest constraint , this conservation law ( eq.([sigmaconserved ] ) ) is valid also when the contraction is performed with respect to the second index .
this property enables us to express the stress field as a double curl of the symmetric two - form @xmath175 ( also known as the elastic gauge field @xcite ) @xmath176 the physical field ( stress ) is invariant under the gauge transformation @xmath177 we are nearly done .
the hamiltonian eq.([hlinear ] ) may now be expressed in terms of the elastic gauge fields @xmath175 .
the leftover term with the singular displacement @xmath178 leads to ( after integrating by parts ) @xmath179 with the final identification of the dislocation and disclination defect densities of eqs.([staticdislocation],[staticdisclination ] ) . as usual ( e.g. , the famous minimal coupling @xmath180 ) in electrodynamics ) , such densities multiplying a gauge potential ( here @xmath181 ) serve as source terms . in eq.([lcoupling ] )
, the charges appearing in the elastic minimal coupling are topological .
the redundancy of the defect densities needed in the coupling term eq .
( [ lcoupling ] ) is obvious : from nine independent topological defect densities ( six dislocation density components @xmath182 and six disclination density components @xmath183 , connected by three relations ) , only three linear combinations couple to the physical degrees of freedom . to overcome the possible confusion ,
the coupling term may be treated differently so that the ehrenfest constraint is utilized at the very start @xmath184 after the integration by parts , the elastic gauge field @xmath175 becomes minimally coupled to the double curl topological defect density of eq.([eta3d ] ) @xmath185 amongst many others , excellent contemporary introductory textbooks on the subject include p. m. chaikin and t. c. lubensky , _ principles of condensed matter physics _ , cambridge university press ( 1995 ) , and d. nelson , _ defects and geometry in condensed matter physics _ , cambridge university press ( 2002 ) j. friedel , _ dislocations _ , pergamon press , new york , 1954 f. r. n. nabarro , _ theory of dislocations _ , clarendon , oxford , 1967 j.p .
hirth and j. lothe , _ theory of dislocations _ , mcgraw - hill , new york , 1968 h. kleinert , _ gauge fields in condensed matter _ , vol .
ii : _ stresses and defects , differential geometry , crystal defects _ ( world scientific , singapore , 1989 ) t. hatano , phys .
lett . * 93 * , 085501 ( 2004 ) a. carpio and l. l. bonilla , phys .
90 * , 135502 ( 2003 ) m. karttunen , m. haataja , k. r. elder , and m. grant , phys .
lett . * 83 * , 3518 ( 1999 ) i. s. aranson , s. scheidl , and v. m. vinokur , phys .
b * 58 * , 14541 ( 1998 ) l. onsager , nuovo cim
* 6 * , 249 ( 1949 ) , ann , n. y. acad .
sci . * 51 * , 627 ( 1949 ) w. shockley , _
letat solide _ ,
proceeding of neuvienne - consail de physique , brussels , 1952 , ed .
r. stoops , ( solvay institut de physique , brussels ) l. burakovsky and d. l. preston , solid state comm . ,
* 115 * , 341 ( 2000 ) j. m. kosterlitz and d. thouless , j. phys .
c * 5 * , l124 ( 1972 ) ; _ ibid_,*6 * , 181 ( 1973 ) t. r. kirkpatrick and p. g. wolynes , phys . rev .
a * 35 * , 3072 ( 1987 ) ; phys .
b * 36 * , 8552 ( 1987 ) ; t. r. kirkpatrick , d. thirumalai and p. g. wolynes , phys .
a * 40 * , 1045 ( 1989 ) ; t. r. kirkpatrick and d. thirumalai , phys .
* 58 * , 2091 ( 1987 ) see , e.g. , k. otsuka and c. m. wayman , _ shape memory materials _ , cambridge university press ( 1998 ) , edited by o. lavrentovich et al .
, springer ( 2001 ) ; _ dynamics and defects in liquid crystals _ , edited by p. e. cladis , and p. palffy - muhoray , taylor & francis ( 1998 ) ; p. j. collins , _ liquid crystals _ , princeton university press ( 2002 ) ; _ liquid crystals _
, authored and edited by s. kumar , cambridge university press ( 2001 ) ; p. g. de gennes and j. prost , _ the physics of liquid crystals _ , oxford university press ( 1995 ) s. limkumnerd and j. p. sethna , cond - mat/0507456 p. lipowsky , m. j. bowick , j. h. meinke , d. r. nelson , and a. r. bausch , cond - mat/050636 e. rodary , d. rodney , l. proville , y. brechet , and g. martin phys .
b * 70 * , 054111 ( 2004 ) m. c. marchetti and l. radzihovsky phys . rev .
b * 59 * , 12001 ( 1999 ) s. scheidl and v. m. vinokur phys . rev .
b * 56 * , r8522 ( 1997 ) r. j. amodeo and n. m. ghoniem phys . rev .
b * 41 * , 6958 ( 1990 ) s. nasuno , s. takeuchi , and y. sawada phys
a * 40 * , r3457 ( 1989 ) j. zaanen , z. nussinov , and s. i. mukhin , annals of physics , * 310/1 * , 181 ( 2004 ) s. a. kivelson , e. fradkin , and v. j. emery , nature * 393 * , 550 ( 1998 )
j. m. tranquada et al . ,
nature * 375 * , 51 ( 1995 ) ; s. more , c. h. chen , and s.- w. cheong , nature * 392 * , 473 ( 1998 ) m. p. lilly , k. b. cooper , j. p. eisenstein , l. n. pfeiffer , k.w .
west , phys .
lett . * 82 * , 394 ( 1999 ) jian - xin zhu , k. h. ahn , z. nussinov , t. lookman , a. v. balatsky , and a. r. bishop , phys .
lett . * 91 * , 057004 ( 2003 ) ; k. h. ahn , jian - xin zhu , z. nussinov , t. lookman , a. saxena , a. v. balatsky , and a. r. bishop , j. of superconductivity * 17 * , 7 ( 2004 ) ; k. h. ahn , t. lookman and a. r. bishop , nature * 428 * , 401 ( 2004 ) d.r . nelson and b.i .
halperin , phys .
b * 19*. 2457 ( 1979 ) a.p .
young , phys .
b * 19 * , 1855 ( 1979 ) c. s. ohern and t. c. lubensky , phys .
e * 58 * , 5948 ( 1998 ) ; l. golubovi and m. golubovi , phys .
lett . * 80 * , 4341 ( 1998);c .
s. ohern , t.c .
lubensky and j. toner , phys .
* 83 * , 2745 ( 1999 ) ; e. fradkin and s. a. kivelson , phys .
b * 59 * , 8065 ( 1999 ) ; a. h. macdonald and m. p. a. fisher , phys . rev .
b * 61 * 5724 ( 2000 ) ; v. j. emery and e. fradkin and s. a. kivelson and t. c. lubensky , phys . rev . lett . *
85 * , 2160 ( 2000 ) ; ashvin vishwanath and david carpentier , phys .
lett . * 86 * , 676 ( 2001 ) ; z. nussinov and e. fradkin , phys .
b * 71 * , 195120 ( 2005 ) ; g. weingarten , _ atti accad . naz
. lincei rc . _
[ 5 ] * 10*,57 ( 1901 ) frank was the first to pose the question of a time - like burgers vector .
a natural example is provided by earth where every point but poles may define a local time ( based on the position of sun ) .
a traveler who makes a trip around the world and notices only smooth changes in time will experience 24h time displacement .
therefore a need for the international date line that acts as a volterra cut connecting two singularities ( poles ) .
e. wigner , phys . rev . * 46 * , 1002 ( 1934 ) b. bernu , d .. m . ceperley , j. phys - condens
. mat . * 16 * , 701 ( 2004 ) v. cvetkovic , j. zaanen , z. nussinov , s. mukhin , _ in preparation _
readers not heavily versed with euclidean space , should not be alarmed by the appearance of an imaginary time ( @xmath186 ) derivative in the continuity equation of eq.([gliderho ] ) . as the euclidean space mass currents @xmath187 are linked by a trivial factor of @xmath188 to their real space counterparts ( see , e.g. , eq.([ji . ] ) ) , an innocuous global factor separates eq.([gliderho ] ) from its more familiar real time representation . | topological defects play an important role in physics of elastic media and liquid crystals .
their kinematics is determined by constraints of topological origin .
an example is the glide motion of dislocations which has been extensively studied by metallurgists . in a recent theoretical study dealing with quantum dualities associated with the quantum melting of solids
it was argued that these kinematic constraints play a central role in defining the quantum field theories of relevance to the description of quantum liquid crystalline states of the nematic type .
this forms the motivation to analyze more thoroughly the climb constraints underlying the glide motions . in the setting of continuum field theory
the climb constraint is equivalent to the condition that the density of constituent particles is vanishing and we derive a mathematical definition of this constraint which has a universal status .
this makes possible to study the kinematics of dislocations in arbitrary space - time dimensions and as an example we analyze the restricted climb associated with edge dislocations in 3 + 1d .
very generally , it can be shown that the climb constraint is equivalent to the condition that dislocations do not communicate with compressional stresses at long distances .
however , the formalism makes possible to address the full non - linear theory of relevance to short distance behaviors where violations of the constraint become possible . |
since the discovery of q0957 + 561 @xcite , about 80 gravitationally lensed quasars have been discovered @xcite .
lensed quasars are not only intriguing phenomena but also have become indispensable astronomical tools , including probes of the cosmological parameters and the structure of galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
in particular , the abundance of gravitational lenses in a well - defined source sample can be used to constrain dark energy ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
unfortunately , the largest existing survey , the cosmic lens all - sky survey ( class ; * ? ? ?
* ; * ? ? ?
* ) , contains only 22 lensed radio sources ( with a well - defined statistical sample of 13 lenses ) discovered from @xmath6 radio sources , which is still insufficient to place tight constraints on dark energy models . the sloan digital sky survey ( sdss ; * ? ? ?
* ) should lead to a significantly larger lens sample for attacking the dark energy problem .
sdss is expected to identify @xmath7 quasars spectroscopically ( e.g. , * ? ? ?
* ) and @xmath8 quasars photometrically ( e.g. , * ? ? ?
* ) , which should lead to a sample of over @xmath9 lensed quasars given a standard lensing probability of @xmath10 @xcite .
indeed , 12 new lensed quasars have been discovered from the sdss quasars so far @xcite , in addition to recovering 5 previously known lensed quasars @xcite .
we can presently construct a well - defined statistical sample of 16 lensed sdss quasars , but there remain many promising sdss lensed quasar candidates for which the necessary follow - up observations are incomplete . in this paper
, we report on the discovery of two more gravitationally lensed quasars , sdss j080623.70 + 200631.9 ( sdss j0806 + 2006 ) and sdss j135306.35 + 113804.7 ( sdss j1353 + 1138 ) .
we present imaging and spectroscopic follow - up observations with the university of hawaii 2.2-meter ( uh88 ) telescope , the w. m. keck observatory s keck i and ii telescopes , and the magellan consortium s landon clay 6.5-m ( lc6.5 m ) telescope .
we model the systems to check that their geometries are consistent with the lensing hypothesis .
the structure of this paper is as follows .
we describe our lens candidate selections from the sdss data in [ sec : sdss ] .
the follow - up observations and mass modeling of sdss j0806 + 2006 and sdss j1353 + 1138 are presented in [ sec:0806 ] and [ sec:1353 ] , respectively .
we finally present a summary and give a conclusion in [ sec : conc ] . throughout the paper we assume a cosmological model with the matter density @xmath11 , cosmological constant @xmath12 , and hubble constant @xmath13 @xcite .
the sdss is conducting a photometric and spectroscopic survey of 10,000 square degrees of the sky approximately centered on the north galactic cap using the dedicated wide - field ( @xmath14 field of view ) 2.5-m telescope @xcite at the apache point observatory in new mexico , usa .
photometric observations @xcite are made in five optical filters @xcite .
after automated data processing by the photometric pipeline @xcite , quasar and galaxy candidates are selected by the spectroscopic target selection algorithms @xcite .
spectra of these candidates are obtained according to the tiling algorithm of @xcite using a multi - fiber spectrograph covering 3800 to 9200 at a resolution of r@xmath15 .
the data are very homogeneous , with an astrometric accuracy better than about @xmath16 rms per coordinate @xcite and photometric zeropoint errors less than about 0.03 magnitude over the entire survey area @xcite .
the data have been released continuously to the public @xcite .
we selected the two objects , sdss j0806 + 2006 and sdss j1353 + 1138 , as lensed quasar candidates from the sdss spectroscopic quasar sample with the same algorithm used for the discovery of most sdss lensed quasars ( e.g. , * ? ? ? * ) .
specifically , the algorithm uses the sdss image parameters dev_l , exp_l and star_l for the likelihood that a source can be modeled as a de vaucouleurs profile , an exponential disk or a point source to quantify the structure of each quasar .
lensed quasars should be modeled poorly by all three profiles , so quasars with small values for all three likelihoods are good lens candidates ( see * ? ? ? * for more details ) .
the sdss @xmath17-band images of sdss j0806 + 2006 and sdss j1353 + 1138 are shown in figure [ fig : field08061353 ] .
both systems are clearly resolved , making them excellent lens candidates since their spectra are clearly those of @xmath18 quasars . for sdss j0806 + 2006 , the total magnitudes ( within @xmath19 aperture radius ) of the system
are @xmath20 , @xmath21 , @xmath22 , @xmath23 , and @xmath24 , in @xmath25 , @xmath26 , @xmath27 , @xmath17 , and @xmath28 , respectively . for sdss j1353 + 1138
, the total magnitudes ( @xmath29 , within @xmath30 aperture radius ) are @xmath31 , @xmath32 , @xmath33 , @xmath34 , and @xmath35 , respectively .
the redshifts of sdss j0806 + 2006 and sdss j1353 + 1138 measured from their sdss spectra are @xmath36 and @xmath37 , respectively . from the sdss spectra , with a fiber aperture of @xmath38 , we know that the components of the candidates can not have greatly dissimilar spectra .
however , we can not resolve the spectra of the individual components .
thus , while the sdss data are sufficient to identify these objects as lens candidates , additional observations are needed to confirm them as lensed quasars .
deeper and higher resolution images are needed to confirm the existence of multiple quasar images and to search for the lens galaxies , and spatially resolved spectra are needed to confirm that the quasar images have identical redshifts .
we present this evidence for sdss j0806 + 2006 in 3 and for sdss j1353 + 1138 in 4 .
we obtained @xmath39 , @xmath40 and @xmath41-band images of sdss j0806 + 2006 using the 8k mosaic ccd camera at the uh88 telescope , on 2004 december 16 .
the pixel scale of the 8k mosaic ccd camera is 0235 @xmath42 .
the exposure time was 360 sec for each band .
the typical seeing in the exposures was @xmath43 , about a half of the typical seeing size of the sdss ( @xmath44 ) .
bias - subtracted and flat - field corrected images are shown in the upper panels of figure [ fig : uh88_0806 ] .
we also obtained an @xmath45-band image using the quick infrared camera ( quirc ) of the uh88 telescope on 2005 february 21 and a @xmath46-band image using the near infrared camera ( nirc ; * ? ? ? * ) of the keck i telescope on 2005 april 24 .
the pixel scale of quirc is @xmath47 @xmath42 , and that of the nirc is @xmath48 @xmath42 .
the seeing was @xmath43 in the quirc observation , and it was @xmath49 in the nirc observation .
the total exposure time was 720 sec and 900 sec for @xmath45-band imaging and @xmath46-band imaging , respectively .
the @xmath45-band image is shown in figure [ fig : uh88_0806 ] , and the @xmath46-band image is shown in figure [ fig : keck_0806 ] . all the images ( @xmath50 ) clearly show two stellar components ; we name these two stellar components a ( eastern component ) and b ( western component ) , with a being the brighter component . to determine if there are any extended objects in between the stellar components , we subtracted point spread functions ( psfs ) from the raw @xmath51 images , adopting a nearby star as a template for the psf .
the results are shown in the lower panels of figure [ fig : uh88_0806 ] ; there is clearly residual flux between components a and b in all four images .
this object ( named g ) is quite red , with colors of @xmath52 and @xmath53 that are similar to those of an early - type galaxy at @xmath54 @xcite .
such a redshift is consistent with the redshift of a absorption line system in the spectra of the quasars ( see [ sec:0806spc ] ) and an estimate based on the faber - jackson relation ( see [ sec:0806model ] ) .
therefore , we conclude that component g is the lens galaxy . in the higher resolution
nirc @xmath46-band image ( left panel of figure [ fig : keck_0806 ] ) , we can identify component g between components a and b even before psf subtraction .
we fit this image using galfit @xcite to find that g is well - modeled by a de vaucouleurs profile of ellipticity @xmath55 and major axis position angle @xmath56 .
the astrometry and photometry of components a , b and g are summarized in table [ tab:0806 ] .
the optical images were calibrated using the standard star pg 0231 + 051 @xcite and the @xmath45-band image was calibrated using the standard star fs 21 @xcite .
we lack a photometric standard stars for the nirc @xmath46-band image , thus the data were used only for astrometry and the flux ratio constraints in the lens models .
when we fit the nirc @xmath46-band image using a model consisting of the two quasar images and the lens galaxy , we find a quasar flux ratio of 0.53 as compared to the mean flux ratio of 0.67 ( derived from fitting only the psf models ) for the optical bands that could be created by a modest amount of dust extinction or chromatic microlensing .
we report the astrometry from using galfit to simultaneously fit components a , b and g in the nirc @xmath46-band image .
the angular separation of components a and b is @xmath57 .
a spectroscopic observation of sdss j0806 + 2006 was conducted with the keck ii telescope on 2005 april 12 in @xmath58 seeing .
we used the echellette mode of the echellette spectrograph and imager ( esi ; * ? ? ?
* ; * ? ? ?
* ) with the mit - ll 2048@xmath594096 ccd camera .
the spectral range was 3900 to 11,000 at a spectral resolution of @xmath60 ( r@xmath6127000 ) .
the exposure time was 900 sec .
the @xmath62-wide slit was oriented to observe components a and b simultaneously .
the spectra of each component was extracted using the standard method ( summing the fluxes in a window around the position of each component and subtracting the sky using neighboring windows on either side of the trace ) .
we show the binned spectra in figure [ fig : spec0806 ] .
both spectra have the and emission lines at the same wavelength , with an estimated velocity difference of @xmath63 km s@xmath64 for the emission line .
we summarize the redshifts calculated from these emission lines in table [ tab:0806em ] .
in addition , the spectral energy distributions ( seds ) are also very similar .
the ratio of the spectra , which is shown in figure [ fig : spec0806 ] , is almost constant ( @xmath65 ) for a wide range of wavelengths , and it is consistent with the mean flux ratio of the optical images ( 0.67 ) .
note , however , that the emission line flux ratios appear to differ slightly from that of the continuum , suggesting that there is some microlensing of the continuum by the stars in the lens galaxy .
in addition to the quasar emission lines , we found a strong absorption line system at @xmath66 in both spectra ( see the inset of figure [ fig : spec0806 ] ) .
we also found h and k ( @xmath67 ) and absorption lines ( @xmath68 ) at @xmath66 in the spectra .
the fact that the redshift of this absorption line system is close to the estimated redshift ( see [ sec:0806img ] ) of the probable lens galaxy ( component g ) and that absorption lines are frequently associated with galaxies ( e.g. , * ? ? ?
* ) suggests that the absorption is due to the lens galaxy .
if this is the case , the lens redshift should be @xmath69 ; we adopt this value for the mass models presented in the next subsection . to explore the lensing hypothesis further , we modeled the system using the two standard mass models : a singular isothermal sphere with an external shear ( sis+shear ) model , and a singular isothermal ellipsoid ( sie ) model .
both models have eight parameters : the einstein radius @xmath70 , the shear @xmath71 or ellipticity @xmath72 and its position angle ( @xmath73 or @xmath74 ) , the position of the lens galaxy , and the position and flux of the source quasar .
we have only eight constraints ( the positions of a , b and g , and the fluxes of a and b ) , so we should be able to fit the data perfectly .
we should , however , be able to do so with sensible values for the shear or ellipticity of the models .
we used the component positions from table [ tab:0806 ] and the flux ratio of 0.53 derived from the galfit model of the nirc @xmath46-band image including the two stellar components and the lens galaxy .
lensmodel _ @xcite to determine the model parameters , with the results summarized in table [ table : model0806 ] .
the required ellipticities of @xmath75 or @xmath76 are typical of other lensed systems and roughly consistent with that of the lens galaxy ( @xmath55 ) , but there is a significant misalignment between the position angle of the major axis of the models ( @xmath77 ) and that of the lens galaxy ( @xmath78 ) .
this might imply that the system is affected by a strong external perturbation @xcite .
indeed , there are at least four galaxies within a 16@xmath79 radius of the lens ( named g14 , with the nearest galaxy only 4@xmath79 from the lens ; see table [ tab:0806gal ] ) , so the position angle of the models may be a compromise between that of the lens galaxy and the shear induced by g1 .
finally , these models predict that the time delay between the images is @xmath80days .
we can estimate the redshift of the lens galaxy based on the faber - jackson relation @xcite . from the table 3 of @xcite
, we estimate that the lens magnitude should be @xmath81 for @xmath2 , @xmath82 , and assuming @xmath83 .
while this is somewhat brighter than the observed @xmath40-band magnitude of galaxy g ( @xmath84 ) , it is roughly consistent given the @xmath85 mag scatter @xcite .
the rough agreement of this estimate further supports for a lens galaxy redshift of @xmath69 .
we obtained @xmath39 , @xmath40 , and @xmath41-band images of sdss j1353 + 1138 using the 8k mosaic ccd camera of the uh88 telescope on 2004 may 25 in @xmath58 seeing .
the exposure times were 120 sec for @xmath39 and 180 sec for @xmath40 and @xmath41 , respectively .
we also acquired an @xmath45-band image on the uh88 telescope using quirc on 2005 february 21 .
the exposure time was 720 sec , and the seeing was @xmath43 .
the images , shown in the upper panels of figure [ fig : uh88_1353 ] , clearly show two stellar components , which we named a ( northern and brighter component ) and b ( southern component ) .
although it is not obvious in the raw @xmath45-band image of figure [ fig : uh88_1353 ] , an extended object can be seen between components a and b even before the psf subtraction .
we also obtained @xmath26 and @xmath17-band images using the magellan instant camera ( magic ) at the lc6.5 m telescope on 2005 april 15 .
the pixel scale of magic is @xmath86 @xmath42 .
the seeing was @xmath87 , and the exposure time was 360 sec for @xmath26-band and 480 sec for @xmath17-band .
after subtracting psf models for the two stellar components , we clearly detect an extended red object positioned between them in all the residual images .
these residual images are shown in figure [ fig : uh88_1353 ] and [ fig : mag_1353 ] for the uh88 and magellan data , respectively .
we identify this extended object , which we denote as g , with the lens galaxy .
we fit the @xmath17-band image using galfit to find that the lens galaxy is well - modeled with a de vaucouleurs profile of ellipticity @xmath88 and major axis position angle @xmath89 .
in addition to component g , there is an additional object which we named component c near image a. it is most easily seen in the @xmath26-band image after subtracting components a and b ( see the middle panel of figure [ fig : mag_1353 ] ) , but it can also be seen at the same position in the @xmath17-band image
. there are four possible interpretations of component c. first , it could be a chance superposition of a foreground star .
it is well fit by the psf ( see the bottom panels of figure [ fig : mag_1353 ] ) , and our lens models require only components a , b and g to reproduce the system with reasonable parameter values ( see [ sec:1353model ] ) .
second , it could be an object related to the lens galaxy .
however , lens models with significant extra mass at the position of component c generally fit the data very badly .
third , it could be a third image of the quasar .
this seems unlikely since its colors are very different from components a and b ( see table [ tab:1353 ] ) .
moreover , identifying it as a quasar image makes little sense as a lens geometry since it is in the wrong place to be a central , odd image ( which should lie between components b and g ) or to be part of a four - image system in which one of the a / b / c components is an unresolved image pair .
fourth , it could be emission from the host galaxy of the source quasar .
probably only the first possibility is permitted , although we can not completely exclude the other possibilities .
we summarize astrometry and photometry of components a , b , c and g in table [ tab:1353 ] .
the optical images were calibrated using the standard stars pg 1528 + 062 @xcite and sa 107 - 351 @xcite , and the @xmath45-band image was calibrated using the standard star fs 21 @xcite
. the flux ratio of components a and b changes significantly with wavelength , from a mean flux ratio in the optical of @xmath90 to an @xmath45-band flux ratio of 0.54 .
the angular separation of components a and b is @xmath91 .
we used esi on the keck ii telescope to obtain spectra of the two components of sdss j1353 + 1138 on 2005 april 12 .
the instrumental set up was the same as in [ sec:0806spc ] and we used an exposure time of 600 sec .
the binned spectra are shown in figure [ fig : spec1353 ] .
both components have and emission lines at identical wavelengths , with a velocity difference of only @xmath92 km s@xmath64 for the emission line .
we summarize the redshifts derived from the emission lines in table [ tab:1353em ] .
the flux ratio of the spectra is nearly constant ( @xmath93 ) and consistent with that in the optical images ( @xmath94 ) .
there is , however , a slight increase in the ratio redwards of about 5300 that may be due to contamination of the spectrum of component b by emission from the lens galaxy . assuming this feature is due to the 4000 break of the lens galaxy
, it implies a lens redshift of @xmath95 .
since the colors of the lens galaxy of @xmath96 , @xmath97 , and @xmath98 are roughly consistent with those of an early - type galaxy at @xmath99 @xcite , we adopt @xmath100 for mass modeling .
the quasar spectra also contain strong absorption line systems at @xmath101 , @xmath102 , and @xmath103 , with their absorption lines .
these are unlikely to be associated with the lens galaxy , given the large difference from the lens redshift estimated from the colors , and also given the fact that many ( unlensed ) quasars show absorption .
we note that an absorption line due to a @xmath104 lens is undetectable in our spectra because it would lie beyond the atmospheric cutoff .
we modeled sdss j1353 + 1138 in the same way as in [ sec:0806model ] .
the positions of the quasars and lens galaxy are taken from table [ table : model1353 ] ( neglecting component c ) . for the flux ratio of components a and b , we adopt @xmath105 , which was derived from the galfit model of the magic @xmath17-band image .
the results are summarized in table [ table : model1353 ] .
the derived shear and ellipticity of @xmath106 and @xmath107 , are again typical for a lensed quasar system .
the derived ellipticity ( 0.15 ) appears to be significantly smaller than the ellipticity of the observed light profile ( @xmath108 ) , but this difference is commonly seen in the lensed quasar systems @xcite .
however , in addition to the disagreement of the ellipticity , there is a significant misalignment between the position angle of the models ( @xmath109 ) and that of the lens galaxy ( @xmath89 ) . while this might be due to external perturbations , there are no nearby galaxies that can easily explain the misalignment
. it could be due to component c , but we find that simple tests of lens models with mass located at the position of component c generally fail to fit the data . the time delay between a and b is predicted to be @xmath110days , assuming the lens redshift of @xmath100 . as in
[ sec:0806model ] we also estimated the lens redshift based on the faber - jackson relation .
again we used table 3 of @xcite , @xmath3 , and @xmath111 to estimate that the lens magnitude should be @xmath112 , assuming @xmath113 .
this estimate is in rough agreement with the observed @xmath40-band magnitude of galaxy g ( @xmath114 ) , again supporting for a lens redshift of @xmath115 .
we report the discovery of two doubly - imaged quasar lenses , sdss j0806 + 2006 and sdss 1353 + 1138 . both were selected from the sdss
spectroscopic quasar sample as lensed quasar candidates and confirmed in subsequent imaging and spectroscopic observations .
sdss j0806 + 2006 consists of two @xmath0 quasar images separated by @xmath2 lensed by a galaxy at @xmath116 .
the lens galaxy is closer to the fainter image as expected , and its redshift , as suggested by its magnitude , colors , and the presence of a absorption feature , is @xmath117 .
several nearby galaxies may perturb this system and indicate that the lens galaxy is part of a small group .
sdss 1353 + 1138 consists of two @xmath1 quasar images separated by @xmath3 with a lens galaxy at @xmath118 .
the redshift of the lens galaxy is estimated based on its magnitude , colors , and the spectral flux ratio between the two quasar images .
there is an additional component , which we have labeled c , superposed on this system whose nature is presently unexplained .
observations using the _ hubble space telescope _ are probably required to clarify its role in the lensed quasar system .
n. i. and m. o. are supported by jsps through jsps research fellowship for young scientists .
a portion of this work was also performed under the auspices of the u.s .
department of energy , national nuclear security administration by the university of california , lawrence livermore national laboratory under contract no
. w-7405-eng-48 .
some of the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the korean scientist group , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
kochanek , c. s. , schneider , p. , wambsganss , j. , 2004 , part 2 of gravitational lensing : strong , weak & micro , proceedings of the 33rd saas - fee advanced course , g. meylan , p. jetzer & p. north , eds .
( springer - verlag : berlin ) lupton , r. , gunn , j. e. , ivezi , z. , knapp , g. r. , kent , s. , & yasuda , n. 2001 , in asp conf .
238 , astronomical data analysis software and systems x , ed .
f. r. harnden , jr .
, f. a. primini , and h. e. payne ( san francisco : astr . soc . pac . ) , p. 269
( astro - ph/0101420 ) crrrrrr a & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + b & @[email protected] & @[email protected] & [email protected] & [email protected] & [email protected] & [email protected] +
g & @[email protected] & @[email protected] & [email protected] & [email protected] & [email protected] & [email protected] + crrrrrrrr a & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + b & @[email protected] & @[email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + c & [email protected] & @[email protected] & & & & & [email protected] & [email protected] + g & @[email protected] & @[email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + | we report the discoveries of two , two - image gravitationally lensed quasars selected from the sloan digital sky survey : sdss j0806 + 2006 at @xmath0 and sdss j1353 + 1138 at @xmath1 with image separations of @xmath2 and @xmath3 respectively .
spectroscopic and optical / near - infrared imaging follow - up observations show that the quasar images have identical redshifts and possess extended objects between the images that are likely to be lens galaxies at @xmath4 in sdss j0806 + 2006 and @xmath5 in sdss j1353 + 1138 .
the field of sdss j0806 + 2006 contains several nearby galaxies that may significantly perturb the system , and sdss j1353 + 1138 has an extra component near its einstein ring that is probably a foreground star .
simple mass models with reasonable parameters reproduce the quasar positions and fluxes of both systems . |
m dwarfs are main sequence stars whose spectra display bands of tio and other molecules such as cah , vo , feh , and crh @xcite .
the high molecular opacities are the result of a cool atmosphere with effective temperatures in the range 2400k-3700k @xcite . observed orbital motion of astrometric binaries @xcite
show that m dwarfs have masses ranging from @xmath50.5 m@xmath6 all the way down to the hydrogen - burning limit ( @xmath50.075 m@xmath6 ) .
m dwarfs have relatively small sizes , with radii roughly proportional to their masses .
recent measurements of eclipsing systems @xcite indicate a roughly linear relationship between masses and radii , with a 0.4 m@xmath6 ( 0.1 m@xmath6 ) m dwarf having a radius of 0.4 r@xmath6 ( 0.15 r@xmath6 ) . by number ,
m dwarfs form the bulk of the stars in our galaxy .
the census of stars within 33 parsecs of the sun @xcite shows m dwarfs outnumbering all other hydrogen - burning objects by a factor of three to one .
because of their low mass - to - light ratio , m dwarfs contribute only a small fraction of the light emission in galaxies , but they may be the dominant baryonic component in at least some galaxies @xcite . in recent years
there has been an increasing interest in m dwarfs due to the discovery that they too host exoplanets . as of january 2011 , 25% of the 41 doppler - confirmed planets with @xmath730m@xmath8 are orbiting m dwarfs .
the geneva / harps planet search team reports that @xmath930% of low - mass stars harbor planets with masses @xmath030@xmath10 on orbits shorter than 50 d @xcite and preliminary _ kepler _ results suggest that m dwarfs host just as many , if not more , super - earths and neptunes as solar - mass stars @xcite .
exoplanet candidates have been identified by _ kepler _ in stars having colors consistent with m dwarfs @xcite , despite the fact that very few m dwarfs are being targeted by the mission @xcite .
large exoplanet surveys have now started to monitor sizable numbers of m dwarfs , such as the m2k program which is targeting some 1,600 m dwarfs for radial velocity monitoring @xcite , and the mearth project @xcite which is designed to detect exoplanet transits in nearby late - type m dwarfs .
the two principal methods of exoplanet detection , doppler measurement of radial velocity ( rv ) and photometric detection of transits , are more sensitive to planets around stars of lower mass . for a given planet mass and orbital period , rv amplitude scales as @xmath11 .
in contrast to late - type m stars , the vast majority of early m dwarfs are not rapid rotators or chromospherically active ( w@xmath12 ) @xcite and are thus amenable to these search techniques .
in fact , inactive k and early m stars have weaker p - mode oscillations and lower doppler noise ( `` jitter '' ) than their f or g counterparts @xcite .
photometric transit depth scales approximately as @xmath13 for planets of similar size .
because of their significantly smaller radii , m dwarfs display deeper transits which do not require high precision photometry for detection , as they do for g stars @xcite .
another method of planet detection is high - contrast imaging at infrared wavelengths where a ( giant ) planet is self - luminous @xcite .
m dwarf systems present a much more favorable planet / star contrast ratio compared to solar - mass stars , due to the significantly lower luminosity of m dwarfs .
young systems are of particular interest because the luminosity of giant planets is predicted to be orders of magnitude higher at ages of @xmath14 yr @xcite providing a more favorable contrast , even as the host star itself is moderately more luminous at this younger age .
young solar mass stars can be identified by their enhanced rotation , dynamo activity , and chromospheric emission @xcite , although the trend for late m stars for which the mechanism of dynamo operation is thought to be different is less clear @xcite .
another challenge for this approach is the apparent rarity of giant planets on long - period orbits around solar - mass stars @xcite combined with the paucity of giant planets around m dwarfs compared to g dwarfs @xcite .
it is also possible that young giant planets may be less luminous and more difficult to detect than previously predicted @xcite .
low luminosity m dwarfs are also attractive targets because planets within their diminutive habitable zones are easier to detect @xcite .
the correlation between high metallicity and the presence of giant planets found among solar - mass stars @xcite has been recently extended to m dwarf stars @xcite , and metal - rich m dwarfs , if they can be identified @xcite , would be especially propitious targets for planet searches .
despite the attractiveness of m dwarfs for exoplanet surveys , there still is no systematic , all - sky census of these objects to a useful limiting magnitude .
part of the problem is that the low - mass m dwarfs are stars of relatively low luminosities ( m@xmath15 ) .
all but the closest m dwarfs are too faint ( v@xmath912 ) to be part of the _ hipparcos _ catalog @xcite , which means that existing m dwarf samples are selected from proper motion catalogs with fainter magnitude limits .
these include the tycho-2 catalogue @xcite , which extends the hipparcos catalog a few magnitudes fainter , and also the older and deeper luyten half second ( lhs ) and new luyten two - tenths ( nltt ) catalogs
@xcite which were the primary source material for the well - known _ catalog of nearby stars _ @xcite .
newer and more complete proper motion catalogs , such as the lspm - north @xcite have the potential to expand the list of available m dwarf targets . in this paper , we present the first all - sky catalog of bright ( j@xmath010 ) m dwarfs .
the sample is extracted from the all - sky superblink proper motion catalog of stars with @xmath140 mas yr@xmath2 , an extension of lspm - north currently in development .
the selection method is detailed in 2 .
the catalog itself is presented in 3 .
the completeness of the catalog is evaluated in 4 .
conclusions follow in 5 .
the superblink proper motion survey is an all - sky search for stars with large proper motions based on a re - analysis of images from the digitized sky surveys ( dss ) using a specialized image - differencing algorithm and software .
the basic search method is described in @xcite .
quality control procedures , including cross - correlation with other catalogs and the compilation of astrometric and photometric results , is discussed at length in @xcite .
stringent quality control procedures , including visual confirmation of most objects , guarantees a very low level of false detections .
the survey has a fixed , low proper motion threshold of @xmath16 mas yr@xmath2 .
new stars with very large proper motions ( @xmath17 mas yr@xmath2 ) discovered in the survey have been published in @xcite , @xcite , @xcite , and @xcite .
a complete list of northern stars with proper motions @xmath18 mas yr@xmath2 was published in @xcite ; a similar list covering the southern sky has been completed and is available from the authors .
those complete lists incorporate data from the hipparcos and tycho-2 catalogs at the bright end , and the catalogs are estimated to be @xmath19 complete to a visual magnitude @xmath20 , and extend to @xmath21 .
the completeness is significantly higher at high galactic latitudes , where field crowding is less of an issue .
dense fields at low galactic latitudes have a lower completeness due to faint stars often being lost to non - linearity in the crowded areas of photographic plates ; saturation also prevents the detection of proper motion sources in the extended glare ( up to several arc minutes ) of extremely bright stars ( e.g. vega , sirius , canopus ) . as of july 2011 ,
the full superblink catalog ( stars with proper motions @xmath16 mas yr@xmath2 ) comprised 2,270,481 objects , with the most extensive sky coverage north of decl.=-20@xmath22 .
the catalog was complete for the northern sky ( decl.@xmath90@xmath22 ) , but the survey was still in progress for large parts of the southern sky .
the southern survey was however complete for stars with proper motions @xmath18 mas yr@xmath2 .
this version of the superblink catalog is the one used here .
the use of a proper motion catalog is a convenient way to identify nearby stars in general , and m dwarfs in particular , because proper motion - selected samples discriminate against distant background sources including m giants that would otherwise significantly contaminate a sample of cool stars selected by color alone .
however , proper motion selection carries its own inherent bias : stars with small velocities in the plane of the sky tend not to be detected in proper motion surveys , depending on the distance to the source .
nearby , slow - moving stars can be included by lowering the proper motion limit of the sample , but this has the effect of increasing the contamination from background sources . there is a fine balance to achieve in adjusting the proper motion threshold , which would ideally have the sample incorporate the largest possible fraction of stars up to a specified distance , minimizing the kinematic bias , while at the same time limiting the contamination from background objects .
the superblink catalog , with a proper motion limit @xmath16 mas yr@xmath2 , achieves such a balance for stars located within 100 parsecs of the sun .
this can be demonstrated with an analysis of proper motion in the _ hipparcos _ catalog .
figure 1 shows the fraction of hipparcos stars at a given distance from the sun which have proper motions @xmath18 mas yr@xmath2 and @xmath16 mas yr@xmath2 .
the former value is the proper motion limit of the lspm - north and south catalogs ( which is also the approximate proper motion limit of the nltt catalog and , by extension , the catalog of nearby stars ) , while the latter is the proper motion limit of the extended superblink catalog .
the diagram shows that operating with a proper motion limit of @xmath18 mas yr@xmath2 will only detect half the stars at 40pc and very few stars ( only those with very large components of motion ) at 100pc .
however a sample with a proper motion limit @xmath16 mas yr@xmath2 will include @xmath595% of the stars at 40pc and @xmath570% of the stars at 100pc . background giant stars , on the other hand , tend to have small proper motions , which exclude them from the superblink catalog .
figure 2 shows the distribution of proper motions for red giant stars in the hipparcos catalog , selected with the color and absolute magnitude cuts @xmath231.0 , @xmath243.0 .
the bulk of the giants have proper motions below the superblink threshold .
this particular subset of giants , because it is from the hipparcos catalog , also represents the nearest red giants to the sun , since they all have magnitudes @xmath25 .
more distant giants would have even smaller proper motions , and would also be excluded from our sample .
figure 3 shows the distribution of optical - to - infrared @xmath26 colors for a subset of m dwarfs with spectral classification from the sloan digital sky survey database @xcite .
infrared @xmath27 magnitudes are from 2mass , while @xmath28 magnitudes are the superblink visual magnitudes , estimated from palomar sky survey photographic plate measurements as described in @xcite .
the subset includes all stars in sdss which have counterparts in the superblink proper motion catalog . for this subset of nearby m dwarfs ,
@xmath26 color ranges from @xmath29 at subtype m0 , to @xmath30 at subtype m9 .
we adopt an m dwarf color - cut of @xmath31 , which is expected to include most stars of subtype m0 and later , at the expense of some contamination from late - k stars .
in addition , we select stars that are bright enough to be monitored by current and planned doppler radial velocity instruments . stars as faint as @xmath32 are routinely monitored using the high resolution spectrometer ( hires ) on the 10 m keck i telescope , e.g. @xcite , and fainter but highly meritorious stars with transiting planets , e.g. gj 1214 @xcite , have been observed .
m dwarf stars are brighter in the near - infrared , and future high - resolution spectrometers should be able to detect planets around stars as faint as @xmath33 . anticipating such developments , we select stars with @xmath34 , which also includes all m dwarfs with @xmath35 . to summarize ,
we select bright m dwarfs among high proper motion stars ( @xmath16 mas yr@xmath2 ) in the superblink catalog using these color / magnitude cuts : @xmath36 figure 4 shows the full magnitude - color distribution of superblink stars , and the region targeted by our selection . out of 2,270,481 stars in the superblink catalog , our color and magnitude cuts select 10,659 bright red stars .
this we use as our initial sample to identify bright m dwarfs .
the sky distribution of the color - selected stars is shown in figure 5 .
separate plots show the stars listed in the tycho-2 and hipparcos catalogs ( top ) , stars from the superblink catalog which are not listed in tycho-2 or hipparcos _ and _ have proper motions @xmath1 150 mas yr@xmath2 ( center ) , and finally stars from the superblink catalog not listed in tycho-2 or hipparcos _ and _ with proper motions 40 mas yr@xmath37 150 mas yr@xmath2 ( bottom ) .
the subsample from the tycho-2 and hipparcos catalogs includes 3,319 of the brighter stars , spread over the entire sky .
their distribution shows a slight overdensity in the direction of the galactic bulge , which is likely due to the larger number of red giants found in that part of the sky . the very high proper motion ( @xmath1 150 mas yr@xmath2 ) subsample from superblink includes 4,398 stars , and extends uniformly over the entire sky .
the moderately high proper motion ( 40 mas yr@xmath37 150 mas yr@xmath2 ) subsample from superblink includes 2,941 stars , and clearly suffers from incompleteness in the southern hemisphere ; the subsample also displays an overdensity around ( @xmath38,@xmath39)@xmath5(80,+20 ) which is associated with the nearby hyades cluster .
giant stars constitute the main source of contaminant in the sample . indeed , an excess of red sources in the apparent magnitude range @xmath40 ( see figure 4 ) points to a modest population of very red m giants in the superblink catalog ( probably including agb stars ) . while these giants are only a tiny fraction of the full superblink catalog , they are systematically selected in our initial color / magnitude cut . to eliminate these giants from our sample
we apply three selection criteria .
the first is based on evaluation of absolute magnitudes , and is used on stars with existing parallax data .
the second method identifies giants from their location in a reduced proper motion diagram , and is used for stars for which there are no existing parallax measurements .
a third method selects for m dwarfs based on @xmath41 and @xmath42 colors , further eliminating giants and interlopers with bad optical and/or infrared photometry .
we searched the literature to find astrometric parallaxes for stars in our subsample of m dwarf candidates .
parallaxes for 1,622 stars were recovered from the updated version of the hipparcos catalog @xcite .
ground - based astrometric parallaxes were additionally recovered for 10 stars in @xcite , 46 stars in @xcite , 15 stars in @xcite , 6 stars in @xcite , 3 stars in @xcite , 1 star in @xcite , 2 stars in @xcite , 18 stars in @xcite , 4 stars in @xcite , 14 stars in @xcite , 15 stars in @xcite , 3 stars in @xcite , and 3 stars in @xcite . whenever parallaxes were found for one star in more than one bibliographical source
, we adopted the most recent measurement , following the order above .
the numbers quoted above refer to incremental additions to the list of parallaxes .
in addition , we found parallaxes for 3 stars in the sky2000 master catalog , version 4 @xcite , and for 315 more from the nstars database . in all cases we only retained parallax measurements with errors @xmath020% .
the final tally comprises a total of 2,080 stars with astrometric parallaxes . from the parallax @xmath43
we calculate absolute visual magnitudes using : @xmath44 the resulting hertzsprung - russel ( h - r ) diagram is displayed in figure 6 .
the stars populate two very distinct loci .
the more luminous group includes all m giants and the less luminous group consists of all the m dwarfs . for stars with @xmath29 , the giants are typically 7 - 8 magnitudes more luminous than the dwarfs .
the difference in luminosity only increases at redder colors , with the reddest giants almost 20 magnitudes more luminous than dwarfs of similar @xmath26 color . from the absolute magnitude @xmath45 and optical - to - infrared color @xmath26 , we define the following criterion for a star to be considered an m dwarf : @xmath46 this cut identifies 561 stars as most likely to be red giants .
the remaining 1,519 objects are identified as m dwarfs .
we emphasize that this selection works only for the @xmath47 stars in our initial sample for which astrometric parallax measurements are available .
note that the color - magnitude relationship for the m dwarfs in our sample closely follows : @xmath48 which implies that we are eliminating stars that are more than 4.5 magnitudes overluminous compared to the average m dwarf of the same color .
this cut could possibly eliminate some very young m dwarfs , which tend to be overluminous in an optical - to - infrared color - magnitude diagram @xcite .
indeed a number of stars with parallaxes appear to be a few magnitudes over - luminous compared to average field stars .
however a close examination of figure 6 indicates that our magnitude cut would , in the worst case , eliminate only a handful of objects that could have been overluminous dwarfs . for stars without parallax measurements , it is possible to identify giants based on their location in a reduced proper motion diagram .
the reduced proper motion is defined as : @xmath49 a quantity which is analogous to the absolute magnitude @xmath45 ( eq.3 ) except that the proper motion @xmath50 is substituted for the parallax @xmath43 .
the two quantities are mathematically connected , with : @xmath51 where @xmath52 is the projected motion in the plane of the sky , i.e. the transverse velocity , here expressed in km s@xmath2 .
this means that the reduced proper motion diagram is basically an h - r diagram where the ordinates are modulated by the individual components of transverse velocities .
just as in the h - r diagram , m giants and m dwarfs tend to occupy distinct loci , with the giants all having relatively low values of @xmath53 and clustering in the upper part of the diagram .
we demonstrate this in the upper panel of figure 7 , which shows the reduced proper motion diagram of the stars in our sample for which there are parallax measurements .
stars identified as giants based on their parallax are plotted in red , dwarfs are plotted in black .
we define a demarcation line which separates stars of the two groups quite well . from this
, we postulate that m dwarfs can be separated from giants following this criterion : @xmath54 because the reduced proper motion depends on both absolute magnitude _ and _ transverse velocity , our selection may eliminate some main sequence m dwarfs if their transverse motion is small . based on equations 5 and 7
, we see that stars on the mean main sequence will be eliminated if their transverse velocity @xmath55 km s@xmath2 .
given that the velocity dispersion for stars in the solar neighborhood is @xmath540 km s@xmath2 @xcite , our adopted criterion will exclude relatively few m dwarfs .
conversely , giants with large transverse velocities will also have large values of @xmath53 which may place them beyond the limit defined in equation 8 . among the parallax - confirmed giants , we find 4 stars which fall within our defined reduced proper motion cut for m dwarfs .
one of the stars has a inferred transverse velocity @xmath56 km s@xmath2 based on its hipparcos parallax and proper motion .
however the other three stars have suspiciously faint absolute magnitudes ( [email protected] ) for m giants , and may instead be m dwarfs with erroneous ( underestimated ) parallaxes measurements .
the reduced proper motion diagram of stars with no parallax measurements is shown in the bottom panel of figure 7 .
of the 8,580 stars with proper motions but no parallax measurements , we identify 626 as probable m giants , and those are eliminated from the sample .
the remaining 7,954 stars are retained as m dwarf candidates .
this leaves 9,473 stars , with or without parallax , on the list of m bright m dwarf candidates .
a significant number of sources are found to have very red colors ( @xmath585 ) and reduced proper motion in the range typical for dwarf stars ( @xmath5912 ) and yet fall above the line we defined as the giant - dwarf boundary .
a close examination of their infrared colors shows that most of these stars fall outside the normal range for m dwarfs ( see below ) and are likely either stars with erroneous ( overestimated ) v magnitudes , or distant f - g - k stars with large components of reddening .
most of these stars are eliminated after the infrared cuts described below .
giant stars are notable for having infrared colors different from the m dwarfs , most apparent in a [ ( @xmath41),(@xmath60 ) ] color - color diagram .
this has been known since @xcite .
dwarf stars have comparatively bluer @xmath41 colors than giants .
this is due to opacity differences , mostly from molecular bands of h@xmath61o , which yield significantly higher @xmath62-band and @xmath63-band opacities in m dwarfs @xcite .
figure 8 shows the [ ( @xmath41),(@xmath60 ) ] color - color distribution for the stars in our initial sample .
stars with parallax measurements are shown in the upper panel , stars without parallax in the lower panel . in both panels ,
all objects we have identified as probable giants , based on either absolute magnitude ( 2.4.1 ) or reduced proper motion(2.4.2 ) , are plotted in red .
the remaining m dwarf candidates are plotted in black .
the giants are clearly segregated from the dwarfs along j - h , although there is a small overlap between the two groups . in fact
m dwarfs appear to be confined to a relatively compact region around @xmath64 and @xmath65 .
a small number of alleged m dwarfs fall outside the cluster ; some have red @xmath41 colors more consistent with the m giants , and indeed they most probably are .
other stars have bluer colors which are more typical of f - g - k stars ; we suspect these probably are more massive main sequence stars which made it into our sample probably due to an overestimated photographic magnitude @xmath28 , which would also overestimate the @xmath26 color term , bringing them into the m star subsample . in order to eliminate
these leftover misidentified m giants and misclassified f - g - k stars , we further restrict our m dwarf selection to the region in [ ( @xmath41),(@xmath60 ) ] diagram bounded by : @xmath66 among stars with trigonometric parallaxes , we find that 97 of the m dwarf candidates ( black dots in figure 8 upper panel ) fall outside the limits of the region defined above .
we conclude that these stars likely have significant errors in their estimated optical or infrared magnitudes , and we exclude them from our census of m dwarfs . among stars with no parallax measurements identified as probable m dwarfs from the reduced proper motion cut ( black dots in figure 8)
, we find 487 which fall outside the defined m dwarf boundaries ; these stars are eliminated as well .
some of the stars have color consistent with f - g - k main sequence stars ( [email protected] , [email protected] ) and most likely made the initial @xmath582.7 color cut due to an underestimated v magnitude .
other excluded sources have @xmath690.7 , and are either giants or distant main sequence stars with large reddening . after these cuts , we are left with a list of 8,889 probable bright m dwarfs , which we retain for building our catalog . among the giants confirmed through parallax ( stars plotted in red in the upper panel of figure 8) a small number fall within the selection box for m dwarfs .
specifically , 66 of the 561 giants fall within the region defined by equations 9 - 13 .
this means that jhk color cuts alone are susceptible to contamination by background giants .
this makes the reduced proper motion selection critical in minimizing giant contamination .
note that only 4 of the 561 giants have reduced proper motions within the limit defined for m dwarfs , which increases our confidence that giant contamination in our catalog is negligible .
the reduced proper motion cut , on the other hand , may be eliminating bona fide m dwarfs , as mentioned in 2.4.2 above .
the lower panel in figure 8 show evidence for this , as 160 of the 626 stars excluded after the reduced proper motion cut ( plotted in red ) do fall within the m dwarf box .
overall this suggests that our adopted selection criteria places more of an emphasis on avoiding contamination by giants , at the expense of some incompleteness in the m dwarf targets . from the numbers above
, we predict that @xmath70100 m dwarfs might be missing from our catalog because they fail the reduced proper motion cut .
in addition , we note that stars identified as giants based on parallax data ( figure 8 top panel , red dots ) and stars excluded from the reduced proper motion cuts ( figure 8 bottom panel , red dots ) show significantly different distributions in @xmath41/@xmath60 color - color space .
one average the parallax giants tend to have bluer @xmath42 colors .
we suggest that this is due to the fact that most red giants with @xmath140 mas yr@xmath2 are bright enough to have parallaxes from the hipparcos catalog , and that the reduced proper motion cut for stars with no parallax essentially serves to exclude stars with large components of reddening , which are expected to have redder @xmath42 colors than the giants .
lllrrrrrrrc pm i00001@xmath716943 & & & & 0.028531 & 69.717117 & 0.136 & 0.136 & -0.002 & & + pm i00003@xmath720802s & & & & 0.087698 & -8.037146 & 0.100 & 0.029 & -0.096 & & + pm i00005@xmath720533 & & & & 0.144895 & -5.551960 & 0.193 & 0.181 & 0.067 & & + pm i00006@xmath711829 & & tyc 1181 - 1683 - 1 & & 0.163528 & 18.488850 & 0.387 & 0.335 & 0.195 & & + pm i00007@xmath711624 & & & & 0.195877 & 16.402805 & 0.125 & 0.013 & -0.125 & & + pm i00007@xmath723510 & & & & 0.195306 & -35.168343 & 0.316 & 0.312 & -0.048 & & + pm i00007@xmath726243 & & & & 0.189068 &
-62.728802 & 0.199 & 0.166 & -0.110 & & + pm i00012@xmath711358s & & tyc 600 - 1507 - 2 & & 0.303578 & 13.972055 & 0.147 & 0.025 & 0.144 & [email protected] & va95 + pm i00014@xmath714724 & & & gj 1293 & 0.371546 & 47.414665 & 0.172 & 0.171 & -0.008 & & + pm i00014@xmath721656 & hip 112 & tyc 5838 - 784 - 1 & & 0.357633 & -16.948410 & 0.393 & 0.299 & -0.255 & [email protected] & vl07 + pm i00015@xmath720814 & & & & 0.399365 & -8.244875 & 0.124 & 0.097 & -0.077 & & + pm i00016@xmath723258 & & & & 0.408577 & -32.978249 & 0.238 & 0.238 & 0.005 & & + pm i00016@xmath727613 & & & & 0.408879 & -76.230522 & 0.188 & 0.188 & 0.008 & & + pm i00017@xmath723528 & & & & 0.433436 & -35.476082 & 0.502 & 0.502 & -0.023 & & + pm i00024@xmath710440 & & & & 0.613839 & 4.668418 & 0.122 & 0.058 & 0.105 & & + pm i00024@xmath724601 & hip 191 & tyc 8022 - 637 - 1 & & 0.612922 & -46.028893 & 0.194 & 0.192 & -0.022 & [email protected] & vl07 + pm i00025@xmath726324 & & & & 0.639251 & -63.401974 & 0.822 & -0.488 & -0.661 & & + pm i00026@xmath713821 & & & & 0.667151 & 38.362587 & 0.075 & -0.073 & -0.020 & & + pm i00026@xmath721123 & & & & 0.654399 & -11.393147 & 0.584 & -0.413 & -0.413 & & + pm i00026@xmath723919 & & & & 0.654961 & -39.320053 & 0.200 & 0.199 & 0.023 & & +
each of the 8,889 stars in our catalog is identified with the standard proper motion star i d used in the superblink proper motion survey .
these superblink names are tabulate in the first column of both table 1 and table 2 .
the first three letters ( `` pmi '' ) are the superblink catalog identifier .
these are followed by 5 digits which relate to the icrs right ascension of the star in sexagesimal ; the first four digits are the hours and minutes of right ascension , the fifth digit is the seconds of right ascension divided by 6 and rounded down to the nearest integer .
these are followed by the declination sign and then four digits which replicate the icrs declination ( hours , then minutes ) of the source .
a disambiguation letter ( nsew ) is used when the scheme would leave two stars with the same name , with the letter indicating the relative orientation of the stars in the pair .
this scheme follows the recommendation of @xcite and is the convention now adopted in reporting results from the superblink survey , see e.g. @xcite . in table 1
, we additionally provide the names of the stars in both the hipparcos ( column 2 ) and tycho-2 ( column 3 ) catalogs , for all stars which have counterparts in those catalogs .
these are provided for convenience because hipparcos names are now commonly used to refer to very bright stars , and because many exoplanet surveys have adopted the hipparcos names in designating the host stars and their exoplanets .
the tycho-2 names perhaps do not enjoy the same popularity , but are useful in identifying objects for which the proper motion is known to higher accuracy .
our list includes a total of 977 hipparcos and 1,859 tycho-2 stars .
some 779 entries have counterparts in both the hipparcos and tycho-2 catalogs , but 192 stars have hipparcos counterparts only , while 1,074 stars are only listed in tycho-2 . in addition , we identify by their catalog names all the stars listed in the third catalog of nearby stars @xcite or cns3 ( column 4 ) .
these are also convenient because cns3 names have been widely adopted by the exoplanet community to designate nearby stars .
the cns3 counterparts are found after cross - correlation with the catalog of revised positions by @xcite which lists 4,106 objects .
a total of 1,529 stars from our catalog of bright m dwarfs are thus matched to their cns3 counterpart .
stars with cns3 counterparts include most of the very nearest systems .
however , the cns3 does suffer from a significant kinematic bias because it is largely based on the proper motion catalogs of luyten , particularly the lhs catalog which has a relatively high proper motion limit @xmath74450 mas yr@xmath2 . as a result ,
many stars on our list which we identify as very nearby objects do not have cns3 names ; these are typically stars with relatively small ( @xmath75 mas yr@xmath2 ) proper motions .
we note that @xmath76 of the stars in our catalog ( 6,003 entries ) have no counterpart in either the hipparcos , tycho-2 , or cns3 catalog .
use of the pm designations for these stars would be convenient , especially given that the pm name indicates the location of the star on the sky .
coordinates in the table are tabulated in columns 5 and 6 .
right ascension ( @xmath38 ) and declination ( @xmath39 ) are listed in the icrs system , and quoted for the 2000.0 epoch .
this means that current positions must be extrapolated using the tabulated proper motions . for stars listed in the hipparcos catalog ,
we have extrapolated the positions to the 2000.0 epoch , from the values listed in @xcite which were listed for epoch 1991.25 .
for all other stars , the 2000.0 coordinates are based on the position of their 2mass counterparts as listed in @xcite and extrapolated to the 2000.0 epoch from the epoch of the 2mass observations .
the positions are typically accurate to 0.8@xmath77 , which are the quoted 2mass catalog errors on the absolute astrometry ; proper motion errors have little effect on the accuracy of the extrapolated positions because of the proximity of the 2mass survey epoch to the millennium year ( @xmath78 years ) .
the total proper motion @xmath50 of the stars are tabulated in column 7 and listed in seconds of arc per year .
the vector components of the proper motion in the directions of r.a . and decl .
( @xmath79 ) are tabulated in columns 8 and 9 .
both components are given in seconds of arc per year , following the superblink convention .
note that @xmath80 is listed in seconds of arc per year , not seconds of time per year , by which we mean that @xmath81 .
these proper motions are from three separate sources .
stars with hipparcos counterparts are listed with their proper motion from the hipparcos catalog @xcite .
stars not in the hipparcos catalog but with counterparts in the tycho-2 catalog are listed with their proper motion from tycho-2 @xcite .
the hipparcos and tycho-2 proper motions are typically accurate to 0.5 - 2.0 mas yr@xmath2 . for stars not listed in either
the hipparcos or tycho-2 , the proper motions listed are those measured in the superblink proper motion survey .
those proper motions are based on a re - analysis of the digitized sky surveys images , as described in detail in @xcite ; these proper motions have a typical precision of @xmath738 mas yr@xmath2 .
one can determine the source of the proper motion by checking whether the star has a hipparcos or tycho-2 name .
astrometric parallaxes are recovered from the literature for 1,422 of the m dwarfs in our catalog .
see section 2.4.1 for the list of papers and catalogs that were searched to recover the parallaxes .
all values are listed in column 10 whenever available , otherwise the field is left blank .
a flag is added in column 11 to indicate the source of the parallax , with the key provided in the table footnote . in all cases , measurement errors as reported in the relevant source , are noted in column 10 .
lrrrrrrrrrrrrrrc pm i00001@xmath716943 & 0.0461 & & & & & 14.3 & 12.6 & 10.8 & 9.70 & 9.12 & 8.84 & 13.52 & 3.82 & [email protected] & m3 + pm i00003@xmath720802s & & & & & & 13.9 & 11.8 & & 9.12 & 8.47 & 8.27 & 12.93 & 3.81 & [email protected] & m3 + pm i00005@xmath720533 & & & & & & 13.2 & 11.2 & 10.2 & 9.00 & 8.39 & 8.17 & 12.28 & 3.28 & [email protected] & m1 + pm i00006@xmath711829 & & & & 12.91 & 11.28 & 11.9 & 10.3 & 9.5 & 8.44 & 7.79 & 7.64 & 11.28 & 2.84 & [email protected] & k7 + pm i00007@xmath711624 & & & 21.02 & & & 14.0 & 11.9 & & 9.32 & 8.68 & 8.46 & 13.03 & 3.71 & [email protected] & m0 + pm i00007@xmath723510 & & & & & & 13.3 & 11.1 & 10.2 & 9.12 & 8.48 & 8.28 & 12.29 & 3.17 & [email protected] & m1 + pm i00007@xmath726243 & & & & & & 13.7 & 11.7 & 11.0 & 9.89 & 9.23 & 9.07 & 12.78 & 2.89 & [email protected] & m2 + pm i00012@xmath711358s & & & 19.85 & 13.67 & 11.12 & & & & 8.36 & 7.71 & 7.53 & 11.12 & 2.76 & [email protected] & k7 + pm i00014@xmath714724 & & & & & & 13.8 & 11.4 & 10.6 & 9.67 & 9.02 & 8.83 & 12.70 & 3.03 & [email protected] & k7 + pm i00014@xmath721656 & & & & 12.14 & 10.76 & 11.5 & 9.8 & 8.9 & 8.02 & 7.41 & 7.22 & 10.76 & 2.74 & [email protected] & m0 + pm i00015@xmath720814 & & & & & & 14.4 & 11.7 & & 9.79 & 9.12 & 8.91 & 13.16 & 3.37 & [email protected] & m1 + pm i00016@xmath723258 & & & & & & 13.4 & 11.6 & 10.7 & 9.77 & 9.14 & 8.94 & 12.57 & 2.80 & [email protected] & k7 + pm i00016@xmath727613 & & & & & & 14.1 & 12.6 & 10.6 & 9.48 & 8.89 & 8.60 & 13.41 & 3.93 & [email protected] & m3 + pm i00017@xmath723528 & & & & 14.99 & 13.36 & 14.5 & 12.1 & 11.0 & 9.82 & 9.19 & 8.93 & 13.36 & 3.54 & [email protected] & m2 + pm i00024@xmath710440 & & & 21.96 & & & 13.6 & 11.2 & 9.9 & 9.19 & 8.61 & 8.45 & 12.50 & 3.31 & [email protected] & m1 + pm i00024@xmath724601 & & & & 13.92 & 12.43 & & & & 9.18 & 8.52 & 8.34 & 12.43 & 3.25 & [email protected] & m1 + pm i00025@xmath726324 & & & & & & 14.2 & 12.1 & 10.7 & 9.32 & 8.71 & 8.53 & 13.23 & 3.91 & [email protected] & m3 + pm i00026@xmath713821 & & 21.97 & 20.15 & & & 14.8 & 13.1 & 10.6 & 9.71 & 9.20 & 8.91 & 14.02 & 4.31 & [email protected] & m1 + pm i00026@xmath721123 & & & & 14.74 & 13.09 & & & & 9.86 & 9.24 & 9.03 & 13.09 & 3.23 & [email protected] & m0 + pm i00026@xmath723919 & & & & & & 14.0 & 11.4 & 10.9 & 9.84 & 9.19 & 9.00 & 12.80 & 2.96 & [email protected] & m4 + table 2 lists photometric data for the stars , along with photometric parallaxes and estimated spectral types , both based on the photometric measurements .
superblink stars names are repeated in column 1 of table 2 , to facilitate the cross - identification with table 1 , although the number of lines are the same .
we have cross - correlated our list of bright m dwarfs with both the _ rosat _ all - sky bright source catalog @xcite and the _ rosat _ all - sky survey faint source catalog @xcite .
we used a search radius of 15 , which is on the order of the astrometric precision of the rosat catalog .
our search identified 1,065 stars with x - ray counterparts .
flux values in counts s@xmath2 are listed in column 2 .
x - ray fluxes can be useful in identifying young stars in the catalog .
m dwarfs can be x - ray emitters if they have significant chromospheric activity .
high activity levels are typically linked to rapid rotation , which in m dwarfs is indicative of either spin - orbit coupling in a close binary @xcite or of young age @xcite .
it is likely that the youngest m dwarfs in our catalog are to be found among the subsample of objects with significant x - ray emission .
follow - up observations of x - ray bright m dwarfs has been successful in identifying scores of low - mass members to young moving groups @xcite .
the catalog was cross - correlated against the fifth data release ( dr5 ) of the galex mission . using a @xmath82 search radius
, we found galex counterparts for 3,905 of the 8,889 stars on our list .
the @xmath83 and @xmath84 magnitudes are listed in columns 3 and 4 .
some 762 objects have counterparts in both the @xmath84 and @xmath83 , while 3,115 have counterparts only in the @xmath83 , leaving 28 stars with a counterpart in the @xmath84 only .
these ultra - violet ( uv ) magnitudes can also be used to identify young stars .
a high uv luminosity is associated with activity in m dwarfs , and can be used as an age diagnostic @xcite .
we include optical magnitudes @xmath85 and @xmath86 as they are listed in the hipparcos and tycho-2 catalogs .
these are included in columns 5 and 6 respectively .
these are available only for stars with counterparts in those catalogs .
the @xmath85 and @xmath86 magnitudes are useful because they are generally more accurate ( 0.1 mag or better ) than the photographic magnitudes which we list for all the stars .
optical magnitudes are obtained from the usno - b1.0 catalog of @xcite , which is based on scans of historical photographic surveys .
the blue magnitudes are extracted mostly from scans of @xmath87 plates from the palomar sky surveys ( poss - i , poss - ii ) and the southern eso schmidt ( serc ) survey .
red magnitudes are extracted from scans of @xmath88 plates from poss - i and poss - ii and also from the anglo - australian observatory red survey ( aao - red ) .
photographic infrared magnitudes are extracted from @xmath89 plates from poss - ii and serc .
cross - correlation with the usno - b1.0 catalog is performed as part of the superblink quality control procedure , and all ambiguous cases are verified by visual examination using overlays of the usno - b1.0 sources on the dss scans .
all the recovered photographic magnitudes are listed in columns 7 ( blue ) , 8 ( red ) and 9 ( infrared ) . since the optical @xmath90 , @xmath91 , @xmath92
, magnitudes are based on scans of photographic plates , they are generally not as reliable as magnitudes measured on electronic detectors , such as those from 2mass and galex .
photographic magnitudes are typically accurate to only @xmath93 mag . as part of the quality control process ,
all the stars in the superblink proper motion survey have their counterparts identified in the 2mass catalog of @xcite . as described in 3.2.1 above
, the 2mass counterparts provide the absolute astrometry for superblink objects . because of the color restriction of equation 1 , every object in our catalog is required to have a counterpart in 2mass .
the corresponding infrared @xmath27 , @xmath62 , and @xmath94 magnitudes are listed in columns 10 , 11 , and 12 , respectively .
the magnitudes are generally accurate to @xmath730.08 mag , though one should refer to the 2mass documentation for a detailed discussion of magnitude errors and uncertainties .
very bright stars ( @xmath95 ) , in particular , are saturated in 2mass and have larger photometric uncertainties . for convenience , we calculate an effective visual magnitude @xmath28 , which for bright stars is the visual magnitude @xmath86 quoted in the hipparcos or tycho-2 catalogs . for faint stars ,
it is a combination of the photographic @xmath90 , @xmath91 , and @xmath92 magnitudes , following the algorithm described in @xcite where this effective visual magnitude is labeled as @xmath96 . in this paper
, we use the shorter form @xmath28 for convenience , but it should be understood that all the mentions of @xmath28 actually refer to this effective visual magnitude @xmath96 .
all values are listed in column 13 .
photometric distances are estimated for all the stars based on the color magnitude relationship for m dwarfs calibrated in @xcite . for stars with @xmath31
the relationship is : @xmath97.\ ] ] photometric parallaxes @xmath98 are then calculated based on the estimated @xmath45 and the apparent visual magnitude @xmath28 .
values are listed in column 14 . as discussed in @xcite ,
the relationship has a scatter of 0.7 mag about the mean , which typically results in photometric distances with errors of @xmath99 . in figure 9
, we compare the photometric distances to triangulated distances , for stars with available parallaxes . of the 1,422 m dwarfs with parallaxes , some 1,206 stars have photometric distance estimates within @xmath99 of the astrometric distance , but only 772 have photometric distances that are accurate to @xmath100 .
there is also a significant number of outliers which could be unrecognized giants , young field stars , or unresolved doubles , all of which would have distances underestimated by photometry .
metal - poor stars ( m subdwarfs ) on the other hand would have their distances overestimated by photometry .
spectroscopic follow - up would be required to identify all such objects .
thus , while photometric distances are conveniently provided here for all the stars , one should bear in mind that these are mainly for guidance purposes , as these distances carry large uncertainties in many cases .
astrometric parallaxes ( listed in table 1 ) should be preferred for stars for which they are available , and should be measured for the remaining objects in our catalog .
for all the stars , we provide an estimated spectral subtype which is based on the relationship between @xmath26 color and the spectral subtype ( see fig .
we use this simple two - component linear relationship to estimate the subtype @xmath101 : @xmath102,\ ] ] after which the values are rounded up to the nearest integer .
we use lowercase letters ( k , m ) instead of the usual uppercase ( k , m ) to indicate that the spectral type is only an estimate ; a value of @xmath103 yields an estimated subtype of k7 .
a histogram of the estimated subtype distribution is shown in figure 10 , where it becomes clear that our magnitude - limited catalog of bright m dwarfs is biased toward stars of earlier subtypes .
the initial mass function is expected to peak at around 0.3 m@xmath6 or spectral subtype m3-m4 @xcite , so our incompleteness starts at around m0 and becomes severe by m5 .
these subtypes are tabulated in column 15 and may be used as a guide in selecting targets of interest .
some of the stars on the list , in particular stars listed in the cns3 , already have published literature about them , and the interested user would be advised to verify the existence of formal spectral classification(s ) .
we are currently conducting a spectroscopic follow - up survey to obtain formal spectral types for the brightest stars in our catalog , and intend to provide formal spectral classifications for these in the near future .
distribution on the sky of all the stars in our catalog is displayed in figure 11 in a galls - peter equal - area projection . to emphasize differences between the brighter and fainter stars in our catalog , we display the distribution in two panels , with one showing only stars with visual magnitudes @xmath104 ( upper panel ) and the other showing stars with visual magnitudes @xmath105 .
the separation roughly corresponds to the magnitude limit of the tycho-2 catalog . of the 1,993 stars in the brighter sample ,
1,614 are listed in either the tycho-2 or hipparcos catalog .
the fainter sample , on the other hand , contains only 437 hipparcos and tycho-2 stars out of a total of 6,896 objects .
the distribution of the brightest ( @xmath25 ) m dwarfs appears relatively uniform over the sky , except for an overdensity at low galactic latitude in the general direction of the galactic center ( @xmath106 ) .
this overdensity is likely due to contamination by background giants in dense fields .
since our catalog is dependent on the hipparcos and tycho-2 catalogs at the bright end , errors in proper motion measurements due to field crowding also possibly result in stars having overestimated proper motions , which would in turn fall within the reduced proper motion range of m dwarfs .
close examination of a number of stars at low galactic latitudes indeed raises doubt about the accuracy of the proper motion quoted in hipparcos and tycho-2 in at least a few cases .
the distribution of fainter ( @xmath107 ) m dwarfs , on the other hand , shows a net deficit of objects in the southern sky , especially for @xmath108 . as discussed in 2.3 this is due to the limited coverage of superblink in the south , where the survey is only complete for proper motions @xmath18 mas yr@xmath2 , compared to @xmath16 mas yr@xmath2 in the north . the distribution , however , does not show any overdensity / deficit of stars at low galactic latitudes , which suggests both a high completeness and low contamination from background giants .
this demonstrates the high efficiency and quality controls of the superblink survey in high density fields .
we illustrate in figure 12 the effects of kinematic bias and magnitude limit in our catalog .
the upper panels show the estimated absolute magnitudes as a function of distance , while the bottom panels show the transverse motions @xmath109 ( i.e. projected motions in the plane of the sky ) as a function of distance , for all the stars in our catalog .
the astrometric distances are used when available , otherwise we use the photometric distances as calculated in 3.4 .
one can not speak of incompleteness due to the magnitude limit , since the @xmath34 magnitude limit was _ imposed _ as a primary constraint to the catalog .
figure 12 however illustrates the effects of this magnitude limit in restricting the distance range of the stars in the catalog .
all the m dwarfs with @xmath34 populate the nearest 75 parsecs .
the limiting distance is also a function of absolute magnitude thus spectral subtype .
late - type m dwarfs , which have lower absolute magnitudes @xmath45 , have a much shorter distance limit in our catalog , which explains the strong deficit of stars with subtypes later than m4 ( see figure 10 ) .
the main source of incompleteness in our magnitude - limited catalog is kinematic bias .
because all of our stars are selected among sources with proper motions @xmath16 mas yr@xmath2 , nearby stars with low components of their transverse motion @xmath109 ( i.e. projected motion in the plane of the sky ) may be missed .
nearby stars can have low values of @xmath109 either because of a low intrinsic motion relative to the sun , or because their motion vector is nearly perpendicular to the plane of the sky .
for example , a star at 50 parsecs from the sun with @xmath110 mas yr@xmath2 has a transverse velocity @xmath1119.48 km s@xmath2 , which will occur with a probability of about 10% for stars in the solar neighborhood
( see 2.2 and fig.1 ) .
the superblink catalog currently has a higher proper motion limit ( @xmath112 mas yr@xmath2 ) in large swaths of the southern sky , which makes our census much more prone to a kinematic bias .
this explains the deficit of sources in the southern sky ( fig.11 ) .
this is only mitigated at the bright end because of our use of the hipparcos and tycho-2 catalogs . to estimate the overall completeness of our catalog , we proceed in two steps .
first we use monte - caro simulations to estimate the completeness of the northern sky census , which we assume to suffer only from kinematic bias due to the @xmath110 mas yr@xmath2 proper motion limit .
then we assume that the local distribution of field stars is isotropic in the vicinity of the sun , and thus infer how many stars are missing from the southern sky census , i.e. assuming that both the north and south hemisphere should have the same number of bright m dwarfs .
we first assume that the transverse velocity @xmath109 is independent of the distance @xmath113 for stars in the solar neighborhood , and assume a statistical distribution of the form @xmath114 .
we extract from our data the statistical distributions @xmath115 and @xmath116 from appropriate cuts in @xmath113 and @xmath109 . for @xmath117
, we consider only stars in the northern sky with transverse velocities @xmath118 km s@xmath2 ; this boundary is illustrated in figure 12 ( lower - left panel ) , which shows that the stars that satisfy this constraint do not suffer from the kinematic bias , because all the stars in our catalog which have @xmath118 km s@xmath2 also have proper motions @xmath16 mas yr@xmath2 .
the normalized distribution function @xmath117 is shown in figure 13 ( top panel ) . for @xmath119
, we consider only stars with distances @xmath120 parsecs ; again the boundary is illustrated in figure 12 , and shows that these very nearby stars suffer from minimal kinematic bias , because stars with proper motions @xmath110 mas yr@xmath2 ( and missing from the census ) can only have @xmath121 km s@xmath2 , which excludes very few stars .
the normalized distribution function @xmath119 is shown in figure 13 ( bottom panel ) where the limit of completeness ( @xmath122 km s@xmath2 ) is shown for reference .
given the sharp drop in @xmath119 for transverse motions @xmath123 km s@xmath2 , it is safe to assume that the sample of @xmath120 parsecs stars is essentially complete .
we perform monte - carlo simulations based on the statistical distribution @xmath117 and @xmath119 to estimate the probability that an m dwarf with @xmath34 will be detected in our survey .
we thus estimate that our census of the northern sky has a completeness of @xmath124% .
this means that the 5,361 stars in the northern sky sample are most likely drawn from a hypothetical sample of 5,840@xmath7318 stars .
the @xmath125 stars missing due to the kinematic bias are expected to have proper motions @xmath110 mas yr@xmath2 .
the above analysis assumes that the superblink survey is complete within the proper motion range and magnitude limit .
if the survey itself is incomplete , then the number of missing objects would be larger .
the completeness of the superblink survey generally depends magnitude and field density .
however in the magnitude range under consideration ( 8@xmath12615 ) a completeness test based on the fraction of tycho-2 stars recovered by superblink _ and _ the fraction of superblink stars listed in tycho-2 , suggests that the combined tycho-2/superblink catalog has a completeness @xmath127 @xcite .
if we factor in the possibility that @xmath52% stars in that magnitude range have escaped detection by both tycho-2 and superblink , this would bring the total number of hypothetical bright m dwarfs in the northern sky to @xmath55,950 stars .
our southern sky census is significantly less complete than the northern census .
while our catalog has 5,361 stars north of the celestial equator , it only has 3,528 in the south .
the main reason for the incompleteness is the limited coverage of the superblink proper motion survey in the south .
the survey is complete to @xmath16 mas yr@xmath2 for the entire northern sky , but reaches that limit only over a fraction of the south , mainly in the declination range @xmath128 ( see fig.11 ) south of @xmath129 , the survey is only complete to @xmath18 mas yr@xmath2 , but reaches to @xmath130 mas yr@xmath2 in some areas .
the incompleteness of the superblink survey is mitigated by the inclusion of the tycho-2 catalog , which is mostly limited by magnitude and actually includes stars with proper motions smaller than the superblink limit of @xmath13140 mas yr@xmath2 .
however the magnitude limit of the tycho-2 catalog , @xmath13212.0 , is too bright to include all but a fraction of the m dwarfs in the magnitude range of interest ( @xmath133 ) . under the assumption that the northern sky census is fully representative of the entire volume around the sun , i.e. assuming the m dwarf distribution to be _ isotropic _ in the solar neighborhood , we use the results from the monte - carlo simulation above and estimate that the southern - sky census of m dwarfs brighter than @xmath33 should also include 5,950 stars , for a total of 11,900 m dwarfs over the entire sky . given that our southern sample includes only 3,528 stars ,
we thus conclude that we are currently missing @xmath52,420 stars in the southern sky . in other words our southern sky census has a completeness of only @xmath559.3% .
the need to assemble a comprehensive list of _ bright _ m dwarfs over the entire sky is primarily motivated by the search for extra - solar planets . based on current techniques ,
m dwarfs are the most promising stars around which to look for earth - mass planet in a habitable zone .
radial velocity measurements are most sensitive to the reflex motion from planets in close orbits around their parent stars , or to planets orbiting low - mass stars ; for planets in habitable zones , this favors m dwarfs , whose zone is closer in , and whose masses are relatively low .
transit surveys have a higher chance of detecting planets in close orbits , and are more sensitive to planets orbiting smaller stars ; this likewise favors m dwarfs as the most promising targets .
the main caveat is that both radial velocity measurements and transit monitoring are most efficient for relatively bright stars . with current observational resources ,
radial - velocity monitoring can realistically be performed on stars with absolute magnitudes @xmath134 , and near - term developments only promise to increase the depth of surveys by one or two magnitudes , most notably by observing at infrared wavelengths where m dwarfs are noticeably brighter .
m dwarfs with apparent infrared magnitudes @xmath34 will soon be within range of radial velocity surveys .
the motivation of the superblink proper motion survey is to identify the largest possible number of hydrogen - burning objects up to at least 100 parsecs of the sun . with a low proper motion limit of @xmath16 mas yr@xmath2 and astrometric accuracy @xmath135 mas yr@xmath2 ,
the catalog suffers from minimal kinematic bias to a distance of 50 - 60 parsecs , which is the range over which one would expect to find m dwarfs with @xmath34 .
the current superblink catalog is complete over the entire northern sky , but the southern sky survey is still in progress . at this time
, the catalog has a proper motion limit @xmath16 mas yr@xmath2 in the declination range @xmath136 ; at southern declinations the proper motion limit is now @xmath18 mas yr@xmath2 , except for relatively bright stars ( @xmath25 ) for which the catalog extends to @xmath16 mas yr@xmath2 because of the inclusion of data from the tycho-2 and hipparcos catalogs .
our search of the current version of the superblink catalog has identified 8,889 bright m dwarfs , of which 5,361 are in the northern sky and 3,528 in the south .
the lower number of stars catalogued in the south reflects the limited extent of our proper motion survey in the south .
our estimate of the kinematic bias suggests that our northern census is @xmath124% complete . from this
we infer that there should be @xmath137 m dwarfs with magnitudes @xmath34 over the entire sky .
our catalog currently lists @xmath138% of the entire set .
we recommend that these stars be observed by radial - velocity monitoring programs with the highest priority .
in addition to exoplanet surveys , completing the census of the solar neighborhood is important to constrain the stellar density in the galaxy , and to determine the distribution of stellar properties .
our catalog includes many stars which are identified here for the first time as members of the solar neighborhood . in particular ,
most of the sources are not listed in the catalog of nearby stars .
the key to completing the census of bright m dwarfs in the south is to obtain a proper motion catalog with a lower proper motion limit , currently limited to @xmath18 mas yr@xmath2 south of @xmath129 .
we are currently working on an extension of the superblink proper motion survey , which will extend the limits to @xmath16 mas yr@xmath2 to @xmath139 , and to @xmath130 mas yr@xmath2 in the @xmath140 range .
we predict that extended surveys will identify an additional @xmath141 bright m dwarfs , if a proper motion limit @xmath16 mas yr@xmath2 can be reached over the entire sky .
a revised catalog will be produced as the survey reaches completion . at this time
we are conducting a systematic spectroscopic follow - up survey of the brightest stars in this catalog , and searching the literature for previous determinations of spectral subtypes .
results from these investigations will be provided in upcoming papers by our team .
perryman , m. a. c. & esa 1997 , the hipparcos and tycho catalogues .
astrometric and photometric star catalogues derived from the esa hipparcos space astrometry mission , publisher : noordwijk , netherlands : esa publications division | we present an all - sky catalog of m dwarf stars with apparent infrared magnitude j@xmath010 .
the 8,889 stars are selected from the ongoing superblink survey of stars with proper motion @xmath140 mas yr@xmath2 , supplemented on the bright end with the tycho-2 catalog .
completeness tests which account for kinematic ( proper motion ) bias suggest that our catalog represents @xmath3 of the estimated @xmath4 m dwarfs with j@xmath010 expected to populate the entire sky .
our catalog is , however , significantly more complete for the northern sky ( @xmath590% ) than it is for the south ( @xmath560% ) .
stars are identified as cool , red m dwarfs from a combination of optical and infrared color cuts , and are distinguished from background m giants and highly - reddened stars using either existing parallax measurements or , if such measurements are lacking , on their location in an optical - to - infrared reduced proper motion diagram .
these bright m dwarfs are all prime targets for exoplanet surveys using the doppler radial velocity or transit methods ; the combination of low - mass and bright apparent magnitude should make possible the detection of earth - size planets on short - period orbits using currently available techniques .
parallax measurements , when available , and photometric distance estimates are provided for all stars , and these place most systems within 60 parsecs of the sun .
spectral type estimated from v - j color shows that most of the stars range from k7 to m4 , with only a few late m dwarfs , all within 20 pc .
proximity to the sun also makes these stars good targets for high - resolution exoplanet imaging searches , especially if younger objects can be identified on the basis of x - ray or uv excess . for that purpose ,
we include x - ray flux from rosat and fuv / nuv ultraviolet magnitudes from galex for all stars for which a counterpart can be identified in those catalogs .
additional photometric data include optical magnitudes from digitized sky survey plates , and infrared magnitudes from 2mass . |
the global analyses of the free nucleon parton distribution functions ( pdfs ) are grounded on the asymptotic freedom of qcd , factorization and parton evolution .
these features allow to express the hard - process cross - sections formally as @xmath0 where @xmath1s are the scale - dependent pdfs , and @xmath2 denote the perturbatively computable partonic pieces .
the factorization theorem has turned out to work extremely well with increasingly many different types of data included in the latest free proton pdf analyses . for bound nucleons factorization
is not as well - established , but it has anyway proven to do a very good job @xcite in describing the measured nuclear modifications @xmath3 in deep inelastic scattering ( dis ) and drell - yan ( dy ) dilepton production involving nuclear targets .
this talk summarizes our new nlo analysis @xcite .
we define the nuclear pdfs through nuclear modification factors @xmath4 as follows : @xmath5 where @xmath6 refers to a cteq set of the free proton pdfs @xcite in the zero - mass variable flavour number scheme .
the @xmath7 and @xmath8 dependences of @xmath4s satisfying the sum rules , are parametrized considering three modifications : @xmath9 for both valence quarks , @xmath10 for all sea quarks , and @xmath11 for gluons .
the nuclear pdfs at @xmath12 are obtained as solutions to the dglap equations using our own nlo dglap solver based on a semi - analytical method described e.g. in @xcite .
all cross sections are computed using the factorization theorem and the initial parametrization is adjusted to find the minimum of @xmath13 ^ 2 .
\nonumber\end{aligned}\ ] ] for each data set @xmath14 , the @xmath15 denotes the experimental data value with point - to - point uncertainty @xmath16 , and @xmath17 is the theory prediction computed using parameters @xmath18 .
the pion data suffers from an overall @xmath19 normalization uncertainty @xmath20 , and the normalization factor @xmath21 $ ] is determined by minimizing @xmath22 .
the weight factors @xmath23 amplify the importance of those data sets whose content is physically relevant , but whose contribution to @xmath24 would otherwize be too small to be noticed by the automated minimization routine we use .
besides finding the central set of parameters @xmath25 that optimally fits the data , propagating the experimental uncertainties to the pdfs has become an inseparable part of the modern pdf fits .
the hessian method @xcite , which we use , trusts on a quadratic approximation @xmath26 the eigenvectors @xmath27 of the hessian matrix @xmath28 serve as an uncorrelated basis for building the pdf error sets @xmath29 .
these are obtained by displacing the fit parameters to the positive / negative direction along @xmath27 such that @xmath24 grows by a certain amount @xmath30 from the minimum @xmath31 . using these sets , the upper and lower uncertainty of a quantity @xmath32
can be written e.g. as @xmath33 ^ 2 \label{eq : error_best}\end{aligned}\ ] ] where @xmath34 denotes the value of the quantity @xmath32 computed using the set @xmath29 .
requiring each data set to remain close to its 90%-confidence range , we obtain @xmath35 .
at @xmath36 and at @xmath37 .
the black lines indicate the best - fit , whereas the dotted green curves denote the individual error sets which combine to the shaded bands like in eq .
( [ eq : error_best ] ) . ]
now , we briefly summarize the main results from the present analysis , starting with fig .
[ fig : pbpdfs ] where we show the obtained modifications for lead the nucleus relevant for the lhc at two scales .
interestingly , even the rather large uncertainty at small-@xmath7 gluons becomes notably smaller in the scale evolution .
this is a prediction that might be testable in the future colliders . as the dis and dy data constitute our main data constraints , we display in fig . [
fig : rf2a1 ] some examples of the measured nuclear modifications with respect to deuterium , @xmath38 for different nuclei and compare with the eps09 . and @xmath39 compared with the nmc @xcite and e772 @xcite data.,title="fig : " ] and @xmath39 compared with the nmc @xcite and e772 @xcite data.,title="fig : " ] the shaded blue bands always denote the uncertainty derived from the eps09 error sets .
their size is comparable to the experimental errors , supporting our choice for @xmath30 . at @xmath40 from hkn07 @xcite , nds @xcite and this work , eps09 @xcite .
right : the computed @xmath41 for @xmath42 yield compared with the phenix @xcite and star @xcite data multiplied by @xmath43 and @xmath44 respectively.,title="fig : " ] at @xmath40 from hkn07 @xcite , nds @xcite and this work , eps09 @xcite .
right : the computed @xmath41 for @xmath42 yield compared with the phenix @xcite and star @xcite data multiplied by @xmath43 and @xmath44 respectively.,title="fig : " ] compared with the nmc data @xcite . ]
the nuclear modification for pion production is defined as @xmath45 where @xmath46 denotes the number of binary nucleon - nucleon collisions and @xmath47 the pion s transverse momentum and rapidity .
the comparison with the phenix and star data plotted in fig .
[ fig : phenix ] , shows that the shape of the spectrum which in our calculation is a reflection of the similar shape in @xmath48 gets well reproduced by eps09 .
figure [ fig : phenix ] also presents a comparison of the eps09 gluon modifications @xmath49 with the earlier nlo analyses .
the significant scatter of the curves highlights the difficulty of extracting the nuclear modifications from the dis and dy data alone .
consequently , also the predictions for pion @xmath41 differ significantly as is easily seen in fig . [
fig : phenix ] .
this is actually good news as this type of data , especially with better statistics , may eventually discriminate between different proposed gluon modifications .
attention should be also paid to the experimentally observed scaling - violations and that the dglap dynamics is able to reproduce them well .
such effects are most transparent e.g. in the small-@xmath7 structure function ratios versus @xmath50 , of which fig .
[ fig : rf2_slopes ] shows an example . to summarize , the excellent agreement with the experimental data , @xmath51 , and especially the correct description of the scale - breaking effects
we argue is evidence for the applicability of collinear factorization in nuclear environment .
the best fit and all the 30 nlo npdf error - sets are available for practical use from @xcite . also the leading - order eps09 sets are provided .
we thank the v. , y. , & k. visl foundation , the m. ehrnrooth foundation , and the academy of finland for financial support . 00
k. j. eskola , v. j. kolhinen and p. v. ruuskanen , nucl .
b * 535 * ( 1998 ) 351 [ arxiv : hep - ph/9802350 ] .
k. j. eskola , v. j. kolhinen and c. a. salgado , eur .
j. c * 9 * ( 1999 ) 61 [ arxiv : hep - ph/9807297 ] .
k. j. eskola , v. j. kolhinen , h. paukkunen and c. a. salgado , jhep * 0705 * ( 2007 ) 002 [ arxiv : hep - ph/0703104 ] .
k. j. eskola , h. paukkunen and c. a. salgado , jhep * 0807 * ( 2008 ) 102 [ arxiv:0802.0139 [ hep - ph ] ] . m. hirai , s. kumano and t. h. nagai , arxiv:0709.3038 [ hep - ph ] .
d. de florian and r. sassot , phys .
d * 69 * ( 2004 ) 074028 [ arxiv : hep - ph/0311227 ] .
k. j. eskola , h. paukkunen and c. a. salgado , jhep * 0904 * ( 2009 ) 065 [ arxiv:0902.4154 [ hep - ph ] ] .
d. stump _ et al .
_ , jhep * 0310 * ( 2003 ) 046 [ arxiv : hep - ph/0303013 ] .
p. santorelli and e. scrimieri , phys .
b * 459 * ( 1999 ) 599 [ arxiv : hep - ph/9807572 ] .
h. paukkunen , phd thesis , arxiv:0906.2529 [ hep - ph ] .
j. pumplin _ et al .
_ , phys .
d * 65 * ( 2001 ) 014013 [ arxiv : hep - ph/0101032 ] .
m. arneodo _ et al . _
[ new muon collaboration .
] , nucl .
b * 441 * ( 1995 ) 12 [ arxiv : hep - ex/9504002 ] .
p. amaudruz _ et al .
_ [ new muon collaboration ] , nucl .
b * 441 * ( 1995 ) 3 [ arxiv : hep - ph/9503291 ] .
d. m. alde _ et al .
_ , phys .
* 64 * ( 1990 ) 2479 .
s. s. adler _ et al .
_ [ phenix collaboration ] , phys .
* 98 * ( 2007 ) 172302 [ arxiv : nucl - ex/0610036 ] .
j. adams _ et al . _ [ star collaboration ] , phys . lett .
b * 637 * ( 2006 ) 161 [ arxiv : nucl - ex/0601033 ] .
m. arneodo _ et al . _
[ new muon collaboration ] , nucl .
b * 481 * ( 1996 ) 23 . | in this talk , we present our recent next - to - leading order ( nlo ) nuclear parton distribution functions ( npdfs ) , which we call eps09 . as an extension to earlier nlo analyses , we supplement the deep inelastic scattering and drell - yan dilepton data by inclusive midrapidity pion measurements from rhic in order to reduce the otherwize large freedom in the nuclear gluon densities .
our hessian - type error analysis leading to a collection of npdf error sets , is the first of its kind among the npdf analyses . |
ultracold fermi gases have been subject of many experimental and theoretical studies during recent years ( see e.g. @xcite ) .
they provide a unique system to study key concepts of condensed matter theory .
this is because in these systems many parameters such as the particle density , the fermi energy , the confinement potential , or the interaction strength between the fermions , which in a solid state system are typically fixed quantities , can be externally controlled in a wide range @xcite . in particular , magnetic - field feshbach resonances provide the means for controlling the interaction strength between fermions by varying an external magnetic field .
the tunability of the s - wave scattering length , which is the dominant interaction channel , makes ultracold fermi gases ideal for exploring different regimes of interacting many - body systems in a single system .
this includes the limiting regimes of weakly attracting fermions , which condense into cooper pairs forming a bardeen - cooper - schrieffer ( bcs ) phase below a certain temperature @xmath0 , and repulsive dimers formed by two fermions , which can undergo a bose - einstein condensation ( bec ) .
these two limiting regimes are separated by the strongly interacting bcs - bec crossover regime where the scattering length diverges and the system exhibits unitary properties @xcite .
in addition to the variable interaction strength , ultracold atomic gases offer a unique opportunity to explore the influence of a confinement on the pairing correlations , because dimensionality and confinement can be precisely controlled by tuning external parameters @xcite . varying the confinement ,
which is often well approximated by a harmonic confinement potential , allows one to access new degrees of freedom . restricting the fermi gases to quasi - low dimensionality may
, e.g. , offer the possibility for an experimental evidence of unconventional phases , like the fulde - ferrell - larkin - ovchinnikov ( fflo ) state @xcite .
moreover it may help to study and get experimental insight into shape resonances , theoretically predicted for quasi - low dimensional conventional superconductors @xcite . in ref .
@xcite the first quantitative measurements of the transition from 2d to quasi-2d and 3d in a weakly interacting fermi gas has been reported . at low atom numbers , the shell structure associated with the filling of individual transverse oscillator states has been observed . on the theoretical side the ground state properties of a @xmath1li gas confined in a cigar - shaped laser trap
have been investigated predicting size - dependent resonances of the superfluid gap @xcite , similar to the case of superconducting nanowires @xcite , yielding an atypical bcs - bec crossover .
an important effort is now devoted to the exploration of the out - of - equilibrium behavior of trapped ultracold atomic fermi gases and , in particular , to the determination of their dynamical properties .
the dynamics has been studied in the normal as well as in the condensed phase , observing second sound @xcite and soliton trains @xcite , and showing a low - frequency oscillation of the cloud after a change of the system confinement or optical excitation @xcite .
furthermore , state - of - the - art technology allows one to change the coupling constant on such short time scales that it is possible to explore the regime where the many - body system is governed by a unitary evolution with nonequilibrium initial conditions . in ultracold atomic fermi gases the dynamics
may be initiated by readjusting the pairing interaction through switching an external magnetic field in the region of a feshbach resonance ( i.e. , a quantum quench ) or by a rapid change of the confinement potential of the trap @xcite . due to the small energies in the trapping potential
the dynamics in the fermi gases take place on a millisecond timescale . therefore ,
in contrast to metallic superconductors , where sub - picosecond excitations are required to achieve non - adiabatic dynamics @xcite , in atomic gases the non - adiabatic regime can be reached already by excitations in the sub - millisecond range .
spontaneous symmetry breaking gives rise to collective modes of the order parameter which are classified into the higgs amplitude mode , and the goldstone mode , the latter corresponding to a phase oscillation of the gap @xcite .
the higgs mode has been subject of intensive theoretical and experimental @xcite research efforts in the past . on the theoretical side the non - adiabatic temporal response of the order parameter to ( quasi-)instantaneous perturbations has been studied .
different regimes of an oscillatory temporal behavior of the pairing potential were theoretically predicted in homogeneous fermionic condensates @xcite .
it was shown that the amplitude of the order parameter oscillates without damping when the coupling constant is increased above a certain critical value @xcite .
on the other hand , the gap vanishes when the coupling constant is decreased below another critical value . in between these two limiting regimes
the amplitude exhibits damped dephased oscillations and the system goes to a stationary steady state with a finite gap @xcite . in extended systems
the approach to a stationary state occurs in an oscillatory way with an inverse square root decay in time of the amplitude of the oscillations @xcite .
a similar evolution was predicted for conventional bulk superconductors @xcite where the non - adiabatic regime is reached by excitation with short , intense terahertz pulses .
an experimental realization was reported in @xcite .
in contrast , in finite length superconducting nanowires a breakdown of the damped oscillation and a subsequently rather irregular dynamics has been predicted @xcite .
in this paper we present a theoretical analysis of the short - time bcs dynamics of a @xmath1li gas confined in a cigar - shaped laser trap .
the excitation is modeled by a sudden change of the interaction strength which can be achieved through a feshbach resonance by an abrupt change of the external magnetic field @xcite . applying the well - known bcs theory in mean field approximation to ultracold fermi gases
we show that the change of the coupling strength induces a collective oscillation of the bogoliubov quasiparticles close to the fermi level .
this results in a damped amplitude oscillation of the bcs gap , which corresponds to the higgs mode .
like in the case of confined bcs superconductors this oscillation breaks down after a certain time revealing rather chaotic dynamics afterwards .
we explain these dynamics in terms of coupled harmonic oscillators which can be derived by linearizing the quasiparticle dynamics obtained from the heisenberg equation of motion . in doing
so we first derive the quasiparticle equations of motion from the inhomogeneous bogoliubov - de gennes hamiltonian ( sec .
[ sec : formalism ] ) which , because of using the standard contact - type interaction , requires a proper regularization of the gap equation .
starting from the ground state calculated according to ref .
@xcite we then calculate the dynamics of the superfluid gap after an instantaneous change of the coupling constant
. the numerical results as well as their explanation follow in sec . [ sec : results ] , where we first discuss a rather small system and then proceed to a larger , experimentally more easily accessible system . finally , in sec .
[ sec : conclusions ] we summarize our results and give some concluding remarks .
our approach aims at describing the dynamics of the superfluid order parameter @xmath2 of an ultracold @xmath1li fermi gas , confined in a cigar - shaped , axial symmetric harmonic trapping potential @xmath3 here , @xmath4 is the mass of the @xmath1li atoms and @xmath5 ( @xmath6 ) is the confinement frequency in the @xmath7-@xmath8-plane ( @xmath9-direction ) , respectively . choosing @xmath10 yields an elongated cigar - shaped trap where the oscillator length @xmath11 provides a measure of the system length .
the eigenvalues @xmath12 are measured with respect to the fermi energy @xmath13 .
the index @xmath4 refers to the combination of quantum numbers @xmath14 , @xmath15 , and @xmath16 . for this geometry
the one - particle states form one - dimensional subbands , labeled by @xmath17 [ cf .
[ fig : states ] ] , while the states within each subband are labeled by @xmath16 .
each subband has a constant one - particle density of states and thus the overall density of states exhibits finite jumps whenever a new subband appears .
we consider the gas to be composed of two spin states , @xmath18 and @xmath19 , and start from the inhomogeneous bcs hamiltonian at @xmath20 . within the anderson approximation
we then derive equations of motion for the corresponding bogoliubov quasiparticle excitations .
the usual inhomogeneous bcs hamiltonian for an effective bcs - type contact interaction reads @xcite @xmath21 \ , d^3 r\notag \\ & - g \int \ , \psi_{\uparrow}^{\dagger}({\mathbf{r } } ) \psi_{\downarrow}^{\dagger}({\mathbf{r } } ) \psi_{\downarrow}^{}({\mathbf{r } } ) \psi_{\uparrow}^{}({\mathbf{r } } ) \ , d^3 r , \end{aligned}\ ] ] where @xmath22 and @xmath23 are the field operators for up and down spin , respectively , @xmath24 is the one - particle hamiltonian and @xmath25 is the coupling constant of the contact interaction @xmath26 . in the limit of low temperatures
the main contribution to the interaction between two fermionic atoms in different internal spin states is given by scattering processes at low momentum .
the description of those can be replaced by the widely known pseudopotential only depending on the scattering length @xmath27 @xcite , which yields @xmath28 . a bcs - like mean field expansion in terms of anomalous expectation values and a particle - hole transformation , leaving spin - up operators unchanged , @xmath29 , while interchanging spin - down ones , @xmath30 , leads to the bogoliubov - de gennes ( bdg ) hamiltonian @xcite @xmath31 where @xmath32 is the bcs order parameter . from eq .
it becomes apparent that the corresponding eigenvalue equation can be written as the one - particle bogoliubov - de gennes equation @xmath33 @xmath34 can thus be diagonalized by bogoliubov s transformation , using the eigenfunctions @xmath35 and @xmath36 .
this introduces non - interacting quasiparticles with energy @xmath37 obeying fermionic commutation relations with the corresponding creation operator @xmath38 \ , d^3 r.\ ] ]
the spectrum of the bdg equation is symmetric with respect to the fermi energy and thus the eigenstates of the bdg equation occur in pairs .
labeling @xmath39 for states @xmath40 and @xmath41 for @xmath42 , respectively , one finds the relations @xmath43 and @xmath44 for the eigenstates .
therefore , all quantities can be expressed solely using the @xmath45 state wave functions @xcite . in the following
we omit the @xmath45 , @xmath46 index of the eigenfunctions while they are still necessary for the quasiparticle operators .
hereafter in the case of the eigenfunctions the index @xmath4 refers to the @xmath45 states . for our further calculations it is convenient to transform to the excitation picture ( @xmath47 , @xmath48 ) with @xmath49 and @xmath50 , where all quasiparticle excitations vanish in the ground state .
we can rewrite the order parameter in the basis given by the eigenfunctions , where it reads : @xmath51 this yields the well - known result for the ground state order parameter @xmath52 which has to be solved self - consistently with eq .
focusing on the underlying physics , we exploit the anderson approximation ( a.a . )
@xcite , choosing the bdg wave functions @xmath53 and @xmath54 proportional to the one - particle wave functions of the confinement potential @xmath55 , i.e. , @xmath56 here the amplitudes of the bdg wave functions @xmath57 , @xmath58 are obtained from the bdg eq . and read @xmath59 with the quasiparticle energies given by @xmath60 and the one - particle energies @xmath61 given by eq . .
applying the anderson approximation to eq . additionally yields @xmath62 , where @xmath63 are the one - particle eigenfunctions .
the anderson approximation has been tested in several nanostructured geometries and no qualitative deviations have been found @xcite .
it is applied to all our calculations . from eqs .
and we obtain the well - known bcs - like self - consistency equation , also referred to as gap equation .
the ground state order parameter in the state @xmath64 is given by @xmath65 here @xmath66 is the interaction matrix element @xmath67 which exhibits maxima for states at the subband minimum ( i.e. , states with low @xmath16 ) .
the contact interaction used here leads to a well - know ultraviolet divergence in the summation over all states , i.e. , in eq . , which can be regularized by applying a scattering length regularization @xcite .
this has been established for the homogeneous gap equation and subsequently extended to the inhomogeneous gap equation , where ref .
@xcite gives a careful derivation for confined systems .
however , refs .
@xcite state that a much simpler regularization is sufficient since the results are not sensitive to the details of the method used .
the corresponding regularized gap equation reads @xmath68 which can be rewritten as a multiplication by a factor @xmath69 thus , each quasiparticle state @xmath70 is weighted by a factor @xmath71 . in order to extend this in a consistent manner to the nonequilibrium case ,
in which @xmath72 deviates from its ground state value @xmath73 and is determined by the full eq .
, the same procedure has to be applied to the nonequilibrium version of eq .
as will be discussed below [ see eq . ] . and thus the ground state gap @xmath74 in the dynamical equations .
since in the dynamical case the time - dependent gap @xmath75 differs from the ground state value this may seem unusual .
however , we have checked that the method is not sensitive on the detailed parameters used , e.g. , on using one constant value for all @xmath74 in order to omit the newly introduced dependence .
thus the proposed regularization scheme is a suitable extension since the physical behavior of the system remains unchanged . ]
figure [ fig : dispersion ] shows the dependence of the quasiparticle energies @xmath76 on the one - particle energies @xmath61 . for all subbands crossing the fermi energy finite minima of the quasiparticle energy evolve at @xmath77 .
subbands with the minimum close to the fermi energy ( i.e. , @xmath78 ) exhibit a larger @xmath74 due to the larger interaction matrix elements @xmath66
. this leads to a rather shallow parabolic - like minimum in these subbands and , thus , to a maximum in the corresponding density of states ( see fig . [
fig : densityofstates ] ) .
subbands showing this feature will be called resonant in the following and systems in which such a subband exists are referred to as resonant systems .
the behavior can be compared to the well - known parabolic - like dispersion relation in the homogeneous bcs theory with @xmath79 .
a similar behavior has also been obtained for nanostructured superconducting systems @xcite .
overall the density of states combined with the interaction matrix @xmath66 leads to the quantum size oscillations of the order parameter upon changing the lateral size of the system ( i.e. , @xmath5 ) found in ref .
@xcite .
in this paper we consider an excitation of the fermi gas by a quantum quench , i.e. , a sudden change of the coupling constant @xmath80 . since in the region of a feshbach resonance
the coupling constant strongly depends on an external magnetic field @xmath81 , this can be achieved experimentally by rapidly switching @xmath81 from an initial value @xmath82 to a final value @xmath83 .
we assume that this switching process occurs on a time scale much faster than the characteristic time scale of the order parameter dynamics , such that the excitation can be taken to be instantaneous .
this assumption is realistic since fast linear magnetic ramps with rates of @xmath84g / ms are experimentally available @xcite while the timescale of the gap dynamics is of the order of @xmath85ms @xcite .
the excitations considered in this paper require a shift in the magnetic field of a few gauss , which indeed can be assumed to be instantaneous on the typical ms time scale in ultracold fermi gases . in this case during the switching of the magnetic field the state of the system remains unchanged .
as usual , the dynamics of a quantum mechanical system can be described in different basis systems , which from a mathematical point of view are all equivalent . in our case , to calculate the dynamics of the order parameter after a quench from the _ initial _ system ( @xmath86 , @xmath87 , @xmath88 ) to the _ final _ system ( @xmath53 , @xmath54 , @xmath25 ) we choose a time - independent basis rather than remaining in the diagonal basis . for convenience
we take the basis corresponding to the eigenstates of the system after the switching , i.e. , to the coupling constant @xmath25 .
all our calculations are thus carried out in the basis @xmath53 , @xmath54 .
the initial state , which is characterized by the ground state order parameter @xmath89 , corresponding to the coupling constant @xmath88 and the basis functions @xmath86 , @xmath87 , has to be expressed in terms of the basis @xmath53 , @xmath54 , which in particular gives rise to non - vanishing quasiparticle excitations in this basis .
since the confinement potential is unchanged , also the corresponding one - particle wave functions @xmath55 remain unchanged . according to eq .
, in the present case only the bdg amplitudes @xmath57 and @xmath58 change while the spatial shapes of @xmath53 and @xmath54 remain unchanged .
therefore , all orthogonality relations are preserved and only diagonal quasiparticles are populated . for the initial values of the normal and anomalous expectation values , respectively , one finds @xmath90 in addition @xmath91 holds for all times @xmath92 .
since the instantaneous order parameter @xmath93 deviates from the ground state value of the final system @xmath73 , the hamiltonian in the basis @xmath53 , @xmath54 becomes non - diagonal depending on the difference @xmath94 .
it thus becomes explicitly time dependent according to @xmath95 \gamma_{ma}^{\dagger}\gamma_{na}^ { } \notag \\ & + \sum_{m , n } \left[\left(\delta-\delta_{gs}\right)_{u_m^ * u_n^ * } - \left(\delta^*-\delta_{gs}^*\right)_{v_m^ * v_n^ * } \right ] \gamma_{ma}^{\dagger}\gamma_{nb}^{\dagger } \notag \\ & - \sum_{m , n } \big[\left(\delta-\delta_{gs}\right)_{v_m v_n } - \left(\delta^*-\delta_{gs}^*\right)_{u_m u_n } \big ] \gamma_{mb}^{}\gamma_{na}^ { } \notag \\ & - \sum_{m , n } \left[\left(\delta-\delta_{gs}\right)_{v_m u_n^ * } + \left(\delta^*-\delta_{gs}^*\right)_{u_m v_n^ * } \right ] \left(1-\gamma_{mb}^{\dagger}\gamma_{nb}^{}\right ) .
\end{aligned}\ ] ] with @xmath96 v_n({\mathbf{r } } ) d^3 r \notag \\ & \underbrace{=}_{\text{a.a . } } \left(\delta - \delta_{gs } \right)_{m } u_m^ { } v_m^ { } \delta_{mn}^{}. \end{aligned}\ ] ] here , the anderson approximation has been applied to the dynamical equations as proposed in ref .
@xcite , yielding @xmath97 .
the time evolution of the system is thus described by the time - dependent quasiparticle expectation values .
the corresponding equations of motion can be obtained via heisenberg s equation of motion . for the considered instantaneous change of the coupling constant only diagonal expectation values
are excited .
the required equations of motion read @xmath98 where @xmath99 , \label{eq : e_ren}\ ] ] @xmath100 and @xmath101 v_{mk } \chi_k .
\end{aligned}\ ] ] in eq .
again the regularization factor @xmath102 has been introduced .
equations - represent a finite set of coupled ordinary differential equations that we solve numerically .
it is interesting to note that the evolution of the anomalous expectation values [ eq . ] corresponds to a set of harmonic oscillators with energies approximately given by @xmath103 [ first term in eq . ] while eq . as well as the second terms of eqs .
and contain nonlinear couplings to all other oscillators via the factor @xmath104 , which vanishes when the order parameter agrees with its ground state value .
we will come back to this separation into linear and nonlinear terms in sec .
[ sec : linearized ] .
in the following the temporal evolution of the amplitude of the spatially averaged gap @xmath105 will be shown and analyzed for different system parameters . here , the normalization volume @xmath106 is set to @xmath107 with @xmath108 being the oscillator length in @xmath45 direction . in order to concentrate on the physics
we will start our analysis by investigating a very small system : all the main features occurring in the dynamics of larger , experimentally accessible confinements arise in small systems , too , but with a strongly reduced degree of numerical complexity .
thus , the dynamics of the superfluid gap will at first be explained on the basis of small systems .
the results for a larger system will be shown afterwards .
an exemplary result for the gap dynamics after a quantum quench , obtained by changing the scattering length from @xmath109 nm to @xmath110 nm for a system with the confinement frequencies @xmath111khz and @xmath112hz , is shown in fig .
[ fig : dynamik+ft ] .
the fermi energy has been set to @xmath113 yielding @xmath114 according to ref .
as can be seen the amplitude of the gap shows an initial drop corresponding to the decreased coupling and thus decreased ground state gap .
afterwards a smoothly damped oscillation around the new ground state value of the gap occurs , which after a certain transition time @xmath115 turns into an irregular , rather chaotic oscillation . here
@xmath115 is defined as the time of the first deviation is assumed to be reached if the distance between two adjacent maxima shows a deviation of more then 50@xmath116 compared to the averaged distance of all preceding maxima . ] from a smooth oscillation .
nm to @xmath117 nm ; inset : fourier transform of the gap dynamics .
the confinement frequencies are @xmath118khz and @xmath119hz . ]
the inset of fig .
[ fig : dynamik+ft ] suggests that this irregular oscillation after @xmath115 is the result of a superposition of several frequencies . here
a segment of the fourier spectrum of the gap dynamics is shown .
the spectrum is khz .
the arrows mark the transition time @xmath115 . ] composed of a series of sharp peaks in the range of @xmath120pev to about @xmath121pev , which can each be assigned to a corresponding quasiparticle state , i.e. , @xmath122 .
the main peaks at the lower end of this series correspond to the frequency of the initial damped oscillation and to the dominant frequencies of the irregular dynamics afterwards .
their values are given by the quasiparticle energies closest to the fermi level .
these lie in the vicinity of a quasiparticle subband minimum .
the corresponding frequencies are thus given by @xmath123 , where @xmath124 is the @xmath9 quantum number referring to the state with minimal quasiparticle energy , i.e. , the state at the fermi energy .
the other peaks belong to higher quasiparticle states and decrease continuously with increasing energy .
while the qualitative picture of the gap dynamics is the same for all investigated systems , the quantitative values of the features mentioned above crucially depend on the system parameters . on the one hand the ground state gap and thus the mean value of the oscillation and
its frequency contributions strongly depend on the perpendicular confinement @xmath125 ( due to the size - dependent superfluid resonances @xcite ) and on the scattering length @xmath27 .
the transition time , on the other hand , increases with decreasing parallel confinement @xmath126 i.e . , with increasing system length as can be seen in fig .
[ fig : uebergangszeit_entwicklung ] . here
the gap dynamics is shown for the same perpendicular confinement and excitation as in fig .
[ fig : dynamik+ft ] but for increasing system length , i.e. , decreasing @xmath127 ( from bottom to top ) .
figure [ fig : uebergangszeit_entwicklung ] shows that the transition time @xmath115 moves to larger times as the length of the system increases .
in addition a revival of the oscillation can be seen for the two largest systems with @xmath128hz and @xmath129hz , which for smaller systems would occur after the breakdown . a quantitative analysis of the transition time for different perpendicular confinements @xmath125 over a wide range of parallel confinements
is shown in fig .
[ fig : uebergangszeit ] . here
the transition time is plotted against the inverse parallel confinement frequency @xmath130 .
one can see that @xmath115 is independent of the size of the gas in the @xmath7-@xmath8-plane , since the values for every @xmath125 lie on the same curve .
it is only influenced by the confinement in @xmath9-direction , where a linear dependence on @xmath131 can be observed .
this is in full agreement with the behavior found for superconducting quantum wires @xcite .
even a beating - like pattern was found for thin quantum wires which corresponds to the revivals seen in fig .
[ fig : uebergangszeit_entwicklung ] . compared to the prediction from the linearized theory ( dashed curve ) . ] to investigate the smooth regime of the gap dynamics fig .
[ fig : resonant_vergleich ] shows calculations for a rather large system length and two different perpendicular confinements . for this case of large lengths in ref .
@xcite it was found that thick quantum wires exhibit a damping of the gap oscillation given by a power law @xmath132 with @xmath133 for resonant systems and @xmath134 for off - resonant ones . however , thin quantum wires were found to differ from this power law showing a more irregular oscillation but still a rather fast decay of the gap oscillation when resonant subbands are present .
khz ; red curve ) and an off - resonant system ( @xmath135khz ; blue curve ) ; in both cases @xmath136hz .
inset : fourier spectra . ]
figure [ fig : resonant_vergleich ] shows that this situation applies to ultracold fermi gases as well . here a calculation of the gap dynamics is shown for a resonant system ( upper , red curve ) , which is again characterized by the same perpendicular confinement as in figs .
[ fig : dynamik+ft ] and [ fig : uebergangszeit_entwicklung ] , as well as for a system far away from resonance ( lower , blue curve ; @xmath135khz ) .
the excitation is the same as before and the parallel confinement frequency is chosen as @xmath137hz , which corresponds to a rather long cloud .
it can be seen that both systems show an initial decay of the gap oscillation until a minimal amplitude is reached . in the resonant case
this initial decay is rather strong and fast . here
, the amplitude of the oscillation falls close to zero .
afterwards it exhibits revivals until the smooth oscillation breaks down .
in contrast , the off - resonant system shows an only moderate , comparatively slow decay of the oscillation , which after a short time exhibits a nearly constant amplitude .
thus , on the one hand the decay of the oscillation is much stronger in the resonant than in the off - resonant case . on the other hand revivals and
a beating like pattern occur for the resonant case before the breakdown while systems far away from resonance exhibit a nearly constant oscillation amplitude . the inset of fig .
[ fig : resonant_vergleich ] shows that these different temporal evolutions correspond to different fourier spectra . here the fourier transforms of the resonant ( positive @xmath8-axis ) and the off - resonant ( negative @xmath8-axis ) system are shown .
the resonant spectrum is composed of several strong components at the lower end and weaker peaks towards higher energies .
in contrast , the off - resonant spectrum it is shifted due to the smaller gap contains only one dominant frequency part at low energies while the rest of the spectrum is strongly suppressed . thus , resonant systems on the one hand correspond to spectra with several strong modes and weaker high - frequency parts .
off - resonant systems on the other hand are strongly dominated by one single mode with only minor contributions from other energies .
the features described above the damped oscillation , the breakdown , and the irregular dynamics will be explained in the following section . as mentioned before
, they also occur in bcs - superconductors in a very similar way , where cigar - shaped fermi gases correspond to thin , short quantum wires .
the following explanations therefore apply to both ultracold fermi gases and confined bcs - superconductors thus showing the close relation between these systems . to analyze the mechanisms underlying the gap dynamics and its features we introduce a linearized set of equations of motion .
this can be derived by neglecting all terms of second and higher order in the quasiparticle excitations ( this is valid due to the weak excitation investigated in this paper , which leads to @xmath138 ) . in doing so , eq .
can be neglected since by inserting eqs . and into this equation
only terms of at least second order in the normal excitations or products of anomalous and normal excitations contribute , which are to be neglected in the linearized case .
khz and @xmath119hz . ]
equation reduces to a closed set of equations for the anomalous excitations which , by performing the derivative in time of one equation and inserting the other , can be separated into its real and imaginary parts leading to @xmath139\chi_k , \hspace{0.01 cm } \label{eq : bwg - gl - lin } \end{aligned}\ ] ] where again all terms nonlinear in the quasiparticle expectation values have been neglected .
this equation describes a set of linearly coupled harmonic oscillators with the uncoupled frequencies @xmath140 and @xmath141 the coupling strength of the oscillators @xmath142 with @xmath143 therein is weak due to the in this case mostly small matrix elements @xmath144 .
the shift of the eigenfrequencies of the coupled system [ eq . ] with respect to the uncoupled frequencies [ eq . ] should therefore be small , as should be the shift of the uncoupled frequencies compared to the bare ones @xmath145 .
figure [ fig : linear_vergleich ] shows the dynamics of the bcs gap obtained from eq .
compared to the full dynamics .
, in the dynamical calculations ( i.e. , in eqs . , and ) the coupling has to be slightly lowered by a small amount , i.e. , @xmath146 .
it should be mentioned that the full ( nonlinear ) equations never lead to divergences . ]
the parameters correspond to the system shown in fig .
[ fig : dynamik+ft ] and are exemplary for all investigated systems .
the linearized equations clearly reproduce the full dynamics and all its features in very good agreement .
the inset shows that the positions as well as the strengths of most of the frequencies in the fourier spectra are well described by this approximation .
the spectrum corresponding to the full equations of motion ( upper , red curve ) and the one corresponding to the linearized equation ( lower , black curve ) show only slight differences . only one low lying weak frequency component close to zero , which is present in the linearized version ( outside the range shown in the inset of fig .
[ fig : linear_vergleich ] ) as well as weak side peaks that occur adjacent to every main peak in the full dynamics are not fully reproduced .
the latter can be attributed to the nonlinear couplings .
their influence on the dynamics , however , is obviously negligible .
the main features of the gap dynamics can thus be explained on the basis of eq . , which can be solved analytically .
the analytic solution can be expressed in terms of a linear superposition of simple ( co)sine - oscillators .
the corresponding frequencies are determined by the uncoupled oscillator frequencies @xmath147 and the coupling @xmath142 .
they are thus completely fixed by the the system parameters and the bcs gap , but they are independent of the initial conditions . in the considered case of weak coupling
the coupled spectrum is only slightly shifted compared to the uncoupled frequencies .
the eigenfrequencies of the coupled dynamics are thus approximately given by twice the quasiparticle energies , as observed above .
the amplitudes of the different eigenmodes of the coupled system are determined by the initial values of the dynamics and thus depend on the details of the excitation . in general
, one observes that each of the quasiparticle oscillators @xmath148 carries strong contributions from oscillators in the vicinity of its uncoupled frequency @xmath147 and in areas of a high density of states .
these areas are located close to the minimum of a quasiparticle subband ( see section [ sec : formalism ] ) , i.e. , near @xmath149 .
in addition one observes that the contributions from each oscillator to the lowest energies are mostly in phase while these to higher energies are more and more out of phase .
a sum of all quasiparticle oscillators , which according to eq .
yields the bcs gap , thus leads to dynamics dominated by low - lying frequencies at a quasiparticle band minimum .
these can be understood as cumulative peaks created by collective oscillations of all quasiparticles in the system .
all spectra obtained by the full equations indeed show such a spectrum with rather dense , strong frequency contributions near @xmath150 and relatively widespread suppressed frequency contributions at higher energies .
the explanations given above have shown that the gap dynamics can be understood as a linear superposition of quasiparticle oscillations which themselves are given by a sum of simple oscillators . to finally explain the main features of the gap dynamics in the time domain the damping , the transition and the irregular oscillation one can therefore use an even simpler picture of a set of independent cosine - oscillators with the frequencies @xmath147 . due to the abrupt excitation
all these oscillators start in phase at maximum deflection . starting to oscillate they will soon dephase .
a sum of all oscillators ( i.e. , the gap ) thus performs a damped oscillation ( cf .
@xcite ) . here
systems with several strong frequency contributions i.e . , resonant systems show a rather fast and persistent damping since a large part of the cumulative amplitude is able to dephase . off - resonant systems in contrast exhibit only a slight decay of the oscillation since the main part of the cumulative oscillation is carried by one single mode . proceeding in time the damped oscillation continues until all oscillators are completely dephased and the amplitude of the oscillation is minimal .
then the oscillators start to rephase , the amplitude grows and the oscillation reappears ( see fig .
[ fig : uebergangszeit_entwicklung ] ; for off - resonant systems this effect is strongly suppressed since one single frequency dominates the spectrum ) .
as soon as the first adjacent pair of oscillators goes back in phase again this beating - like pattern is interrupted .
at this moment a spike in the cumulative oscillation indicates the breakdown of the regular oscillation and thus the transition time @xmath115 . afterwards
all other frequencies rephase successively and create a rapid sequence of spikes which leaves a picture of an irregular oscillation .
the preceding argumentation suggests that the time of this breakdown should be inversely proportional to the maximum spacing of adjacent quasiparticle energies . strictly speaking the transition times are found to be determined by adjacent quasiparticles from the same subband @xcite , i.e. , @xmath151 it thus increases @xmath152 with decreasing parallel confinement frequency since the energy spacing of the atomic spectrum then decreases .
this is in full agreement with the relation found before .
in fact , eq . gives exactly the linear curve shown in fig . [
fig : uebergangszeit ] .
this indicates that the breakdown of the smooth initial oscillation of the bcs gap indeed is due to adjacent frequencies rephasing in time . as a final example the gap dynamics of a rather large system with @xmath153hz and @xmath154khz again after a sudden change of the scattering length from @xmath155 nm to @xmath117 nm
is shown in fig .
[ fig : dynamik_gross ] .
this corresponds to a gas with @xmath156 and @xmath157 and is thus on an experimentally accessible length - scale @xcite .
the fermi energy is chosen as @xmath158 which corresponds to @xmath159 atoms in the trap .
nm to @xmath117 nm ; inset : fourier transform of the gap dynamics .
the confinement frequencies are @xmath160khz and @xmath161hz . ]
figure [ fig : dynamik_gross ] shows that the qualitative behavior of the gap dynamics is the same as for the smaller systems : a slowly decaying oscillation of the gap occurs .
the transition time @xmath162 calculated from eq .
is in good agreement with a small bump in the curve at @xmath163ms , the first deviation from a smooth oscillation .
afterwards more and more deviations occur and the gap dynamics becomes successively irregular .
when looking at the damping of the gap oscillation one can see that although the system is resonant one subband is close to the fermi energy the strength of the damping is situated somewhere between the resonant and the off - resonant case of fig .
[ fig : resonant_vergleich ] .
this is due to the weak perpendicular confinement and therefore large number of states contributing to the condensate : compared to the overall number of relevant states the resonant ones give only a small contribution to the gap . for larger systems
the resonances are thus less pronounced @xcite .
the fourier transform in the inset of fig .
[ fig : dynamik_gross ] shows a familiar picture , too , with strong contributions at the lower end of the spectrum and successively decaying peaks towards higher energies .
the difference with respect to the spectra shown before is on the one hand the high density of peaks for the large system .
this is due to the weaker confinement and thus higher density of bare and quasiparticle states . on the other hand
the gap in the fourier spectrum and thus the main frequency of the oscillation is comparatively small .
this is because of the larger ratio of the perpendicular to the parallel length @xmath164 . with the fermi energy fixed at @xmath165 this leads to a comparatively low particle density of the trapped gas and thus a weaker condensate and a smaller gap .
in conclusion , we have calculated the higgs amplitude dynamics in the bcs phase of an ultracold @xmath166li gas confined in a cigar - shaped trap .
the dynamics is induced by a quantum quench resulting from a sudden change of an external magnetic field .
we have shown that the amplitude of the spatially averaged gap performs a damped oscillation breaking down after a certain time @xmath115 , which is determined by the parallel confinement frequency @xmath127 , i.e. , by the length of the cloud . afterwards
a rather irregular oscillation involving many different frequencies occurs .
we have investigated the influence of the confinement on the gap dynamics and the impact of the size - dependent superfluid resonances on its qualitative behavior .
it turned out that in the case of a resonant system , i.e. , a system where the fermi energy is close to a subband minimum , the dynamics of the order parameter exhibits a strong damping and , for sufficiently long systems , a revival before eventually the irregular regime is reached .
in contrast , in an off - resonant system the damping is much less pronounced and the oscillation before the transition to the irregular regime is mainly determined by a single frequency . by analyzing the linearized version of the equations of motion for the quasiparticle excitations we were able to interpret the observed features of the dynamics .
it turned out that for the excitations studied in this paper the linearized equations well reproduce the dynamical behavior of the gap , except for some slight details resulting from the nonlinearities in the full equations of motion . from the linearized model it becomes evident that the system approximately behaves like a set of weakly coupled harmonic oscillators .
the frequencies as well as the couplings of these oscillators are completely determined by the system parameters after the excitation while the amplitudes of the different eigenmodes depend on the details of the excitation . the analysis revealed in particular that the transition time to the irregular dynamics is directly related to the energy separation of the one - particle energies while the differences between resonant and non - resonant systems is caused by the different densities of states and coupling efficiencies close to the subband minima . | the higgs amplitude mode of the order parameter of an ultracold confined fermi gas in the bcs regime after a quench of the coupling constant is analyzed theoretically .
characteristic features are a damped oscillation which at a certain transition time changes into a rather irregular dynamics .
we compare the numerical solution of the full set of nonlinear equations of motion for the normal and anomalous bogoliubov quasiparticle excitations with a linearized approximation . in doing so the transition time as well as the difference between resonant systems , i.e. , systems where the fermi energy is close to a subband minimum , and off - resonant systems
can be well understood and traced back to the system and geometry parameters . |
plasma rotation has become one of the key ingredients of fusion plasma performance , after the discovery of a reduction of turbulence and transport @xcite through a sheared @xmath2 rotation .
with regard to stability , large toroidal rotations can stabilize resistive wall modes and neoclassical tearing modes .
for these reasons , the physics of toroidal momentum transport , and the so - called intrinsic rotation , have become a very active area of research , mainly for their importance in the operation of a reactor plasma , for which the external torque will be negligible @xcite .
thus , reliable rotation measurements are requiered . in this sense , the charge exchange recombination spectroscopy ( cxrs ) technique @xcite has developed into a basic tool for diagnosing plasma rotation over the last three decades , and will be a key diagnostic for iter @xcite .
one of the basic results of the kinetic theory of strongly magnetized plasmas with small ( compared to thermal ) @xmath2 velocity , is that first order ion and electron flows are parallel to flux surfaces and incompressible @xcite , _ i.e. _ , their velocity fields satisfy @xmath3 .
this can be understood as a consequence of particle number conservation @xmath4 with flux - constant density @xmath5 and zero first - order radial velocity @xmath6 ( here , @xmath7 is a radial coordinate labelling a magnetic surface ) .
such a velocity field is fully determined by two flux functions ( _ i.e. _ , that depend on spatial coordinates through @xmath7 only ) .
note that the incompressibility of flows follows from the usual neoclassical ( and gyrokinetic ) ordering schemes but it neither assumes nor implies the flows to be of neoclassical origin. indeed , the spatial variation of an incompressible flow has been recently employed to estimate the poloidal velocity from inboard and outboard toroidal measurements in tcv and diii - d @xcite .
nevertheless , in the alcator c - mod @xcite and asdex - upgrade @xcite tokamaks , it has been observed that as the main ion pressure increases , a significant poloidal variation of the impurity density appears .
indeed , a redistribution of the density of impurities on flux surfaces is expected to be driven by the parallel friction caused by large ion diamagnetic velocities @xcite , making impurity flows compressible .
there is no clear picture concerning the neoclassical ( nc ) nature of poloidal flows in tokamaks .
early measurements on tftr @xcite showed a large discrepancy with the nc theory .
however , a deeper understanding of cxrs measurements @xcite led to better agreement between experimental poloidal flows and nc expectations on jt-60u , mast or nstx ( see @xcite , and references therein ) , in h - mode plasmas or even in the presence of an internal transport barrier ( itb ) .
moreover , recent measurements on alcator c - mod @xcite and asdex - upgrade @xcite have been demonstrated to follow nc theory in the absence of impurity density asymmetries .
nevertheless , poloidal velocity measurements on jet within an itb @xcite , and on diii - d during h - mode and quiescent h - mode @xcite , are still not fully understood . on the other hand ,
toroidal flows in tokamaks are expected to be dominated by mechanisms other than neoclassical , since toroidal viscosity vanishes for axisymmetric systems .
nevertheless , the presence of a sizeable toroidal ripple and/or non - axisymmetric magnetic perturbations for elm control in tokamak reactors could induce a non - negligible toroidal viscosity .
this demands experimental studies on rotation in stellarator devices , since in such machines plasma can not rotate freely @xcite , and the radial electric field is determined by the ambipolarity condition on the nc radial particle fluxes .
detailed comparisons of the nc radial electric field with cxrs measurements have been performed on the w7-as stellarator , showing reasonable agreement for most of the plasma scenarios studied @xcite .
in addition , in the lhd and chs stellarators , qualitative agreement was found between radial electric field measurements and calculations with simplified nc theories @xcite .
the effect of the magnetic field ripple on the toroidal flows was also investigated in chs @xcite and lhd @xcite , showing qualitative effects of helical and toroidal ripple on spontaneous toroidal flow .
finally , a recent work in hsx has shown a clear discrepancy between the experimental and nc radial electric fields @xcite .
the authors claim that the monoenergetic approximation used by dkes might be inadequate for radial electric fields close to the helical resonance .
nonetheless , the parallel flow was well reproduced after including momentum - correction techniques and impurities in the calculations . in the tj - ii stellarator
@xcite , the radial electric field has been studied by means of the heavy ion beam probe ( hibp ) @xcite and doppler reflectometry @xcite diagnostics .
measurements from the former showed consistency with nc calculations in the electron root , within a factor a two @xcite
. however , the experimental values systematically exceed the nc predictions in the ion - root operation @xcite .
in addition , passive spectroscopy has been utilized to measured poloidal and toroidal rotations of impurities @xcite .
qualitative agreement was reported between the experimental poloidal flow and nc values , mainly in the change of root as density increases @xcite .
in contrast , significant discrepancies were found in the toroidal flows for the different species studied and with nc calculations @xcite . in this work
we use cxrs to measure fully ionized carbon impurity flows in low density plasmas of the tj - ii stellarator and compare them with neoclassical calculations .
first , we present in section [ sec : experimental_setup ] the diagnostic technique , together with a general treatment of the sightlines and flow geometry , which allows the extraction of the two flux - surface averaged ( fsa ) flows from two velocity measurements .
these flux constants define an incompressible flow within a flux surface and are related to the radial electric field and the bootstrap velocity . in section [ sec :
incompressible_flows ] , the method is applied to two pairs from three measurements ( poloidal , outboard - toroidal and inboard - toroidal ) to show that the spatial variation of the flow is consistent with incompressibility .
finally , the measured flows are compared with neoclassical predictions ( section [ sec : comparison_nc ] ) and good agreement is found for the radial electric field ( in all the cases studied ) and bootstrap flows ( in ecrh plasmas without momentum injection ) .
the tj - ii is a four - period heliac - type stellarator with major radius of @xmath81.5 m and averaged minor radius of @xmath90.22 m @xcite .
hydrogen plasmas are created and heated by 2 gyrotrons operated at 53.2 ghz ( p@xmath10600 kw ) . during this heating phase central electron densities ,
n@xmath11 , and temperatures , t@xmath11 , up to @xmath12 m@xmath1 and @xmath13 kev , respectively , are achieved .
in addition , plasmas can be maintained by two tangential neutral beam injectors ( nbi ) , each providing @xmath14500 kw . as a result ,
plasmas with n@xmath15 m@xmath1 and t@xmath16400 ev are attained .
the machine is operated with lithium wall coating , which allows density control , but results in a strong reduction in impurity concentrations and hence in impurity photon emission fluxes . in order to undertake cxrs measurements in such conditions ,
a high - throughput optical system and spectrograph , plus a high - efficiency detector , are needed @xcite .
these are described next .
the cxrs process of interest in tj - ii involves electron capture from accelerated hydrogen by fully ionized carbon ions into a highly excited state of c@xmath17 , followed by spontaneous decay via photon emission , _
i.e. _ the c vi line at 529.07 nm ( @xmath18 ) . for this , a compact diagnostic neutral beam injector ( dnbi ) provides a 5 ms long pulse of neutral hydrogen accelerated to @xmath19 kev ( the ratio of its full , half and third energy components is @xmath20 ) @xcite .
its @xmath21-radius at focus is @xmath22 mm . for light collection ,
a bidirectional optical system , consisting of commercial camera lenses and 12-way fibre optic bundles , collects the c@xmath17 light across the outer half of the plasma diameter with @xmath23 cm spatial separation between lines of sight .
these components are installed in top and bottom access ports located in the same machine sector as the dnbi .
a second optical system with a 16-way fibre bundle , ( 12 equally spaced fibres along a row plus four additional fibres that straddle the row on both sides ) , views almost the complete beam / plasma interaction volume from a nearby tangential viewing access port .
this provides @xmath24 cm spatial separation between sightlines .
note : a retractable in - vacuum mirror located in the mouth of the access port is needed to redirect light to the lens as the port is shared with a neutral particle analyser diagnostic .
schematic diagram of cxrs diagnostic sightlines for the standard magnetic configuration
. only 8 poloidal top lines of sight are shown , as the outboard 4 do not provide useful data for this configuration .
inboard and outboard toroidal measurements are highlighted .
in addition , poloidal cuts of magnetic surfaces are presented , together with the directions of concern for velocity measurements : radial ( @xmath25 ) , parallel ( @xmath26 ) and @xmath27 . note that none of these vectors is contained in the poloidal plane .
the magnetic field vector , @xmath26 , points into the page.,width=7 ] a schematic layout of the diagnostic sightlines is depicted in figure [ fig : sight - lines ] .
as observed , poloidal cuts of tj - ii magnetic surfaces are bean shaped and exhibit large flux compressions within a surface .
the plasma minor radius region scanned by nearly symmetric poloidal views is @xmath28 in the standard magnetic configuration , @xmath29 ( where the nomenclature reflects currents in the central , helical and vertical field coils , respectively ) . here , a normalised radius @xmath30 is defined , where @xmath31 and @xmath32 are the volumes enclosed by the surface of interest and the last closed magnetic surface , respectively . as explained in section [ sec : data_analysis ] , poloidal top fibres are used to correct for uncertainties in the c@xmath17 rest emission wavelength , and thus , do not provide additional information about the c@xmath33 distribution function . on the other hand ,
the toroidal fibres cover both sides of the magnetic axis , from @xmath34 to @xmath35 at 10 locations ( 2 of the 12 sightlines are obstructed by the vacuum chamber ) .
the region in which both poloidal and toroidal measurements are taken is labelled as outboard , in analogy with tokamaks ; while the zone where only toroidal measurements are made is labelled as inboard .
the nomenclature @xmath36 and @xmath37 is also utilized to define these regions . in the outboard region ,
poloidal and toroidal fibres view the same surfaces at @xmath38 and @xmath39 .
therefore , the 2d - flow velocity is completely determined at these locations .
the redundant inboard - toroidal measurements will be used in section [ sec : incompressible_flows ] to verify whether the incompressibility assumption holds or not .
light collected by all systems is transferred to the input of a light dispersion element , an holographic imaging spectrograph , modified with three @xmath40m - wide curved entrance slits ( to compensate for the short focal length ) and equipped with a transmission grating sandwiched between two bk7 prisms .
it provides a focal - plane dispersion of @xmath41 nm / mm at @xmath42 nm @xcite .
in addition , a narrow bandpass filter ( @xmath43 nm at full - width at half - maximum ) prevents spectral overlapping of the light from the three fibre arrays at the image plane .
the set - up also includes a high - efficiency back - illuminated ccd camera and a fast mechanical shutter ( @xmath44 ms time window ) . using on - chip binning ,
multiple spectra are collected during discharges ( @xmath45 ms ) . in order to obtain neutral
beam induced spectral line data , and hence spatially localized information , it is necessary to remove background -passive- c@xmath17 light from spectra .
this is done by alternatively injecting and not injecting the dnbi into reproducible plasmas ( shot - to - shot technique ) @xcite ) .
then , by subtracting a background spectrum from the stimulated plus background -active plus passive- spectrum the cxrs data is obtained . for ecrh plasmas ,
the active emission is a factor of @xmath46 smaller than the passive one , see figure [ fig : raw_signals ] . in nbi
heated plasmas , this ratio is reduced to a factor of @xmath47 .
a multi - gaussian fit to the active signal is also shown , and will be explained in section [ sec : data_analysis ] .
dnbi - excited raw c - vi spectral line data ( active plus passive ) and background emission ( passive ) , for two reproducible ecrh discharges
. the active signal is obtained from the difference .
a multi - gaussian fit to the latter is presented.,width=7 ] correctly performed instrument wavelength calibration and correction are essential if experimental uncertainties are to be minimized .
indeed , wavelength calibration is done regularly during tj - ii operation , as thermal variations in the instrument can be significant , _
e.g. _ , @xmath48 @xmath49 m k@xmath50 at the focal plane , this being equivalent to an offset of @xmath51 km s@xmath50 k@xmath50 . for this
, spectral lines from a neon pencil - type lamp are used to establish the wavelength dispersion at each fibre location on the focal plane @xcite .
finally , several well - studied effects are negligible for tj - ii or require correction before the doppler shift of the beam induced c vi spectral lines can be determined .
these effects , together with machine specific effects such as magnetic geometry and beam / sightline volumes , are handled by an in - house software package developed to analyse the data and obtain impurity ion temperature and velocities .
these are described in more detail in section [ sec : data_analysis ] .
alignment of the bidirectional poloidal sightlines is performed by illuminating each fibre bundle with an intense light source and observing , through a nearby view - port , the orientation and location of the resultant bright spots with respect to markings on the inside of the opposing vacuum flange . from these
, the sightline paths through the neutral beam can be determined using a cross - sectional machine drawing and magnetic configuration maps .
on the other hand , alignment of the tangential viewing sightlines involves aligning the in - vacuum mirror as well as orienting the fibre bundle ferrule .
for this , a preliminary alignment is made by illuminating the fibres with the intense light source and observing , through a view - port located in a sector downstream of the neutral beam sector , the location of bright spots on the inside of the vacuum chamber . then
, after removing the light source , the fibre ends that attach to the spectrograph input fibre bundle , are connected to avalanche photodiodes ( apd ) whose bias voltages were previously adjusted to a common gain .
then , by repeatedly injecting the dnbi into the tj - ii chamber , the mirror and fibre bundle are adjusted in - turn until all apd signals are maximized . moreover , as an additional check , light from the sightlines that straddle the 12 principal sightlines are checked for symmetry about the beam centre . finally , by repeating the fibre illumination procedure described before for the toroidal fibres , the location of bright spots on the vessel inboard wall are noted and cross - checked .
then , using a 3 dimensional computer aided design ( cad ) software with the geometrical coordinates of the vacuum chamber , neutral beam and sightlines as inputs , the beam / sightline crossing points are found for the magnetic configuration of interest . in figure
[ fig : raw_signals ] , a multi - gaussian fit to an active signal obtained with the shot - to - shot technique is made using the expression : @xmath52 where @xmath53 is the relative amplitude of each of the fine structure components of the c@xmath17 spectral line when assuming a statistical population of the initial _ j_-levels , and normalized to the sum , @xmath54 .
@xmath55 stands for the wavelength separation of the fine - structure components respect to the rest wavelength : @xmath56 in vacuum @xcite .
in addition , @xmath57 measures the line centroid .
carbon temperatures are obtained from the parameter @xmath58 , after considering zeeman broadening and a numerical deconvolution of the instrumental function @xcite .
a representative temperature profile of the ecrh phase is displayed in figure [ fig : ecrh_t_profile ] .
it is found that cxrs temperature measurements from the three fibre arrays coincide , within error bars , on both sides of the magnetic axis .
indeed , the consistency of the temperature measurements from the three fibre arrays is used usually as an additional check of the diagnostic optical alignment . for completeness , the majority ion temperature measured at @xmath59 by the poloidal neutral particle analyser ( npa ) , is shown .
fully - ionized carbon ion temperature profile for a representative ecrh discharge ( discharge @xmath60 , with @xmath61 m@xmath1 and @xmath62 kev ) , as measured by poloidal top ( blue circles ) , bottom ( red asterisks ) and toroidal sightlines ( black squares ) . the central majority ion temperature measured by the npa is shown as well ( triangles).,width=7 ] carbon impurity density measurements are more challenging , since the beam density , @xmath63 , needs to be known .
the latter can be roughly estimated by considering a simple pencil - like beam attenuation model @xcite , together with analytical expression to account for the beam profile in vacuum .
besides , the optical system is not absolutely calibrated at present
. however , an estimate can be made by assuming that the relative efficiency of each fibre is proportional to the total intensity recorded during the instrument calibration procedure , _
i.e. _ , @xmath64 where the integral of the beam atoms is performed in the sightlines measurement volume , @xmath65 .
finally , the velocity along sightlines , @xmath66 , is obtained from the line centroid shift , @xmath57 : @xmath67 here , @xmath68 accounts for possible deviations in the populations of the fine structure level from a complete statistical mixing , that might cause a false velocity .
this is removed by adding the redundant measured wavelengths from top and bottom poloidal sightlines , since they are almost symmetrically opposite .
it is well documented that the daughter distribution function of c@xmath17 ions may exhibit some undesired velocities , usually named _
pseudo - velocities _ , which arise from the energy dependence of the beam / carbon charge exchange cross - section as well as the gyromotion of the impurity during the de - excitation process @xcite . in tj - ii , due to the relatively small magnetic field strength and temperatures , the only significant contribution comes from the former , which is orientated in the dnbi direction .
these effects are only substantial for the toroidal view ( since the poloidal sightlines view the beam almost perpendicularly ) , being of the order of @xmath69 km s@xmath50 , at most .
the general form of the impurity velocity field is @xcite : @xmath70 where the perpendicular velocity is given by @xmath71 and diamagnetic contributions , and thus , @xmath72 here , @xmath73 and @xmath74 are the density , charge and pressure of the impurity , respectively . in addition , @xmath75 is the electric potential and @xmath76 denotes flux - surface average .
the last term in equation is the well - know pfirsch - schlter ( ps ) flow , which arises from the toroidicity of the magnetic field strength , and satisfies : @xmath77 in order to obtain flux - surface averaged ( fsa ) flows from poloidal and toroidal velocity measurements , it is convenient to define dimensionless vectors , which store the variation of the flow within the surface .
these are : @xmath78 note that the vector @xmath79 has both perpendicular and parallel components , the latter coming from the ps contribution : @xmath80 . then , impurity flows are expressed as : @xmath81 here , the following perpendicular and parallel fsa flows are defined : @xmath82 where the fsa radial electric field is given by : @xmath83 . finally , taking the projections of the velocity vector @xmath84 over the poloidal - bottom and toroidal sightlines ( @xmath85 and @xmath86 , respectively )
the fsa mean velocities are obtained from the measured flows : @xmath87 here , @xmath88 , and the sub - index ( p , t ) denote the projection of a vector ( @xmath89 or @xmath79 ) over the poloidal and toroidal lines of sight .
accordingly , the fsa flows can be obtained from two different measurements performed on a surface , but not necessarily at the same location .
the results of this procedure are presented in the next section .
dimensionless factors that store the variation of the flows on the surface for the ( a ) poloidal - bottom and ( b ) toroidal fibre arrays .
the vector @xmath79 has been split into their parallel and perpendicular components ( @xmath90 and @xmath91 , respectively).,title="fig:",width=7 ] dimensionless factors that store the variation of the flows on the surface for the ( a ) poloidal - bottom and ( b ) toroidal fibre arrays .
the vector @xmath79 has been split into their parallel and perpendicular components ( @xmath90 and @xmath91 , respectively).,title="fig:",width=7 ] the projection of the geometrical vectors , @xmath79 or @xmath92 , over the poloidal - bottom and toroidal sightlines is depicted in figure [ fig : geometrical_cxrs_factors ] .
the vector @xmath79 has been split into their parallel and perpendicular components , to illustrate their contributions on the velocity field .
as expected , the poloidal velocity is dominated by the perpendicular contribution , although the parallel one ( through the @xmath93 factor ) has to be accounted for .
the sample volumes located at the outboard are close to stagnation points of the ps flow and therefore its contribution to these measurements is modest , unlike in the inboard region where it can become dominant for toroidal velocities .
in addition , the mean values of a geometrical quantity , @xmath94 , have been obtained through an average over the measurement volume for a given fibre , @xmath65 : @xmath95 note that the integral should be weighted with the local carbon impurity density , rather than the electron density .
although the former is missing , we consider this approximation sufficiently good to obtain mean values .
the errors associated with geometrical quantities , @xmath96 , are given by its standard deviation in the measurement volume : @xmath97 , and are propagated through the code .
it can be observed that when moving towards the plasma centre , the dispersion in these factors increases , since the poloidal views become more perpendicular to the surfaces .
the errors in the results will consist of the statistical ones , plus those due to the average in the measurement volume .
profiles of @xmath98 , @xmath99 , @xmath100 and @xmath101 ( from top to bottom ) for the ecrh scenario , shot @xmath60 ( blue diamonds ) ; and the nbi case , discharge @xmath102 ( red circles).,width=264 ] in this work , two plasmas scenarios are considered , both being operated in the standard magnetic configuration .
the first one , heated by ecrh , is characterized by flat electron density ( @xmath61 m@xmath1 ) and peaked electron temperature profiles ( @xmath62 kev ) , see figure [ fig : plasma_profiles ] .
however , ions remain cooler , @xmath103 ev ( @xmath104 is assumed , see the comparison in figure [ fig : ecrh_t_profile ] ) , with slightly hollow temperature and impurity density profiles . the data presented here corresponds to tj - ii discharge @xmath60 , which is representative of several similar discharges that are briefly discussed at the end of this section .
in the second scenario , plasmas are heated with a tangential nbi injected in the direction of the toroidal magnetic field .
the resultant density profile is more peaked , with @xmath105 m@xmath1 , whilst the electron temperature decreases to @xmath106 ev .
in contrast , the ion channel reaches temperatures up to @xmath107 ev .
the discharge chosen for this case , @xmath102 , is not representative of fully developed nbi heated plasmas , as it corresponds to the first steps of the nbi phase , where the density grows rapidly .
nevertheless , it is included in this paper to highlight that for these low density plasmas the results are independent of the heating method .
finally , the measurements presented are restricted to the region @xmath108 , because cxrs measurements are not reliable at the edge of tj - ii , due to the poor statistics and the possible presence of suprathermal ions @xcite .
the assumption of an incompressible velocity field like the one in equation is widely used in present day cxrs data analysis . in this subsection
we verify this by applying the method developed in the previous section , equation , to the poloidal and toroidal velocities measured in the discharges presented in figure [ fig : plasma_profiles ] .
carbon impurity ion flows observed by toroidal ( black squares ) and poloidal bottom ( blue circles ) sightlines , for the : ( a ) ecrh plasma , discharge @xmath60 ; and ( b ) nbi plasma , discharge @xmath102.,width=7 ] the observed velocity pattern , after correcting for fine structure effects and pseudo - velocities , is represented in figure [ fig : raw_velocity ] for the ecrh and nbi heating scenarios under study .
poloidal top velocity measurements are not shown , since they were already used to correct for fine structure effects , and so do not provide further information about poloidal flow . in this
figure , positive velocities corresponds to particles moving away from the observer . in the ecrh plasma ,
the toroidal rotation is opposite to the magnetic field direction , while the toroidal flow is reversed in the nbi case .
in addition , ecrh toroidal measurements exhibit a strong in / out asymmetry that is explained next . as commented before , cxrs
toroidal velocity measurements are performed on both sides of the magnetic axis in tj - ii .
therefore , they are suitable for the verification of flow incompressibility , due to its redundancy on several magnetic surfaces . in figure
[ fig : fsa_flows ] the perpendicular , @xmath109 , and parallel , @xmath110 , fsa flows obtained are plotted .
the values that result from inboard toroidal measurements are presented as open circles , whilst the ones deduced from outboard toroidal velocities are squares . in both cases ,
the poloidal measurements inserted in equation come from the outboard region .
flux - surface averaged flows @xmath109 and @xmath110 , for the : ( a ) ecrh discharge , @xmath60 ; and ( b ) nbi plasma , @xmath102 .
the circles correspond to toroidal measurements performed in the inboard region , and the squares to those taken in the outboard one.,width=7 ] the agreement found between inboard / outboard measurements confirms the main hypothesis of the applied method , _
i.e. _ , that impurity flows are incompressible for the low density plasmas presented here .
indeed , to the authors knowledge , this is the first time that this fundamental assumption of kinetic theory is checked with independent measurements at different locations of a surface ( recent work at tcv and diii - d @xcite extract the poloidal flow from toroidal in / out measurements , but are not compared with direct poloidal measurements , due to difficulties associated with such measurement in these tokamaks ) .
the strong asymmetry in the toroidal velocity observed in the ecrh scenario , figure [ fig : raw_velocity ] ( a ) , is explained by the larger -compared to the nbi case- perpendicular flow , which generates significant variations in the parallel velocity , through the local pfirsch - schlter flow . as expected
, there is a change from electron to ion root in the radial electric field with increasing density @xcite , and therefore , a change of sign in @xmath111 ( the impurity diamagnetic contribution is negligible , except for @xmath112 in the nbi case , where it reaches @xmath23 km s@xmath50 ) .
finally , a positive parallel velocity is observed in the nbi plasmas , which is compatible with the injection of the beam in the direction of the magnetic field .
this will be discussed in the next subsection . as an indication of the importance of this previous check on the flow spatial variation
prior to its comparison with neoclassical estimates , we note that in higher density nbi plasmas a _
compressible _ flow variation is observed in tj - ii .
this makes the comparison of parallel velocities with standard neoclassical theory not straightforward .
indeed , impurity density variations within a flux surface have been theoretically predicted in the presence of large ion pressure gradients @xcite , and recently observed in several tokamak devices ( alcator c - mod @xcite and asdex - upgrade @xcite ) .
the asymmetry is related to ion - impurity friction and ultimately to impurity radial fluxes , which makes its comprehension particularly relevant for nuclear fusion .
the analysis of such plasmas will be presented in a subsequent article .
the drift kinetic equation is solved by using the numerical code dkes @xcite complemented with momentum correction techniques .
dkes allows to calculate the monoenergetic transport coefficients of the magnetic configuration .
these coefficients are convoluted with a maxwellian distribution function which in turn depends on local density and temperature .
inclusion of the thermodynamical forces ( gradients of density and temperature and radial electric field ) allows the fsa neoclassical fluxes to be estimated .
the neoclassical radial electric field itself is found iteratively by imposing ambipolarity on the particle radial fluxes .
specific details of the procedure are found in @xcite .
the inputs for this calculation are the magnetic field equilibrium and tj - ii density and temperature profiles .
the former is taken to be the vacuum equilibrium , for the low @xmath113 plasmas studied here .
the latter require integration of data from the thomson scattering , helium beam probe , reflectometry , interferometry ( see @xcite and references there in ) and cxrs diagnostics . the uncertainties in the transport coefficients and in the measured profiles are propagated to the final theoretical estimate , as in @xcite . in order to estimate accurately the parallel and poloidal momentum balance , one should make the calculation by including protons , electrons and main impurities present in the plasma .
however , for the lithium - coated plasmas studied in this work , the impurity concentration is low @xcite and does not modify the poloidal momentum balance , leaving @xmath114 unchanged .
since the impurity diamagnetic term is small , this completely determines the poloidal impurity rotation . on the other hand ,
once one ensures momentum conservation in the calculations ( by including interspecies friction ) , the bootstrap flow of the trace impurities is shown to follow that of the bulk ions , which are calculated as in @xcite .
finally , one may wonder if values of @xmath114 close to resonances could be found , as in @xcite . in such a case ,
the poloidal @xmath2 and magnetic drifts cancel out , producing a peaking of the radial fluxes , and the local ansatz underlying dkes computations fails . at tj - ii
, this happens for very large values of @xmath114 @xcite .
the agreement found in the present comparison supports the hypothesis in our calculation .
some general considerations can be made , following the results obtained for model plasma profiles @xcite .
low density ecrh plasmas in tj - ii are in the electron root : @xmath114 is positive and large , and therefore it partially cancels the large contribution of the @xmath115 gradient to the electron radial transport ( see figure [ fig : plasma_profiles ] ) .
this large radial electric field drives a large ion parallel flow , especially close to the centre , where the fraction of trapped particles is small and @xmath114 is considerable . on the contrary , nbi plasmas are in the ion root . in order to reduce the ion particle radial transport , driven by the ion temperature gradients , to the electron level
, @xmath114 is small and negative .
this partial cancellation of the ion channel has consequences in the parallel momentum balance as well , where the negative @xmath114 leads to a small ion bootstrap current . in the following ,
the results of a comparison between nc theory and cxrs measurements are presented . since cxrs
active data has been obtained from two different discharges ( shot - to - shot technique ) , we compare the experimental data with the average of nc outcomes for both plasmas .
comparison of neoclassical ( grey shading ) and measured ( red circles ) @xmath114 , for the : ( a ) ecrh discharge , @xmath60 ; and ( b ) nbi plasma , @xmath102.,width=7 ] in figure [ fig : neo_er ] , the measured and calculated radial electric field are shown .
the experimental values are in agreement with previous results obtained in similar plasmas with the hibp diagnostic @xcite , and agree with the nc predictions agree within the error bars . nonetheless , the mean values observed at the core of the ecrh plasma are somewhat larger than that of the neoclassical estimate ; however , the error bars are large . moreover , the contribution to the experimental @xmath114 from toroidal measurements , @xmath116 , equation , is similar to that of the poloidal flows , @xmath117 , due to the high toroidal velocities observed in the core region , see figures [ fig : geometrical_cxrs_factors ] and [ fig : raw_velocity ] ( a ) .
this increases the uncertainties in @xmath114 , since toroidal sightlines are almost aligned with the magnetic field lines , and thus , the extraction of its perpendicular contribution is less reliable . in the nbi plasma ,
the experimental errors are reduced , see figure [ fig : neo_er ] ( b ) .
the inferred radial electric field is in good quantitative agreement with neoclassical calculations .
the error bars about the nc estimates in figure [ fig : neo_er ] correspond to dispersion in the ambipolar condition when one considers the not insignificant uncertainties in the plasma profiles , as in @xcite .
in particular , for @xmath118 , both electron and ion roots can result , as different values of density and temperature are considered within the error bars of figure [ fig : plasma_profiles ] , hence the large uncertainties .
finally , the agreement found for this discharge in the radial electric field suggests that fast nbi ions do not contribute significantly to the radial ambipolar balance , although they inject parallel momentum , as discussed in the following text .
comparison of neoclassical and measured parallel flow , @xmath119 , for the : ( a ) ecrh discharge , @xmath60 ; and ( b ) nbi plasma , @xmath102.,width=7 ] figure [ fig : neo_boot ] ( a ) shows calculations of neoclassical bootstrap ion flow during the ecrh phase , together with the c@xmath33 parallel flux - surface flow , @xmath120 . a strong parallel rotation driven by the radial electric field
is observed at the core region , that is well reproduced by the neoclassical predictions . in the higher density unbalanced nbi case , the parallel flow consists of the neoclassical one ( bootstrap ) , plus the rotation induced by the injected neutrals .
indeed , these measurements show that the latter dominates , being able to reverse sign .
since this is a low density nbi plasma , the injected fast neutrals can penetrate into the core .
this would explain why the deviation of the parallel flow is maximum towards @xmath121 , whilst measured parallel flow approaches bootstrap calculations for @xmath118 .
comparison of nc and measured radial electric field and bootstrap flow , for a set of ecrh discharges .
circles correspond to cxrs measurements , and grey shading to nc estimates.,width=7 ] finally , to highlight the reproducibility of the results in the ecrh scenario , the comparison of the measured radial electric field and bootstrap flow with nc calculations have been depicted in figure [ fig : neo_ecrh_complete ] , for a set of similar ecrh discharges .
the features observed in figures [ fig : neo_er ] ( a ) and [ fig : neo_boot ] ( a ) are replicated in these discharges , and the agreement with the nc calculations is repeated as well .
in summary , a comparison between cxrs carbon impurity flow measurements and nc calculations has been done .
in addition , the basic form of the velocity field , _ i.e. _ a divergence - free flow tangent to magnetic surfaces , has been verified . to this end , a general treatment of sightlines and
flow geometry has been applied to two pairs of three independent velocity measurements performed at different locations of a flux surface , and consistency with a 2d incompressible flow has been demonstrated .
in addition , good quantitative agreement has been found between measured and nc radial electric fields for low density ecrh plasmas .
furthermore , a comparison of impurity parallel rotation with nc bootstrap flow has been done for the first time in tj - ii , showing consistency with the calculations in these plasmas . in an unbalanced nbi discharge with relatively low density
, a similarly good agreement has been found for the radial electric field , while a change in the direction of parallel rotation ( with respect to the nc calculations , which do not account for momentum injection ) has been observed .
this points to a minor role of the fast ion - driven radial current in the ambipolar balance ( or poloidal momentum balance ) of the tj - ii stellarator .
the authors are indebted to the tj - ii experimental group .
they would like to thank b. zurro for useful discussions , and c. hidalgo for his support .
j. arvalo acknowledges financial support from the fpi grant awarded by ciemat ( boe resolution n 171 , 24/06/2008 ) . | poloidal and toroidal velocities of fully - ionised carbon are measured by means of charge exchange recombination spectroscopy ( cxrs ) in the tj - ii stellarator .
we present a detailed treatment of the 3d geometry and show that flow measurements , performed at different locations of the same flux surface , are compatible with flow incompressibility for the low density plasmas under study ( line averaged electron densities @xmath0 m@xmath1 ) .
furthermore , comparison with neoclassical calculations shows quantitative agreement with the measured radial electric field and ion bootstrap parallel flow in the absence of an external momentum input . |
let @xmath0 be a connected 4-valent graph and @xmath1 a planar realization(the words realization and configuration will be used interchangeably ) . the graph @xmath2 is always assumed to be finite and simple .
the rigidity matrix , which we denote by @xmath3 , is a matrix of size @xmath4 .
the rank of @xmath3 , which we denote by @xmath5 , can be taken as the definition of the rank of the _ infinitesimal rigidity matroid _ of the framework @xmath6 .
when @xmath7 is a _ generic realization _
, one can show @xmath5 is independent of @xmath7 as long as it is generic , this can be taken as the definition of the rank of the _ generic rigidity matroid _ of @xmath2 and we will write @xmath8 for it . for more details on the definitions and terminologies in rigidity theory , we refer the readers to @xcite
. as we pointed out in @xcite , the question of finding the lower bound of the rigidity matrix is closed related to a question in symplectic geometry .
we will not repeat it here but refer the readers to @xcite and references therein .
the main results in the paper are summarized as the following four theorems .
[ thm : generic ] assume @xmath9 is a connected 4-valent graph , then @xmath10 [ thm : general ] assume @xmath9 is a connected 4-valent graph and @xmath1 is a configuration in general position , i.e. , the image under @xmath7 of any three points in @xmath11 do not lie on the same line , then @xmath12 [ thm : connected generic ] assume @xmath9 is a connected 4-valent graph which is also @xmath13-edge - connected ( see definition [ def : edge - connected ] ) , then @xmath14 [ thm : connected general ] assume @xmath9 is a connected 4-valent graph which is also @xmath13-edge - connected and @xmath1 is a configuration in general position , then @xmath15 [ def : edge - connected ] a graph @xmath0 is called @xmath16-edge - connected for some @xmath17 if @xmath2 remains connected upon removing any @xmath18 edges .
in particular , a connected graph is @xmath19-edge - connected .
theorem [ thm : generic ] was theorem 1 in @xcite , here we will give an alternative proof of it by first proving theorem [ thm : connected general ] and then theorem [ thm : general ] .
theorem [ thm : general ] , theorem [ thm : connected general ] and theorem [ thm : connected generic ] provide answers to question 1.9 and question 1.10 in @xcite in case of 4-valent graphs .
it should also be pointed out that theorem [ thm : connected general ] is essentially proved in section 5 of @xcite , it is the @xmath20 version of theorem 5.1(or proposition 5.7 ) there .
however , @xcite was written in a language ( mathematically ) different from here , so despite the repetition , we will present the full details of the proof in this paper .
[ fig:15vertices ] [ example : generic ] the graph @xmath2 with configuration @xmath7 shown in figure [ fig:15vertices ] is example 1.3 in @xcite . the configuration is slightly modified so that it is in general position .
it shows the order of the bound given in theorem [ thm : generic ] and theorem [ thm : general ] are sharp .
let s briefly recall the construction . take @xmath21 copies of complete graphs on @xmath22 vertices and delete one edge from each , then connect them together to form a loop as shown in the figure .
both @xmath8 and @xmath5 are @xmath23 for this graph .
this example can be easily generalized to a graph @xmath24 of @xmath25 vertices and configuration @xmath26 in general position such that both @xmath27 and @xmath28 are @xmath29 . in this example , @xmath28 does not depend on @xmath26 as long as @xmath26 is in general position , but it should be emphasized that this is not usually the case .
the graph @xmath2 shown in figure [ fig : genericmaxs ] is obtained by connecting @xmath13 copies of complete graphs on @xmath13 vertices .
it is 4-edge - connected and of @xmath30 .
this example can be easily generalized to a graph @xmath24 with @xmath31 vertices , such that @xmath32 if @xmath33 , @xmath34 if @xmath35 .
this example shows the order of bound given in theorem [ thm : connected generic ] is sharp . in order to introduce the example that shows the order of the bound given in theorem [ thm : connected general ] is sharp , we first introduce the definition of cartesian product in graph theory .
the cartesian product @xmath36 of graphs @xmath37 and @xmath38 is a graph such that * the vertex set of @xmath39 is the cartesian product @xmath40 ; and * any two vertices @xmath41 and @xmath42 are adjacent in @xmath39 if and only if either * * @xmath43 and @xmath44 is adjacent with @xmath45 in @xmath46 , or * * @xmath47 and @xmath48 is adjacent with @xmath49 in @xmath50 .
[ lemma : cartesianproduct ] assume @xmath51 and @xmath52 are two connected graphs with planar configuration in general position .
for any non - zero constants @xmath53 , we can define a planar configuration of @xmath54 by @xmath55 if @xmath7 is also a configuration in general position , then we have @xmath56 this follows immediately from proposition 3.18 and proposition 3.4 in @xcite , noting @xmath57 was denoted by @xmath58 there . the configuration @xmath59 given in the above lemma , which we assume to be in general position ,
will be denoted by @xmath60 .
the notation is not perfect in that it does not suggest the dependence of @xmath7 on the choices of @xmath53 , but this should not bother us since the quantity @xmath5 does not rely on @xmath61 as the above lemma suggests .
the following example is communicated to me by prof .
bob connelly .
if we use @xmath62 to denote the the complete graph on @xmath63 vertices and @xmath64 to denote a @xmath65-gon , i.e. , a regular 2-valent graph of @xmath65 vertices . then @xmath66 is a 4-edge - connected 4-valent graph .
figure [ fig : k_3 ] and [ fig : p_6 ] give a planar configurations in general position for @xmath67 and @xmath68 respectively .
figure [ fig : k_3productp_6 ] shows an induced configuration in general position for @xmath69 in the way described in lemma [ lemma : cartesianproduct ] .
denote by @xmath2 the product @xmath69 and by @xmath7 the configuration illustrated in figure [ fig : k_3productp_6 ] , then according to lemma [ lemma : cartesianproduct ] , @xmath70 this example is easily generalized to @xmath71 , which is a 4-edge - connected 4-valent graph , equipped with a planar configuration @xmath26 in general position that is induced from a configuration of @xmath67 and a configuration of @xmath64 , and @xmath72 .
this class of graphs shows the order of the bound given in theorem [ thm : connected general ] is sharp .
it is natural to ask the generalizations of these theorems about regular graphs of degree @xmath73 for @xmath74 .
the @xmath75 version of theorem [ thm : generic ] was proved in @xcite .
we ask the following questions . [
question : generic ] assume @xmath9 is a connected @xmath73-valent graph , @xmath76 , do we have @xmath77 note that if the bound holds , then the order of it is sharp . because we may take @xmath78 copies of @xmath79 , delete one edge from each , then connect them to form a loop as we did in example [ example : generic ] , call the resulting graph @xmath2 . then @xmath80 .
assume @xmath9 is a connected @xmath73-valent graph , @xmath74 , and @xmath1 is a configuration in general position , do we have @xmath81 note that a positive answer to this question would imply the positive answer for question [ question : generic ] as @xmath82 . also , if the bound holds , the order of it is sharp .
assume @xmath9 is a connected @xmath73-valent graph which is also @xmath73-edge - connected and @xmath1 is a configuration in general position , do we have @xmath83 again , we note that if the bound holds , then the order of it is sharp .
we may consider @xmath84 with the planar configuration @xmath7 in general position induced from that of @xmath85 and @xmath64 as in lemma [ lemma : cartesianproduct ] .
then by lemma [ lemma : cartesianproduct ] , we have @xmath86 also note that the equality holds when @xmath87 .
other than edge - connectivity , there is another kind of connectivity in graph theory called _ vertex - connectivity_. a graph @xmath0 is called @xmath16-vertex - connected if for any @xmath88 with @xmath89 , the graph @xmath90 defined by @xmath91 is still connected , where @xmath92 is the subset of @xmath93 consisting of edges incident to some vertex in @xmath94 . note that a connected graph is automatically @xmath19-vertex - connected .
we can ask similar questions about the rank of rigidity matrix with vertex - connectivity assumptions .
assume @xmath0 is a regular @xmath73-valent graph that is also @xmath16-vertex - connected , @xmath95 .
what is the lower bound for @xmath8 ?
assume @xmath7 is a planar configuration in general position , what is the lower bound for @xmath5 ? in the case of 4-valent graphs , the answer to
the above question follows trivially from our results about @xmath8 and @xmath5 when assuming ( or not ) edge - connectivities .
we will simply state the result without proof .
assume @xmath2 is a 4-valent graph and @xmath7 a planar configuration in general position .
if @xmath2 is @xmath96-vertex - connected , then @xmath97 . if @xmath2 is @xmath21-vertex - connected or @xmath13-vertex - connected , then @xmath98 and @xmath99 .
the order of all the above bounds are sharp .
* acknowledgement : * i would like to thank bob connelly and tara holm for some helpful discussions .
the content in this section is essentially the work in section 5 of @xcite .
first we briefly recall some notations we used and results we obtained in @xcite , the readers are referred to section 2 of @xcite for more details and proofs about these ( up to lemma [ lemma : disconnect2 ] here ) .
let @xmath9 be a graph , the vertices are labeled as @xmath100 , the edge incident to @xmath101 and @xmath102 is denoted by @xmath103 .
we do not distinguish between @xmath103 and @xmath104 , but most of the time we will make the first coordinate smaller than the second one .
we will use @xmath105 to denote the degree of @xmath106 .
when the graph @xmath2 needs to be emphasized ( this could be the case if @xmath106 is the vertex of two graphs under discussion ) , we use @xmath107 to denote the degree of @xmath106 as a vertex of @xmath2 .
we will use @xmath108 to be subset of @xmath93 consisting of edges incident to @xmath101 . for any @xmath88
, we will use @xmath92 to denote the set of edges incident to some vertex in @xmath94 .
the configuration @xmath1 will _ always _ be assumed to be in general position .
if @xmath24 is a subgraph of @xmath2 , then by abuse of notation , we will also use @xmath7 to denote the configuration of @xmath24 induced from that of @xmath2 .
assume @xmath0 is a graph and @xmath1 is a planar configuration in general position .
we use @xmath109 to denote the _ number of stress _ of @xmath2 and use @xmath110 to denote the _ number of stress _ of @xmath6 .
the following four lemmas and corollary are stated and proved in @xcite under generic configuration assumption , but the proof could be carried over to configuration in general position without any change and we are not going to bother rewriting the proofs . the following lemma is lemma 2.3 in @xcite and called the _ deleting lemma_. [ lemma : delete ] given @xmath0 and @xmath1 a planar configuration in general position , assume there is a vertex @xmath101 with @xmath111 or @xmath96 . define @xmath24 to be the graph obtained from @xmath2 by deleting @xmath101 and @xmath108 , then @xmath112 and @xmath113 .
[ corollary : delete ] given @xmath0 and @xmath1 a planar configuration in general position , let @xmath114 .
define @xmath24 to be the graph obtained from @xmath2 by deleting @xmath101 and @xmath108 , then @xmath115 and @xmath116 the following two lemmas are corollary 2.8 and corollary 2.9 of @xcite .
[ lemma : disconnect1 ] given @xmath6 a graph with planar configuration in general position , assume deleting one edge @xmath103 increases the number of connected component of @xmath2 by @xmath19 .
denote by @xmath24 the graph obtained from @xmath2 by deleting @xmath103 . then @xmath112 and @xmath113 .
[ lemma : disconnect2 ] given @xmath6 a graph with planar configuration in general position , assume deleting some two edges @xmath117 and @xmath118 increases the number of connected components of @xmath2 by @xmath19 , but deleting any one of these two edges does not change the number of connected components . denote by @xmath24 the graph obtained from @xmath2 by deleting @xmath117 and @xmath118
. then @xmath112 and @xmath113 .
[ def : trim ] a graph @xmath0 is called _ trimmed _ if * each vertex of @xmath2 is of degree at least @xmath21 ; * each connected component of @xmath2 is @xmath21-edge - connected . for any graph @xmath0 , we can obtain a trimmed subgraph @xmath119 of @xmath2 in the following way .
if @xmath2 is trimmed , then let @xmath120 .
if @xmath2 is not trimmed , then we first obtain a subgraph @xmath50 of @xmath2 in one of the following two ways : * if there is vertex @xmath101 of @xmath2 of degree @xmath19 or @xmath96 , let @xmath121 .
the choice of @xmath101 may not be unique ; * if every vertex is of degree at least @xmath21 , then there exists @xmath122 , such that @xmath123 , and removing @xmath124 would increase the number of connected components of @xmath2 , but removing any proper subset of @xmath124 would not increase the number of connected components of @xmath2 , then let @xmath125 . again , the choice of @xmath124 may not be unique .
if @xmath50 is trimmed , then we let @xmath126 . otherwise , we obtain @xmath46 as a subgraph of @xmath50 in the same way as described above . following this process we get a sequence of graphs @xmath127 , where @xmath128 is a subgraph of @xmath129 obtained in one of two ways we just described , and @xmath130 , which could be empty , is trimmed .
we let @xmath119 be @xmath130 .
[ def : trimming ] we call the process of getting @xmath119 from @xmath2 described above the _ trimming process_. it follows from lemma [ lemma : delete ] , lemma [ lemma : disconnect1 ] and lemma [ lemma : disconnect2 ] that @xmath131 and @xmath132 for any planar configuration in general position .
evan we may have different choices for each graph @xmath133 in the middle of the trimming process , the graph @xmath119 in fact does not depend on the choice of @xmath133 s , it is the maximal trimmed subgraph of @xmath2 .
we will not need this , hence will not bother with the proof .
we make the following definition so that the statements and proofs in the rest of this section could be made more concise .
[ def : a_4 ] we call a graph @xmath0 is _ of type @xmath134 _
if * each vertex of @xmath2 is of degree @xmath21 or @xmath13 ; * each connected component of @xmath2 has at least one vertex of degree @xmath21 ; * each connected component of @xmath2 is @xmath21-edge - connected .
the following proposition is a special case of proposition 5.7 of @xcite .
[ prop : s for a_4 ] assume @xmath0 is a graph of type @xmath134 and @xmath1 a planar configuration in general position , then @xmath135 where @xmath136 is the number of vertices of degree @xmath13 , @xmath137 is the number of connected components of @xmath2 .
we are going to use induction on the size of @xmath138 .
the graph of type @xmath139 with lease number of vertices is the empty graph , inequality holds trivially in this case .
now we consider @xmath0 with @xmath140 and assume holds for @xmath141 .
if @xmath142 , then each connected compnent of @xmath2 is also of type @xmath139 and of less vertices .
so the induction hypothesis says holds for each connected component .
we may then add them up to show holds for @xmath2 as well .
now we assume @xmath2 is connected .
since @xmath2 is of type @xmath139 , there exists a vertex @xmath143 of degree @xmath21 .
define @xmath90 by @xmath144 and @xmath145 . by corollary [ corollary :
delete ] , we have @xmath146 .
apply the trimming process to @xmath24 to obtain @xmath147 , then @xmath148 and @xmath147 is of type @xmath139 .
assume @xmath149 and @xmath150 .
since @xmath2 is @xmath21-edge - connected , for each @xmath133 , there must be at least @xmath21 vertices @xmath106 in @xmath151 , such that @xmath152 and @xmath153 .
so @xmath154 since @xmath147 is of type @xmath139 and of less vertices than @xmath2 , by the induction hypothesis we have @xmath155 so @xmath156 & \leq \dfrac{n_{4}(\tilde{g'})}{3}+c(\tilde{g'})+1\\[1.2ex ] & = \dfrac{1}{3}{\displaystyle}{\sum_{i=1}^a n_{4}(g_i ) } + a+1\\ & \leq \dfrac{1}{3}(n_{4}(g)-3a)+a+1\\[1.2ex ] & = \dfrac{n_{4}(g)}{3}+1=\dfrac{n_{4}(g)}{3}+c(g ) .
\end{array}\ ] ] this completes the induction step . pick one edge @xmath157 and form a new graph @xmath158 . by corollary [ corollary :
delete ] we have @xmath146 .
the graph @xmath24 is of type @xmath139 , so by proposition [ prop : s for a_4 ] , we have @xmath159 , hence @xmath160 . therefore @xmath161
if @xmath2 is the empty graph , the theorem holds trivially .
now we may assume @xmath162 .
if @xmath2 is @xmath13-edge - connected , then by theorem [ thm : connected general ] , we have @xmath163 because @xmath162 . if @xmath2 is not @xmath13-edge - connected , then there exists @xmath164 with @xmath165 , such that the graph @xmath166 is disconnected .
we may assume @xmath124 has the least cardinality among the subsets of @xmath93 with this property .
we observe @xmath167 can not be @xmath21 or @xmath19 , because the sum of degrees of the vertices of one connected component of @xmath24 has to be an even number .
so @xmath168 .
now apply the trimming process , which was defined in definition [ def : trimming ] , to @xmath2 , we get @xmath119 .
then @xmath119 is a proper subgraph of @xmath2 and is of type @xmath139 .
we put the connected components of @xmath119 into two categories : the type of four vertices and the type of at least five vertices .
assume @xmath119 decomposes into disjoint union of connected components @xmath169 where each @xmath133 is of four vertices , hence has to be the complete graph on four vertices , each @xmath170 is of at least five vertices and of type @xmath139 .
assume @xmath171 has @xmath172 vertices , then @xmath173 , so by proposition [ prop : s for a_4 ] , we have @xmath174 so @xmath175 and because @xmath176 \leq & -\dfrac{1}{15}(|v|-5b)\leq 0 \end{array}\ ] ] we immediately see that @xmath177 , so @xmath178 this follows trivially from theorem [ thm : general ] because we always have @xmath82 .
the proof uses a similar idea with that in the proof of theorem [ thm : connected general ] .
first we recall two results about generic rigidity from section 2 of @xcite .
[ lemma : disconnect3 ] given a graph @xmath0 , assume @xmath164 with @xmath179 and the edges in @xmath124 do not share a common vertex .
let @xmath166 .
if removing @xmath124 increases the number of connected components of @xmath2 , but removing any proper subset of @xmath124 will not do so , then @xmath112 .
[ lemma : one extension ] given @xmath0 , assume there is a vertex @xmath143 of degree @xmath21 and the three vertices adjacent to @xmath143 are @xmath180 and @xmath181 .
assume @xmath182 .
we define a new graph @xmath90 by @xmath183 and @xmath184
. then @xmath185 .
* it is trimmed ( see definition [ def : trim ] ) ; * if there exists @xmath164 , such that @xmath179 and removing @xmath124 increases the number of connected components of @xmath2 , then all the edges in @xmath124 share a common vertex .
similar to the trimming process , we may define the _ generic trimming process_. for any graph @xmath0 , we can obtain a generically trimmed subgraph @xmath186 of @xmath2 in the following way .
if @xmath2 is generically trimmed , then let @xmath187 .
if @xmath2 is not generically trimmed , then we first obtain a subgraph @xmath50 of @xmath2 in one of the following three ways : * if there is vertex @xmath101 of @xmath2 of degree @xmath19 or @xmath96 , let @xmath121 .
the choice of @xmath101 may not be unique ; * if every vertex is of degree at least @xmath21 , and there exists @xmath122 , such that @xmath123 , and removing @xmath124 would increase the number of connected components of @xmath2 , but removing any proper subset of @xmath124 would not increase the number of connected components of @xmath2 , then let @xmath125 .
the choice of @xmath124 may also not be unique . *
if @xmath2 is trimmed , and there exists @xmath164 with @xmath179 , such that the vertices in @xmath124 do not share a common vertex and removing @xmath124 will increase the number of connected components of @xmath2 , then let @xmath125 .
again , the choice of @xmath124 may not be unique .
if @xmath50 is generically trimmed , then we let @xmath188 . otherwise , we obtain @xmath46 as a subgraph of @xmath50 in the same way as described above . following this process we get a sequence of graphs @xmath127 , where @xmath128 is a subgraph of @xmath129 obtained in one of three ways we just described , and @xmath130 , which could be empty , is generically trimmed .
we let @xmath186 be @xmath130 .
[ def : generic trimming ] we call the process of getting @xmath186 from @xmath2 described above the _ generic trimming process_. it follows from lemma [ lemma : delete ] , lemma [ lemma : disconnect1 ] , lemma [ lemma : disconnect2 ] and lemma [ lemma : disconnect3 ] that @xmath189 .
[ lemma : k_4 and b_4 ] given @xmath0 is a connected graph of type @xmath191 , assume @xmath192 is of degree @xmath21 and the three vertices adjacent to @xmath143 are @xmath193 .
denote by @xmath194 the complete graph on @xmath195 and @xmath196 . if @xmath197 is a subgraph of @xmath2 , then @xmath198 . if @xmath199 , then @xmath200 is not empty .
deleting @xmath124 from @xmath2 would disconnect @xmath2 .
because @xmath2 is generically trimmed , we must have @xmath179 and there exists @xmath201 , such that @xmath202 . because @xmath2 is of type @xmath191 , either @xmath203 or @xmath204 .
if @xmath203 , then @xmath2 only have five vertices : @xmath205 , and only @xmath96 of them are of degree @xmath21 , this contradicts the assumption that @xmath2 is of type @xmath191 . if @xmath204 , assume @xmath206 is the other vertex adjacent to @xmath207
. then deleting @xmath208 will disconnect @xmath2 , contradicting the assumption that @xmath2 is generically trimmed .
so @xmath198 .
we are going to use induction on @xmath138 .
the base case is @xmath211 , holds trivially in this case .
now assume @xmath212 and holds for graphs with less than @xmath213 vertices .
if @xmath142 , then each connected component of @xmath2 is of type @xmath191 and of less vertices , so holds for each connected component by the induction hypothesis , we may then add them to show holds for @xmath2 as well .
now we assume @xmath2 is connected . * if @xmath216 , then by lemma [ lemma : k_4 and b_4 ] we have @xmath198 and we can check holds in this case . * if @xmath217 , then without loss of generality we may assume @xmath218 .
define @xmath38 by @xmath219 and @xmath220 . then by lemma [ lemma : one extension ] , we have @xmath221 . because both @xmath2 and @xmath50 are of type @xmath191 , @xmath46 must be of type @xmath191 as well . since @xmath222 , by the induction hypothesis , we have @xmath223 .
so @xmath224 .
we claim in each @xmath171 , there are at least @xmath13 vertices @xmath106 such that @xmath229 and @xmath152 . if not , because @xmath2 is @xmath21-edge - connected , there exist exactly three vertices @xmath230 such that @xmath231 for @xmath232 .
assume the edges in @xmath2 but not in @xmath225 and incident to some @xmath233 are @xmath234 and @xmath235 . because @xmath2 is of type @xmath191 ,
hence generically trimmed , we must have @xmath236 .
so if @xmath237 , then @xmath207 is a vertex of @xmath225 , but this is not the case , so @xmath238 , hence @xmath239 .
then @xmath240 , this contradicts the assumption that @xmath50 is not of type @xmath191 .
so we have proved the claim .
so each @xmath171 , hence @xmath225 , is of type @xmath191 , by the induction hypothesis , we have @xmath241 so @xmath242 & \leq \dfrac{n_{4}(\tilde{g_1}^g)}{4}+c(\tilde{g}^g)+1\\[0.6ex ] & \leq \dfrac{1}{4}(n_{4}(g)-4a)+a+1\\[1.2ex ] & = \dfrac{n_{4}(g)}{4}+1=\dfrac{n_{4}(g)}{4}+c(g ) . \end{array}\ ] ] this completes the induction steps .
define @xmath37 by @xmath243 and @xmath244 .
then by corollary [ corollary : delete ] , @xmath245 .
apply the generic trimming process to @xmath50 , we obtain @xmath225 , then @xmath226 .
assume @xmath227 and @xmath246 for each @xmath171 , there must exist at least @xmath13 vertices of degree @xmath21 , because @xmath2 is @xmath13-edge - connected .
so each @xmath171 , hence @xmath225 , is of type @xmath191 . by proposition [ prop : s for b_4 ] , we have @xmath247 so @xmath248 & \leq \dfrac{n_4(\tilde{g_1}^g)}{4}+a+2\\ & \leq \dfrac{1}{4}(|v|-1 - 4a)+a+2\\[1.5ex ] & = \dfrac{|v|+7}{4 } \end{array}\ ] ] hence @xmath249 | in this paper we study the rank of planar rigidity matrix of 4-valent graphs , both in case of generic realizations and configurations in general position , under various connectivity assumptions on the graphs . for each case
considered , we prove a lower bound and provide an example which shows the order of the bound we proved is sharp .
this work is closed related to work in @xcite and answers some questions raised there . |
transformation optics @xcite has paved the way for the development of optical devices that can realize functionalities that were thought to be possible only in science fictions @xcite-@xcite .
one such conceptual device that has attracted great public interest is a gateway that can block electromagnetic waves but that allows the passage of other entities .
this device can be viewed as an implementation of a `` hidden portal '' mentioned in fictions @xcite .
however , the feasibility of such devices is limited by the very complex material parameters and the narrow bandwidth . here , we show that gateway - type devices can actually be realized with simple parameters and they can have wider band widths such that the concept is closer to reality @xcite-@xcite than previously thought .
the structure can be implemented by using the magnetic photonic crystal structures that are field tunable , resulting in an invisible electromagnetic gateway that can be open or shut using magnetic fields .
we start from a simple implementation of transformation optics . in fig .
[ fig.1](a ) , an object ( colored in blue ) is placed to the left side of a double negative medium ( dnm ) with @xmath0 .
let the object be of permittivity @xmath1 and permeability @xmath2 .
the detailed shape and the length scale of the structure are shown in fig . [ fig.1](a ) . from the viewpoint of transformation optics ,
the whole structure is optically equivalent to another object described in fig . [
fig.1](b ) for far field observers .
the equivalent permittivity and permeability tensors are @xmath3 and @xmath4 with a constant tensor @xmath5 .
we introduce some parameters for ease of reference following the coordinates @xmath6 , @xmath7 , @xmath8 , @xmath9 and the angle @xmath10 in fig .
[ fig.1 ] .
let @xmath11 be the waist of the object in the @xmath12-direction , @xmath13 be a coordinate compression ratio and @xmath14 where @xmath15 are the position coordinates of the equivalent transformation medium in fig .
[ fig.1](b ) .
the inequalities for the definition of the parameter , @xmath16 , come from three different coordinate transformations and the three regions are distinguished by the two dashed lines in fig . [
fig.1](b ) .
it can be proved that ( see appendix a ) : @xmath17.\ ] ] the coordinate transformation is based on the recent work on transformation optics with complementary medium ( or folded geometries ) @xcite .
the original object is mapped into the equivalent transformation medium .
the dnm together with its mirror image ( the boundary is marked by the dashed line in fig . [ fig.1](a ) ) of air form a pair of complementary media , in the sense the phase accumulated in one segment is exactly cancelled by another .
a more detailed discussion of the optical property of this specific geometry and the associated coordinate transformation can be found in appendix a. the structure in fig .
[ fig.1](a ) can be treated as a form of a scattering amplifier @xcite because its scattering cross section can be much larger than its geometric cross section .
the amplification effect originates from the excitation of surface plasmons in the dnm and air interface . as a concrete example
, we suppose that the original isotropic material ( the blue region in fig . [
fig.1](a ) ) has a large value of permittivity , @xmath18 and @xmath19 , which may be treated approximately as a perfect electric conductor ( pec ) .
figure [ fig.2](a ) shows the scattering field pattern of the structure in fig . [ fig.1](a ) .
the plane wave is incident from the top to the bottom and has transverse electric ( te ) polarization , for which the e field is along @xmath20-direction . in this paper , we only consider the te modes for simplicity . note that the same idea works for transverse magnetic ( tm ) modes as well .
the frequency is @xmath21 and @xmath22 , @xmath23 and @xmath24 .
the structure behaves like an equivalent material with parameters described by equation ( [ eq2 ] ) to the far - field observers . due to the large mismatch of the impedance of the equivalent material with the air background
, we expect that the equivalent object ( the yellow domain in fig . [
fig.1](b ) ) should scatter like a pec , and for that reason we choose for comparison in fig .
[ fig.2](b ) the scattering field pattern of a perfect conductor filling up the entire domain of the virtual object in fig .
[ fig.1](b ) .
the similar far - field scattered patterns between fig . [ fig.2](a ) and fig .
[ fig.2](b ) confirm the strong scattering effect .
the permittivity and permeability of the dnm are actually chosen as @xmath25 in the simulations to avoid numerical divergence problems @xcite .
the pec - like scatters are always chosen as the materials of permittivity at @xmath26 and the permeability is taken to be @xmath27 in this paper for simplicity .
the amplified scattering effect can be utilized to make an invisible gateway @xcite .
suppose that a pec wall separates the whole space into two regions , the upper domain and the lower domain .
if there are channels ( or gateways ) opened in the pec wall , people in the two different spaces can communicate with each other , both physically and through em waves .
however , if we replace the doors with the above - described configuration at a specific frequency , the communication for that frequency will be blocked because the systems behave like pecs .
the most amazing fact is that the channel is in fact physically empty .
there is nothing but air in the channel so objects can `` walk through '' but the channel is blocked as perceived by the eye because light at the designated frequency can not penetrate .
figure [ fig.3 ] is a schematic plot of such a gateway based on the above idea . with the same scale as the one in fig . [
fig.2](a ) , we demonstrate the properties of such a gateway .
we suppose that there is a line source located at @xmath28 with a frequency of @xmath21 in the upper domain .
figure [ fig.4](a ) shows that the waves can not pass through the gateway and are excluded from the lower domain .
however , without the dnm , the waves can propagate into the lower region as shown in fig . [ fig.4](b ) .
we now consider the band - width of such a device . to be more concrete , we consider the following dispersions for the dnm : @xmath29 where @xmath30 , @xmath31 , @xmath32 , @xmath33 and @xmath34 . when @xmath35 , the relative permittivity and permeability return to @xmath25 used in the above simulations .
when @xmath36 , the impedances do not match at the interface of the air and the dnm while the refractive index of the dnm is about @xmath37 .
figure [ fig.4](c ) shows the electric field pattern while fig .
[ fig.4](d ) shows the field without the dnm .
the dnm can reduce the penetration of the waves from the upper region when compared with fig .
[ fig.4](c ) and fig .
[ fig.4](d ) .
however , as @xmath38 , there is still a `` slit '' between the virtual boundary and the right pec , which allows the waves to propagate somewhat into the lower region . if @xmath39 , the virtual boundary is simply the mirror of the interface of the dnm and pec on the left with @xmath40 as the mirror .
the position of the virtual boundary can be obtained heuristically from the image - forming principle . when @xmath41 , the impedances are still mismatched while the refractive index of the dnm is about @xmath42 .
as the absolute value of @xmath43 decreases , the virtual boundary of the image expands and there will be no passage for the waves to penetrate because the virtual boundary overlaps with the pec on the right - band side .
for example , we plot the electric field pattern in fig . [ fig.4](e ) and the case without the dnm in fig . [ fig.4](f ) .
we find that the dnm can eliminate the penetration of the waves from the upper region .
that means that the negative band of the dnm beyond @xmath35 ( or @xmath44 ) is the working frequency of the designed gateway .
however , we also find that the wave blocking effect will become weaker with higher frequencies ( in this case , about @xmath45 ) in our simulations . as negative index media
are intrinsically dispersive , the bandwidth has to be finite as the parameters will deviate progressively from those required by transformation optics .
but the present gateway is shown to be relatively robust and has a broad operation bandwidth of about @xmath46 , which can be regarded as a broadband device . to have a broader bandwidth
, we can simply reduce the distance of the dnm and the right pec to enhance the overlapping while the functionality of the gateway is sacrificed .
as there are extensive designs of the dnms at various wavelengths ( both theoretically and experimentally ) @xcite-@xcite , it would be reasonably feasible for the present gateway to be realized in the future .
if the losses of the metamaterial are large ( i.e. , the imaginary parts of the parameters of the dnm are larger than @xmath47 ) , the bandwidth of the gateway will be small and the functionality will be compromised .
the absorption of the dnm is the key difficulty for both the gateway described here as for other devices such as the perfect lens .
all our simulation results are calculated using the comsol multiphysics finite element - based electromagnetics solver .
very recently , it has been demonstrated that the dnm can be realized with a simple array of ferrite rods without any metallic components @xcite , which can be used to implement the present gateway .
one of the unique merits of this magnetic metamaterial is that its optical properties are magnetically tunable . as such
, the gateway can be manipulated using external magnetic fields .
the magnetic metamaterial is a periodic square array of subwavelength ferrite rods with radii of @xmath48 and a lattice constant of @xmath49 .
the permittivity is taken to be @xmath50 , and the permeability of the ferrite rods has the form @xmath51 with @xmath52 and @xmath53 , where @xmath54 is the damping coefficient , @xmath55 is the resonance frequency with @xmath56 being the gyromagnetic ratio ; @xmath57 is the sum of the external magnetic field and shape anisotropy field along the @xmath20-direction , @xmath58 is the characteristic frequency with @xmath59 the saturation magnetization and it is taken as @xmath60 , typical for single - crystal yttrium - iron - garnet ( yig ) . as the absorption of single - crystal yig is extremely low , we can set @xmath61 .
the numerical simulations on the gateway under different conditions are performed by using the multiple scattering method @xcite .
figure [ fig.5 ] shows the electric field intensity in the logarithmic scale .
the bulk pec is replaced by a discrete system of the periodically arranged pec rods ( marked with dark green solid circles in the figure ) , and the ferrite rods are denoted as hollow circles . under @xmath62 and the frequency @xmath63 , the effective refractive index of the magnetic metamaterial is @xmath64 . in fig .
[ fig.5](a ) , it can be seen that the electric field is excluded from the air channel , so that the open channel appears to be blocked to the eye at this frequency .
this demonstrates that the gateway can be implemented by a photonic crystal type structure in which each element ( each ferrite rod in the periodic array ) is identical . by changing the external field to @xmath65 ,
the refractive index of the magnetic metamaterial becomes @xmath66 ( @xmath67 . the electric field intensity under such conditions is shown in fig .
[ fig.5](b ) , which shows that the channel is open for em wave passage , or the passage appears to be open to the eye .
our results demonstrate that the gateway can be implemented by a very simple configuration and the effect is tunable by an external field .
as shown in appendix b , the effect is fairly robust to the disorder of the position of ferrite cylinders .
figures [ fig.5](c ) and [ fig.5](d ) illustrate the fact that the channel is electromagnetically open if the ferrite rods are replaced by air and by pec rods .
in addition to the simple geometry and weak absorption , the magnetic metamaterial has a reasonably broad bandwidth as a dnm .
in conclusion , we showed that a robust and tunable electromagnetic gateway can be constructed using simple material parameters .
the idea is conceived through consideration of transformation optics and can be realized using a photonic crystal type structure .
we thank dr .
yun lai and dr .
jack ng for their helpful discussions .
this work was supported by hong kong central allocation grant no .
computation resources were supported by the shun hing education and charity fund .
s.y.l and z.f.l were supported by cnkbrsf , nnsfc , pcsirt , moe of china ( b06011 ) , and shanghai science and technology commission .
in this appendix , we give the detailed coordinate transformation to produce the folded geometry used in the text .
let us first consider the following coordinate transformation , @xmath68 where @xmath69
this coordinate transformation maps the virtual space ( @xmath12-space ) described in fig .
[ fig.1](b ) into the physical space ( @xmath70-space ) described in fig .
[ fig.1](a ) .
if the virtual space is a vacuum , we can obtain the corresponding material parameters in the physical space .
it contains two kinds of materials in the vacuum background , one a double negative medium ( dnm ) with @xmath71 ( shown in red in fig . [ fig.1](a ) ) , the other an anisotropic medium ( shown in blue in fig . [ fig.1](a ) ) whose parameters are derived in the following section .
we can easily obtain the jacobian transformation matrix , @xmath72\ ] ] from equations ( [ eqa1 ] ) and ( [ eqa2 ] ) for the mapping between the blue region in fig . [
fig.1](a ) and the yellow region in fig .
[ fig.1](b ) , and we can also prove that @xmath73 and @xmath74 in the text . from transformation optics ,
we can obtain the material parameters for the blue region , @xmath75.\ ] ] figure [ fig.a1](a ) shows the propagation of the incident transverse electric ( te ) polarization plane wave from the left to the right but at an angle of 45 degrees from the @xmath12-direction in free space .
the frequency is @xmath21 .
we consider the object described in equation ( [ eqa4 ] ) together with an dnm domian
. the scales of the structure in our simulations are the same as the ones used in the text .
figure [ fig.a1](b ) shows the scattering pattern of such a structure and demonstrates that the structure itself is invisible to the far - field observers .
a modification of the above folded geometry can lead to a variety of conceptual devices .
for example , if the virtual space is not a vacuum but an object of permittivity @xmath76 and permeability @xmath77 , then the parameters in the blue region should be a permittivity tensor , @xmath78 , and a permeability tensor , @xmath79 .
the object in the blue region together with the dnm domain will have a scattering cross section that can be much larger than its geometric cross section .
figure [ fig.a2](a ) shows the scattering pattern of an object with permittivity @xmath80 and permeability @xmath81 .
figure [ fig.a2](b ) shows the scattering pattern of the structure including the object with permittivity tensor @xmath78 and permeability tensor @xmath79 and the dnm .
the two similar far - field patterns in figs . [ fig.a2](a ) and [ fig.a2](b ) confirm the enhanced scattering effect . in the text
, we apply this structure in an inverse manner .
we suppose that the object in the blue region is of permittivity @xmath1 and permeability @xmath2 .
then , the whole structure is equivalent to another object with its permittivity and permeability tensors , @xmath3 and @xmath4 ( described in the text ) .
we can easily obtain equation ( [ eq2 ] ) by performing an inverse transformation .
in this appendix , we demonstrate the robustness for the implementation of the gateway using the magnetic photonic crystal structure . we introduce disorder by displacing the ferrite rods from their original positions with @xmath82 where @xmath83 is the lattice constant
, @xmath84 is the radii of the ferrite rods , @xmath85 is a uniform random variable on the interval @xmath86 ( assures no overlap of cylinders ) , @xmath87 is the strength of disorder .
figure [ fig.b1](a ) corresponds to fig . [ fig.5](a ) but with a disorder of @xmath88 introduced .
the gateway remains blocked to the eye for this strength of disorder .
when @xmath57 changes to @xmath89 , the gateway is open again as shown in fig .
[ fig.b1](b ) .
figure [ fig.b1](c ) corresponds to fig . [ fig.5](a ) but with a strong disorder of @xmath90 introduced .
the functionality of the gateway is now slightly compromised but the blocking is at least partially effective at such a strength of disorder . with @xmath91 , the gateway is open again as shown in fig . [ fig.b1](d ) .
the results indicate that the functionality is robust to structural perturbations .
10 leonhardt u 2006 _ science _ * 312 * 1777 pendry j b , schurig d and smith d r 2006 _ science _ * 312 * 1780 greenleaf a , lassas m and uhlmann g 2003 _ math .
* 10 * 685 schurig d , mock j j , justice b j , cummer s a , pendry j b , starr a f and smith d r 2006 _ science _ * 314 * 977 cai w , chettiar u k , kildishev a v and shalaev v m 2007 _ nat .
photonics _ * 1 * 224 rahm m , schurig d , roberts d a , cummer s a , smith d r and pendry j b 2008 _ photon .
applic . _ * 6 * 87 chen h y and chan c t 2007 _ appl .
lett . _ * 90 * 241105 greenleaf a , kurylev y , lassas m and uhlmann g 2007 _ phys .
lett . _ * 99 * 183901 kildishev a v and narimanov e e 2007 _ opt .
lett . _ * 32 * 3432 li j and pendry j b 2008 _ phys
* 101 * 203901 liu r , ji c , mock j j , chin j y , cui t j and smith d r 2009 _ science _ * 323 * 366 leonhardt u and tyc t 2009 _ science _ * 323 * 110 ng j , chen h y and chan c t 2009 _ opt .
lett . _ * 34 * 644 lai y , chen h y , zhang z q and chan c t 2009 _ phys .
lett . _ * 102 * 093901 rowling j r 1998 _ harry potter and the chamber of secrets _
( london : bloomsbury ) luo x d , yang t , gu y , chen h y and ma h r 2008 conceal an entrance by means of superscatterer _ preprint _
physics.optics/0809.1823 shelby r a , smith d r and schultz s 2001 _ science _ * 292 * 77 smith d r , pendry j b and wiltshire m c k 2004 _ science _ * 305 * 788 soukoulis c m , linden s and wegener m 2007 _ science _ * 315 * 47 lezec h j , dionne j a and atwater h a 2007 _ science _ * 316 * 430 yao j , liu z , liu y , wang y , sun c , bartal g , stacy a and zhang x 2008 _ science _ * 321 * 930 valentine j , zhang s , zentgraf t , ulin - avila e , genov d a , bartal g and zhang x 2008 _ nature _ * 455 * 376 leonhardt u and philbin t g 2006 _ new j. phys . _ * 8 * 247 pendry j b and ramakrishna s a 2003 _ j. phys .
condens . matter _ * 15 * 6345 yang t , chen h y , luo x d and ma h r 2008 _ opt .
express _ * 16 * 18545 liu s , chen w , du j , lin z f , chui s t and chan c t 2008 _ phys .
lett . _ * 101 * 157407 | transformation optics is used to design a gateway that can block electromagnetic waves but allows the passage of other entities .
our conceptual device has the advantage that it can be realized with simple materials and structural parameters and can have a reasonably wide bandwidth . in particular , we show that our system can be implemented by using a magnetic photonic crystal structure that employs a square ray of ferrite rods , and as the field response of ferrites can be tuned by external magnetic fields , we end up with an electromagnetic gateway that can be open or shut using external fields .
the functionality is also robust against the positional disorder of the rods that made up the photonic crystal . |
large - scale spectroscopic galaxies surveys , such as the sloan digital sky survey ( sdss ; @xcite ) and galaxy and mass assembly ( gama ; @xcite ) survey , are well established and powerful tools for investigations of galaxy evolution
. however , they are limited by only observing a single aperture in the centre of each galaxy , so are inherently unable to investigate the spatially distributed properties of galaxies .
integral field spectrographs ( ifss ) provide the information missed by traditional fibre - based surveys .
the spatially resolved spectroscopic data produced by an ifs can be used to measure diverse quantities such as rotation curves , spatial distributions of star formation , and radial variation in stellar populations . until recently , technical limitations made ifs observations difficult and time - consuming , limiting sample sizes to a few hundred at most , e.g. 260 in atlas-3d @xcite and @xmath0600 in califa @xcite
. a massive step forward is now being taken by instruments that can make ifs observations of multiple objects at once .
the sydney
aao multi - object integral field spectrograph ( sami ) is one of the first of such instruments , with 13 integral field units ( ifus ) , each with a field of view of 15 , that can be deployed across a 1-degree patrol field @xcite .
each ifu consists of a bundle of 61 optical fibres lightly fused to have a high ( @xmath075% ) filling factor @xcite .
sami is installed on the 3.9-m anglo - australian telescope ( aat ) , feeding the existing aaomega spectrograph @xcite .
for the first time , sami allows the rapid generation of large samples of ifs observations .
the sami galaxy survey is an ongoing project using sami to obtain integral field spectroscopic data for @xmath03400 galaxies , with an expected completion date of mid-2016 . at the time of writing , @xmath01000 galaxies
have been observed , including the pilot survey of @xmath0100 galaxies .
early science investigations have included studies of galactic winds @xcite , the kinematic morphology
density relation @xcite , and star formation in dwarf galaxies ( richards et al . , mnras accepted ) ; see also the related contributions by j. bland - hawthorn , l.m.r .
fogarty , i .-
t . ho , and n. scott in this volume .
the target galaxies for the sami galaxy survey are split into two samples : a field and group sample from the galaxy and mass assembly ( gama ) survey g09 , g12 and g15 fields , and a cluster sample from a set of eight galaxy clusters .
the inclusion of a dedicated cluster sample extends the range of environmental densities to higher values than a simple mass- or luminosity - selected sample of galaxies would provide .
full details of the target selection are provided in @xcite .
the target selection for the gama regions , consisting of a tiered set of volume - limited samples , is illustrated in fig .
[ fig : target_selection ] .
the figure also marks the galaxies included in the early data release ( edr ) , discussed in section [ sec : edr ] . the initial steps in the sami data reduction , up to the production of row - stacked spectra ( rss ) ,
are carried out using version 5.62 of 2dfdr . the dedicated sami pipeline
@xcite then performs flux calibration using separate observations of spectrophotometric standard stars , and corrects for telluric absorption based on simultaneous observations of a secondary standard star in the galaxy field .
the set of @xmath07 dithered frames in each field are then resampled onto a regular grid and combined to produce a pair of datacubes ( representing the blue and red arms of aaomega ) for each galaxy .
[ fig : example_galaxy ] illustrates the finished product for a single galaxy , showing the nature and quality of the data obtained .
the complete data reduction process is described in @xcite . and @xmath1 .
upper panel : flux for a central spaxel ( blue ) and one 375 to the north ( red ) .
lower panels , from left to right : sdss @xmath2 image ; continuum flux map ( @xmath3ergs@xmath4@xmath5@xmath4 ) ; stellar velocity field ( kms@xmath4 ) ; h@xmath6 flux map ( @xmath3ergs@xmath4@xmath5 ) ; h@xmath6 velocity field ( kms@xmath4 ) .
the two velocity fields are each scaled individually . for the stellar velocity map , only spaxels with per - pixel signal - to - noise ratio @xmath75 in the continuum are included .
each panel is 18 square , with north up and east to the left .
the grey circle in the second panel shows the fwhm of the psf . ] a comprehensive assessment of the quality of the data produced by sami is presented in @xcite ; here we provide some key results . through the use of dome flat fields we achieve a flat field accuracy of 0.51% in most of the blue arm and 0.30.4% in the red arm .
we use a combination of cuar arc lamps and sky lines to achieve a wavelength calibration accuracy of @xmath80.1 pixels or better .
the continuum sky emission is typically subtracted with an accuracy of @xmath01% . comparing synthetic @xmath9 colours to established photometry
, we find a mean offset of 0.043 mag ( sami data redder ) with standard deviation 0.040 mag .
the observed @xmath10-band magnitudes have a mean offset of @xmath11 mag relative to sdss photometry ( sami data brighter ) , with a scatter of 0.27 mag , although a significant contribution to this scatter comes from difficulties in defining magnitudes for extended , irregular sources .
the median full width at half maximum ( fwhm ) of the final point spread function in the datacubes is 21 , with a typical range from 15 to 30 .
the sami galaxy survey is backed up by a wide range of ancillary datasets . by targetting the gama fields , we have available deep spectroscopic data as well as imaging from current and future surveys ranging from the uv to radio , including galex mis , vst kids , vista viking , wise , herschel - atlas , gmrt and askap @xcite .
the spectroscopic information is crucial for robustly defining the environments of the sami galaxy survey targets @xcite , while the multi - wavelength imaging data provides a wealth of information on processes ( e.g. dust - obscured star formation , radio agn ) that may not be apparent in the optical data .
the cluster sample also has a variety of ancillary data including archival x - ray observations from _ xmm - newton _ and _ chandra _ , and we have carried out a dedicated redshift survey to characterise the clusters and give robust classifications of cluster membership ( owers et al .
, in prep . )
we have recently made available a set of fully calibrated datacubes for 107 galaxies , forming the sami galaxy survey early data release ( edr ; @xcite ) .
the galaxies were selected from the gama regions of the survey . also available
are the datacubes of the corresponding secondary standard stars and a table of ancillary data including quantities such as stellar mass , effective radius and surface brightness .
all data can be downloaded from the sami galaxy survey edr website .
the sami galaxy survey now includes @xmath01000 galaxies with ifs observations , and will grow to @xmath03400 galaxies by mid-2016 .
this unique dataset is providing new insight into a broad range of topics in galaxy evolution , from star formation in dwarf galaxies to stellar kinematics in galaxy clusters . as the sample grows
, further results will come into reach based on robust statistical analyses of the galaxy population , something that has rarely been possible with ifs data in the past . as a starting point
, we have publicly released the datacubes for 107 galaxies , allowing the research community access for scientific investigations and preparation for the full dataset .
jta acknowledges the award of an arc super science fellowship and a sief john stocker fellowship . | the sydney aao multi - object integral field spectrograph ( sami ) galaxy survey is an ongoing project to obtain integral field spectroscopic observations of @xmath03400 galaxies by mid-2016 . including the pilot survey , a total of @xmath01000 galaxies have been observed to date , making the sami galaxy survey the largest of its kind in existence .
this unique dataset allows a wide range of investigations into different aspects of galaxy evolution
. the first public data from the sami galaxy survey , consisting of 107 galaxies drawn from the full sample , has now been released . by giving early access to sami data for the entire research community
, we aim to stimulate research across a broad range of topics in galaxy evolution .
as the sample continues to grow , the survey will open up a new and unique parameter space for galaxy evolution studies . |
recent experiments dedicated to the study of the anisotropy of the cosmic microwave background have found evidence for galactic microwave emission in the range 1090 ghz correlated at high galactic latitudes with thermal emission ( dirbe 100 @xmath0 m map ) from interstellar dust ( kogut et al . 1996 ; leitch et al . 1997 ; de oliveira - costa et al . 1999 , 2002 , 2004 ) .
an explanation for this anomalous dust - correlated microwave emission based on electric dipole emission from fast rotating carbon - based molecules has been proposed by draine & lazarian 1998a , b ) .
these models appear to reproduce the major features of the so - called anomalous microwave emission ( finkbeiner et al .
2004 ) but do not identify the actual carrier of the emission . iglesias - groth ( 2004 ) has recently shown that photoabsorption by fullerenes and buckyonions ( multishell fullerenes ) in the interstellar medium ( ism ) can account for the uv bump ( 217.5 nm ) of the interstellar extinction curve and possibly for some of the diffuse interstellar bands .
it is suggested that fullerenes and buckyonions are broadly distributed in the galaxy and contain a large fraction of interstellar carbon .
the intrinsic dipole moment of these highly symmetric molecules is rather small , thus it is unlikely that they contribute significantly to the anomalous microwave emission .
however , it is expected that a fraction of fullerenes in the interstellar medium will be hydrogenated ; the so - called fulleranes would have a much higher dipole moment because of the polar nature of the c - h bond .
experimental results on the electric dipole moment of isolated alkali - c@xmath1 molecules show an increase from 12.4 d ( debye ) for c@xmath1li to 21.5 d for c@xmath1cs and indicate a strong electron transfer from the alkali atom to the c@xmath1 cage that is almost complete for the largest alkalis ( antoine et al .
given this behavior it is reasonable to expect that the dipole moment of fulleranes is higher than tha of the c - h bond ( 0.3 d ) and therefore , these molecules are potential carriers of electric dipole emission .
fulleranes have deserved attention as potential carriers of diffuse intestellar bands and other interstellar and circumstellar features ( webster 1991 , 1992,1993a , b ) .
both , fullerenes and fulleranes have also been detected in samples of the allende meteorite ( becker & bunch 1997 ) which suggests their existence in the interstellar medium . in this paper
, we study the electric dipole emission of fulleranes with various level of hydrogenation in representative physical conditions in the interstellar medium and show their potential as carriers of the anomalous microwave emission in the galaxy . the predicted emissivities are compared with very recent measurements of anomalous microwave emission in the perseus molecular complex ( watson et al .
we will consider here the various families of fulleranes c@xmath2 h@xmath3 that can be formed from icosahedral fullerenes with @xmath4 = 60 , 180 , 240 , 540 , 960 , 1500 , and 2160 carbon atoms .
fullerenes can also adopt multilayered configurations in which one is encapsulated inside another with separation of @xmath5 3.43.5 .
these multishell fullerenes , also called buckyonions , can also be hydrogenated .
we will consider here only those buckyonions with a complete number of shells and with hydrogen atoms bonded only in the outermost shell . at most , one hydrogen atom may be bound to each of the carbon atoms of a fullerene ; thus , the number , @xmath6 , of hydrogen atoms of any fullerane or buckyonion will lie in the range @xmath7 to @xmath4 .
the ratio between the number of hydrogen and carbon atoms will be , @xmath8 .
we will deal with the following types of chemical species : c@xmath9h@xmath10 , c@xmath11h@xmath10 , c@xmath12h@xmath10 , c@xmath9h@xmath13 , and the corresponding hydrogenated buckyonion analogs ( c@xmath14c@xmath15c@xmath16h@xmath17 , etc . ) .
fullerene radii , @xmath18 , were adopted as in iglesias - groth et al .
they range between 3.55 and 21.5 for c@xmath1 and c@xmath19 , respectively .
the hydrogen atoms of the fulleranes are expected to be located radially outward from the carbon atoms .
the length of the c h bond will be taken as in benzene 1.07 ( see , for example , braga et al .
we assume the moment of inertia , @xmath20 , of a fullerene as that of a spherical cage with radius @xmath21r : @xmath22 , where @xmath23 is the mass of the carbon atom .
for the ensemble of fulleranes with @xmath6 hydrogen atoms we adopt as moment of inertia the sum of the moment of inertia of the relevant fullerenes , i@xmath24 , with the moment of inertia of an hypothetical spherical cage of mass @xmath6 times the mass of the hydrogen atom , and radius @xmath25 .
the moment of inertia of the hydrogenated buckyonion will be taken as the sum of the moments of inertia of all the individual fullerenes conforming the molecule plus that of the layer of hydrogen atoms .
we will assume in what follows that the dipole moment of a fullerane ( or hydrogenated buckyonion ) is proportional to the dipole moment of the c h bond ( @xmath260.3 d ) and that the hydrogen atoms are randomly distributed on the surface of the cage in such a way that the intrinsic dipole moment of the fullerane with @xmath6 hydrogen atoms can be approximated by @xmath27 .
we will adopt a value @xmath28 in the calculations .
owing to collisions with ions and electrons and also because of photoelectric emission , fullerenes in interstellar space may be charged .
an additional electric dipole component could be expected as a consequence of any displacement between the centroid of the molecule charge and the center of mass ( purcell 1979 ) .
given the high symmetry of our molecules we will assume these displacements produce negligible additional dipole moments and will thus adopt the same value for the dipole moments of c@xmath29h@xmath17 , c@xmath30h@xmath17 and c@xmath2h@xmath17 .
fulleranes of the form c@xmath2 h@xmath31 , where a proton has been adsorbed instead of a neutral hydrogen atom , will significantly increase their intrinsic dipole moment .
given the absence of experimental information , we have estimated what would be the permanent dipole moment for this type of molecule and also for protonated buckyonions , using the formalism in rittner ( 1951 ) .
this takes into account the relatively high polarization contribution of the fullerene cage .
values for the polarizabilities of the fullerenes and buckyonions have been adopted from the semiempirical model of iglesias - groth et al .
( 2002 , 2003 ) .
in particular , this formalism gives 5.9 d for the dipole moment of c@xmath1h@xmath13 .
since this formalism may overestimate the dipole moment of these molecules we will conservatively assume in these cases that the value of the dipole moment is that of the c
h@xmath32 bond ( 1.68 d , follmeg et al .
to study the rotational damping and excitation mechanisms affecting fulleranes and hydrogenated buckyonions in the ism we will use the formalism developed by draine & lazarian ( 1998a , b ) .
we will consider two representative environments in the interstellar medium : a ) the so - called cold neutral medium ( cnm ) characterized by low hydrogen density of @xmath33= 30 @xmath34 , gas temperature of @xmath35 k , and low ionization fraction @xmath36(h@xmath32)@xmath37n@xmath38 ; and b ) the warm ionized medium ( wim ) with very low density @xmath33 = 0.1 @xmath34 , gas temperature of @xmath39 k , and high ionization fraction @xmath36(h@xmath32)/@xmath33 = 0.99 .
in thermal equilibrium , fulleranes share the same temperature as the gas and estimation of the rms rotation frequency is straightforward .
when the temperature of fulleranes and gas are different , some processes will produce rotational damping and others will lead to rotational excitation .
draine and lazarian discuss both types of processes comprehensively .
following their treatment , we compute the rotational damping produced by _
collisional drag _ , associated with h , h@xmath40 , and he species ( neutral and ions ) temporarily stuck and subsequently desorbed from the surface of the fullerenes ; _ plasma drag _ caused by the the interaction of the electric dipole moment , @xmath0 , of the fullerene and the electric dipole of the passing ions ; and the _ infrared emission _ , due to the thermal emission of photons that previously heated the molecule .
we also compute rotation excitation rates associated with the impact of neutral particles and ions , plasma drag , and infrared emission .
we have assumed the temperature of the fulleranes is 20 k in the cnm and wim phases of the ism . in figure 1
we show results for the effective rotation rates , @xmath41 of fulleranes with different degree of hydrogenation .
we illustrate the case for the cnm phase of the ism .
the rotation rate decreases as the radius of the molecule increases .
buckyonions show slower rotation rates due to the higher moment of inertia .
increasing the level of hydrogenation for molecules of a given radius ( increasing the dipole moment ) leads to lower rotation rates .
we display results for three levels of hydrogenation @xmath42 , 1/10 , and 1/3 , where @xmath4 is the number of carbon atoms in the fullerane or in the most external shell of the buckyonion .
assuming @xmath28 , in the case of c@xmath1 these hydrogenation ratios correspond to dipole moments of 0.3 , 0.74 , and 1.34 d , respectively ; and 0.3 , 4.4 , and 8.1 d for c@xmath19 , the largest fullerane in this study .
hydrogenation ratios as high as @xmath43 are not rare among other carbon based molecules , for example pahs , and as discussed by webster ( 1992 ) may be rather common in fulleranes . for comparison , we also display the rotation rates of singly protonated fullerenes in wim conditions .
we find lower rotation rates than in the cnm case .
emissivity curves for individual fulleranes and buckyonions of various sizes are plotted in figure 2a for cnm conditions . for these computations , we assume that the densities of c@xmath1h@xmath3 and c@xmath44h@xmath3 are 0.2 @xmath45 10@xmath46 @xmath33 , as obtained for similar size fullerenes in iglesias - groth ( 2004 ) . for the larger fulleranes abundances are scaled down according to the size distribution law @xmath47 , with @xmath48 = 3.5 @xmath49 1.0 proposed in the previous paper .
we are assuming @xmath28 and illustrate results for the case @xmath43 .
we find that for a given level of hydrogenation the major emitters are the smallest fulleranes , and that the frequency of peak emission decreases as the radius increases .
we also plot the total emissivity curve for a mixture of monoshell fulleranes following the proposed size distribution law .
the computations have been extended up to fulleranes with n=2160 , but clearly the largest fulleranes have negligible contribution to the total emissivity curve .
remarkably , this curve shows significant emission ( above 10% of the peak ) in the range 10 to 65 ghz with a peak near 25 ghz . for comparison
, we also plot emissivity curves for a mixture of hydrogenated ( same @xmath50 value ) icosahedral buckyonions with a complete number of shells , starting with c@xmath14c@xmath51h@xmath3 .
we assume the same abundance for this molecule as for the single c@xmath51h@xmath3 .
the emissivity of these buckyonions is similar to that of the indilvidual fullerane of same radius .
we also plot the emissivity curve that would correspond to a mixture of hydrogenated buckyonions following the previous size distribution . in this case , most of the emission lies in the frequency range 1030 ghz with a peak at @xmath518 ghz . for the same level of hydrogenation ,
the buckyonion mixture has a much lower emissivity than the mixture of single - shell fulleranes . in panel
b of figure 2 we show how the predictions are sensitive to the level of hydrogenation and to the physical conditions of the medium .
we present total emissivity curves for a mixture of fulleranes and hydrogenated buckyonions ( following the size distribution of index , @xmath52 ) with three different levels of hydrogenation ( @xmath42 , 1/10 , and 1/3 ) for both cnm and wim conditions . increasing the level of hydrogenation ( i.e. , increasing the average dipole moment of the molecules ) leads to a narrower frequency range for the emission .
the abundances of the smaller fullerenes ( c@xmath1 and c@xmath44 ) have been assumed the same @xmath53 @xmath33 .
wim emissivity curves are found narrower and shifted to lower frequencies . under wim conditions it would be plausible to expect some degree of protonization of the fulleranes . as an indication of how significant this could be regarding electric dipole emission , we show in the figure the emissivity curve obtained for the case @xmath42 , but considering a proton instead of a hydrogen atom . in this case , the much larger dipole moment of the molecules makes the peak emission to shift down almost a factor 2 . in summary ,
under cnm conditions we expect the hydrogenated fullerene / buckyonion mixture produces electric dipole emission with a peak in the range 2050 ghz which shifts to lower frequencies as increases the hydrogenation level . in wim conditions
, this emission tends to be narrower and shifted to lower frequency .
very recently , using data from the cosmosomas experiment ( 1117 ghz ) in combination with wmap data , watson et al .
( 2005 ) show that on scales of a few degrees , the anomalous microwave emission detected towards the perseus molecular complex dominates any other known galactic emission process ( synchotron , free free , or dust emission ) in the frequency range 1050 ghz .
the anomalous emission peaks at 22 ghz ( with an integrated flux of @xmath540 jy ) very close to the peak emission that we have found for a mixture of fullerenes under cnm conditions .
watson et al .
show how a mixture of cnm and wim of draine and lazarian models of electric dipole emission by spinning dust particles may explain the results .
the nature of the particles in the draine and lazarian models are very general , so it is interesting to study whether the anomalous microwave emission can be explained by fulleranes with a minimum number of hypotheses . in figure 3
we plot the empirical results obtained by watson et al .
( 2005 ) and emissivity curves for mixtures of fulleranes and buckyonions computed as described above .
the measurements are corrected following the prescriptions in watson et al . for the contribution of vibrational dust ( @xmath54 k and emissivity index 1.55 ) and for the contribution of free - free in the perseus region .
we consider all the measurements available in the frequency range where anomalous emission appears to make a major contribution , i.e. from 10 to 90 ghz and overplot emissivity curves computed for several minimal cases either using cnm conditions or a combination of cnm and wim .
we plot a few examples that show how the predicted emissivities are indeed very close to those observed .
fitting the observations require some fine tuning between the number of molecules under cnm and wim conditions and the level of hydrogenation . for simplicity
, we have assumed that in wim conditions the fullerenes are singly protonated .
first , it is apparent that a significant level of hydrogenation is required if fulleranes are to explain the observed emission . increasing the level of hydrogenation leads to a higher dipole moment and
shifts the bulk of emission to lower frequencies .
however , the tail emission at high frequencies ( @xmath55 40 ghz ) also requires a contribution of fulleranes with a small dipole moment .
molecules under wim conditions alone would not be able to explain this tail .
the dashed - dotted curve in figure 3 corresponds to the emissivity of a mixture of fulleranes and buckyonions in cnm , where any possible degree of hydrogenation has equal weight .
the curve has been scaled to match approximately the measurement at 22 ghz , but the shape is inconsistent with the observations .
the dotted line present the results for a mixture assuming a range of hydrogenation peaking at @xmath56 , such that 40% of the molecules of a given radius have @xmath56 , 30% have a factor 2 higher dipole moment and the remaining 30% a factor 2 lower .
the emissivity curves obtained for other hydrogenation ranges do not give significantly better fits .
the experimental data are best reproduced ( solid line ) for a combination ( with equal weight ) of the previous curve and the emissivity curve of singly protonated fullerenes in wim conditions , again scaled to match the measurement at 22 ghz .
this factor is close to 3 and results from a possible underestimation of the dipole moments and/or number density of fulleranes in the region .
the dashed line illustrates the same combination between cnm and wim but modifying the hydrogenation range of the molecules in cnm , which are now assumed to be uniformly distributed among @xmath56 , 1/20 , and 1/30 . in summary , it is possible to understand the perseus microwave anomalous emission in terms of electric dipole emission of fulleranes if they follow a size distribution similar to that proposed in the study of the uv extinction bump .
the dominant microwave emission would be associated to the smaller fullerenes .
lazarian and draine ( 2000 ) have suggested that fast rotating ultrasmall ( @xmath57 10 ) grains may be aligned with the interstellar magnetic field via paramagnetic resonance relaxation and produce polarized electic dipole emission .
recent reports that carbon based material like graphite ( esquinazzi et al .
2002 ) polymerized fullerenes ( makarova et al . 2001 ) and c@xmath1h@xmath58 hydrofullerites ( antonov et al .
2002 ) exhibit a weak ferromagnetic moment suggests that the magnetic properties of fulleranes could also lead to polarized rotational electric dipole emission .
precise measurements of the intensity and polarization of the anomalous microwave emission and its relation with uv extinction in the perseus molecular complex and other similar regions will be essential to establish the role of these molecules .
we have computed effective rotation rates of hydrogenated fullerenes in the ism and find values in the 165 ghz range with a clear trend toward lower rotation as the radius of the molecule increases . increasing the level of hydrogenation shifts the peak of electric dipole emission to lower frequencies .
we predict emissivity curves for fullerane mixtures assuming the abundance and size distribution @xmath59 proposed for fullerenes in the ism .
we find that the dominant dipole electric emission would be due to fulleranes with radii smaller than @xmath5 1015 and dipole moments of order @xmath5 1 d. changes in the degree of hydrogenation moves the peak of the emissivity curves in the range 1060 ghz .
the resulting curves for mixtures of fulleranes in cnm and wim conditions reproduce the recent observations of anomalous microwave emission in the perseus molecular complex .
the bulk of this emission can be explained in terms of electric dipole radiation of small icosahedral fullerenes ( c@xmath1 to c@xmath51 ) with a modest hydrogenation level .
laboratory measurements of the dipole moment of fulleranes are crucial to the further study of these molecules as potential carriers of the dust - correlated microwave emission detected by cosmic microwave background experiments .
draine , b.t . & lazarian , a. 1998a , , 494 , l19 draine , b.t . & lazarian , a. 1998b , , 508 , 157 esquinazzi , p. et al .
2002 phys .
b 66,024429 finkbeiner , d.p . , langston , g.i and minter , a.h .
2004 , , 617 , 350 follmeg , b. , rosmus , p. & werner , h - j 1987 , chem .
, 136 , 562 iglesias - groth , s. , ruiz , a. , bretn , j. & gmez llorente , j.m .
2002 , j. chem .
116 , 1648 iglesias - groth , s. , ruiz , a. , bretn , j. & gmez llorente , j.m .
2003 , j. chem .
118 , 7103 iglesias - groth , s. 2004 , , 608 , l37 kogut , a et .
1996 , , 460 , 1 rittner 1951 , j. chem . phys .
19 , 1030 watson , r.a . et .
al 2005 , , 624 , l89 webster , a.s . , 1991 , nature , 352 , 412 webster , a.s.1992 , a&a , 257 , 750 webster , a.s 1993a , mnras , 263 , 385 webster , a.s 1993b , mnras , 265 , 421 , as a function of radius for neutral fulleranes with various levels of hydrogenation ( filled triangles ) , hydrogenated buckyonions ( open triangles ) in cnm conditions .
the case of a fullerenes and buckyonions with one proton is shown as filled circles or open circles respectively . ]
sr@xmath60 h@xmath60 ) of fulleranes ( solid line)and hydrogenated buckyonions ( dashed line ) with hydrogenation ratio @xmath43 in cnm conditions .
results are plotted for c@xmath1 , c@xmath44 , and c@xmath51 , and for a mixture of fulleranes and buckyonions .
b ) electric dipole emissivity curves for a mixture of fulleranes and hydrogenated buckyonions with three different levels of hydrogenation for both cnm ( solid line ) and wim ( dashed line ) conditions .
the dotted line is the emissivity due to a mixture of singly protonated fullerenes and buckyonions for wim . ]
sr@xmath60 h@xmath60 ) of a mixture of fulleranes and hydrogenated buckyonions in cnm conditions ( dotted and dotted - dashed line ) and a combination of cnm and wim conditions ( dashed line and solid line ) .
see text for details . ] | we study the rotation rates and electric dipole emission of hydrogenated icosahedral fullerenes ( single and multishell ) in various phases of the interstellar medium . using the formalism of draine and lazarian for the rotational dynamics of these molecules in various astrophysical environments , we find effective rotation rates in the range 165 ghz with a trend toward lower rotational frequency as the radius of the molecule increases .
owing to the moderately polar nature of the c h bond , hydrogenated fullerenes ( fulleranes ) are expected to have a net dipole moment and produce electric dipole radiation . adopting the same size distribution proposed for fullerenes in the study of the uv extinction bump ( 2175 ) we predict the dipole electric emission of mixtures of fulleranes for various levels of hydrogenation .
we find that these molecules could be the carriers of the anomalous microwave emission recently detected by watson et al . in the perseus molecular complex . |
many mathematical models used in science and technology contain parameters for which a direct observation is very difficult .
a good example is subsurface geophysics .
the aim in subsurface geophysics is the reconstruction of subsurface properties such as density and permeability given measurements on the surface.using the laws of physics , these properties can be used as parameters of a forward model mapping them to the measurements which we subsequently call data .
inverting such a relationship is non - trivial and lies in the focus of the area of inverse problems .
classically , these parameters are estimated by minimisation of a regularised least squares functional which is based on the data output mismatch ( tikhonov ) .
the idea of this approach is to use optimisation techniques aiming at parameters that produce nearly the same noiseless output as the given noisy data while being not too irregular .
however , it is difficult to quantify how the noise in the data translates into the uncertainty of the reconstructed parameters for this method .
uncertainty quantification is much more straightforward in the bayesian approach .
the basic idea of the bayesian method is that not all parameter choices are a priori equally likely . instead , the parameters are artificially treated as random variables by modelling their distribution using a priori knowledge .
this distribution is accordingly called the prior . for a specific forward model and
given the distribution of the observational noise , the parameters and the data can be treated as jointly varying random variables . under mild conditions , the prior
can then be updated by conditioning the parameters on the data .
the posterior is one of the main tools for making inference about the parameters .
possible estimates include approximation of the posterior mean or the maximum a posteriori ( map ) estimator .
moreover , it is possible to quantify the uncertainty of the reconstructed parameter by posterior variance or posterior probability of a set around for example an estimate of the parameters under consideration .
the main focus of this article lies on posterior consistency which quantifies the quality of the resulting posterior in a thought experiment . as for any evaluation for an approach to inverse problems , an identical twin experiment is performed , that is for a fixed set of parametersand artificial data is generated .
it is conceivable to expect that , under appropriate conditions , the posterior concentrates around this set of true parameters .
results of this type are called posterior consistency .
it justifies the bayesian method by establishing that this method recovers the true parameters sometimes with a specific rate .
so far , there are only posterior consistency results available for linear forward models and mainly gaussian priors @xcite . in this article
, we prove posterior consistency of nonlinear inverse problems with explicit bounds on the rate . the main idea behind our posterior consistency results is to use stability properties of the deterministic inverse problem to reduce posterior consistency of a nonlinear inverse problem to posterior consistency of a bayesian non - parametric regression problem .
our guiding example is the inverse problem of reconstructing the diffusion coefficient from measurements of the pressure .
more precisely , we assume that the relation between the diffusion coefficient @xmath0 and the pressure @xmath1 satisfies the following partial differential equation ( pde ) with dirichlet boundary conditions @xmath2 where @xmath3 is a bounded smooth domain in @xmath4 for this guiding example the required stability results are due to @xcite .
however , our methods are generally applicable to inverse problems with deterministic stability results .
these are often available in the literature because they are also needed for convergence results of the tikhonov regularisation ( consider for example theorem 10.4 . in @xcite ) .
finally , we complete our reasoning by proving appropriate posterior consistency results for the corresponding bayesian non - parametric regression problem . in section [ sec : preliminary ] , we both review preliminary material and give a detailed exposition of our main ideas , steps and results . in section [ sec : general - results ] , we provide novel posterior consistency results for bayesian non - parametric regression . in order to evaluate the rate for the regression problem , we compare our rates to those for gaussian priors for which optimal rates are known . these results are needed in order to obtain posterior consistency for the elliptic inverse problem in section [ sec : results - for - elliptic ] . we obtain explicit rates for priors based on a series expansion with uniformly distributed coefficients . in section [ sec : conclusion ] , we draw a conclusion and mention other inverse problems to which this approach is applicable .
the appendix contains a detailed summary of relevant technical tools such as gaussian measures and hilbert scales which are used in the proofs of our main results .
the author would like to thank professor martin hairer , professor andrew stuart , dr .
hendrik weber and sergios agapiou for helpful discussions .
sjv is grateful for the support of an erc scholarship .
our crucial idea for proving posterior consistency for a nonlinear bayesian inverse problem is the use of stability results which allow us to break it down to posterior consistency of a bayesian regression problem . because the proofs are quite technical , it is worth becoming familiar with the outline of our main ideas first .
therefore this section is intended to motivate , review and summarise our investigation of posterior consistency for a nonlinear inverse problem leaving technical details to the sections [ sec : general - results ] and [ sec : results - for - elliptic ] . for the convenience of the reader
we also repeat the relevant material on bayesian inverse problems in section [ sub : reviewbayesian ] without proofs , thus making our exposition self - contained . in section [ sub : posterior - consistency ] , we precisely define posterior consistency in this setting and place it within the literature . * * s**ubsequently , we introduce an elliptic inverse problem as guiding example for which we apply our method using stability results from @xcite .
finally , we conclude our exposition by giving a general abstract theorem of posterior consistency for nonlinear inverse problems with stability results in section [ sub : consistency - through - stability ] .
the key idea of bayesian inverse problems is to model the input @xmath5 of a mathematical model , for example the initial condition of a pde , as random variable with distribution @xmath6 based on a priori knowledge .
this distribution is called the prior which is updated based on the observed data @xmath7 .
the resulting distribution @xmath8 is called posterior and lies in the focus of the bayesian approach .
we assume that the data is modelled as @xmath9 with @xmath10 being the forward operator , a mapping between the hilbert spaces @xmath11 and @xmath12 , and with the observational noise @xmath13 .
the aim of the inverse problem is the reconstruction of @xmath14 given the data @xmath15 .
because @xmath10 might be non - injective and @xmath16 is unknown , the problem is not exactly solvable as stated . * * if the distribution of the noise @xmath13 is known , then @xmath14 and @xmath15 can be treated as jointly varying random variables . under mild assumptions on the prior , the distribution of the noise and the forward operator
, there exists a conditional probability measure on @xmath14 , called the posterior @xmath17 .
it is an update of the prior using the data and models the a posteriori uncertainty .
therefore it can be viewed as the solution to the inverse problem itself . in this way it is possible to obtain different explanations of the data corresponding to different modes of the posterior . in this article
, we assume that the law of the observational noise @xmath18 is a mean - zero gaussian with covariance @xmath19 . in this case bayes rule
can be generalised for any @xmath10 mapping into a finite dimensional space @xmath12 .
it follows that @xmath20 by @xmath21 we denote the norm of the cameron - martin space @xmath22 of @xmath23 that is the closure of @xmath12 with respect to @xmath24 ( see [ sec : notation ] for more details ) . a proper derivation of equation ( [ eq : generalposterior ] ) , including the fact that its last line is also valid for functional data , and an appropriate introduction to bayesian inverse problems can be found in @xcite and @xcite .
all in all the bayesian approach can be summarised as @xmath25 \mbox{noise } & \noise\:\sim\mathcal{n}(0,\gamma ) \\[10pt ] \mbox{posterior } & \frac{d\mu^{y}}{d\mu_{0}}\propto\exp\left(-\frac{1}{2}\bigl\vert\opobs(\input)\bigr\vert_{\obscov}^{2}+\left\langle y,\opobs(\input)\right\rangle _ { \obscov}\right ) .
\end{array}\label{eq : bayesinv}\ ] ] as one can see in this example , the posterior can usually only be expressed implicitly as an unnormalised density with respect to the prior .
thus , in order to estimate the input parameters or perform inference using the posterior , it has to be probed using either * sampling methods , such as mcmc which aim at generating draws from the posterior or * variational methods for determining the location of an infinitesimal ball with maximal posterior probability .
the second approach is also called the maximum a posteriori probability ( map ) estimator .
it can be viewed as an extension to many classical methods for inverse problems .
for example , it can be linked to the @xmath26-tikhonov regularisation by considering a gaussian prior and noise @xcite .
this relates the choice of norms in the tikhonov regularisation to the choice of the covariance of the prior and the noise .
these regularisation techniques can be justified by convergence results . similarly , inference methods based on the posterior can be justified by posterior consistency , a concept which we introduce in the next section .
as for any approach to inverse problems , the bayesian method can be evaluated by considering an identical twin experiment .
therefore a fixed input @xmath27 , called the _
truth _ , is considered and data is generated using a sequence of forward models @xmath28 which might correspond to the increasing amount of data or diminishing noise .
for each @xmath29 we denote the posterior corresponding to the prior @xmath30 , the noise distribution @xmath31 and the forward operator @xmath32 by @xmath33 . under appropriate assumptions ,
the posterior @xmath34 is well - defined for @xmath35 given by bayes rule in equation ( [ eq : bayesinv ] ) for @xmath30-a.e . @xmath14 and @xmath31-a.e .
@xmath36 ( c.f .
this bayes rule does not give rise to a well - defined measure for arbitrary @xmath7 .
however , we will pose assumptions such that the normalising constant in the bayes rule will be bounded above and below for every @xmath37 belonging to a particular set and @xmath31-a.e .
we will denote these posteriors by @xmath39 .
this sequence of inverse problems is called posterior consistent if the posteriors @xmath40 concentrate around the truth @xmath37 .
we quantify the concentration by the posterior probability assigned to the ball @xmath41 . here
@xmath41 denotes a ball of radius @xmath42 with respect to a metric @xmath43 . in the following we define this concept precisely and
place it within the literature before closing this section by relating posterior consistency to small ball probabilities for the prior .
* [ def : posterior - consistency]*(analogue to @xcite ) a sequence of bayesian inverse problems @xmath44 is posterior consistent for @xmath37 with rate @xmath45 and with respect to a metric @xmath43 if for @xmath46 there exists a constant @xmath47 and a sequence @xmath48 such that @xmath49 we simply say that @xmath44 is posterior consistent if the above holds for any fixed constant @xmath50 .
two important special cases of this definition are * posterior consistency in the small noise limit : @xmath51 * posterior consistency in the large data limit : @xmath52 in the above formulation @xmath53 corresponds to different measurements while @xmath54 denotes the law of a random variable . *
* there exists a variety of results for * * posterior consistency and inconsistency for statistical problems .
two important examples are the identification of a distribution from ( often i.i.d . ) samples or density estimation @xcite .
the former is concerned with considering a prior on a set of probability distributions and the resulting posterior based on @xmath29 samples of one of these probability distributions . in @xcite
, doob proved that if a countable collection of samples almost surely allows the identification of the generating distribution , then the posterior is consistent for almost every probability distribution with respect to the prior .
this very general result is not completely satisfactory because it does not provide a rate and the interest may lie in showing posterior consistency for every possible truth in a certain class .
moreover , some surprisingly simple examples of posterior inconsistency have been provided for example by considering distributions on @xmath55 @xcite .
the necessary bounds for posterior consistency ( c.f . equation ( [ eq : posteriorconsistency ] ) ) can be obtained using the existence of appropriate statistical tests which are due to bounds on entropy numbers .
these methods are used in a series of articles , for example in @xcite .
this idea has also recently been applied to the bayesian approach to linear inverse problems in @xcite . in general , posterior consistency for infinite dimensional inverse problems
has mostly been studied for linear inverse problems in the small noise limit where the prior is either a sieve prior , a gaussian or a wavelet expansion with uniform distributed coefficients @xcite . except for @xcite
, all these articles exploit the explicit structure of the posterior in the conjugate gaussian setting , that means that we have a gaussian prior as well as a gaussian posterior .
in contrast , we consider general priors , general forward operators and gaussian noise in this article .
usually , the posterior has a density with respect to the prior as in equation ( [ eq : bayesinv ] ) .
however , it is possible to provide examples where both the prior and posterior are gaussian but not absolutely continuous .
this can be achieved using for example proposition 3.3 in @xcite .
subsequently , we assume that the posterior has a density with respect to the prior implying that the posterior probability of a set is zero whenever the prior probability of this set is zero .
therefore it is necessary that @xmath37 is in the support of the prior giving rise to the following definition .
[ ass : nondegeneratedef]the support of a measure @xmath56 in a metric space @xmath57 is given by @xmath58 it is natural to expect that the posterior consistency rate depends on the behaviour of @xmath59 as @xmath60 .
asymptotics of this type are called small ball probabilities .
we recommend @xcite as a good survey and refer the reader to @xcite for an up - to - date list of references . in this article
, we consider algebraic rates of posterior consistency , that means we take @xmath61 in definition [ def : posterior - consistency ] . in order to establish these rates of posterior consistency , we consider small ball asymptotics of the following form @xmath62 where @xmath63 and with the notation as in appendix [ sec : notation ] .
both posterior consistency and the contraction rate depend on properties of the prior .
this suggests that we should choose a prior with favourable posterior consistency properties . from a dogmatic point of view
the prior is only supposed to be chosen to match the subjective a priori knowledge . in practice priors
are often picked based on their computational performance whereas some of their parameters are adapted to represent the subjective knowledge .
an example for this is the choice of the base measure and the intensity for a dirichlet process @xcite .
finally , we would like to conclude this section by mentioning that it has been shown in @xcite that posterior consistency is equivalent to the property that the posteriors corresponding to two different priors merge .
the yet unpublished book @xcite contains a more detailed discussion about the justification of posterior consistency studies for dogmatic bayesians .
the aim of this section is to set up the elliptic inverse problem for which we will prove posterior consistency ( c.f .
section [ sub : posterior - consistency ] ) both in the small noise and the large data limit . in a second step
we describe the available stability results and how they can be used to reduce the problem of posterior consistency of a nonlinear inverse problem to that of a linear regression problem .
we end this section by stating a special case of our posterior consistency results in section [ sec : results - for - elliptic ] .
our results do not only apply to this particular elliptic inverse problem but to any nonlinear inverse problem with appropriate stability results ( c.f .
section [ sub : consistency - through - stability ] ) .
however , the results for the elliptic inverse problem are of particular interest because it is used in oil reservoir simulations and the reconstruction of the groundwater flow @xcite .
the forward model corresponding to our elliptic inverse problem is based on the relation between @xmath64 and @xmath14 given by the elliptic pde in [ eq : postconellipticforward ] .
we would like to highlight that the relation between @xmath14 and @xmath64 is nonlinear .
under the following assumptions , the solution operator @xmath65 to the above pde is well - defined @xcite .
( [ ass : ellipticforward]forward conditions ) suppose that 1 .
@xmath3 is compact , satisfies the exterior sphere condition ( see @xcite ) and has a smooth boundary ; 2 .
@xmath66 and f is smooth in @xmath67 ; 3 .
@xmath68 and @xmath69 in equation ( [ eq : postconellipticforward ] ) .
under these assumptions , the regularity results from @xcite yield the following forward stability result . [
prop : forwardstability]if @xmath70 and @xmath71 satisfy assumption [ ass : ellipticforward ] and are elements of @xmath72 for @xmath73 , then @xmath74 the inverse problem is concerned with the reconstruction of @xmath0 given the data @xmath75 which is related to @xmath1 in the following way .
[ ass : forwardsplitting]the forward operator @xmath10 can be split into a composition of the solution operator @xmath64 and an observation operator @xmath76 , that is @xmath77 the bayesian approach to the elliptic inverse problem ( eip ) summarises as @xmath78 \mbox{prior } & \prior\mbox { on } \input \\[7pt ] \mbox{data } & \data=\opobs_{n}(\input)+\noise_{n}=\mathcal{o}_{n}(\pressure(\cdot,\input))+\noise_{n},\,\noise_{n}\sim\mathcal{n}(0,\gamma_{n } ) \\[7pt ] \mbox{posterior } & \frac{d\mu^{n}}{d\prior}(\input)\propto\exp\left(-\frac{1}{2}\bigl\vert\opobs_{n}(\input)\bigr\vert_{\obscov_{n}}^{2}+\left\langle y,\opobs(\input)\right\rangle _ { \obscov_{n}}\right ) .
\end{array}}\quad\left(\mbox{eip}\right)\ ] ] a rigorous bayesian formulation of this inverse problem , with log - gaussian priors and besov priors has been given in @xcite and @xcite respectively .
in @xcite the problem is considered with a prior based on a series expansion with uniformly distributed coefficients ( see section [ sub : posterior - consistency - rate ] ) . in the same article , a generalised polynomial chaos ( gpc )
method is derived in order to approximate posterior expectations .
we consider posterior consistency as set up in definition [ def : posterior - consistency ] in the following cases : * the small noise limit with @xmath79 corresponding to a functional observation and an additive gaussian random field as noise such that @xmath80 * the large data limit with @xmath81 where @xmath82 are evaluations at @xmath83 . in this case
the data takes the form @xmath84 posterior consistency in both cases are based on a stability result which can be derived by taking @xmath14 as the unknown in equation ( [ eq : postconellipticforward ] ) .
this leads to the following hyperbolic pde @xmath85 imposing assumption [ ass : ellipticforward ] , it has been established that there exists a unique solution @xmath14 to this pde without any additional boundary conditions : [ prop : richter]suppose @xmath64 arises as a solution to equation ( [ eq : postconellipticforward ] ) with @xmath14 as diffusion coefficient satisfying assumption [ ass : ellipticforward ]
. then equation ( [ eq : hyperbolicpde ] ) is uniquely solvable for any @xmath86 and @xmath14 such that @xmath87 moreover , if @xmath70 and @xmath71 satisfy these assumptions , then the stability result above and a change of variables ( theorem [ thm : indpendentofdiscription ] ) implies @xmath88 this statement reduces posterior consistency of the eip in @xmath89 to posterior consistency of the following bayesian regression problem ( brp ) in @xmath90 @xmath91 \text{data } & \data=\mathcal{o}_{n}(\pressure)+\noise_{n},\,\noise_{n}\sim\mathcal{n}\left(0,\gamma_{n}\right ) \\[7pt ] \text{posterior } & \frac{d\tilde{\mu}^{y_{n}}}{d\tilde{\prior}}(\pressure)\propto\exp\left(-\frac{1}{2}\bigl\vert\mathcal{o}(\pressure)\bigr\vert_{\obscov_{n}}^{2}+\left\langle y,\mathcal{o}(\pressure)\right\rangle _ { \obscov_{n}}\right ) \\[7pt ] & \text{with } \mathcal{o}_{n}=\text{id}\text { or } \mathcal{o}_{n}=\left(e_{x_{i}}\right)_{i=1}^{n } \end{array}}\quad\left(\text{brp}\right)\ ] ] where @xmath1 is now treated as an variable , that is the prior and the posterior are now formulated on the pressure space .
moreover , @xmath92 denotes the push forward of the prior under @xmath1 .
note that for @xmath93 the brp can also be viewed as the simplest linear inverse problem .
the required posterior consistency results for the brp can be derived from those in section [ sec : general - results ] using interpolation inequalities . in this way
we obtain posterior consistency results in section [ sec : results - for - elliptic ] a special case of which is the following theorem : suppose that the prior @xmath30 satisfies @xmath94 let the noise be given by @xmath95 . if @xmath96 and @xmath97 , then ( eip ) is posterior consistent for any @xmath98 in the small noise limit with respect to the @xmath99-norm for any @xmath100 .
this approach is not limited to the eip as the following section shows . in section [ sub : introelliptic ] , we present our main idea , that is the reduction of the problem of posterior consistency of the eip to that of the brp .
the main ingredients of this reduction are the stability result that was summarised in proposition [ prop : richter ] and the posterior consistency results for the brp .
this approach is not limited to the eip but it is applicable to any inverse problem for which appropriate stability results are available .
this is the case for many inverse problems such as the inverse scattering problem in @xcite or the calderon problem in @xcite .
we would like to point out that these stability results are also crucial for proving the convergence of regularisation methods ( see theorem 10.4 in @xcite ) .
suppose @xmath101 with @xmath102 and @xmath103 .
moreover , we assume that * there exists a stability result of the form @xmath104 * the sequence of bayesian inverse problems @xmath105 is posterior consistent with respect to @xmath106 for all @xmath107 with rate @xmath108 .
then @xmath44 is posterior consistent with respect to @xmath109 for all @xmath110 with rate @xmath111 using the notation of section [ sub : introelliptic ] , we denote the posteriors for the bayesian inverse problems @xmath44 and @xmath105 by @xmath40 and @xmath112 , respectively . then a change of variables ( c.f .
theorem [ thm : indpendentofdiscription ] ) implies @xmath113
as described in the previous section , for many inverse problems posterior consistency can be reduced to posterior consistency of a brp ( c.f .
section [ sub : consistency - through - stability ] ) using stability results .
thus , with the results obtained in this section we may conclude posterior consistency for apparently harder nonlinear inverse problems .
for the eip this is achieved by an application of the results in theorem [ thm : notail ] and [ thm : largedata ] . because the derivation of these two results is quite technical
, we first give a summary and we recommend the reader to become familiar with both theorems but to skip the technical details on the first read .
it is classical to model the response as @xmath114 in the following we consider two bayesian regression models with * @xmath93 and the noise is a gaussian random field that is scaled to zero like @xmath115 or * @xmath81 and @xmath116 corresponding to evaluations of a function with additive i.i.d .
gaussian noise .
these models represent the large data and the small noise limit , respectively . *
* we prove posterior consistency for both problems under weak assumptions on the prior .
this is necessary because the brps resulting from nonlinear inverse problems are usually only given in an implicit form .
for both cases we are able to obtain a rate assuming appropriate asymptotic lower bounds on the small ball probabilities of the prior around @xmath37 ( see section [ sub : posterior - consistency ] )
. moreover , posterior consistency with respect to stronger norms can be obtained using prior or posterior regularity in combination with interpolation inequalities which is the subject of section [ sub : convergence - in - stronger ] . for the large data limit , that is @xmath81
, we obtain posterior consistency with respect to the @xmath89-norm in section [ sub : point - wise - observations - in ] .
we assume an almost sure upper bound on a hlder norm for the prior and an additional condition on the locations of the observations .
the latter is justified by construction of a counterexample .
for the small noise limit , that is @xmath93 , we prove posterior consistency with respect to the cameron - martin norm of the noise in section [ sub : the - small - noise ] .
this norm corresponds to the @xmath117-norm in the hilbert scale with respect to the covariance operator @xmath118 .
both the cameron - martin norm and hilbert scales are introduced in [ sec : notation ] . if an appropriate @xmath119-norm is @xmath30-a.s
. bounded , we obtain an explicit rate of posterior consistency .
otherwise , the rate is implicitly given as a low - dimensional optimisation problem . however , the condition for mere posterior consistency takes a simple form .
( see corollary [ cor : postconsmallnorate ] for the case of general noise ) + suppose that the noise is given by @xmath120 and @xmath121 for @xmath122 and @xmath123 .
then the posterior is consistent in @xmath124 for any @xmath125 if @xmath126 and @xmath127 satisfy the following conditions [ rem : noratecorollary - gaussian]if the prior is gaussian , then the above inequality is satisfied because @xmath128 and the rhs is less than @xmath129 for any @xmath130 .
thus , the only remaining condition is @xmath122 .
[ rem : noratecorollary - logconcave]it is worth pointing out that for the large class of log - concave measures it is known that @xmath131 , for details consult @xcite . in the statistics literature regression models
are mainly concerned with pointwise observations . despite its name
this is also true for _ functional data analysis _ * * ( see @xcite ) .
however , the regression problem associated with @xmath93 can be viewed as a particular linear inverse problem . as described in the introduction ,
this has been studied for gaussian priors in @xcite and @xcite .
although our focus lies on establishing posterior consistency for general priors and non - linear models , we also obtain rates which in the special case of gaussian priors are close to the optimal rates given in the references above . in the following
we study posterior consistency for a bayesian regression problem assuming that the data takes values in the hilbert space @xmath132 .
in particular we deal with the regression model @xmath133 with @xmath134 and @xmath16 all being elements of @xmath132 . moreover
, we suppose that the observational noise @xmath13 is a gaussian random field @xmath18 on @xmath132 and we assume that it satisfies the following assumption .
[ ass : noisetracesumm]suppose there is @xmath135 such that @xmath136 is trace - class for all @xmath137 , that is @xmath138 imposing this assumption , it becomes possible to quantify the regularity of the observational noise in terms of the hilbert scale defined with respect to the covariance operator ( c.f . [ sec : notation ] ) .
more precisely , this is possible due to lemma [ lem : regularitynoise ] . from @xcite .
the regression model in equation ( [ eq : postconinvers ] ) is a special case of a general inverse problem as considered in equation ( [ eq : ip ] ) .
hence the corresponding posterior takes the following form ( c.f . equation ( [ eq : bayesinv ] ) ) .
@xmath139 assuming that the data takes values in the hilbert space @xmath132 , equation ( [ eq : brpposterior ] ) can simply be derived by an application of the cameron - martin lemma in combination with the conditioning lemma ( lemma 5.3 in @xcite ) .
we generate data for a fixed truth @xmath37 @xmath140 by changing the normalising constant , we may rewrite the posterior in the following way @xmath141 the normalising constant is bounded above and below for @xmath142 for @xmath31-a.e .
@xmath16 . in fact , this holds under weaker assumptions than needed for our results .
[ lem : brpnormalising]suppose @xmath143 for @xmath144 and @xmath145 .
then the normalising constant in equation ( [ eq : identityposterior ] ) is bounded for @xmath146-a.s . and every @xmath147 above and away form zero .
see [ sec : normal ] .
the expression above suggests that the posterior concentrates in balls around the truth in the cameron - martin norm .
first , we make this fact rigorous for priors which are a.s .
uniformly bounded with respect to the @xmath148-norm . in a second step , we assume that the prior has higher exponential moments .
considering gaussian priors , we show that our rate is close to the optimal rate obtained in @xcite .
the following theorem can be viewed as a preliminary step towards theorem [ thm : postconsmallgeneral ] which contains our most general posterior consistency result for the bayesian regression problem in the small noise limit .
while containing our main ideas , the following result also establishes an explicit rate for posterior consistency which will be used for the eip in section [ sec : results - for - elliptic ] .
[ thm : notail ] suppose that the noise satisfies assumption [ ass : noisetracesumm ] and @xmath149 for @xmath150 if @xmath151 and then @xmath40 is consistent in @xmath152 . additionally , if the following small ball asymptotic is satisfied @xmath153 then this holds with rate @xmath154 for any @xmath155 with @xmath156 .
our proof is based on the observation that posterior consistency is implied by the existence of a sequence of subsets @xmath157 such that @xmath158 and @xmath159 where @xmath160 .
this implication holds because @xmath161 and thus @xmath162 which together with @xmath158 implies posterior consistency , for details see equation ( [ eq : posteriorconsistency ] ) .
fix @xmath163 .
then @xmath164 with @xmath165 as @xmath166 sufficiently slow .
we notice that lemma [ lem : regularitynoise ] implies that @xmath167 as @xmath166 .
the remainder of the proof will be devoted to showing that equation ( [ eq : ratiolimit ] ) holds .
we bound @xmath168 by smoothing @xmath16 at the expense of @xmath169 @xmath170 interpolating between @xmath152 and @xmath171 for @xmath172 ( c.f .
lemma [ lem : hscaleinterpolation ] ) yields @xmath173 with @xmath174 .
an application of equation ( [ eq : identityposterior ] ) yields the following upper bound @xmath175\nonumber\\
\hspace{-0.4 cm } & \ge \hspace{-0.1cm}z(n,\xi)\exp\left[-n^{1 - 2\kappa}\left[\frac{\epsilon\bigl\vert a - a^{\dagger}\bigr\vert_{1}}{2}\right]^{2}-k_{n}n^{\frac{1}{2}-\lambda\kappa}\left(\frac{\epsilon}{2}\right)^{\lambda}\right]\hspace{-0.1cm}\prior\hspace{-0.1cm}\left[b_{\frac{\epsilon}{2}n^{-\kappa}}^{1}\left(\truth\right)\right]\hspace{-0.1cm}.\label{eq : postconsmallnottaillower}\end{aligned}\ ] ] similarly , we obtain the following upper bound @xmath176 the expression in the exponential in equation ( [ eq : identityposterior ] ) can be rewritten as a function @xmath177 of @xmath178 which is decreasing on @xmath179 . if @xmath180 then @xmath181 for @xmath182 and @xmath29 large enough leading to @xmath183 we now derive sufficient conditions for @xmath184 to be the dominant term in the exponential in the equations ( [ eq : postconsmallnottaillower ] ) and ( [ eq : postconsmallnottailupper ] ) implying equation ( [ eq : ratiolimit ] ) .
this is the case if , in addition to inequality ( [ eq : postconsmallnotailrate1 ] ) , @xmath185 hold .
the first line is equivalent to inequality ( [ eq : postconsmallnotailrate1 ] ) and using inequality ( [ eq : yhmnotailsmallball ] ) the second line is implied by @xmath186 thus , the inequalities ( [ eq : postconsmallnotailrate1 ] ) and ( [ eq : postconsmallnotailrate2 ] ) imply that @xmath187 is the dominant term in the inequalities ( [ eq : postconsmallnottaillower ] ) and ( [ eq : postconsmallnottailupper ] ) establishing equation ( [ eq : ratiolimit ] ) .
letting @xmath188 concludes the proof .
in the following we weaken the assumptions of theorem [ thm : notail ] by assuming that the prior has exponential moments of @xmath189 .
the price we pay is that the algebraic rate of convergence is implicitly given as a low - dimensional optimisation problem .
[ thm : postconsmallgeneral]suppose that the noise satisfies assumption [ ass : noisetracesumm ] , the prior satisfies the small ball asymptotic @xmath190 and @xmath191@xmath192 for @xmath123 and @xmath193 for @xmath144 .
if the following optimisation problem has a solution @xmath194 , then for any @xmath195 the posterior @xmath40 is consistent in @xmath152 for @xmath37 in @xmath171 with rate @xmath196 .
@xmath197 where @xmath198
. see appendix [ sec : proof - of - theorem ] .
[ rem : postcongenerals]in general , @xmath199 might depend on @xmath172 for @xmath191@xmath192 to hold . therefore the rate might be improved by optimising over different @xmath144 .
whereas the algebraic rate in theorem [ thm : postconsmallgeneral ] is implicit , the following corollary yields a simple condition implying posterior consistency . [ cor : postconsmallnorate]suppose that the noise satisfies assumption [ ass : noisetracesumm ] , and @xmath191@xmath192 for @xmath123 , @xmath193 and @xmath144 .
if one of the following two conditions holds @xmath200 then @xmath40 is posterior consistent for @xmath37 in @xmath171 .
it follows from the proof of theorem [ thm : postconsmallgeneral ] that we only have to find @xmath201 , @xmath202 , @xmath203 and @xmath172 such that the inequalities ( [ eq : postconsmallrate1 ] ) , ( [ eq : postconsmallrate2 ] ) , ( [ eq : postconsmallrate5 ] ) , ( [ eq : postconsmallrate7 ] ) and ( [ eq : postconsmallrate8 ] ) are satisfied . choosing @xmath201 as large as inequality ( [ eq : postconsmallrate1 ] ) permits , that is @xmath204 , extends the range of solutions of the other inequalities ( ( [ eq : postconsmallrate2 ] ) and ( [ eq : postconsmallrate8 ] ) ) containing @xmath201
similarly , choosing @xmath205 as large as ( [ eq : postconsmallrate2 ] ) permits , that is @xmath206 , extends the range of solutions of inequality ( [ eq : postconsmallrate8 ] ) .
letting @xmath60 in ( [ eq : postconsmallrate8 ] ) yields @xmath207@xmath208 now it is left to perform a case - by - case analysis .
starting from inequality ( [ eq : postconnorate1 ] ) , the first two cases are @xmath209 and @xmath210 . for these cases we have to treat @xmath211 and @xmath212 separately in order to rearrange equation ( [ eq : postconnorate1 ] ) to a quadratic inequality in @xmath1 .
the details are tedious but straightforward algebra .
we would like to point out that the remarks [ rem : noratecorollary - gaussian ] and [ rem : noratecorollary - logconcave ] are also valid for this more general corollary [ cor : postconsmallnorate ] .
in the special case of jointly diagonalisable prior and noise covariance , we evaluate the consistency rate in theorem [ thm : postconsmallgeneral ] by comparing it with the optimal rates obtained in @xcite . by numerically solving the optimisation problem in theorem [ thm : postconsmallgeneral ] , we indicate that our rates are close to the optimal rate . in the following we first derive a gaussian prior and noise for a regression problem before reformulating our result in this context . in a second step
we reformulate the problem in the notation of @xcite and state the corresponding result .
we conclude this section by an actual comparison between the posterior consistency rate obtained in @xcite and the results of this paper .
we suppose that the prior is gaussian @xmath213 and that the covariance operators @xmath214 of the prior and @xmath118 of the noise are jointly diagonalisable over @xmath215 denoting an orthonormal basis of eigenvectors .
furthermore , we assume that the eigenvalues @xmath216 and @xmath217 of @xmath214 and @xmath118 satisfy @xmath218 respectively .
the inner product of the hilbert scale with respect to @xmath118 can now explicitly be written as @xmath219 moreover , we remark that assumption [ ass : noisetracesumm ] is satisfied with @xmath220 .
the covariance operator @xmath221 of @xmath30 on @xmath171 has eigenvalues @xmath222 which can be seen by denoting @xmath223 and calculating @xmath224 & = \left\langle \priorcov s^{2sr}u , s^{2sr}u\right\rangle _ { \mathcal{h}}=\left\langle s^{2sr}\priorcov u , v\right\rangle _ { \mathcal{h}^{s}}.\nonumber \end{aligned}\ ] ] in order to conclude that @xmath221 is trace - class on @xmath225 , we need to impose that @xmath226 . in this case , we know from example 2 and proposition 3 in section 18 of @xcite that the small balls asymptotic @xmath190 is satisfied for @xmath30 with @xmath227 .
for this problem we adapt theorem [ thm : postconsmallgeneral ] by optimising over @xmath172 in the appropriate range as described in remark [ rem : postcongenerals ] .
moreover , fernique s theorem @xcite for gaussian measures motivates us setting @xmath128 and @xmath227 as discussed above .
[ cor : gaussiancase ] let the prior and the observational noise be specified as in equation ( [ eq : postconsmallgaussprior ] ) and ( [ eq : postconsmallgaussnoise ] ) . if the following optimisation problem has a solution @xmath194 , then for any @xmath195 the posterior @xmath40 is consistent in @xmath152 for @xmath37 in @xmath171 with rate @xmath196 .
@xmath228 @xmath229 where @xmath230 .
we now recast our problem reformulating it in the setting and notation of @xcite .
letting @xmath231 be @xmath132-valued white noise , our problem corresponds to recovering @xmath14 from @xmath232 this problem is equivalent to @xmath233 where @xmath234
let @xmath235 be an orthonormal basis of eigenvectors of @xmath19 on @xmath132 .
in order to adapt the notation of @xcite , we write @xmath236 and note that @xmath237 will be equivalent to the cameron - martin space which takes the form @xmath238 with orthonormal basis @xmath239 .
moreover , let @xmath240 be defined as @xmath241 in order to match assumption 3.1 in @xcite , we have to bound the eigenvalues @xmath242 of @xmath243 as follows @xmath244 we determine these eigenvalues by noting that @xmath245 the calculation above yields @xmath246 and thus @xmath247 as in equation ( [ eq : determinecov ] ) , we identify the covariance operator of @xmath30 on @xmath237 through its eigenvalues @xmath248 by theorem 4.1 in @xcite the posterior contraction rate is given by @xmath249 where @xmath250 ( compare equation ( 3.5 ) in @xcite ) and @xmath251 is the regularity of the truth .
as above , we suppose that @xmath252 resulting in @xmath253 in figure [ fig : comparison ] , we use numerical optimisation to compare our rate to the optimal one for @xmath254 with varying @xmath255 . just considering inequality ( [ eq : postconsmallnoisesmallball ] ) ( essential to our approach since this implies that the cameron - martin term dominates the prior measure c.f .
equation ( [ eq : postconhilbertscaledeltaball ] ) ) yields @xmath256 which coincides with the rate @xmath257 obtained by solving the optimisation problem in corollary [ cor : gaussiancase ] .
thus , even if we are able to improve our bounds , there is a genuine gap between our rate and the optimal rate in the case of gaussian priors .
the reason for this gap is that theorem [ thm : postconsmallgeneral ] is applicable to any prior satisfying the stated regularity and small ball assumptions .
nevertheless , figure [ fig : comparison ] indicates that the obtained rates are quite close .
in contrast , @xcite is only applicable to gaussian priors for which the gaussian stucture of the prior and the posterior are explicitly used .
posterior consistency rate for the bayesian regression model with the noise and prior given in equations ( [ eq : postconsmallgaussprior ] ) and ( [ eq : postconsmallgaussnoise ] ) .
we denote the rate obtained in @xcite and the one based on corollary [ cor : gaussiancase ] as @xmath258 and @xmath259 , respectively .
we also plot @xmath260 an upper bound on the rate that is obtainable with our method which is based on the small ball asymptotics of the prior.,scaledwidth=79.0% ] we consider the following non - parametric bayesian regression problem @xmath261 with @xmath262 , @xmath3 a bounded domain and @xmath263 .
we assume that a prior @xmath30 is supported on @xmath264 resulting in a posterior of the form @xmath265 subsequently , we will prove posterior consistency for this problem for the case @xmath266 $ ]
. however , the same reasoning applies to any bounded domain @xmath267 but the actual posterior consistency rate depends on @xmath43 . as in the previous section ,
we suppose that the data @xmath268 in equation ( [ eq : largedatamodel ] ) is generated for a fixed truth @xmath37 .
hence @xmath269 in this setup posterior consistency depends on the properties of the prior as well as on the sequence @xmath270 . in the following ,
we discuss appropriate assumptions on both giving rise to theorem [ thm : largedata ] .
moreover , we relate this result with its assumptions to the literature .
[ ass : postcondatapriorass]there exist @xmath271 $ ] and @xmath272 such that @xmath273 as @xmath29 increases , we gain more and more information about the function @xmath14 . in particular , if @xmath270 is dense in @xmath274 $ ] it is even possible to reconstruct the value of @xmath275 from @xmath276 .
more precisely , let @xmath277 $ ] be arbitrary , then there are @xmath278 such that @xmath279 however , we will see that this is not sufficient for posterior consistency .
in fact , we will give an example of posterior inconsistency for this case .
so far , the problem of posterior consistency for this type of regression problems has mainly been investigated for random evaluation points @xmath280 which are known as random covariates .
appropriate results of this type can be found in * @xcite*. an exception is @xcite where posterior consistency without a rate with respect to the @xmath281-norm for deterministic @xmath280 is shown .
this result is obtained under the following assumption .
@xcite[ass : choi ] suppose that there exists a constant @xmath282 such that whenever @xmath283 for @xmath284 there is at least one @xmath285 such that @xmath286 .
the above condition guarantees that the number of observations in each interval satisfies a lower bound proportional to its size .
more precisely , an interval of size @xmath196 has at least order @xmath287 for @xmath29 being large enough .
this can be seen by chopping the interval into intervals of size @xmath288 .
in contrast to @xcite , we are also able to obtain a posterior consistency rate under this assumption .
posterior consistency without a rate can be concluded under the following weaker assumption .
[ ass : equally spaced ] we suppose that for @xmath270 there exists a @xmath289 such that for any @xmath290 $ ] there is an @xmath291 such that @xmath292 where @xmath293 denotes the empirical distribution of @xmath294 [ thm : largedata]suppose that the assumptions [ ass : postcondatapriorass ] and [ ass : equally spaced ] are satisfied .
then @xmath295 is posterior consistent with respect to the @xmath89-norm for any @xmath296 .
moreover , if assumption [ ass : choi ] and the small ball asymptotic @xmath297 are satisfied , then @xmath295 is posterior consistent with respect to the @xmath89-norm with any rate @xmath196 and @xmath298 as in theorem [ thm : notail ] , posterior consistency is implied by @xmath299 for increasing sets @xmath157 such that @xmath300@xmath301 . for notational convenience
we write @xmath302 @xmath303 and we denote by @xmath201 a generic @xmath304 random variable .
this allows us to rewrite the posterior in equation ( [ eq : postconlargedata ] ) as @xmath305 since @xmath306 is finite dimensional , it is easy to see that @xmath307 is bounded from above and below . again , fixing @xmath163 , we only need to consider @xmath308 . thus ,
for @xmath309 we have a lower bound on @xmath310 in order to derive an upper bound on @xmath311 , let @xmath312 be chosen arbitrarily and notice that @xmath313 is decreasing for @xmath314 .
the upper bound on @xmath311 therefore boils down to a lower bound on @xmath315 that is larger than @xmath316 in fact , there is @xmath317 such that @xmath318 . applying hlder continuity yields @xmath319\ ] ] for @xmath320 .
let @xmath321 be the following index set @xmath322\right\ } .\ ] ] for @xmath29 larger than @xmath323 it follows that @xmath324 if we only consider @xmath280 with @xmath325 , we obtain that @xmath326 which gives rise to the following upper bound @xmath327.\label{eq : thmlargedatapostconupper}\ ] ] by choosing @xmath328 small enough , we also know that @xmath329 in order to obtain a rate of posterior consistency , we use @xmath330 and hence @xmath331.\label{eq : postconlargelower}\end{aligned}\ ] ] thus , equation ( [ eq : thmlargedatapostconupper ] ) implies that @xmath332 the first term in the exponential in equation ( [ eq : postconlargelower ] ) is dominant over the corresponding term in equation ( [ eq : postconlargeupper ] ) by choosing @xmath333 moreover , the first * * term in the equations ( [ eq : postconlargelower ] ) and ( [ eq : postconlargeupper ] ) is dominant over the other terms respectively if @xmath334 these three inequalities are respectively implied by @xmath335 choosing @xmath336 small enough , we see that @xmath40 is consistent in @xmath89 with any rate @xmath337 the assumptions in the theorem above can be justified because a slight violation leads to the example of posterior inconsistency in the next section . in this section ,
we construct a counterexample to illustrate that despite the strong assumption [ ass : postcondatapriorass ] it is not sufficient for @xmath338 to be dense in order to establish posterior consistency .
given such a sequence it is always possible to extract a subsequence satisfying assumption [ ass : equally spaced ] .
even though all the other observations can be viewed as additional , we will choose a prior so that the posterior sequence is not consistent . in the following ,
we choose the prior concentrated on functions @xmath339 which are continuous , satisfy @xmath340 and are linear on @xmath341 $ ] and * * @xmath342 $ ] . by identifying @xmath343 and @xmath344 with the first and second component respectively the following two - dimensional example
can be extended to the setting of equation ( [ eq : largedatamodel ] ) .
this extension is an example of posterior inconsistency with respect to the @xmath345-norm for @xmath346 because any of these norms is equivalent to @xmath347 for an arbitrary norm on @xmath348 .
we consider the following prior on @xmath348 @xmath349 and we choose @xmath350 as truth. the data consists of @xmath29 and @xmath351 with @xmath352 being measurements of the form @xmath353 respectively .
consequently , the posterior takes the form @xmath354 here , posterior consistency of @xmath355 is equivalent to the statement that for any @xmath282 there is @xmath356 such that @xmath357 we will not only show that @xmath355 is posterior inconsistent but also that there is @xmath358 such that for @xmath359 @xmath360 because of @xmath361 , we may proceed as in the proofs of the theorems [ thm : notail ] and [ thm : largedata ] and thus it is enough to construct sets of increasing @xmath362-probability such that on these sets @xmath363 & = & \frac{\sum_{k}\exp\left(-\frac{1}{2}\sum_{j=1}^{n}\frac{1}{k}-\frac{2}{k}\xi_{j}-2k^{2}\right)}{\sum_{k}\exp\left(-\frac{1}{2}\sum_{j=1}^{n}\frac{1}{4k}-\frac{1}{\sqrt{k}}\xi_{j}-\frac{1}{2}\sum_{j=1}^{n^{\theta}}1 - 2\tilde{\xi}_{k}-k^{2}\right)}\rightarrow0.\end{aligned}\ ] ] the @xmath362-probabilities of @xmath364 and @xmath365 are exponentially small in @xmath29 .
thus , it is enough to consider @xmath366 & \leq\max\left\ { \exp\left(-\frac{3}{4}\sqrt{n}\right),\exp(-n)\right\ } \rightarrow0\text { as } n\rightarrow\infty.\end{aligned}\ ] ] hence we have shown that @xmath40 is not posterior consistent .
this example relies on the prior having strong correlations between its two components .
therefore it seems an interesting question how the assumptions on @xmath30 can be strengthened in order to relax those on @xmath367 we conclude this section by showing that interpolation inequalities can be used in order to strengthen the norm in which the posterior concentrates .
in particular we consider the small noise limit as described in section [ sub : the - small - noise ] .
suppose we know that the posterior concentrates around the truth @xmath37 in the cameron - martin norm @xmath368 . in order to show consistency in @xmath369 ,
we write @xmath370 the posterior probability of the first set is small due to the posterior consistency in @xmath152 .
the posterior probability of the second set is small due to the tails of the prior and the posterior .
obtaining estimates of this type can be done similarly to the steps subsequent to equation ( [ eq : postconhilbertscalesposteriormoment ] ) in the proof of theorem [ thm : postconsmallgeneral ] . using this technique ,
it is also possible to apply the results of this section to the eip in the next section .
a similar technique based on interpolation inequalities between hlder spaces applies to the large data limit and is also used for the eip .
in section [ sub : introelliptic ] , we introduced the idea of reducing posterior consistency of the eip to that of the brp . for this example
we demonstrate our method for both the small noise and the large data limit .
we start by giving the proof for the small noise limit in detail before sketching the same steps for the large data limit . *
* we emphasise the case of posterior consistency in the small noise limit because of its analogy with convergence results for regularisation methods . using theorem [ thm : notail ] from section [ sec : general - results ] to conclude posterior consistency of the eip ( c.f .
section [ sub : introelliptic ] ) is not entirely straightforward because we have to lift the posterior consistency for the brp to @xmath90 .
moreover , we have to find appropriate assumptions on the prior @xmath30 so that the push forward prior @xmath92 satisfies the assumptions of the theorem [ thm : notail ] . again
, a rate of posterior consistency is obtained if the prior satisfies appropriate small ball asymptotics . in a second step we verify those for the so - called uniform priors which are based on a series expansion with uniformly distributed coefficients , for details see below or consider @xcite . in order to formulate assumptions on @xmath30 implying that @xmath92 satisfies the assumptions of theorem [ thm : notail ] , we assume for simplicity that @xmath371 where @xmath372 denotes the laplacian with homogeneous dirichlet conditions . in this case the abstract hilbert scale @xmath171 ( c.f . [
sec : notation ] ) corresponds to the standard sobolev space @xmath373 .
thus , the almost sure bounds in theorem [ thm : notail ] are implied by the appropriate assumptions on the prior and classical results from @xcite .
moreover , the choice @xmath371 also implies that assumption [ ass : noisetracesumm ] holds for @xmath374 .
this is due to the fact that the operator @xmath375 has eigenvalues @xmath376 with @xmath377 ( see section [ sec : notation ] for notation ) where @xmath43 denotes the dimension of the domain @xmath3 .
these results are called weyl asymptotics and further details can be found in @xcite and @xcite .
the following theorem summarises the consequences for the posterior consistency of the eip .
[ thm : ellipticnorate]suppose that the noise is given by @xmath378 and that the prior @xmath30 satisfies @xmath379 if @xmath96 , @xmath380 and @xmath381 , then the eip is posterior consistent with respect to the @xmath99-norm for any @xmath100 .
additionally , if @xmath382 then the eip is posterior consistent with respect to the @xmath89-norm with rate @xmath196 for any @xmath383 such that @xmath384 before proving theorem [ thm : ellipticnorate ] , we notice that forward stability results ( as proposition [ prop : forwardstability ] ) can be used to transfer small ball asymptotics from @xmath30 to @xmath385 .
[ lem : pushforwardsmallball]if the prior satisfies the small ball asymptotic @xmath386 then @xmath387 proposition [ prop : forwardstability ] implies that @xmath388 \text { and } p\left(b_{\epsilon}^{c^{\beta}}\left(\truth\right)\right ) & \subseteq & b_{c\epsilon}^{c^{\beta+1}}\left(\pressure^{\dagger}\right).\end{aligned}\ ] ] hence the statement follows .
having established lemma [ lem : pushforwardsmallball ] , we are now in the position to prove the main theorem of this section .
subsequently , @xmath47 will denote a generic constant in different contexts that may change form line to line .
we will first prove posterior consistency in @xmath89 before we use an interpolation inequality to bootstrap it to @xmath99 . in order to prove posterior consistency in the @xmath89-norm
, it is enough to show posterior consistency of the brp in the @xmath90-norm because @xmath389 which follows by an application of proposition [ prop : richter ] and a change of variables ( see theorem [ thm : indpendentofdiscription ] ) . using theorem 6.19 from @xcite
, we may conclude that @xmath390 since @xmath391 , @xmath1 is @xmath392-a.s .
an element of the cameron - martin space of @xmath23 as it corresponds to @xmath124 .
posterior consistency of the brp with respect to the @xmath124-norm is now implied by theorem [ thm : notail ] .
its conditions are satisfied because @xmath393 with @xmath394 furthermore , proposition [ prop : forwardstability ] and the fact that @xmath381 imply that @xmath395 . in order to bootstrap to posterior consistency in the @xmath90-norm
, we use a generalisation of the sobolev embedding theorem for besov spaces and an interpolation inequality between besov spaces on domains ( for details consult @xcite ) .
we first note that @xmath396 and @xmath397 for @xmath398 .
in particular theorem 4.33 in @xcite implies that @xmath399 for @xmath163 being small .
if @xmath400 , we can conclude posterior consistency in the @xmath90-norm because @xmath401 holds for @xmath336 small enough .
otherwise , we use the interpolation inequality between besov spaces subject of theorem 4.17 in @xcite @xmath402 for @xmath336 small enough and with @xmath403 .
similar to equation ( [ eq : ellipticsmallnoiseenough ] ) , it follows that @xmath404 the equations ( [ eq : ellipticsmallnoiseenough ] ) and ( [ eq : ellipticsmallnoiseinterp ] ) allow us to bootstrap the posterior consistency of @xmath405 to @xmath406 equation ( [ eq : thmnottailnoratestability ] ) implies posterior consistency of @xmath407 in the @xmath89-norm .
similarly , we bootstrap to posterior consistency in @xmath99 for @xmath100 using the same interpolation technique as above .
in order to obtain a rate for posterior consistency , we first note that @xmath408 implies @xmath409 due to lemma [ lem : pushforwardsmallball ] .
now theorem [ thm : notail ] implies posterior consistency of the sequence of posteriors @xmath405 in @xmath124 with any rate @xmath383 such that @xmath410 using the interpolation inequality as above gives rise to posterior consistency for * * @xmath405 in @xmath411 with rate @xmath196 for any @xmath383 such that @xmath384 as above , this implies the same rate of posterior consistency for @xmath40 in @xmath412 in this section , we establish a rate of posterior consistency for the eip with the so - called uniform prior introduced in @xcite .
this choice of the prior was motivated by the preceding analysis in the uncertainty quantification literature , see for instance @xcite .
it is given by @xmath413\label{eq : uniformprior}\ ] ] where @xmath414 denotes the law of a random variable .
moreover , we suppose that @xmath415 , @xmath416 and @xmath417 such that @xmath418 in order to obtain a rate for the eip with this prior , we derive a small ball asymptotic under an appropriate assumption on the decay of @xmath419 . [
ass : coefficientdecay]there exists @xmath420 such that for all @xmath421 @xmath422 since the series in equation ( [ eq : uniformprior ] ) is absolutely convergent , we assume without loss of generality that @xmath423 is decreasing .
this allows us to use the following classical inequality from approximation theory @xcite @xmath424 [ thm : smallballprior]suppose that @xmath30 is given as in equation ( [ eq : uniformprior ] ) , assumption [ ass : coefficientdecay ] is satisfied with @xmath425 and @xmath426.\ ] ] then for any @xmath421 @xmath427 we obtain an asymptotic lower bound on the small ball probability by choosing an appropriate subset @xmath428 of @xmath429 .
we denote a generic element of this set by @xmath430 choosing @xmath431 such that @xmath432 , the corresponding terms contribute at most @xmath433 to the difference @xmath434 .
the subset @xmath428 prescribes intervals for @xmath435 @xmath436 such that this contribution is at most @xmath433 , too .
more precisely , let @xmath437 , then equation ( [ eq : summabilityexp ] ) implies @xmath438 for @xmath439 .
let the subset @xmath440 be given by @xmath441 then @xmath442 combining lemma [ thm : smallballprior ] and theorem [ thm : ellipticnorate ] results in the following theorem which characertises posterior consistency for this class of priors .
let the prior @xmath30 be defined as in equation ( [ eq : uniformprior ] ) and let assumption [ ass : coefficientdecay ] be satisfied .
additionally , we assume that @xmath443 , @xmath96 and @xmath444 then the posterior @xmath40 is consistent for any @xmath445\ ] ] with respect to with rate @xmath446 for any @xmath383 such that @xmath447 in the following we show that the results for the brp can be transferred to posterior consistency results in the large data limit for the eip .
we consider only the case @xmath448 with @xmath266 $ ] as the general case is similar .
furthermore , assuming that the observations are of the form @xmath449 the sequence of posteriors is given by @xmath450 posterior consistency of the eip in @xmath89 can then be derived on the basis of theorem [ thm : largedata ] .
suppose that the sequence @xmath451 satisfies assumption [ ass : equally spaced ] , @xmath452 @xmath30-a.s . with @xmath453 and @xmath454
@xmath30-a.s .. if @xmath455 , then the eip is posterior consistent in the large data limit with respect to @xmath456 for any @xmath457 an application of theorem 6.13 in @xcite yields the existence of @xmath458 so that for all @xmath14 satisfying @xmath459 there is a unique solution @xmath1 such that @xmath460 .
thus , @xmath461 satisfies the assumptions of theorem [ thm : largedata ] implying that @xmath462 is posterior consistent in @xmath412 using the interpolation inequality between @xmath89 and @xmath463 , we also obtain consistency in @xmath90 . as in theorem
[ thm : ellipticnorate ] , proposition [ prop : richter ] can be used in order to conclude posterior consistency of @xmath464 in @xmath89 .
we can bootstrap from @xmath465 to @xmath456 by interpolating between @xmath89 and @xmath466 .
in this article , we have established a novel link between stability results for an inverse problem and posterior consistency for the bayesian approach to it .
we have explicitly shown this link for an elliptic inverse problem ( c.f .
eip ) but the same method is also applicable for the general case .
an instance is electrical impedance tomography ( caldern problem ) for which stability results are available @xcite .
this example would lead to a very slow posterior consistency rate since its stability results are weak .
essentially , we would have to redo all the calculations on a log - scale instead of an algebraic scale .
so far , we need exponential moments of the prior for the bayesian regression of functional response and for pointwise observations ( see also section 4.2.2 in @xcite ) . for this reason it is harder to prove posterior consistency for example for log - gaussian priors .
log - gaussian measures have moments of arbitrary order but no exponential moments .
this is a problem that we would like to pursue further in the future .
we use the following notation for asymptotic inequalities : let @xmath467 and @xmath468 be sequences in @xmath469 .
we denote by @xmath469 @xmath470 that there are @xmath471 and @xmath472 such that @xmath473 moreover , if @xmath474 , we write @xmath475 in order to measure the smoothness of the noise and samples of the prior , we introduce hilbert scales following @xcite .
let @xmath19 be a self - adjoint , positive - definite , trace - class linear operator with eigensystem @xmath476 .
we know that @xmath477 is a densely defined , unbounded , symmetric and positive - definite operator because @xmath478 we define the _ hilbert scale _ by @xmath479 with @xmath480 for @xmath481 we will denote balls with respect to the @xmath482-norm by @xmath483 moreover , these collection of norms satisfies an interpolation inequality [ lem : hscaleinterpolation](proposition 8.19 in @xcite ) let @xmath484 then the following interpolation inequality holds @xmath485 our definition here is slightly different from the literature in order to match it to the sobolev spaces for @xmath486 in this section , we set out our notation for some standard results about infinite dimensional gaussian measures which can be found in the following textbooks and lecture notes @xcite .
let @xmath336 be a gaussian measure on a hilbert space @xmath487 .
it is characterised by its _
mean _ given by the bochner integral @xmath488 and the _ covariance operator _
@xmath489 characterised by the relation @xmath490 from this it is clear that the covariance operator is positive - definite and self - adjoint .
moreover , we note that @xmath19 is necessarily trace - class and the gaussian can be expressed through eigenvalues @xmath376 and the corresponding eigenbasis @xmath491 @xmath492 the _ cameron - martin space _ associated with @xmath336 is @xmath493 equipped with the inner product @xmath494 where @xmath495 and @xmath496.this space characterises the support as well as the direction such that @xmath497 where @xmath498 is the translation operator @xmath499 .
we also consider the hilbert scale @xmath500 generated by @xmath336 and the regularity of a draw @xmath501 can be expressed as follows .
[ lem : regularitynoise](@xcite ) imposing assumption [ ass : noisetracesumm ] the following statements hold : 1 .
let @xmath231 be a white noise , then @xmath502 for all @xmath137 .
2 . let @xmath503 , then @xmath504 @xmath505-a.s .
for every @xmath137 .
the state of a model can be described in several ways . in this section , we present the resulting relationship between two different descriptions of the same model .
[ thm : indpendentofdiscription]suppose @xmath101 with @xmath102 and @xmath506 .
furthermore , assume that the posterior @xmath40(@xmath112 ) is well - defined for the forward operator @xmath507(@xmath508 ) , the prior @xmath509 ( @xmath510 ) and the noise @xmath511 .
it is given by @xmath512 in this case @xmath513 .
it is sufficient to show that both measures agree on all sets @xmath514 @xmath515 by the transformation rule @xmath516
we follow the same steps as in the proof of theorem [ thm : notail ] up to equation ( [ eq : postconnotailinterp ] ) reading @xmath517 with @xmath174 .
we now separate the product using young s inequality with @xmath518 @xmath519 where , for simplicity of notation , we used @xmath520 .
+ the following lower bound on @xmath521 is based on equation ( [ eq : postconsmallnoisecminner ] ) @xmath522\right].\label{eq : postconhilbertscaledeltaball}\end{aligned}\ ] ] the term @xmath184 has to be dominant in equation ( [ eq : postconhilbertscaledeltaball ] ) because the same exponent is appearing in equation ( [ eq : postconhilbertscaledeltaballcomplement ] ) except for a larger coefficient .
choosing @xmath523 and substituting the expression for @xmath524 , this is the case if @xmath525 we need small ball probabilities and the exponential moments of @xmath30 in order to obtain explicit sufficient conditions on @xmath383 .
we first note that @xmath526 equation ( [ eq : postconsmalltmp1 ] ) holds if @xmath527 we bound @xmath529 by @xmath530 we denote by @xmath532 the following supremum @xmath533 which is finite if @xmath534 the first two summands above can be rewritten as a function @xmath535 of @xmath536 where @xmath537 by considering @xmath538 , we see that @xmath535 is decreasing for @xmath539 .
thus , for @xmath540 the following inequality holds @xmath541.\label{eq : postconhilbertscaledeltaballcomplement } \end{aligned}\ ] ] then for large @xmath29 , equation ( [ eq : postconhilbertscalenecdecreasing ] ) is implied by @xmath542 in this section , we bound @xmath544 using markov s inequality in combination with the exponential moments of the prior @xmath545 we denote the term appearing in the exponential in the second line by @xmath546 .
it can be bounded similar to the upper bound on * @xmath531 * @xmath547 we denote by @xmath548 an inequality with a multiplicative constant not involving @xmath29 or @xmath383 . in order to get an upper bound for equation ( [ eq : postconhilbertscalesposteriormoment ] ) , we bound the exponential moment by @xmath549 introducing @xmath550 and performing an integration by parts , it follows that @xmath551 + 1d\prior(\input)\\ & \lesssim\int_{0}^{\infty}g^{\prime}(r)\exp\left(g(r)+\mathfrak{u}_{t_{0}}\right)d\prior\left(\left\vert \input\right\vert _ { s}>r\right)dr\\ & \lesssim\int_{0}^{\infty}g^{\prime}(r)\exp\left(n^{\frac{1}{2}-\hilbertscalesyoungs}r^{(1-\hilbertscalesinter)q}-2fr^{e}\right)dr.\end{aligned}\ ] ] the above can only be expected to be finite if @xmath552 moreover , we assume that @xmath553 since otherwise @xmath554 in order to achieve an upper bound , we split the term in the exponential into @xmath555 and @xmath556 .
the first term is negative whenever @xmath557 for @xmath29 large enough @xmath558 holds . on the interval
@xmath559 $ ] an upper bound @xmath560 on the maximum value of @xmath561 can be derived as follows @xmath562 putting everything together gives rise to @xmath563 for some @xmath0 . using markov s inequality , this yields
@xmath564 again substituting @xmath523 , this is asymptotically smaller than @xmath565 if @xmath566 collecting the inequalities from above , we see that the results follow by letting @xmath188 .
in order to bound @xmath567 in equation ( [ eq : identityposterior ] ) , we rewrite it as @xmath568 where @xmath569 .
we bound @xmath570 using the cauchy - schwarz inequality @xmath571 the following steps are quite similar to the steps in the proof of the theorems [ thm : notail ] and [ thm : postconsmallgeneral ] .
we treat @xmath572 by smoothing @xmath16 at the expense of @xmath0 @xmath573 we use the interpolation inequality for hilbert scales with @xmath174 ( see lemma [ lem : hscaleinterpolation ] ) and hlder s inequality with @xmath518 to obtain @xmath574 combining these bounds yields @xmath575 the first three terms are bounded in @xmath0 because they are dominated by the first if @xmath576 .
this is implied by choosing @xmath577 note that @xmath578 is @xmath579-a.s .
bounded due to lemma [ lem : regularitynoise ] .
thus @xmath580 is bounded below if @xmath581 .
letting @xmath582 in @xmath583 we see that this is the case for @xmath584 an upper bound on @xmath580 follows from a simple lower bound on @xmath570 on @xmath585 and the prior measure of this set . | in the bayesian approach , the a priori knowledge about the input of a mathematical model is described via a probability measure .
the joint distribution of the unknown input and the data is then conditioned , using bayes formula , giving rise to the posterior distribution on the unknown input . in this
setting we prove posterior consistency for nonlinear inverse problems : a sequence of data is considered , with diminishing fluctuations around a single truth and it is then of interest to show that the resulting sequence of posterior measures arising from this sequence of data concentrates around the truth used to generate the data .
posterior consistency justifies the use of the bayesian approach very much in the same way as error bounds and convergence results for regularisation techniques do . as a guiding example
, we consider the inverse problem of reconstructing the diffusion coefficient from noisy observations of the solution to an elliptic pde in divergence form .
this problem is approached by splitting the forward operator into the underlying continuum model and a simpler observation operator based on the output of the model . in general ,
these splittings allow us to conclude posterior consistency provided a deterministic stability result for the underlying inverse problem and a posterior consistency result for the bayesian regression problem with the push - forward prior .
moreover , we prove posterior consistency for the bayesian regression problem based on the regularity , the tail behaviour and the small ball probabilities of the prior . |
in their study of multiple - access collision channel without feedback , massey and mathys show that the capacity region can be achieved by deterministic channel access method @xcite .
it contrasts with the more traditional multiple - access scheme like pure aloha or slotted aloha@xcite , where transmission times of packets form a random process . in this paper , we extend the transmission scheme by massey and mathys to tandem network , in which nodes are located on a straight line .
nodes that are more than two hops away do not interfere with each other .
if a node receives two packets that overlap in time , either partially or completely , both packets are assumed erased and unrecoverable .
this model is applicable to wireless sensor network along a highway or river for instance .
since antenna system that transmits and receives at the same time is often too costly to implement , we assume that each node operates in half - duplex mode ; when a node is not transmitting , it listens to its two neighboring nodes . in our proposed transmission scheme ,
the transmission times of packets of each node follow a pre - assigned deterministic and periodic pattern .
implementation of this kind of deterministic channel accessing scheme is particularly easy . we can simply store the whole pattern of transmission times in memory , read it out repeatedly , and send out packets accordingly . no coordination between nodes and no centralized packet scheduling
is needed .
network operations are thus fully distributed .
this feature is especially suitable for low - complexity wireless sensor network .
related work on tandem network can be found in @xcite . in @xcite ,
the authors consider one source node and one destination node , that are connected by a series of intermediate relay nodes in tandem , and model each hop as an erasure channel . a coding scheme which can approach the min - cut bound @xcite is proposed .
similar system setting with general discrete memoryless channel in each hop is considered in @xcite , and scaling laws for capacity is discussed .
the networks considered in @xcite and @xcite are full - duplex wireline network . in @xcite ,
half - duplex wireless network is investigated , and the effect of random multiple - access on network coding @xcite is addressed . in this paper , we devise a transmission scheme that combines multiple - access protocol , erasure - correction coding and network coding efficiently .
we will show that this cross - layer design can achieve the ( zero - error ) capacity region of bi - directional tandem collision network , and is thus optimal .
this paper is organized as follows .
the system model is introduced in section [ sec : system ] .
we consider both slot - synchronous and slot - asynchronous case .
a transmission scheme that incorporates multiple - access , erasure correction and network coding for slot - synchronous system is presented in sections [ sec : achievable ] .
then we show that it can be extended to slot - asynchronous system . in order to show that the proposed transmission scheme is optimal for the bi - directional tandem collision network , we derive an outer bound on capacity region in section [ sec : converse ] , and
show that the achievable rate region and the outer bound coincide .
comparison with some random access schemes is carried out in section [ sec : example ] .
we close with some concluding remarks in section [ sec : conclusion ] .
we consider @xmath0 nodes located on a straight line . each node broadcasts signal to its two neighbors , one on the left and one on the right .
it is assumed that the transmit range of each node is adjusted so that there is no interference to the nodes that are two or more hops away .
we model the network as a directed graph with vertex set @xmath1 and edge set @xmath2 we assume that all nodes operate in half - duplex mode , meaning that each of them can not transmit and receive at the same time .
suppose that there are @xmath3 independent data sources , and each data source is associated with a node . for @xmath4
, we let @xmath5 be the vertex to which the @xmath6-th source is attached .
the function @xmath7 is called the _ source mapping_. each source is multicast to a subset of nodes in @xmath8 . define the _ destination mapping _ , @xmath9 , which is a function from @xmath10 to @xmath11 , such that for @xmath4 , source @xmath6 is demanded by all nodes belonging to set @xmath12 .
node @xmath13 is called a _ source node _ if @xmath14 for some source @xmath6 .
if @xmath15 for some @xmath6 , then node @xmath13 is called a _ destination node_. we remark that a node may be a source node and a destination node simultaneously .
a node that is neither a source nor a destination node is called a _ pure relay node_. ( two - way network ) the two nodes at the two ends send data to each other , and the nodes in the middle are pure relay nodes . fig .
[ fig : linear_network1 ] illustrates an example for @xmath16 .
there are two data sources , and so we have @xmath17 .
nodes 1 wants to send message @xmath18 to node 4 , and node 4 wants to send message @xmath19 to node 1 . in this example
, we have @xmath20 , @xmath21 , @xmath22 , @xmath23 .
the message in square bracket signifies that it is demanded by the associated node .
nodes 2 and 3 are pure relay nodes .
( -1,-.5)(7,2 ) ( 0,0)n11 ( 2,0)n22 ( 4,0)n33 ( 6,0)n44 ( 0,0.7)@xmath24 $ ] ( 0,1.2)@xmath18 ( 6,0.7)@xmath25 $ ] ( 6,1.2)@xmath19 ( bi - directional multicast network ) the network consists of five nodes .
two sources are associated with nodes 2 and 4 .
node 2 wants to send message @xmath18 to nodes 1 and 5 , and node 4 wants to send message @xmath19 to nodes 1 and 5 ( fig .
[ fig : linear_network2 ] ) .
the system parameters are @xmath26 , @xmath17 , @xmath27 , @xmath21 , and @xmath28 . node 3 acts as a pure relay node .
( -.5,-.5)(6.5,2 ) ( 0,0)n11 ( 1.5,0)n22 ( 3,0)n33 ( 4.5,0)n44 ( 6,0)n55 ( 0,0.7)@xmath29 $ ] ( 6,0.7)@xmath29 $ ] ( 1.5,0.7)@xmath18 ( 4.5,0.7)@xmath19 we assume that the data stream is packetized , and the durations of all packets are @xmath30 seconds .
each packet may assume @xmath31 possible values , for some positive integer @xmath31 .
thus , each packet carries @xmath32 bits .
consider node @xmath13 and its two neighboring nodes @xmath33 and @xmath34 .
if either node @xmath33 or @xmath34 transmits a packet , while the other remains silent during the whole packet duration , the packet is assumed to be received with no error at node @xmath13 .
however , if both node @xmath33 and @xmath34 transmit and the two packets overlap either partially or completely , then both packets are considered lost and unrecoverable at node @xmath13 . in this case , we say that there is a collision at node @xmath13 . as an example , suppose that node @xmath33 transmits a packet at time @xmath35 .
this packet is successfully received by node @xmath13 if and only if node @xmath13 is in receive mode between time @xmath35 and @xmath36 , and node @xmath34 does not transmit any packet between time @xmath37 and @xmath36 .
we call this network a _
tandem collision network_. we adopt the approach in @xcite and impose the restriction that the packet transmission times are independent of the messages to be sent or forwarded , and independent of how the other nodes access the channel .
this can be accomplished by statically assigning each node a _ protocol signal _ for channel access .
the protocol signal for node @xmath13 , @xmath38 , is a deterministic and periodic signal of period @xmath39 second , and is equal to either zero or one for all @xmath40 .
a protocol signal is equal to one over some semi - open intervals whose lengths are integral multiple of @xmath30 .
node @xmath13 is required to transmit packets whenever @xmath41 , and remain silent whenever @xmath42 . due to the lack of common time reference among the nodes ,
the protocol signals are subject to delay offsets .
we denote the delay offset of node @xmath13 by @xmath43 , which takes value between 0 and @xmath39 .
we assume that the delay offsets are arbitrarily chosen but fixed throughout the communication session .
two protocol signals with delay offsets @xmath44 and @xmath45 are plotted in fig .
[ fig : protocol_signal ] .
a packet is sent within the duration of each `` square pulse . ''
we can see from fig .
[ fig : protocol_signal ] that the first two packets of both nodes are collided , but the third packet from node 2 can be received successfully at node 1 .
we define the duty factor of @xmath38 by @xmath46 it measures the fraction of time that node @xmath13 is transmitting .
( -1,-1)(8,4 ) ( 0,-1)(0,3 ) ( 0,0)(1.8,0)(1.8,1)(2.8,1)(2.8,0)(4.2,0)(4.2,1)(5.2,1)(5.2,0)(7.5,0 ) ( 0,2)(1.5,2)(1.5,3)(2.5,3)(2.5,2)(3.7,2)(3.7,3)(4.7,3)(4.7,2)(5.7,2)(5.7,3)(6.7,3)(6.7,2)(7.5,2 ) ( -0.2,0)@xmath47 ( -0.2,2)@xmath48 ( 0.9,-0.5)@xmath44 ( 0,-0.2)(1.8,-0.2 ) ( 2.3,-0.5)@xmath30 ( 1.8,-0.2)(2.8,-0.2 ) ( 4.7,-0.5)@xmath30 ( 4.2,-0.2)(5.2,-0.2 ) ( 0.75,1.5)@xmath45 ( 0,1.8)(1.5,1.8 ) ( 2,1.5)@xmath30 ( 1.5,1.8)(2.5,1.8 ) ( 4.2,1.5)@xmath30 ( 3.7,1.8)(4.7,1.8 ) ( 6.2,1.5)@xmath30 ( 5.7,1.8)(6.7,1.8 ) we consider both slot - synchronous and slot - asynchronous systems . in the _ slot - synchronous _ case , the delay offsets @xmath43 , for @xmath49 , are integral multiples of packet duration @xmath30 . in the _ slot - asynchronous case _
, the delay offset @xmath43 is an arbitrary real number between 0 and @xmath39 , for @xmath49 .
a time interval in the form @xmath50 for some integer @xmath51 , with reference to the local clock , is called a _
time slot_. we say that a system is _ time - slotted _ if each packet is sent within a time slot .
this means that the protocol signal in a time - slotted system satisfies @xmath52 where @xmath53 refers to the smallest integer larger than or equal to @xmath54 .
it can be easily deduced from that in a time - slotted system , the period @xmath39 of protocol signal @xmath38 is an integral multiple of @xmath30 .
the notions of slot - synchronous system and time - slotted system are _ not _ the same and should be distinguished .
`` time - slotted '' is an attribute pertaining to the protocol signal set .
`` slot - synchronous '' is about synchronization of clocks among the nodes .
[ fig : protocol_signal ] provides an example that is neither slot - synchronous nor time - slotted .
it is not slot - synchronous because the difference between the two delay offsets , @xmath55 , is not an integral multiple of @xmath30 .
it is not time - slotted because the gap between the two pulses in @xmath47 is not an integral multiple of @xmath30 . in a time - slotted but slot - asynchronous system ,
partial overlap of packets is inevitable .
however , if a system is both time - slotted and slot - synchronous , then whenever two packets overlap , they overlap completely . for time - slotted system
, we can compactly specify the protocol signals by binary sequences , called _
protocol sequences_. a protocol signal of period @xmath56 in a time - slotted system , where @xmath57 is an integer , corresponds to a discrete - time periodic sequence with period @xmath57 , denoted by @xmath58 $ ] , such that @xmath59 = 1 \text { if and only if } s_i(t ) = 1 \text { for } t \in [ kt , ( k+1)t).\ ] ] we will abuse the use of language and denote both protocol signal and protocol sequence by the letter `` @xmath60 '' . nevertheless , we can distinguish protocol signal and protocol sequences by using parenthesis for continuous time index and square bracket for discrete time index .
we assume that the protocol signals are jointly designed and known to all nodes , and will not be changed throughout the communication session in concern . in this paper , we do _ not _ assume any packet header , and hence a packet only carries data and does not contain any sender s information .
as we will see in a later section , the sender of a packet can be identified by some sliding - window algorithm , which requires the knowledge neighboring nodes protocol signals .
we will also show that the maximal system throughput can be achieved without any packet header .
nevertheless , if packet headers are present , as in most practical systems , we can relax the requirement that the protocol signals are known to all nodes . the block diagram of a node which is both a source and a destination is shown in fig . [ fig : block_diagram1 ] .
the encoder produces packets as a function of the data source and all previously received packets .
the transmission times of the packets are determined by the protocol signal generator , which is a stand - alone device without any input from the source , the decoder , or the channel .
the computation of decoder s output is based on the content of the successfully received packets as well as the data from the local source . in fig .
[ fig : block_diagram1 ] , flow of data is indicated by solid arrow , and flow of control signal is indicated by dashed arrow .
the block diagram of a source node is the same as in fig .
[ fig : block_diagram1 ] except that the data sink is absent .
[ fig : block_diagram2 ] shows the block diagram of a destination node .
the output of the decoder is feedback and re - encoded . for a pure relay node ,
the block diagram is the same as in fig .
[ fig : block_diagram2 ] except that there is no sink .
( -1,-1)(7,6 ) ( 0,0 ) ( 0,5 ) ( 2,0 ) ( 4,5 ) ( 4,2 ) ( 5.8,0 ) ( 5.8,5 ) ( 5.8,2.5 ) ( -1,-1)(7,6 ) ( 0,0 ) ( 2,0 ) ( 4,5 ) ( 4,2.5 ) ( 5.8,0 ) ( 5.8,5 ) ( 5.8,2.5 ) we describe the channel model formally as follows .
we consider @xmath61 seconds of transmission time , where @xmath62 is an integer and @xmath30 is the time duration of a packet .
each data source is discrete memoryless over a @xmath31-ary alphabet @xmath63 . for @xmath4 , source
@xmath6 produces source symbols at a rate of @xmath64 symbols per packet duration .
each source symbol is chosen uniformly and independently from @xmath63 .
let @xmath65 denote a message from source @xmath6 . in a duration of @xmath62 packet durations .
@xmath65 may assume any value in @xmath66 .
since it is assumed that each packet can store a @xmath31-ary symbol as well , we can store one source symbol in a packet . for given delay offsets @xmath67 ,
the encoding and decoding functions are specified as follows .
if node @xmath13 is a source node , associated with source @xmath6 , the packet transmitted by node @xmath13 at time @xmath40 is obtained by applying the encoding function @xmath68 where @xmath69 is the number of successfully received packets by node @xmath13 up to time @xmath40 . as the delay offsets are fixed , @xmath69 is known and well - defined .
if node @xmath13 is not a source node , then the encoding function for the packet at time @xmath40 is @xmath70 suppose source @xmath6 is demanded by node @xmath13 .
if node @xmath13 is associated with another source , say source @xmath71 , then at the end of @xmath62 packet durations , node @xmath13 decodes @xmath65 by @xmath72 if node @xmath13 is not a source node and wants to decode the data from source @xmath6 , then the decoding function has the form @xmath73 for @xmath15 , let the estimated value of the message from source @xmath6 by node @xmath13 be denoted by @xmath74 .
we say that there is a decoding error if @xmath75 for some @xmath15 and @xmath76 .
a rate vector @xmath77 with @xmath78 for all @xmath6 is said to be _ achievable _ if there are @xmath0 protocol signals @xmath79 , and encoding and decoding functions , such that for all @xmath6 , source @xmath6 can be multicast to all nodes in @xmath12 at a rate of @xmath64 symbols per packet duration with no decoding error , regardless of the delay offsets @xmath67 .
the closure of the set of all achievable rate vectors , denoted by @xmath80 , is called the _ zero - error capacity region _ , or simply the _ capacity region_. the subscript `` @xmath81 '' signifies that the capacity region is for slot - asynchronous system . for slot - synchronous system , we define achievable rate vector similarly , except that the delay offsets are restricted to integral multiples of @xmath30 .
the corresponding capacity region is denoted by @xmath82 .
we say that a tandem collision network is _ bi - directional _ if for all @xmath6 such that @xmath83 , we can find @xmath13 and @xmath84 in @xmath12 such that @xmath85 . in words
, it means that unless a source is associated with the left - most or right - most node , its message has to be sent to a node to the left and a node to the right .
the two tandem collision networks in examples 1 and 2 are both bi - directional .
one of the main result in this paper is that , for bi - directional tandem collision network , the capacity region for both slot - synchronous and slot - asynchronous case can be achieved by the transmission scheme given in the next section .
once the delay offsets @xmath43 are fixed , the transmission times of all packets in the future are also fixed .
each link between two nodes becomes a deterministic erasure channel ; the number of successfully received packets at each node per period is determined .
so , for each fixed combination of delay offsets , the tandem collision network reduces to an _ aref network _ @xcite .
as we are interested in the worst - case throughput over all possible delay offsets , the analysis in this paper can be interpreted as taking the minimum throughput of a collection of aref networks .
the assumption that the protocol sequences are periodic is not restrictive , since in practice , all pseduo - random number generators are periodic , and in theory , the period can be arbitrarily large .
we have the assumption that the transmission times of packets are independent of the messages to be transmitted or forwarded , and independent of the transmission of the other nodes .
the capacity region in this paper is obtained under the condition that channel accessing is done by protocol signal , so that collision avoidance algorithm and re - scheduling of packet transmission times etc .
are not allowed .
we first give a transmission scheme for slot - synchronous system , and then describe an extension to the slot - asynchronous case at the end of this section .
the transmission scheme to be devised is time - slotted , and is based on a special class of protocol sequence , called shift - invariant protocol sequences , and a joint channel - network coding scheme .
we remark that this transmission scheme can be applied to tandem collision network which may or may not be bi - directional .
however , for bi - directional network , we will show in the next section that the achievable rates are indeed optimal .
as mentioned in section [ sec : system ] , a protocol signal in time - slotted system can be specified by a zero - one sequence , @xmath86 $ ] , so that a packet is transmitted in the @xmath51-th time slot @xmath50 if and only if @xmath86 = 1 $ ] .
let the smallest common period of the @xmath0 protocol sequences be denoted by integer @xmath87 .
the duty factor of a zero - one sequence @xmath86 $ ] of period @xmath87 is defined as @xmath88.\ ] ] it can be easily seen that this is compatible with the notion of duty factor for continuous - time protocol signal . for slot - synchronous system ,
the delay offsets are all integral multiples of packet duration @xmath30 . with a slight abuse of language we model the delay offsets by integers , instead of real numbers .
we will denote the delay offsets of node @xmath13 by an integer @xmath89 , for @xmath49 .
the actual delay is @xmath90 .
[ [ shift - invariant - protocol - sequences ] ] shift - invariant protocol sequences + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + shift - invariant protocol sequences are first used by massey and mathys @xcite to achieve the capacity region of the multiple - access collision channel without feedback .
the following definition is from @xcite . given a set of @xmath0 zero - one sequences , @xmath58 $ ] , for @xmath49 , with common period @xmath87 , and a subset @xmath91 , the _
generalized hamming cross - correlation _ is defined by @xmath92.\ ] ] we note that for a subset @xmath93 of @xmath94 consisting of two elements , the notion of generalized hamming cross - correlation reduces to the usual hamming cross - correlation for a pair of sequences . for @xmath93 that is a singleton , it reduces to the hamming weight . a set of @xmath0 protocol sequences
is said to be _ shift - invariant _ if for each subset @xmath95 , the generalized hamming cross - correlation is independent of delay offsets . for the application in this paper , we do not need the full force of shift invariance . in a network with linear topology ,
any transmission from a node two or more hops away does not cause any interference .
this justifies the restriction of our attention to generalized hamming cross - correlation for subset @xmath93 consisting of three or less elements .
we say that a set of @xmath0 protocol sequences is _
consecutively 3-wise shift - invariant _
if for each subset @xmath93 of size three or less , consisting consecutive integers from @xmath94 , the generalized hamming cross - correlation @xmath96 is independent of the delay offsets .
we define the _ throughput _ from node @xmath13 to node @xmath34 , denoted by @xmath97 , as the number of packets from node @xmath13 to node @xmath34 without collision in a period divided by @xmath87 , @xmath98 ( 1-s_{i+1}[k-\tau_{i+1 } ] ) ( 1-s_{i+2}[k-\tau_{i+2 } ] ) .
\label{eq : throughput2}\ ] ] here , @xmath89 , @xmath99 and @xmath100 are integers representing the delay offsets of node @xmath13 , @xmath34 and @xmath101 , respectively .
we note that the throughput from node @xmath13 to node @xmath34 is affected only by the protocol signal of nodes @xmath13 , @xmath34 and @xmath101 , and hence is a function of @xmath89 , @xmath99 and @xmath100 .
the throughput from node @xmath13 to node @xmath33 is similarly defined as @xmath98 ( 1-s_{i-1}[k-\tau_{i-1 } ] ) ( 1-s_{i-2}[k-\tau_{i-2 } ] ) \label{eq : throughput3}\ ] ] and denoted by @xmath102 .
if consecutively 3-wise shift - invariant protocol sequences are used in a time - slotted tandem collision network , then for all delay offsets @xmath103 , @xmath104 , @xmath89 , @xmath99 and @xmath100 , we have @xmath105 and @xmath106 where @xmath107 is the duty factor of the @xmath13-th protocol sequence . [ lemma : si ] the above lemma says that the throughput function does not depend on the relative delay offsets , provided that the protocol sequences are consecutively 3-wise shift - invariant . the proof is based on an elementary property of zero - one sequences @xcite , which is included here for the sake of completeness .
we prove only the statement for @xmath97 .
the proof for the second one is similar and omitted .
the summand in can be expanded as a linear combination of four terms , namely @xmath108 $ ] , @xmath108s_{i+1}[k-\tau_{i+1}]$ ] , @xmath108s_{i+2}[k-\tau_{i+2}]$ ] , and @xmath109s_{i+1}[k-\tau_{i+1 } ] s_{i+2}[k-\tau_{i+2}].\ ] ] after taking the summation over @xmath51 , we can see that @xmath110 is equal to the sum of four generalized hamming cross - correlations , each of which is independent of the delay offsets @xmath89 s by the shift - invariant assumption .
hence the linear combination is also independent of the delay offsets .
this proves that @xmath110 is independent of delay offsets .
let this value be denoted by @xmath111 .
next , we sum over @xmath89 , @xmath99 and @xmath100 . after exchanging the order of summations ,
we obtain @xmath112 \sum_{\tau_{i+1}=1}^p ( 1-s_{i+1}[k-\tau_{i+1 } ] ) \\ & \qquad \cdot \sum_{\tau_{i+2}=1}^p ( 1-s_{i+2}[k-\tau_{i+2 } ] ) \\ & = \frac{1}{p } \sum_{k=1}^p ( p f_i ) ( p ( 1-f_{i+1 } ) ) ( p ( 1-f_{i+2 } ) ) \\ & = p^3 f_i ( 1-f_{i+1 } ) ( 1-f_{i+2}).\end{aligned}\ ] ] this proves the first part of the lemma .
we write the duty factor of the protocol sequence @xmath58 $ ] as @xmath113 , where @xmath114 is a common denominator .
we next construct consecutively 3-wise shift - invariant sequences with period @xmath115 .
given positive integer @xmath62 and @xmath114 , let @xmath116 denote a @xmath114-dimensional row vector whose @xmath62 components on the left are 1 , and the @xmath117 entries on the right are 0 , @xmath118.\ ] ] * construction : * given @xmath0 fractions @xmath113 for @xmath49 , we construct @xmath0 protocol sequences of period @xmath115 as follows . for @xmath119 ,
let @xmath58 $ ] be the concatenation of @xmath120 copies of @xmath121 , @xmath122.\ ] ] for @xmath123 , let @xmath58 $ ] be the concatenation of @xmath114 copies of @xmath124 , @xmath125.\ ] ] for @xmath126 , let @xmath58 $ ] be @xmath127 .
suppose that @xmath26 and the the duty factors are @xmath128 for @xmath129 , and @xmath130 for @xmath131 .
the protocol sequences constructed by the above method are @xmath132:\ & 100\ 100\ 100\ 100\ 100\ 100\ 100\ 100\ 100 \\ s_2[k]:\ & 111\ 000\ 000\ 111\ 000\ 000\ 111\ 000\ 000 \\ s_3[k]:\ & 111\ 111\ 111\ 000\ 000\ 000\ 000\ 000\ 000 \\ s_4[k]:\ & 110\ 110\ 110\ 110\ 110\ 110\ 110\ 110\ 110 \\ s_5[k]:\ & 111\ 111\ 000\ 111\ 111\ 000\ 111\ 111\ 000\end{aligned}\ ] ] the protocol sequences by the above construction method are consecutively 3-wise shift - invariant .
[ prop : si ] the proof of prop .
[ prop : si ] is similar to theorem 8 in @xcite , and is omitted .
[ [ identifying - the - senders - of - a - packet ] ] identifying the senders of a packet + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we will use an interesting property of consecutively 3-wise shift - invariant sequences , called _
identifiability_. consider node 1 .
all successfully received packets at node 1 come from node 2 .
also , it is also clear that all successfully received packets at node @xmath0 come from node @xmath133 . identifying the sender of packets is trivial for the two nodes at the ends .
for the nodes in between , it is not immediate to tell whether the packets are from the left or from the right .
one method of providing the identity of senders is to attach a header at the beginning of each packet .
however , if consecutively 3-wise shift - invariant protocol sequences are used , the receiver can identify the sender of all non - collided packets by observing the channel activities , without looking into the content of the packets .
hence theoretically , we can dispense with packet header .
after a period of @xmath87 time slots , a node records the status of channel in each time slot . consider a given node , say node @xmath13 .
we indicate the time slot when node @xmath13 is transmitting by @xmath134 . for the time slot when node @xmath13 is receiving , we use symbols `` 0 '' , `` 1 '' and `` @xmath135 '' to indicate whether there is 0 , 1 , or more than 1 packets is received . the symbol `` @xmath135 '' represents a collision .
we call this sequence of `` @xmath134'',``0 '' , `` 1 '' and `` @xmath135 '' the _ channel activity signal_. if we can always determine the senders of all non - collided packets from the channel activity signal , regardless of what the delay offsets are , then we say that the protocol sequence set is _ identifiable_. as an example , suppose that all delay offsets are zero , and the protocol sequences in example 3 are used .
the channel activity signal observed at node 2 is @xmath136 if the protocol sequences are identifiable , the receiver at node 2 is able to deduce from the above channel activity signal that the packets at time slots 5 , 6 , 8 and 9 are from node 3 , and the packets at time slots 13 , 16 , 22 and 25 are from node 1 .
we remark that identifiability does not mean that the delay offsets are determined .
in fact , for the channel activity signal in , we can cyclically shift @xmath137 $ ] by any multiple of 3 , without any change in the channel activity signal .
therefore , determining the delay offsets uniquely using only the channel activity signal is not possible in general .
we now show that the property of identifiability is implied by consecutively 3-wise shift invariance .
let @xmath138 $ ] denote the channel activity signal observed by node @xmath13 .
suppose that the delay offsets of nodes @xmath33 and @xmath34 are @xmath104 and @xmath99 , which is unknown to node @xmath13 .
given @xmath139 and @xmath140 from 0 to @xmath141 , node @xmath13 compute @xmath142 = \begin{cases } \delta & \text{if } s_i[k+\tau_i ] = 1 \\ 0 & \text{if } s_i[k+\tau_i ] = 0 \text { and}\\ & \qquad s_{i-1}[k+\tau_{i-1 } ' ] = s_{i+1}[k+\tau_{i+1 } ' ] = 0 , \\ * & \text{if } s_i[k+\tau_i ] = 0 \text { and } \\ & \qquad s_{i-1}[k+\tau_{i-1 } ' ] = s_{i+1}[k+\tau_{i+1 } ' ] = 1 , \\ 1 & \text{if } s_i[k+\tau_i ] = 0 \text { and either } \\ & \ \ \text{$s_{i-1}[k+\tau_{i-1}']=1 $ or $ s_{i+1}[k+\tau_{i+1}']=1$}. \\
\end{cases}\ ] ] node @xmath13 would observe @xmath143 $ ] if the delay offsets of node @xmath33 and @xmath34 were @xmath139 and @xmath140 respectively .
we search for @xmath139 and @xmath140 such that the corresponding @xmath143 $ ] is the same with the true channel activity signal @xmath138 $ ] .
there is always at least one solution to this search problem , because @xmath144 and @xmath145 is one such solution . in general
, there may be multiple solutions .
after one such pair of @xmath146 is found , we then declare that the packet with time index @xmath51 satisfying @xmath143 = s_{i\pm 1 } [ k+\tau_{i\pm 1}']=1 $ ] is sent from node @xmath147 .
the correctness of this algorithm , provided that the protocol sequences @xmath148 $ ] , @xmath149 $ ] and @xmath150 $ ] are consecutively 3-wise shift - invariant , is given in the appendix .
[ [ determining - the - delay - offsets ] ] determining the delay offsets + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have seen that the identifiability property is insufficient to determine the delay offsets . nevertheless , it is pointed out in @xcite that we can find the delay offsets by the following initialization mechanism at the beginning of the communication session .
suppose a node began to transmit information starting from some finite time in the past , and before that , the zero packet was transmitted in the infinite past when a node is required to transmit a packet by the protocol sequence .
following the notation in @xcite , we call the @xmath151 packets from node @xmath13 within a period of @xmath87 time slots as a _ frame_.
before any information packets are transmitted , node @xmath13 first sends the following @xmath152 frames @xmath153 $ ] , @xmath154 $ ] , @xmath155 , \ldots , [ 0,\ldots , 0 , 1]$ ] . in the first frame ,
all packets contain the value @xmath156 in the alphabet set @xmath63 .
in each of the remaining @xmath157 frames , there is exactly one packet which equals 1 .
we now describe how node @xmath34 determine the delay offset of node @xmath13 .
before node @xmath13 starts transmitting , node @xmath34 can only observe idle time slots , or packets from node @xmath101 .
when node @xmath34 first receives a packet containing the symbol 1 from user @xmath13 , node @xmath34 buffers @xmath158 periods of packets .
( node @xmath34 can distinguish the packets from node @xmath13 and @xmath101 by the identifiability property of shift - invariant sequences . )
node @xmath34 then tries to find a pair of packets from node @xmath13 , containing the value `` 1 '' , such that their time difference is an integral multiple of @xmath87 .
suppose that there are two such packets separated by @xmath159 time indices for some integer @xmath160 .
the first packet should belong to the very first frame @xmath161 $ ] , and the second to the frame @xmath162 $ ] with a `` 1 '' in the @xmath160-th position .
since node @xmath34 knows the protocol sequence of node @xmath13 , @xmath58 $ ] , the time index of the @xmath160-th `` 1 '' in @xmath58 $ ] is also known .
the time indices of this pair of packets , reduced modulo @xmath87 , is the relative delay offset of node @xmath13 .
[ [ joint - channel - network - coding ] ] joint channel - network coding + + + + + + + + + + + + + + + + + + + + + + + + + + + + the encoder in each node has two objectives : combat the erasures caused by collisions ( channel coding ) and facilitate information flow in both directions ( network coding ) .
this is achieved by a joint channel - network coding scheme called _ nested coding _ , which is a coding technique found useful in cooperative relaying @xcite .
the main difference between the nested coding scheme in this paper and those in @xcite is that our target is to correct erasures , while the nested codes in @xcite are aimed at correcting errors .
we continue the notation that the duty factor at node @xmath13 , @xmath107 , is equal to a fraction @xmath113 , for @xmath49 . with the use of the consecutively 3-wise shift - invariant protocol sequences constructed earlier ,
the common period of the protocol sequences is @xmath163 , and the number of packets sent out by node @xmath13 in a period is @xmath164 . in this section
, we assume that @xmath31 is a prime power and @xmath165 for @xmath49 , and assume that @xmath63 is the finite field of @xmath31 elements .
let the element of @xmath63 be ordered in some arbitrary way , @xmath166 we use reed - solomon ( rs ) code @xcite as a building block of the nested coding .
let @xmath167 be the set of all polynomials over @xmath63 with degree less than or equal to @xmath168 , @xmath169 the coefficients of a polynomial in @xmath167 are treated as information symbols to be encoded .
given any subset @xmath170 of @xmath63 with @xmath171 elements , we let @xmath172 be the @xmath171-tuple with the component indexed by @xmath173 equal to the value of @xmath174 evaluated at @xmath173 . the vector @xmath172 is called a _ codeword_.
we define the @xmath51-dimensional rs code on @xmath170 over @xmath63 as the set of all codewords @xmath175 the encoding function maps a polynomial of degree strictly less than @xmath51 , with coefficient in @xmath63 , to @xmath172 .
the encoding map is denoted by @xmath176 . for any @xmath177
, if we know the value of @xmath174 evaluated at any @xmath51 elements in @xmath170 , then we can recover @xmath174 by interpolation , and thus the coefficients of @xmath174 can be uniquely determined .
in fact , given @xmath51 points @xmath178 where @xmath179 are distinct for @xmath180 , we can obtain the coefficients of @xmath181 by solving @xmath182 the matrix on the left hand side is a vandermonde matrix , and is invertible because @xmath183 , @xmath184 are distinct . to facilitate the discussion on the encoding and decoding procedure , we make the following definitions .
for @xmath185 , let @xmath186 be the set of sources associated with a node to the left of node @xmath13 and demanded by a node to the right of node @xmath13 , @xmath187 also , for @xmath185 , let @xmath188 be the set of sources associated with a node to the right of node @xmath13 and demanded by a node to the left of node @xmath13 , @xmath189 we set @xmath190 .
the superscripts `` @xmath191 '' and `` @xmath192 '' stand for `` forward '' and `` backward '' respectively .
we note that @xmath186 and @xmath188 may be empty . for @xmath185 ,
let @xmath193 be the data rate through node @xmath13 from left to right , and @xmath194 be the data rate through node @xmath13 from right to left . to illustrate the notation
, we consider the four - node network in example 1 .
the data from source 1 passes through nodes 2 and 3 in the forward direction , and the data from source 2 passes through nodes 2 and 3 in the backward direction .
we have @xmath195 , @xmath196 , @xmath197 , @xmath198 and @xmath199 .
the encoding scheme is a decode - and - forward scheme . to simplify the description , we assume that all delay offsets @xmath89 are equal to zero , i.e.
, all protocol sequences are aligned , and consider the operations at node @xmath13 . in a period of @xmath87 packet durations ,
@xmath200 source symbols which are going to be sent through node @xmath13 in the forward direction are generated .
call these symbols @xmath201 likewise , let @xmath202 be source symbols to be sent to node @xmath33 through node @xmath13 in the backward direction .
suppose that @xmath203 , i.e. , source @xmath204 is associated with node @xmath13 . in a time period of @xmath205 seconds
, source @xmath204 produces @xmath206 symbols @xmath207 the encoder at node @xmath13 maps these @xmath208 symbols to @xmath209 symbols and send them out according to a protocol sequence with duty factor @xmath113 .
define the following three polynomials , @xmath210 the degrees of polynomials @xmath211 , @xmath212 and @xmath213 are no more than @xmath214 , @xmath215 and @xmath216 , respectively .
let @xmath217 be the set of the first @xmath164 elements in @xmath63 .
we transmit the following frame of packets @xmath218 in a period . here , the addition and multiplication are polynomial arithmetics over the finite field @xmath63 .
( -.3,-.5)(6,1 ) ( 0,0)n1@xmath219 ( 1.5,0)n2@xmath33 ( 3,0)n3@xmath220 ( 4.5,0)n4@xmath34 ( 6,0)n5@xmath101 ( 3,-.7)@xmath221 ( 2.25,0.4)@xmath222 ( 3.75,0.4)@xmath223 ( 5.25,0.4)@xmath224 ( .75,-0.4)@xmath225 ( 2.25,-0.4)@xmath226 ( 3.75,-0.4)@xmath227 suppose source @xmath228 is attached to node @xmath33 and produces symbols at a rate of @xmath229 symbols per packet duration . at the end of a period , node @xmath13 receives from node @xmath33 the following frame of packets @xmath230 with some of the packets erased due to collision .
the degrees of @xmath231 , @xmath232 and @xmath233 are no more than @xmath234 , @xmath235 and @xmath236 , respectively .
the coefficients of @xmath232 are the symbols sent from node @xmath33 to node @xmath13 , and the coefficients of @xmath233 are the symbols sent from node @xmath33 to node @xmath219 .
the coefficients of @xmath231 are the symbols from source @xmath228 .
since @xmath233 is known to node @xmath13 , the decoder of node @xmath13 can subtract @xmath237 from , and obtain @xmath238 with some components missing due to erasures . here
, the subtraction is done using arithmetics in finite field @xmath63 .
the vector in is an rs codeword , corresponding to a polynomial of degree no more than @xmath239 .
provided that the number of non - collided packets is larger than or equal to @xmath240 , then node @xmath13 can recover @xmath231 and @xmath232 .
suppose there is no source associated with node @xmath34 .
then the frame of packets transmitted by node @xmath34 is in the form of @xmath241 the coefficients of @xmath242 are the symbols sent from node @xmath34 to node @xmath101 , and the coefficients of @xmath243 are the symbols sent from node @xmath34 to node @xmath13 .
since @xmath242 is known to node @xmath13 , the decoder of node @xmath13 subtracts @xmath244 from , and obtain an erased version of @xmath245 . note that @xmath246 is a polynomial of degree no more than @xmath247 .
provided that node @xmath13 receives at least @xmath248 non - collided packets from node @xmath34 , then the decoder of node @xmath13 can recover @xmath243 .
( see fig .
[ fig : encoding ] . ) after the coefficients of @xmath231 , @xmath232 and @xmath243 are recovered , the encoder of node @xmath13 produces the frame of packets for the next period , and the process continues .
if the delay offsets are not zero , then the nodes need to buffer the decoded symbols for one period . with extra delay due to the buffering
, the joint channel - network coding scheme works in a similar way as in the case with zero delay offset .
we characterize the rate region achieved by this transmission scheme in the following theorem . for @xmath49 ,
let @xmath186 and @xmath188 be defined as in and .
a rate vector @xmath249 is achievable in a slot - synchronous tandem collision network if for @xmath49 , @xmath250 when node @xmath13 is a source node associated with source @xmath204 , and @xmath251 when node @xmath13 is not a source node , for some non - negative real numbers @xmath252 between 0 and 1 .
[ thm : synchronous ] ( in theorem [ thm : synchronous ] , @xmath253 , @xmath254 , @xmath255 and @xmath256 are set to zero . )
let @xmath257 be a rate vector that satisfies to for some real numbers @xmath258 , @xmath259 between 0 and 1 . for @xmath49 , we approximate @xmath107 by rational @xmath260 , where @xmath114 is a common denominator .
construct @xmath0 consecutively 3-wise shift - invariant protocol sequences by the construction described at the beginning of this section , with common period @xmath115 and duty factor @xmath261 , @xmath49 .
consider the link from node @xmath13 to node @xmath34 .
node @xmath34 use the initialization mechanism described in paragraph ( c ) earlier in the section to determine the delay offset pertaining to node @xmath13 .
this mechanism can always find the delay offset because the protocol sequences are consecutively 3-wise shift - invariant , and hence identifiable .
after the delay offset of node @xmath13 is known , node @xmath34 can determine the time indices of the packets sent from node @xmath13 . by lemma [ lemma : si ] , node @xmath34 always receives @xmath262 successful packets per slot duration from node @xmath13 .
we recall that by the shift - invariant property , the number of successful packets from node @xmath13 to node @xmath34 in a period of @xmath115 slot times is a constant .
if node @xmath13 is a source node , with source @xmath204 attached to it , the number of information packets from node @xmath13 to node @xmath34 in a period of @xmath115 time slots is @xmath263 .
provided that condition @xmath264 is satisfied , the channel - network coding scheme described earlier is able to recover the information packets with zero decoding error . on other hand
, if node @xmath13 is not a source node , the joint channel - network coding scheme can decode the messages with zero error provided that @xmath265 the rate constraint for the link from node @xmath13 to node @xmath33 is either @xmath266 or @xmath267 depending on whether node @xmath13 is a source node or not . since @xmath261 is an approximation to @xmath107
, we can find a rate vector @xmath268 close to @xmath269 , which satisfies to .
the deviation of @xmath270 from @xmath271 depends on the differences between @xmath261 and @xmath107 , @xmath49 .
as we take @xmath261 approaching @xmath107 for @xmath272 , the corresponding rate vector @xmath268 approaches @xmath269 .
hence @xmath269 is a limit point of a sequence of achievable rate vectors .
this proves that @xmath269 lies in the achievable rate region .
the coding scheme described above can be modified as in @xcite and operates in slot - asynchronous system with a slight loss of data rates .
the idea is to replace each zero in a protocol sequence by @xmath273 consecutive zeros , and each one by @xmath274 consecutive ones followed by a single zero .
all duty factors are then multiplied by a factor @xmath275 after this process .
the codewords from the joint channel - network coding scheme are interleaved @xmath274 times .
for example , if @xmath276 , the two protocol sequences @xmath277 $ ] and @xmath278 $ ] are mapped to two sequences of length 12 , @xmath279 \\ & [ 1\ 1\ 0\ 1\ 1\ 0\ 0\ 0\ 0\ 0\ 0\ 0\ 0].\end{aligned}\ ] ] let @xmath280 be a rate vector which is achievable when the system is slot - synchronous .
it is shown in ( * ? ? ?
* lemma 5 ) that the resulting transmission scheme is error - free in the slot - asynchronous case with rate @xmath281 .
we refer the readers to @xcite for the details of argument .
since @xmath273 can be arbitrarily large , we conclude that the rate vector @xmath282 is achievable in the slot - asynchronous case .
we have thus proved the following . any rate vector that satisfies the conditions in theorem [ thm : synchronous ]
is also achievable in the slot - asynchronous case .
[ thm : asyn ] theorem [ thm : asyn ] says that there is essentially no loss of achievable data rates when we compare slot - synchronous and slot - asynchronous system .
nevertheless , if we approach the boundary of the achievable rate region by increasing the value of @xmath273 , decoding delay also increases .
in this section , we derive an outer bound on the achievable rate region . in fact , we will give an outer bound on achievable rate vectors for slot - synchronous system .
this also yields an outer bound for slot - asynchronous system .
we recall that a slot - synchronous system is not necessarily a time - slotted system . in the derivation of the outer bound
, we do _ not _ assume that the system is time - slotted . for @xmath283 ,
let @xmath284 be the set of sources associated with node @xmath13 or a node to the left of node @xmath13 , and demanded by a node to the right of node @xmath13 , @xmath285 for @xmath286 , let @xmath287 be the set of sources associated with node @xmath13 or a node to the right of node @xmath13 , and demanded by a node to the left of node @xmath13 , @xmath288 @xmath289 and @xmath290 are defined as the empty set .
in contrast to the definition of @xmath291 and @xmath292 in and , we have `` @xmath293 '' and `` @xmath294 '' in and instead of strict inequality in and ; both @xmath284 and @xmath295 include the source associated with node @xmath13 .
it is easy to see that @xmath296 and @xmath297 .
given source mapping @xmath298 and receiver mapping @xmath9 in a tandem collision network with @xmath0 nodes and @xmath3 sources , let @xmath299 with the union taken over all real numbers @xmath300 , @xmath49 .
( the variables @xmath253 , @xmath254 , @xmath255 , @xmath256 are taken to be zero . )
the main result in this section is the following .
all achievable rate vectors for slot - synchronous tandem collision network are contained in @xmath301 .
[ thm : converse ] for a _ bi - directional _ tandem collision network , if node @xmath13 is a source node associated with source @xmath204 , then @xmath302 and @xmath303 for @xmath49 .
if node @xmath13 is not a source node , then @xmath304 and @xmath305 . comparing the outer bound and and the rate constraints to in theorem [ thm : synchronous ] , we see that the outer bound @xmath301 coincides with the achievable rate region for slot - synchronous and slot - asynchronous system described in the previous section . we have thus found the capacity region for bi - directional tandem collision network . for bi - directional tandem collision network , the capacity region in the slot - asynchronous and slot - synchronous case is equal to @xmath301 .
[ thm : capacity ] the proof roughly follows the same line as in ( * ? ? ?
* section iv ) .
we assume that for all @xmath13 , the period @xmath39 of protocol signal @xmath38 is a rational multiple of the duration of a packet , i.e. , @xmath306 for some integers @xmath307 and @xmath308 , where @xmath30 denotes packet duration .
there is no loss of generality in this assumption , because we can approximate any real number by rational numbers with arbitrarily small error .
let @xmath309 be the least common multiple of @xmath310 .
then @xmath311 is a common period of all protocol signals , because @xmath312 for all @xmath40 and for all @xmath13 . in order to simplify notation ,
it is convenient to define @xmath313 for all @xmath40 . in the slot - synchronous system model , the delay offsets @xmath314
are fixed integral multiples of @xmath30 in the channel model . for the purpose of our proof
, we impose a fictitious probability distribution on the delay offsets , and assume that @xmath314 are independent random variables , uniformly distributed over @xmath315 .
suppose the duty factor of @xmath38 , defined in , is equal to @xmath107 , for @xmath49 .
we want to prove the following claim : for all @xmath40 , we have @xmath316 = f_i , \label{eq : mean_of_s}\ ] ] with the expectation taken over @xmath43 .
we note that the left hand side of the above equation is a discrete sum , while the right hand side is an integration .
the definition of @xmath107 implies that within a period of @xmath311 , there are @xmath317 non - overlapping semi - open intervals of length @xmath30 , in which @xmath38 is equal to 1 .
let @xmath35 be a fixed real number between 0 and @xmath311 , and consider the following set of @xmath309 time instants @xmath318 with the subtraction performed modulo @xmath311 .
this is a set of @xmath309 evenly spaced points in @xmath319 . because the length of each semi - open interval is @xmath30 , each of these @xmath320 intervals contains exactly one time instant in
if we evaluate @xmath38 at the time instants in , exactly @xmath321 of them equal one . at the remaining time instants ,
the values of @xmath38 are zero .
thus , we get @xmath322 hence , @xmath323 = \frac{1}{c}\sum_{k=0}^{c-1 } s_i(t_0 - k t ) = f_i.\ ] ] this completes the proof of the claim .
consider an arbitrary semi - open time interval @xmath324 , for some constant @xmath35 . within this semi - open time interval
, packets can be sent from node @xmath13 to node @xmath34 only when node @xmath13 is in transmit mode , node @xmath34 is in receive mode , and node @xmath101 is not transmitting anything .
let @xmath325 be a subset of @xmath324 containing time instants that satisfy @xmath326 or equivalently , @xmath327 the set @xmath325 is the union of some non - overlapping sub - intervals of @xmath324 .
we define the length of @xmath325 as the summation of the length of the constituent sub - intervals .
every non - collided packet from node @xmath13 to node @xmath34 must be transmitted totally within @xmath325 .
otherwise , it partially overlaps with other packets and is lost due to collision .
the total number of non - collided packets from node @xmath13 to node @xmath34 is thus no larger than the length of @xmath325 divided by @xmath30 .
we remark that the number of non - collided packets may be strictly less than the length of @xmath325 divided by @xmath30 , because we only assume slot - synchronous system , which may not be time - slotted . by the independence of @xmath43 , @xmath328 and @xmath329 , the expected value of the left hand side of , over random variables @xmath43 , @xmath328 and @xmath329 , is @xmath330 \\ & = e \big [ s_i(t-\delta_i)\big ] e\big[(1 - s_{i+1}(t -\delta_{i+1}))\big ] \\ & \qquad \cdot e\big [ ( 1 - s_{i+2}(t -\delta_{i+2 } ) ) \big ] \\ & = f_i(1-f_{i+1 } ) ( 1-f_{i+2}).\end{aligned}\ ] ] the last equality follows from .
let @xmath331 be the indicator function , @xmath332 the expected length of @xmath325 , taken over random variables @xmath43 , @xmath328 and @xmath329 , is thus @xmath333 & = e\big [ \int_{t_0}^{t_0 + ct } \mathbb{i}(t \in \mathcal{t}_{i , i+1 } ) \ , dt \big]\\ & = \int_{t_0}^{t_0 + ct } e\big [ \mathbb{i}(t \in \mathcal{t}_{i , i+1 } ) \big]\ , dt \\ & = \int_{t_0}^{t_0 + ct } f_i(1-f_{i+1 } ) ( 1-f_{i+2 } ) \ , dt \\ & = ct f_i(1-f_{i+1 } ) ( 1-f_{i+2}).\end{aligned}\ ] ] we can find some realization of the random variables @xmath43 , @xmath328 and @xmath329 such that the length of @xmath325 is less than or equal to the expected value @xmath334 .
therefore , there are some specific values of @xmath43 , @xmath328 and @xmath329 such that the number of non - collided packets from node @xmath13 to node @xmath34 in a duration of @xmath311 seconds is no more than @xmath335 .
suppose that there is a friendly genie who in advance informs nodes @xmath13 and @xmath34 which packets will be collided , and which packets will be received successfully . consider a cut of the network by the edge @xmath336 .
the decoding of the sources in @xmath337 by nodes @xmath34 , @xmath338 is a function of the packets from node @xmath13 to node @xmath34 . with the help of the genie
, the link from node @xmath13 to node @xmath34 reduces to a noiseless discrete memoryless channel with packet rate less than or equal to @xmath339 packets per slot duration . by the max - flow bound for multicast network codes (
* chapter10 ) ( * ? ? ?
* chapter 21 ) , we deduce that if @xmath249 is achievable no matter what the delay offsets are , then @xmath340 a rate vector is not achievable with the help of genie is certainly not achievable in the presence of genie .
this proves the rate constraint in .
the derivation of for traffic in the backward direction is similar as above , with @xmath34 and @xmath101 replaced by @xmath33 and @xmath219 , respectively .
the above argument holds for any fixed duty factors .
the outer bound follows by taking the union over all duty factors .
we compare the capacity region with three random access schemes .
the first two schemes do not have the network coding feature , while the third one does .
in all these three schemes , each node maintains two queues , one for incoming packets from the left and one for packets from the right .
we consider the heavy traffic scenario and assume that the queues are saturated for simplified analysis in the first scheme , each node sends packets in the fashion of pure aloha , and we will call this _ uncoded pure aloha _
transmission of the nodes are independent from each other .
we adopt a simplifying assumption that the transmission times of packets of node @xmath13 follow a poisson process with intensity @xmath341 packets per packet time .
we assume that each packet has a header which stores the identity of the sender and packet numbers .
packet number of collided packets are piggybacked to the transmitting node for re - transmission .
the overhead due to packet header is neglected .
consider a source node , say node @xmath13 , that is associated with source @xmath204 .
when a transmission is initiated , node @xmath13 transmits a source packet with probability @xmath342 , a packet to be relayed to the left with probability @xmath343 and a packet to be relayed to the right with probability @xmath344 , where @xmath342 , @xmath343 and @xmath344 are non - negative real numbers such that @xmath345 .
we consider the transmission of a source packet successful if it is received successfully by both node @xmath33 and node @xmath34 .
when a packet is transmitted by node @xmath13 at time @xmath35 , it will be received by node @xmath33 and @xmath34 if nodes @xmath219 , @xmath33 , @xmath34 and @xmath101 do not transmit any packet in the time interval @xmath346 $ ] of length @xmath347 .
this leads to the following rate constraint , @xmath348 ( we define @xmath349 , @xmath350 , @xmath351 and @xmath352 to be zero . ) in order to forward packets to nodes @xmath34 and @xmath33 with rates @xmath353 and @xmath354 respectively , it is required that @xmath355 for a node which is not a source node , the rate requirement is the same as and , with @xmath356 . in the second scheme , called _
uncoded slotted aloha _ , slot - synchronization is assumed .
node @xmath13 transmits a packet in a time slot with probability @xmath107 .
the protocol is similar to the uncoded pure aloha scheme . for a source @xmath204 which is associated with node @xmath13
, we have the following rate requirement , @xmath357 the rate constraints for forward and backward traffic at node @xmath13 are @xmath358 ( we define @xmath253 , @xmath254 , @xmath255 and @xmath256 as zero . ) for a non - source node @xmath13 , we have two rate constraints as in and with @xmath359 . the third scheme , which is described in @xcite , is a random access scheme with network coding .
it is assumed that the system is time - slotted and slot - synchronous , and node @xmath13 transmits a packet in a time slot with some fixed probability @xmath107 . when a transmission is initiated at node @xmath13 , with probability
@xmath342 a source packet is transmitted , and with probability @xmath360 the xor of two packets from opposite direction to be relayed through node @xmath13 is transmitted . if source @xmath204 is associated with node @xmath13 , we have @xmath361 and @xmath362 for non - source node @xmath13 , the rate constraints are and with @xmath363 set to zero .
we call the third scheme _
network - coded slotted aloha_. by theorem [ thm : capacity ] , the capacity region for the two - way tandem network in example 1 consists of the rate pairs that satisfy @xmath364 for some duty factors @xmath365 $ ] .
( @xmath366 and @xmath367 in and are set to 0 . ) the inequality in is deduced from the requirement that @xmath368 is less than all rate constraints in the forward direction , and corresponds to the backward direction .
the achievable rate region is plotted in fig .
[ fig : graph ] .
we can verify that the point corresponding to @xmath369 and @xmath370 is achievable by putting @xmath371 , @xmath372 , @xmath373 and @xmath374 in , @xmath375 by setting @xmath376 , we can check that maximal symmetric rate is @xmath377 , @xmath378 thus the point @xmath379 is the maximal symmetric rate pair in the achievable rate region .
the achievable rate region by the network - coded slotted aloha scheme is _ identical _ to the capacity region defined by and .
this feature holds for all two - way tandem networks , i.e. , only the two end nodes are source nodes , while the nodes in the middle are non - source nodes .
we can also observe that all rate regions are not convex .
this is because time - sharing requires inter - node coordination in the multiple - access layer , which is not available in our setting . the achievable rate regions for uncoded pure aloha and uncoded slotted aloha are plotted as the dashed lines in fig . [ fig : graph ] .
the symmetric rate achieved by uncoded pure aloha and uncoded slotted aloha are 0.0678 and 0.1058 , respectively , which are much smaller than the maximal symmetric rate 0.1481 achieved by the scheme proposed in this paper and by network - coded slotted aloha .
we observe from fig . [
fig : graph ] that the two extreme points @xmath380 and @xmath381 are achievable by all schemes described in this paper except the uncoded pure aloha scheme . in this example , the rate region achieved by network - coded slotted aloha is strictly smaller than the capacity region .
by theorem [ thm : capacity ] , the capacity region consists of rate pairs @xmath382 that satisfy @xmath383 for some duty factors @xmath384 $ ] .
it is easy to see that we can set @xmath385 in the above inequalities without affecting the result .
the set of rate constraints that defines the capacity region reduces to @xmath386 where @xmath387 , @xmath388 and @xmath389 are real numbers between zero and one .
the capacity region is plotted in fig .
[ fig : graph2 ] .
the maximal symmetric rate is 0.1716 .
the reason why the network - coded slotted aloha has smaller throughput than our proposed scheme is that the source packets and on - going traffic are treated separately .
nested coding using rs code is capable of encoding packets in an effective way . as in the previous example , we observe that the two extreme points @xmath380 and @xmath381 are achievable by all schemes except the uncoded pure aloha scheme .
when two sources are active , uncoded pure aloha , uncoded slotted aloha and network - coded slotted aloha are all suboptimal . unlike the previous example , the network - coded slotted aloha scheme is suboptimal in this cases .
the main functionality difference between the network - coded slotted aloha scheme and the capacity - achieving scheme is that there is no channel coding in the network - coded slotted aloha scheme .
thus , we see that nested coding is essential in improving throughput .
in this paper we present a transmission scheme for tandem collision network .
the construction involves three different ideas : shift - invariant protocol sequences for multiple - access , nested coding for erasure correction , and network coding for bi - directional information flow .
the resulting transmission system does not require any coordination or synchronization among the nodes , and yet is able to achieve optimal throughput in bi - directional tandem collision network . for wireless sensor networks , achieving time synchronization
is costly .
the proposed scheme can alleviate this problem and operate slot - asynchronously .
numerical examples show that the protocol - signal approach can achieve higher data rate than random - access transmission schemes .
a key feature of the proposed transmission that makes it superior to other random - access schemes is erasure - correcting coding .
the regularity of channel accessing pattern dictated by protocol sequences make it easy to incorporate channel coding .
this feature is not available if the channel accessing is stochastic .
suppose that the protocol sequences are consecutively 3-wise shift - invariant .
also , suppose @xmath390 and @xmath391 are chosen such that the associated channel activity signal @xmath392 $ ] is the same as @xmath393 $ ] for all @xmath51 .
the number of packets declared to be sent from node @xmath33 is @xmath394 , because the number of time indices @xmath51 such that @xmath395 = s_{i-1}[k-\tau'_{i-1 } ] = 1\ ] ] is equal to the throughput @xmath396 , which is a constant by the shift - invariant property .
similarly , the number of packets declared to be sent from nodes @xmath34 is equal to @xmath397 . also , the number of symbol `` * '' in the channel activity signal @xmath392 $ ] is equals @xmath398 .
\(b ) @xmath399 and @xmath400 .
if @xmath392 = 1 $ ] for some @xmath51 , then by construction , either @xmath401 = 1 $ ] or @xmath402 = 1 $ ] the packet at time @xmath403 is decided to be sent from node @xmath33 if @xmath404 = s_{i-1}[\tilde{k}+\tau_{i-1 } ] = 1 .
\label{eq : identify1}\ ] ] thus the decisions for the packets that are declared to be from node @xmath33 are all correct .
we can find exactly @xmath394 such @xmath403 that satisfies by the shift - invariant property .
the remaining @xmath397 successfully received packets can not be sent from node @xmath33 , and so must be sent from node @xmath34 .
incidently , they are all declared to be sent from node @xmath34 , and hence there is no identification error .
\(d ) @xmath405 and @xmath400 .
if the output of the algorithm was incorrect , then a packet from node @xmath33 would be mistakenly decided to be from node @xmath34 , and a packet from node @xmath34 would be mistakenly decided from node @xmath33 . in this case , the number of symbol `` * '' contained in the channel activity signal when the delay offsets of nodes @xmath33 and @xmath34 were @xmath104 and @xmath140 , would be strictly larger than @xmath398 .
this contradicts the assumption that the protocol sequences are shift - invariant .
k. w. shum , c. s. chen , c. w. sung , and w. s. wong , `` shift - invariant protocol sequences for collision channel without feedback , '' _ ieee trans .
inform . theory _
55 , no . 7 , pp . 33123322 , jul . | in aloha - type packetized network , the transmission times of packets follow a stochastic process . in this paper , we advocate a deterministic approach for channel multiple - access .
each user is statically assigned a periodic protocol signal , which takes value either zero or one , and transmit packets whenever the value of the protocol signal is equal to one . on top of this multiple - access protocol , efficient channel coding and network coding schemes are devised .
we illustrate the idea by constructing a transmission scheme for the tandem collision network , for both slot - synchronous and slot - asynchronous systems .
this cross - layer approach is able to achieve the capacity region when the network is bi - directional .
collision channel , protocol sequence , tandem network , bi - directional network , network coding . |
the recent experimental demonstration @xcite of the transition from a superfluid state to a mott insulating state of atoms in an optical lattice has opened the way to novel realizations of effective quantum lattice models with widely tunable control parameters .
quantum magnetic systems can be realized by spinor atoms in an optical lattice , e.g. @xmath3na with a total @xmath2 moment .
confining @xmath2 atoms to an optical lattice there are two scattering channels for identical atoms with total spin @xmath4 which can be mapped to an effective bilinear and biquadratic spin interaction @xcite : h=_i , j , [ eqn : biquadhamiltonian ] where we adopt the standard parametrization @xmath5 and @xmath6 . in one dimension , the bilinear - biquadratic spin - one model has a rich phase diagram ( see fig . [
fig : phasediagram ] ) with some well established phases : the haldane gap phase @xcite , a dimerized phase @xcite , a ferromagnetic phase and some less well understood phases : a critical phase with period three correlations @xcite , which we will characterize as having dominant spin nematic correlations , and possibly a gapped spin nematic phase between the dimerized and ferromagnetic phase @xcite , which however remains controversial . on the square and the simple cubic lattice , the bilinear - biquadratic spin - one model is well understood for the case @xmath7 .
it exhibits a ferroquadrupolar spin nematic phase for @xmath8 @xcite .
adjacent to it are an antiferromagnetic nel phase and a ferromagnetic phase @xcite .
the region of purely antiferromagnetic couplings @xmath9 remains to be understood , and could possibly contain a spin liquid region @xcite .
recently the hamiltonian ( [ eqn : biquadhamiltonian ] ) on the triangular lattice attracted some interest @xcite , as a possible explanation of the unconventional magnetism of niga@xmath10s@xmath11 @xcite . by having dominant @xmath12 spin quadrupolar correlations .
the possible occurence of a spin nematic like phase close to @xmath13 is investigated and critically discussed .
[ fig : phasediagram ] ] for a single chain chubukov @xcite suggested the existence of a gapped , nondimerized phase showing dominant spin nematic correlations close to the ferromagnetic region of the phase diagram .
subsequent numerical work @xcite could however not substantiate this claim and it was therefore believed for a while that the dimerized phase would extend up to the ferromagnetic phase boundary .
recent quantum monte carlo calculations @xcite and field theoretical work @xcite suggest that this picture might need to be reconsidered , especially in the light of possible experimental verifications in bose - einstein condensate systems @xcite . in the meantime
there has been a considerable number of publications in favor or against a spin nematic phase close to @xmath14 @xcite , but leaving the final answer still open .
the aim of the present paper is twofold .
first we give a characterization in terms of the dominant correlation function for the extended gapless critical phase , whose mere existence is well established .
we find in this case that the predominant correlations are of spin quadrupolar nature , with a wavevector of @xmath15 .
the second aim is to shed some more light on the region @xmath16 $ ] of the single chain .
we present extensive numerical simulation results based on different numerical methods and discuss possible interpretations . due to the difficulty in obtaining a firm conclusion regarding the spin nematic phase , we have also studied bilinear - biquadratic ladders . in this case
the dimerized phase plays only a minor role in the phase diagram and we report an extended gapped spin nematic phase , very close in spirit to the proposal by chubukov @xcite for the single chain .
surprisingly this phase seems to be adiabatically connected to the well known haldane phase of the isolated chain .
the outline of the paper is the following : in section [ sec : period3 ] we show that the established gapless phase @xmath17 has dominating spin quadrupolar correlations , a characterization which was lacking before .
we then move on to the region @xmath16 $ ] and discuss the possible existence of a spin nematic intermediate phase between the dimerized and the ferromagnetic phase .
we report several anomalous physical properties encountered upon approaching @xmath18 , which could be interpreted as a phase transition to a spin nematic state .
however , due to numerical limitations it has remained elusive to pinpoint such a phase transition , and an alternative scenario where a a very large crossover scale emerges as one approaches the @xmath19 point at @xmath13 becomes more plausible . in section
[ sec : twochains ] we introduce an extension of the single chain model by coupling two chains to form a ladder .
we show that the dimerized phase gives very rapidly way to an extended gapped phase .
this short - ranged ordered phase encompasses the haldane phase at @xmath20 as well as its ladder extension , the so called `` plaquette singlet solid '' state@xcite , and crosses over to a gapped spin nematic state close to @xmath18 .
the existence of an extended critical phase in the interval @xmath21 has first been discussed in ref . [ ] .
later work in refs .
[ ] agreed on the gapless nature of the phase due to soft modes at @xmath22 .
the special point @xmath23 has an enlarged @xmath19 symmetry and is solvable by bethe ansatz @xcite ( uimin - lai - sutherland model ) , proving the existence of soft modes at @xmath22 as a rigourous result . despite the consensus on the presence of an extended gapless phase ,
the nature of the dominant correlations in this phase has not been clearly worked out .
initally xiang @xcite proposed an almost `` trimerized '' groundstate , with dominant singlet correlations involving three consecutive spins . although models can be constructed which have exactly trimerized groudstates @xcite , in the present model these are not the dominant fluctuations , as shown earlier in ref .
[ ] . here
we show that the dominant correlations are not of singlet , but of _ spin nematic _ ,
i.e. quadrupolar character .
we build upon the field theoretical work of itoi and kato @xcite in which they showed that throughout the critical region there are three fields with spin @xmath24 whose scaling dimensions are all equal to @xmath25 .
however apart from @xmath26 it is the @xmath27 mode at @xmath12 which is subject to correlation-_enhancing _ logarithmic corrections which will show the dominant correlations .
note that this is consistent with the observation @xcite that the lowest energy level at @xmath12 on finite chains carries @xmath27 . as a function of @xmath28 for different system sizes .
at the su(3 ) point @xmath23 the two correlation functions are related by symmetry . by going deeper into the critical region the spin correlations get weaker , while the quadrupolar correlations are enhanced .
lower plot : the central charge , the excitation velocity and the scaling dimension of the quintuplet field in the critical region .
the filled red square is the exact result for the velocity at @xmath26 .
[ fig : trimerizedcorrs ] ] as a function of @xmath28 for different system sizes . at the su(3 )
point @xmath23 the two correlation functions are related by symmetry . by going deeper into the critical region
the spin correlations get weaker , while the quadrupolar correlations are enhanced .
lower plot : the central charge , the excitation velocity and the scaling dimension of the quintuplet field in the critical region .
the filled red square is the exact result for the velocity at @xmath26 .
[ fig : trimerizedcorrs ] ] in order to work out the correlation content of this @xmath27 mode we note that this mode will only show up in a dynamical correlation function if it is targeted with an @xmath27 operator .
a natural operator is the irreducible quadrupolar part of the rank two tensor @xmath29 . in order to test this conjecture ,
we calculate a component of the static spin _
quadrupolar _ structure factor ( k)= ( s^z)^2(-k ) ( s^z)^2(k ) and similar for the static spin structure factor ( k)= s^z(-k ) s^z(k ) both at the wavevector @xmath30 .
the results shown in the upper plot of fig . [ fig : trimerizedcorrs ] for system sizes up to @xmath31 display nicely that the quadrupolar correlations indeed become more important than the spin - spin correlations when one moves beyond the @xmath19 symmetric point at @xmath26 .
although all correlation exponents ( dimer , spin , quadrupolar ) are equal to @xmath32 , logarithmic corrections @xcite render the quadrupolar correlations clearly dominant in this phase .
note that our notion of a dominant correlation function in this case is analogous to the @xmath33 antiferromagnetic heisenberg chain , where both staggered spin and staggered dimer correlations have equal correlation exponent @xmath34 , but the logarithmic corrections enhance the spin - spin correlations and then decay more slowly .
hence , in the same spirit as one argues that the dominant correlations of the heisenberg chain are the spin - spin correlations , the dominant correlations of the bilinear - biquadratic spin - one chain for @xmath17 are of spin quadrupolar nature with wavevector @xmath35 . for completeness
we show selected parameters of this critical region in the lower panel of fig .
[ fig : trimerizedcorrs ] .
the results are in good agreement with field theoretical predictions@xcite ( @xmath36 and @xmath37 ) , exact results@xcite [ @xmath38 $ ] , and previous numerical work @xcite .
the numerical estimate of the scaling dimension @xmath39 is slightly smaller than the expected value of @xmath40 , due to the presence of logarithmic corrections . to close this section
we note that the occurrence of dominant period - three quadrupolar correlations on a single chain calls for a fully ordered three sublattice structure when the chains are coupled in an appropriate two - dimensional fashion .
the triangular lattice perfectly fulfills this prerequisite , and indeed recent works @xcite reported analytical and numerical evidence for a three sublattice quadrupolar ordered phase in the region @xmath0 .
after having elucidated the nature of the dominant correlations in the extended critical region beyond the haldane phase , we now turn to the behavior of the bilinear - biquadratic chain between the point where the dimerized groundstate is firmly established ( @xmath41 ) @xcite and the point where the ferromagnetic state takes over ( @xmath14 ) .
let us note that this point @xmath14 is special due to a @xmath19 symmetry .
this symmetry has the consequence that the conventional ferromagnetic multiplet is degenerate with a fully ordered ferroquadrupolar spin nematic state @xcite .
chubukov @xcite suggested the existence of an intermediate phase which was supposed to be completely gapped and to have dominant short - ranged nematic correlations . since then several studies @xcite most of them numerical tried to pin down the existence of an intermediate phase .
presently it seems most likely that the gapped nematic phase in its original form is not realized in the single chain model , mainly because the gap data do not show evidence for a closing and reopening .
nevertheless the spin nematic correlations grow dramatically as one approaches @xmath42
. it could therefore in principle be possible that the dimerized phase gives way to a critical spin nematic phase without a gap @xcite .
we will critically discuss this possibility in the following .
we track the evolution of the energy gap to the first magnetic excitation using density matrix renormalization group ( dmrg ) calculations@xcite on long open chains of up to 512 spins and retaining up to 1000 states .
a phase transition is signaled by the closing of the gap .
the lowest excited state on open chains carries spin two for @xmath43 $ ] @xcite .
our results shown in fig .
[ fig : dmrggaptheta ] are clearly consistent with a finite gap of the @xmath27 excitation for @xmath44 .
however , for @xmath45 the extrapolated gap becomes very small ( of the order of @xmath46 for @xmath47 ) which could suggest the existence of a phase transition around @xmath48 below which the gap is zero .
the data shows no evidence for a reopening of the @xmath27 gap in the interval @xmath49 $ ] , at variance with the initial proposal @xcite . in order to investigate the possible closing of the gap by different means we have calculated strong coupling series expansions@xcite of the @xmath27 single particle gap starting from the dimerized limit up to 10th order in the interdimer coupling @xmath50 for fixed @xmath28 .
a _ direct _ evaluation of the quintuplet gap shows a closing of the gap around @xmath51 which is illustrated in the left panel of fig .
[ fig : serieslambda ] . using dlog pad approximants
we have determined the critical interdimer coupling @xmath52 where the gap closes , see upper right panel of fig .
[ fig : serieslambda ] .
consistent with previous estimates by chubukov @xcite these extrapolations suggest that the gap closes for the uniformly coupled ( @xmath53 ) chain around @xmath54 . gaps as a function of @xmath28 .
the gray region denotes the estimated onset of the critical behavior .
the filled diamond is the exactly known gap result at @xmath41 .
inset : finite size extrapolation of the @xmath27 gaps ( 32 to 512 sites ) .
@xmath55 varies from @xmath56 to @xmath57 with decrements of @xmath58 from top to bottom .
[ fig : dmrggaptheta ] ] gap .
left panel : @xmath27 gap as a function of @xmath28 .
the gap closes at @xmath59 .
upper right panel : critical value @xmath60 where the gap vanishes .
the uniformly coupled chain of interest here corresponds to @xmath53 ( dashed line ) .
lower right panel : critical exponent of the @xmath27 gap opening as a function of the dimerization @xmath61 calculated by various dlog pad approximants .
[ fig : serieslambda ] ] as a further indication for the possible closure of the gap around @xmath51 , we now consider the existence of a generalized bkt phase transition . to determine the critical point we use phenomenological level spectroscopy .
we calculate the level crossing @xmath62 of the lowest singlet excitation at @xmath63 with the lowest spin-2 level at @xmath64 for different system sizes @xmath65 with ed and extrapolate @xmath66 .
the results are shown in fig .
[ fig : criticalgaptheta ] .
the extrapolation is performed with @xmath67 and @xmath68 corrections , fitting only the largest four system sizes .
the extrapolated critical point is @xmath69 , in good agreement with estimates obtained by other techniques @xcite .
the coefficient of the @xmath67 correction is small , and fitting without this term yields slightly larger values @xmath70 . and
the quintuplet ( @xmath27 ) gap at momentum 0 .
the extrapolation yields a critical value @xmath71 in the thermodynamic limit .
[ fig : criticalgaptheta ] ] dimerization in the phase diagram of the spin - one chain has been firmly established by exact results obtained for the @xmath72 point @xcite . it should be noted that although the system is dimerized , it is rather poorly described by a simple product wavefunction of alternating singlets , as can be seen from the small gap and the large correlation length over the whole dimerized phase .
the dimerization operator considered here is : ( k ) _ j e^i k r_j ( _ j_j+1 ) with @xmath63 .
the second kind of correlations expected to be important in this region of the phase diagram @xcite are the _ ferroquadrupolar _ spin fluctuations ( k ) _
j e^i k r_j [ ( s^z_j)^2 - 2/3 ] with @xmath64 .
note that we considered similar fluctuations with @xmath73 in section [ sec : period3 ] . ] based on these fluctuation operators we have calculated the standard static structure factor ( sf ) : ^(k)= || ( k)|0||^2 and the generalized nonlinear susceptibility ( i.e. the real part of the dynamical correlation function at zero frequency ) ^(k)=_0^+ 0|^(-k ) ( k)|0for both kinds of correlations ( @xmath74 ) .
the nonlinear susceptibility quantifies the perturbative response of the system upon explicitly coupling the symmetry breaking operator to the hamiltonian @xcite .
we present ed results concerning the dimerization in the left panel of fig .
[ fig : susceptibilities ] . the sf and the gns both diverge as @xmath75 for @xmath41 , where long range dimer order is established @xcite .
similar behavior is found for @xmath76 .
for @xmath77 however , we find that the sf and the gns both _ decrease _ with system sizes , pointing to a possible absence of spontaneous dimerization in this region .
in the right panel we display the same kind of observables for the @xmath64 ( _ ferroquadrupolar _ ) mode of the quadrupolar correlations . here
the behavior seems different : while there are only short range correlations deep in the dimerized phase , the ferroquadrupolar correlations increase drastically with system sizes for @xmath28 close to @xmath13 .
\a ) @xmath78 chain .
upper panel : dynamical staggered dimer correlation function @xmath78 plotted for a range of values of @xmath28 . there is an important transfer of spectral weight upon lowering @xmath28 from a low energy level collapsing to the groundstate as a function of system size to a level which will converge to the energy @xmath79 at @xmath14 .
lower panel : dynamical ferroquadrupolar correlations @xmath80 for @xmath81 \pi$ ] .
[ fig : dimerdynamics],title="fig : " ] @xmath82 \b ) @xmath80 chain .
upper panel : dynamical staggered dimer correlation function @xmath78 plotted for a range of values of @xmath28 . there is an important transfer of spectral weight upon lowering @xmath28 from a low energy level collapsing to the groundstate as a function of system size to a level which will converge to the energy @xmath79 at @xmath14 .
lower panel : dynamical ferroquadrupolar correlations @xmath80 for @xmath81 \pi$ ] .
[ fig : dimerdynamics ] ] in order to shed more light on the excitations and their nature close to @xmath14 we have calculated in addition the _ dynamical _ dimer and quadrupolar structure factor using a continued fraction technique .
this will allow us to track the evolution of the energy of the important levels and their spectral weight .
the dynamical structure factor is defined as follows : ( k,)=- _ 0^+ 0|^(-k ) ( k)|0 , where we focus again on the dimerization and the ferroquadrupolar correlations . the plots discussed below show the normalized intensity for a given @xmath28 .
the overall weight can be obtained by multiplying with the structure factors displayed in fig .
[ fig : susceptibilities ] the results for the dimerization shown in the upper panel of fig .
[ fig : dimerdynamics ] are very interesting .
they show a strong crossover with a redistribution of spectral weight on the way from @xmath41 to @xmath13 .
while there is a single state ( the lowest energy singlet ) exhausting almost all of the @xmath63 dimer weight around @xmath41 ( as expected in a truly dimer ordered phase ) , the weight in this state entirely fades out and is transfered to a higher energy state at @xmath83 as @xmath84 .
the results for the dynamical ferroquadrupolar correlations do nt show such a crossover but show a steady lowering of the finite size @xmath27 gap upon approaching @xmath13 , accompanied by an accumulation of all the weight in that single state .
this is a manifestation of the continuity of the finite size groundstate wavefunction as one approaches the fully ferroquadrupolar _ ordered _ state at @xmath14 .
we return to the special behavior of the spin - two gap in subsection [ sec : sma ] .
we believe that it is this strong crossover seen in the dynamical dimer correlations which renders numerical calculations very difficult . on finite size systems sufficiently close to @xmath14
the system looks basically like a ferroquadrupolar ordered system , even in one dimension , due to the very strong influence of the su(3 ) point @xmath13 . as a consequence a crossover scale seems to emerge , which rapidly grows close to @xmath13 , and one has to go to huge systems in order for the dimerization to win , i.e. for the lowest singlet at @xmath63 to collapse onto the groundstate , while the lowest @xmath27 exitation at @xmath64 should remain gapped .
this scenario would then imply that the curve in fig .
[ fig : criticalgaptheta ] finally would have to bend down for large systems and touch the @xmath85-axis only at @xmath14 . ) , and the quintuplet ( @xmath64 ) field are shown .
inset : effective central charge .
the value at the boundary of the critical region ( @xmath54 ) is @xmath86 .
[ fig : scalingdims ] ] in order to discriminate between potential conformal theories describing the phase transition and the extended gapless region , we calculate relevant field theory parameters , i.e. the scaling dimensions ( sd ) @xmath87 of several fields and the central charge @xmath88 ( fig .
[ fig : scalingdims ] ) .
the calculation of these parameters has been performed with ed and relies on finite size scaling properties of groundstate and excited states energies @xcite .
leading logarithmic corrections have been taken into account .
the sd of the @xmath63 singlet field ( @xmath89 ) and the @xmath64 quintuplet field ( @xmath39 ) are both close to @xmath90 at the onset of criticality near @xmath91 .
@xmath89 significantly increases , while @xmath39 decreases and seems to approach @xmath92 as @xmath18 , compatible with quadrupolar long range order precisely at @xmath14 @xcite .
our results for the effective central charge are unexpected .
the central charge @xmath88 does not seem to be constant throughout the potentially critical region .
the smallest value is found at the onset ( @xmath93 ) , and then seems to increase monotonously .
while this behavior can not directly be ruled out on field theoretical grounds , it is rather uncommon .
it remains to be understood whether @xmath88 is really continuously increasing or whether we are facing a crossover phenomenon .
the hypothetical critical theory at @xmath94 is however surprisingly well characterized by a level two @xmath95 wess - zumino - witten model given that both the scaling dimensions ( @xmath90 ) and the central charge ( @xmath96 ) are in agreement with such a theory .
some more support for this claim comes from the indirect calculation of @xmath89 within our series expansions .
there the critical exponent @xmath97 of the @xmath27 gap is related to the scaling dimension @xmath89 by @xmath98 . in the lower right panel of fig .
[ fig : serieslambda ] we show the critical exponent calculated by dlog pad approximants to the 10th order series . in the vicinity of @xmath94 the various approximants
reveal only small spreading . within the precision of a 10th order calculation
the exponents comply with a scaling dimension of @xmath99 ( dashed line ) . in the following we present an argument based on the single mode approximation ( sma ) which aims at explaining the anomalously strong suppression of the @xmath27 gap upon approaching @xmath13 . in the single mode approximation one
starts by constructing a trial state upon the application of a structure factor operator on the groundstate : |(k)=(k ) |0 , where ( k ) = _ j e^i k r_j _ j , and @xmath100 is an operator acting on site @xmath101 .
one can now relate the variational energy of the state @xmath102 to a groundstate expectation value : _
k=. note that the energy @xmath103 is a strict upper bound on the gap in the momentum @xmath104 sector .
the power of the sma comes from the fact that it can be used to prove the absence of a gap under certain conditions .
there are two prototypical cases for the absence of a gap : i ) the denominator diverges as a function of systems size , while the numerator diverges more slowly or stays finite .
this is the conventional situation for systems which are critical or exhibit spontaneous symmetry breaking in the thermodynamic limit . in this case
the diverging static structure factors drives the gap to zero , but only for the infinite system .
ii ) the numerator vanishes , while the denominator does not vanish . a sufficient condition for the numerator to vanish is if the structure factor operator @xmath105 commutes with the hamiltonian . if we now choose for the operator @xmath105 the spin quadrupolar structure factor at zero momentum : i.e. @xmath106 $ ] we realize scenario ii ) at the special @xmath19 point @xmath14 . for this specific value
@xmath107 commutes with the hamiltonian and therefore the numerator is zero .
numerically we find that the denominator is different from zero .
this means that there is at least a level with @xmath27 which is degenerate with the groundstate already on finite size samples @xcite .
the reason which makes the numerical analysis for @xmath108 so difficult is that there is a continuity in the wavefunction as a function of @xmath28 when approaching the @xmath19 point from above .
on one hand we know that the commutator @xmath109 $ ] will vanish continuously as we approach @xmath110 , on the other hand we find numerically that the denominator ( the ferroquadrupolar structure factor ) grows rapidly as a function of the accessible system sizes , c.f .
full symbols in the right panel of fig [ fig : susceptibilities ] .
so these two factors cooperate in giving an anomalously small sma energy @xmath103 , which is an upper bound to the measured gap .
the dimerization along the chains is rapidly suppressed by a finite interchain coupling .
the dominating phase is a unique gapped phase which is adiabatically connected to the haldane phase of two isolated heisenberg chains .
[ fig : twochains ] ] we have seen considerable difficulty in the preceding section to actually decide whether the dimerization vanishes before reaching @xmath111 .
one potential way of circumventing the spontaneous dimerization of a single chain is to consider two coupled chains in a ladder geometry .
it seems rather natural to assume that the dimerization of a single chain disappears once the coupling on the rung is suffiently strong .
we will substantiate this claim using numerical simulations below .
the ladder model is also a first step towards a realistic setup of spin-1 bosonic atoms in optical lattices , where a finite interchain coupling can easily be generated .
the ladder hamiltonian is given as follows : h&= & j _
i , n ( _ i , n_i+1,n ) + ( _ i , n _
i+1,n ) ^2 + & + k & _ i ( _ i,1_i,2 ) + ( _ i,1 _ i,2 ) ^2 , [ eqn : biquadladderhamiltonian ] where @xmath112 denotes a spin - one operator at position @xmath113 on chain @xmath114 . in the following we choose @xmath115 and vary @xmath28 as well as @xmath116 , where the limit @xmath117 corresponds to decoupled chains and @xmath118 to decoupled rung dimers . for @xmath20 the hamiltonian ( [ eqn : biquadladderhamiltonian ] ) describes a conventional @xmath2 heisenberg spin ladder @xcite .
interestingly , for this special case todo _
et al . _
@xcite have shown numerically that the finite gap of the single chain does _ not _ close for any @xmath119 , i.e two points on the @xmath119 axis can be reached without any second order quantum phase transition in between .
they also generalized the string order parameter of a single chain to a more complicated nonlocal operator which is non - zero for any finite value of @xmath119 , therefore showing that this phase is also topologically ordered @xcite . in the following we will show that this gapped spin liquid state at @xmath20 extends deeply into the region @xmath120 , and for suffiently large @xmath119 even to @xmath110 . while the groundstate is always protected by a finite gap , the nature of the lowest excitation changes as a function of @xmath28 . for @xmath121
$ ] the lowest excitation carries @xmath2 , while for @xmath16 $ ] it carries @xmath27 . in the absence of dimerization such a state on the ladder where the lowest excitation is of ferroquadrupolar nature ,
is very close in spirit to the long - sought gapped nematic phase initially proposed by chubukov @xcite for the single chain .
the phase diagram of two coupled chains with @xmath122 $ ] and @xmath123 $ ] is shown in fig . [
fig : twochains ] .
we expect the extended gapped phase to continue up to the rung dimer limit @xmath118 , similar to the results for @xmath20 @xcite . .
it can be clearly seen that that the dimerization extrapolates to a finite value for @xmath124 , while it tends to zero for @xmath125 .
[ fig : ladderdimerization ] ] let us now describe the simulation results which lead to the phase diagram presented in fig .
[ fig : twochains ] .
first we discuss the behavior of the dimerization of the single chain upon coupling two chains . to this end
we performed dmrg calculations on ladder systems of sizes up to @xmath126 sites .
we have determined the boundary of the dimerized phase using finite size extrapolations of the remnant dimerization in the middle of a chain @xcite , assuming a columnar dimer arrangement on the two chains .
spin models on the square lattice dimerize this way for @xmath127 @xcite , while our model at @xmath41 corresponds to @xmath128 .
] an example for such finite size data is shown for @xmath41 in fig .
[ fig : ladderdimerization ] .
we expect the dimerization in the middle of the chain to converge to a finite value for @xmath129 in a dimerized phase , and to exhibit an exponential drop upon reaching the correlation length of a disordered phase .
these two distinct behaviors can be seen in fig .
[ fig : ladderdimerization ] for the values @xmath130 and @xmath131 respectively .
this leads us to the conclusion of a critical ratio @xmath132 at @xmath41 , below which the ladder is still dimerized .
the evolution of the values of @xmath133 with @xmath28 is shown in fig .
[ fig : twochains ] .
the dimerized phase is most stable close to @xmath41 , and seems to vanish linearly upon approaching the bt point at @xmath134 , similar to the case of explicit dimerization for a single chain @xcite . on the other side
however the dimerization boundary drops rapidly and becomes very small once @xmath135 .
similar to the single chain case it is very difficult to decide whether the dimerized phase ceases to exist before @xmath13 or not . in the ladder case
however this implies that for a finite but very small interchain coupling the spontaneous dimerization ( if any ) of a single chain vanishes . as a function of @xmath136 $ ] .
the nature of the gap changes at the @xmath19 point @xmath41 , from a @xmath137 state for @xmath138 to a a @xmath139 state for @xmath140 .
the qmc value of the gap at @xmath20 is from ref .
[ fig : laddergap ] ] next we investigate the behavior of the spin gap in the non - dimerized phase .
we performed dmrg calculations for a fixed value of @xmath141 .
the results for the gap to the lowest @xmath142 state are plotted in fig .
[ fig : laddergap ] for system sizes @xmath143 with @xmath144 .
starting at @xmath20 where we find good agreement with the high - precision qmc gap from ref .
[ ] , the gap grows monotonously until reaching the bipartite @xmath19 point at @xmath41 .
there we observe a level crossing , and the gap decreases for @xmath140 .
note that at this special point the total spin quantum number of the lowest level changes . while it is a @xmath2 state for @xmath138 , the lowest levels carries @xmath27 for @xmath140 .
similar to the single chain case discussed in sec .
[ sec : chainspintwogap ] , we detect a very small gap for @xmath28 close to @xmath13 .
this anomalously small @xmath27 gap can again be understood using a single mode approximation argument , along the lines of sec .
[ sec : sma ] , the extension of the argument to the ladder case being straightforward .
we take this dichotomy of the gap as a first indication that the @xmath19 line ( @xmath41 ) constitutes a crossover line inside the extended gapped phase , where the spin - spin correlations dominate for @xmath145 , while we expect dominant short - ranged spin nematic ferroquadrupolar correlations for @xmath146 .
our gap results give evidence that the groundstate for @xmath28 close to @xmath13 adiabatically connects to the `` plaquette singlet solid '' phase , which is itself connected to the haldane phase of the isolated chains .
this would also imply that the the whole phase is characterized by a finite nonlocal order parameter .
it would be interesting to investigate this order in a future work .
to summarize , we have shown that spin nematic correlations in the form of quadrupolar correlations play an important role in the phase diagram of single or coupled @xmath2 bilinear - biquadratic chains .
we first addressed the long - standing question of the nature of the dominant correlations in the gapless period - three phase , where we uncovered the spin quadrupolar correlations at @xmath147 as the leading ones .
we critically discussed the phase diagram close to the @xmath19 point at @xmath14 and by assembling several numerical and analytical results concluded that an unconventional crossover phenomenon is at the heart of the considerable difficulty in settling the issue on the existence or absence of the gapped nematic phase put forward by chubukov @xcite .
finally we studied two coupled bilinear - biquadratic chains in a ladder geometry and found an extended gapped phase which is surprising in two respects : first it realizes a variant of chubukov s nematic phase , i.e. a gapped , non - dimerized phase with dominant spin - nematic correlations for @xmath146 , and second , this phase is adiabatically connected to the standard @xmath2 heisenberg ( @xmath20 ) ladder , and therefore also connected to the haldane phase of single chains , as shown by todo _
et al . _
@xcite .
we would like to thank f. alet , f.d.m .
haldane , k. harada , n. laflorencie , p. lecheminant , f. mila , d. roggenkamp , m. troyer and s. wessel for valuable discussions .
we are grateful to s. todo for providing us his ladder qmc gap data .
the authors acknowledge support from the swiss national science foundation .
part of the numerical calculations have been performed on the `` pliades '' cluster at epfl . c. orzel ,
science * 291 * , 2386 ( 2001 ) ; m. greiner , nature * 415 * , 39 ( 2002 ) .
yip , phys . rev .
90 * , 250402 ( 2003 ) .
a. imambekov , m. lukin , and e. demler , phys .
a * 68 * , 063602 ( 2003 ) .
garcia - ripoll , m.a .
martin - delgado , and j.i .
cirac , phys .
* 93 * , 250405 ( 2004 )
. f.d.m .
haldane , phys . lett . *
93a * , 464 ( 1983 ) ; phys .
lett . * 50 * , 1153 ( 1983 ) .
barber and m.t .
batchelor , phys .
b * 40 * , 4621 ( 1989 ) ; a. klmper , europhys .
* 9 * , 815 ( 1989 ) ; y. xian , phys .
a * 183 * , 437 ( 1993 ) .
g. fth and j. slyom , phys .
b * 44 * , 11836 ( 1991 ) .
chubukov , phys .
b * 43 * , 3337 ( 1991 ) . n. papanicolaou , nucl . phys . *
b305 * , 367 ( 1988 ) .
k. harada and n. kawashima , phys .
b * 65 * , 052403 ( 2002 )
. k. tanaka , a. tanaka and t. idogaki , j. phys .
gen . * 34 * , 8767 ( 2001 ) .
p. li and s .- q .
shen , new journal of physics * 6 * , 160 ( 2004 ) . h. tsunetsugu , and m. arikawa , cond - mat/0512209
. a. luchli , f. mila , and k. penc , cond - mat/0605234 . s. bhattacharjee , v.b .
shenoy , and t. senthil , cond - mat/0605277 .
s. nakatsuji , science * 309 * , 1697 ( 2005 ) .
g. fth and j. slyom , phys .
b * 51 * , 3620 ( 1995 ) .
n. kawashima , prog .
* 145 * , 138 ( 2002 ) .
b.a . ivanov and a.k .
kolezhuk , phys .
b * 68 * , 052401 ( 2003 )
. a. luchli , g. schmid , and s. trebst , cond - mat/0311082 ( unpublished ) .
d. porras , f. verstraete , and j.i .
cirac , phys .
b * 73 * , 014410 ( 2006 ) .
m. rizzi , d. rossini , g. de chiara , s. montangero , and r. fazio , phys rev . lett . * 95 * , 240404 ( 2005 ) .
k. buchta , g. fth , . legeza , and j. slyom , phys . rev .
b * 72 * , 054433 ( 2005 ) .
s. bergkvist , i. mcculloch , and a. rosengren , cond - mat/0606265 .
s. todo , m. matsumoto , c. yasuda , h. takayama , phys .
b * 64 * , 224412 ( 2001 ) .
y. xian , j. phys .
matter * 5 * 7489 , ( 1993 ) .
p. reed , j. phys . a : math .
* 27 * l69 ( 1994 ) .
bursill , j. phys .
a : math . gen . * 28 * 2109 ( 1995 ) . c. itoi and m. kato , phys .
b * 55 * , 8295 ( 1997 ) .
a. schmitt , phys .
b * 58 * , 5498 ( 1998 ) .
uimin , pisma zh .
* 12 * , 332 ( 1970 ) ; jetp lett .
* 12 * , 225 ( 1970 ) .
lai , j. math .
* 15 * , 1675 ( 1974 ) .
b. sutherland , phys . rev .
b * 12 * , 3795 ( 1975 ) .
j. slyom and j. zittartz , europhys .
lett . * 50 * 389 ( 2000 ) .
batista , phys .
b * 65 * , 180402(r ) ( 2002 ) . s.r .
white , phys .
* 69 * , 2863 ( 1992 ) .
gelfand and r.r.p .
singh , adv .
phys . * 49 * , 93 ( 2000 ) .
the generalized nonlinear susceptibility is defined as the second derivative of the groundstate energy with respect to the addition of a symmetry breaking perturbation .
this approach has been discussed in l. capriotti , int .
b * 15 * , 1799 ( 2001 ) .
blte , phys .
* 56 * , 742 ( 1986 ) ; i. affleck , phys .
* 56 * , 746 ( 1986 ) ; j.l .
cardy , j. phys . a : math .
* 19 * , l1093 ( 1986 ) .
d. snchal , phys .
b * 52 * , 15319 ( 1995 ) ; d. allen and d. snchal , phys . rev .
b * 61 * , 12134 ( 2000 ) ; m. sato , phys . rev .
b * 71 * , 024402 ( 2005 ) . s. r. white and i. affleck , phys . rev .
b * 54 * , 9862 ( 1996 ) .
k. harada , n. kawashima , and m. troyer , phys .
* 90 * , 117203 ( 2003 ) .
a. kitazawa and k. nomura , phys .
b * 59 * , 11358 ( 1999 ) . | we present an extensive numerical study of spin quadrupolar correlations in single and coupled bilinear - biquadratic spin one chains , using several methods such as exact diagonalization , density matrix renormalization group and strong coupling series expansions . for the single chain we clarify the dominant correlation function in the enigmatic gapless period - three phase for @xmath0 , which is of spin quadrupolar nature with a period three spatial structure .
then we revisit the open problem of the possible existence of a ferroquadrupolar phase between the dimerized and the ferromagnetic phases .
although an extended critical region is in principle compatible with the numerical results , a scenario with a huge crossover scale is more plausible . finally we study the fate of the dimerized phase upon coupling two chains in a ladder geometry .
the dimerized phase rapidly vanishes and an extended gapped phase takes over .
this gapped phase presumably has dominant short - ranged ferroquadrupolar correlations for @xmath1 and suprisingly seems to be adiabatically connected to the plaquette single solid phase of the heisenberg @xmath2 ladder and therefore also with the haldane phase of isolated chains . |
long s equation [ 1,2,3,4 ] models the flow of stratified incompressible fluid in two dimensions over terrain .
when the base state of the flow ( that is the unperturbed flow field far upstream ) is without shear the numerical solutions ( in the form of steady lee waves ) of this equation over simple topography ( i.e. one hill ) were studied by many authors [ 5 - 13 ] .
the most common approximation in these studies was to set brunt - visl frequency to a constant or a step function over the computational domain .
moreover the values of two physical parameters which appear in this equation were set to zero .
( these parameters control the stratification and dispersive effects of the atmosphere - see @xmath0 . ) in this ( singular ) limit the nonlinear terms and one of the leading second order derivatives in the equation drop out and the equation reduces to that of a linear harmonic oscillator over two dimensional domain .
careful studies [ 8 ] showed that these approximations set strong limitations on the validity of the derived solutions [ 9 ] .
long s equation also provides the theoretical framework for the analysis of experimental data [ 14,15,29 ] under the assumption of shearless base flow .
( an assumption which , in general , is not supported by the data ) . an extensive list of references appears in [ 16,17,18 ] .
an analytic approach to the study of the solutions of this nonlinear equation was initiated recently by the current author [ 19,20,21 ] .
we showed that for a base flow without shear and under rather mild restrictions the nonlinear terms in the equation can be simplified .
we also identified the `` slow variable '' that controls the nonlinear oscillations in this equation . using phase averaging approximation we derived for self similar solutions of this equation a formula for the attenuation of the stream function perturbation with height .
this result is generically related to the presence of the nonlinear terms in long s equation .
the impact that shear has on the generation and amplitude of gravity waves was investigated by us in [ 20 ] .
a new representation of this equation in terms of the atmospheric density was derived in [ 21 ] .
one of the weak aspects of long s equation is related to the fact that the terrain is represented by the shape of the domain and the boundary conditions . as a result
the impact of different terrains on the solution of this equation can only be studied numerically .
furthermore discretization errors which occur in the representation of the terrain render it impractical to consider complex terrain . in part
these errors are due to the scale of the terrain relative to the computational domain . accordingly
only simple topographies which were represented by one hill were considered in the literature . furthermore even for these simple topographies only approximate boundary conditions were applied at the terrain .
( see discussion in @xmath1 ) . with this motivation
it is our objective in this paper to derive a terrain following formulation of long s equation in which the terrain is incorporated as part of the coefficients of the differential equation , and the computational domain is always a rectangle .
this new representation makes it possible to derive new analytic insights about the solution of this equation in some limiting cases .
it will make it easier also to study how the solution varies as a function of the terrain and other parameters that appear in the equation .
the plan of the paper is as follows : sec .
2 presents a short review of long s equation and some aspects of its solutions . in sec .
3 we derive the new formulation of this equation .
sec 4 considers some analytic and geophysical aspects of this new formulation while sec 5 compares its numerical solution over three different terrains .
these simulations are motivated by recent experiments in the alps region to educe properties of gravity waves from experimental data [ 14 ] .
we end up in sec 6 . with a summary and conclusions .
in two dimensions @xmath2 the flow of a steady inviscid and incompressible stratified fluid is modeled by the following equations : @xmath3 @xmath4 @xmath5 @xmath6 where subscripts indicate differentiation with respect to the indicated variable , @xmath7 is the fluid velocity , @xmath8 is its density @xmath9 is the pressure and @xmath10 is the acceleration of gravity .
we can non - dimensionalize these equations by introducing @xmath11 where @xmath12 represents a characteristic horizontal length , and @xmath13 represent respectively the free stream velocity and averaged base density ( i.e. here @xmath14 is a constant ) .
@xmath15 is an averaged value of the brunt - visl frequency @xmath16 where @xmath17 is the base density . in these new variables
eqs ( [ 2.1])-([2.4 ] ) take the following form ( for brevity we drop the bars ) @xmath18 @xmath19 @xmath20 @xmath21 where @xmath22 @xmath23 @xmath24 is the boussinesq parameter [ 13 ] ( this name has nothing to do with the `` boussinesq approximation '' ) which controls stratification effects ( assuming @xmath25 ) and @xmath26 is the long wave parameter which controls dispersive effects ( or the deviation from the hydrostatic approximation ) . in the limit @xmath27 the hydrostatic approximation is fully satisfied , [ 10,11 ] . in view of eq .
( [ 2.7 ] ) we can introduce a stream function @xmath28 so that @xmath29 using this stream function we can rewrite eq .
( [ 2.8 ] ) as @xmath30 where for any two ( smooth ) functions @xmath31 @xmath32 eq . (
[ 2.14 ] ) implies that the functions @xmath33 are dependent on each other and we can express each of them in terms of the other .
thus we can write @xmath28 as @xmath34 ( or @xmath8 as @xmath35 [ 21 ] ) .
after a long algebra one can derive the following equation for @xmath28 [ 22,1,13 ] @xmath36=s(\psi)\ ] ] where @xmath37 is the nondimensional brunt - visl frequency .
we observe that in this definition @xmath38 is a function of @xmath28 .
( as a result it can be an additional source of nonlinearity in eq .
( [ 2.21 ] ) ) this is in contrast to the previous definition of this quantity in eq .
( [ 2.6 ] ) which depends only on the base state . in the following
we assume without loss of generality that the direction of base flow is from left to right along the x - axis .
furthermore we assume it to be a function of @xmath39 only .
@xmath40 is some unknown function which is determined from the base flow . to carry out this determination of @xmath41 we consider eq .
( [ 2.21 ] ) as @xmath42 and express the left hand side of this equation in terms of @xmath28 only .
( assuming that disturbances do not propagate far upstream [ 16,18 ] ) .
( [ 2.21 ] ) is referred to as long s equation .
for example if we let @xmath43 i.e consider a shearless base flow with @xmath44 then @xmath45 and eq .
( [ 2.21 ] ) becomes : @xmath46 = 0.\ ] ] it is evident from this derivation that different profiles for the base flow as @xmath42 will lead to different forms of @xmath47 $ ] . for a general base flow in an unbounded domain over topography with shape @xmath48 and maximum height @xmath49 the following boundary conditions
are imposed on @xmath28 @xmath50 @xmath51 where the constant in eq.([2.27 ] ) is ( usually ) set to zero . as to the boundary condition at @xmath52 it is appropriate to set @xmath53 ( in spite of the fact that long s equation contains no dissipation terms ) .
however over finite computational domain only radiation boundary conditions can be imposed in this limit .
similarly as @xmath54 it is customary to impose ( following [ 7 ] ) radiation boundary conditions .
( the imposition of these boundary conditions is discussed in detail in @xmath55 ) . for the perturbation from the shearless base flow @xmath56 eq .
( [ 2.25 ] ) becomes @xmath57 where @xmath58 we observe that when @xmath59 the boundary condition ( [ 2.27 ] ) can be approximated by @xmath60 when @xmath61 is constant eq .
( [ 2.29 ] ) is invariant with respect to translations in @xmath62 and hence admits self - similar solutions of the form @xmath63 [ 19 ] .
these solutions are interpreted as gravity waves that are generated by the flow over the topography . from a numerical point of view
it is a common practice [ 7,8,13 ] to solve eq .
( [ 2.29 ] ) in the limit @xmath64 and @xmath27 with constant @xmath61 over the domain .
however observe that the definition of @xmath61 in long equation is given by ( [ 2.22 ] ) and it depends on @xmath28 . in some other numerical simulations
the computational domain is divided into subdomains where @xmath61 is constant in each subdomain but this led to numerical instabilities at the interface between these subdomains . in these limits eq .
( [ 2.29 ] ) reduces then to a linear equation @xmath65 we observe that the limit @xmath64 can be obtained either by letting @xmath66 or @xmath67 . in the following
we assume that this limit is obtained as @xmath66 ( so that stratification persists in this limit and the leading term in @xmath68 is not zero ) .
( [ 2.32 ] ) is a singular limit of long s equation as one of the leading second order derivatives drops when @xmath27 and the nonlinear terms drops out when @xmath64 and @xmath61 is constant .
this approximation and its limitations were considered numerically and analytically [ 6,7,19,20 ] and was found to be justified only under strong restrictions even under the assumption that the base flow is shearless .
nevertheless it is used routinely in the actual analysis of atmospheric data [ 14,15,16 ] . the general solution of eq .
( [ 2.32 ] ) is @xmath69 where the functions @xmath70 ) have to be determined so that the the boundary conditions derived from eq .
( [ 2.27]),([2.31 ] ) and the radiation boundary conditions are satisfied . these lead in general to an integral equation for @xmath71 and @xmath72 and it easy to show [ 13 ] that @xmath73 $ ] where @xmath74 $ ] is the hilbert transform of @xmath72 .
the boundary condition on the terrain becomes ; @xmath75\sin(\tau nf(x ) ) = -\tau f(x)\;.\ ] ] this integral equation has to be solved numerically [ 6,7,13 ] .
to derive a terrain following formulation of long s equation which incorporates the terrain in the coefficients of the differential equation ( rather than the shape of the domain ) we introduce gal - chen transformation .
if the height of the ( bottom ) terrain is described by a sufficiently smooth function @xmath76 and the height of the computational flow region is finite , i.e. @xmath77 , where @xmath78 is a constant , then this transformation is given by @xmath79 under this transformation we have @xmath80 where @xmath81 furthermore the expression of the laplace operator becomes @xmath82\displaystyle\frac{\partial^2 } { \partial\bar{z}^2 } + 2g^{12}\displaystyle\frac{\partial^2}{\partial\bar{x } \partial\bar{z } } + \left[\displaystyle\frac{\partial g^{12}}{\partial\bar{x } } + g^{12 } \displaystyle\frac{\partial g^{12}}{\partial\bar{z } } \right ] \displaystyle\frac{\partial}{\partial\bar{z}}.\ ] ] under this transformation the continuity equation ( [ 2.7 ] ) becomes @xmath83 however if we introduce @xmath84 then it is a simple algebra to show that eq .
( [ 3.5 ] ) can be rewritten as @xmath85 from this equation we see that we can introduce a `` terrain following stream function '' @xmath86 so that @xmath87 multiplying eq .
( [ 2.8 ] ) by @xmath88 we can rewrite this equation in the following form : @xmath89 using eq .
( [ 3.8 ] ) this can be rewritten as @xmath90 where @xmath91 is defined as in eq .
( [ 2.15 ] ) but with differentiations with respect to @xmath92 .
equation ( [ 3.10 ] ) implies that @xmath93 ( and vice versa ) . to eliminate the pressure term from eqs .
( [ 2.9 ] ) and ( [ 2.10 ] ) we differentiate ( [ 2.9 ] ) by @xmath39 and apply the operator @xmath94 to ( [ 2.10 ] ) and subtract .
we obtain @xmath95 using eq .
( [ 2.8 ] ) the first two terms in this equation can be written as @xmath96 \\ \nonumber & & = \frac{\beta}{2}\left[\rho_x(u^2+\mu^2w^2)_z -\rho_z(u^2+\mu^2w^2)_x\right ] = \frac{\beta}{2\sqrt{g}}\left[\rho_{\bar{x}}(u^2+\mu^2w^2)_{\bar{z}}- \rho_{\bar{z}}(u^2+\mu^2w^2)_{\bar{x}}\right ] = \\ \nonumber & & \frac{\beta}{2\sqrt{g}}\bar{j}\{\rho , u^2+\mu^2w^2\}.\end{aligned}\ ] ] using eqs ( [ 3.6 ] ) and ( [ 3.8 ] ) to re - express @xmath97 we have @xmath98(\psi_{\bar{z}})^2 \right\}\ ] ] the third and the fourth terms in eq .
( [ 3.11 ] ) can be rewritten using ( [ 2.7 ] ) as @xmath99 = -\frac{\beta\rho}{\sqrt{g}}\bar{j}\{\psi,\chi\}\ ] ] where @xmath100 is the vorticity .
expressing @xmath101 in terms of @xmath86 we have @xmath102 where @xmath103 \displaystyle\frac{\partial}{\partial\bar{z}}\right\}+ \left[\frac{1}{g } + \mu^2(g^{12})^2\right]\displaystyle\frac{\partial^2 } { \partial\bar{z}^2}\ ] ] is the `` terrain following laplace operator '' . finally for the right hand side of eq .
( [ 3.11 ] ) we have @xmath104 where @xmath105 combining all the results contained in eqs .
( [ 3.12])-([3.17 ] ) we can re - express eq .
( [ 3.11 ] ) in the following form : @xmath106 -n^2(\psi)g(\bar{x},\bar{z})\right\ } = 0\ ] ] where @xmath107 is defined as in eq . ( [ 2.22 ] ) .
hence it follows that , @xmath108 -n^2(\psi)g(\bar{x},\bar{z } ) = s(\psi).\ ] ] this is the terrain following form of long s equation . at this juncture
it might be asked why one can not `` save '' this derivation and apply the terrain following transformation ( [ 3.1 ] ) directly to ( [ 2.21 ] ) .
doing so will yield an extremely complicated equation .
this has been avoided in our derivation by the use of the `` terrain following stream function '' in ( [ 3.8 ] ) . to determine the function @xmath109 in eq .
( [ 3.19 ] ) we assume that @xmath110 and that ( as an example ) @xmath86 satisfies @xmath111 it follows then that @xmath112 and long s equation becomes : @xmath113 \\ \nonumber & & -n^2(\psi)\left[g(\bar{x},\bar{z})-\psi-\frac{\beta}{2}\right ] = 0 . \end{aligned}\ ] ] in this representation the flow domain is a rectangle @xmath114\times [ 0,h]$ ] or an infinite stripe @xmath115 \times [ 0,h]$ ]
. the boundary condition at the bottom topography is @xmath116 where @xmath117 is the normal to the topography which is described by the curve @xmath118 .
hence this normal is given by @xmath119 . using ( [ 3.6]),([3.8 ] )
this leads to the boundary condition @xmath120 and this constant can be chosen to be zero .
the other boundary condition that has to be imposed on @xmath86 is a radiation boundary condition at @xmath121 ( which implies that the outgoing wave is not reflected by the boundary ) . to obtain an equation for the perturbation from the base state we set @xmath122 substituting this in eq .
( [ 3.21 ] ) we obtain the following ( exact ) equation for @xmath123 @xmath124 \right ] \\
\nonumber & & = -\mu^2\left(\displaystyle\frac{\partial g^{12}}{\partial\bar{x } } + g^{12 } \displaystyle\frac{\partial g^{12}}{\partial\bar{z}}\right)+ n^2(\eta)\left\{h(\bar{x})(1-\frac{\bar{z}}{h})+\frac{\beta}{2}\left[(\frac{1}{g}+\mu^2(g^{12})^2)-1\right ] \right\ } .
\end{aligned}\ ] ]
in the traditional representation of long s equation the topography determines the shape of the flow domain and as a result it is not feasible to obtain analytic solutions to this equation even in some limits of the parameters @xmath24 and @xmath26 .
we now show that this problem can be overcome in some limiting cases when the terrain following formulation of this equation is used .
we consider two limiting cases @xmath125 and @xmath126 we also assume @xmath127 . for brevity we drop in the following the bars over @xmath62 . in this case
( [ 3.21 ] ) simplifies to @xmath128\ ] ] whose general solution is @xmath129 .\ ] ] here @xmath130 and @xmath131 are functions which have to be determined from the boundary conditions .
the boundary condition ( [ 3.22 ] ) implies @xmath132 . to determine @xmath133
we must apply the radiation boundary condition as @xmath54 on the solution . to this end
we must insure that the vertical group velocity of the wave is positive . using
the dispersion relation for hydrostatic flow given in [ 16,p.181 ] this group velocity is : @xmath134 where @xmath135 is the horizontal wave number .
we deduce then that the vertical group velocity is positive when @xmath136 . to impose this condition on the solution ( [ 4.2 ] )
we express @xmath131 in fourier integral form @xmath137 where @xmath135 is the horizontal wave number .
we deduce then that the solution ( [ 4.2 ] ) can be written as @xmath138\ ] ] to satisfy the radiation boundary condition for @xmath54 the first and second integral must vanish for @xmath139 and @xmath140 respectively
. therefore @xmath141 and @xmath142 must satisfy @xmath143 which implies that @xmath133 is the hilbert transform of @xmath132 .
this represents a complete analytic solution of long s equation for this limiting case .
we compare now these analytic results with the solution methodology that has been used previously in the literature as was discussed @xmath1 .
first we note that this analytic solution requires only the direct ( and simple ) computation of the hilbert transform of the terrain function @xmath118 .
this is a straightforward procedure even if it has to be done numerically . on the other hand to compute @xmath72 using ( [ 2.34 ] ) requires in general the solution of an integral equation .
to do so one must use an iterative algorithm which might turn out to be unstable or non - convergent over complex terrain .
furthermore there is the issue of applying the boundary conditions on @xmath86 at the terrain . to this end the procedure discussed in @xmath1 requires the use of the approximations that lead to ( [ 2.31 ] ) . as a result
the equation that is used to compute @xmath72 ( eq .
( [ 2.34 ] ) ) is also an approximate equation which will yield at best approximate solution for this function . on the other hand
the application of the boundary conditions using the procedure discussed in this section is exact and does not place constraints on the height of the terrain . from an experimental geophysical point of view
it has been a common practice to assume that the gravity wave generated by a flow over terrain is of the form @xmath144 ( or similar)[14,15,28 ] .
this has led to difficulties in the eduction of this wave from experimental data .
our results show that this form of the wave is incorrect ( at least in principle ) .
furthermore the numerical simulations that we carry in the next section demonstrate that complex terrain can alter drastically the shape and amplitude of this wave due to interference effects . in this limiting case
( [ 3.21 ] ) becomes @xmath145 + z + h(x)(1-\frac{z}{h } ) \right\ } = 0.\ ] ] since this is a nonlinear equation we can find an approximate analytical solution using first order perturbation expansion under the assumption that @xmath146 ( which is satisfied in most practical situations ) .
expressing @xmath86 approximately as @xmath147 and substituting this expression in ( [ 4.7 ] ) we obtain to order zero and one in the parameter @xmath148 the following equations @xmath149\ ] ] @xmath150 the boundary conditions on @xmath151 are given by eq .
( [ 3.22 ] ) at @xmath152 and radiation boundary conditions as @xmath54 . solving these ( linear ) equations for @xmath153 and @xmath154
we obtain the following expressions for their solutions @xmath155\ ] ] @xmath156 where @xmath157 @xmath158 @xmath159 ^ 2.\ ] ] the determination of the functions @xmath160 from the boundary conditions can be done using the same procedure outlined in the previous subsection .
however as it is algebraically cumbersome we omit the details .
in the previous section we presented analytical solutions to long s equation in some limiting cases . in general one has to resort to numerical simulations of long s equation .
however as it was said in the introduction the terrain following form of long s equation obviate the need to discretize the terrain in these numerical procedures and therefore lead to better representations of complex terrains .
also the application of the boundary condition at the terrain is being simplified considerably and is treated exactly contrary to the approach discussed in @xmath161 ( see eqs .
( [ 2.31 ] ) and ( [ 2.34 ] ) ) .
these simulations are carried in order to validate our equations and explore the different flow patterns ( and possible special effects ) that can be predicted over complex topography .
the geophysical motivation for these simulations is related to recent experiments to measure and educe gravity waves and their properties over the alps from balloon data .
this endeavor faced several difficulties . in part
these difficulties can be traced to the fact that the terrain over which the measurements were made contained ( at least ) two summits rather than one [ 15 ] . in this section
we carry simulations over three terrains . these consist of one , two and three hills with the following shape functions
@xmath162 @xmath163 @xmath164 in all cases we solved ( [ 3.24 ] ) for the perturbation with the following parameters ; @xmath165 here @xmath61 represents the nondimensional brunt - visl frequency which was defined in ( [ 2.22 ] ) . to solve for the perturbation @xmath123 over a finite two dimensional domain which after the transformation ( [ 3.1 ] )
is represented by @xmath166 \times [ 0,h]$ ] we imposed at the boundary @xmath152 the condition @xmath167 .
radiation boundary conditions were imposed at @xmath168 and @xmath169 .
these are necessary to avoid reflection of the outgoing wave . to implement these boundary conditions we used `` sponge boundaries '' at @xmath168 and @xmath169 [ 23,24 ] .
the sponge damping constant was calibrated to suppress wave reflection at these boundaries .
in addition open boundary conditions were used at these boundaries themselves viz .
we let @xmath170 in the simulations described below we let ( in nondimensional units ) @xmath171 , @xmath172 and @xmath173 with a grid of @xmath174 points .
the sponge layer ( at @xmath175 ) consists of @xmath176 grid points . to solve ( [ 3.24 ] ) under these settings we used matlab[28 ] .
central finite differences approximations were used to discretize long s equation on the domain grid and a fixed point iterative algorithm was implemented to solve the resulting equations .
the convergence criteria for the iterations was that the step error @xmath177 was less than @xmath178 where @xmath179 is the iteration number .
convergence was achieved in less than @xmath180 iterations .
- it should be noted however that @xmath123 in these simulations represents the perturbation from the `` terrain following stream function '' which was defined in ( [ 3.8 ] ) . using ( [ 3.6 ] ) , ( [ 3.8 ] ) we find that the perturbation to the base flow field @xmath181 is given by @xmath182 @xmath183 where @xmath184 is defined in ( [ 3.1 ] ) . from this vector field
we can compute the `` regular '' stream function @xmath185 which is defined as @xmath186 where @xmath187 satisfies @xmath188 a zoom - plot of @xmath185 ( on part of the computational domain ) on the region around the three topographies discussed above is presented in @xmath189 .
we observe that the phase lines of the stream function in all these figures tilts with height . in all three figures
the gravity waves are in the lee of the topography .
these figures demonstrate the impact that complex terrain can have on the structure of gravity waves .
@xmath190 displays the results of the simulation for one hill ( described by ( [ 5.1 ] ) ) .
the plot corresponds to the classical results that appeared in the literature for this case .
the results for two hills in @xmath191 show some wave activity between the two hills and less organized waves with somewhat smaller amplitude in the lee of the two hills .
this might be due to interference effects between the waves generated by the two hills . for three hills
the simulation shows the same wave activity between the two hills ( on the left ) as in @xmath191 but then a `` quiet '' zone between the second and third hills in which @xmath192 is almost constant .
the waves in the lee of the three hill appear regular as in @xmath190 .
we derived in this paper a terrain following formulation of long s equation in which the topography is `` absorbed '' in the coefficients of the differential equation representing the flow rather than being part of the boundary conditions .
we used this representation to solve long s equation analytically in some limiting cases and numerically over complex topography .
the new formulation also opens the possibility to develop analytical estimates which compare the solutions of this equation over different topographies . from a geophysical point of view
it well known that some present models for the generation of gravity waves over estimate this effect [ 25,27 ] .
partially , this is due to the fact that they use oversimplified representation of the terrain .
furthermore they do not take into account the effects that are due to complex terrain ( as demonstrated by our simulations ) .
we believe that the new form of long s equation will make it easier to consider more realistic representations of the terrain and its effect on the generation and propagation of gravity waves .
dewan , r.h .
picard , r.r .
oneil , h.a .
gardiner , j. gibson , j.d .
mill , e. richards , m. kendra , and w. o. gallery , 1998 , .
msx satellite observations of thunderstorm - generated gravity waves in mid - wave infrared images of the upper stratosphere , geophysical research letters 25 pp .
939 - 942 .
fig . 1 zoom - contour plot of the stream function @xmath192 over one hill + centered at @xmath193 with @xmath194 , @xmath195 , @xmath196 , @xmath197 .
+ + fig .
2 same as @xmath198 with two hills centered at @xmath193 and @xmath199 .
+ + fig .
3 same as @xmath198 with three hills centered at @xmath200 and @xmath201 . | long s equation describes two dimensional stratified atmospheric flow over terrain which is represented by the geometry of the domain .
the solutions of this equation over simple topography were investigated analytically and numerically by many authors . in this paper
we derive a new terrain following formulation of this equation which incorporates the terrain as part of the differential equation rather than the geometry of the domain .
this leads to new analytic insights about the solutions of this equation and enable us to compute steady state gravity wave patterns over complex topography .
pacs 92.60.gn , 92.60.dj , 02.30.ik |
among the nearly 2000 exoplanets discovered so far , more than half are transiting systems . due to the rotation of the host star , the observed stellar radial velocity ( rv )
is expected to change as the planet transits different parts of the rotating stellar surface .
this is called the rossiter - mclaughlin ( rm ) effect and was initially measured for eclipsing binary stars @xcite .
after the first observation of the exoplanetary rm effect by @xcite , there are now more than 80 exoplanet systems with this effect observed , e. g. @xcite .
this effect is normally used to measure the projected angle between the stellar spin and the planetary orbit .
@xcite was the first to propose using the wavelength - dependent rm effect to probe the atmospheres of transiting exoplanets .
this method exploits the fact that the planetary effective radius is wavelength - dependent due to differential atmospheric absorptions .
@xcite theoretically modeled the rm effect for the na absorption of giant exoplanets .
the advantage of the rm method compared to the traditional spectrophotometry is that it does not demand the use of photometric reference stars .
the rm method promises to become a powerful technique for future transmission spectrum measurements , especially as the next generation of very large ground - based telescopes are likely to have a small field - of - view , making it difficult to employ suitable reference stars . with more and more terrestrial exoplanets being discovered , the characterization of their atmospheres will become a major goal for exoplanet research in which the rm method can play an important role . although it is currently difficult to observe the rm effect of these terrestrial exoplanets due to instrumental limitations , a lunar eclipse provides us with an opportunity to observe the earth transiting the sun and so explore the effectiveness of the rm method .
the concept of regarding the earth as an exo - earth using lunar eclipses has been applied in the past by @xcite .
these studies obtained the transmission spectrum directly from the ratio spectrum before and during a lunar eclipse . here
we present for the first time the use of the rm effect of an earth transit during a lunar eclipse to retrieve the transmission spectrum .
we observed the 15-april-2014 lunar eclipse with the high accuracy radial velocity planet searcher ( harps ) mounted on the eso la silla 3.6 m telescope @xcite . a consecutive sequence of observations over one entire night covered all of the eclipse stages , i.e. the penumbral and umbral eclipse and out of eclipse ( called hereafter the bright moon ) .
the classical fiber spectroscopy mode was used with the fiber located at the center of the tycho crater ( see fig .
[ nasa - tycho ] for the trajectory ) .
the telescope used non - sidereal tracking and the tracking velocities were updated every few minutes .
in addition , manual guiding on crater structures was performed .
this worked reasonably well and limited the pointing coordinate drift to @xmath0 1.5 arcmin .
we used varying exposure times during the eclipse because the lunar surface brightness changed dramatically . in total ,
382 lunar spectra were obtained ( table [ observation ] ) . [ cols="^,^,^,^,^",options="header " , ] data reduction was performed using the harps pipeline .
each exposure frame comprises 72 spectral orders that cover the wavelength range from 378 nm to 691 nm with a spectral resolution of @xmath1 @xmath2 115,000 .
radial velocities are measured , for each spectral order as well as for the overall spectral range , using a cross - correlation with a g2 stellar spectral template .
the overall measured and theoretical rv curves are shown in fig .
[ rme - curve ] .
the theoretical rv is the combination of two components : the motion of the sun with respect to tycho and the motion of tycho with respect to the observer .
these velocities are calculated with the jpl horizon ephemeris , which considers the orbital motions , rotations of the earth and the moon , and the light travel time corrections .
there is an offset between the theoretical rv and the measured rv , which probably originates from the rv zero - point of the spectral template used by the harps pipeline @xcite .
we correct this offset by using the data points taken during the bright moon as a baseline .
[ rme - curve ] shows the measured rv with the motion rv corrected and the baseline offset subtracted .
this is the final rv curve of the rm effect for the earth transiting the sun . at the bright moon stage ,
the rv is essentially corrected to zero , but there is still a small slope with an amplitude of about 4 m / s which is probably due to the instrumental drift and the non - perfect telescope guiding when observing the moon .
since the rm amplitude is on the order of 2 km / s , this small residual slope does not significantly affect our analysis .
when the moon enters the 1st penumbra , the observed rv becomes negative ( blueshift ) as the earth begins to transit the redshifted rotating part of the solar disk . at the umbral stage ,
the rv gradually changes from a blueshift to a redshift .
when the moon enters the 2nd penumbral stage , the rv is redshifted since the earth obscures mainly the blueshifted solar region .
the rv gradually decreases during this stage as the moon moves out of the penumbral shadow .
the umbral rv has a relatively complex structure and its details are determined by a combination of both the refracted part of the solar disk and the properties of the earth s atmosphere refracting the sunlight . in this work , the umbral part is not studied since we focus only on the penumbral parts for the rm effect .
to retrieve the wavelength - dependent rm effect due to the earth s atmospheric absorption , we firstly establish an rm model to take into account the wavelength - dependent solar parameters .
our model follows the method described in @xcite and @xcite .
we divide the solar disk into elements with a size of @xmath3 and the radial velocity due to solar rotation for each element is : @xmath4 where @xmath5 is the rotation angular velocity , @xmath6 is the inclination of the solar spin axis towards tycho , x and y are in the sky - plane coordinate system of which the origin is at the projected solar center and the y - axis is along the projected solar rotation axis .
the coordinate values are in units of the solar radius @xmath7 .
the rv anomaly @xmath8 due to the rm effect is then calculated by integrating the intensity - weighted rvs of the visible solar disk : @xmath9 the rm curve for white - light ( i.e. the entire wavelength range of the harps spectrum ) is generated from this model ( the black line in fig .
[ white - light - model ] ) .
the parameters are further described as follows : 1 .
the coordinates of the earth s center and the solar center as seen from tycho are generated using the horizon ephemeris .
we use the earth radius of @xmath10 km and the solar radius of @xmath11 km .
these data are used to determine the visible part of the solar disk at a given time .
the inclination @xmath6 is calculated using the solar north pole position angle given by horizon .
the mean value of @xmath6 is @xmath12 and changes by about @xmath13 during the eclipse .
the solar differential rotation used is @xcite : @xmath14 where @xmath15 is the heliographic latitude and is calculated with : @xmath16 the coefficients adopted are a = 13.46 , b = -2.99 @xcite .
the quadratic limb - darkening coefficients @xcite are used for the white - light model to calculate the intensity @xmath17 .
we adopt the coefficients used by @xcite , i.e. @xmath18 and @xmath19 . for an exoplanet transit , the projected angle between the stellar spin and the planetary orbit
is calculated by fitting the observed rv with the rm model .
however , we adopt the actual data from the horizon ephemeris and so the model curve in fig .
[ white - light - model ] is not the result of such a fit .
the projected spin - orbit " angle for this observation is @xmath20 calculated using the trajectory of the earth as seen from tycho . because this trajectory is the combination of the earth and lunar orbits , the angle is not the real spin - orbit angle of the sun - earth system of @xmath21 . the stellar convective motion results in a blueshift of the lines which is called the convective blueshift ( cb ) @xcite .
as the cb radial component varies across the stellar disk , the observed stellar rv changes when a planet transits the stellar surface .
@xcite discusses the cb during an exoplanet transit and its effect on the rm effect .
we follow the cb model in @xcite and adopt a typical cb value of the sun @xmath22 m / s for the white - light rv calculation .
the blueshift value for each element on the solar disk is @xmath23 , where @xmath24 . the rv change caused by cb ( blue line in fig .
[ white - light - model ] ) is then calculated in a similar way to the rm model and added to the rv value caused by the rm effect to give the red line in fig .
[ white - light - model ] .
the effects of atmospheric refraction are important for interpreting the transits of terrestrial planets @xcite and affect the rm effect .
as the atmosphere refracts the light from the stellar disk region that should be obscured by the planet , the apparent effective radius of the planet will appear smaller than without refraction .
thus in general , refraction would result in a smaller rm amplitude than without refraction . in our rm model
, we do not include atmospheric refraction , which means the modelled rm amplitude should be larger than the actual observed rm amplitude .
this is consistent with the result in fig .
[ white - light - model ] in which the modelled rv value without refraction ( red line ) is generally larger than the observed rv ( green points ) . in the rm model
described above , we assume the line profile is dominated by the stellar rotation and use the first - moment approximation @xcite .
this assumption gives us the rv anomaly expressed by equation [ equ - deltarv ] .
however , other mechanisms like the microturbulence , macroturbulence and instrumental broadening also affect the line profile and can change the shape of the rv curve @xcite .
this may contribute to the discrepancy between the modelled rv and the observed rv in fig .
[ white - light - model ] .
the harps pipeline produces directly a rv value for each of the 72 spectral orders .
three orders do not have useful rv data due to the lack of stellar lines in the corresponding spectral mask .
after correcting for the earth - moon - sun motion and the baseline offset described in section 2 , we obtain 69 rm curves each representing a different wavelength range . to obtain the rm effect differences caused by the earth s atmosphere ,
the wavelength - dependent solar parameters need to be considered .
the first aspect concerns the stellar convection blueshift .
the actual rv caused by cb depends on where the line forms , e.g. the strong lines or low - excitation lines may have a small cb because they usually form high in the stellar atmosphere where the granulation is weak @xcite .
as each of the harps spectral orders comprises stellar lines with different cb values , the @xmath25 of each spectral order varies .
therefore , unlike the white - light cb model for which we can assume @xmath22 m / s , the actual cb effect for a given harps order is difficult to model .
instead , we use two symmetric parts on the rv curve to cancel this effect utilizing the fact that the convection induced rv anomaly during transit is symmetric with respect to the mid - transit as shown by the blue line in fig . [ white - light - model ] .
two regions of the rm curve ( labelled as pm1 and pm2 in fig .
[ white - light - model ] ) are then chosen .
the pm1 and pm2 are symmetric regions of the eclipse , i.e. they are at the same distance from mid - eclipse .
[ rme - amplitude]a shows the rv values of pm1 and pm2 and the rv difference between them ( i.e. pm2 - pm1 ) .
we use this pm2 - pm1 value to represent the rm effect amplitude for each spectral order . from fig .
[ rme - amplitude]a , it can be seen that there is a correlated systematic difference between pm1 and pm2 .
we believe that this correlated pattern results from the measured effect of the cb depending on the type and number of spectral lines present in each spectral order but is well - cancelled by using pm2 - pm1 . the second wavelength - dependent solar parameter is the limb - darkening . here
we use the empirical power - law limb - darkening coefficients from @xcite . for each of the harps orders , we interpolate a limb - darkening coefficient at the corresponding wavelength and calculate a value of pm2 - pm1 with a fixed earth radius ( 6378 km ) using our rm model .
the model values of pm2 - pm1 are shown in fig .
[ rme - amplitude]b . the final rm effect at different wavelength ranges , after the correction of the limb - darkening ,
is presented in fig .
[ rme - amplitude]c .
the limb - darkening correction made here is mainly for the continuum , however , since the solar lines have different limb - darkening compared to the adjacent continuum @xcite , this could introduce some extra noise in the final rm amplitudes . the rv curve of the wavelength - dependent rm amplitudes in fig.[rme - amplitude]c results from the different effective radius at different wavelengths , and it can be regarded as a mapping of the low resolution transmission spectrum of the earth s atmosphere into effective radius and hence radial velocity space . to interpret the observed atmospheric features , we build a transmission spectral model following the methods of @xcite and @xcite .
we calculate the overall transmission spectrum of the earth atmosphere from 0 to 80 km altitude considering the ozone absorption and rayleigh scattering extinction ( blue line in fig .
[ rme - amplitude]c ) .
this transmission spectrum is overlaid with the rm amplitude curve to compare their shapes .
we emphasize here that the transmission spectrum model is not used to fit the observed rm amplitude curve but to demonstrate the presence of the atmospheric features mapped into it . by comparing the spectral shapes in fig .
[ rme - amplitude]c , we interpret the rm amplitude curve as follows : the rm amplitude is larger towards the blue due to rayleigh scattering extinction that makes the atmosphere more opaque and the atmospheric effective thickness larger at shorter wavelengths .
the broad peak around 600 nm results from the ozone chappuis band absorption while the rm amplitude towards the red becomes smaller since both the ozone absorption and the rayleigh scattering become weaker , rendering the atmosphere more transparent .
the rm model we have used for this work suffers from several defects and incompleteness .
the first is observational in that there is a slight , but barely significant , drift in the telescope guiding on the lunar surface .
the second is that other line broadening mechanisms like the stellar micro / macro - turbulence are not included in the rm model ( cf .
subsection [ sec - turbulence ] ) .
the third is more fundamental and its full solution is beyond the scope of this letter .
this is the determination of the detailed mapping between the rm amplitude and the atmospheric exctinction as a function of wavelength .
this depends on the effects of atmospheric refraction in the exoplanetary atmosphere which will be influenced not only by the absorption coefficients of its gaseous constituents but also by the effects of screening due to clouds and aerosols .
the combination of these effects can explain the slight deviation of the modelled rv from the observed rv values .
it can also explain why the modelled rm amplitudes for the limb darkening correction are larger than the observed rm amplitudes ( shown in fig .
[ rme - amplitude]b ) .
however , this has a very limited effect on our retrieval of the final rm amplitude curve since the limb - darkening correction produces only a small rv variation with wavelength so that the shape of the final rm amplitude curve is little affected by the rm model .
future modelling work containing a proper treatment of refraction and stellar turbulence will allow us to combine the transmission spectral model with the rm model , and to fit the observed rm amplitude curve directly instead of just comparing its shape with the transmission spectrum as presented in the current work .
we have observed for the first time the rossiter - mclaughlin effect of the earth transiting the sun using a lunar eclipse . the rm effect curve has been obtained using high accuracy rv observations and an rm effect model is built to analyze the observed result .
separate rm curves at different wavelengths are obtained from 69 harps spectral orders .
after the correction of the wavelength - dependent limb - darkening of the sun and the convective blue - shift of the solar lines , we retrieve the wavelength - dependent rm amplitudes due to the transmission of the earth s atmosphere . the ozone chappuis band and the rayleigh scattering signatures are clearly detectable .
the rm method can be used to detect broad features , such as the rayleigh scattering , in exoplanet atmospheres .
the advantage is that no reference stars are needed in contrast to the requirements for the traditional spectrophotometric method . since the next - generation ground - based telescopes such as e - elt will have a relatively small field - of - view , limiting the access to nearby reference stars
, the rm method will provide a promising technique for the characterization of planetary atmospheres .
thus , in the future , this method can be applied to detect the atmospheres of terrestrial planets and particularly to search for the bio - signature gas ozone . extending the observation to near - infrared ( nir )
wavelengths will yield more absorption features .
however , the ability of the rm method is limited by the number of suitable spectral lines of the parent star .
the sun , for example , does not have sufficient lines in the nir .
however , for m - type stars which are expected to be promising targets for exoplanet atmosphere studies , there are many stellar molecular lines in the nir which can be exploited .
we greatly appreciate the excellent support from the lasilla / paranal science operations team . in particular , we thank valentin ivanov for performing the observations on site and his and lorenzo monaco s support during the preparation of the observing run .
we would also like to thank gaspare lo curto and andrea chiavassa for helpful discussions and the referee for useful suggestions .
the study is supported by the national natural science foundation of china under grants nos .
11390371 and 11233004 .
_ facilities : _ . | due to stellar rotation , the observed radial velocity of a star varies during the transit of a planet across its surface , a phenomenon known as the rossiter - mclaughlin ( rm ) effect .
the amplitude of the rm effect is related to the radius of the planet which , because of differential absorption in the planetary atmosphere , depends on wavelength .
therefore , the wavelength - dependent rm effect can be used to probe the planetary atmosphere .
we measure for the first time the rm effect of the earth transiting the sun using a lunar eclipse observed with the eso harps spectrograph .
we analyze the observed rm effect at different wavelengths to obtain the transmission spectrum of the earth s atmosphere after the correction of the solar limb - darkening and the convective blueshift . the ozone chappuis band absorption as well as
the rayleigh scattering features are clearly detectable with this technique .
our observation demonstrates that the rm effect can be an effective technique for exoplanet atmosphere characterization .
its particular asset is that photometric reference stars are not required , circumventing the principal challenge for transmission spectroscopy studies of exoplanet atmospheres using large ground - based telescopes . |
distance - scale determination of gamma - ray bursts ( grb ) has been one of the most important achievements of astrophysics in recent years .
accurate and fast localization of the prompt and afterglow emission bursts ( i.e.@xcite ) by the italian / dutch x ray satellite bepposax ( @xcite ) led to the identification of optical counterparts ( @xcite ) and ultimately to spectral measurements of a redshift ( @xcite ) . while the extragalactic origin of grb has gathered solid evidence in its support , the source of the large energy implied by their distance is still speculative .
direct information on the nature of the central engine of grb can be derived by studying the nearby environment ; for example using line spectroscopy .
the star forming region and massive pre - explosion winds associated with the hypernova scenario ( @xcite ) imply a mass rich environment down to short ( @xmath3 ) distance scales .
in contrast , ns - ns merging should happen in a fairly clean environment , because such objects are expected to form with significant speeds which can lead them at least a few parsecs away from the regions of star formation in which they formed .
current line measurements are inconclusive , because all the spectral features observed so far are signatures produced by the host galaxy rather than at the burst site .
grb980425 is an exception , but its tantalizing association with sn 1998bw ( @xcite ) requires clarification . the measurement of x - ray lines emitted directly by the grb or its afterglow could provide a direct measurement of the distance and probe into the nature of the central environment ( @xcite ) .
fe k line emission is very promising in this respect .
theoretical computations have been carried out by several authors in the framework of grb ( @xcite ) .
most of this work deals with features generated within the interstellar medium ( ism ) of the host galaxy , i.e. on scale larger than several parsecs .
however , even in the favourable case of dense regions of stellar formation ( e.g. the site of grb formation in the hypernova scenario ) , iron emission lines would be hard to detect with current x - ray instrumentation .
lately , mszros and rees ( 1998 ) have shown that the circumburst environment created by the stellar wind before the explosion of the hypernova could yield a line of substantial intensity .
a similarly favourable situation should be expected in a variation of the hypernova scenario , the supranova ( @xcite ) , where the grb is shortly preceded by a supernova explosion with the ejection of an iron rich massive shell . in these cases , and for a isotropic distribuition of material around the source , a tight upper limit on the column density @xmath4@xmath5 derives from the requirement that the thomson optical depth be less than a few , to avoid smearing out the the short time - scale structure of the grb by thomson scattering ( @xcite ) .
furthermore , fe absorption features of stregth comparable to that in emission should also be detected .
it is also conceivable that the impact of the relativistic shell that produced the original grb on these ejecta could provide an additional energy input in the afterglow .
grb970508 is the most promising candidate of the bepposax afterglows for the search of line emission .
it is characterized by a large outbursting event during its afterglow phase ( @xcite ) and has the highest signal to noise ratio of the bepposax grb afterglows . in this letter
we present results of a spectral analysis of the afterglow of grb970508 and discuss their implications on the nature of the central engine of grb .
the gamma - ray burst grb970508 was detected by the gamma - ray burst monitor ( grbm ) aboard bepposax ( @xcite ) .
the wide field camera ( wfc ) provided a few arcmin position acquired with the narrow field instruments 6 hours after the burst , which led to the identification of the x - ray afterglow associated with the grb .
three further observations were carried out , the last performed 6 days after the burst ( fig.1 , tab.1 ) .
lecs(0.1 - 3.5 kev ) and mecs(1.8 - 10.0 kev ) spectra of the observations were obtained using standard procedures ( bepposax web pages : _
www.sdc.asi.it_ ) , with 15 - 20 counts ( source + background ) per bin .
inspection of the light curve ( fig.1 ) shows that in the first half period of observation number 1 the flux decays , followed by a rising trend in the second half .
we have thus split this observation into two parts ( 1a & 1b ) at the time of the minimum of the light curve , to investigate spectral evolution in different states , producing 5 spectra in total .
we fitted each spectrum with a power law with variable absorption , keeping the relative normalization of lecs vs mecs free to vary in the range 0.6 - 0.9 .
this model provides a good fit to all but dataset 1a . in this case
we have @xmath7 for @xmath8 degrees of freedom , which corresponds to a probability of 1.4% however the major contribution to @xmath9 is not randomly distributed but is concentrated in a line feature around 3.5 kev ( fig .
2a ; note that in figure we show for clarity only data points below 2 kev for lecs and above 2 kev for mecs ) .
the possible association of this feature with fe k line emission is obvious , considering that its energy coincides with - k@xmath10 emission at the redshift z=0.835 , i.e. that of the putative host galaxy ( bloom et al .
we have therefore added to the power law model a narrow line ( width @xmath11 fwhm@xmath12 kev ) at the energy of 3.5 kev .
the relative normalization lecs / mecs that depends primarily on the position of the source in the lecs detector has , in all datasets , an error greater than the nominal range 0.6 - 0.9 . we have hereafter fixed it to 0.8 , the value averaged over all data sets .
the best fit now yields @xmath13 with @xmath14 .
the improvement of @xmath15 corresponds to a confidence level of 99.3% ( f test ) .
we stress that this confidence level relies on the a priori knowledge of the energy of the fe line available from an independent redshift measurement .
were it not available , the confidence level for the presence of a line with free energy would have been 97% .
the best fit line intensity is @xmath16 ph @xmath5s@xmath17 ( hereafter all errors and upper limits correspond to 90% confidence level for a single parameter ) .
upper limits on line intensity in datasets 1b ( fig.2b ) , 2 3 and 4 are 2.5 , 2 , 2.5 and 0.5 @xmath18 ph @xmath5s@xmath17 respectively . the power law parameters obtained in dataset 1a and 1b are consistent with each other ( see tab
. 1 ) and the absorption column densities are consistent with the hydrogen column in the line of sight of our galaxy , i.e. the intrinsic absorption is consistent with zero .
we next performed a simultaneous fit to both datasets with the same continuum parameters , and the addition of a line to dataset 1a only . in this case
we left the energy of the line as free parameter , deriving @xmath19 , corresponding to @xmath20 in the source rest frame .
the error range is too large to identify the ionization state of iron . in fig.2 ( inset )
we show the contour plot of i@xmath21 vs e. we note that the absence of the line in dataset 1b explains why a power law model provides an acceptable fit to the summed spectrum ( @xcite ) .
we have also investigated whether the spectral continuum could be fitted by a black body or a thermal bremsstrahlung .
we find that thermal models provide worse results than that of a power law continuum , as already shown in the case of grb970228 ( @xcite ) .
spectral variability of the afterglow spectra supports the idea that a refreshed energy input produced the outburst observed in the light curve . the spectral index in the first observation ( tab .
1 ) is @xmath22 . in the second and third observation
the flux lies above an extrapolation of the light curve expected from a power law evolution from the primary event .
the spectrum is harder ( @xmath23 ) , indicating electron reacceleration , as expected in a newly formed shock . in the last observation
the flux decreases and the spectrum is very steep ( @xmath24 ) : the new energy source has been exhausted .
we have noticed previously that spectra 1a and 1b are consistent with each other .
since the transition 1a-1b identifies the beginning of the outburst , one would then expect an hardening of the spectrum in 1b .
however such an effect could be masked by the continuation of the power law decay law ( fig.1 ) , which accounts for about 60% of the flux observed in dataset 1b .
the minimum amount of mass @xmath25 present can be estimated ( lazzati 1998 ) simply by counting how many ionizing photons a given iron atom can absorb , assuming that recombination is much faster than the time interval between successive ionizations , and by demanding that the existing iron atoms thus account for the fluence in the line : @xmath26 where @xmath27 is the iron abundance in units of the solar value , @xmath28 is the line duration , @xmath29 is the distance of the medium from the central source , and @xmath30 is the fraction of the total energy @xmath31 which is recycled into the line . since the line disappears after about @xmath32 , we expect the emitting region to have a size @xmath33 ; in fact , since the line width is mostly instrumental , we can bar the existence of relativistic effects .
then the typical density is @xmath34 , and typical thompson and iron optical depths are @xmath35 and @xmath36 .
> from the above we deduce that the emitting material can not lie along the line of sight : for @xmath37 , @xmath38 , and thomson scattering would smear out the time structure of the burst , contrary to observations ( @xcite ) .
even if @xmath39 , we would still have @xmath40 , which implies an iron edge not seen in the spectrum .
for @xmath41 , we notice that the fireball would load itself with so many baryons to spoil its relativistic expansion , again contradicting observations .
having established that the emitting medium lies sideways from the line of sight , we also notice that we can not have @xmath42 , because then the line would appear much later than 1 day .
we conclude that @xmath43 .
the ionization parameter , using the minimum density derived above , is found to be @xmath44 where @xmath45 is the luminosity in the range 0.016 - 10 kev , whose temporal evolution is approximately described by a power law ( @xcite ) .
> from eq .
( 2 ) one sees that for early times ( @xmath46 ) , the material will be fully ionized , and thus the line may be due to recombination , while at later times , when @xmath47 , fluorescence may account for the emission ( @xcite ) .
the exact moment of the transition can not be determined , because we established above that the emitting material lies sideways with respect to the observer , and we can not exclude some amount of beaming , but the following discussion still stands . at early times , the emitting material will probably reach close to the inverse compton temperature in the burst radiation field , which clearly exceeds @xmath48 ; for these high temperatures and low densities , recombination is not an effective emitter of line photons , and it is most likely that most of the line we see is emitted by fluorescence in the time range @xmath49 , during which the total observed fluence is @xmath50 . putting this into eq .
[ min ] , we find a minimum mass of @xmath51 which is our best estimate for the total mass present . in principle , the line might arise from the normal interstellar medium surrounding the burst volume , or from a normal stellar wind emitted by the burst progenitor .
however , a very large density ( @xmath52 ) is necessary to account for the observed spectral feature ; this is obvious when one considers that the line reported here is as bright as a typical one from clusters of galaxies at @xmath53 .
winds of massive stars are also unlikely to be responsible for the iron line : even for rather extreme wind parameters ( @xmath54 , and @xmath55 ) , the wind density exceeds the minimum density only for @xmath56 . however , for this small distance , the ionization parameter ( eq . [ xi ] )
is @xmath57 even a day after the burst and , even including beaming , the wind material will be totally ionized ; under these conditions , only recombination can provide the line photons , but the recombination rates are very small ( @xmath58 ) , and thus the conditions of instantaneous de excitation under which eq .
[ min ] was established are no longer satisfied : much larger masses are required than can be accounted for by the winds .
furthermore , such small distances ( @xmath59 ) can not explain the line duration .
we thus conclude that the material surrounding the burst must have been pre ejected .
@xmath60 is likely to be close to an upper limit to the iron abundance of the ism or of a stellar wind at @xmath61 .
scenarios concerning mergers of binaries then begin to appear unlikely , regardless of whether the binary is made of two neutron stars , a black hole and a neutron star , or a black hole and a white dwarf , because it has been established ( mszros and rees 1992 ) that they can not pre eject , even under the rosiest assumptions , more than @xmath62 due to their tidal interactions , about three orders of magnitude less than the limit in eq .
we remark that the supranova ( @xcite ) provides the most appealing scenario , because of the large amount of iron material ejected by the supernova explosion preceding the burst .
in fact , theoretical computations of the explosion of massive stars indicate that @xmath63 of iron are ejected by the sn ( @xcite ) . in the case of sn1987a ,
whose progenitor had @xmath64 , a similar amount of iron was derived from optical observations ( @xcite ) . in future ,
it is conceivable that simultaneous observations of the ly@xmath10 line may help to determine @xmath27 , and thus uniquely determine the total illuminated mass , and by implication the adequacy of different formation scenarios .
in this paper we have obtained the following results : * we have searched the x ray spectrum of gb970508 s afterglow for an iron line , located at the system s redhift ( z = 0.835 ) ; we found such a line with limited statistical significance ( @xmath65 ) in the early part ( first @xmath66 ) of the afterglow ; the line decreases in the later part of the observations ( @xmath67 day after the burst ) by at least a factor 2 , enough to make it undetectable with current apparatus ; * simultaneously with the line disappearance , the x ray flux both rises and hardens ( @xmath68 , while @xmath69 before the reburst ) , consistent with the appearance of a new shock .
then , at the end of the outburst , the spectrum steepens .
we showed that these observations indicate that a mass @xmath70 should be located at a distance of @xmath71 and sideways with respect to the observer , and is moving at subrelativistic speed . in order to reach such distance ,
this material must have been pre ejected by the source originating the burst , shortly ( perhaps a year , for typical sn expansion speeds ) before the burst .
we stress that these observations contain two coincidences : on the one hand , this is the only burst in which a reburst and a line have been observed by bepposax ; on the other , the iron line disappears exactly at the moment of the reburst .
we also point out that a line feature , with a similar significance , has been found by asca in another burst , grb970828 , which also shows an event of rebursting during the x - ray afterglow ( @xcite ) .
we thank the referee , dr .
m. boettcher , for his critical suggestions that improved the paper substantially , f. matteucci for helpful discussions and the bepposax team for the support with observations .
bepposax is a program of the italian space agency ( asi ) with the participation of the dutch space agency ( nivr ) .
bloom j.s . ,
djorgovski s.g .
, kulkarni s. r. & frail d. a. 1998 , , * 507 * , 25 .
boella g. 1997 , , * 122 * , 299 .
boettcher m. , dermer c.d.d .
, crider a. w. & liang e. d. 1999 , , in press , preprint : astro - ph/9809156 .
costa e. 1997 , , * 387 * , 783 danziger j. & bouchet , p. 1993 , in new aspects of magellanic cloud research , proc . of _ evolution of galaxies _ , heidelberg , june 15 - 17 1992 , p.208 frontera f. 1998 , , * 493 * , 67 galama t.j .
1998 , * 395 * , 670 .
ghisellini g. , , submitted ; preprint : astro - ph/9808156 .
hirano t. 1987 , 39 , 619 kallmann t. r. & mccray r. 1982 , , 50 263 kulkarni s. r. 1998 , , * 395 * , 663 .
lazzati d. , ghisellini g. & campana s. , , submitted .
mszros p. & rees m. j. 1992 , , * 397 * , 570 .
mszros p. & rees m. j. 1998 , , * 299 * , l10 .
metzger m.r .
1997 , * 387 * , 878 .
owens a. 1998 , , * 339 * , l37 .
paczynski , b. 1998 , , * 494 * , l45 .
perna r. & loeb a. , 1998 , , * 501 * , 467 .
piro l. , scarsi l. & butler r.c . in _ x - ray and euv / fuv spectroscopy and polarimetry _
s. fineschi ) spie 2517 , 169 - 181 ( 1995 ) .
piro l. 1998 * 331 * , l41 .
thieleman f - k . ,
nomoto , k. & hashimoto , m. 1996 , , * 460 * , 408 van paradijs 1997 , , * 386 * , 686 .
vietri m. & stella l. , * 507 * , l45 ( 1998 ) woosley s. e. & weaver , t. a. 1995 , * 101 * , 181 .
yoshida a. 1998 , proc .
of _ gamma - ray bursts in the afterglow era _ , rome , nov . 3 - 6 1998 , frontera f. & piro l. eds . ,
a&as , submitted llllllllll 1a & 0.24 & 8 & 11 & & @xmath72 & @xmath73 & @xmath74 & @xmath75 & 6.5/7 1b & 0.65 & 7 & 16 & & @xmath76 & @xmath77 & @xmath78 & @xmath79 & 6/9 1a+b & 0.24 & 15 & 28 & & @xmath80 & @xmath81 & @xmath82 & @xmath83 & 14.2/18 2 & 2.74 & 6 & 24 & & @xmath84 & @xmath85 & @xmath86 & @xmath87 & 6.2/9 3 & 4.14 & 3 & 12 & & @xmath88 & @xmath89 & @xmath90 & @xmath91 & 5.2/6 4 & 5.7 & 14 & 73 & & @xmath92 & @xmath93 & @xmath94 & @xmath95 & 6.3/10 | we report the possible detection ( 99.3% of statistical significance ) of redshifted fe iron line emission in the x - ray afterglow of gamma - ray burst grb970508 observed by bepposax .
its energy is consistent with the redshift of the putative host galaxy determined from optical spectroscopy .
the line disappeared @xmath0 day after the burst .
we have also analyzed the spectral variability during the outburst event that characterizes the x - ray afterglow of this grb .
the spectrum gets harder during the flare , turning to steep when the flux decreases .
the variability , intensity and width of the line indicate that the emitting region should have a mass @xmath1 ( assuming the iron abundance similar to its solar value ) , a size of @xmath2 , is distributed anisotropically , and is moving with sub - relativistic speed .
in contrast to the fairly clean environment expected in the merging of two neutron stars , the observed line properties would imply that the site of the burst is embedded in a large mass of material , consistent with pre - explosion ejecta of a very massive star .
this material could be related with the outburst observed in the afterglow 1 day after the grb and with the spectral variations measured during this phase .
1s^-1 2cm^-2 2r^2 _ |
the over determination of the ckm matrix , using the new data from the b factories at slac and kek , is a sensitive test of the standard model @xcite .
the determination of the ckm matrix elements from experiment depends critically on hadronic parameters , such as @xmath0 , @xmath1 , @xmath2 and @xmath3 , most of which are inaccessible to experiment , but can be calculated from lattice qcd . in a recent survey of the latest lattice
qcd results for the @xmath0 ( the decay constant of the b meson ) bernard @xcite quotes : @xmath4 mev from quenched qcd and @xmath5 mev for the value of the b decay constant in full qcd .
although all lattice qcd calculations @xcite have seen an increase in @xmath0 between quenched and unquenched qcd , the effect in the world `` average '' is only at the one @xmath6 level .
the increase in @xmath0 between quenched and full qcd is more significant for an individual collaboration s results , for example cp - pacs , obtain @xcite @xmath7 .
as unquenched simulations are so computationally demanding , it seems useful to review the additional arguments that support the increase in decay constants due to the inclusion of dynamical fermions .
lattice qcd is a `` clever '' finite difference approximation to continuum qcd @xcite .
lattice qcd calculations involve computing the partition function @xmath8 where @xmath9 describe the gauge fields , @xmath10 is the lattice representation of the gauge action ( @xmath11 ) and @xmath12 is a lattice representation of the dirac operator for quarks .
the quark fields have been integrated out .
the dynamics of the gluon fields depends on the determinant of the dirac operator .
the determinant describes the dynamics of the sea quarks and is very expensive to compute . until recently
, most phenomenological lattice calculations did not include the determinant in the dynamics ( quenched qcd ) .
surprisingly , quenched qcd calculations describe experiment reasonably well .
the biggest recent study of quenched qcd by the cp - pacs collaboration @xcite found that the spectrum of light hadrons disagreed with experiment at the 10% level .
the results from an individual calculation depend on the lattice spacing @xmath13 and physical volume .
these errors may be removed by repeating the calculation at different lattice spacings and volumes and then extrapolating the results to the continuum and infinite volume limit .
for example , in the recent cp - pacs calculation @xcite they obtain @xmath0 = @xmath14 mev , @xmath15 mev , and @xmath16 mev , at lattice spacings : 0.22 fm , 0.16 fm , and 0.11 fm respectively , from a calculation of @xmath17 qcd .
cp - pacs prefer to quote @xmath18 @xmath19 mev as their continuum result , however bernard @xcite prefers to extrapolate the cp - pacs data to the continuum assuming a quadratic dependence on the lattice spacing ( there are good , but not totally rigorous arguments for this type of extrapolation ) and obtains @xmath20 mev .
this kind of ambiguity in the final analysis of lattice results is the cause of the large systematic errors in the final results of lattice calculations .
the cost of lattice qcd calculations that include dynamical fermions goes ( something ) like @xcite @xmath21 where @xmath13 is the lattice spacing , so halving the lattice spacing is very computationally expensive .
there is a simple model of the effect of unquenching that is based on the quark model @xcite .
consider the richardson heavy quark potential @xcite .
@xmath22 where @xmath23 is the number of flavours .
equation [ eq : richardpot ] or the potential extracted from a lattice calculation is used in schrdinger s equation to calculate the wave function of mesons .
the decay constant is computed using the van royen - weisskopf formulae @xmath24 in position space the potential in equation [ eq : richardpot ] at small radial separations is deeper in the full theory ( @xmath23=3 ) relative to the quenched theory ( @xmath23=0 ) . so the decay constant ( computed from equation [ eq : vrqfdecay ] ) is higher in the full theory than in quenched qcd .
there is evidence from many lattice qcd calculations of the heavy quark potential , that dynamical fermions produce a similar effect to the @xmath23 dependence of the richardson potential @xcite .
the milc collaboration @xcite have systematically studied this model in an unquenched simulation using staggered fermions . from the graphs in the paper @xcite , at @xmath25 , the effect of unquenching is 3% for @xmath26 and 7% for @xmath0 .
the milc collaboration found that unquenching was smaller on the ratio of decay constants that on individual decay constants in this model .
a key qualitative prediction of this model is that the decay constant in unquenched qcd is greater than the quenched value .
also the unquenching effect should be larger for @xmath0 than @xmath26 . another way to understand
the effect of unquenching is to use a quenched lagrangian @xcite .
the idea is similar to chiral lagrangian s , where a lagrangian is written in terms of hadron fields with the same symmetries as qcd .
a number of unknown parameters enter the chiral lagrangian , that must be fixed from experiment ( or from lattice calculations ) .
once the parameters are fixed , the lagrangian can be used to make predictions . for the case of quenched qcd ,
ghost fields are introduced to cancel the quark determinants .
the resulting theory has a different symmetry to qcd , but the idea is basically the same as for continuum chiral perturbation theory .
the formalism for heavy - light mesons was developed by booth @xcite and sharpe and zhang @xcite .
the quenching errors are estimated by comparing the chiral logs ( some of the loop effects ) in quenched and full qcd .
this is thought to be an upper bound on the magnitude of the quenching effects @xcite .
unfortunately , the estimate of the quenching errors involves 7 parameters that are hard to determine accurately .
the message from this analysis was that @xmath0 in full qcd could be either greater than or less than the quenched value .
also , it was possible that the ratio @xmath27 had large quenching errors ( this has not been found in simulations @xcite ) .
the results of the first systematic studies of decay constants in quenched qcd have been reviewed by sharpe @xcite .
sharpe concluded that @xmath28 was lower than experiment , after the continuum limit had been taken .
the recent large cp - pacs collaboration @xcite study of quenched qcd found that in quenched qcd @xmath29 ( @xmath30 ) = @xmath31 ( @xmath32 ) mev , that are smaller than experiment by @xmath33 ( @xmath34 ) .
after the continuum extrapolation , the light decay constants from quenched qcd are lower than experiment , consistent with the quark model in section [ se : model ] .
the sesam collaboration @xcite found at one lattice spacing that the pion decay constant was approximately one sigma larger with dynamical fermions , than from an equivalent quenched simulation .
the opposite trend was seen in the qcdsf / ukqcd @xcite data .
it will be difficult to determine the effect of dynamical fermions on light decay constants , until a continuum extrapolation is done . including heavy quarks ( charm and bottom ) in lattice qcd calculations
requires the introduction of new techniques , such as effective field theories for the heavy quarks @xcite . as a test of the new heavy quark methods , the @xmath35 decay constant
is computed and compared against experiment .
the value for @xmath35 quoted in the particle data table @xcite is @xmath36 mev . in table
[ tab : fdssummary ] , i have collected some results for @xmath35 from lattice gauge theory calculations ( all the various errors have been added in quadrature ) .
all the lattice results for quenched qcd are lower than the experimental value .
this is consistent with the picture that unquenching raises the value of a decay constant .
the latest results @xcite from lattice calculations show a 3 - 8% increase in @xmath35 , when dynamical fermions are included . .
summary of recent quenched lattice results for .
the @xmath6 column is the number of lattice errorbars below the central experimental value of 280 mev . [
cols="^,^,^,^",options="header " , ]
a consistent picture does seem to emerge from lattice qcd calculations , that unquenching does raise the value of decay constants .
however , the errors on the calculations need to be reduced .
ideally , a calculation of @xmath0 similar to the recent cp - pacs calculation is required @xcite , but at lighter sea masses and smaller lattice spacings .
improvements in lattice techniques and faster computers will reduce the errors on @xmath0 .
10 y. nir , 2000 , preprint hep - ph/0008226 . c. bernard , 2000 , preprint hep - lat/0011064 .
cp - pacs , a. a. khan _ et al . _ , 2000 , preprint hep - lat/0010009
. r. gupta , 1999 , preprint hep - lat/9905027 .
cp - pacs , s. aoki _ et al . _ , phys . rev .
lett . * 84 * , 238 ( 2000 ) , preprint hep - lat/9904012 . s. r. sharpe , 1998 , preprint hep - lat/9811006
. a. x. el - khadra , nucl .
proc . suppl . *
30 * , 449 ( 1993 ) , preprint hep - lat/9211046 .
f. butler , h. chen , j. sexton , a. vaccarino , and d. weingarten , nucl .
* b421 * , 217 ( 1994 ) , preprint hep - lat/9310009 . c. bernard _
et al . _ ,
phys . rev . * d62 * , 034503 ( 2000 ) , preprint hep - lat/0002028 .
j. l. richardson , phys .
b82 * , 272 ( 1979 ) .
g. s. bali , 2000 , preprint hep - ph/0001312 .
s. r. sharpe , phys . rev .
* d46 * , 3146 ( 1992 ) , preprint hep - lat/9205020 . c. w. bernard and m. f. l. golterman , phys
* d46 * , 853 ( 1992 ) , preprint hep - lat/9204007 . m. j. booth , phys . rev . * d51 * , 2338 ( 1995 ) , preprint hep - ph/9411433 . s. r. sharpe and y. zhang , phys . rev . *
d53 * , 5125 ( 1996 ) , preprint hep - lat/9510037 .
sesam , n. eicker _
_ , phys . rev .
* d59 * , 014509 ( 1999 ) , preprint hep - lat/9806027 .
ukqcd , r. horsley , 2000 , preprint hep - lat/0010059 .
d. e. groom _ et al .
j. * c15 * , 1 ( 2000 ) .
t. draper , nucl .
. suppl . * 73 * , 43 ( 1999 ) , preprint hep - lat/9810065 .
jlqcd , k .-
et al . _ ,
phys . rev .
* d61 * , 074501 ( 2000 ) , preprint hep - lat/9905036 .
c. bernard _
et al . _ ,
lett . * 81 * , 4812 ( 1998 ) , preprint hep - ph/9806412 .
s. collins _
_ , 2000 , preprint hep - lat/0007016 .
ukqcd , k. c. bowler _
_ , 2000 , preprint hep - lat/0007020 . | i review our qualitative understanding of the increase in the value of the b meson decay constant ( @xmath0 ) , when dynamical fermions are included in lattice qcd calculations . |
the experiment @xcite was designed with the goal to measure cp violation in the b - system .
it started operation at desy , hamburg in spring 2000 .
neutral b - mesons are produced by interactions of 920 gev protons on a stationary nuclear target followed by a magnetic spectrometer . since the b cross section is very low , the experiment was planed for an interaction rate of 40 mhz to obtain an acceptable b production rate .
the detectors therefore were designed to withstand the corresponding high particle rates and high radiation levels . to cope with the high particle flux which drops roughly
as one over the distance from the beam axis squared , the main tracking system has been subdivided into two parts with different rate capabilities : the inner tracker ( itr ) near the beam pipe and at larger distance the outer tracker ( otr ) consisting of drift chambers composed of honeycomb structures .
the outer dimensions of the sensitive area of the itr are given by the requirement that the occupancy per channel of the otr does not exceed a critical value of 20% .
the high intensity area can be covered by four msgc detectors with wafer size of which can be realized in industry .
the basic geometry of a msgc detector is shown in figure [ detlay ] .
the detectors have a round recess which permits to position four msgcs around the beam pipe .
they have to sustain a particle flux of up to and radiation doses of up to 1 mrad / year .
these requirements are very similar to those of the inner tracking in the atlas and cms experiments for lhc and are a big challenge for the tracking detector technology .
the total itr system requires 46 detector planes , each consisting of four chambers .
each of the 184 chambers has 752 anode strips leading to a total of about 140000 readout channels .
this basic design has remained unchanged , the detector technology , however , had to be modified several times as a consequence of r&d work and , especially , the results of beam tests . at the time of approval of the experiment in 1995 ,
the msgc technology was claimed to be ready for high rate applications @xcite .
very promising test results had been obtained and published with the claim that msgcs on ordinary borosilicate glass wafers can be operated at high rates with good stability and homogeneous response . moreover , aging tests with ar - dme ( argon - dimethylether ) mixtures and clean gas systems were reported which were interpreted to demonstrate that these detectors can survive up to of lhc operation ( equivalent to about the same time at ) without severe problems . finally , the production of large size wafers with small error rate at rather low cost was promised .
the construction of an inner tracker for based on the msgc technology therefore looked rather straight forward when the experiment was approved .
in the course of our development work we faced however several serious and unexpected problems : * the production of large size msgc wafers with acceptable quality and a small fraction of broken anodes and cathodes turned out to be difficult .
after extensive r&d at the company imt , series production of these wafers now has high yield with less than 5% interrupted anodes and no shorts .
* msgcs based on bare glass ( desag 263 ) wafers as proposed were unable to satisfy the requirements on gain homogeneity , rate stability and radiation tolerance .
these problems were solved by the introduction of `` diamond like coating '' ( dlc ) which provides a well defined electronic surface conductivity of the substrate . *
msgcs with dlc exhibit gas discharges between anodes and cathodes when operated in a hadron beam at the required high gas gain .
every discharge damages electrodes and in the long run this effect leads to an intolerable number of broken anodes .
this problem could be solved by the introduction of an additional amplification step using the gem technology @xcite . *
the msgc / gem technology is vulnerable .
we have experienced a series of problems with gas discharges at the gem which subsequently lead to damages of the msgc electrodes .
* moreover , we were repeatedly faced with aging problems due to traces of impurities which are very hard to avoid reliably during production of a large series and during long term operation . above all , the ar - dme mixture originally foreseen as counting gas in most msgc developments showed rapid aging in large area x - ray and test beam illumination of detectors .
finally , an ar - co@xmath3 mixture had to be used . in the following , major development steps and the most important results
are described .
the electrode structure chosen for the msgc ( type a ) has a pitch of , anode width of and an anode cathode gap of .
48 chambers ( type b ) located at the downstream end of the tracking have to cover a larger sensitive area .
they have a pitch of allowing to use the same readout configuration as for the smaller chambers .
the pitch is a compromise between cost of the electronics and maximum tolerable occupancy .
the relatively large anode width was chosen to minimize the number of anodes interrupted during production .
the maximum length of an anode in the experiment is .
the height of the drift gap was chosen to be 3.3 mm .
this guarantees that all primary electrons have reached the anodes within 60 ns even for slow ar - dme mixtures and that the anode signal occupies only one bunch crossing .
subsequent bunch crossings are 96 ns apart .
the high voltage is distributed via a resistor network to groups of 16 cathodes . in this way ,
a short between an anode and a cathode strip leads to an inefficient region of about 2 % of the total chamber area .
experience with first prototypes of msgcs @xcite showed that bare glass with its relatively low ionic conductivity can not be used in a high rate and high dose environment .
local gain variations up to a factor of three were found even at low counting rates .
the gas gain was rate dependent and the detectors did not survive an x - ray dose equivalent to about half a year .
glass with electronic conductivity was excluded both by cost and its short radiation length .
the obvious solution was to coat the surface with an electronically conducting compound .
many different coatings have been tested by various groups .
finally , dlc proved to be adequate .
dlc surfaces had been developed simultaneously by the cern group together with the company surmet , and by the siegen group together with a fraunhofer institut .
the two coatings are very similar in their chemical composition .
they consist of mixtures of the elements c , si , h , n deposited on the glass by chemical vapor deposition ( cvd ) using gas mixtures of ch@xmath4 , sih@xmath5 , n@xmath3 .
in fact , the coating consists mainly of amorphous carbon and has very little diamond content , if any .
the relative abundance of nitrogen and hydrogen and the production temperature determine the electric conductivity .
silicon improves the adhesion of the coating layer which has a typical thickness of 80 nm .
the resistivity of the coating can be increased by tempering the wafers .
tempering also increases the stability of the resistivity with time .
wafers coated by surmet were delivered with typical resistivity of @xmath6 @xmath7 whereas the resistivity of the fraunhofer coating was originally lower by one to two orders of magnitude .
the surmet coating had reasonable electrical properties and very good chemical stability .
the mechanical quality of the large area coating was however not satisfactory most likely due to problems in handling and cleaning .
for the main series production the dlc was therefore produced by the fraunhofer institut on af45 glass ( alkali `` free '' , produced by desag ) .
this type of glass with little alkali content was chosen to reduce the risk of electrolytic modifications of the substrate , caused by the continuous alkali ion current over several years @xcite .
the coating has an average surface conductivity of @xmath8 @xmath7 which varies over the area by about 20 % .
the lithographic production of the electrode pattern was realized by the company imt on the coated wafers of 0.4 mm thickness and outer dimensions of about . to avoid overlapping hits in subsequent bunch crossings
, we have to avoid electronic broadening of the signals propagating along the anode beyond about 50 ns .
the signal shape at the amplifier input depends on the distance of the avalanche to the amplifier end of the anode and on the product of anode - cathode resistivity and capacitance .
delta shaped voltage pulses generated at the far end of the anode produce exponentially decaying signals at the amplifier input with a time constant @xmath9 , with @xmath10 the total anode resistance and @xmath11 the anode cathode capacitance .
thus the signal width is proportional to the strip length squared . for the given anode strip length of 23 cm and anode cathode capacitance of 20 pf an anode material of resistivity below
is required .
this requirement excludes nickel and chromium and points to aluminium or gold .
aluminium is preferable for lithography and has also rather good resistance to sparks ( see section 2.4 ) .
aging tests with aluminium wafers however were discouraging ( see section 4 ) . the aluminium option was therefore abandoned .
a thick gold msgc pattern was produced by a lift - off technique .
the thickness of the different wafer materials is summarized in table 1 .
._summary of the material thickness for the different wafer layers .
[ wafermat ] _ [ cols="<,^,^,^,^ " , ] splitting of the gas amplification into two steps allowed to operate the msgc at moderate voltage .
typical voltage settings during the testing phase were @xmath12 v , @xmath13 v , @xmath14 kv which correspond to gas amplification factors of about 35 for the gem and 250 for the msgc . for comparison to reach the same amplification with a single msgc structure , one would have to apply about 100 v higher cathode voltages .
the detectors were tested intensively in the laboratory and at hadron beams at psi and in the hera - b experiment .
the gem shows the characteristic behavior of devices containing dielectric materials . after switching on the hv ,
the gas amplification of the gem rises by a factor of about two within a few hours and stabilizes after a few days , when the polyimide is fully polarized .
the gain is rather uniform .
it varies by less than 20 % over the full area of the chamber . during irradiation
the polyimide charges up and the gain increases by up to 20 % .
the gain variations with time and rate can be inhibited by coating the gem foils @xcite . since they are not critical for the operation at hera we avoided the complication , risk and cost of an additional production step . with the gem alone , gas amplification up to could be obtained with photon irradiation before sparking started .
a part of the itr chambers are located inside the spectrometer magnet with a field of 0.85 t. initial worries that the gem functionality would suffer in moderate magnetic fields were not confirmed by a test at the nominal hera - b field perpendicular to the hole axis .
no decrease of the gain was observed within 5 % @xcite .
first laboratory tests at the psi pion beam in 1998 of four msgc / gem chambers operated with ar - dme 50/50 at a gem gain of about 50 gave satisfactory results @xcite .
a second test one year later , however , was similarly disastrous as our first msgc tests at a hadron beam .
gem sparks were observed which induced secondary sparks at the microstrip structure , and sometimes led to discharges between the gem and the microstrip structure or between gem and the drift electrode .
the electrodes of the msgc were severely damaged . in extensive laboratory studies
the secondary sparking was reproduced and investigated .
the details of the complex sparking process were never fully understood .
however , it is obvious that a discharge of the gem with its high capacitance generates a huge amount of ions and electrons .
even a small fraction of the electrons drifting towards the msgc , with additional gas amplification there , are likely to produce secondary sparks .
thus the rate of those strongly depends on the strength of the transfer field .
this was confirmed by laboratory measurements presented in figure [ gemtransfer ] and it explained also the different behavior of the chambers at the two psi tests which were performed at slightly different hv settings . a significant dependence of the secondary spark rate on the cathode voltage has not been observed ( see figure [ gemtransfer ] ) . with the voltage between anodes and lower gem plane limited to below 1250 v in the hera - b experiment ,
the msgc / gem detectors were able to cope with the intense hadron flux without suffering from fatal damage by discharges ( see section 5 ) .
two gas mixtures were systematically studied , ar - dme 50/50 and ar - co@xmath3 70/30 .
the dme mixture has better quenching properties hence less problems with gas discharges .
it had also been extensively tested with respect to aging properties by several groups , which consistently reported no aging problems up to large charge depositions on the strips corresponding to many years of lhc or operation @xcite .
compared to ar - co@xmath3 70/30 it also has larger primary ionization , 24 primary electrons in average for a minimum ionizing particle in the 3 mm gas gap for ar - dme compared to 18 for ar - co@xmath3 which has also significantly larger transverse diffusion . however , there are also disadvantages of ar - dme : the ar - dme mixture is flammable and constitutes therefore a safety risk .
dme is not commercially available with guaranteed purity .
thus every bottle would have to be checked carefully in aging tests to ensure that the detectors are not polluted .
in addition , we realized that dme is absorbed by kapton leading to a reduction of the gem tension with time .
after flushing the chamber for 2 weeks , we observed a sag of a foil by 800 @xmath15 m .
the original tension could be recovered by changing the gas @xcite .
the use of dme would have required additional constructional efforts to keep the gem foil in place .
comparative sparking tests in the laboratory with @xmath16 particles in the gas indeed demonstrated that the spark protection of ar - co@xmath3 is significantly worse for the same gas gain .
this is illustrated in figure [ co2sparks ] which shows the spark rates for ar - co@xmath3 and ar - dme versus gas gain .
the result can not be safely extrapolated to the low operation voltages in a msgc / gem detector .
it is nevertheless obvious that ar - co@xmath3 offers less safety with respect to gas discharges than ar - dme .
protection against sparking can be strongly improved for ar - co@xmath3 by adding small amounts of water to the gas . however , as discussed below , the addition of water led to fast gas aging and was therefore ruled out for .
big efforts have been made by several groups to study gas aging phenomena in msgcs @xcite .
these studies are difficult and hard to reproduce because gas aging depends on a multitude of different parameters .
it was however clear from the beginning that the msgc structures are very vulnerable and that even tiny amounts of pollutions by organic materials can cause fatal deposits on the anodes .
therefore , most of the counting gases which have been used in wire chambers involving organic quenchers can not be used in msgcs for high rate applications .
moreover , we made a very careful selection of materials like composites , glues and applied elaborated cleaning and outgassing procedures to avoid pollution of the counting gas .
the gas system was entirely based on stainless steel tubing .
the sealing rings were made of kalrez , materials like teflon were avoided .
all materials were also tested for out - gassing both using gas chromatography and by exposing large surfaces of them to the input stream of counting gas used in gas aging tests of msgc detectors .
an x - ray tube with copper anode was used to produce a high flux of photons .
the x - rays were collimated by a lead collimator such that a round area of 113 mm@xmath17 was irradiated .
the detector was operated at a total gas gain of about 3500 .
the gas flow was arranged such that the gas in the detector was exchanged twice per hour .
the typical counting rate was about 100 khz / cm strip length .
one x - ray count led to an average avalanche size of about 840000 @xmath18 and this rate corresponds to a charge deposition rate per cm strip length which is about 20 times higher than that of the hottest region at .
this acceleration factor was varied between 1 and 40 for different tests but the large integrated charge depositions with illuminations over several months used acceleration factors of not more than 20 .
the position of the cu @xmath19 line was evaluated in regular intervals during these measurements to track possible changes in the gas gain .
aging tests were carried out with a large number of detectors over a period of three years . with the full size msgc / gem pre - series detectors we observed neither for ar - dme nor for ar - co@xmath3 any significant change of the pulse height up to an integrated charge of 45 mc / cm which corresponds to about 6 years of operation .
this result was supported by optical inspection of the irradiated areas which showed no sign of depositions on the anodes .
this result was interpreted as sufficient reassurance for our choice of counting gas and constructional materials to start mass production of the detectors in spring 1998 .
it should be noted that our result was in line with similar measurements of other groups @xcite .
all their results were obtained with small irradiated areas of sometimes only 7 mm@xmath20 . in autumn 1998
a beam test was carried out at psi with two full size msgc / gem detectors using ar - dme 50/50 as counting gas .
the goal was twofold : during daytime the pion beam was used to study efficiencies , resolution and the performance of our trigger chain . during night time the detectors were exposed to an intense proton beam to accumulate a higher charge dose .
the beam spot had a gaussian profile with a typical full width at half maximum of . within few days a total accumulated charge of was reached in the center region .
we observed a rather fast reduction of the drift current during proton irradiation for constant beam current .
this was even seen online as shown in figure [ psicurrent ] for a 12 hour period of proton running at maximal intensity ( about 20 times charge deposition rate ) .
the detectors were subsequently investigated in the laboratory . within the illuminated area the detectors still showed acceptable gas gain but outside the illuminated area they did not count at all .
optical inspection showed severe deposits on those anodes which were connected to ground and therefore experienced the nominal electrical field .
disconnected anodes showed no deposits .
depositions were found both in the irradiated area and over the rest of the detector .
parts of an irradiated wafer were sent to the fraunhofer institut fr schicht- und oberflchentechnik which carried out a surface analysis shown in figure [ fraunhoferanalyse ] .
the anodes in the irradiated area were covered by a layer of pure carbon .
a picture of these deposits as seen with an electron microscope is shown in figure [ psideposit ] . the areas outside , which were not irradiated were covered by an insulating layer of about composed of hydrogen and carbon .
this unexpected result prompted a series of new laboratory aging tests .
the final outcome is very surprising .
the ar - dme gas mixture shows gas aging under x - ray irradiation if the area which is irradiated is large enough .
this fact is illustrated in figure [ dmeaging ] which shows the gas gain versus irradiation time for a collimator with an area of 113 mm@xmath20 as used before compared to one with an area of .
the large area was irradiated at half the local intensity of the smaller one .
identical chambers operated with ar - co@xmath3 show no aging independent of the size of the irradiation area . apparently , irradiation of a large area with ar - dme as a counting gas leads to fast aging and visible deposits on the anodes .
these deposits are not limited to the irradiated area .
we therefore have to draw the conclusion that gas aging effects depend strongly on the size of the irradiated area .
all evidence supports the assumption that the origin of these deposits is the dme itself .
we have no explanation for this experimental fact but it invalidates all aging tests done before .
subsequently ar - co@xmath3 was chosen as the main candidate for the counting gas .
as this gas mixture has low primary ionization and large transverse electron diffusion leading to large cluster width in the detector , we tested several other gas mixtures without hydrocarbons .
these were ne - co@xmath3 , kr - co@xmath21 and he - co@xmath3 mixtures .
the results were unfortunately discouraging .
measurements with msgcs showed that for the same avalanche charge on the anodes these gases led to comparable cluster widths but poorer performance with respect to discharge protection .
we therefore chose the ar - co@xmath3 70/30 gas mixture .
as explained in section 3 , the safety against gas discharges became a major concern .
we therefore tested different combinations of electrode materials and counting gases which promised better protection .
the operation characteristics of a msgc can be significantly improved by adding small amounts of water , typically 0.3% , to the counting gas .
this avoids discharges near insulators e.g. near the detector frames and anode ends and improves the overall protection against sparking , probably by reducing the surface resistivity of plastic materials which tend to absorb water .
addition of water to the counting gas shows however very severe and fast gas aging for both , ar - dme and ar - co@xmath3 .
this is shown in figure [ wateraging ] which compares the pulse height as a function of accumulated charge for ar - dme with and without an admixture of water .
for the gas with water admixture the pulse height drops by a factor three for an accumulated charge of less than .
moreover the anodes show deposits in the irradiated area .
we also tested msgcs with aluminium electrodes .
aluminium electrodes are easier to manufacture and sparking tests showed that the electrodes are more robust against discharges than gold but aluminium is known to introduce aging in msgcs @xcite .
our aging tests with aluminium msgc / gem detectors confirmed that this material can not be used with ar - dme or with ar - co@xmath3 . in both cases
we observed a fast reduction of the pulse height by more than 20% for an accumulated charge of only .
moreover , the electrodes were severely damaged .
especially the cathodes showed bubbles and craters in the irradiated area .
a rather interesting phenomenon was observed for very long irradiation times for chambers operated with gas mixtures without indication for gas aging like ar - co@xmath3 and ar - co@xmath3-cf@xmath5 which is used in the outer tracker of hera - b . after a stable period corresponding to about one year
these chambers show a strong increase of pulse height with time of up to a factor four . for some intermediate time
the gas gain becomes very inhomogeneous , changing locally by more than a factor two . after
an exposure corresponding to about five years of operation the gain becomes stable again but is about four times higher than at the beginning .
the gain in this state depends strongly on the counting rate which indicates that the surface near the anodes has become insulating such that charging - up affects the gain .
the origin of these effects was clarified by a detailed surface analysis .
the intense plasma which is created during the aging tests destroys the dlc layer near the anodes .
it is gradually etched away up to the point that there is an insulating strip near the anode edges .
this is illustrated in figure [ diamondaging ] which shows the measured thickness of the coating as a function of the distance between anode and cathode for an area where a charge of was accumulated .
the coating is etched away almost completely near the anode and is reduced over the whole gap compared to the starting thickness of about 80 nm . with ar - co@xmath3
this effect is relatively slow under the conditions of our aging tests , for a gas mixture including cf@xmath5 the effect is much faster such that the coating would be destroyed completely for a collected charge of only .
this gas can therefore not be used .
the observed etching effect is of course likely to depend strongly on the local current and plasma density and could be very different for the running conditions of the experiment .
it is therefore impossible to predict if and how fast such a change would appear at .
the final beam test at psi in 1999 using low energy protons was used to accumulate a deposited charge equivalent to 1/3 year of operation . during this relatively small exposure no signs of gas aging or damage of the dlc were observed .
during production , the different parts of the itr went through several tests : the wafers were checked at zurich directly after production ( see section 2.2 ) .
interrupted anodes were recorded .
all gems were tested in heidelberg before chamber construction .
the gems had to sustain 500 v in a nitrogen atmosphere for 12 hours .
the individual chambers were transported to siegen for bonding of the electronics . at siegen
they were flushed with the final chamber gas and put to high voltage . within a period of 24 hours
the voltages were ramped up to 520 v for the cathode and to 410 v across the gem .
all currents were recorded and put into a reference file .
about half of the chambers showed anode cathode shorts .
these shorts were eliminated by disconnecting the anode . in 10 chambers
fatal gem shorts were observed .
chambers fulfilling the requirements went back to heidelberg .
there the stations were put together and fully assembled stations were shipped to desy . before installation another high voltage test similar to the one at siegen was performed .
the required voltages were 490 v for the cathodes and 400 v across the gems .
in 12 out of 144 tested chambers an anode cathode short was detected .
one chamber showed a gem short .
apparently , each construction or transportation step introduced some damage to part of the detectors and the intermediate high voltage tests could not exclude additional problems occurring later in the hadron beam of hera - b . during the period from july 1999 to may 2000
a total of 150 chambers were installed and operated at .
the first few months of the year 2000 were mainly devoted to commissioning of the detectors .
data were routinely taken from april to the end of the running period by august 2000 .
the maximum operation time of 1100 hours was reached by 56 chambers . during the last two months 90 % of the installed chambers were routinely participating in data taking without major problems . the interaction rate at the target varied between 3 mhz and 40 mhz .
most of the data were taken at 5 mhz . a typical distribution of the hit density in a detector is shown in figure [ hitdist ] for 5 mhz interaction rate .
the hit density is about 0.006 per anode and event .
the shape of the distribution is determined by the beam recess of the msgc wafer ( see fig . 1 ) and the radial decrease of particle density .
the raw pulse height distribution can not be used to determine the gas gain because it is dominated by soft signals which do not belong to tracks .
we therefore need to reconstruct tracks first to measure a signal over noise distribution and the efficiency . since at the beginning of the run the track reconstruction program was not yet well tuned , it was possible only for the chambers in front of the magnet to estimate during running the efficiencies which have to be known in order to adjust the high voltage .
typical voltage settings were -510 v for the cathode and 420 to 450 v for the gem .
individual gem voltage settings were necessary to correct for the gain variations of up to a factor 2.5 from gem to gem which are caused by slight alignment imperfections of the copper holes at the two sides of the gem .
the high voltage settings for the remaining chambers could not be adjusted and therefore the chambers were operated at gem voltages below 440 v and correspondingly more moderate gain .
while at the end of the data taking period running was rather smooth , several problems occurred after installation : * four gems developed conductive paths between their two copper surfaces .
the resistivity dropped to values of the order of m@xmath22 , values which effectively short the two gem electrodes .
* in 43 % of the chambers the msgc electrodes showed one or two short circuits between anode and cathode strips .
these problems occurred shortly after installation and were probably due to spurious dust or production defects .
sparking in gems can not be avoided completely , but under normal conditions the rate is quite low , of the order of a single spark per day ( see ) , a rate which is tolerable .
only six chambers showed sparking at rates above ten per day . the spark rate increases exponentially with the gem voltage rising from 1 to 5 sparks / day when the gem voltage is increased from 400 v to 450 v. this is shown in figure [ sparkvsgem ] for chambers behaving normally .
chambers with exceptional high rates are excluded in this figure to avoid a biased picture . as expected , the sparking rates increased with the beam intensity .
similar results have been reported in refs .
@xcite where a linear dependence of the sparking probability on the beam intensity and a strong dependence on the beam composition has been observed . while under normal conditions the spurious sparking does not disturb the performance of the detectors , in some cases continuous sparking produced gem `` shorts '' by carbonizing the polyimide surface in the hole . in some cases
when the resulting conductivity across the gem was low , the `` short '' could be burned away by applying a short current pulse which evaporates the copper near the affected hole .
when this was not possible , the whole chamber was lost . at one occasion , the damage of a gem could be correlated to a spike in the target rate .
a total of 79 anode cathode shorts were observed during the run period in 2000 . due to
the hv grouping of the cathodes this corresponds to a loss of 1.1% of all anode strips . out of the total
, 29 shorts were produced already within the first 10 hours after switching on and only 16 occurred in the second half of the running period .
most of the late shorts coincided with an increase in the high voltage .
inspection of the chambers during the hera luminosity upgrade shut - down revealed that some shorts are correlated to anodes with lithographic defects .
interrupted anodes show a tenfold higher probability to produce a short than perfect anodes .
interrupted anodes have now been disconnected electrically from ground to avoid this problem .
the corresponding loss in efficiency is completely negligible .
most problems occurred at the beginning of the operation in the beam .
they can be avoided or at least reduced to a tolerable level by the following measures : * operating the detectors for some days at overvoltage before installation , * carefully training the chambers in the beam before applying the full high voltage , * hv control based on monitoring of the currents and on spark detection , * applying a coordinated ramping of the different chamber high voltages , * avoiding unstable beam conditions . during the 2000 running period a training procedure has been worked out which has led to a considerable improvement of the reliability of the detectors .
the voltages are raised slowly in ten steps over a period of two to four weeks ( 400 hours beam time ) depending on the chamber behavior .
very few problems were observed with chambers treated in this way .
a control system based on microprocessors inside the high voltage distribution boxes has been optimized .
it continuously monitors the drift and cathode currents and the gem voltages .
in addition , it detects sparking through rapid voltage changes .
when a spark occurs , the voltages are decreased thereby avoiding sequential sparks . after some delay
, the voltages are automatically ramped up again .
the behavior of all chambers is continuously monitored online .
meanwhile all damaged chambers have been exchanged .
cathode groups have been re - activated by cutting the anode responsible for the short .
extrapolation from the experience gained last year indicates that we should be able to operate the itr with less than 1% of channels lost per year .
the detectors show a noise distribution with a mean value of 2500 electrons .
this value is well compatible with the estimate from the strip capacitance and the amplifier input resistor value .
a signal over noise ( s / n ) plot is shown in figure [ s_ndist ] .
efficiencies of the chambers in front of the magnet could be determined during running using well reconstructed tracks extrapolated from the vertex detector .
they are presented in figure [ effhera ] as a function of the gem voltage .
the efficiencies are above 90 % which is sufficient for tracking .
as mentioned above , there are considerable variations in the gem performance from chamber to chamber which are compensated by adjusting the gem high voltage which can be individually set for each chamber .
the local gain variations across a single chamber , however , are quite small , namely below 10 % .
the strip multiplicity distribution shown in figure [ stripmult ] corresponds to a mean value of 2.6 for a threshold of 5000 electrons , which is twice the average noise level .
this rather high value is compatible with the large transverse diffusion in ar - co@xmath3 along the long drift path of the electrons .
the itr chambers have been geometrically aligned using an iterative fitting method .
a residual distribution for fitted tracks of particles with momenta between 35 and 50 gev / c is presented in figure [ residuals ] . in average 15 chambers contribute to one track . from a comparison of the observed root mean error of with a monte carlo simulation
we conclude that the intrinsic spatial resolution of the detectors is better than .
after many unexpected difficulties the development of the itr for has led to a detector which is working to the required specifications .
efficiencies greater than 90% were achieved for part of the chambers and are expected for the full detector during the next running period .
extrapolation of the present experience indicates that inefficiencies due to anode cathode shorts will remain below the 1% level .
irradiation with x - rays have revealed that aging for ar - dme gas mixtures depend strongly on the size of the irradiated area .
measurements with ar - co@xmath3 , 70/30 did not show the same behavior and indicate that our detectors would adequately perform for more than six years .
irradiation also affects the dlc which is etched away starting from the anode borders .
this leads to gain variations with time and beam intensity .
the msgc with the dimensions necessary at and the beam conditions of hera can not be operated reliably without the addition of a gem . at the gas amplification required to obtain sufficient efficiency
the spark rates would be lethal .
the situation here is different from that at lhc where with much smaller detectors simpler solutions would have been possible @xcite .
our system is quite fragile and expensive and requires a complicated hv steering . with the knowledge gained in the last few years ,
it has become clear that new , more robust devices like multi gem detectors @xcite are better suited for harsh beam environments than msgc based systems .
we acknowledge the strong effort of the technical stuff of the collaborating institutions .
in particular , we wish to thank s. henneberger , r. rusniak , a. rausch , ch .
rummel , d. gieser , r. eitel and s. rabenecker ( heidelberg ) , o. meyer , g. schmidt ( siegen ) , k. boesiger , k. esslinger , b. schmid and s. steiner ( zuerich ) , serguei cheviakhov ( desy ) .
our work profited very much from the close collaboration with the vertex detector group of the mpi heidelberg , especially in the areas of readout and monitoring . | the results of five years of development of the inner tracking system of the hera - b experiment and first experience from the data taking period of the year 2000 are reported .
the system contains 184 chambers , covering a sensitive area of about ( 20 @xmath0 20 ) @xmath1 each .
the detector is based on microstrip gas counters ( msgcs ) with diamond like coated ( dlc ) glass wafers and gas electron multipliers ( gems ) .
the main problems in the development phase were gas discharges in intense hadron beams and aging in a high radiation dose environment .
the observation of gas discharges which damage the electrode structure of the msgc led to the addition of the gem as a first amplification step .
spurious sparking at the gem can not be avoided completely .
it does not affect the gem itself but can produce secondary damage of the msgc if the electric field between the gem and the msgc is above a threshold depending on operation conditions .
we observed that aging does not only depend on the dose but also on the spot size of the irradiated area .
ar - dme mixtures had to be abandoned whereas a mixture of 70% ar and 30% co@xmath2 showed no serious aging effects up to about 40 mc / cm deposited charge on the anodes .
x - ray measurements indicate that the dlc of the msgc is deteriorated by the gas amplification process . as a consequence
, long term gain variations are expected .
the inner tracker has successfully participated in the data taking at hera - b during summer 2000 .
msgc , msgc - gem , gas aging , discharges , hera - b 29.40.g , 29.40.c , 07.85.f , 81.40.c , 52.80 |
the origin for the dramatic slowing down of the relaxation dynamics of a liquid close to the glass transition is still a matter of debate @xcite .
different theoretical approaches to describe the dynamic transition from the liquid to the glassy state have been presented so far .
some concentrate on thermodynamic aspects and propose a first or second order phase transition ( e.g. free volume theory @xcite or entropy models @xcite ) , while others , such as mode - coupling theory @xcite , focus on the microscopic dynamics that shows non - linear feedback effects which are responsible for the slow dynamics .
many of these theories are using in some form the concept of cooperative motion within a liquid . as already mentioned by kauzmann @xcite it is reasonable to assume that in a liquid neighboring particles can not move independently of each other , in particular at low temperatures .
instead one must expect that the motion of two particles that are separated by a distance that is short is cooperative , i.e. there exist cooperatively rearranging regions ( crr s ) within the liquid .
however , how this cooperative motion is realized microscopically is not that clear .
e.g. one can discuss the motion of the particles by means of the so called cage effect @xcite : a liquid particle is surrounded by its nearest neighbors , which form a cage . at low temperatures this cage is quite rigid and therefore at short and intermediate times a particle will mainly rattle around within this cage . for longer times the cage will open up and allow the particle to escape . with decreasing temperature
this opening will take more and more time , and hence the relaxation dynamics of the particles will slow down .
the cage effect already includes the concept of cooperativity since other particles have to move as well before the cage breaks up and the trapped particle can escape . under the assumption that there is some kind of cooperativity within the liquid
the next question is whether the size of such crr s does indeed increase with decreasing temperature , as proposed in several models for the glass transition @xcite .
usual thermodynamic second order phase transitions , who also show a strong slowing down of the dynamics , are accompanied by the divergence of some static length scale . while in spin glasses higher order correlation functions
do show a divergence @xcite , no diverging _
static _ length scale has been found so far in structural glasses .
the analysis of two point correlations such as the structure factors or radial distribution functions shows that the length scales involved are hardly growing with decreasing temperature . in particular
, the structure of a glass is almost identical to the one of a liquid . on the other hand
there has been evidence for the existence of _ dynamic _ length scales which grow much faster , although evidence for a divergence at some finite critical temperature has been given only very recently @xcite .
e.g. computer simulations @xcite and measurements on colloidal systems @xcite have shown that the dynamics is heterogeneous .
it was found , that the fastest particles within the system are not distributed randomly but tend to form clusters .
the size of these clusters grows on lowering the temperature although by a factor that is much smaller than the increase found in typical structural relaxation times @xcite .
furthermore the precise connection between heterogenous and cooperative dynamics is a question that is still open . since in most experimental setups it is not possible to access directly the dynamics of the individual particles , people have proposed and used methods to study the cooperative dynamics in supercooled liquids in an indirect way
the idea is to investigate _ confined _ systems and thus to introduce a new length scale , the thickness of the film or the diameter of the pore , and to see how this scale influences the relaxation dynamics .
if crr s do exist and grow with decreasing temperature , the dynamics should differ from the bulk behavior as soon as the size of the crr s at a given temperature becomes comparable to the system size @xcite .
one large class of experiments are done on molecular liquids ( such as salol , glycol , ortho - terphenyl ) confined to porous material such as vycor glass , sol - gel glass and zeolithes with pore sizes ranging from several hundred to only a few nanometer .
typical experimental methods are differential scanning calorimetry ( dsc ) @xcite , dielectric spectroscopy @xcite , neutron scattering @xcite , solvation dynamics @xcite and many more .
as expected , most experimental data show a deviation from bulk behavior below a certain threshold temperature .
the main problem with these experiments is the correct interpretation of the results , which is not at all universal @xcite .
depending on the investigated liquid , the realization of the confinement and even the experimental method , there is a great variety in experimental findings .
some give evidence for an accelerated dynamics in confinement @xcite , i.e. the glass transition temperature @xmath3 goes down , while others see an increase of @xmath3 @xcite . in many cases in which the dynamics is slowed down ,
experimental spectra suggest the existence of a secondary relaxation process and often the interpretation of this result is based on the interaction between liquid and surface .
the idea is that there is an immobile particle layer which is bound to the surface ( slow process ) and the remaining liquid is then often accelerated with respect to the bulk @xcite .
a systematic study of supported and free standing polymer films by ellipsometry @xcite supports this interpretation . while free standing films show a systematic decrease of @xmath3 with decreasing thickness , the dynamics is partially slowed down in the presence of an attractive surface ( e.g. in supported films ) .
the experimental data can be described by a simple three layer model @xcite , where one distinguishes between surface layers for supported and free interfaces and a liquid like layer at larger distance from the interface .
other general trends found in experiments include a broadening of the @xmath4relaxation peaks in dielectric @xcite but also in neutron scattering @xcite experiments .
as mentioned above , an important input parameter , which however is hard to control in many experiments , is the interaction between surface and liquid and the local structure .
if the liquid is bound to the surface the mobility is of course reduced and due to the existence of cooperativity this slowing down could also propagate further into the system .
the opposite effect is seen , if the particle mobility is enhanced close to the interface .
furthermore it is also important to know the precise geometry of the confining material .
e.g. in vycor glass one could think of a distribution of different pore sizes which leads to a superposition of relaxation processes on slightly different time scales which could explain the broadening of the relaxation peak for the whole system . since in an experiment
only the _ average _ particle density can be controlled , it is always possible that the _ local _ structure shows strong deviations from bulk behavior , such as , e.g. , density oscillations .
lower local densities should of course accelerate the dynamics and vice versa . that such oscillation in the density do indeed occur is a result of many computer simulations of confined liquids @xcite .
depending on the investigated liquid and the surface interaction , the amplitude of such oscillations can be very large @xcite , but there are also situations where the density is almost constant @xcite .
thus it is not very clear so far whether the change of the relaxation dynamics of confined systems is a real effect of the confinement , or whether it is just a `` secondary '' effect , i.e. due to the interaction of the liquid with the surface and hence the resulting change in structure and thus in the dynamics .
since in principle computer simulations offer the possibility to control the most important input parameters such as the geometry of the confinement and the interactions with the surface , and furthermore allow to investigate the relaxation dynamics also on a local scale , they are an ideal method to study the dynamics of confined systems and in the present paper we report the results of such a study . in our simulation
we realize a situation where the static properties are _ identical _ to the one in the bulk and therefore the dynamics is not influenced by `` secondary '' effects coming from a different structure .
thus this helps to obtain a better understanding of the relaxation dynamics of such confined systems and also to interpret some experimental findings . in the following we
first introduce a model for the investigated liquid and the implementation of a rough wall with an amorphous structure .
subsequently we investigate the dynamic properties using as observables the mean squared displacement and the intermediate scattering function .
we characterize the influence of the wall on the local particle dynamics and extract length scales of cooperativity from the structural relaxation times .
finally we discuss finite size effects which show up only at low temperatures or for very thin films , and analyze the average relaxation dynamics , i.e. the only quantities that are accessible in a real experiment .
we investigate a binary mixture of 80% a and 20% b particles that interact via a lennard - jones ( lj ) potential @xmath5 $ ] ( @xmath6 ) .
the interaction parameters are chosen as @xmath7 , @xmath8 , @xmath9 , @xmath10 , @xmath11 , and @xmath12 . the potential is truncated and shifted at cut - off radii @xmath13@xmath14 . within this paper
all results will be given in reduced lj units , i.e. length in units of @xmath15 , energy in units of @xmath16 , and time in units of @xmath17 .
the introduction of two different length scales for the size of the particles strongly suppresses the crystallization of the liquid at low temperatures .
previous simulations in the bulk have shown that this system is a very good glass - former , i.e. is not prone to crystallisation @xcite . as a reference for the relevant temperature scale
we note that the system is in a normal liquid state for @xmath18 , that the critical temperature of the mode - coupling theory @xcite is around 0.435 , and that the kauzmann temperature is around 0.3 @xcite the equations of motion have been integrated with the velocity form of the verlet algorithm , using a step size of 0.01 and 0.02 for @xmath19 and @xmath20 , respectively . during
these runs the accessible volume to the particles was keep constant . as already mentioned in the introduction , it is most important to be able to decide whether the change in the relaxation dynamics is due to the _ confinement _ or whether it is related to the _ interaction _ of the fluid particles with the wall , i.e. a change of the local structure close to the wall . since in the present work
we want to study only the influence of the confinement on the dynamics , we have to avoid a change of the structural properties due to the confining walls . in ref .
@xcite we presented a simulation of a liquid confined in a tube in which the positions of the wall particles were given by a frozen configuration of the same liquid at a _ fixed _ temperature @xmath21 ( which corresponded to a temperature at which the bulk liquid was slightly supercooled ) .
since the structure of the wall was thus almost identical to the enclosed liquid in its bulk phase , the difference of the structure of the confined liquid from the one of the bulk was fairly small . nevertheless , due to the slightly different structure , the confined liquid was not quite in equilibrium after the introduction of the walls and hence it was necessary to equilibrate the confined system . due to the strong slowing down of the confined liquid close to the wall , this equilibration took a very long time and hence it was possible to access only intermediate temperatures , i.e. the relaxation dynamics of the confined system at low temperatures could not be studied .
it is , however , possible to avoid this problem by setting the `` wall temperature '' @xmath21 , i.e. the temperature at which this configuration was equilibrated , identical to the temperature of the contained liquid @xcite .
in contrast to the simulation of ref .
@xcite the wall structure is now also a ( weak ) function of temperature .
furthermore it is necessary to add a hard core potential directly at the interface in order to prevent that the fluid particles penetrate into the wall . in the following
we will briefly demonstrate that with this setup the structure of the confined liquid is indeed identical to the one of the bulk system .
although we will restrict ourself to the situation in which the liquid is confined between two parallel plane walls , the derivation can easily be expanded to an arbitrary confining geometry . due to this fact it is hence not necessary to equilibrate the liquid and thus the above mentioned problem is avoided .
consider a static variable @xmath22 where @xmath23 is a @xmath24-dimensional vector that characterizes the position of all particles in a bulk system . in the canonical ensemble the expectation value of the observable @xmath25
is given by : @xmath26 \ , , \label{eq1}\ ] ] where @xmath27 is the total potential between all particles and @xmath28 $ ] the partition function .
the configuration average in eq .
( [ eq1 ] ) can be split into contributions that come from domains in which the particles are inside the fluid region , denoted by @xmath29 , and outside of it ( denoted by @xmath30 ) .
( note that at this stage this division is purely formal . ) : @xmath31 \ldots \left [ \int_f { \rm d}{\bf r}_n + \int_w { \rm d}{\bf r}_n\right ] \ , x({\bf r } ) \
, \exp \left [ -\beta u({\bf r } ) \right ] .
\label{eq2}\ ] ] if we now relabel the particles such that the ones that are in the wall - domain have index @xmath32 , with @xmath33 , and thus the remaining @xmath34 particles are in the fluid - domain , we can write @xmath35 as follows : @xmath36 . \label{eq3}\ ] ] in the following we will denote by @xmath37 the @xmath38 dimensional vector that describes a configuration of @xmath39 particles that characterize a given wall . the canonical average of an observable @xmath25 that depends only on the liquid particles in the fluid domain and that is confined by a wall given by the @xmath39 particles at position @xmath37 is given by : @xmath40 .
\label{eq4}\ ] ] here @xmath41 is the potential energy of the system if the position of the first @xmath39 particles are given by @xmath37 , and @xmath42 is the partition function for the liquid system with walls @xmath37 .
( note that here we have made use of the fact that we have at the interface a hard core potential and that thus the domain of integration is restricted to @xmath29 . ) to determine the expectation value of an observable @xmath25 for a confined system , one has to average over all realizations of the walls : @xmath43 where @xmath44 stands for the sum over all possible wall configurations with @xmath39 particles and @xmath45 is the statistical weight of such a configuration .
this weight is just given by the trace of the boltzmann factor over the fluid particles : @xmath46 .
\label{eq6}\ ] ] by means of eqs .
( [ eq3])-([eq6 ] ) it is now easy to see that we have @xmath47 , i.e. that by averaging over all possible walls the static properties of the confined system are the same as the one of the bulk . in practice
we proceeded as follows to determine the dynamical properties of the confined liquid in a film geometry with thickness @xmath48 : first we equilibrated a rectangular ( @xmath49 ) bulk configuration with @xmath50 .
to create the film of thickness @xmath48 ( in @xmath1direction , @xmath51 ) we introduced two rigid walls , at @xmath52 and @xmath53 .
the configuration for the wall particles is given by the periodic images of the same configuration which are translated in @xmath1direction by @xmath48 and @xmath54 .
since we use cut - off radii , only wall particles within a slice of thickness @xmath55 have to be taken into account . the liquid interaction and also the average particle density of @xmath56
is identical to the bulk simulations that have been done earlier @xcite .
most of the simulations have been done for @xmath57 , although thinner films have been considered as well ( see below ) .
( note that for @xmath57 the given density corresponds to 3000 particles . )
we have found that at low temperatures the relaxation dynamics shows strong finite size effects , see section [ sec3f ] .
therefore we restricted our investigations to the temperature range @xmath58 for which such effects are not present .
the remarkable advantage of the just discussed method to confine the liquid is that we only have to equilibrate bulk systems where relaxation times are orders of magnitude smaller than in films with rough surfaces @xcite .
of course the production run has to be much longer than the equilibration runs in order to investigate the structural relaxation also close to the surface .
a remaining problem is the realization of the average over the different walls , see eq .
( [ eq5 ] ) . in practice
we calculated this mean by averaging over 16 independent walls , a number that , for the system sizes used in the present simulation , turned out to be sufficient .
in particular we found that the average particle density as a function of @xmath0 is a constant , as it has to be for a homogeneous system and that structural quantities like the partial structure factors are identical to the ones in the bulk @xcite .
since we already know from previous simulations that the dynamic properties will strongly depend on the particle distance @xmath0 from the wall @xcite we will concentrate on observables that take into account this spatial dependence . in particular we will investigate the single particle dynamics such as mean squared displacements and the self part of the van hove correlation functions .
subsequently we will use the decay of density fluctuations to extract characteristic relaxation times and study their temperature dependence .
finally we will study collective quantities and dynamic properties averaged over all particles .
one very useful observable to characterize the dynamics of a tagged particle is the mean squared displacement ( msd ) @xcite . since this function
is expected to depend on @xmath0 we generalize the usual definition of the msd in the following way : @xmath59 thus @xmath60 corresponds to the dynamics of particles of type @xmath61 which at time @xmath62 had a distance @xmath0 to one of the walls .
( note that for @xmath63 the particle will change its distance from the wall , i.e. @xmath64 for @xmath63 . )
furthermore we will see that it is useful to distinguish between displacements in different directions , namely @xmath65 perpendicular to the wall and @xmath66 for displacements parallel to the film plane . to calculate @xmath60 we divided the film into 10 slices ( of thickness @xmath67 ) and averaged over the two slices that have the same distance to one of the walls .
parallel component of the mean squared displacement @xmath68 for a particles in layers with different distance to one of the walls ( @xmath69 ) at @xmath70 and @xmath71 .
the bold dashed lines are the bulk data at the same temperatures.,width=264 ] in fig .
[ fig_msd_p_center_vs_wall ] we show the time dependence of @xmath68 for two temperatures and the five values of @xmath0 ( see caption for details ) .
also included in the figure are the corresponding curves for the bulk ( bold dashed lines ) and in order to understand the data for the confined system it is useful to start the discussion with their time dependence . at short times
the bulk curves show a @xmath72 dependence since on that time scale the particles move just ballistically . at high
@xmath73 this time dependence crosses over immediately to the diffusive regime , i.e. the msd increases linearly with time .
the same short and long time behavior is found also at low @xmath73 .
however , we see that at these temperatures the two regimes are separated by a time - window in which the msd increases only slowly .
the physical reason for this behavior is the so - called `` cage - effect '' , i.e. the fact that at intermediate times the particles are temporarily trapped by their surrounding neighbors .
this time window is often also called `` @xmath74relaxation '' @xcite .
the curves for the confined system for particles that are in the center of the film ( thin lines ) are basically identical to the one of the bulk system , showing that for large distances from the wall the relaxation dynamics is not changed .
this is in contrast to the dynamics of the particles close to the walls , dotted lines , which seems to be significantly slower than the one of the bulk . already at high temperatures we see that the msd has a weak plateau at intermediate times , thus demonstrating that the cage - effect is already present
this becomes much more pronounced if the temperature is lowered .
e.g. we find that at @xmath71 the msd for the smallest @xmath0 shows a plateau that extends over several decades in time and only at the very end of the run the particles start to show a diffusive behavior .
( note that at very long times the msd for the different values of @xmath0 must come together since ultimately the system will be ergodic and hence the dynamical properties of a particle can not depend on its location at time zero .
however , for the lowest temperatures considered these `` mixing '' times are well beyond the time scale of our simulation . ) thus we conclude from the figure that the mobility of a particle at the surface is strongly suppressed and that the typical time scale to leave its cage is orders of magnitude larger than the one of a particle in the bulk .
anisotropy of the mean squared displacement for a particles at @xmath70 and @xmath75 .
displacements @xmath76 in film plane and @xmath77 perpendicular for particles with @xmath78 ( center of film , curves close to bulk data ) and @xmath79 ( region close to the surface ) .
also shown are the isotropic bulk data ( bold dashed lines).,width=275 ] the curves shown in fig .
[ fig_msd_p_center_vs_wall ] are the component of the msd that is parallel to the confining wall .
it can be expected , however , that the relaxation dynamics is not isotropic and that therefore it is of interest to investigate also the component of the motion perpendicular to the wall .
this is done in fig .
[ fig_msd_isotropy ] where we show for two temperatures and two values of @xmath0 the components @xmath68 and @xmath65 . from this graph
we see that for @xmath80 , i.e. in the center of the film , the parallel and perpendicular component are the same and that they are identical to the one in the bulk ( bold dashed lines ) . for @xmath81
this is , however , not the case in that both components show a significantly slower @xmath82dependence than the one of the bulk .
furthermore we see that the relaxation in @xmath0 direction is at intermediate time scales slower than the one in @xmath83 direction .
also the size of the cage in @xmath0 direction is significantly smaller than the one in @xmath83 direction , as can be recognized by the fact that the height of the plateau at intermediate times is smaller for @xmath65 than the one seen in @xmath68 .
this observation is related to the fact that the particles of the wall are completely frozen , i.e. do not oscillate around an `` equilibrium '' position , and hence are not able to make space for the vibrations of the fluid particles .
in order to investigate the relaxation dynamics of the particles in more detail it is useful to investigates not only their averaged displacements but their distribution at different times , i.e. to study the self part of the van hove correlation function @xcite . also here
we have to generalize this function in order to take into account its dependence on @xmath0 : @xmath84 ) \delta(z_k(0)-z ) \rangle \ , .
\label{eq8}\ ] ] in order to improve the statistics of our data , we do not distinguish between the different orientations of the vector @xmath85 and instead average over the angular dependence of @xmath86 .
( see ref .
@xcite for a discussion of this angular dependence . ) the @xmath87dependence of @xmath88 for the a particles is shown in fig .
[ fig_vanhove_layerdependence ] for three characteristic times . at the first time , @xmath89 , the particles are still in the ballistic regime and thus the distribution is independent of @xmath0 , see fig .
[ fig_vanhove_layerdependence]a .
figure [ fig_vanhove_layerdependence]b shows that for a time in which the bulk system is in the @xmath74relaxation regime , the van hove function does depend strongly on the distance from the wall . for large values of @xmath0 , and
hence also for the bulk system , the distribution has a peak around 0.3 and non - negligible values even for @xmath90 larger than 1.0 , i.e. the distance that corresponds to the nearest neighbor distance .
this means that these particles have left the cage in which they were at time @xmath62 .
in contrast to this , the curves for small values of @xmath0 are still localized at distances around 0.1 , i.e. the particles close to the wall have not yet left their initial cage . only at very long times
a substantial number of particles at the surface start to leave their original cage , see fig .
[ fig_vanhove_layerdependence]c . from the presence and the nature of the time evolution of the secondary peak in the distribution function ( located around @xmath91 )
we conclude that the responsible mechanism is a jump type of motion , i.e. these particles do not move continuously but hop to a new position that has a typical distance 1.0 from the initial site . in the bulk system
such a hopping motion has been observed only for the b particles and even this only at very low temperature @xcite . in general such type of relaxation dynamics
is more characteristic for strong glass formers , such as , e.g. , silica @xcite . in these systems
the typical `` jump distance '' corresponds to the first peak in the radial distribution function and it is possible to recognize even the existence of peaks of higher order and which correspond to subsequent jumps .
the reason why the confined system shows a hopping motion that is not present in the bulk system is related to the presence of the walls .
since these walls consist of immobile particles they give rise to a ( static ! ) local potential energy landscape with energy minima for positions in between those particles and the observed jumps are just the motion of the particles between these local minima .
note that these jumps are not necessarily parallel to the surface and are not confined to the first layer of the fluid since they are also observed for particles that are in the second and third layer ( see fig .
[ fig_vanhove_layerdependence]c and ref .
the reason for this is that the above mentioned energy landscape propagates also somewhat into the liquid and hence affects the relaxation dynamics also at larger values of @xmath0 .
we now address the question how to quantify the slowing down of the dynamics close to the surface .
as we will see in the following it is possible to extract from the decay of time dependent correlation functions characteristic relaxation times which show a pronounced @xmath0-dependence . we will concentrate on wave vector dependent intermediate scattering functions in reciprocal space @xcite , but the investigation of similar density correlators in real space shows the same qualitative and even quantitative results @xcite . here
we will consider the two following functions : @xmath92 \delta(z_j(0 ) - z ) \rangle \ , , \label{eq10}\ ] ] and its diagonal part @xmath93 \delta(z_j(0 ) - z ) \rangle \ , .
\label{eq11}\ ] ] thus @xmath94 and @xmath95 are just the coherent and incoherent intermediate scattering functions generalized in such a way to take into account their @xmath1dependence .
note that , in contrast to the case of a bulk liquid that is isotropic , we have to take into account the orientational dependence of these functions . in the following
we will concentrate on @xmath96 and @xmath97 , the functions in which one has taken the angular average over @xmath98 for wave - vectors that are parallel to the walls . although most of the results presented below are for the a particles and the wave - vector @xmath99 , the location of the maximum in the partial a - a structure factor @xcite , we have found qualitatively similar results for the b particles and/or for other wave - vectors @xcite .
self part @xmath100 of the intermediate scattering function at @xmath101 and @xmath99 for a particles .
the different curves correspond to the following distances from the wall : @xmath102 ) .
the bold solid curve is the average over the whole film and the bold dashed curve is the bulk curve .
the horizontal dashed line is used to define a relaxation time @xmath103 .
, width=321 ] let us first consider the typical behavior of the incoherent intermediate scattering function at low temperatures in bulk systems ( dashed curve in fig .
[ fig_fsqzt ] ) .
the decay of @xmath104 happens in two steps , first the decay to a plateau which corresponds to the relaxation inside the cage , see the above discussion of the mean squared displacement , and a second step that corresponds to the @xmath4relaxation .
note that for @xmath101 the mentioned plateau is not yet very well developed , but its width increases rapidly with decreasing temperature @xcite .
as expected , @xmath100 for particles in the center of the film are identical to the bulk curve . upon approach to the wall
we recognize again the continuous slowing down of the dynamics .
compared to the @xmath1dependence of the slowing down found in the mean squared displacements the effect is even more pronounced , since here the time scales change by many decades .
( note that for the smaller values of @xmath0 considered , the correlators do not decay to zero anymore since the dynamics is so slow .
hence we see that it is indeed crucial to have starting configurations that are already equilibrated in order to avoid aging effects . ) from the figure we also recognize that the reason for the unusual shape of the _ average _ correlator , bold solid curve in fig .
[ fig_fsqzt ] , is that the curves close to the wall are decaying so much slower than the one in the center of the film .
time dependence of @xmath100 as a function of rescaled time @xmath105 ( in bold ) .
the values of @xmath0 are @xmath106 , and 7.5 .
the thin dashed lines are fits to the data in the @xmath4relaxation regime with a kww - law ( see eq .
( [ eq12 ] ) ) .
, width=321 ] figure [ fig_fsqzt ] shows that the typical time scale for the relaxation increases rapidly for decreasing @xmath0 .
the question arises whether it is only the time scale that depends on @xmath0 or whether also the shape of the curve depends on the distance from the wall . to check this
we can define a relaxation time @xmath103 by requiring that at this time the correlator has decayed to @xmath107 ( see horizontal dashed line in fig .
[ fig_fsqzt ] ) . in fig .
[ fig_fsqzt_superpos ] we show a plot of @xmath100 vs. @xmath105 for various values of @xmath0 .
if the shape of the curves would be independent of @xmath0 we would find that this rescaling of time produces a master curve .
however , from the figure we see that this is not the case since the correlators for small values of @xmath0 are more stretched . in order to quantify this effect
we have fitted the correlators by a kohlrausch - williams - watts function ( kww ) , @xmath108 \ , , \label{eq12}\ ] ] a functional form often used to describe the @xmath4relaxation of supercooled liquids . that this function is able to characterize well also the correlators in the confined liquid
is demonstrated in fig .
[ fig_fsqzt_superpos ] where we have included the kww fits to the data as well ( thin dashed lines ) .
@xmath1dependence of the kww parameters @xmath109 , @xmath110 and @xmath111 as obtained from fits to @xmath100 for @xmath112 ( thin lines ) and @xmath101 ( bold lines ) .
, width=321 ] in fig .
[ fig_fsqzt_kwwpar ] we show the @xmath1dependence of the fit parameters obtained by fitting @xmath100 with a kww function .
the two set of curves correspond to @xmath112 and @xmath101 ( thin and bold lines , respectively ) .
we start the discussion with the relaxation times @xmath110 ( dashed lines in the figure ) . from the figure it becomes clear that for large @xmath0 this relaxation time is basically independent of @xmath0 and is the same as the one found in the bulk . for decreasing @xmath0
it starts to rise rapidly and at the smallest @xmath0 for which we were able to determine it , it is more than two decades higher than the bulk value .
the parameter @xmath109 , i.e. the height of the plateau in the @xmath74relaxation regime , shows only a relatively mild @xmath1dependence in that it increases by about 20% in the @xmath1range considered ( dotted curves ) .
in contrast to this , the third parameter of the kww - function , the stretching parameter @xmath111 , shows a relatively strong @xmath1dependence in that it decreases from a value around 0.8 to @xmath113 at small @xmath0 .
thus we see that not only do the relaxation times increase with decreasing @xmath0 , but that also the stretching increases significantly .
the likely reason for this strong stretching is that very close to the surface the system is very heterogeneous in the sense that a slight modification of the local ( frozen ) environment , and such variations are certainly present , will give rise to very strong fluctuations in the relaxation behavior .
hence if one averages over these fluctuations the result is a very pronounced stretching .
note that this small value of @xmath111 is also the reason why the correlators for small @xmath0 decay by 4 - 5 orders of magnitude slower than the ones at large @xmath0 ( see fig . [ fig_fsqzt ] ) , whereas the relaxation time @xmath110 increases only by a factor of @xmath114 .
a comparison of the data sets for @xmath112 and @xmath101 shows that the @xmath1dependence for @xmath109 and @xmath111 are , within the noise of the data , the same .
this implies that also for the confined system the so - called time - temperature superposition is valid , i.e. that the shape of the curves is independent of temperature .
this feature is often found in glass forming liquids and in particular also for the present binary lennard - jones system @xcite .
however , we already point out at this place that the relaxation times for the two temperatures do _ not _ differ by just a constant factor .
we will come back to this point in sec .
[ sec3c ] .
since the relaxation dynamics of the present lennard - jones system in the _ bulk _ is described well by means of mode - coupling theory , we will try to test to what extend some of the predictions of this theory hold for our confined systems .
we have already seen that for a given value of @xmath0 the time - temperature superposition principle holds , in agreement with the prediction of mct .
a further important prediction of the theory is the so - called factorization property for the @xmath74relaxation : consider any correlator @xmath115 , where @xmath116 stands for wave - vector , particle species , etc .
mct predicts that in the time window of the @xmath74regime the time dependence of @xmath115 can be written as follows : @xmath117 here @xmath118 and @xmath119 are the height of the plateau and an amplitude , respectively .
the remarkable statement of mct is that the whole time and @xmath120dependence of the correlator is given by the function @xmath121 , the so - called @xmath74correlator , and that this function is system universal , i.e. it does not depend on @xmath116 .
one possibility to check whether or not the factorization property holds is to calculate the following function @xcite : @xmath122 where @xmath123 and @xmath124 are two arbitrary times in the @xmath74regime .
if the factorization property holds , the function @xmath125 should be independent of @xmath116 .
for the bulk system it was found that the factorization property does indeed hold very well in that the data for @xmath125 for all sort of different correlators do indeed fall on top of each other @xcite . in the following we
thus will check to what extend this holds also true for @xmath1dependent correlators .
time dependence of @xmath100 for the a particles at @xmath101 for different wave - vectors ( see legend ) for particles in the center ( @xmath126 , bold lines ) and at the surface ( @xmath127 , thin lines ) .
, width=321 ] to begin we investigate the wave - vector dependence of these functions .
figure [ fig_fsqzt_qdep ] shows that qualitatively the time dependence of the correlators is independent of @xmath128 but that the typical relaxation time as well as the height of the plateau increases rapidly with decreasing @xmath128 . for large values of @xmath0 ( bold curves ) and
large wave - vectors it is in fact difficult to see that there is indeed a plateau ( see curve for @xmath129 ) since the curve decays very quickly .
this is due to the fact that large wave - vectors correspond to small displacements of the order @xmath130 and that thus a decorrelation of the particle position can take place already on a very short time scale .
no such fast decay is seen for small values of @xmath0 ( thin curves ) since there the relaxation dynamics is very slow even on small length scales .
since the factorization property is predicted to hold only in the @xmath74relaxation regime , it is necessary that correlators show a well developed plateau . from fig .
[ fig_fsqzt_qdep ] we recognize that at @xmath101 this condition is not fulfilled for large values of @xmath0 and therefore it is not possible to check whether or not the factorization property holds .
however , since these correlators are very similar to the one in the bulk , and it has been shown that for the bulk system the factorization property holds @xcite , we can assume that it does so also for the confined system at large @xmath0 , _ if _ the temperature is sufficiently low .
for small values of @xmath0 it is instead possible to make a direct check on the validity of the factorization property by calculating the function @xmath125 .
the results are presented in fig .
[ fig_fsqzt.factorization ] where we show the time dependence of the function @xmath125 for different values of @xmath0 ( spaced by @xmath131 ) and three different wave - vectors ( panels ( a)-(c ) ) .
the times used were @xmath132 and @xmath133 .
we recognize that in the @xmath74regime the correlators for the different values of @xmath0 fall nicely onto a master curve , thus showing that the factorization property holds .
note that the fact that these curves fan out at short and long times shows that the existence of such a master curve is by no means a trivial matter . the existence of the master curves shown in figs .
[ fig_fsqzt.factorization]a - c demonstrates that the shape of the correlators is independent of @xmath0 , as predicted by mct .
however , this shape should also be independent of @xmath128 . in order to check this prediction we have calculated for several values of @xmath128 a mean master curve by averaging the correlator @xmath100 over @xmath134 .
the time dependence of these master curves for the different wave - vectors are shown in fig .
[ fig_fsqzt.factorization]d . from this figure
we see that the shape of the master curve does depend on @xmath128 , i.e. the factorization property does not hold .
it might of course be that the factorization property does hold at even lower temperatures , but due to the finite size effects discussed in sec .
[ sec3f ] this can presently not be checked .
we will now discuss in more detail the spatial dependence of structural relaxation times in the liquid . in fig .
[ fig_tau_z]a we plot @xmath135 , defined by @xmath136 , for all temperatures investigated . from this graph we recognize that the thickness of the film is sufficiently large since at all temperatures the relaxation times become independent of @xmath0 and are very close to the value in the bulk ( open diamonds at @xmath137 ) . at the highest temperature bulk behavior
is realized already for @xmath138 . with decreasing temperature the influence of the wall reaches further into the system
, i.e. , the region which is affected by the wall grows with decreasing temperature and at the lowest temperature considered it extends up to @xmath139 . from the plot it also becomes evident that if the temperature would be decreased somewhat more , the relaxation times in the middle of the film would no longer coincide with the ones of the bulk , i.e. that @xmath135 becomes a function of @xmath48 .
our main goal is now to extract a characteristic dynamic length scale from the data . since there are no reliable theoretical concepts on how this should be done , we are forced to use just an empirical description of our data which includes such a length scale . in simulations of the same lennard - jones liquid within a narrow cylindrical _ pore _
@xcite we have been able to find such a description , at least for the particles close to the walls , i.e. for small @xmath0 .
the proposed ansatz had the form @xmath140 \ , , \label{eq15}\ ] ] where @xmath141 is a fit parameter that gives some sort of `` penetration depth '' of the particles into the wall @xcite .
that this functional form is indeed able to describe the data very well is shown in fig .
[ fig_tau_z]b .
the solid inclined lines are fits with the function given by eq .
( [ eq15 ] ) and we have chosen @xmath142 . since the slope of these lines are the parameter @xmath143 , we see that this quantity is increasing with decreasing temperature ( from 2.8 at @xmath70 to 7.5 at @xmath144 ) .
we note that this result is in contrast to the one found for the narrow pore in that there we found that the fit works well even if @xmath143 and @xmath141 are kept constant , i.e. independent of @xmath73 @xcite .
the reason for this difference might be that in ref . @xcite the temperature at which the wall was frozen was always the same , in contrast to the present case where @xmath21 is set equal to @xmath73 , see sec .
( we remark that it is also possible to obtain good fits to the data by keeping @xmath143 fixed to 7.53 and to make @xmath141 depend on temperature @xcite . )
we now extract a characteristic length scale for the influence of the wall by determining the length at which the fit with eq .
( [ eq15 ] ) crosses the bulk value ( horizontal lines in fig .
[ fig_tau_z]b ) . in the following
we will denote this so determined length scale by @xmath145 and in the next subsection we will discuss its @xmath120dependence .
( we remark that the values of @xmath145 depend only weakly on the way one does the fit to the data with eq .
( [ eq15 ] ) , i.e. whether @xmath143 or whether @xmath141 is kept constant . ) since eq .
( [ eq15 ] ) is only capable to describe the data close to the wall we looked for alternative possibilities to parameterize the relaxation times .
if one thinks of a characteristic length scale @xmath146 for a decay process , a natural ansatz is the exponential form @xmath147 $ ] .
however , an analysis of the difference of @xmath135 to the bulk value showed that the increase is much stronger than purely exponential , and therefore we made an exponential ansatz for the logarithm of the ratio between @xmath135 and a reference value @xmath148 : @xmath149 = a(t ) \cdot \exp \left [ - \frac{z}{\xi_0(t ) } \right ] \ , .
\label{eq16}\ ] ] in an ideal situation @xmath148 should of course correspond to the bulk value since for large distances from the wall we have @xmath150 .
however , remaining small density differences within the simulation as well as finite size effects ( see section [ sec3f ] ) lead to small deviation from this theoretical value . hence it was possible to improve the fit , in particular close to the surface where the @xmath0-dependence is most pronounced , by extracting @xmath148 from an extrapolation of @xmath135 to large @xmath0 or by using @xmath148 as a free parameter .
this was done to obtain the data shown in fig .
[ fig_tau_z]c , where we plot in a logarithmic way the left hand side of eq.([eq16 ] ) as a function of @xmath0 .
( we mention that the values for @xmath148 differ only slightly from the bulk values . )
we see that the data are indeed compatible with straight lines , and this for all distances , thus validating the ansatz .
the negative of the slope of the straight lines is the inverse of @xmath146 and from the fact that this slope decreases with decreasing temperature we see that @xmath146 becomes larger if @xmath73 is lowered . in the next section
we will discuss the @xmath120dependence of @xmath146 in more detail .
finally we mention that the length @xmath146 decreases slightly with increasing wave - number @xmath128 , but that this effect is only weak @xcite .
fit of the simulation data for @xmath100 for a particles at @xmath99 and @xmath101 according to eq .
( [ eq17 ] ) for several values of @xmath0 ( @xmath151 0.3 , 0.5 , 0.7 , 1.0 , 1.5 , 2.0 , 3.0 , and 7.5).,width=321 ] in ref .
@xcite we attempted not only to describe the @xmath1dependence of the relaxation times but the whole time and space dependence of the scattering function @xmath100 .
the ansatz used in that study had the following form : @xmath152 \ , .
\label{eq17}\ ] ] here @xmath153 is the intermediate scattering function in the bulk and thus the second term of the left hand side of eq .
( [ eq17 ] ) describes the deviation of @xmath100 from this correlator . that this type of ansatz is able to give a very good description of the time dependent data is demonstrated in fig .
[ fig_fsqzt_parisifit ] where we show @xmath100 for different values of @xmath0 ( symbols ) and the corresponding fit ( dashed lines .
( we mention that if one plots the data versus @xmath0 for different times @xmath154 the agreement between data and fit is very good as well ( see ref .
@xcite ) . ) time and temperature dependence of the fit parameter @xmath155 from eq .
( [ eq17 ] ) at different temperatures ( @xmath156 1.0 , 0.8 , 0.7 , 0.6 , 0.55 , and @xmath157).,width=302 ] the time and @xmath120dependence of the stretching parameter @xmath158 is discussed in ref .
@xcite . here
we focus on the dependence of the length scale @xmath155 on @xmath154 and @xmath73 .
this dependence is shown in fig .
[ fig_fsqzt_parisifitpar ] from which we recognize that @xmath155 is a smooth function of time with a local maximum @xmath159 at @xmath160 .
the location of this maximum shifts to larger times if @xmath73 is decreased and also its value increases .
thus we conclude that there exists a characteristic time scale at which the influence of the wall extends farthest into the liquid , i.e. for which the system is relatively stiff .
this time scale corresponds roughly to the time of the @xmath161-relaxation in the bulk system @xcite , a result which is plausible as can be seen as follows : it is well known that for simple liquids the @xmath4relaxation is largest for wave - vectors at the peak of the static structure factor ( see , e.g. , data in ref .
@xcite ) , i.e. the length scale that corresponds to the typical distance between the particles .
since it is this time scale which is really relevant for the structural relaxation , since it corresponds to the typical time needed for the particles to leave their cage , it can be expected that any external perturbation , such as the presence of the wall , will affect the dynamics at long distances on this time scale which is comparable with this relaxation time .
finally we remark that the qualitative behavior as well as the height of the maximum in @xmath155 does not change significantly if one fits the data with an exponential function , i.e. @xmath162 in eq .
( [ eq17 ] ) , ( see the subsequent section and fig . [ fig_dynamic_lengthscales ] ) , although we note that such a fit is not able to describe the data very well @xcite .
comparison of different dynamic length scales as extracted from the spatial dependence of scattering functions @xmath100 and structural relaxation times : @xmath145 from fits with eq .
( [ eq15 ] ) , @xmath146 from fits with eq .
( [ eq16 ] ) and @xmath159 from fits with eq .
( [ eq17 ] ) ( see also legend and text ) .
, width=302 ] from the results described so far we were able to extract three different dynamic length scales : @xmath145 from eq .
( [ eq15 ] ) , @xmath146 from eq .
( [ eq16 ] ) , and @xmath159 from the ansatz given by eq .
( [ eq17 ] ) .
we now discuss how these different length scales depend on temperature .
this @xmath120dependence is shown in fig .
[ fig_dynamic_lengthscales ] where we have plotted @xmath145 , @xmath146 , and @xmath159 as a function of inverse temperature .
first of all we recognize that the different definition of length scale give basically all the same result and that thus the precise definition is not relevant .
furthermore we see that it is also not crucial how exactly one defines a length scale : for the case of @xmath145 it does not manner whether one keeps @xmath143 or @xmath141 independent of @xmath73 ( see discussion of eq .
( [ eq15 ] ) ) ; for the case of @xmath146 it does not matter whether the relaxation time @xmath148 is the bulk value , the value obtained from extrapolating @xmath135 to large @xmath0 , or whether it is just a fit parameter ; for the case of @xmath159 it is irrelevant at which wave - vector one uses the ansatz ( [ eq17 ] ) or whether @xmath158 is set to unity or not .
note that we have plotted all data in an arrhenius plot , i.e. logarithmically versus inverse temperature .
although to our knowledge there is no theoretical reason why one should plot the data in this manner , the graph shows that this presentation does rectify the data .
thus we conclude that the length increases like @xmath163 and that the value of @xmath164 is independent of the length scale . in ref .
@xcite we have discussed the temperature dependence in more detail and made also a comparison with other characteristic length scales in the liquid .
the conclusion is that basically all relevant length scales are very similar in size and show a quite similar @xmath120dependence .
finally we remark that this @xmath120dependence is significantly larger than the one that one obtains from the one of static quantities like , e.g. , the envelope of the radial distribution function @xcite . to get a better understanding of the collective dynamics close to the surface
we investigate the intermediate scattering function @xmath165 for all ( a and b ) particles which is given by eq .
( [ eq10 ] ) . in this quantity
the initial position of a particle is correlated with the position of any other particle at a later time @xmath154 .
the main consequence for the behavior close to the surface is the following . since the configuration of the wall particles is frozen there exist a static landscape in the potential energy and particles of the liquid prefer to occupy the local minima in this landscape .
therefore it can be expected that as soon as a liquid particle close to the surface leaves its position this vacancy will be filled by another particle .
this implies that the collective correlation function close to the wall does not decay to zero even at long times .
normalized coherent intermediate scattering function @xmath165 for @xmath99 for both types of particles at @xmath166 .
the thin curves correspond to @xmath167 .
the bold dashed line is the data for the bulk .
inset : @xmath1dependence of the height of the plateau @xmath168 for different temperatures.,width=302 ] that this expectation is indeed met is shown in fig .
[ fig_fqzt ] where we plot the time dependence of @xmath165 as a function of temperature .
note that , in order to improve the statistics of the data , we do not distinguish between a and b particles and in addition we have made a spherical average .
we mention , however , that the data in which we averaged only over wave - vectors parallel to the surfaces look qualitatively quite similar @xcite .
also included in the figure is the corresponding correlator for the bulk system ( bold dashed line ) and we recognize that this function coincides with @xmath165 for large values of @xmath0 .
furthermore we see that the height of the plateau in @xmath165 at long times increases rapidly with decreasing @xmath0 and that for the smallest @xmath0 we do no longer find the two - step relaxation usually seen in supercooled liquids .
the @xmath1dependence of the height of this plateau , @xmath168 , is shown in the inset of the figure ( in a log - lin plot ) .
we see that it shows basically an exponential decay , although within the noise of the data it is difficult to exclude other functional forms .
the slope of the straight lines in the inset gives a _
static _ length scale over which the liquid starts to become uncorrelated with the `` pinning - field '' generated by the wall .
thus it can be expected that this length scale is comparable to the one found in other static two - point correlation functions , such as , e.g. the decay of the correlation in the radial distribution function . in agreement with this expectation
we see that this length scale is indeed on the order or 2 - 3 and that it shows only a weak dependence on temperature .
in particular this @xmath120dependence is significantly weaker than the one found for the dynamical length scales discussed above . due to the presence of the plateau at long times in @xmath165 ,
it is difficult to extract a relaxation time from this function . in an attempt to define such a time scale ,
we have subtracted the long time limit of @xmath165 and normalized the function by @xmath169 .
for the so obtained resulting normalized correlator it is possible to define a relaxation time . however , we have found that this time scale depends only relatively weakly on @xmath0 , and in particularly much weaker than the relaxation time for @xmath170 , and therefore this approach was no longer pursued @xcite . until now we have considered only values of @xmath48 and @xmath73 such that the system under investigation is sufficiently large to guarantee that in the center it has the same dynamics as in the bulk .
thus strictly speaking we have so far only investigated the dynamics of a liquid in the vicinity of _ one _ wall and the fact that the overall geometry is a film was completely irrelevant .
however , from fig . [ fig_tau_z ] we have concluded that with decreasing temperature the dynamical length scales for cooperativity are growing and that the influence of the wall reaches further into the system .
therefore it is obvious that for a system of a given thickness also the particles in the center will be influenced by the walls , if the temperature is sufficiently low .
hence it can be expected that the dynamics should be slowed down even more and the goal of this subsection is to investigate this effect in more detail . local relaxation times in films with different thickness ( @xmath171 10.0 , and 5.0 ) .
plot of the left hand side of eq .
( [ eq16 ] ) versus distance to the nearest wall ( symbols ) .
the lines correspond to the ansatz given by eq .
( [ eq16 ] ) with parameter that have been determined in systems with @xmath57.,width=302 ] instead of lowering the temperature we investigate films that have a smaller thickness ( @xmath172 and @xmath173 ) for temperatures down to @xmath71 . we have repeated the procedure described in the previous subsection for @xmath57 in order to obtain the @xmath1dependence of the relaxation times of the incoherent intermediate scattering function and we present this dependence in fig . [ fig_tau_z_finitesize ] . at high temperatures , where each particle is influence by at most _ one _ wall , the local dynamics is only a function of its distance to the wall .
this can be see by the data for @xmath70 since the curves for the different values of @xmath48 fall on top of each other and can , e.g. , be described by the exponential ansatz of eq .
( [ eq16 ] ) , solid line .
if we decrease temperature , @xmath166 , the data for the thinnest film starts to deviate from the master curve because the particles in the center already feel the presence of both walls . at even lower temperatures , @xmath71 ,
this effect becomes more pronounced and also the data for @xmath172 starts to deviate from the one of @xmath57 .
note that since at this @xmath73 the data for @xmath172 shows strong deviations from the one for @xmath57 at @xmath174 , it follows that in a film of thickness @xmath173 the particles close to a wall are also influenced by the presence of the _ opposite _ wall .
the question is now if this enhancement of the slowing down is just a _ linear _ effect , i.e. it is given by the sum of the contributions from both walls , or if the situation is more complicated . under the assumption that the local dynamics at a distance @xmath0 from a rough surface is given by eq .
( [ eq16 ] ) a reasonable generalization is given by the ansatz @xmath149 = a(t ) \cdot \left\lbrace \exp \left [ - \frac{z}{\xi_0(t ) } \right ] + \exp \left [ - \frac{d - z}{\xi_0(t ) } \right ] \right\rbrace \ , .
\label{18}\ ] ] using the parameters @xmath175 and @xmath176 as determined from the fits to the data for @xmath57 , see sec .
[ sec3c ] , we thus can use this expression to calculate the @xmath1dependence of @xmath177 for the smaller films .
the resulting predictions for @xmath172 and @xmath173 are included in fig .
[ fig_tau_z_finitesize ] as well ( dashed and dotted lines ) .
we see that although the predicted curves look qualitatively similar to the data for these films , they severely underestimate the relaxation times at large values of @xmath0 .
qualitatively the same result is obtained if we use eqs .
( [ eq15 ] ) or ( [ eq17 ] ) to predict @xmath135 for thin films @xcite .
hence we conclude that the presence of two walls affects the relaxation dynamics of the liquid particles in a non - additive manner , i.e. that the reason for the slowing down must be non - linear .
this result is not that surprising since mode - coupling theory , a very successful theoretical framework to describe the slowing down of the dynamics of glass - forming systems , predicts that this slowing down is related to a non - linear feedback mechanism @xcite . according to mct the relaxation dynamics of a particle
is strongly coupled to the motion of its neighboring particles that form the cage of the tagged particle . since this tagged particle itself
is part of the cage of other particles , a decrease in @xmath73 , i.e. a stronger coupling of the motion , leads to a strong ( i.e. non - linear ) stiffening of the cage ( feedback effect ) . in our case the particles that form the wall
are completely frozen and thus they do no longer participate to help to relax the fluid particles in their neighborhood .
thus from a qualitative point of view the lack of fluctuations of the cage of the particles near the walls explains the slowing down of their dynamics and the feedback effect rationalizes why the presence of two walls affects the system stronger than just the linear sum of two individual walls .
finally we mention that this non - linear feedback effect should affect the relaxation dynamics even more in the case of systems in which the confinement is given by a tubular geometries or in cavities .
having characterized in great detail the local dynamics and understood how the structural relaxation depends on the distance from the wall , we now can investigate the dynamic properties of the whole system , i.e. if one averages the correlators over @xmath0 .
these averaged correlation functions are exactly the quantities that are accessible in most experiments . in contrast to experiments we are , however , in the favorable position to be able to relate these averaged quantities to the time dependence of the local ones .
self part @xmath178 of the intermediate scattering function for a particles in a film with @xmath57 for @xmath99 and temperatures @xmath179 , and @xmath157 ( solid lines ) .
the dashed lines are fits to the data with a superposition of two stretched exponential laws.,width=302 ] in fig .
[ fig_fsqt.film ] we show the time dependence of @xmath180 , i.e. the incoherent intermediate scattering function for the a particles averaged over the whole film ( solid lines ) .
( the wave - vector is again parallel to the walls . )
although at a first glance these correlators look quite similar to the ones in the bulk @xcite , a closer inspection shows that there are important differences . already at @xmath70
one sees that after a rapid decay to a small value around 0.05 , the correlator shows a tail that extends to relatively long times . at low temperatures
the function shows at intermediate times the usual plateau that is related to the cage effect but at longer times the mentioned tail starts now to become very pronounced .
the reason for this behavior follows directly from the results of the local correlation functions .
since the typical time scales for the structural relaxation as a function of @xmath0 extend over several decades in time , the average over all those curves should be very stretched ( see also fig .
[ fig_fsqzt ] ) . for @xmath70 particles with @xmath181 ,
i.e. more than 70% of the particles , behave more or less bulk - like and contribute to the fast decay seen in fig .
[ fig_fsqt.film ] at short times .
the correlators for particles at the surface decay much slower and give therefore rise to the long time tail for @xmath182 . with decreasing temperature the number of particles with a relaxation that is significantly slower than the one in the bulk becomes larger and therefore the amplitude of the long time tail increases . as a consequence
the shape of the correlators change with temperature , i.e. the time - temperature superposition principle is not valid , in contrast to the case of the bulk system @xcite .
therefore the definition of characteristic relaxation times for the whole decay is not really possible . of course one can calculate the time dependence of @xmath180 from the knowledge of the @xmath1dependent correlators which we are able to describe by a set of kww functions ( see fig .
[ fig_fsqzt_superpos ] ) . on the other hand , in experiments
this information is usually not available and therefore it is interesting to check how one can characterize the averaged data in a different way .
we have found that for all temperatures considered , the @xmath4relaxation of the average correlator can be described very well by the sum of two kww - laws ( having different amplitudes ) : one that is for the early part of the relaxation and the second one for the mentioned tail . that this type of ansatz does indeed give a very good description of the data
is shown in fig .
[ fig_fsqt.film ] where we have included the results of such fits as well ( dashed lines ) .
temperature dependence of the relaxation times of the two processes extracted from the self part of the intermediate scattering function averaged over all a particles in the system ( filled triangles ) .
the open circles are the @xmath4relaxation time for the bulk.,width=294 ] in fig .
[ fig_fsqt_fitpar ] we show the temperature dependence of the relaxation times , obtained from the kww fits , for the fast and slow process .
we see that within the accuracy of the data i ) these @xmath120dependencies are very similar , although the time scale for the slow process is around 40 times larger , and ii ) that the fast process shows basically the same @xmath120dependence as the bulk system . for the fast process
the values for the kww - parameter @xmath111 is around 0.8 and depends only weakly on @xmath73 .
in contrast to this we find for the slow process that @xmath111 is only around 0.25 - 0.3 , i.e. a very small value @xcite .
the reason for such a strong stretching , which is basically independent of @xmath73 , is that the relaxation dynamics is extremely heterogeneous , i.e. that close to the wall the typical time scales differ by orders of magnitude ( see fig .
[ fig_fsqzt ] ) .
thus we find that the long time tail is basically given by the second process which has a clearly separated time scale with a smaller amplitude .
note that this conclusion comes from our data in the _ time _ domain .
however , it is also possible to make a time - fourier transform of our data and to calculate the dynamical susceptibility @xmath183 .
this has been done in ref .
@xcite where it was shown that these two processes can also be seen in the frequency domain in that one finds that the imaginary part of @xmath183 seems to be indeed the sum of two peaks .
such a double peak structure is often found in real spectroscopy experiments where one observes a second peak at frequencies well below the peak corresponding to the @xmath4relaxation of the bulk system .
thus from such _ averaged _ data one could easily come to the conclusion that there are indeed two distinct processes , e.g. one corresponding to particles in the center and the other to a layer of slow particles at the surface and which give rise to the second process .
however , the analysis of the @xmath1dependent correlators presented here shows that such a conclusion might be wrong .
time dependence of the intermediate scattering function @xmath104 for all a particles in films with different thickness ( @xmath57 , 10.0 , 5.0 ) for @xmath99 and different temperatures ( @xmath70 , 0.8 , 0.55 ) .
the dotted curves correspond to the bulk data .
, width=302 ] finally we discuss how the averaged time correlation function depends on the thickness of the film . in fig .
[ fig_fsqt.film.size ] we show @xmath104 for different temperatures and three values of @xmath48 as well as for the bulk system .
we see that in these averaged curves one can notice already at the highest temperature a significant deviation of the correlators for the confined systems from the one of the bulk .
this is due to the presence of the particles that are slowed down near _
one _ of the walls .
( recall that in fig .
[ fig_tau_z_finitesize ] we have seen that at this @xmath73 none of the particles of even the smallest system are affected by _ both _ walls . )
thus the amplitude of the tail at long times is given only by twice the fraction of the particles that are strongly affected by one wall .
( note that the value of this fraction depends on @xmath48 . ) with decreasing @xmath73 the fraction of particles affected by one wall increases , due to the growing dynamical length scale and hence the curves for the confined system decay significantly slower than the bulk curves . for even lower temperatures we have the additional effect that the particles close to the center are affected in a non - linear way by the presence of _ both _ walls , which leads to an additional slowing down .
thus we see that the @xmath120 and @xmath184dependence of the relaxation function is quite complex since there are three main causes : i ) the general slowing down of the dynamics with decreasing temperature ; ii ) the presence of a growing length scale over which the slow dynamics close to the wall affects the relaxation dynamics ; iii ) the non - linear effects occurring if a particle is influenced by more than one wall .
the sum of these causes has the effect that the shape of the correlation function depend in a strong and highly non - trivial way on @xmath73 and @xmath48 .
this dependence also hampers a sensible definition of a relaxation time which is the reason why we do not discuss the @xmath120 and @xmath184dependence of it .
in this paper we have presented the results of a molecular dynamics computer simulation of a binary lennard - jones liquid confined between two rough surfaces .
the walls are given by a frozen amorphous configuration of the same liquid and as we have shown here that the static properties of the confined fluid are identical to the ones of the bulk system .
we have found that the relaxation dynamics of particles close to the interface are slowed down by orders of magnitude with respect to the one in the bulk . due to the cooperativity of the particle motion within a liquid ,
also particles at some distance from the wall are influenced by the presence of the wall .
therefore the properties of the structural relaxation become a function of distance @xmath0 to the wall .
all these properties are a smooth function of @xmath0 , although some of them show a very strong @xmath1dependence which results that the dynamics of the system is very heterogeneous .
the time dependence of the incoherent intermediate scattering functions @xmath2 is described well by a kww - function with a stretching parameter @xmath111 and a relaxation time @xmath110 that depends strongly on @xmath0 .
the @xmath1dependence of the relaxation times can be described well by various phenomenological laws . from these laws it is possible to extract various dynamical length scales over which the dynamics of the system is influenced by the walls .
these length scales are of purely dynamical origin and show a @xmath120dependence which is compatible with an arrhenius law with an activation energy that is independent of the details of the definition of the length scale .
note that this @xmath120dependence is in contrast to other dynamical length scales that have been discussed in the literature before , such as , e.g. , the dynamical heterogeneities @xcite for which a divergence at a temperature close to @xmath185 of mode - coupling theory has been found .
we have tested whether or not the @xmath120dependence of our length scales are compatible with a divergence at a finite temperature and have found , see ref .
@xcite , that in principle such a scenario is possible at @xmath185 or at @xmath186 , the kauzmann temperature of the system .
however , the noise in the data is unfortunately too large to make strong statements on this point .
as long as the dynamical length scales are significantly smaller than the system size , the discussed results are just an effect due to the interface , i.e. the dynamics is only a function of distance to the surface and independent of the film thickness . for a film with given thickness
there exists however a threshold temperature below of which particles are influenced by both interfaces since the size of the crr s are larger than @xmath187 . in that case
we find that the slowing down is much stronger than one might expect from a simple linear superposition of the influence of two single surfaces .
this is evidence that the mechanism for the slowing down of the particles is a strongly non - linear effect , such as proposed , e.g. , by mode - coupling theory .
finally we have studied the time dependence of the correlators averaged over the whole film .
these correlators are a quite complex function of @xmath73 and @xmath48 since one has to distinguish between the normal slowing down of the liquid with decreasing temperature ( which is present in the bulk system as well ) , the influence of one interface on the dynamics with a dynamical length scale that increases with decreasing @xmath73 , and last not least with the mentioned non - linear effects . in view of these different reasons for the slowing down of the average dynamics
, it is clear that it is rather difficult to come up with an expression for the @xmath184dependence of the glass transition temperature , since it is not even obvious how the typical relaxation time of the system should be defined . despite this uncertainty
it is clear that the glass transition temperature , which could be defined as the temperature at which a relaxation time or the viscosity attains a given predescribed value , will increase with decreasing thickness of the film .
however , we point out that this qualitative result holds only true for the case of rough walls as they have been considered in the present work . in a different study
we have also investigated the relaxation dynamics of a system with _ smooth _ walls where we have again made sure that the structural properties of the confined liquid is not changed by these interfaces @xcite . in that study we have found that the dynamics close to the wall is _ accelerated _ which in turn leads to a decrease of the glass transition temperature with decreasing @xmath48 . hence we conclude that the relaxation dynamics is strongly affected by the nature of the confining walls , in agreement with experimental results .
thus it seems that it would be of interest to avoid these confining walls altogether in order to disentangle the influence of the wall on the dynamics and the effect of the finite extension of the system . within a simulation
this is indeed possible since one can investigate the relaxation dynamics of a given system for different system size ( using periodic boundary conditions in all three directions ) . in the past such simulations have been done and it has been found that at a given temperature the relaxation dynamics slows down with decreasing temperature @xcite . hence we conclude that intrinsically confinement leads to a slower dynamics but that the presence of fluid - wall interactions can completely mask this general trend .
last not least we comment on the necessity for a more complete theoretical understanding of the relaxation dynamics of confined systems .
the present simulations have shown that it is possible to have a slowing down of the dynamics without a change of the structural properties ( and in refs .
@xcite it was shown that also an acceleration can be observed ) .
hence we conclude that it is not sufficient to know the structure in order to predict the dynamics .
this is in contrast to the situation of the bulk where it is indeed possible , using mode - coupling theory , to predict the relaxation dynamics using as input only structural quantities @xcite .
for the moment it is not clear to us how this theory has to be modified in order to rationalize the results found in this work .
of course it is probably possible to use some phenomenological theories in order to describe these results @xcite but having a theory at hand that allows a full microscopic calculation would certainly be much more preferable .
w. gtze , p. 287 in _
liquids , freezing and the glass transition _ eds . : j. p. hansen , d. levesque , and j. zinn - justin , les houches .
session li , 1989 , ( north - holland , amsterdam , 1991 ) ; w. gtze , j. phys . : condens .
matter * 10 * , a1 ( 1999 ) . | we present the results of molecular dynamics computer simulations of a binary lennard - jones liquid confined between two parallel rough walls .
these walls are realized by frozen amorphous configurations of the same liquid and therefore the structural properties of the confined fluid are identical to the ones of the bulk system .
hence this setup allows us to study how the relaxation dynamics is affected by the pure effect of confinement , i.e. if structural changes are completely avoided .
we find that the local relaxation dynamics is a strong function of @xmath0 , the distance of the particles from the wall , and that close to the surface the typical relaxation times are orders of magnitude larger than the ones in the bulk . because of the cooperative nature of the particle dynamics
, the slow dynamics also affects the dynamics of the particles for large values of @xmath0 . using various empirical laws , we are able to parameterize accurately the @xmath1dependence of the generalized incoherent intermediate scattering function @xmath2 and also the spatial dependence of structural relaxation times .
these laws allow us to determine various dynamical length scales and we find that their temperature dependence is compatible with an arrhenius law .
furthermore , we find that at low temperatures time and space dependent correlation function fulfill a generalized factorization property similar to the one predicted by mode - coupling theory for bulk systems . for thin films and/or at sufficiently low temperatures
, we find that the relaxation dynamics is influenced by the two walls in a strongly non - linear way in that the slowing down is much stronger than the one expected from the presence of only one confining wall .
finally we study the average dynamics of all liquid particles and find that the data can be described very well by a superposition of two relaxation processes that have clearly separated time scales .
since this is in contrast with the result of our analysis of the local dynamics , we argue that a correct interpretation of experimental data can be rather difficult . |
experimental measurements of the inclusive momentum distribution of charged particles in collisions provide important insight into the process of how quarks turn into hadrons .
this distribution is commonly normalised to the total hadronic cross section @xmath48 and presented as a function of the scaled momenta @xmath49 of the charged hadrons , where is the centre - of - mass energy . in this form ,
the spectrum is usually referred to as the fragmentation function and can be obtained experimentally from the total number of hadronic final states , @xmath50 , and the number of charged particles in each bin , @xmath51 : @xmath52 the charged particle momentum spectrum can also be studied as the distribution of @xmath1 .
the distribution emphasises the low momentum component and the distribution the high momentum component of the momentum spectrum . in the nave
quark parton model , the scaled momentum distribution is expected to be independent from the centre - of - mass energy . a violation of this scaling is expected due to gluon radiation in the final state . experimentally , scaling violation in fragmentation functions had indeed been observed by combining measurements at different centre - of - mass energies and could be used to determine @xcite .
the position of the maximum of the distribution , , has been studied in the past in various experiments ( see for example @xcite and references therein ) .
the energy dependence of the position of the maximum provides an important test of the qcd prediction for the emission of soft gluons @xcite . in events with a heavy primary quark
, the possibility of cascade decays of bottom or charm hadrons results in more particles sharing the same energy than in light quark events and a softer momentum spectrum can be expected . since the flavour composition of the primary quarks in @xmath53
is predicted by the electroweak theory to change with centre - of - mass energy , this flavour dependence of the momentum spectra affects the energy dependence of the and distributions of the inclusive event sample . to correct for this contribution , in @xcite not only the inclusive fragmentation function
was measured but also fragmentation functions in event samples with different flavour compositions were studied . in @xcite , measurements of fragmentation functions in samples with different flavour composition were used to extract flavour dependent fragmentation functions for events with primary light ( uds ) , charm ( c ) or bottom ( b ) quarks . here
we present a measurement of flavour dependent fragmentation functions , based on the methods developed for the opal measurement of charged particle multiplicities in uds , c and b quark events @xcite .
events were divided into two hemispheres by a plane perpendicular to the thrust axis .
secondary vertices and reconstructed mesons were used to tag hemispheres to create samples of events with different quark flavour mixtures ( section 2 ) . to reduce the biases induced by the tagging ,
the measurement of the fragmentation functions was based on the momentum spectrum of charged particles in the event hemisphere opposite to the tag .
corrections for hemisphere correlations and for distortions due to detector effects are described in section 3 .
the flavour dependent and distributions were obtained from a simultaneous fit to the momentum spectra of the different hemisphere samples ( section 4 ) .
for the first time , the measured position of the maximum of the distribution , , is presented separately for uds , c and b events .
a measurement of the inclusive distribution of all five flavours was also performed , based on the track momentum spectrum of all events , i.e. , without considering any flavour tagging .
a complete description of the opal detector can be found elsewhere @xcite .
this analysis relied on the precise reconstruction of charged particles in the central detector , consisting of a silicon microvertex detector , a vertex drift chamber , a large volume jet chamber and chambers measuring the @xmath54-coordinate along the electron beam direction and with positive @xmath55 pointing towards the centre of the lep ring .
the polar angle @xmath56 is defined relative to the @xmath57 axis and the azimuthal angle @xmath58 relative to the @xmath59 axis .
] of tracks as they leave the jet chamber .
this analysis used data recorded with the opal detector in the years 1990 to 1995 at centre - of - mass energies around 91.2 gev comprising an integrated luminosity of about 177 pb@xmath60 .
decays were selected using the criteria described in @xcite . to ensure that most charged particles were well contained in the detector ,
the polar angle of the thrust axis was required to satisfy @xmath61 . to reduce systematic errors in the application of the secondary vertex tag
, it was only based on a homogeneous data sample taken in the year 1994 , representing an integrated luminosity of about 34 pb@xmath60 .
the full integrated luminosity was used in the case of the meson tag .
charged tracks used in the measurement of the fragmentation function were required to have a measured momentum in the plane , @xmath62 , of at least 0.150 gev/@xmath63 and to satisfy @xmath64 cm , where @xmath65 is the distance of closest approach to the origin in the @xmath55-@xmath66 plane .
simulated hadronic decays were generated with the jetset 7.4 monte carlo program @xcite tuned to opal data @xcite .
the events were passed through a detailed simulation of the opal detector @xcite and processed using the same reconstruction and selection algorithms as the data
. samples with varying purity of b quark events were selected by reconstructing secondary vertices , following the procedure described in @xcite .
events were divided into two hemispheres by the plane perpendicular to the thrust axis and comprising the interaction point .
jets were reconstructed by combining charged tracks and electromagnetic clusters not associated to tracks , using the scaled invariant mass algorithm described in @xcite with the jade - e0 recombination scheme and the invariant mass cut - off being set to @xmath67 . a vertex fit
was then attempted in the highest energy jet in each hemisphere separately .
each track used in these vertex fits was required to have at least one hit in the silicon microvertex detector .
all such tracks in the jet were fitted to a common vertex point in the @xmath55-@xmath66 plane and the track with the largest contribution to the @xmath68 was removed if this contribution was greater than four .
the remaining tracks were then refitted until either all tracks contributed less than four to the @xmath68 or there were fewer than four remaining tracks .
for each successfully reconstructed secondary vertex , the projected decay length @xmath69 in the @xmath55-@xmath66 plane with respect to the primary vertex was calculated , where the primary vertex was reconstructed from all tracks in the event together with a constraint to the average beamspot position as in @xcite .
the decay length significance , i.e. the decay length divided by its uncertainty was used to obtain three event samples @xmath70 of varying b flavour purity . according to the simulation , the b purities @xmath71 in these samples vary from @xmath72 to @xmath73 ( see table [ table1 ] ) .
fig [ plot8](a ) shows the distribution of the decay length significance in data and in monte carlo . | fragmentation functions for charged particles in @xmath0 events have been measured for bottom ( b ) , charm ( c ) and light ( uds ) quarks as well as for all flavours together .
the results are based on data recorded between 1990 and 1995 using the opal detector at lep .
event samples with different flavour compositions were formed using reconstructed mesons and secondary vertices .
the @xmath1 distributions and the position of their maxima are also presented separately for uds , c and b quark events .
the fragmentation function for b quarks is significantly softer than for uds quarks .
european laboratory for particle physics cern - ep/98 - 089 + 29 may 1998 * measurements of flavour dependent fragmentation functions in @xmath2 events * the opal collaboration + submitted to european physics journal c + the opal collaboration k.ackerstaff@xmath3 , g.alexander@xmath4 , j.allison@xmath5 , n.altekamp@xmath6 , k.j.anderson@xmath7 , s.anderson@xmath8 , s.arcelli@xmath9 , s.asai@xmath10 , s.f.ashby@xmath11 , d.axen@xmath12 , g.azuelos@xmath13 , a.h.ball@xmath14 , e.barberio@xmath3 , r.j.barlow@xmath5 , r.bartoldus@xmath15 , j.r.batley@xmath6 , s.baumann@xmath15 , j.bechtluft@xmath16 , t.behnke@xmath3 , k.w.bell@xmath17 , g.bella@xmath4 , s.bentvelsen@xmath3 , s.bethke@xmath16 , s.betts@xmath18 , o.biebel@xmath16 , a.biguzzi@xmath6 , s.d.bird@xmath5 , v.blobel@xmath19 , i.j.bloodworth@xmath11 , m.bobinski@xmath20 , p.bock@xmath21 , j.bhme@xmath16 , m.boutemeur@xmath22 , s.braibant@xmath3 , p.bright-thomas@xmath11 , r.m.brown@xmath17 , h.j.burckhart@xmath3 , c.burgard@xmath3 , r.brgin@xmath20 , p.capiluppi@xmath9 , r.k.carnegie@xmath23 , a.a.carter@xmath24 , j.r.carter@xmath6 , c.y.chang@xmath14 , d.g.charlton@xmath25 , d.chrisman@xmath26 , c.ciocca@xmath9 , p.e.l.clarke@xmath18 , e.clay@xmath18 , i.cohen@xmath4 , j.e.conboy@xmath18 , o.c.cooke@xmath3 , c.couyoumtzelis@xmath24 , r.l.coxe@xmath7 , m.cuffiani@xmath9 , s.dado@xmath27 , g.m.dallavalle@xmath9 , r.davis@xmath28 , s.de jong@xmath8 , l.a.del pozo@xmath26 , a.de roeck@xmath3 , k.desch@xmath3 , b.dienes@xmath29 , m.s.dixit@xmath30 , m.doucet@xmath31 , j.dubbert@xmath22 , e.duchovni@xmath32 , g.duckeck@xmath22 , i.p.duerdoth@xmath5 , d.eatough@xmath5 , p.g.estabrooks@xmath23 , e.etzion@xmath4 , h.g.evans@xmath7 , f.fabbri@xmath9 , a.fanfani@xmath9 , m.fanti@xmath9 , a.a.faust@xmath28 , f.fiedler@xmath19 , m.fierro@xmath9 , h.m.fischer@xmath15 , i.fleck@xmath3 , r.folman@xmath32 , a.frtjes@xmath3 , d.i.futyan@xmath5 , p.gagnon@xmath30 , j.w.gary@xmath26 , j.gascon@xmath31 , s.m.gascon-shotkin@xmath14 , c.geich-gimbel@xmath15 , t.geralis@xmath17 , g.giacomelli@xmath9 , p.giacomelli@xmath9 , v.gibson@xmath6 , w.r.gibson@xmath24 , d.m.gingrich@xmath33 , d.glenzinski@xmath7 , j.goldberg@xmath27 , w.gorn@xmath26 , c.grandi@xmath9 , e.gross@xmath32 , j.grunhaus@xmath4 , m.gruw@xmath19 , g.g.hanson@xmath8 , m.hansroul@xmath3 , m.hapke@xmath24 , c.k.hargrove@xmath30 , c.hartmann@xmath15 , m.hauschild@xmath3 , c.m.hawkes@xmath6 , r.hawkings@xmath19 , r.j.hemingway@xmath23 , m.herndon@xmath14 , g.herten@xmath20 , r.d.heuer@xmath3 , m.d.hildreth@xmath3 , j.c.hill@xmath6 , s.j.hillier@xmath11 , p.r.hobson@xmath34 , a.hocker@xmath7 , r.j.homer@xmath11 , a.k.honma@xmath35 , d.horvth@xmath36 , k.r.hossain@xmath28 , r.howard@xmath12 , p.hntemeyer@xmath19 , p.igo-kemenes@xmath21 , d.c.imrie@xmath34 , k.ishii@xmath10 , f.r.jacob@xmath17 , a.jawahery@xmath14 , h.jeremie@xmath31 , m.jimack@xmath11 , a.joly@xmath31 , c.r.jones@xmath6 , p.jovanovic@xmath11 , t.r.junk@xmath3 , d.karlen@xmath23 , v.kartvelishvili@xmath5 , k.kawagoe@xmath10 , t.kawamoto@xmath10 , p.i.kayal@xmath28 , r.k.keeler@xmath37 , r.g.kellogg@xmath14 , b.w.kennedy@xmath17 , a.klier@xmath32 , s.kluth@xmath3 , t.kobayashi@xmath10 , m.kobel@xmath38 , d.s.koetke@xmath23 , t.p.kokott@xmath15 , m.kolrep@xmath20 , s.komamiya@xmath10 , r.v.kowalewski@xmath37 , t.kress@xmath21 , p.krieger@xmath23 , j.von krogh@xmath21 , p.kyberd@xmath24 , g.d.lafferty@xmath5 , d.lanske@xmath16 , j.lauber@xmath18 , s.r.lautenschlager@xmath39 , i.lawson@xmath37 , j.g.layter@xmath26 , d.lazic@xmath27 , a.m.lee@xmath39 , e.lefebvre@xmath31 , d.lellouch@xmath32 , j.letts@xmath8 , l.levinson@xmath32 , r.liebisch@xmath21 , b.list@xmath3 , c.littlewood@xmath6 , a.w.lloyd@xmath11 , s.l.lloyd@xmath24 , f.k.loebinger@xmath5 , g.d.long@xmath37 , m.j.losty@xmath30 , j.ludwig@xmath20 , d.lui@xmath8 , a.macchiolo@xmath9 , a.macpherson@xmath28 , m.mannelli@xmath3 , s.marcellini@xmath9 , c.markopoulos@xmath24 , a.j.martin@xmath24 , j.p.martin@xmath31 , g.martinez@xmath14 , t.mashimo@xmath10 , p.mttig@xmath32 , w.j.mcdonald@xmath28 , j.mckenna@xmath12 , e.a.mckigney@xmath18 , t.j.mcmahon@xmath11 , r.a.mcpherson@xmath37 , f.meijers@xmath3 , s.menke@xmath15 , f.s.merritt@xmath7 , h.mes@xmath30 , j.meyer@xmath19 , a.michelini@xmath9 , s.mihara@xmath10 , g.mikenberg@xmath32 , d.j.miller@xmath18 , r.mir@xmath32 , w.mohr@xmath20 , a.montanari@xmath9 , t.mori@xmath10 , k.nagai@xmath32 , i.nakamura@xmath10 , h.a.neal@xmath8 , b.nellen@xmath15 , r.nisius@xmath3 , s.w.oneale@xmath11 , f.g.oakham@xmath30 , f.odorici@xmath9 , h.o.ogren@xmath8 , m.j.oreglia@xmath7 , s.orito@xmath10 , j.plinks@xmath29 , g.psztor@xmath40 , j.r.pater@xmath5 , g.n.patrick@xmath17 , j.patt@xmath20 , r.perez-ochoa@xmath3 , s.petzold@xmath19 , p.pfeifenschneider@xmath16 , j.e.pilcher@xmath7 , j.pinfold@xmath28 , d.e.plane@xmath3 , p.poffenberger@xmath37 , b.poli@xmath9 , j.polok@xmath3 , m.przybycie@xmath3 , c.rembser@xmath3 , h.rick@xmath3 , s.robertson@xmath37 , s.a.robins@xmath27 , n.rodning@xmath28 , j.m.roney@xmath37 , k.roscoe@xmath5 , a.m.rossi@xmath9 , y.rozen@xmath27 , k.runge@xmath20 , o.runolfsson@xmath3 , d.r.rust@xmath8 , k.sachs@xmath20 , t.saeki@xmath10 , o.sahr@xmath22 , w.m.sang@xmath34 , e.k.g.sarkisyan@xmath4 , c.sbarra@xmath12 , a.d.schaile@xmath22 , o.schaile@xmath22 , f.scharf@xmath15 , p.scharff-hansen@xmath3 , j.schieck@xmath21 , b.schmitt@xmath3 , s.schmitt@xmath21 , a.schning@xmath3 , t.schorner@xmath22 , m.schrder@xmath3 , m.schumacher@xmath15 , c.schwick@xmath3 , w.g.scott@xmath17 , r.seuster@xmath16 , t.g.shears@xmath3 , b.c.shen@xmath26 , c.h.shepherd-themistocleous@xmath3 , p.sherwood@xmath18 , g.p.siroli@xmath9 , a.sittler@xmath19 , a.skuja@xmath14 , a.m.smith@xmath3 , g.a.snow@xmath14 , r.sobie@xmath37 , s.sldner-rembold@xmath20 , m.sproston@xmath17 , a.stahl@xmath15 , k.stephens@xmath5 , j.steuerer@xmath19 , k.stoll@xmath20 , d.strom@xmath41 , r.strhmer@xmath22 , r.tafirout@xmath31 , s.d.talbot@xmath11 , s.tanaka@xmath10 , p.taras@xmath31 , s.tarem@xmath27 , r.teuscher@xmath3 , m.thiergen@xmath20 , m.a.thomson@xmath3 , e.von trne@xmath15 , e.torrence@xmath3 , s.towers@xmath23 , i.trigger@xmath31 , z.trcsnyi@xmath42 , e.tsur@xmath4 , a.s.turcot@xmath7 , m.f.turner-watson@xmath3 , r.van kooten@xmath8 , p.vannerem@xmath20 , m.verzocchi@xmath20 , p.vikas@xmath31 , h.voss@xmath15 , f.wckerle@xmath20 , a.wagner@xmath19 , c.p.ward@xmath6 , d.r.ward@xmath6 , p.m.watkins@xmath11 , a.t.watson@xmath11 , n.k.watson@xmath11 , p.s.wells@xmath3 , n.wermes@xmath15 , j.s.white@xmath37 , g.w.wilson@xmath16 , j.a.wilson@xmath11 , t.r.wyatt@xmath5 , s.yamashita@xmath10 , g.yekutieli@xmath32 , v.zacek@xmath31 , d.zer-zion@xmath3 @xmath11school of physics and astronomy , university of birmingham , birmingham b15 2tt , uk @xmath9dipartimento di fisica dell universit di bologna and infn , i-40126 bologna , italy @xmath15physikalisches institut , universitt bonn , d-53115 bonn , germany @xmath26department of physics , university of california , riverside ca 92521 , usa @xmath6cavendish laboratory , cambridge cb3 0he , uk @xmath23ottawa - carleton institute for physics , department of physics , carleton university , ottawa , ontario k1s 5b6 , canada @xmath30centre for research in particle physics , carleton university , ottawa , ontario k1s 5b6 , canada @xmath3cern , european organisation for particle physics , ch-1211 geneva 23 , switzerland @xmath7enrico fermi institute and department of physics , university of chicago , chicago il 60637 , usa @xmath20fakultt fr physik , albert ludwigs universitt , d-79104 freiburg , germany @xmath21physikalisches institut , universitt heidelberg , d-69120 heidelberg , germany @xmath8indiana university , department of physics , swain hall west 117 , bloomington in 47405 , usa @xmath24queen mary and westfield college , university of london , london e1 4ns , uk @xmath16technische hochschule aachen , iii physikalisches institut , sommerfeldstrasse 26 - 28 , d-52056 aachen , germany @xmath18university college london , london wc1e 6bt , uk @xmath5department of physics , schuster laboratory , the university , manchester m13 9pl , uk @xmath14department of physics , university of maryland , college park , md 20742 , usa @xmath31laboratoire de physique nuclaire , universit de montral , montral , quebec h3c 3j7 , canada @xmath41university of oregon , department of physics , eugene or 97403 , usa @xmath17rutherford appleton laboratory , chilton , didcot , oxfordshire ox11 0qx , uk @xmath27department of physics , technion - israel institute of technology , haifa 32000 , israel @xmath4department of physics and astronomy , tel aviv university , tel aviv 69978 , israel @xmath10international centre for elementary particle physics and department of physics , university of tokyo , tokyo 113 , and kobe university , kobe 657 , japan @xmath34institute of physical and environmental sciences , brunel university , uxbridge , middlesex ub8 3ph , uk @xmath32particle physics department , weizmann institute of science , rehovot 76100 , israel @xmath19universitt
hamburg / desy , ii institut fr experimental physik , notkestrasse 85 , d-22607 hamburg , germany @xmath37university of victoria , department of physics , p o box 3055 , victoria bc v8w 3p6 , canada @xmath12university of british columbia , department of physics , vancouver bc v6 t 1z1 , canada @xmath28university of alberta , department of physics , edmonton ab t6 g 2j1 , canada @xmath39duke university , dept of physics , durham , nc 27708 - 0305 , usa @xmath40research institute for particle and nuclear physics , h-1525 budapest , p o box 49 , hungary @xmath42institute of nuclear research , h-4001 debrecen , p o box 51 , hungary @xmath22ludwigs - maximilians - universitt mnchen , sektion physik , am coulombwall 1 , d-85748 garching , germany @xmath43 and at triumf , vancouver , canada v6 t 2a3 @xmath44 and royal society university research fellow @xmath45 and institute of nuclear research , debrecen , hungary @xmath46 and department of experimental physics , lajos kossuth university , debrecen , hungary @xmath47 on leave of absence from the university of freiburg |
so far , every 3d effort to perform 3 + 1 numerical simulations of space - times containing black - hole singularities , or compact objects such as neutron stars , has encountered the frustration of producing short - lived evolutions .
the impediment is due to the presence of sever instabilities .
there is a common agreement among researchers that the origin of these instabilities is not purely numerical .
inconsistent boundary conditions far away from the holes or neutron stars , unsuitable choices of gauge or coordinate conditions , inappropriate algorithms to excise the black hole singularity , etc .
are all contributing factors . since approximately the mid 90 s
, the numerical relativity community has been accumulating evidence supporting the view that a crucial aspect determining the life - time of simulations is the form of the einstein equation that is implemented in the numerical codes .
when viewed as an initial value problem , the einstein equation can be cast in what is commonly known as a 3 + 1 form ; that is , the einstein equation becomes a set of constraint and evolution equations that govern the history of the geometry of space - like hypersurfaces , geometrodynamics .
there are many ways of writing the einstein equation in a 3 + 1 form . during the early years of numerical relativity ,
the most popular form was that introduced by arnowitt , deser and misner ( adm ) @xcite .
it is safe to say that all of the attempts to obtain long - term stable and convergent , 3d simulations of black holes using codes based on the adm formulation have failed .
these efforts have produced short lived evolutions lasting @xmath0 , with @xmath1 the mass of the black hole .
unfortunately , the reasons behind this failure are still not known . nonetheless , because of all of these failed attempts , the numerical community has practically abandoned the adm system .
one alternative to adm has been formulations of the einstein equation that are explicitly hyperbolic @xcite .
these formulations have been introduced in part to facilitate handling both the outer boundary of the computational domain and the excision boundary in the vicinity of the black hole singularity .
hyperbolic formulations certainly have the advantage that , by knowing the characteristics of each of the evolved fields , one is able to apply the correct boundary conditions to those fields that require them .
hyperbolic formulations also enjoy the advantage of the mathematical machinery to prove existence , uniqueness and well - posedness .
unfortunately , hyperbolic formulations have so far been able to produce evolutions of single black holes lasting @xmath2 @xcite .
another formulation of the einstein equation that has been relatively successful is the one re - introduced by baumgarte and shapiro @xcite and originally developed by shibata and nakamura @xcite .
this formulation , commonly known as the bssn system , belongs to the class of conformal traceless formulations ( see next section for a detailed derivation ) .
there are several ingredients that go into the derivation of these equations . besides the use of conformal transformations and a traceless decomposition of the extrinsic curvature , perhaps the most important aspect is the introduction of a connection . with this connection
, the formulation contains some hyperbolic flavor .
studies have shown , however , that 3d , single black hole evolutions are still limited to life - times @xmath3 @xcite .
because of the relative success of the hyperbolic and bssn systems over the adm formulation , there seems to be a certain degree of consensus that a successful formulation should have some hyperbolic flavor .
the formulation we introduce in this paper retains this property .
the innovative aspect of this system enters in shifting the focus from the principal part of the evolutions equations , i.e. those terms containing differentiation , to non - linear algebraic terms .
our work is motivated by a recent numerical study of evolutions of single black holes in spherical symmetry @xcite using the adm system . in ref .
@xcite , long - term stable and convergent evolutions were achieved by `` adjusting '' the adm system .
the adjusting consisted of eliminating or switching the sign of non - linear terms involving the extrinsic curvature .
the stability that was obtained was remarkable .
unfortunately , an extension to 3d is not possible because the adjustment took advantage of the spherical symmetry in the problem . as we will show , it is possible to achieve a similar elimination or switching of sign of non - linear terms . instead of working with the adm system , one needs to consider a modified form of the bssn formulation .
we will present 3d numerical simulations of single , non - rotating black holes in which we demonstrate a remarkable improvement in the stability and duration of the evolutions when compared with those based on bssn .
in order to gain a better understanding of the elements present in the conformal traceless system here considered , we will review the derivation of the bssn formulation .
the traditional derivation of the bssn system starts with the adm formulation .
the adm primary variables are the spatial metric , @xmath4 , and extrinsic curvature , @xmath5 , of the space - like hypersurfaces .
the equations that govern the dynamics of @xmath4 and @xmath5 are _ o g_ij&= & - 2 k_ij [ gdot ] + _ o k_ij&= & - _ i _ j + r_ij + & + & ( k k_ij - 2 k_ik k^k_j ) , [ kdot ] where @xmath6 and @xmath7 although we concentrate our attention on the vacuum case , the inclusion of matter sources is straightforward .
above , @xmath8 and @xmath9 denote covariant differentiation and ricci tensor associated with @xmath4 , respectively .
the lapse function , @xmath10 , and shift vector , @xmath11 , encapsulate the diffeomorphism freedom of general relativity and are thus freely specifiable . in terms of the adm variables ,
the einstein constraints read r + k^2 - k_ij k^ij & = & 0 [ hamcons ] + _ j k^ij - ^i k & = & 0 .[momcons ] as customary in numerical relativity , evolutions are performed without ensuring that the evolved data continue to satisfy the constraints . eqs .
( [ hamcons ] ) and ( [ momcons ] ) are only used to monitor the accuracy and stability of the simulations .
the first step in obtaining the bssn formulation is to abandon @xmath4 and @xmath5 as primary variables and work instead with @xmath12 and @xmath13 .
the relationships between these and the adm variables are [ eq : phi ] & = & g^1/2 + [ eq : ghat ] g_ij & = & e^- 4 g_ij + [ eq : a ] a_ij & = & e^- 4 a_ij , where @xmath14 . the first point to notice
is that ( [ eq : phi ] ) and ( [ eq : ghat ] ) imply that the conformal metric @xmath15 has unit determinant . enforcing this condition in numerical evolutions
has an important impact on the stability of the simulations . unless noted , indices will be raised and lowered with the conformal metric @xmath15 . given ( [ eq : phi])-([eq : a ] ) , it is not difficult to show that the adm system ( [ gdot])-([kdot ] ) of evolution equations take the form [ phidot ] _ o & = & - k + [ gdot2 ] _ o g_ij & = & - 2 a_ij + [ kdot2 ] _ o k & = & - _ i ^i + ( a_ij a^ij + k^2/3 ) + [ adot ] _ o a_ij & = & e^- 4 ( - _ i _ j + r_ij ) ^tf + & + & ( k a_ij - 2 a_il a^l _ j ) . before continuing , there are several points to be made about the derivation and structure of these equations .
first , in deriving eq .
( [ kdot2 ] ) , the hamiltonian constraint ( [ hamcons ] ) was used to eliminate the ricci scalar . in eq .
( [ adot ] ) , the superscript @xmath16 denotes the trace - free part of the tensor between brackets , e.g. @xmath17 for any tensor @xmath18 . when calculating @xmath19 , the trace @xmath20 is again eliminated using the hamiltonian constraint ( [ hamcons ] ) .
to simplify notation , we did not explicitly transform the terms involving covariant derivatives of the lapse or the ricci tensor , @xmath9 .
these terms should be understood as being given respectively by : [ lapse ] _ i _ j & = & _ i _ j - 4_(i_j ) + & + & 2 g_ij ^l_l + [ ricci ] r_ij & = & r_ij -2 _ i _ j - 2 g_ij ^l _
l + & + & 4 _ i _ j - 4 g_ij ^l _ l , where @xmath21 is the ricci tensor computed from @xmath15 .
one should also keep in mind that @xmath15 and @xmath13 are tensor densities of weight @xmath22 , so their lie derivatives are _
a_ij = ^k _ k a_ij + a_ik _ j
^k + a_kj _ i ^k - a_ij _ k ^k and a similar expression for @xmath15 .
the function @xmath23 is not a true scalar , see eq .
( [ eq : phi ] ) .
its lie derivative is given by _
= ^k _ k + _ k ^k .
next is perhaps the key ingredient of the bssn system .
one introduces a conformal connection [ cgsf ] ^i g^jk ^i_jk = - _ j g^ij , where the @xmath24 are the connection coefficients associated with @xmath15 . the last equality holds because of the condition @xmath25
. the motivation for introducing these connections is to substitute in the ricci tensor @xmath21 all the derivatives of @xmath26 in favor of derivatives of @xmath27 . in doing so , the ricci tensor yields a system of evolution equations @xcite with `` hyperbolic flavor . ''
expanding the number of primary variables in the system to include the conformal connections implies the need of an additional evolution equation . from definition ( [ cgsf ] )
, one can show that the evolution equation for @xmath28 is given by [ eq : gamma ] _
o ^i & = & g^jk_jk^i + g^ij_jk^k - 2a^ij_j + & + & 2^i_jka^jk + 12a^ij_j - ^ik . in deriving this equation , the momentum constraint ( [ momcons ] ) was used to eliminate derivatives of @xmath29 . a final point to be made
, @xmath28 is a vector density of weight 2/3 , so _ ^i = ^k _ k ^i - ^k_k ^i + ^i_k ^k . in general terms , the bssn system involves : conformal transformations , a trace - free decomposition of the extrinsic curvature , the introduction of a conformal connection and the use of constraints to eliminate the ricci scalar and derivatives of the trace - free extrinsic curvature .
we are now prepared to derive the new system of evolution equations .
because this system retains the main ingredients of the bssn formulation , its derivation proceeds along a similar path .
however , the starting point is modified . instead of ( [ eq : phi])-([eq : a ] )
, we use [ nct : phi ] & = & g^1/2 + [ nct : ghat ] g_ij & = & e^- 4 g_ij + [ nct : k ] k & = & e^6nk + [ nct : a ] a^i_j & = & e^6na^i_j + [ nct : lapse ] n & = & e^-6n , with @xmath30 a parameter to be determined later . notice the main differences .
we allow for the trace of the extrinsic curvature and the lapse function to be conformally transformed .
@xmath31 and @xmath32 are densities .
also , we use as primary variable the mixed - index form of the trace - less part of the extrinsic curvature , @xmath33 . if @xmath34 , one recovers the bssn scalings . with the above transformations , the evolution equations take the following form [ nct : phidot ] _ o & = & - n k + [ nct : gdot2 ] _ o g_ij & = & - 2 n a_ij + [ nct : kdot2 ] _ o k & = & -e^6n _ i ^i + na^i_j a^j_i + & + & ( 1 - 3n)nk^2/3 + [ nct : adot ] _
o a^i_j & = & e^6n ( - ^i _ j + r^i_j ) ^tf + & + & ( 1-n)nk a^i_j + [ nct : gamma ] _ o ^i & = & g^jk_jk^i + g^ij_jk^k - 2a^ij _ j n + & + & 2n^i_jka^jk - ng^ij_jk + & + & 12(1-n)na^ij_j . here again , the terms involving @xmath10 and @xmath9 should be understood as being given by ( [ lapse ] ) and ( [ ricci ] ) , with the index raised with @xmath35 .
in addition , the lapse expression ( [ lapse ] ) should be modified with the appropriate substitution in terms of the densitized lapse @xmath32 . by fixing the attention to eq .
( [ nct : adot ] ) , one can understand the motivation behind the particular form of the conformal scalings and the use of @xmath33 as a primary variable . if we choose @xmath36 , the term @xmath37 in eq .
( [ nct : adot ] ) could be eliminated or have its sign switched . in the spherically symmetric numerical study in ref .
@xcite , it was demonstrated that this algebraic term was responsible for the instabilities of the adm - based simulations . in ref .
@xcite , the changes in sign or eliminations of these terms were obtained by substitution of the momentum constraint .
this adjustment of the adm system is not possible in non - spherical symmetry .
however , it is clear that the conformal scalings we have introduced accomplish the same effect .
the importance of eliminating @xmath37 can be understood by the following hand - waving argument .
equation ( [ nct : adot ] ) has the following structure [ eq : hand ] _
t a^i_j & = & c a^i_j + & + & a + a where c n_k^k + ( 1-n)nk . at the continuum level , for a single , static black hole , the r.h.s .
( [ eq : hand ] ) should vanish identically .
however , because of truncation errors , this will not be the case in numerical evolutions .
if this cancellation does not take place , a solution of the form @xmath38 is in principle allowed .
thus , @xmath39 has the potential of yielding an exponential growth .
almost all the numerical evolutions of single black holes have been performed with space - time foliations such that @xmath40 .
since the simulations of physical interested are future directed ( i.e. @xmath41 ) , then @xmath42 . as a consequence ,
the second term in @xmath43 has the `` wrong '' sign . because of the conformal scalings we have introduced , values of @xmath44 correct this sign problem .
the first term in @xmath43 gives the impression that it also has the wrong sign .
this is not the case .
all of the single black hole solutions of interest have shift vectors for which @xmath45 . for more generic , multiple black hole situations
, it should be possible to construct shift vectors that preserve this property .
if not , a large enough value of @xmath30 should be able to yield the appropriate sign of @xmath43 . finally , we point out that this hand - waving argument provides a reasonable explanation of why the simulations in @xcite with @xmath46 were significantly more stable than those with positive @xmath47 .
the use of a mixed - index , traceless extrinsic curvature in 3d studies is not new .
one of us used this approach in numerical studies of inhomogeneous inflationary space - times @xcite .
the evolution equations used in @xcite did not include a conformal connection @xmath27 nor a densitized lapse .
the use of densitized variables had a different motivation in @xcite . with those densitized variables , the shift terms in the evolution equations
could be written in an almost flux - conservative form .
however , it was quickly realized that , because during an inflationary epoch the determinant of @xmath4 grows exponentially , working with positive weight , densitized variables becomes eventually problematic .
our attempts to obtain stable 3d , single black hole evolutions with a code based on the evolution system in ref .
@xcite have failed .
given the success of the results in the present work , we believe that the failure of the system in @xcite is likely due to the absence of the conformal connection @xmath27 .
there are indications that in spherical symmetry , using flux - conservative equations seems to extend the life of the simulations @xcite .
this view is also supported by recent work using a hyperbolic formulation @xcite in which the evolution equations are written in a flux - conservative form .
finally , mixed - index extrinsic curvature , not traceless , as a primary variable has been used in 1d evolutions of plane - symmetric cosmologies @xcite and gravitational waves @xcite , supporting the view that this form is more advantageous .
to test the stability of the new system , we concentrate our attention on a single , non - rotating black hole .
we use the ingoing - eddington - finkelstein form of the black hole solution .
the black hole singularity is handled via excision .
that is , a region inside the horizon , containing the black hole singularity , is removed from the computational domain .
our excision has a cubical shape . at its boundary , the values of the primary variables
@xmath48 are obtained by third order extrapolation of the corresponding evolved values . at the outer boundary
, the evolved data are blended to the analytic solution .
temporal updating is performed via the iterated crank - nicholson method . after each time - step , we impose the constraints @xmath25 and @xmath49 as follows : a^i_j & & a^i_j - g^i_j a^k_k/3 + g_ij & & g^-1/3g_ij . once the new @xmath15 is obtained , we use definition ( [ cgsf ] ) to also reset @xmath28 .
we do not write the shift terms in an almost flux - conservative form as done in ref . @xcite .
all the discretization of spatial operators is done with centered finite differences .
the exception is shift terms of the form @xmath50 .
those terms are approximated using an up - wind discretization described in ref .
@xcite .
since it is well known that the choice of gauge or coordinate conditions has a fundamental impact on the duration of the evolutions , we decided to test the new system on the most challenging case , namely the gauge condition in which @xmath32 and @xmath11 are obtained from the exact analytic solution .
figure [ fig1 ] shows the l2 norm of the hamiltonian constraint and the error in the mass of the black hole from a 3d evolution .
the location of the outer boundary of the computational domain was set to @xmath51 and the grid - spacing to @xmath52 .
we imposed quadrant symmetry , namely reflection symmetry across the @xmath53 and @xmath54 planes .
this choice was motivated by the work in ref .
@xcite , in which long term stability was only achieved in octant symmetry and with live gauge conditions .
we performed evolutions with higher resolutions in octant symmetry only for the purpose of testing second order convergence in our code .
an identical , quadrant - symmetric , run using the bssn system crashed at @xmath55 .
short - lived , bssn evolutions with analytic lapse and shift were also reported in ref . @xcite .
as one can see from fig [ fig1 ] , the new system is able to extend the life - time of the simulation an order of magnitude beyond that of bssn . to test the effects of the symmetry conditions , we repeated a simulation using bitant or equatorial symmetry .
figure [ fig2 ] shows the outcome of the simulation . a comparison of the results in figs .
[ fig1 ] and [ fig2 ] indicate that the duration and quality of the data still depend on the symmetries imposed . although we have managed to extend the duration of the simulation , as seen in fig .
[ fig2 ] , there is a late - onset of an instability .
preliminary experiments indicate that the instability is likely to be connected to the type of outer boundary condition and its placement , as well as the symmetries imposed .
our results are , nonetheless , comparable to evolutions in ref .
@xcite based on a parametrized hyperbolic formulation . in that study , evolutions lasting @xmath56 were obtained for similar computational domain sizes , but without any symmetries applied , using analytic densitized lapse and shift conditions .
a direct comparison with the results in @xcite is not straightforward due to the numerical technique employed , namely spectral methods . finally , we performed evolutions in octant symmetry for @xmath57 .
we obtained similar stability to that with @xmath58 .
the trend observed in these simulations suggests that for values @xmath59 it is very likely to experience a deterioration on the stability of the evolutions .
this should not be surprising .
looking at the conformal transformations ( [ nct : k])-([nct : lapse ] ) , one sees that large values of @xmath30 have the effect of increasing the spatial gradients of these primary variables .
this paper introduces a conformal - traceless formulation of the einstein equation that yields a substantial improvement in the duration of evolutions .
this system has various elements in common with the bssn system .
however , it has the following new aspects : ( 1 ) modification of the conformal scalings of the extrinsic curvature , ( 2 ) use of the mixed - index form of the traceless extrinsic curvature as a primary variable , and ( 3 ) densitization of the lapse function .
the reason behind these modifications is to eliminate or switch the sign of non - linear terms involving the extrinsic curvature .
our results support the view that these terms are partially responsible for rendering the evolutions unstable .
we are currently investigating the degree to which each of the above elements of the formulation here introduced impacts the stability of the simulations@xcite .
we have demonstrated that our 3d simulations of a single black hole exhibit a remarkable improvement when compared with those using the bssn formulation .
this is especially apparent when one considers we have performed these tests in what may be the most troublesome gauge choice , namely obtained directly from the analytic black hole solution .
furthermore , we did not impose outgoing boundary conditions .
these conditions have been shown to improve the duration of simulations in hyperbolic and bssn systems .
we are currently extending our study to investigate whether the stability of this new system is preserved and improved using live gauge conditions and in dynamical situations such as distorted and binary black holes .
this work was supported by nsf grants phy-9800973 and phy-0114375 .
we want to thank a. ashtekar for comments and helpful discussions . we also want to thank the rest the maya project development team : d. fiske , b. kelly , e. schnetter and k. smith . for reviews on hyperbolic methods for einstein s equations :
`` the cauchy problem for the einstein equations '' , h. friedrich and a. rendall , in _ einstein s field equations and their physical interpretation _
, ed . by b. g. schmidt , springer , berlin , 2000 . published in _ lect .
notes phys . _
* 540 * , 127 ( 2000 ) .
e - print archive : gr - qc/000207 . also
: `` hyperbolic methods for einstein s equations '' , o. reula , _ living reviews in relativity _ ( http://www.livingreviews.org ) | there is strong evidence indicating that the particular form used to recast the einstein equation as a 3 + 1 set of evolution equations has a fundamental impact on the stability properties of numerical evolutions involving black holes and/or neutron stars .
presently , the longest lived evolutions have been obtained using a parametrized hyperbolic system developed by kidder , scheel and teukolsky or a conformal - traceless system introduced by baumgarte , shapiro , shibata and nakamura .
we present a new conformal - traceless system .
while this new system has some elements in common with the baumgarte - shapiro - shibata - nakamura system , it differs in both the type of conformal transformations and how the non - linear terms involving the extrinsic curvature are handled .
we show results from 3d numerical evolutions of a single , non - rotating black hole in which we demonstrate that this new system yields a significant improvement in the life - time of the simulations . |
one of the main questions concerning symbolic dynamics and algebraic structures was asked by r. bowen : characterize group shifts , that is shifts supporting a group structure so that the shift map is an automorphism .
this question was answered by b. kitchens @xcite , who showed that any group shift is conjugated to the product of a full shift with a finite set .
a more general case was studied by n.t .
sindhushayana , b. marcus and m. trott @xcite , who proved the analogous result for a homogeneous shift , that is a shift space @xmath0 on the alphabet @xmath1 for which there exist a group @xmath2 of permutations of @xmath1 and a group shift @xmath3 , such that @xmath0 is invariant under the action of any element of @xmath4 .
this work concentrates on quasigroups , often called cancellation semi - groups , thus with left and right cancellable operations . in [ ii2 ] we present sufficient and necessary conditions to a compact zero - dimensional quasi - group @xmath5 , where is defined an expansive automorphism @xmath6 , to be conjugated and isomorphic to a markov shift with a 1-block operation .
furthermore , we give examples of zero - dimensional quasigroups which verify such conditions .
these are quasi - group versions of results of @xcite , and their proofs use a quasi - group version of compact zero - dimensional groups ( @xcite , theorem 16 , pg.77 ) .
we will show that the unique shift spaces which can support a 1-block quasi - group operation are markov shifts .
so , [ ii12 ] is dedicated to study the case when @xmath7 is a markov shift and @xmath8 is a 1-block operation .
there , we characterize completely its structure by supplying a conjugacy with a product of a finite quasigroup with a full shift . in the last section we use amalgamations and state splittings operations ( @xcite,@xcite and @xcite ) , to characterize any isomorphism between two quasi - group shifts as in kitchens
let @xmath1 be a finite alphabet and @xmath9 be the _ two - sided full shift _ endowed with the product topology ( it is a hausdorff compact space ) .
let @xmath10 be a _ shift space _ , that is a closed shift - invariant set , and denote by @xmath11 the alphabet used by @xmath7 .
let @xmath12 be the set of all words or blocks with length @xmath13 which are allowed in @xmath7 ( often we simply write @xmath14 instead of @xmath12 ) .
given @xmath15\in{\mathcal{w}}(\lambda , n)$ ] , we write @xmath16 , or simply @xmath17 , the follower set of @xmath18 : @xmath19\in{\mathcal{w}}(\lambda , n+1)\right\}}.\ ] ] in the same way , we define @xmath20 , or simply @xmath21 , the set of predecessors of @xmath18 .
+ for @xmath22 , @xmath23 , we denote @xmath24:=[x_m , x_{m+1},\ldots , x_n]\in{\mathcal{w}}(n - m+1)$ ] .
+ let @xmath25 be the _ shift map _ defined on @xmath7 , when the context is clear we simply put @xmath26 instead of @xmath25 .
+ we say that @xmath7 is a _ shift of finite type _
( sft ) if there exists @xmath27 such that for any @xmath28 and for all @xmath29 we have @xmath30)={\mathcal{f}}(\mathbf{x}[-n,0])$ ] . in this case
we refer to @xmath7 as a @xmath31-step sft .
if @xmath32 is a transition matrix on the alphabet @xmath1 , denote by @xmath33 the _ two sided markov shift _ and by @xmath34 the alphabet used by @xmath35 . without lost of generality , we can assume that all rows and columns of @xmath32 are not null , what is equivalent to say that @xmath36 .
a markov shift is a @xmath37-step sft .
+ let @xmath38 be a set and @xmath8 a binary operation on @xmath38 .
we say that @xmath39 is a _ quasigroup _ if @xmath8 is left and right cancellable : @xmath40 if , in addition , @xmath38 is a topological space and @xmath8 is continuous with respect to topology of @xmath38 , we say that @xmath39 is a _
topological quasigroup_. when the context is clear , for @xmath41 , we write @xmath42 instead of @xmath43 . a partition @xmath44 of @xmath38 is said to be compatible with @xmath8 if defining @xmath45 , so for all @xmath46 there exists @xmath47 such that @xmath48 , which is equivalent to say @xmath49 is also a quasigroup .
+ suppose that @xmath50 is a topological quasigroup .
then , the shift map is a continuous isomorphism if and only if @xmath8 is given by a @xmath51-block local rule , i.e. , there exists @xmath52 and @xmath53 , such that @xmath54,\mathbf{y}[j-\ell , j+r]).\ ] ] in this case , we say that @xmath55 is the memory and @xmath56 the anticipation of @xmath8 .
when @xmath57 , @xmath8 is a 1-block operation .
+ @xmath58 is a _ topological dynamical system _ if @xmath0 is a compact space and @xmath6 a homeomorphism .
if there exists @xmath59 , such that @xmath60 is dense in @xmath0 , @xmath58 is said to be _ transitive _ or _ irreducible_. two topological dynamical systems @xmath58 and @xmath61 are _ topologically conjugated _ if and only if there exists a homeomorphism @xmath62 , such that @xmath63 .
+ if @xmath5 is a topological quasigroup and @xmath58 a topological dynamical system , such that @xmath6 is an automorphism for @xmath8 , we will denote it by @xmath64 .
we will say that @xmath64 and @xmath65 are isomorphic if and only if there exists @xmath62 , which is both a topological conjugation between @xmath58 and @xmath61 , and an isomorphism between @xmath5 and @xmath66 .
+ if @xmath58 is a topological dynamical system , then its topological entropy @xcite will be denoted by @xmath67 . when we refer to the entropy of a shift @xmath68
, we will write @xmath69 .
in @xcite kitchens proved that if @xmath70 is a topological group and @xmath6 is an automorphism , such that : ( h1 ) : : @xmath0 is compact ( hausdorff ) , zero - dimensional and has a numerable topological basis , that is , each element @xmath71 has a clopen fundamental neighborhood @xmath72 : @xmath73 ( h2 ) : : @xmath74 is an expansive automorphism ; then , * @xmath75 is isomorphic by a 1-block code to @xmath76 , where @xmath77 is a finite group with 1-block operation ; @xmath78 is a full n shift ; and @xmath79 is a @xmath80-block operation , with memory 0 and anticipation @xmath81 . *
if @xmath82 , then @xmath83 , that is , the full shift is trivial . *
if @xmath74 is irreducible , then @xmath84 , that is , @xmath77 is trivial .
recall that expansivity means that _ there exists @xmath85 , a partition of @xmath0 by clopen sets ( which is finite since @xmath0 is compact ) , such that @xmath86 , @xmath87 , there exists @xmath88 such that @xmath89 and @xmath90 belong to distinct sets of @xmath85_. our aim is to extend the previous result to the case when @xmath70 is a topological quasigroup .
now , since a quasigroup has fewer assumptions about its structure , we need some additional hypotheses on @xmath70 .
in particular , it is reasonable to assume the following property : ( h3 ) : : @xmath91 : @xmath92 .
( h3 ) is equivalent to : @xmath93 , @xmath94 , such that @xmath95 and @xmath96 .
furthermore , since @xmath97 is a quasigroup , the elements @xmath98 and @xmath99 are unique . notice that if @xmath70 is a finite quasigroup , then ( h3 ) holds .
under ( h3 ) we can define on @xmath0 the following quasi - group operations @xmath100 and @xmath101 : @xmath102 and @xmath103 also , for any @xmath71 , we can define the functions @xmath104 and @xmath105 by @xmath106 and @xmath107 . in the same way we define @xmath108 and @xmath109 , using the operation @xmath100 , and the functions @xmath110 and @xmath111 , using the operation @xmath101 .
it is easy to check that all of these functions are homeomorphisms .
we recall the identity element plays a fundamental role in the study of zero - dimensional groups ( see @xcite , and @xcite , theorem 16 , pg.77 ) . in the case of zero - dimensional quasi - groups
we will need the hypothesis ( h3 ) to define a substitutive notion : [ ii3 ] for an arbritarily fixed element @xmath112 , given @xmath71 we define @xmath113 and @xmath114 as the unique elements in @xmath0 ( which there exist due ( h3 ) ) , such that @xmath115 and @xmath116 .
we say @xmath113 and @xmath114 are respectively the left and right inverses of @xmath117 with respect to @xmath118 .
notice that @xmath119 and @xmath120 .
moreover , we have that @xmath121 and the maps @xmath122 and @xmath123 are homeomorphisms . in order to reach our goal ,
some additional hypotheses over @xmath70 will be needed .
assume that hypotheses ( h1 ) and ( h2 ) hold for @xmath75 .
we shall prove that there exists a quasi - group shift @xmath50 such that @xmath75 is isomorphic to @xmath124 .
[ ii4 ] let @xmath75 as above .
then @xmath75 is isomorphic to @xmath124 , where @xmath7 is a shift and @xmath8 is a @xmath80-block operation , for some @xmath125
. from expansibility and 0-dimensionality there exists a partition @xmath85 of @xmath0 , a shift space @xmath126 , and a homeomorphism @xmath127 , which is a topological conjugacy between @xmath58 and @xmath68 . in @xmath7 , we define the quasi - group operation @xmath8 , given by : @xmath128 we have that @xmath50 is isomorphic by @xmath129 to @xmath70 .
furthermore , since @xmath130 is an automorphism for @xmath8 , then @xmath8 is @xmath80-block , for some @xmath125 .
in particular we will be interested in the case @xmath8 being a 1-block operation . from the proof of lemma [ ii4 ]
we deduce that @xmath8 is a 1-block operation if and only if the partition @xmath85 is compatible with @xmath97 . for instance , if @xmath0 is a shift space with a 1-block operation @xmath97 , then any partition @xmath85 of @xmath0 by cylinders defined by the same coordinates is compatible with @xmath97 ( which means @xmath75 is isomorphic to @xmath124 , where @xmath8 is 1-block ) .
the natural problem consists in finding such partitions compatible with the operation for any topological quasigroup in which ( h1 ) , ( h2 ) ( and additionally ( h3 ) ) hold .
this problem remain open .
therefore , we can ask for the kind of quasi - group structures allowing to obtain analogous results .
suppose @xmath75 is a topological quasigroup , verifying ( h1 ) , ( h2 ) , and such that the following properties hold : ( h1 ) : : @xmath70 is commutative ( that is @xmath131 ) ; ( h2 ) : : @xmath97 has period 2 , this means , @xmath132 , @xmath133 has period 2 ; ( h3 ) : : the aforementioned element @xmath134 has a fundamental neighborhood system @xmath135 , such that + @xmath136 ( h4 ) : : @xmath70 has the medial property : + @xmath137 + notice that [ ( h1 ) and ( h2 ) ] is equivalent to [ @xmath97 , @xmath100 and @xmath101 are identical ] . thus , for all @xmath71 : @xmath138 .
moreover , these two hypotheses imply that ( h3 ) holds and that the hypothesis ( h3 ) is equivalent to @xmath139 , @xmath140 . furthermore , from ( h3 ) , @xmath141 , which implies @xmath142 .
we notice @xmath118 is _ not _ an identity element , since in general @xmath143 .
hence , by using ( h4 ) , we deduce that for all @xmath144 , the left inverse with respect to @xmath118 verifies : @xmath145 let @xmath146 and let a 2-block operation @xmath97 with local rule @xmath147 be defined for @xmath148 , @xmath149 $ ] and @xmath150 $ ] , by : @xmath151 where @xmath152 denote the sum @xmath153 . then , @xmath70 verifies all of the previous hypotheses , but it is not a group . the next result is a quasi - group version of a construction done on topological zero - dimensional groups ( @xcite , theorem 17 , pg.77 ) . +
[ ii6 ] let @xmath70 be a topological quasigroup , such that ( h1 ) , ( h1 ) , ( h2 ) , ( h3 ) , ( h4 ) hold .
then , given a neighborhood @xmath154 of @xmath118 , there exists a clopen neighborhood @xmath155 of @xmath118 , such that @xmath156 is a finite partition of @xmath0 compatible with @xmath97 . * step 1 : * let @xmath135 be the neighborhood system over @xmath118 , given by ( h3 ) .
we can suppose @xmath157 , @xmath139 .
let @xmath158 be a compact subset such that @xmath159 .
we say that @xmath160 can be connected to @xmath118 over @xmath161 by a chain of order @xmath13 , if and only if there exists a sequence @xmath162 , such that @xmath163 let @xmath164 be the set of all points in @xmath161 , which can be connected to @xmath118 over @xmath161 by a chain of order @xmath13 .
it is straightforward to see that @xmath165 .
moreover , every point in @xmath164 can be connected to @xmath118 over @xmath164 by a chain of order @xmath13 .
in fact , if @xmath166 , then there exist @xmath167 such that @xmath168 , @xmath169 . hence , for any @xmath170 , @xmath171 can be connected to @xmath118 over @xmath161 by a chain of order @xmath13 .
therefore , for any @xmath170 , @xmath172 .
thus , @xmath117 can be connected to @xmath118 over @xmath164 by a chain of order @xmath13 . + * step 2 : * we will show that @xmath164 is a relative open set of @xmath173 the induced topology in @xmath161 .
given @xmath166 , we search for a relative open set of @xmath161 , neighborhood of @xmath117 , that is a subset of @xmath164 .
let @xmath174 , which is a relative open set of @xmath161 , since the continuity of @xmath175 implies that @xmath176 is open .
moreover , @xmath177 , which implies that @xmath178 .
now , let @xmath179 , and @xmath167 be a chain of order @xmath13 connecting @xmath117 to @xmath118 over @xmath161 . since @xmath179 , it is direct to see that @xmath180 is a chain of order @xmath13 connecting @xmath181 to @xmath118 over @xmath161 , because @xmath182 .
thus , @xmath183 , so @xmath184 . +
* step 3 : * we need to prove that @xmath164 is a closed set of @xmath0 or equivalently @xmath185 is open . since @xmath186 , and @xmath187 is open , it is sufficient to show that @xmath188 is open .
let @xmath189 , put @xmath190 , which is an open set since @xmath191 is continuous . notice that @xmath192 and , since @xmath193 , it implies that @xmath194 .
let @xmath195 , which is an open set containing @xmath117 because @xmath196,and @xmath197 .
the set @xmath198 can not intersect @xmath164 .
in fact , if the contrary @xmath199 holds , then it is possible to take @xmath200 such that @xmath201 is a chain of order @xmath13 which connects @xmath181 to @xmath118 over @xmath161 .
but @xmath202 , so @xmath203 and then @xmath204 .
since @xmath160 , @xmath205 is a chain of order @xmath13 which connects @xmath117 to @xmath118 over @xmath161 .
hence , @xmath166 , a contradiction with the assumption @xmath189 .
since @xmath206 is an open neighborhood of @xmath117 , we deduce that @xmath188 is an open set . + * step 4 : * let @xmath207 . since @xmath208 is a collection of closed sets and @xmath209 , @xmath210 , we have that @xmath211 .
let us show that @xmath212 .
it is sufficient to show that @xmath213 , @xmath214 , there exists @xmath215 , such that @xmath216 , which implies @xmath217 .
in fact , given @xmath218 , and since @xmath161 is a closed set and @xmath0 is a zero - dimensional hausdorff set , we can write @xmath219 , a disjoint union of closed sets where @xmath220 and @xmath221 .
since @xmath222 and @xmath223 are compacta , we have that @xmath224 is also compact , so it is a closed set . furthermore , @xmath225 because @xmath222 and @xmath223 are disjoint . then , we can take @xmath226 a neighborhood of @xmath118 , such that @xmath227 .
we have that @xmath216 .
in fact , if the contrary @xmath228 holds , there would be a chain of order @xmath229 connecting @xmath181 to @xmath118 over @xmath161 @xmath230 + this would imply that there exists @xmath231 such that @xmath232 and @xmath233 , and so @xmath234 . +
* step 5 : * let @xmath235 be a neighborhood of @xmath118 .
since @xmath97 is continuous , there exists @xmath236 , open neighborhood of @xmath118 , such that @xmath237 .
we put @xmath238 ( the closure of @xmath154 ) and since @xmath212 , there exists @xmath215 , such that @xmath239 .
in fact , if there did not exist such @xmath229 , then for each @xmath240 the set @xmath241 would be closed and not empty .
then , @xmath242 would be not empty , what is a contradiction with @xmath212 .
@xmath243 is a relative open set of @xmath161 , so there exists @xmath244 an open set of @xmath0 , such that @xmath245 . since @xmath246
, we have that @xmath247 , that is , @xmath248 is an intersection of two open sets , so itself is an open set . let us show that @xmath249 . by construction , @xmath250 . if @xmath251 , then @xmath252 , with @xmath253 , and there exist two chains of order @xmath229 , @xmath254 and @xmath255 which connect respectively @xmath117 and @xmath181 to @xmath118 over @xmath243 .
thus , we can take the chain @xmath256 : @xmath257 using the medial property and @xmath258 we get that for all @xmath259 follows @xmath260 .
in fact , @xmath261 then , this is a chain of order @xmath229 connecting @xmath262 to @xmath118 over @xmath161 , and then @xmath263 .
we have proved @xmath264 . since @xmath97 , @xmath100 and @xmath101 are identical , it follows that @xmath265 + * step 6 : * let @xmath266 , we will show that @xmath267 is a partition of @xmath0 compatible with @xmath97 .
it follows straightforwardly from hypothesis ( h3 ) that @xmath268 is an open cover of @xmath0 and , by medial property , @xmath269 .
then we only need to prove that @xmath268 is a partition of @xmath0 . to do this ,
we introduce the following relation over @xmath0 : @xmath270 we claim that @xmath271 is an equivalence relation .
clearly @xmath271 is reflexive . to prove that @xmath271 is symmetric and transitive
, we use that @xmath272 ( @xmath273 ) in fact , @xmath274 is equivalent to saying that @xmath275 .
so , for all @xmath276 , let @xmath277 be such that @xmath278 . by multiplying this equation by the left by @xmath279 , and by using the medial property , we get @xmath280 . therefore , since @xmath281
, we deduce that @xmath282 and so @xmath283 .
thus , we can conclude that @xmath284 . by symmetric reasoning @xmath285
; hence @xmath286 .
( @xmath287 ) if @xmath286 , then for all @xmath288 , there exists @xmath289 , such that @xmath290 . again , multiplying this equation by the left by @xmath291 , and by using the medial property , we deduce that @xmath275 . +
it is not hard to see that @xmath292:\mathbf{b}\in x\right\}}$ ] , where @xmath293:={\left\{\mathbf{c}\in x:\mathbf{c}\sim\mathbf{b}\right\}}$ ] is the equivalence class of @xmath181 .
then , @xmath268 is a partition of @xmath0 into equivalence classes .
moreover , since @xmath0 is compact , @xmath268 is finite . using last theorem
, the following result has a proof similar to the one of proposition 2 at @xcite .
[ ii9 ] let @xmath75 be a topological quasigroup , such that all hypotheses of theorem [ ii6 ] hold , and @xmath6 is an expansive automorphism ( that is , ( h2 ) holds ) .
then @xmath75 is isomorphic to @xmath124 , where @xmath7 is a shift and @xmath8 is a 1-block operation .
the hypotheses ( h1)-(h4 ) are strongly restrictive .
in fact , for a finite quasigroup @xmath70 , dnes - keedwell ( @xcite , theorem 2.2.2 , p.70 ) showed that the medial property implies the quasi - group operation comes from a abelian group operation , that is , there exist an abelian group operation @xmath152 on @xmath0 , two automorphisms @xmath294 and @xmath295 on @xmath0 , and @xmath296 , such that @xmath297 for all @xmath298 .
for our case of zero - dimensional quasigroups , theorem [ ii6 ] , and propositions [ ii9 ] and [ ii10 ] , allow us to get an analogous result whenever there exists an expansive automorphism ( or endomorphism ) on @xmath0 .
let @xmath10 be a shift space .
suppose that @xmath50 is a quasigroup where @xmath8 is a 1-block operation .
in particular , since @xmath8 is @xmath37-block , @xmath26 is an automorphism over @xmath50 .
if @xmath50 is as above and in addition ( h3 ) holds , then : 1 . there exists an operation @xmath299 over @xmath300 , such that @xmath301 is a quasigroup which induces @xmath50 ; 2 .
@xmath302 , @xmath303 , we have @xmath304 . furthermore , if @xmath305 , then @xmath306 and @xmath307 ; 3 .
@xmath7 is a sft .
moreover @xmath124 is isomorphic to a markov shift with a 1-block operation . \i .
since @xmath8 is a 1-block operation , there exists @xmath308 a local rule of @xmath8 .
for @xmath309 , put @xmath310 .
let us show that @xmath301 is a quasigroup . since @xmath300 is finite , this property is equivalent to the fact that for all @xmath309 there exist @xmath311 which are the unique solutions of @xmath312 and @xmath313 .
in fact , if we take @xmath314 , such that @xmath315 , @xmath316 , there exists @xmath28 a unique solution of @xmath317 .
so @xmath318 , which means that @xmath319 is solution of @xmath312 .
since @xmath300 is finite , if we fix @xmath320 , for each @xmath321 there exists a distinct solution @xmath322 .
so , we can deduce that @xmath299 is right permutative . using the same argument
we also deduce the left permutativity .
then , @xmath301 is a quasigroup .
the proof of this fact uses similar arguments as in the proofs of proposition [ ii15 ] and claim [ ii16 ] after .
fix @xmath323 .
we have that @xmath324)\supseteq{\mathcal{f}}(\mathbf{u}[-1,0 ] ) \cdots\supseteq{\mathcal{f}}(\mathbf{u}[-n,0])\supseteq{\mathcal{f}}(\mathbf{u}[-n-1,0])\neq\emptyset$ ] .
since @xmath300 is finite , there exists @xmath325 , such that @xmath326)={\mathcal{f}}(\mathbf{u}[-n,0])$ ] , for all @xmath29 .
if @xmath327\in{\mathcal{w}}(\lambda , n+1)$ ] and @xmath328\in{\mathcal{w}}(\lambda , n+k+1)$ ] , then @xmath329)\subseteq { \mathcal{f}}([g_0,\ldots , g_n])$ ] and they are both cosets of @xmath330)$ ] ( by part ( ii ) ) . then , @xmath329)= { \mathcal{f}}([g_0,\ldots , g_n])$ ] and we deduce that @xmath7 is a @xmath31-step sft .
+ to conclude the proof , we define @xmath35 as the @xmath31-block presentation of @xmath7 , and consider the @xmath37-block quasi - group operation induced by @xmath7 .
we notice that there is no evidence about existence of quasigroups @xmath50 such that @xmath8 is a @xmath37-block operation but ( h3 ) does not hold .
the previous result implies that if ( h3 ) holds , then @xmath50 is a subquasigroup of @xmath331 .
the following proposition reproduce the result of proposition [ ii10 ] using the hypothesis that @xmath50 is a subquasigroup : [ ii11 ] if @xmath50 is a subquasigroup of @xmath331 , where @xmath8 is a @xmath37-block operation , then there exists an operation @xmath299 over @xmath300 , such that @xmath301 is a quasigroup which induces @xmath50 . from the fact of @xmath331
is quasigroup follows that for any constant sequences @xmath332 , @xmath333 and @xmath334 , there exist unique @xmath335 , @xmath336 and @xmath337 solutions of the equations @xmath338 hence , denoting the local rule of @xmath8 as @xmath299 , for any @xmath309 the equations @xmath339 and @xmath340 also have unique solutions , which implies @xmath341 is quasigroup .
in this section we consider the case of subquasigroups @xmath342 , where @xmath35 is a topological markov shift , and @xmath8 is a 1-block quasi - group operation . according to the previous section , the operation @xmath8 over @xmath35 is canonically induced by a quasi - group operation @xmath299 over @xmath34 , such that for any @xmath343 , @xmath344 [ ii14 ] let @xmath345 .
then @xmath346 , @xmath347 .
+ it follows from the bipermutativity of @xmath8 .
[ ii15 ] let @xmath348 be a 1-block quasi - group shift .
then , 1
. @xmath349 , @xmath304 and @xmath350 2 .
if @xmath351 , @xmath352 , then @xmath353 @xmath354 \i . since @xmath299 is bipermutative , @xmath355 , such that @xmath356
let @xmath351 , for all @xmath357 we have @xmath358
. then , @xmath359 and from claim [ ii14 ] , we deduce @xmath360 .
now , let @xmath361 be the permutation over @xmath34 , defined by @xmath362
. there exists @xmath363 , such that @xmath364 , so @xmath365 .
thus , for @xmath366\in{\mathcal{w}}(2)$ ] , @xmath367 , we have @xmath368*(\ldots([r , s]*[g , g']))}_{[r , s ] \ appears\ k-1\ times}=[f_r^{k-1}(g),f_s^{k-1}(g')]=[h , f_s^{k-1}(g')]\in{\mathcal{w}}(2)\ ] ] then , @xmath367 , we have @xmath369 and so @xmath370 . from claim [ ii14 ] , we have @xmath371 we conclude the aimed equality for the follower sets . using similar arguments
we deduce the similar equality for the predecessor sets .
it is straightforward from part i. and fact ( [ ii13 ] ) .
[ ii16 ] for any @xmath372 we have that @xmath373 and @xmath374 + @xmath375 where @xmath376 follows from part ( ii ) of proposition [ ii15 ] . for the predecessor sets , we use the same argument .
let @xmath35 and @xmath341 be as before and define * @xmath377 * @xmath378 notice that @xmath379 and @xmath380 are both covers of @xmath34 . on @xmath379 and @xmath380
we consider the operation canonically defined from the operation on @xmath34 which will be also denoted by @xmath299 .
the claim [ ii16 ] guarantees that @xmath299 is closed in @xmath379 and @xmath380 .
[ ii17 ] @xmath381 and @xmath382 are quasi - groups .
we will only show the result for @xmath381 , because the case @xmath382 is entirely analogous .
since @xmath379 is finite , to prove that @xmath381 is right and left cancellable , is equivalent to prove for all @xmath383 , there exist @xmath384 , that verify @xmath385 and @xmath386 .
we have that @xmath387 and @xmath388 , for some @xmath372 . by bipermutativity in @xmath34
, there exist @xmath389 such that @xmath390 and @xmath391 .
then , @xmath392 and @xmath393 are the solutions for above equations .
[ ii18 ] the markov shift @xmath35 has disjoint follower ( and predecessor ) sets , i.e. , @xmath394 if and only if @xmath395 .
suppose @xmath394 .
let @xmath396 and @xmath397 .
we have @xmath398 where @xmath399 is by proposition [ ii15](ii ) . since @xmath381 is bipermutative
we conclude that @xmath395 .
[ ii19 ] @xmath381 and @xmath382 are isomorphic .
in particular , for any @xmath372 we have @xmath400 .
+ let @xmath401 defined by @xmath402 , where @xmath403 is an arbitrary element .
let us show that @xmath404 is well defined , i.e. , it depends not on the choice of @xmath321 .
in fact , @xmath405 where @xmath406 is by corollary [ ii18 ] . also from above expressions it is direct that @xmath404 is one - to - one . on the other hand ,
is easy to see that @xmath404 is onto .
+ now , given @xmath383 , let @xmath407 and @xmath408 .
we have that @xmath409 and @xmath410 to conclude , notice that this isomorphism implies that @xmath411 . since @xmath379 and @xmath380 are both partitions of @xmath34 , each of them containing sets with the same cardinality ( proposition [ ii15 ] ) , we deduce that @xmath400 , @xmath412 .
suppose that @xmath8 is a group operation . in this case
, if we denote @xmath413 as the identity element , we have that @xmath414 and @xmath415 which implies that @xmath416 and @xmath417 are subgroups of @xmath341 .
moreover , since @xmath379 and @xmath380 are the sets of cosets of these subgroups , and @xmath381 and @xmath382 are also groups , we conclude that @xmath418 and @xmath419 are normal subgroups . given @xmath420 , let @xmath421 and @xmath422 be such that @xmath423 .
we define @xmath424 and denote @xmath425 .
+ notice that @xmath426 is well defined because for each @xmath420 there exists a unique @xmath421 and @xmath422 satisfying @xmath427 and @xmath428 .
moreover , we can write @xmath429 . consider the operation @xmath299 over @xmath430 as in @xmath379 and @xmath380 .
the claim [ ii20 ] give us that @xmath299 is closed in @xmath430 .
[ ii20 ] @xmath431 , @xmath432 .
moreover , @xmath433 and for all @xmath420 , @xmath434 .
suppose @xmath435 and @xmath436 .
thus , @xmath437 where @xmath376 comes from proposition [ ii15](ii ) .
since @xmath438 is a non - empty intersection of sets in @xmath379 and @xmath380 , we deduce that it lies in @xmath430 .
moreover , from definition of @xmath426 and @xmath439 it follows that @xmath440 .
then @xmath441 . on the other hand , from equation ( [ ii21 ] )
, we get @xmath442 .
these last equalities implies that any element of @xmath430 can be written as the product of any other element of @xmath430 by some element of @xmath34 , which implies @xmath434 for any @xmath420 .
[ ii22 ] @xmath443 is a quasigroup .
use the same argument as in proposition [ ii17 ] .
[ ii23 ] @xmath444 , @xmath445 .
the relation ( @xmath287 ) is obvious . for the other one @xmath446 ,
put @xmath435 and @xmath436 .
notice that @xmath447 implies @xmath448 and @xmath449 .
thus , by corollary [ ii18 ] , we have @xmath450 and @xmath451 , and the result follows . fix @xmath413 and let @xmath452 , where @xmath453 and @xmath454 .
given @xmath420 , define @xmath455 as the element in @xmath34 that verifies @xmath456 .
[ ii24 ] let @xmath457 be an arbitrary section of @xmath430 , i.e. , an arbitrary map such that @xmath458 , @xmath459 .
+ notice that @xmath458 , @xmath460 .
[ ii25 ] @xmath461 , @xmath462 . by definition of @xmath463
we have @xmath464 .
then , @xmath465 where @xmath376 is by claim [ ii20 ] , @xmath466 follows from definition of @xmath463 , and @xmath467 follows fromdefinition of @xmath468 .
[ ii26 ] the map @xmath469 given by @xmath470 is a bijection .
moreover , @xmath471 is given by @xmath472 , where @xmath473 is the unique element such that @xmath474 .
to check @xmath475 is one - to - one let @xmath372 , then @xmath476 since @xmath34 and @xmath477 are both finite sets with the same cardinality , by claim [ ii20 ] , @xmath475 is also onto .
moreover , given @xmath478 , let @xmath473 be the unique element such that @xmath479 .
we have that @xmath480 where @xmath376 is by claim [ ii20 ] , and @xmath466 is because @xmath481 .
hence , by proposition [ ii22 ] , we get @xmath482 .
then , @xmath483 .
[ ii27 ] define in @xmath477 the operation @xmath484 , given by @xmath485\ ] ] notice that alternatively we can write @xmath486 where @xmath487 are the unique elements which verify @xmath488 @xmath489 notice that on the first coordinate @xmath484 coincides with @xmath299 on @xmath430 .
[ ii28 ] we can identify @xmath341 to @xmath490 .
it follows straightforward from the definition of @xmath484 that @xmath475 is an isomorphism between @xmath341and @xmath490 .
[ ii29 ] define the markov shift @xmath491 on the alphabet @xmath430 , given by transitions : @xmath492 the transitions in definition [ ii29 ] can be defined by @xmath493 for any @xmath494 . in fact , if @xmath495 , then for all @xmath494 , @xmath496 where @xmath376 is due to the fact that for every @xmath497 , we have @xmath498 , hence @xmath499 because the follower sets partition @xmath34 , by corollary [ ii18 ] .
now , consider the map @xmath500 , which is also denoted as @xmath475 .
we shall check that @xmath501 is well defined , i.e. , for every @xmath502 we have @xmath503 . since @xmath504 , and for all @xmath505 we have @xmath506 by claim [ ii25 ]
, it suffices to verify @xmath507 .
this last property is fulfilled because , if @xmath508 and @xmath509 , then @xmath510 , where @xmath399 is by equation ( [ ii30 ] ) .
[ ii31 ] we can identify @xmath511 to @xmath512 , where @xmath97 is the 1-block operation induced by @xmath484 . let @xmath501 be the previous map .
@xmath501 is one - to - one because its local rule is ( proposition [ ii28 ] ) . on the other hand
if @xmath513 , from proposition [ ii28 ] we get @xmath514 .
so , to deduce that @xmath475 is onto and @xmath515 is 1-block , it is sufficient to show that @xmath502 .
now , by definition of @xmath491 , we have @xmath516 , @xmath517 , and so @xmath508 . therefore , @xmath511 is isomorphic to @xmath512 .
[ ii32 ] @xmath518 is a quasigroup , where @xmath8 is the operation induced by @xmath299 over @xmath430 .
@xmath518 is a quasigroup because it is a factor of @xmath519 , which is itself a quasigroup because by proposition [ ii31 ] says it is isomorphic to @xmath348 .
[ ii33 ] the shift @xmath491 verifies @xmath520 , for all @xmath453 and @xmath454 . .
since @xmath521\in{\mathcal{w}}({\sigma_\mathbf{a}},3)$ ] , we have @xmath522\in{\mathcal{w}}({\sigma_\mathbf{\hat{a}}},3)$ ] .
then , @xmath523 . if @xmath524 then @xmath525\in{\mathcal{w}}({\sigma_\mathbf{\hat{a}}},3)$ ] .
let @xmath526 , so that @xmath527 .
by definition of @xmath491 , we have @xmath528 , and so @xmath529 . on the other hand , also from definition of @xmath491 , it follows that @xmath530 . then , @xmath531 , which is equivalent to @xmath532 .
we deduce @xmath533 , and so we conclude @xmath534 .
[ ii34 ] define the shift @xmath535 with alphabet @xmath379 , and whose transitions are given by : @xmath536 let @xmath537 be the map defined by @xmath538 .
it is an onto homomorphism from @xmath341 to @xmath381 , by proposition [ ii10](ii ) .
let @xmath539 be the quasigroup , with the operation @xmath8 on @xmath535 , induced by the operation @xmath299 on @xmath379 .
we also denote by @xmath540 the map @xmath541 .
let us show that this map is well defined .
let @xmath502 , then @xmath542 .
thus , @xmath543 , i.e. , @xmath544 .
[ ii35 ] with the notations above : 1 .
@xmath545 is a homomorphism from @xmath348 onto @xmath539 ; 2 .
if @xmath546 , then the element @xmath547 appearing in the definition of @xmath535 is unique . in such case
, @xmath540 is an isomorphism between @xmath511 and @xmath548 .
@xmath545 is a homomorphism because its local rule is a homomorphism from @xmath341 to @xmath381 .
let us check that @xmath540 is onto .
let @xmath549 , and notice that from definition of @xmath535 , @xmath516 , @xmath550 , such that @xmath551 .
then , @xmath552 , verifying @xmath553 .
suppose @xmath546 .
let @xmath554 and @xmath555 be such that @xmath556 .
let @xmath420 be such that @xmath387 , and let @xmath557 . then
, @xmath558,[a , g_2,h]\in{\mathcal{w}}({\sigma_\mathbf{a}},3)\ ] ] this implies that @xmath559 . since @xmath560 for some @xmath561 , and
since @xmath562 is unitary , we deduce that @xmath563 . in this case , it is trivial to see that there exists @xmath564 .
[ ii36 ] if @xmath546 , then @xmath564 is a 2-block code , with memory 1 : @xmath565 where for all @xmath505 , @xmath566 is the unique element such that @xmath567 .
the following theorems are the analogous statements for quasi - groups as those of theorems stated in @xcite . from our previous results on quasi - groups
these theorems have similar proof than those in @xcite . [ ii37 ]
let @xmath348 be a quasigroup , where @xmath35 is a markov shift and @xmath8 is a 1-block operation .
then , 1 . @xmath511 is isomorphic by a 1-block code to @xmath76 , where @xmath77 is a finite quasigroup with 1-block operation ; @xmath78 is a full n shift ; and @xmath79 is a @xmath80-block quasi - group operation , with memory k-1 and anticipation @xmath568 .
@xmath569 if and only if @xmath570 ( i.e. , the full shift is trivial ) .
3 . @xmath35 is irreducible and has a constant sequence if and only if @xmath84 ( i.e. , @xmath77 is unitary ) .
[ ii38 ] let @xmath348 be an irreducible markov shift , such that @xmath8 is a 1-block quasi - group operation .
let @xmath571 , where @xmath572 is the prime decomposition of @xmath325 .
then @xmath511 is isomorphic to @xmath573 , where @xmath574 is the full @xmath325 shift and @xmath79 is at most @xmath575-block , with anticipation 0 .
let @xmath331 be a quasigroup , where @xmath8 is induced by a @xmath37-block operation @xmath299 on @xmath1 .
let @xmath576 be a topological markov chain .
define @xmath537 by @xmath538 , as before .
then @xmath35 is closed under @xmath8 if and only if @xmath537 is an onto homomorphism .
furthermore , @xmath348 is irreducible ( transitive ) if and only if there exists @xmath420 such that @xmath577 for some @xmath578 , where @xmath579 is defined inductively by @xmath580 .
[ ii39 ] we can define the shift @xmath581 , in the same way than definition [ ii34 ] : @xmath582 if we consider @xmath581 instead of @xmath535 in proposition [ ii35 ] , we obtain analogous results , but @xmath583 will be a 2-block code with anticipation 1 .
moreover , in the theorems [ ii37 ] and [ ii38 ] , @xmath79 will be a @xmath80-block operation with memory 0 and anticipation @xmath81 .
let @xmath348 be a markov shift with a 1-block quasi - group operation .
as before , denote by @xmath299 the quasi - group operation on @xmath34 induced by @xmath8 .
we define the four elementary isomorphisms as in @xcite : - state splitting by successors : given @xmath420 , let @xmath584 be a subset such that @xmath585 is a partition of @xmath34 compatible with @xmath299 . define @xmath586 where @xmath587 denotes the coset of @xmath588 containing @xmath589 .
consider on @xmath590 the operation coinciding with @xmath299 in each coordinate , and let @xmath591 be the shift defined by the following transitions : @xmath592 which is considered with the operation canonically induced by @xmath590 .
the state splitting is the 2-block code , @xmath593 , defined by : @xmath594\in{\mathcal{w}}({\sigma_\mathbf{a}},2)\qquad\mapsto\qquad ( g,{\mathcal{h}}_h)\in{l_\mathbf{\tilde{a}}}.\ ] ] notice that @xmath595 is a 1-block code given by @xmath596 .
the state splitting is an isomorphism between @xmath35 and @xmath591 . - state splitting by predecessors : it is defined as in the previous case , but using @xmath597 instead of @xmath598
. - amalgamation by common predecessors and disjoint successors : given @xmath420 , let @xmath584 be a subset , such that @xmath599 is a partition of @xmath34 compatible with the operation @xmath299 .
moreover , suppose that @xmath600 , we have @xmath601 is either empty or unitary .
define @xmath602 , where it is considered the operation induced by @xmath299 .
let @xmath591 be the shift given by transitions : @xmath603 where is defined the operation induced by @xmath590 .
the amalgamation is the 1-block code , @xmath593 , given by @xmath604 notice that @xmath595 is a 2-block code given by @xmath605\in{\mathcal{w}}({\sigma_\mathbf{\tilde{a}}},2)\mapsto h\in{l_\mathbf{a}}$ ] , where @xmath589 is the unique element belonging to @xmath606 , such that @xmath607 .
it is straightforward to see that the amalgamation is an isomorphism between @xmath35 and @xmath591
. - amalgamation by common successors and disjoint predecessors : it is defined in the same way than the previous case , but changing the roles of the predecessor and the follower sets .
[ ii41 ] two quasi - group sfts , each of them with 1-block quasi - group operation , are isomorphic if and only if it is possible to go from one to other by a finite sequence of elementary isomorphisms .
let @xmath608 and @xmath609 be both quasi - group shifts with 1-block operations .
let @xmath610 be an isomorphism between them . without lost of generality we can consider @xmath35 a markov shift and @xmath475 a 1-block code ( in fact
, we can take the @xmath325-block presentation of @xmath35 , with @xmath325 sufficiently large ) .
furthermore , @xmath341 and @xmath611 are both quasi - groups which induce , respectively , the operations @xmath8 and @xmath612 .
we have that @xmath613 is an onto homomorphism between @xmath341 and @xmath611 ( notice that the local rule of the code is also denoted by @xmath475 ) .
define @xmath614 , which is a partition of @xmath34 compatible with @xmath299 .
this property also holds when we consider for @xmath240 , @xmath615 , that is @xmath616\}):\quad [ \bar{a}_1,\ldots,\bar{a}_n]\in{\mathcal{w}}({\sigma_\mathbf{\bar{a}}},n)\right\}},\ ] ] which is a partition of @xmath617 compatible with @xmath299 .
since @xmath515 is a @xmath325-block code , there exists @xmath618 , @xmath619 , such that given @xmath620\in{\mathcal{w}}({\sigma_\mathbf{\bar{a}}},n)$ ] , @xmath621,[a'_1,\ldots , a'_n]\in\phi^{-1}(\{[\bar{a}_1 , \ldots,\bar{a}_n]\})$ ] we have @xmath622 .
fix @xmath623\in{\mathcal{w}}({\sigma_\mathbf{\bar{a}}},n)$ ] , @xmath624\in\phi^{-1}(\{[\bar{x}_1 , \ldots,\bar{x}_n]\})$ ] and put @xmath625 .
it follows that @xmath626 and @xmath588 is a partition of @xmath34 , compatible with @xmath299 .
furthermore , for every @xmath561 the set @xmath627 has at most one element .
in fact , if there was more than one element in @xmath627 , then we could find two distinct sequences in @xmath35 with the same image by @xmath475 , what is a contradiction about injectivity of this map .
denote by @xmath593 the amalgamation by common predecessors and disjoint successors , where @xmath628 .
we recall @xmath629 is a @xmath37-block code and @xmath595 is a @xmath630-block code .
we define @xmath631 the @xmath37-block map which has local rule ( which we will also denote by @xmath632 ) , given by @xmath633 for any @xmath634 .
now , we have that given @xmath620\in{\mathcal{w}}({\sigma_\mathbf{\bar{a}}},n)$ ] , for all @xmath635,[\tilde{a}'_1,\ldots,\tilde{a}'_n ] \in\tilde{\phi}^{-1}(\{[\bar{a}_1 , \ldots,\bar{a}_n]\})$ ] follows that @xmath636 and @xmath637 .
we repeat the above process until we get a @xmath37-block isomorphism , @xmath631 , such that for any @xmath620\in{\mathcal{w}}({\sigma_\mathbf{\bar{a}}},n)$ ] , for all @xmath635,[\tilde{a}'_1,\ldots,\tilde{a}'_n ] \in\tilde{\phi}^{-1}(\{[\bar{a}_1 , \ldots,\bar{a}_n]\})$ ] follows @xmath638 , @xmath639 . to conclude , we come back and applying the amalgamation by common successors and disjoint predecessors from the entry @xmath640 until the first entry .
we obtain that @xmath631 is a 1-block isomorphism , and for all @xmath641\in{\mathcal{w}}({\sigma_\mathbf{\bar{a}}},n)$ ] we have that @xmath642\})$ ] contains a unique @xmath325-block of @xmath591 .
this implies that @xmath591 and @xmath535 are identical .
i want to thank a. maass , s. martnez , and m. pivato for their useful discussions and advices on the subject .
i also would like to thank the referee and editor for their comments to strengthen the presentation of the paper . | in this work we characterize those shift spaces which can support a 1-block quasi - group operation and show the analogous of kitchens result : any such shift is conjugated to a product of a full shift with a finite shift .
moreover , we prove that every expansive automorphism on a compact zero - dimensional quasi - group that verifies the medial property , commutativity and has period 2 , is isomorphic to the shift map on a product of a finite quasi - group with a full shift .
marcelo sobottka ' '' '' _ this is a pre - copy - editing , author - produced pdf of an article accepted for publication in discrete and continuous dynamical systems - series a ( dcds - a ) , following peer review .
the definitive publisher - authenticated version marcelo sobottka , topological quasi - group shifts . disc . and cont .
dynamic . systems ( 2007 ) , 17 , 1 , 77 - 93 , is available online at : http://www.aimsciences.org / journals / displayarticles.jsp?paperid=2044 . _ ' '' '' |
many researchers investigate optimal control problems with uncertainties . in @xcite ,
gabasov et al .
consider optimal preposterous observation and optimal control problems for dynamic systems under uncertainty with use of a priori and current information about the controlled object behavior and uncertainty . in @xcite , gabasov et al .
investigate for an optimal control problem under uncertainty the positional solutions , which are based on the results of inexact measurements of input and output signals of controlled object . in @xcite ,
gabasov et al .
study a problem of optimal control of a linear dynamical system under set - membership uncertainty .
fuzzy time - optimal control problem is investigated in different forms in @xcite-@xcite . in @xcite
, plotnikov proves necessary maximin and maximax conditions for a control problem , when behavior of the object is described by a controllable differential inclusion with multivalued performance criterion . in @xcite ,
sakawa et al .
propose a fuzzy satisficing method for multiobjective linear optimal control problems . to solve these problems , they discretize the time and replace the system of differential equations by system of difference equations . in @xcite , molchanyuk and plotnikov study the problem of high - speed operation for linear control systems with fuzzy right - hand sides .
for this problem , they introduce the notion of optimal solution and establish necessary and sufficient conditions of optimality in the form of the maximum principle . in this paper
, we consider a time - optimal control problem with crisp dynamics and with fuzzy start and target states .
we interpret the optimal time as a fuzzy variable and propose a numerical method to calculate it .
the paper consists of 5 sections . in section 2 ,
we describe the classical time - optimal control problem . in section 3
, we define the fuzzy time - optimal control problem and propose a method for calculation of fuzzy optimal time . in section 4 ,
we show the proposed approach by an example .
finally , we give concluding remarks in section 5 .
let the behavior of a controlled object is definite ( crisp ) and described by the following linear system of differential equations : @xmath0 here @xmath1 is @xmath2-dimensional vector - function that describes the phase state of the object , @xmath3 is an @xmath4 matrix , @xmath5 is @xmath2-dimensional control vector - function .
let @xmath6 be a nonempty compact set .
if measurable function @xmath5 , defined on the interval @xmath7 $ ] , satisfies the condition @xmath8 for each @xmath9 , then @xmath5 is called as admissible control .
it is known that for any admissible function @xmath5 and for any initial state @xmath10 the initial value problem @xmath11 @xmath12 has a unique solution @xcite .
this solution @xmath1 describes how the phase state changes under the influence of admissible control @xmath5 .
assume that the start time @xmath13 and the start state @xmath10 are given .
if we want to transfer the object to a given state @xmath14 in the shortest time by choosing an appropriate admissible control @xmath5 , we have the following classical time - optimal control problem of 1st type : @xmath15 subject to @xmath16 @xmath17 @xmath18 note , that the finish time @xmath19 is not known beforehand and is determined as a result of solving the problem .
summarizing , 1st type classical time - optimal problem ( [ eqa1])-([eqa4 ] ) is a problem of finding an admissible control @xmath5 , which transfers the system from the initial phase state @xmath10 to the final phase state @xmath14 in the shortest time .
now , let nonempty compact sets @xmath20 and @xmath21 from @xmath22 , an interval @xmath7 , $ ] and an admissible function @xmath5 on this interval are given .
if the system ( [ motion ] ) has a solution @xmath23 such that @xmath24 and @xmath25 , then it is said that the control function @xmath5 transfers the object from the initial phase set @xmath20 to the final phase set @xmath21 on the interval @xmath26 $ ] .
if we want to transfer the object from the set @xmath20 to the set @xmath21 in the shortest time , we have the following classical time - optimal control problem of 2nd type : @xmath27 subject to @xmath28 @xmath29 @xmath30 where @xmath20 and @xmath21 are given start and target sets .
the solution @xmath5 of the problem ( [ eqb1])-([eqb4 ] ) is called optimal control .
the solution @xmath1 of the system ( [ eqb2])-([eqb4 ] ) , corresponding to the optimal control @xmath5 , is called optimal trajectory .
if @xmath31 is an optimal control and @xmath23 is a corresponding optimal trajectory , then @xmath32 is called to be an optimal pair .
we note that the classical problem of 2nd type can also be reformulated as follows : @xmath33 @xmath34 @xmath35 @xmath36 2nd type classical time - optimal problem ( [ eqb1])-([eqb4 ] ) ( or ( eqd1)-([eqd4 ] ) ) is well studied @xcite . below we
givenecessary conditions of optimality for this problem @xcite .
( maximum principle ) .
let @xmath5 be an admissible control defined on an interval @xmath26 $ ] and let @xmath1 be a solution of the system ( eqb2)-([eqb4 ] ) .
we say that the pair @xmath32 satisfies maximum principle on the interval @xmath26 $ ] if the conjugate system @xmath37 has such a nontrivial solution @xmath38 that the following conditions hold : \1 ) maximum condition : @xmath39 for almost any @xmath40 $ ] ; \2 ) transversality condition on @xmath20 : @xmath41 ; \3 ) transversality condition on @xmath21 : @xmath42 . here
@xmath43 denotes the inner product of vectors @xmath5 and @xmath44 from @xmath22 and @xmath45 denotes the support function of the compact set @xmath46 from @xmath47 .
@xcite ( necessary conditions of optimality for the time - optimal control problem ) .
let @xmath20 and @xmath21 be nonempty convex compact sets . also let the function @xmath5 defined on @xmath26 $ ] be an optimal control for the problem ( [ eqb1])-([eqb4 ] ) and @xmath1 be a corresponding optimal trajectory .
then the pair @xmath32 satisfies maximum principle on the interval @xmath26 $ ] .
if the start and target values in classical problem of 1st or 2nd type are fuzzy , we obtain the following fuzzy time - optimal control problem : @xmath48 subject to @xmath49 @xmath50 @xmath51 where @xmath52 and @xmath53 are given fuzzy initial and final vectors ( or sets ) . depending on different definitions for derivative of fuzzy function or different definitions for solution of system of differential equations ,
the problem ( [ eqf1])-([eqf4 ] ) can be interpreted by different ways .
we will interpret the problem ( [ eqf1])-([eqf4 ] ) as a set of 1st type classical problems ( [ eqa1])-([eqa4 ] ) .
each problem is obtained by taking the initial value @xmath10 from @xmath54 and the final value @xmath14 from @xmath55 .
we denote by @xmath56 , @xmath57 and @xmath58 the solutions of the problem ( [ eqa1])-([eqa4 ] ) .
let @xmath59 ( where @xmath60 denotes the membership of @xmath10 in @xmath54 ) .
we call @xmath61 to be a solution of the problem ( [ eqf1])-([eqf4 ] ) with possibility @xmath62
. set of all @xmath56 , defined above , determines a fuzzy number @xmath63 .
we will investigate how to calculate @xmath64 .
functions @xmath65 and @xmath66 which indicate the left and right boundaries of @xmath62-cuts , determine the number @xmath64 fully .
thus , the problem of calculation of fuzzy optimal time is reduced to calculation of the functions @xmath67 and @xmath68 . as it is known , the initial and final values of the optimal solution @xmath23 of the problem ( [ eqb1])-([eqb4 ] ) are achieved on boundaries of the sets @xmath20 and @xmath21 @xcite .
thus , value of @xmath67 can be obtained by solving the problem ( [ eqb1])-([eqb4 ] ) with taking @xmath69 and @xmath70 ( @xmath71 and @xmath72 denote @xmath62-cuts of @xmath54 and @xmath73 , respectively ) , namely the problem : @xmath74 @xmath75 @xmath76 @xmath77 the problem ( [ eqc1])-([eqc4 ] ) is a classical problem of 2nd type .
taking into account ( [ eqd1 ] ) , it can be seen that @xmath78 .
note that , the value @xmath65 means the shortest time between two points , one of them is from the set @xmath79 and another is from @xmath72 , in the best case .
similarly , @xmath80 means the shortest time in the worst case : @xmath81 calculation of @xmath68 can be performed similarly to calculation of @xmath65 .
in this section , we apply the proposed approach to a fuzzy time - optimal control problem .
the problem is a fuzzified version of the crisp problem of damping of mathematical pendulum , presented in @xcite .
solve the fuzzy time - optimal control problem ( note that below @xmath82 ) : @xmath83 @xmath84 @xmath85 @xmath86 @xmath87 @xmath88 here @xmath89 and @xmath90 are triangular fuzzy numbers . from above @xmath91 and @xmath92 .
system s matrix is @xmath93 $ ] .
since , @xmath94 ( here @xmath95 denotes the conjugate matrix of @xmath96 ) , the conjugate system is : @xmath97 @xmath98 support function of @xmath99 is @xmath100 .
then , from the maximum condition @xmath101 we have @xmath102 .
consequently , @xmath103 if @xmath104 ; @xmath105 if @xmath106 ; @xmath107 if @xmath108 .
let us find the solution of the conjugate system corresponding to an initial condition @xmath109 , where @xmath110 is the unit circle .
initial point can be represented in the form of @xmath111 with @xmath112 .
then the solution of the conjugate system is @xmath113 . the function @xmath114 changes its sign for first time at @xmath115 @xmath116 , if @xmath117 and @xmath118 , if @xmath119 and then after each @xmath120 time period .
depending on @xmath62 , the sign of the function @xmath121 in the interval @xmath122 $ ] is either positive or negative .
thus , according to the maximum condition , the initial value of the optimal control @xmath123 is either @xmath124 or @xmath125 .
after @xmath126 time units , it switches from @xmath124 to @xmath125 or vice versa .
then , it repeatedly changes its sign after each @xmath120 time period . below we interpret the behavior of the object as a motion of the object in the phase plane @xmath127 .
solutions of dynamic system corresponding to @xmath128 are in the form @xmath129 . in the phase plane
@xmath130 these solutions constitute concentric circles with center at @xmath131 ( fig .
the motion on these circles is clockwise with constant speed and whole turn takes @xmath132 time units . and
@xmath133.,width=257,height=168 ] similarly , solutions of dynamic system corresponding to @xmath134 are in the form @xmath135 . in @xmath130
these solutions constitute circles with center at @xmath136 ( fig .
the motion on these circles is clockwise with constant speed and whole turn takes @xmath137 time units .
note that angular speed is @xmath138 for both motions mentioned above .
so , the angle formed by the object during its motion and passed time are equal in value .
let us emphasize two facts which will be used in arguments below .
1 ) in circular motion with @xmath138 after @xmath120 time period the object will be in the position which is central symmetric point of the previous position .
2 ) the symmetric point of @xmath139 is @xmath140 with respect to center point @xmath141 .
if center is @xmath133 , then the symmetry of point @xmath142 is @xmath143 .
now we investigate how is a motion of the object corresponding to an optimal control in the phase plane for a start point @xmath46 and a target point @xmath144 .
let us consider the case when the object starts with control @xmath145 ( the case with start control @xmath146 can be investigated similarly ) .
let @xmath147 denote the number of control switches .
we consider the cases @xmath148 ( motion without switch ) and @xmath149 separately . in the case
@xmath148 , running from the start position @xmath46 and moving along a circle with center @xmath141 the object reaches the target position @xmath144 .
this case occurs , only if @xmath150 ( here @xmath151 denotes the length of the segment @xmath152 ) .
the motion time is @xmath153 ( here @xmath154 denotes the value of the angle @xmath155 ) .
now let @xmath149 .
we differ the cases when @xmath147 is odd and when @xmath147 is even .
let us consider the case that @xmath147 is odd number and take @xmath156 for clarity .
the object runs from the point @xmath46 along a circle with center @xmath141 and after @xmath157 time period arrives a point @xmath158 ( fig .
the points @xmath46 and @xmath159 are on the same circle .
consequently : @xmath160 at the point @xmath159 the control switches for the first time and becomes @xmath146 . under this control ,
the object moves along a circle with center @xmath133 .
after @xmath161 time units it arrives a point @xmath162 . here the control switches for the second time and under new control @xmath145 ( moving on circle with center @xmath141 ) after @xmath120 time the object reaches a point @xmath163 . at the point @xmath164 the control switches for last time and becomes @xmath146 .
the object continues its motion on a circle with center @xmath165 up to the target point @xmath144 .
for the aforementioned motion , the points @xmath166 and @xmath144 must be on the same circle with center @xmath133 , i.e. , @xmath167 it can be seen from table 1 that for an odd @xmath147 ( including @xmath168 ) the last point of control switch is @xmath169 .point of @xmath147-th control switch for optimal motion [ cols="^,^,^,^",options="header " , ] let @xmath170 and @xmath171 . to calculate unknown coordinates @xmath1 and @xmath172 we use equations ( [ c1 ] ) and ( [ c2 ] ) .
using ( [ xk ] ) , these equations can be rewritten in coordinates as follows : @xmath173 subtracting ( [ c2n ] ) from ( [ c1n ] ) we have : @xmath174
. then we can determine @xmath1 and @xmath172 as follows : @xmath175 if @xmath1 and @xmath172 have been determined we can calculate the passed time : @xmath176 let us find an evaluation for @xmath147 . from ( [ x ] ) and ( [ y ] )
we have @xmath177@xmath178hence , we obtain the following evaluation @xmath179where @xmath180 and @xmath181 denote ceiling and floor of @xmath1 , respectively . by taking @xmath182 , we have a feasible motion .
hence , using formula ( [ t1 ] ) , we get : @xmath183then , we have @xmath184 .
consequently , we obtain the following upper evaluation for @xmath147 , by using ( [ k_eval ] ) : @xmath185 the case when @xmath149 and @xmath147 is even can be investigated by similar way . in this case
the last point of the control switch is ( table 1 ) : @xmath186 the last control is @xmath145 and , consequently , the object finishes its motion on a circle with center @xmath141 .
hence , @xmath187 . except this value , the formulas for @xmath1 and @xmath172 become the same as ( [ x ] ) and ( [ y ] ) .
the motion time is : @xmath188 above we have investigated the case when the start control @xmath5 equals to @xmath125 . in the case where @xmath5 is
@xmath124 we have the following final formulas : @xmath189 the above formulas , given for different situations , were obtained on the base of the necessary conditions for optimality .
therefore , every solution constructed on these formulas may not be optimal .
however , the optimal solution is among all solutions , constructed for different start controls and for different values of @xmath147 .
based on the above arguments and formulas a computer program is implemented to calculate the optimal control for a given pair of start point @xmath46 and target point @xmath144 .
firstly , by taking start control @xmath145 , after taking @xmath146 and in both cases by changing the value of @xmath147 from @xmath190 to @xmath191 a solution is constructed ( if there is any ) .
the solution with the shortest time is the optimal solution , transferring the object from @xmath46 to @xmath192 .
now , let us describe how we calculate the fuzzy optimal time @xmath63 numerically . to calculate the value @xmath65 ,
we place equally spaced nodes on the boundaries of the regions @xmath193 and @xmath72 .
the shortest time among all possible start - destination node pairs @xmath194 gives the approximate value of @xmath65 .
to calculate the function @xmath68 we discretize the problem ( [ t1_up ] ) and solve it numerically .
the membership function of fuzzy optimal time @xmath19 , obtained from calculations , is depicted in fig .
the value @xmath195 with possibility @xmath124 corresponds to the solution of the crisp problem ( @xmath196 and @xmath197 ) .
the least value @xmath198 with possibility @xmath199 occurs when @xmath200 and @xmath201 .
the largest value @xmath202 with possibility @xmath199 corresponds to the pair @xmath203 and @xmath204 . .
, width=486,height=247 ]
in this paper , we investigate the problem of time - optimal control with fuzzy initial and final states .
we interpret the problem as a set of crisp problems .
we perform the calculation of fuzzy optimal time by solving crisp optimal control problems of two types .
we exhibit the proposed approach on a numerical example .
r. gabasov , f.m .
kirillova , e.i .
poyasok , optimal control of linear systems under uncertainty , proceedings of the steklov institute of mathematics , 268 ( supplement 1 ) ( 2010 ) 95 - 111 .
doi : 10.1134/s0081543810050081 | in this paper , we consider a time - optimal control problem with uncertainties .
dynamics of controlled object is expressed by crisp linear system of differential equations with fuzzy initial and final states .
we introduce a notion of fuzzy optimal time and reduce its calculation to two crisp optimal control problems .
we examine the proposed approach on an example .
* keywords : * optimal - time control , fuzzy set , maximum principle , mathematical pendulum . |
in this work we develop spectral theory for the @xmath0-boson ( stochastic ) particle system - boson totally asymmetric diffusion model while in @xcite it was also referred to as @xmath0-tazrp .
the term `` stochastic '' is included here to differentiate this with the non - stochastic quantum particle system considered in earlier work @xcite under the name @xmath0-bosons . that earlier studied system is a special limit of the more general system presently considered ( see section [ vandiejen ] ) . despite this , in what follows we will generally suppress the term `` stochastic '' , though still always referring to the stochastic particle system ] .
this is an interacting particle system whose generator is a stochastic representation of the generalization of the @xmath0-boson hamiltonian introduced by sasamoto - wadati in 1998 @xcite ( see section [ afortiori ] for more details ) .
the system ( in fact , a totally asymmetric zero range process ) consists of @xmath1 particles on @xmath2 with locations labeled by @xmath3 . in continuous time , each cluster of particles with the same location transfers one particle to the left by one at rate @xmath4 , where @xmath5 is the size of the cluster and @xmath0 is a parameter fixed between 0 and 1 . in order to preserve the ordering of @xmath6 ,
the highest index particle in a cluster is always the one which moves left .
this particle system can be understood as being a discrete space , @xmath0-deformation of the continuum delta bose gas on @xmath7 with attractive coupling constant ( see section [ deltabosesec ] ) .
this delta bose gas has a rich history , going back to the foundational work of lieb - liniger in 1963 @xcite , and it has recently played an important role in the physics literature @xcite surrounding the kardar - parisi - zhang ( kpz ) equation and universality class ( which includes random growth models , interacting particle systems and directed polymers see the review @xcite ) .
more exactly the moments ( for a fixed time @xmath8 but possibly different spatial locations @xmath9 ) of the solution to the stochastic heat equation ( whose logarithm is the kpz equation that models a randomly growing interface ) satisfy the delta bose gas with initial data corresponding to the initial data of the stochastic heat equation . for more details
see section [ deltabosesec ] .
the moment problem for the solution to the stochastic heat equation is not well - posed since its moments , though all finite , grow too fast to characterize the distribution of the solution . despite this mathematical limitation , there has been a significant amount of non - rigorous work using these moment formulas to extract distributional information about the solutions to kpz equation this sometimes goes by the name of the polymer replica method . in the instances for which rigorous results were available via other means @xcite it has been checked that these computations have yielded the correct answer .
there are now many non - rigorous kpz distribution computations such as those involving different types of initial data ( flat / half - flat @xcite , stationary @xcite , or more general @xcite ) or different times @xcite which are based on this technique and which do not yet have rigorous counterparts .
such computations involve some level of guessing ( as to how to sum certain divergent series ) which varies problem to problem , hence it is hard to be confident ( let alone prove ) whether the outcome of each additional computation will yield the correct answer .
one way to put this line of work on a firm , even rigorous footing is to find discrete regularizations of the kpz equation which are integrable and well - posed in the sense that formulas for moments characterize the distribution .
one such system is the @xmath0-deformed totally asymmetric simple exclusion process ( @xmath0-tasep ) which was introduced by borodin - corwin @xcite in 2011 via the framework of macdonald processes .
it was later observed by borodin - corwin - sasamoto @xcite that moments of @xmath0-tasep solve the @xmath0-boson particle system with initial data corresponding to the initial condition of @xmath0-tasep .
drawing on @xcite , in @xcite this system was explicitly solved for one family of initial data ( corresponding to @xmath0-tasep started from half - stationary initial condition ) . until now , it was not clear how to analyze this system for general initial data , as would be necessary to rigorously approach the many problems studied via the polymer replica method in the past few years .
the reason why the delta bose gas can be solved for general initial data is that it is known how to diagonalize the hamiltonian .
though the eigenfunctions for this hamiltonian have been known via coordinate bethe ansatz @xcite since the work of lieb - liniger @xcite , alone they do not suffice to solve the time evolution equation .
ultimately this requires a plancherel formula which shows how to decompose general initial data onto a subset of the algebraic eigenfunctions ( and hence shows completeness of the bethe ansatz as well ) .
various forms of such a result have been proved in @xcite for the delta bose gas . in this present work
we prove a plancherel formula for the @xmath0-boson particle system .
this allows us to explicitly compute the moments of @xmath0-tasep for general initial data which will hopefully serve as the basis for further rigorous asymptotic work confirming the results of the previously made non - rigorous calculations about the kpz equation .
our proof of the plancherel formula is fairly simple , relying in part on the _ contour shift argument _ used in heckman - opdam s proof of the delta bose gas plancherel formula @xcite ( see section [ plansec ] for more history on this argument ) . in that previous work ,
the hermiticity of the hamiltonian played an important rule . in our work
we consider a non - hermitian hamiltonian , however we find that a symmetry ( known sometimes as pt - invariance or joint space - reflection and time - reversal symmetry ) between our hamiltonian and its adjoint is a suitable replacement for hermiticity , see remark [ commrem ] .
a large part of the paper is devoted to modifications , implications and degenerations of this core result .
for example , the plancherel formula immediately implies a ( spatial ) orthogonality of the left and right eigenfunctions and suggests a second ( spectral ) orthogonality which we prove separately by ultimately appealing to the cauchy - littlewood formula for hall - littlewood ( multivariate symmetric ) polynomials ( though we should emphasize that our eigenfunctions are not hall - littlewood polynomials and only degenerate to them in a certain limit ) . with minor modifications
all results ( though not necessarily the probabilistic interpretations ) apply for more general parameter @xmath10 .
such @xmath0 ( especially on the complex unit circle ) arise naturally from a quantum physics perspective ( see section [ complexq ] ) .
it is possible to take a limit of the @xmath0-boson particle system so as to recover the continuum delta bose gas and its plancherel formula .
there are also two discrete space limits we study
one leads to a delta bose gas and plancherel formula previously studied by van diejen @xcite and another leads to the delta bose gas and plancherel formula related to the semi - discrete stochastic heat equation ( or equivalently the oconnell - yor semi - discrete directed polymer partition function ) .
there exist other integrable discrete regularizations of the kpz equation ( and delta bose gas ) such as the asymmetric simple exclusion process ( asep ) or @xmath0-pushasep @xcite . in a companion paper @xcite we prove ( in a similar manner as here ) a plancherel formula for asep and the heisenberg xxz quantum spin chain on @xmath2 ( thus recovering the xxz results of @xcite in a different manner ) .
let @xmath11 .
the backward generator for the @xmath0-boson particle system is written @xmath12 and defined via its action on functions @xmath13 as @xmath14 here @xmath15 denotes the number of clusters of @xmath16 and @xmath17 denote the sizes of these clusters ( so that @xmath18 ) and @xmath19 . the forward generator of the @xmath0-boson particle system is written @xmath20 and given by the matrix transpose of @xmath12 ( see definition [ bwdgendef ] for its explicit action ) .
we will be concerned with eigenfunctions of @xmath12 and @xmath20 .
define the function @xmath21 by @xmath22 where @xmath17 are as above , and @xmath23 is the @xmath0-factorial .
this is an invariant measure for the @xmath0-boson particle system and the backward and forward generators are related by @xmath24 where @xmath25 and @xmath26 is the multiplication operator @xmath27 .
this relationship is sometimes called pt - invariance , see remark [ commrem ] . for all
@xmath28 , set @xmath29 our first result is that these are the right eigenfunctions of @xmath12 and @xmath20 ( respectively ) . for all @xmath28
, we have @xmath30 the proof of this result is based on the fact that the backward and ( @xmath26-conjugated ) forward generators can be rewritten as free generators subject to @xmath31 two - body boundary conditions .
the coordinate bethe ansatz then readily implies the result . since we are working on @xmath2 ( as opposed to a finite or periodic interval )
there are no bethe equations to be solved .
since @xmath32 is a right eigenfunction for the forward generator @xmath20 , and @xmath20 is the matrix transpose of @xmath12 , it follows that @xmath33 is also a left eigenfunction for @xmath20 .
thus , we use the alternative notation @xmath34 we now proceed to our main result a plancherel type isomorphism theorem .
let @xmath35 be the space of functions @xmath13 of compact support , and let @xmath36 be the space of symmetric laurent polynomials @xmath37 in the variables @xmath38 .
let @xmath39 be positively oriented , closed contours chosen so that they all contain @xmath40 , so that the @xmath41 contour contains the image of @xmath0 times the @xmath42 contour for all @xmath43 , and so that @xmath44 is a small enough circle around 1 that does not contain @xmath0 ( see figure [ circontours ] in section [ plansec ] below ) . define the ( symmetric ) bilinear pairing @xmath45 on functions @xmath46 via @xmath47 and the ( symmetric ) bilinear pairing @xmath48 on functions @xmath49 via @xmath50 here @xmath51 is a partition of size @xmath52 ( i.e. , @xmath53 ) and @xmath54_{i , j=1}^{\ell(\lambda ) } \prod_{j=1}^{\ell(\lambda ) } w_j^{\lambda_j } q^{\frac{\lambda_j^2}{2 } } \frac{dw_j}{2\pi { \mathbf i}};\ ] ] where @xmath55 , the @xmath0-pochhammer symbol is @xmath56 and we use the notation @xmath57 [ smallremarkintro ] the pairing @xmath58 on @xmath36 has a simpler alternative definition : @xmath59 where @xmath60 can be chosen to be a circle containing both 1 and 0 and @xmath61 is the partition with @xmath52 ones .
however , it turns out that the more involved definition above is more useful for our present purposes ( in particular when taking various limits ) .
the _ @xmath0-boson transform _
@xmath62 takes functions @xmath63 into functions @xmath64 via @xmath65 the ( candidate ) _
@xmath0-boson inverse transform _
@xmath66 takes functions @xmath67 into functions @xmath68 via @xmath69 by shrinking the nested contours so that they all lie upon @xmath44 ( cf .
lemma [ expandlem ] ) this can also be written as @xmath70 where @xmath71 is the function which maps @xmath72 .
the @xmath0-boson transform @xmath62 induces an isomorphism between @xmath35 and @xmath36 with inverse given by @xmath66 .
moreover , for any @xmath73 @xmath74 and for any @xmath49 @xmath75 here @xmath76 is defined via its action @xmath77 and is the operator which maps right eigenfunctions to left eigenfunctions .
one immediate corollary of this plancherel isomorphism theorem is the completeness of the coordinate bethe ansatz .
[ compl ] for all @xmath78 , @xmath79 this shows that the space @xmath35 is decomposed onto left ( and right ) eigenfunctions with spectral variables @xmath80 over all @xmath81 and @xmath82 of length @xmath83 , with respect to the ( complex - valued ) plancherel measure given above in ( [ dmulambda ] ) .
the lack of positivity of this measure should be compared to the case of hermitian hamiltonians such as the heisenberg xxz quantum spin chain on @xmath2 or delta bose gas on @xmath7 .
thus our plancherel isomorphism theorem is not an isomorphism between @xmath84 spaces .
there are some degenerations of this system which admit @xmath84 space isomorphisms , namely the system considered in sections [ vandiejen ] and [ deltabosesec ] .
the system considered in section [ semidiscsec ] involves a complex - valued plancherel measure , and likewise does not admit an @xmath84 isomorphism . in applying a contour shift argument ( used , for instance , in heckman - opdam s proof of the delta bose gas plancherel formula @xcite ) in this proof of the plancherel formula
, the non - hermitian nature of our hamiltonian introduces some additional non - triviality .
however , the symmetry given earlier in ( [ ptsym ] ) between @xmath12 and @xmath20 ( see also remark [ commrem ] ) is a suitable replacement for hermiticity .
the plancherel isomorphism theorem also contains within it the biorthogonality ( with respect to the spatial variable @xmath6 and spectral variable @xmath82 ) of the left and right eigenfunctions ( see corollary [ north ] and proposition [ specorth ] below ) . the completeness result above along with the fact that @xmath85 is a right eigenfunction for @xmath12 , enables us to explicitly solve the kolmogorov backward equation for the @xmath0-boson particle system with general initial data in @xmath35 .
the @xmath0-boson particle system plays an analogous role to @xmath0-tasep , as the continuum delta bose gas plays to the kardar - parisi - zhang equation ( cf .
section [ deltabosesec ] ) .
more precisely , due to a markov duality ( see section [ qtasepsec ] or @xcite ) , if we write the trajectory of @xmath0-tasep as @xmath86 , then @xmath87\ ] ] solves the @xmath0-boson particle system kolmogorov backward equation @xmath88 with initial data depending on the initial condition @xmath89 for @xmath0-tasep . since we know how to solve the backward equation ( via the spectral decomposition above ) , this gives us access to exact formulas for moments of @xmath0-tasep started from any initial data .
however , such an exact formula does involve evaluating @xmath90 , where @xmath91 . for certain types of initial data
this summation can be explicitly calculated ( see for instance corollary [ abovecor2 ] and lemma [ belowlemma ] ) .
this is due to the fact that if @xmath92 for some function @xmath93 , then the plancherel isomorphism theorem implies that @xmath94 .
this provides a systematic way to discover and proof many new combinatorial formulas .
in particular , for @xmath0-tasep with step initial condition ( i.e. , @xmath95 for @xmath96 ) , this approach leads to ( see proposition [ dualiyresult ] below ) @xmath97 = ( -1)^k q^{\frac{k(k-1)}{2 } } \oint_{\gamma_1 } \frac{dz_1}{2\pi { \mathbf i}}\cdots \oint_{\gamma_k } \frac{dz_k}{2\pi { \mathbf i } } \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \prod_{j=1}^{k } ( 1-z_j)^{-n_j } \frac{e^{(q-1)tz_j}}{z_j},\ ] ] where @xmath98 are as above , with the additional condition that they do not include @xmath99 . proving this formula
requires us to go beyond the functional spaces @xmath36 and @xmath35 , which in this case is readily doable .
this moment formulas for step initial data has appeared previously ( though not via this spectral approach ) . with all @xmath100 it was first proved in ( * ? ? ?
* section 3.3 ) using the theory of macdonald processes .
the general @xmath6 result was then proved in ( * ? ? ?
* theorem 2.11 ) by simply guessing the above formula and checking that it satisfied the backward equation and initial data .
the macdonald process approach was then extended in @xcite to also cover general @xmath6 . extending from step to general initial data was unclear until the present work .
since @xmath101 , the above moment formulas completely characterize the distribution of @xmath0-tasep at time @xmath8 .
so far , this has been used to write a fredholm determinant formula for the location of a particle @xmath102 ( cf .
* theorem 3.2.11 ) or ( * ? ? ? * theorem 3.12 ) ) . in turn , this formula have served as effective starting point for asymptotic analysis of a variety of models related to @xmath0-tasep ( cf .
@xcite , and related work @xcite ) .
the direct asymptotics ( for instance to the tracy - widom gue distribution and kpz universality class ) of @xmath0-tasep has not yet been performed , though it would seem that the above mentioned fredholm determinant formula is well suited for that .
thus , as pointed out earlier , @xmath0-tasep and the @xmath0-boson particle process serve as discrete regularizations of the kardar - parisi - zhang equation and the continuum delta bose gas , for which the physics replica method can be put on firm mathematical ground .
there are three related motivations which led to the present work : the polymer replica method from physics , tracy - widom s asep transition probability formulas , and measures on partitions and gelfand - tsetlin patterns ( sequences of interlacing partitions ) such as schur , whittaker and macdonald processes .
we have already discussed much of our motivation coming from the polymer replica method .
let us just add that the connection between the moments of the stochastic heat equation and the delta bose gas does not seem to be firmly established in the mathematical literature and some of the types of initial data one wishes to study for the kpz equation ( such as narrow wedge ) fall outside of the realm of the presently proved plancherel formulas .
thus , by working with a discrete regularization , one may hope to avoid such issues as well . in 2008 ,
tracy - widom @xcite computed an exact formula for the transition probability of the @xmath52-particle asep . though the coordinate bethe ansatz was central to their work , they did not diagonalize the @xmath52-particle asep hamiltonian .
we address this diagonalization in the companion paper @xcite as well as the relationship between tracy - widom s work and the asep plancherel formula .
tracy - widom have recorded further progress in computing transition probabilities for variants of asep on @xmath2 ( such as on @xmath103 @xcite ) .
lee @xcite has further developed tracy - widom s methods .
indeed , soon after the first posting of the present work , korhonen - lee @xcite demonstrated how the tracy - widom approach can be employed to find transition probabilities for the @xmath0-boson particle system ( see also section [ transprobsec ] below )
. it would be reasonable to try to develop analogous results to ours in all of these other settings .
measures on partitions and gelfand - tsetlin patterns have played an important role in a wide variety of probabilistic systems including random matrix theory , random growth processes , interacting particle systems , directed polymer models , and random tilings ( see , for example , the recent review @xcite and references therein ) . during the last 15 years
there was significant activity in the analysis of determinantal measures in particular the so - called schur processes @xcite which are written in terms of schur symmetric functions .
it is possible to define generalizations of the schur processes by replacing schur symmetric functions with macdonald symmetric functions ( their two - parameter @xmath104-generalizations ) . until recently ,
little was done with respect to these generalizations , owing largely to the fact that they no longer are determinantal . in 2009 ,
oconnell @xcite introduced the whittaker process and related it to the oconnell - yor semi - discrete directed polymer partition function .
furthermore , he utilized a whittaker function integral identity to compute an exact formula for the laplace transform of the partition function .
as whittaker functions are limits ( for @xmath105 and @xmath106 ) of macdonald symmetric functions @xcite , this development was a source of motivation for borodin - corwin s @xcite subsequent 2011 work on macdonald processes .
the work of borodin - corwin @xcite gives answers to two basic questions about macdonald processes ( and their various degenerations ) : what is their probabilistic content and how can one compute meaningful information and asymptotics . in order to endow the macdonald processes with probabilistic content
, @xcite showed how to construct markov dynamics which map macdonald processes to other macdonald processes ( with evolved sets of parameters ) .
the construction in @xcite is an adaption of one from the schur process context due to borodin - ferrari @xcite ( and based on an idea of diaconis - fill @xcite ) .
there are other markov dynamics ( at various levels of degeneration of macdonald processes ) besides those considered in @xcite
see @xcite .
the @xmath0-tasep was discovered as a one - dimensional marginal of the markov dynamics corresponding to setting the macdonald parameters @xmath105 , and @xmath101 .
when @xmath107 , @xmath0-tasep becomes the usual tasep , and the above macdonald process turns into a schur process . the second challenge in studying macdonald processes was to compute meaningful information and asymptotics . in @xcite
this was accomplished by directly appealing to the integrable structure of the macdonald symmetric polynomials in particular their explicit eigenfunction relationships with the macdonald difference operators .
this structure naturally led to nested contour integral formulas for expectations of various observables of the macdonald processes . with regards to @xmath0-tasep ,
this provided formulas for the expectation @xmath108 $ ] like ( [ page6star ] ) above . as a stochastic process
, @xmath0-tasep has a scaling limit to the logarithm of the semi - discrete stochastic heat equation ( or free energy of the oconnell - yor semi - discrete directed polymer ) and further to the kpz equation ( logarithm of continuum stochastic heat equation ) .
the limits of the observables studied for @xmath0-tasep correspond to products of values of the solution to these stochastic heat equations at fixed time and varying spatial location , and thus @xcite found similar nested contour integral formulas for these limits .
as explained earlier , for the stochastic heat equation , these observables satisfy the delta bose gas and have a significant literature surround them .
the nested contour integral formulas for the delta bose gas seem to have first appeared in 1985 physics work of yudson @xcite and reemerged independently in 1997 work of heckman - opdam @xcite proving the plancherel formula for the delta bose gas .
a limitation of the macdonald process approach to studying @xmath0-tasep is that it only works for initial conditions of @xmath0-tasep which arise as marginals of macdonald processes .
there are certainly many examples of such initial conditions ( an infinite - dimensional family that , in particular , includes the step initial condition ) .
generic initial conditions , and certain important `` examples '' of initial conditions ( such as flat or half - flat ) do not arise in this manner and hence can not be treated via this method .
inspired by the connection between the nested contour integral formulas and the delta bose gas , borodin - corwin - sasamoto @xcite discovered that it was possible to implement the ideas of the polymer replica method at the level of @xmath0-tasep ( and asep ) .
in particular , they showed that for any initial conditions , the expectations @xmath109 solve a discrete @xmath0-deformed ( or regularized ) version of the delta bose gas in which the @xmath0-boson particle system generator replaces the delta bose gas hamiltonian .
the initial data for the @xmath0-boson particle system backward equation directly corresponds with the initial condition for @xmath0-tasep , and @xcite showed that for step and half - stationary @xmath0-tasep initial conditions the solution @xmath109 was given by nested contour integral formulas .
the step solution came directly from @xcite while the half - stationary did not ( though should correspond to a two - sided generalization of macdonald processes such as developed in the schur context in @xcite ) .
the question of how to solve the @xmath0-boson particle system for general initial data ( and hence solve for @xmath109 for @xmath0-tasep with general initial conditions ) is the central motivation of the present work .
the plancherel formula and its various consequences provide a solution to this question .
a fortiori , this work provides some motivation to better understand the relationship between various algebraic structures .
one question which was answered in @xcite was how the theory of macdonald processes implies the fact that @xmath109 solves the @xmath0-boson particle system .
this relationship turned out to be a consequence of a commutation relation satisfied by the macdonald first order difference operators .
however , there remain a few algebraic mysteries which we now touch upon .
the @xmath0-boson particle system on a periodic lattice arises as a stochastic representation of the @xmath0-boson hamiltonian considered in @xcite .
this hamiltonian is a generalization of an earlier version considered by bogoliubov - bullough - timonen @xcite and bogoliubov - izergin - kitanine @xcite .
the algebraic @xmath0-boson hamiltonian from @xcite is written as @xmath110 here @xmath111 , @xmath112 and another set @xmath113 belong to the @xmath0-boson algebra generated by the relations @xmath114
= b_j^{\dag } { \ensuremath{\mathbf{1}}}_{j = k } , \qquad [ n_j , b_k ] = -b_j { \ensuremath{\mathbf{1}}}_{j = k } , \qquad [ b_j , b_k^{\dag } ] = q^{n_j } { \ensuremath{\mathbf{1}}}_{j = k}.\ ] ] one imposes periodic ( of length @xmath15 ) boundary conditions by defining @xmath115 . a periodic version of the @xmath0-boson particle system backward generator corresponds to a particular representation of @xmath116 in which the operators act on functions @xmath117 as @xmath118 one checks that these operators satisfy the @xmath0-boson algebra .
the operator @xmath116 acts on functions @xmath117 as @xmath119 where @xmath120 , and where @xmath121 .
this is the backward generator of the @xmath0-boson particle system ( see section [ stochrepsec ] ) on a periodic portion of @xmath2 .
this algebraic @xmath0-boson hamiltonian arises from the @xmath122-matrix given in ( 5.1 ) of @xcite and its integrability ( in the sense of satisfying the yang - baxter equations ) was shown there as well .
the version studied earlier in @xcite corresponds to @xmath123 and is only stochastic in the limit when @xmath106 . due to our probabilistic motivations
, we primarily consider @xmath124 here , though in section [ epdef ] we consider the general @xmath60 hamiltonian ( under the identification of @xmath125 ) and show how our results extend .
the spectral theory for the periodic @xmath123 case has received attention recently in @xcite , and likewise for the infinite lattice ( as considered herein ) @xmath123 case in @xcite . in these cases ,
hall - littlewood polynomials play the role of left and right eigenfunctions .
another algebraic mystery has to do with two sets of symmetric functions .
the macdonald process related to @xmath0-tasep is defined in terms of macdonald polynomials at @xmath105 and @xmath101 ( these are sometimes called @xmath0-whittaker functions @xcite ) . via the connection to the @xmath0-boson particle system
we naturally arrive at another set of symmetric polynomials the eigenfunctions for the @xmath0-boson particle system which appear to be quite remarkable and ( though not the same ) are close relatives of hall - littlewood symmetric polynomials ( and degenerate to them in the limit corresponding to @xmath123 above ) .
we do not presently have an algebraic explanation for this transition from one type of symmetric polynomials to another .
perhaps a further investigation into the relation of @xmath0-bosons to degenerations of the double affine hecke algebra might shed light on this ( see developments in this direction in @xcite and references therein ) .
the work of @xcite showed that for the particular step and half - stationary initial conditions , the relationship between @xmath0-tasep and the @xmath0-boson particle system persists when inhomogeneous particle jumping rate parameters @xmath126 are introduced .
these @xmath127-parameters have an important meaning from the macdonald perspective as the variables of the macdonald polynomials . from the perspective of the @xmath0-boson particle system , @xcite showed that the backward generator ( similar holds true for the conjugated forward generator ) was equivalent to a free generator with inhomogeneous rates depending on the @xmath127 s and the same two - body boundary condition as occurs when all @xmath128 . despite this evidence
, we do not presently know how to develop a general @xmath127-parameter version of the eigenfunctions and plancherel formula .
one reason why a general @xmath127-version of our present result could be quite enlightening is that we do not have an analog of macdonald processes or @xmath127-parameters in relation to asep and the heisenberg xxz quantum spin chain , and this may help guide that investigation . in @xcite two discrete time versions of @xmath0-tasep were introduced and studied .
their moments solve discrete variants of the @xmath0-boson particle system evolution equations .
it would be interesting to develop the parallel of the results of this paper in the discrete - time context .
there is another generalization of @xmath0-tasep which is called @xmath0-pushasep ( cf .
it seems likely that the analog of the @xmath0-boson particle system for these variants of @xmath0-tasep should have the eigenfunctions as above ( with different eigenvalues ) . as explained in @xcite
, this should also hold true for any operator which arises out of specializing the spectral parameter of the @xmath0-boson transfer matrix @xmath129 ( since transfer matrices commute for different spectral parameters ) .
it is reasonable , then , to hope to show that these discrete systems arise from the transfer matrix . at the same time as the present work was posted , povolotsky @xcite has posted a work in which he introduces a class of models solvable via coordinate bethe ansatz which includes the @xmath0-boson particle system ( and also those of @xcite ) as special cases ( see also earlier related work @xcite ) . similar to here , he computes the eigenfunctions via coordinate bethe ansatz .
he also conjectures their completeness and orthogonality .
section [ appsec ] below shows exactly how our plancherel theorem proves completeness and biorthogonality for the @xmath0-boson particle system .
we hope our methods will extend to the full class of models discussed above .
we collect many of the definitions and notations used throughout the paper ( including some already discussed earlier in the introduction ) .
define @xmath130 for @xmath131 set @xmath132 the backward difference operator @xmath133 and forward difference operator @xmath134 act on functions @xmath135 as @xmath136 for a function @xmath137 , and @xmath138 , we write @xmath139 and @xmath140 as the application of the respective operator in the variable @xmath141 . the space - reflection operator @xmath142 acts on functions @xmath135 as @xmath143 define a symmetric bilinear pairing @xmath45 on functions @xmath144 via @xmath145 ( assuming the series converges ) and note that @xmath146 for any function @xmath147 , and that @xmath148 to @xmath16 of the form @xmath149 we associate the list of ordered cluster sizes @xmath150 ( where @xmath15 is the number of clusters ) .
we also associate to @xmath6 the ordered list @xmath151 of gaps between the clusters of @xmath6 .
thus for @xmath152 , @xmath153 and by convention @xmath154 .
we write these functions of @xmath6 as @xmath155 , @xmath156 and @xmath157 . to illustrate , for @xmath158 , @xmath159 , @xmath160 and @xmath161 .
the pair @xmath162 identifies @xmath6 modulo a global shift .
a partition @xmath163 is a weakly decreasing set of non - negative integers with finitely many nonzero entries .
the size of @xmath164 is @xmath165 and the length of @xmath164 is @xmath166 . if @xmath167 we write @xmath81 . to a partition we associate its multiplicities @xmath168 and
sometimes write @xmath169 . for a partition @xmath81 of length @xmath170 and a set of variables
@xmath171 we define @xmath172 we will also use an additive version of @xmath80 defined as @xmath173 in general , an arrow over a variable ( such as @xmath174 ) denotes a vector @xmath175 whose length is generally @xmath52 ( or else explicitly stated ) .
we define the space @xmath35 of compactly supported functions @xmath13 .
we also define the space @xmath36 of symmetric laurent polynomials in @xmath38 .
in particular , any @xmath176 can be written as @xmath177 or as @xmath178 with all but finitely many coefficients @xmath179 fixed to be zero , and @xmath180 is the vandermonde determinant . throughout
we assume that the parameter @xmath101 , though in section [ complexq ] we explain how our results ( or modifications of them ) hold for more general @xmath10 .
the @xmath0-pochhammer symbol and @xmath0-factorial are defined as @xmath181 the @xmath106 limit of @xmath182 is written @xmath183 .
the symmetric group on @xmath52 elements is denoted by @xmath184 , the indicator function for an event @xmath185 by @xmath186 , and the identity operator by @xmath187 .
this work was started at the simons symposium on the kardar - parisi - zhang equation and advanced during the lebesgue centre s summer school on the kpz equation and rough paths and the cornell probability summer school .
ic appreciates discussions on bethe ansatz for non - hermitian hamiltonians with gunter schtz while visiting the hausdorff institute .
the authors all appreciate illuminating comments and discussions of a draft of this work with jan felipe van diejen , erdal emsiz , christian korff , alexander povolotsky and gunter schtz .
ab was partially supported by the nsf grant dms-1056390 .
ic was partially supported by the nsf through dms-1208998 as well as by microsoft research through the schramm memorial fellowship , and by the clay mathematics institute through a clay research fellowship .
lp was partially supported by the rfbr - cnrs grant 11 - 01 - 93105 .
ts was supported by kakenhi ( 22740054 ) .
in section [ qbosonsyst ] we use the coordinate bethe ansatz to determine algebraic eigenfunctions for the forward and backward generators of the @xmath0-boson particle system .
the main result of this paper the plancherel isomorphism theorem involving these eigenfunctions is proved in section [ plansec ] .
that section also contains results about the completeness and biorthogonality of these eigenfunctions .
section [ appsec ] applies the plancherel formula in order to solve the kolmogorov forward and backward equations , and as an application compute moments of the @xmath0-tasep .
section [ complexq ] records how the earlier results persist when @xmath0 is complex .
section [ semidisc ] describes two limits of the @xmath0-boson particle system , one of which relates to earlier work of van diejen @xcite and the other of which relates to the semi - discrete stochastic heat equation .
finally , in section [ appendixsec ] ( appendix ) we briefly touch on the continuum limit of the work of this paper ( as relates to the continuum delta bose gas and stochastic heat equation ) , prove a proposition related to the expansion of nested contour integrals , and provide a direct combinatorial proof of lemma [ belowlemma ] .
in this section we introduce the @xmath0-boson particle system and use the bethe ansatz to write down algebraic eigenfunctions for its forward and backward generators .
the main idea here is that it is possible ( and straightforward ) to rewrite the generator of the system in terms of a constant coefficient , separable _ free generator _ with @xmath31 two - body boundary conditions imposed .
the coordinate bethe ansatz is then readily applied to provide the desired eigenfunctions .
it is worth recalling that since we work on @xmath2 ( as opposed to a finite or periodic interval ) there are no bethe equations which must be solved in computing the algebraic eigenfunctions .
the usefulness of this approach for interacting particle systems goes back to the work of @xcite .
the @xmath0-boson particle system is a stochastic interacting particle system @xmath188 ( in fact , a certain totally asymmetric zero range process ) with state space @xmath189 ( see @xcite for general background ) . here
@xmath190 represents the non - negative number of particles at a given site @xmath191 . in continuous time , independently for each @xmath192 , one particle may move from site @xmath193 to site @xmath194 at rate @xmath195 ( i.e. , according to an exponential waiting time of that rate ) .
this corresponds to changing @xmath196 to @xmath197 .
we will focus on the restriction of this process to a finite number @xmath1 of particles .
since the dynamics conserve the number of particles , this restriction is itself a stochastic interacting particle system ( see figure [ qboson ] for an illustration of @xmath198 ) . to any @xmath199 such that @xmath200
we may associate a ( weakly ) ordered list of particle locations @xmath201 .
we write @xmath202 to denote this function .
for example , if @xmath203 and @xmath204 except for @xmath205 , @xmath206 and @xmath207 , then @xmath208 .
thus , we may write @xmath209 as the @xmath0-boson particle system at time @xmath8 in these coordinates
. we may write , in these @xmath6-coordinates , the backward and forward generators for the @xmath52 particle restriction of the @xmath0-boson particle system ( see also section [ appsec ] ) . in what follows
we write @xmath210 , @xmath211 , and @xmath212 ( recall from section [ notations ] ) thus suppressing the @xmath6 dependence .
-boson particle system with 9 particles . in the left - most state ,
@xmath213 , @xmath214 , @xmath215 , @xmath216 , @xmath217 and @xmath218 .
the value of @xmath6 associated to this state is given by reading off the location of the labels @xmath219 .
in the middle state , the particle corresponding with @xmath220 has decreased location by 1 ( as occurs at rate @xmath221 , and in the right - most state the particle corresponding with @xmath222 has decreased location by 1 ( as occurs at rate @xmath223 ) . ] [ bwdgendef ] the @xmath0-boson backward generator acts on functions @xmath13 as @xmath224 where @xmath225 is defined in section [ notations ] .
the @xmath0-boson forward generator ( the matrix transpose of @xmath12 ) acts on function @xmath13 as @xmath226 where by convention , for @xmath227 we set @xmath228 .
the function @xmath21 depends only on the list @xmath229 of cluster sizes for @xmath6 via @xmath230 let @xmath231 denote @xmath232 .
we will also write @xmath26 and @xmath233 as ( diagonal ) multiplication operators defined so that @xmath234 and @xmath235 .
one readily sees that @xmath26 and the space - reflection operator @xmath142 commute .
the @xmath0-boson conjugated forward generator is defined as @xmath236 [ equalafwdlem ] we have that @xmath237 since @xmath238 is defined via conjugation of @xmath20 by the diagonal matrix @xmath26 , it suffices to check the coincidence of the off - diagonal matrix elements of @xmath238 with those of the right - hand side of ( [ equalafwd ] ) . observe that if @xmath239 then @xmath240 . from the definition of @xmath26
it then follows that @xmath241 similarly , if @xmath242 then @xmath243 . then @xmath244 these are exactly the ratios of the off - diagonal matrix elements of ( [ afwddef ] ) and ( [ equalafwd ] ) .
let us note that in the above proof the constant @xmath245 in the definition of @xmath26 played no role .
this choice of constant is fixed in the proof of theorem [ kqbosonid ] below , by ( [ kqbosonthmend ] ) and the computation following it .
another choice of constant would result in proving that @xmath246 acts as a constant ( not equal to one ) times the identity operator .
[ commrem ] lemma [ equalafwdlem ] implies that @xmath247 or equivalently @xmath248 showing that @xmath12 and @xmath20 are related via a similarity transform .
up to a constant , the function @xmath249 is the invariant measure for the @xmath0-boson particle process ( see ( * ? ? ?
* corollary 3.3.12 ) ) . in the study of interacting particle systems ,
the conjugated forward generator @xmath238 is sometimes known as the adjoint or time - reversed generator .
the above relationship shows that our present model is pt - invariant , i.e. , invariant under joint space - reflection and time - reversal .
this type of invariance is of fundamental significance in many areas of physics and features prominently in some papers on interacting particle systems such as @xcite . it may be possible to extend the present approach to similarly study and develop spectral theory for other pt - invariant bethe ansatz solvable models .
we are grateful to gunter schtz for noting that the above relation is equivalent to the pt - invariance of the model .
pt - invariance is a property which holds true for all spatially homogeneous nearest neighbor asymmetric zero range processes , so long as @xmath26 is replaced by the product invariant measure for the process .
the proof of this general fact follows the same lines as lemma [ equalafwdlem ] .
the key to finding the eigenfunctions for the backward and conjugated forward generators is to reexpress them in terms of constant - coefficient , separable _
free generators _ subject to @xmath31 two - body boundary conditions .
it is certainly not typical that the generators of an interacting particle system can be reexpressed in this manner ( see however @xcite for asep ) .
this is the hallmark of integrable systems of bethe - ansatz type .
recall the definitions of @xmath139 and @xmath140 from section [ notations ] .
[ freedef ] the @xmath0-boson backward free generator @xmath250 acts on functions @xmath251 as @xmath252 we say that the function @xmath251 satisfies the @xmath31 @xmath0-boson backward two - body boundary conditions if @xmath253 the @xmath0-boson forward free generator @xmath254 acts on functions @xmath251 as @xmath255 we say that the function @xmath251 satisfies the @xmath31 @xmath0-boson forward two - body boundary conditions @xmath256 [ freetrueequiv ] if @xmath251 satisfies the @xmath31 @xmath0-boson backward ( respectively , forward ) two - body boundary conditions , then for @xmath16 , @xmath257 assume that @xmath258 satisfies ( [ star1 ] ) .
if @xmath6 has @xmath259 , then @xmath260 the same reasoning applies to the other clusters of @xmath6 readily implying that @xmath261 .
likewise assume that @xmath258 satisfies ( [ star2 ] ) .
if @xmath6 has @xmath259 , then it follows that @xmath262 the same reasoning applies to the other clusters of @xmath6 readily implying ( we have also appealed here to lemma [ equalafwdlem ] ) that @xmath263 .
[ eigdeffree ] a function @xmath264 is called an eigenfunction of the @xmath0-boson backward free generator with @xmath31 two - body boundary conditions if @xmath265 is an eigenfunction for the @xmath0-boson backward free generator and it satisfies the @xmath31 @xmath0-boson backward two - body boundary conditions ( [ star1 ] ) .
we likewise define what it means for a function @xmath266 to be an eigenfunction of the @xmath0-boson forward free generator with @xmath31 two - body boundary conditions ( [ star2 ] ) .
the following is a corollary of proposition [ freetrueequiv ] .
[ eigcorgen ] any eigenfunction @xmath264 for the @xmath0-boson backward free generator with @xmath31 two - body boundary conditions is , when restricted to @xmath16 , an eigenfunction for the @xmath0-boson backward generator @xmath12 with the same eigenvalue .
similarly any eigenfunction @xmath266 for the @xmath0-boson forward free evolution equation with @xmath31 two - body boundary conditions is , when restricted to @xmath16 , an eigenfunction for the @xmath0-boson conjugated forward generator @xmath238 with the same eigenvalue . in turn
, @xmath267 is an eigenfunction for the @xmath0-boson forward generator @xmath20 with the same eigenvalue .
we briefly review the coordinate bethe ansatz @xcite .
given an operator @xmath122 acting on functions from @xmath268 ( where @xmath269 is an arbitrary space ) we form the @xmath52-particle operator @xmath270 which acts on functions @xmath271 as @xmath272 where @xmath273 acts as @xmath122 on the coordinate @xmath274 .
write @xmath275 to denote any eigenfunction with eigenvalue @xmath276 such that @xmath277 ( we are not assuming all @xmath278 correspond to an eigenfunction or that all eigenvalues are simple ) .
then it follows that @xmath279 is an eigenfunction of @xmath270 with eigenvalue @xmath280 .
presently @xmath281 can be arbitrary .
consider an operator @xmath282 which acts on functions @xmath283 .
we say that @xmath284 satisfies the two - body boundary condition corresponding to @xmath282 if for all @xmath285 , @xmath286 .
let @xmath287 act as @xmath282 in the variables @xmath274 and @xmath288 and as the identity for all other variables .
we wish to find @xmath289 such that @xmath290 this can be accomplished by choosing @xmath291 suitably .
define the function @xmath292 where @xmath293 is defined as @xmath282 applied to the function which takes @xmath294 .
[ abetheconst ] if @xmath295 then @xmath296 from ( [ psibethe ] ) is an eigenfunction of the operator @xmath270 from ( [ cala ] ) with eigenvalue @xmath280 , for which the @xmath31 two - body boundary conditions in ( [ bcbethe ] ) hold . for any
@xmath297 consider the pair of terms in ( [ psibethe ] ) corresponding to @xmath298 and @xmath299 where @xmath300 represents the transposition of @xmath193 and @xmath301 .
we will show that @xmath287 applied to this pair @xmath302 is identically zero if @xmath303 .
if this holds for all pairs , this implies the same holds true for @xmath296 .
call the first term above @xmath304 and the second @xmath305 .
then it follows from the definition of the function @xmath306 that @xmath307 for @xmath308 such that @xmath303 @xmath309 and thus @xmath310 we may apply the coordinate bethe ansatz of section [ corbetherev ] to the @xmath0-boson backward and forward free generators with @xmath31 two - body boundary conditions of section [ freegensec ] .
one should take care to notice that the eigenfunction associated to @xmath174 has eigenvalue @xmath311 and not @xmath312 as in the above general construction .
[ eigdefn ] for all @xmath28 , set @xmath313 observe that for any fixed @xmath16 these are symmetric laurent polynomials in @xmath314 thus elements of @xmath36
. [ prop211 ] for all @xmath28 , @xmath33 is an eigenfunction for the @xmath0-boson backward free generator with @xmath31 two - body boundary conditions ( definition [ eigdeffree ] ) with eigenvalue @xmath315 .
the restriction of @xmath33 to @xmath16 is consequently an eigenfunction for the @xmath0-boson backward generator @xmath12 with the same eigenvalue .
similarly , for all @xmath28 , @xmath316 is an eigenfunction for the @xmath0-boson forward free generator with @xmath31 two - body boundary conditions ( definition [ eigdeffree ] ) with eigenvalue @xmath315 .
the restriction of @xmath316 to @xmath16 is consequently an eigenfunction for the @xmath0-boson conjugated forward generator @xmath238 with the same eigenvalue .
the restriction of @xmath32 to @xmath16 is likewise an eigenfunction for the @xmath0-boson forward generator @xmath20 with the same eigenvalue .
let us focus on the backward case first .
observe that @xmath317 is an eigenfunction for @xmath318 with eigenvalue @xmath319 .
we may apply lemma [ abetheconst ] to construct an eigenfunction for the backward free generator which satisfies the two - body boundary conditions of definition [ freedef ] . in this application of lemma [ abetheconst ]
we find that @xmath320 .
comparing the outcome of the lemma to the expression above for @xmath265 we see that they differ by an overall multiplicative factor @xmath321 however , since the @xmath278-variables are fixed , this factor is just a constant , and hence the desired result follows .
the implication for the backward generator @xmath12 follows from corollary [ eigcorgen ] .
similar reasoning applies in the forward case . observe that @xmath322 is an eigenfunction for @xmath323 with eigenvalue @xmath319 .
the forward two - body boundary condition can be multiplied by @xmath324 yielding @xmath325 we may apply lemma [ abetheconst ] to construct an eigenfunction for the forward free generator which satisfies the above two - body boundary conditions . in this application of lemma [ abetheconst ]
we find that @xmath326 . just as in the backward case
, this differs from @xmath327 by an overall multiplicative constant ( that depends only on the @xmath278-variables ) hence the desired result follows .
the implication for the conjugated forward generator @xmath238 and the forward generator @xmath20 follows from corollary [ eigcorgen ] .
[ qbosonhamiltoniansym ] we may extend the eigenfunctions of definition [ eigdefn ] so as to be defined for all of @xmath328 ( rather than @xmath329 ) by fixing that the value for a general @xmath131 is the same as the value of @xmath330 ( where @xmath297 is a permutation of the elements of @xmath6 taking it into @xmath329 ) .
it is possible to write down an operator on all of @xmath328 for which these extensions ( which we write with the same notation ) are still eigenfunctions . for instance ( cf .
* proposition 2.7 , c ) ) one readily observes that @xmath331 { \ensuremath{\psi^{\textrm{bwd}}}}_{\vec{z}}(\vec{n } ) = ( 1-q ) ( z_1+\cdots + z_k)\ , { \ensuremath{\psi^{\textrm{bwd}}}}_{\vec{z}}(\vec{n}).\ ] ] the extension of @xmath33 to @xmath328 is through symmetric extension and it is not clear how to modify the formula for the eigenfunction so that the @xmath6 coordinates do not require permuting . for the case of the delta bose gas ( a particular limit of the present system considered later in section [ deltabosesec ] )
such a formula can be seen in ( * ? ? ?
* equation ( a.12 ) ) .
proposition [ prop211 ] along with the fact that @xmath20 is the transpose of @xmath12 implies that @xmath332 showing the @xmath32 and @xmath33 are ( respectively ) right and left eigenfunctions for @xmath20 with eigenvalue @xmath315 .
this motivates the following .
[ leftrighteig ] for any @xmath333 define @xmath334 where @xmath249 is given in ( [ cqdef ] ) .
likewise define @xmath335 we could have just as well defined the right and left eigenfunctions with respect to the operator @xmath12 ( see remark [ wrtbwdremark ] for the implications of this ) .
finally , observe the symmetry of @xmath336 and @xmath337 with respect to the space - reflection operator @xmath338
in this section we prove a plancherel formula ( theorem [ kqbosonid ] ) related to the @xmath0-boson particle system .
we define the transform @xmath62 in which functions are paired with right eigenfunctions .
the ( candidate ) inverse @xmath66 is defined in terms of nested contour integrals , though can also be expressed via a ( different ) pairing with left eigenfunctions with respect to a ( complex ) plancherel measure .
the composition of these transforms is written as @xmath339 .
theorem [ kqbosonid ] shows that on @xmath35 , @xmath246 acts as the identity operator .
the proof of this theorem is quite simple and relies on two steps . in the first step
we demonstrate a certain symmetry of @xmath246 and in the second step we use elementary residue considerations and this symmetry to prove that @xmath246 acts as the identity . a key step in our proof
has a history and is sometimes called the _ contour shift argument_. in the related continuum delta bose gas ( cf .
section [ deltabosesec ] ) , heckman and opdam @xcite used a similar approach to prove a plancherel formula for that system .
they , in turn , attribute it to van den ban and schlichtkrull @xcite and ultimately to helgason , gangolli and rozenberg s argument in the proof of the plancherel theorem for a riemannian symmetric space @xmath340 @xcite .
it seems , in fact , that the original instance of this argument goes back to helgason s 1966 work @xcite .
the other result we prove in this section is a dual plancherel formula ( theorem [ kqbosoniddual ] ) which shows that on @xmath36 , @xmath341 acts as the identity operator .
the proof of this dual plancherel formula goes through a spectral orthogonality of the eigenfunctions @xmath337 and @xmath336 which is given in proposition [ specorth ] . combining the placherel formula and
the dual plancherel formula leads to the plancherel isomorphism given in theorem [ isothm ] .
we introduce the @xmath0-boson transform and the @xmath0-boson inverse transform , as well as the contours @xmath39 , and @xmath342 .
recall the definitions of the function spaces @xmath35 and @xmath36 from section [ notations ] .
[ kqtasepdef ] the @xmath0-boson transform @xmath62 takes functions @xmath63 into functions @xmath64 via @xmath65 fix any set of positively oriented , closed contours @xmath98 chosen so that they all contain @xmath40 , so that the @xmath41 contour contains the image of @xmath0 times the @xmath42 contour for all @xmath43 , and so that @xmath44 is a small enough circle around 1 so as not to contain @xmath0 .
let us also fix contours @xmath60 and @xmath343 where @xmath60 is a positively oriented closed contour which contains 1 and its own image under multiplication by @xmath0 , and @xmath343 contains @xmath60 and is such that for all @xmath344 and @xmath345 , @xmath346 .
see figure [ circontours ] for an illustration of a possible set of such contours . and
@xmath101 , a possible set of contours @xmath347 as well as possible contours for @xmath60 and @xmath343 ( see definition [ kqtasepdef ] ) . ] the ( candidate ) @xmath0-boson inverse transform @xmath66 takes functions @xmath67 into functions @xmath68 via @xmath348
the composition of the transform and ( candidate ) inverse transform takes functions @xmath78 into functions @xmath349 via @xmath350
the composition of the ( candidate ) inverse transform and the transform takes functions @xmath176 into functions @xmath351 via @xmath352 it is an easy residue calculation to check that @xmath62 maps @xmath35 into @xmath36 and @xmath66 maps @xmath36 into @xmath35 .
it is convenient to work with these function spaces since it avoids certain issues such as showing convergence of the summation defining @xmath62 .
we expect ( though do not attempt to prove ) that the results which we now develop can be shown to hold true with respect to larger function spaces .
the operator @xmath66 can be written in terms of @xmath85 in two ways .
[ 321prime ] consider a symmetric function @xmath353 and positively oriented , closed contours @xmath39 and @xmath60 such that : * the contour @xmath60 is a circle around 1 and 0 which contains the image of @xmath60 multiplied by @xmath0 ; * for all @xmath354 , the interior of @xmath41 contains @xmath40 and the image of @xmath42 multiplied by @xmath0 ; * for all @xmath355 , there exist deformations @xmath356 of @xmath357 to @xmath60 so that for all @xmath358 with @xmath359 for @xmath360 , and @xmath361 for @xmath362 , the function @xmath363 is analytic in a neighborhood of the area swept out by the deformation @xmath356 .
then , @xmath364 where @xmath61 is the partition with @xmath52 ones and @xmath365 is defined in ( [ dmulambda ] ) . due to the hypothesis on @xmath366 and the contours , the nested contours in ( [ jqbosontrans ] ) can be sequentially ( starting with @xmath367 to @xmath44 ) deformed to the single contour @xmath60 without changing the value of the integral .
since all contours are now the same , we may symmetrize the integrand . towards this end
we rewrite the product @xmath368 where the first product on the right - hand side is symmetric .
the eigenfunction @xmath85 comes from this symmetrization , and the cauchy determinant identity yields the term @xmath369 , thus completing the proof .
it is useful ( especially for later asymptotic purposes ) to also record how the operator @xmath66 can be expressed in terms of contour integrals along the single contour @xmath44 .
the proof of this result is more involved that that above , and given in the appendix .
[ expandlem ] consider a symmetric function @xmath37 and positively oriented , closed contours @xmath98 such that : * the contour @xmath44 is a circle around 1 and small enough so as not to contain @xmath0 ; * for all @xmath354 , the interior of @xmath41 contains @xmath40 and the image of @xmath42 multiplied by @xmath0 ; * for all @xmath355 , there exist deformations @xmath356 of @xmath357 to @xmath44 so that for all @xmath358 with @xmath361 for @xmath360 , and @xmath370 for @xmath362 , the function @xmath363 is analytic in a neighborhood of the area swept out by the deformation @xmath356 .
then , @xmath371 where the definition of the notation @xmath372 , and @xmath373 is given in section [ notations ] and where @xmath365 is defined in ( [ dmulambda ] ) .
this lemma follows immediately from proposition [ 321 ] .
it follows from lemma [ 321prime ] that for @xmath78 @xmath374 and likewise it follows from lemma [ expandlem ] that @xmath375 [ bilinearprime ] define the ( symmetric ) bilinear pairing @xmath48 on functions @xmath49 via @xmath376 the equivalence of the two expressions on the right - hand sides above is due to lemmas [ 321prime ] and [ expandlem ] .
furthermore they imply that we can write @xmath70 where @xmath71 is the function which maps @xmath72 .
the operator @xmath66 applied to @xmath366 amounts to integrating @xmath377 against a measure which is supported on a disjoint sum of subspaces ( or contours and strings of specializations ) indexed by @xmath81 .
this measure may be called the plancherel measure , though it is most certainly complex valued .
this should be compared to the case of the heisenberg xxz quantum spin chain on @xmath2 @xcite , or the continuum delta bose gas @xcite in which the underlying hamiltonian is hermitian and the plancherel measure is positive , and where one can hope to prove @xmath84 isometries with respect to this measure .
we may now state and prove the main results of this work .
[ kqbosonid ] the @xmath0-boson transform @xmath62 induces an isomorphism between the space @xmath35 and its image with inverse given by @xmath66 .
equivalently , @xmath246 acts as the identity operator on @xmath35 . in order to prove this theorem
we must show that on functions in @xmath35 the operator @xmath378 . before proving that we show the following property ( which follows from the pt - invariance of the eigenfunctions , cf .
remark [ commrem ] ) of @xmath246 : for any functions @xmath73 , @xmath379 in order to prove this claim we utilize the formula ( [ kexpformsmall ] ) for @xmath380 ( we could just as well use ( [ kexpform ] instead with immediately modifications to the next set of equations ) . from this as well as
( [ bilinflip ] ) and ( [ psisym ] ) it follows that @xmath381 moving the @xmath382 operator to the right - hand side and using the fact that @xmath383 , yields ( [ claimedkbos ] ) .
we now show that for any @xmath384 , if @xmath385 then for @xmath386 @xmath387 this claim ( which by linearity of @xmath246 completes the proof of the theorem ) is shown in two parts : first we show that @xmath388 is zero unless @xmath389 , and then we compute the value of @xmath390 .
let @xmath391 . then showing the first part is equivalent to showing that @xmath392 may only be non - zero if @xmath389 . with these choices for @xmath393 and @xmath284
we have that @xmath394 since we may ( without crossing poles ) expand the @xmath367 contour to infinity , the integral in @xmath395 can be evaluated by taking the residue at @xmath396 . from this and the definition of @xmath337
we find that this integral may be non - zero only if @xmath397 . on the other hand , by
( [ claimedkbos ] ) , @xmath398 since we may ( without crossing poles ) shrink the @xmath44 contour to @xmath40 , the integral in @xmath399 can be evaluated by taking the residue at @xmath400 . from this and the definition of @xmath337
we find that this integral may be non - zero only if @xmath401 . combining the above two observations
, we find that for @xmath392 to be non - zero , we must have @xmath402 . now assuming that @xmath402
, we may evaluate the residue at @xmath396 in ( [ z1res ] ) and proceed to expand ( without crossing poles ) the @xmath403 contour to infinity . from this
we find that this integral may be non - zero only if @xmath404 . likewise
, we may evaluate the residue at @xmath400 in ( [ zkres ] ) and proceed to shrink the @xmath405 contour to 1 .
from this we find that this integral may be non - zero only if @xmath406 .
thus @xmath407 .
proceeding in this manner we arrive at the first part of the claim , that @xmath388 is zero unless @xmath389 .
now we evaluate @xmath390 .
observe that by sequentially deforming the contours @xmath367 through @xmath44 to infinity , we find that @xmath408 the residue will be non - zero only for those terms in the summation over @xmath184 in which @xmath298 only permutes within the clusters of @xmath308 ( i.e. , if @xmath203 and @xmath409 , then @xmath298 should stabilize the set @xmath410 and the set @xmath411 ) .
let @xmath150 record the cluster sizes of @xmath308 ( see section [ notations ] ) , and define the set @xmath412 .
we may write such a @xmath297 as a product @xmath413 where each @xmath414 permutes the elements of @xmath415 and fixes @xmath416 .
we denote the set of all such permutations @xmath417 .
owing to this decomposition ( and the fact that for such @xmath298 , @xmath418 ) we may rewrite @xmath419 note the identity ( * ? ? ? * iii.1(1.4 ) ) that for any @xmath420 , @xmath421 it follows from the above identity that @xmath422 the residue is now easily computed since the expression has simple poles at @xmath423 coming from the @xmath424 terms : it equals @xmath425 the limit of the first product is @xmath40 due to the order in which the limits are taken . for the second multiplicative term , since @xmath426 it implies that @xmath427 , each factor of @xmath428 limits to @xmath324 . for each @xmath426
there are a total of @xmath429 such factors .
thus ( [ starstars ] ) is equal to @xmath430 note that @xmath431 and , since @xmath432 , that @xmath433 combining these observations with ( [ kqbosonthmend ] ) shows that @xmath434 , and thus completes the proof of the theorem .
[ wrtbwdremark ] the fact that @xmath378 obviously implies that @xmath435 as well .
this alternative form of the plancherel formula effectively switches the roles of @xmath336 and @xmath337 .
we turn now to a dual plancherel formula .
[ kqbosoniddual ] the inverse @xmath0-boson transform @xmath66 induces an isomorphism between the space @xmath36 and its image with inverse given by @xmath62 .
equivalently , @xmath436 acts as the identity operator on @xmath36 .
observe that for any @xmath176 , @xmath437 can be written ( due to lemma [ 321prime ] ) as @xmath438 thus it suffices to prove that the right - hand side above is equal to @xmath93 . towards this aim , consider any complex valued laurent polynomial @xmath439 in the variables @xmath38 . in order to finish the proof of the lemma it suffices to show the following integrated version holds for all such @xmath440 : @xmath441 from the cauchy determinant identity @xmath442 interchanging the @xmath174 integration and @xmath6 summation the right - hand side of ( [ gidint ] )
can be brought to @xmath443 the equality between the first and second lines is an application of proposition [ specorth ] ( given below and proved independently in section [ semidisc ] ) .
this matches the left - hand side of ( [ gidint ] ) and completes the proof of this theorem .
[ conjrem ] it should be possible to extend the dual plancherel formula to more degenerate classes of functions such as those of the form @xmath444 where @xmath81 and @xmath445 has @xmath83 variables . in this direction we conjecture that the following holds when the contour @xmath60 and class of functions is chosen suitably : @xmath446 this result is related to the conjectured spectral orthogonality given in remark [ orthconj ] .
we may combine theorems [ kqbosonid ] and [ kqbosoniddual ] we arrive to the following result .
[ isothm ] the @xmath0-boson transform @xmath62 induces an isomorphism between @xmath35 and @xmath36 with inverse given by @xmath66 .
moreover , for any @xmath73 @xmath447 and for any @xmath49 @xmath448 here @xmath76 is defined via its action @xmath77 and is the operator which maps right eigenfunctions to left eigenfunctions .
the isomorphism follows immediately by combining theorems [ kqbosonid ] and [ kqbosoniddual ] . to show ( [ res1 ] ) and ( [ res2 ] ) consider an arbitrary @xmath449 , and let @xmath450 ( for @xmath451 ) be the function @xmath452
also , let @xmath453 be the function @xmath454 .
then , @xmath455 the first equality follows from definition [ bilinearprime ] as well as lemma [ expandlem ] .
the second equality follows from theorem [ kqbosonid ] and the final equality follows immediately from the definition of the bilinear pairing .
any function @xmath78 is a ( finite ) linear combination of the functions @xmath456 ( over @xmath451 ) , hence the desired result ( [ res1 ] ) follows by linearity of the bilinear pairings and the fact that @xmath457 .
the result ( [ res2 ] ) follows due to an application of theorem [ kqbosoniddual ] .
theorem [ kqbosonid ] has two immediate corollaries .
the first ( corollary [ completenesscor ] ) is the completeness of the coordinate bethe ansatz for the @xmath0-boson particle system ( with respect to a certain complex plancherel measure ) , and the second ( corollary [ north ] ) is the orthogonality of @xmath458 and @xmath459 when integrated against the plancherel measure .
we call this a _ spatial _ orthogonality since it corresponds to an orthogonality in the variables @xmath6
. there is a second orthogonality which we call _ spectral _ as it is in the variables @xmath174 .
the spectral orthogonality ( given as proposition [ specorth ] below ) is not a consequence of theorem [ kqbosonid ] and , as we have just seen , is the main input in the proof of the dual plancherel formula of theorem [ kqbosoniddual ] .
it follows by taking @xmath460 in proposition [ trueorththm ] , which , in turn , is proved by ultimately appealing to the cauchy - littlewood identity for hall - littlewood symmetric polynomials . for what follows ,
recall the contours @xmath461 from definition [ kqtasepdef ] .
[ completenesscor ] any function @xmath63 can be expanded as @xmath462 and also as @xmath463 the expansion in ( [ completenessfwd ] ) follows from applying @xmath246 to the function @xmath393 . on the one hand ,
theorem [ kqbosonid ] shows that this returns @xmath393 , while ( [ kexpform ] ) provides the above expansion .
the expression in ( [ completenessbwd ] ) comes from the previous expansion as well as the relations ( [ psisym ] ) .
equivalently , it is related to the expansion of the nested contour integral corresponding to the left - hand side of ( [ altplan ] ) . due to lemma [ 321prime ]
we also have the expansion @xmath464 and similarly when @xmath465 and @xmath170 are switched .
the following spatial orthogonality is also an immediate consequence of the plancherel formula .
[ north ] for @xmath466 , regarding @xmath71 and @xmath467 as functions @xmath468 and @xmath469 we have that @xmath470 this follows by applying ( [ completenessfwd ] ) to the function @xmath471 .
the following spectral orthogonality does not seem to follow from the plancherel formula ( theorem [ kqbosonid ] ) .
it is , in fact , the key input in proving the dual plancherel formula ( theorem [ kqbosoniddual ] ) , so we provide an independent proof of it .
the spaces of functions with which we deal are more general than necessary for the dual plancherel formula .
[ specorth ] consider a function @xmath439 such that for @xmath15 large enough , @xmath472 is analytic in the closed exterior of @xmath60 , and consider another function @xmath473 such that @xmath474 is analytic in the closed region between @xmath60 and @xmath343 .
then we have that @xmath475 this is the @xmath460 case of proposition [ trueorththm ] . [ orthconj ] the above identity may be formally rewritten as @xmath476_{i , j=1}^{k}\ ] ] where @xmath477 is the dirac delta function for @xmath478 . owing to ( [ 23star ] ) , this can be further ( even more formally ) rewritten as @xmath479_{i , j=1}^{k } \frac{\prod_{i=1}^{k } \frac{1}{1-z_i } dz_i}{d\mu_{(1)^k}(\vec{z})}.\ ] ] it is natural to try to extend the above spectral orthogonality to all eigenfunctions which arise in the completeness results of corollary [ completenesscor ] .
this prompts the conjecture that for any two partitions @xmath480 , if @xmath481 and @xmath482 ( for @xmath445 of length @xmath83 and @xmath483 of length @xmath484 ) then @xmath485_{i , j=1}^{k},\ ] ] where the product @xmath486 means that we omit factors that are zero ( by virtue of the definitions of @xmath174 and @xmath82 ) .
presumably , the above formal identity should be understood in a similar manner as proposition [ specorth ] .
note that the above may be expressed via @xmath487 owing to the fact that @xmath488 in particular , for @xmath489 , we expect that @xmath490 note that in the determinant @xmath491_{i , j=1}^{2}$ ] , only one summand survives , namely , @xmath492 .
we have checked the above @xmath489 version of the conjecture for test functions of the form @xmath493 and @xmath494 . for such functions , any integration contours which enclose @xmath40 suffice ( since there is only a pole at 1 to consider ) .
in this section we show how the plancherel formula ( in particular the completeness result of corollary [ completenesscor ] ) provides means to solve the kolmogorov backward and forward equations for the @xmath0-boson particle system . in the case of the backward equation in section [ qtasepsec ]
we explain how , through a duality with another particle system called @xmath0-tasep , this enables us to calculate exact formulas for expectations of certain observables of @xmath0-tasep with general initial data ( thus extending special cases of initial data studied earlier in @xcite ) .
this should be thought as parallel to the work of dotsenko @xcite and calabrese , le doussal and rosso @xcite in which they use the eigenfunction decomposition of the delta bose gas to calculate exact formulas for joint moments of the solution of the stochastic heat equation ( cf .
section [ deltabosesec ] ) .
the kolmogorov backward and forward equations can be readily solved when the initial data is expressed as a sum over eigenfunctions .
the plancherel formula ( specifically the completeness result of corollary [ completenesscor ] ) provides such a decomposition . for the forward generator @xmath458 is the right eigenfunction , whereas for the backward generator @xmath85 is the right eigenfunction .
the following solutions of the backward and forward equations for general ( compactly supported ) initial data are consequences of corollary [ completenesscor ] .
[ bwdapp ] for any @xmath495 the backward equation @xmath496 with @xmath497 is uniquely solved by @xmath498 where @xmath499 and where the contours @xmath98 are as in definition [ kqtasepdef ] .
the uniqueness of solutions to the backward equation with initial data in @xmath35 is evident because of the triangular nature of the generator @xmath12 . from definition
[ leftrighteig ] and proposition [ prop211 ] , @xmath500 is a right eigenfunction for @xmath12 with eigenvalue @xmath501 .
this implies that the right - hand side of the first line of ( [ ftvecn ] ) solves the desired backward equation .
the fact that it satisfies the initial data follows from corollary [ completenesscor ] , equation ( [ completenessfwd ] ) .
the equivalence of the first and second lines of ( [ ftvecn ] ) is a consequence of proposition [ 321 ] ( see also lemma [ expandlem ] ) .
[ fwdapp ] for any @xmath495 the forward equation @xmath502 with @xmath497 is uniquely solved by @xmath503 where @xmath499 and the contours @xmath98 are as in definition [ kqtasepdef ] .
this is proved in the same manner as corollary [ bwdapp ] with the replacements of @xmath12 with @xmath20 and @xmath500 with @xmath504 .
the initial data is satisfied by virtue of corollary [ completenesscor ] , equation ( [ completenessbwd ] ) .
the kolmogorov forward equation governs the evolution of the transition probabilities of a markov process .
in particular , for the @xmath0-boson particle system @xmath505 ( recall from section [ stochrepsec ] ) define @xmath506 for @xmath507 and @xmath508 .
then @xmath509 solves the kolmogorov forward equation @xmath510 where @xmath20 ( definition [ bwdgendef ] ) acts in the @xmath308 variables
. we may directly utilize corollary [ fwdapp ] to show : [ transprobcor ] the @xmath0-boson particle system transition probability @xmath509 from initial state @xmath386 to terminal state @xmath384 in time @xmath508 is given by both the right - hand side of the first and second line of ( [ ftvecnfwd ] ) with @xmath511 .
this result may be compared to tracy - widom s asep transition probability formulas @xcite .
those formulas involve @xmath512 , @xmath52-fold contour integrals . with some additional work
, the above result ( which is a single @xmath52-fold nested contour integral ) can be brought closer to this form .
in fact , in work of korhonen - lee @xcite posted soon after the present work was first post , the authors utilize the tracy - widom approach to arrive directly at such a formulas as indicated above . in a companion paper @xcite we develop parallel results for asep and the heisenberg xxz quantum spin chain and remark on the relationship to tracy - widom s work on asep .
the kolmogorov backward equation governs the evolution of the expectation of functions of a markov process .
in particular , for the @xmath0-boson particle system @xmath505 and a function @xmath513 , we define @xmath514,\ ] ] which always stays well - defined ( at least for bounded functions @xmath515 ) .
then @xmath109 solves the kolmogorov backward equation @xmath516 where @xmath12 ( definition [ bwdgendef ] ) acts in the @xmath6 variables
. we may directly utilize corollary [ bwdapp ] to show : [ exactthm ] for @xmath495 , the expectation @xmath109 is given by the right - hand side of the first and the second line of ( [ ftvecn ] ) . in ( * ? ? ?
* section 2 ) it is shown that there exists a markov duality between the @xmath0-boson particle system and @xmath0-tasep . as a result ,
expectations of certain observables of @xmath0-tasep can be computed by solving the above backward equation .
the @xmath0-deformed totally asymmetric simple exclusion process ( @xmath0-tasep ) is a continuous time , discrete space interacting particle system @xmath86 .
it was first introduced and studied by borodin - corwin @xcite ( see also @xcite for subsequent developments ) .
particles occupy sites of @xmath2 with at most one particle at any site at a given time .
the location of particle @xmath193 at time @xmath8 is written as @xmath517 .
particles are ordered so that @xmath518 for @xmath519 .
the rate at which the value of @xmath517 increases by one ( i.e. the particle jumps right by one ) is @xmath520 ( here @xmath521 is the size of the gap between particle @xmath274 and @xmath522 at time @xmath8 ) .
all jumps occur independently of each other according to exponential clocks .
we will focus on @xmath0-tasep with only @xmath523 particles , though since particles only depend on those to their right ( i.e. , smaller index ) this restriction can be weakened to include systems with a right - most particle ( though possible infinite particles in total ) .
define @xmath524 as the set of all @xmath16 such that @xmath525 .
for some possibly random initial data @xmath89 for @xmath526 particle @xmath0-tasep define @xmath527 as @xmath99 outside of the compact set @xmath524 and otherwise @xmath528,\ ] ] where the expectation is over the possibly random initial data @xmath529 .
then it follows from proposition 2.7 ( and definition 2.6 ) of @xcite that , @xmath530\ ] ] solves the @xmath0-boson particle system backward equation with initial data @xmath531 . here
the expectation is over both the possibly random initial data and the random evolution of @xmath0-tasep .
in particular , this , along with corollary [ bwdapp ] implies the following .
[ abovecor ] for @xmath0-tasep with @xmath526 particles and initial data @xmath89 , @xmath109 is given by the right - hand side of the first and the second lines of ( [ ftvecn ] ) with @xmath532 . in order to apply corollary [ abovecor ]
it is necessary to compute @xmath533 . for purposes of asymptotics
it is desirable to concisely sum the series given by this bilinear pairing .
the dual plancherel formula given in theorem [ kqbosoniddual ] shows how this can be achieved for a certain class of functions .
[ abovecor2 ] consider any function @xmath67 , if @xmath534 , then @xmath535 .
the follows immediately from the application of theorem [ kqbosoniddual ] . to illustrate ,
let us apply this corollary to the two types of @xmath0-tasep initial data studied previously in @xcite and @xcite .
step initial data for @xmath0-tasep is when @xmath536 for @xmath537 .
half stationary initial data with parameter @xmath538 is when @xmath539 and , for @xmath540 , @xmath541 . here
the @xmath542 s are independent random variables with common distribution @xmath543 see ( * ? ? ?
* section 3.3 ) for a justification of why this is half - stationary initial data . when @xmath544 , half - stationary reduces to step initial data .
for step initial data it is immediate to see that the corresponding initial data @xmath531 is given by @xmath545 . due to the fact that the operator @xmath12 is triangular ,
the solution @xmath109 for @xmath546 does not depend on the initial data @xmath547 for @xmath6 such that @xmath548 . therefore in order to solve for @xmath109 we may just as well work with initial data @xmath549 although observe that this function is no longer compactly supported .
for half - stationary initial data , finding @xmath531 requires a calculation involving the above common distribution of the @xmath542 .
this is done in equation ( 12 ) of @xcite and gives ( after similar considerations involving the indicator functions ) the ( non - compactly supported ) initial data @xmath550 with the convention above that @xmath551 .
a particular instance of corollary [ abovecor2 ] ( in fact , an immediate extension in the function space to allow for non - compact support ) implies the following .
[ belowlemma ] consider any @xmath552 .
assume that there exist positively oriented closed contours @xmath39 and @xmath553 which both satisfy the conditions of definition [ kqtasepdef ] but also do not include @xmath554 and are such that for all @xmath555 and @xmath556 , @xmath557 .
then , for @xmath174 with @xmath361 for @xmath138 , @xmath558 define @xmath559 . we claim that @xmath560 , where the definition of @xmath66 is extended to this function @xmath366 ( which is not in @xmath36 ) by assuming that the contours @xmath39 have the additional condition that they do not include @xmath554 .
checking this is easily achieved through residue considerations .
indeed , when @xmath561 , the @xmath399 integral in @xmath562 has no residue at @xmath40 and hence evaluates to 0 , enforcing the condition @xmath563 , for @xmath138 .
otherwise , we find the desired equality by sequentially deforming the @xmath367 through @xmath44 contours to infinity and evaluating the integral @xmath562 via its residue at infinity .
if corollary [ abovecor2 ] were valid for @xmath366 as above , then the proof of the lemma would be complete .
though we have not proved that this is the case , we may still complete the proof via an approximation argument .
it is possible to find a sequence of functions @xmath564 so that as @xmath565 , @xmath566 converges uniformly over @xmath174 with @xmath361 for @xmath138 , to @xmath93 . in order to extend the result of corollary [ abovecor2 ]
, we must show that @xmath567 for all @xmath174 with @xmath361 for @xmath138 .
theorem [ kqbosoniddual ] implies that this holds for all @xmath568 .
consider the @xmath6 term arising in the summation when @xmath62 is applied to @xmath569 .
call @xmath82 the integration variables involved in the formula for @xmath569 .
we can choose the integration contours for @xmath66 to be @xmath570 .
in that case , the @xmath6 term we are considering can be uniformly ( in @xmath571 ) bounded by a constant times @xmath572 for some @xmath573 ( which is the maximal ratio of @xmath574 over @xmath361 , @xmath575 and @xmath138 ) .
since @xmath366 does not have a pole at @xmath40 , @xmath569 is supported on the non - negative @xmath6 .
the above bound along with uniformity in the convergence of @xmath576 to @xmath366 implies that we can take the @xmath565 limit inside the dual plancherel formula relation and conclude that @xmath567 as desired .
the method of the above proof shows one way to try to extend the plancherel and dual plancherel formulas to apply to larger classes of functions .
corollary [ abovecor2 ] ( in fact theorem [ kqbosoniddual ] ) immediately yields many combinatorial formulas like the one in lemma [ belowlemma ] above .
we provide , in addition to the above proof , a direct inductive proof of lemma [ belowlemma ] in section [ inductiveproof ] the particular identity of lemma [ belowlemma ] is akin to one which arises as ( * ? ? ?
* lemma 5 ) in considering the half - brownian kpz equation , as well as bares similarity to the identity in ( * ? ? ?
* section iii ) . in applying corollary [ abovecor2 ] ,
it is not a priori so clear if given @xmath393 , how to find @xmath366 such that @xmath534 . on strategy
is to try to explicitly compute @xmath577 for @xmath578 and then guess @xmath366 for general @xmath52 . in order to check such a guess ,
one need only compute the integrals in applying @xmath66 to @xmath366 ( a opposed to implementing an inductive proof ) .
there is a class of initial data for @xmath0-tasep which arises as marginal distributions of macdonald processes for which @xmath366 is clear ( the notation and concepts below are explained in detail in @xcite , see also @xcite ) .
consider @xmath0-tasep initialized according to @xmath579 where @xmath164 is a gelfand - tselin pattern distributed according to an ascending ( macdonald parameter @xmath105 and @xmath0 the same in the @xmath0-tasep ) macdonald process with parameters @xmath580 , and @xmath581 a macdonald non - negative specialization characterized by ( non - negative ) parameters @xmath582 , @xmath583 and @xmath60 .
then @xmath584 where the @xmath356 are macdonald first difference operators and @xmath585 is the normalizing constant for the macdonald process . these difference operators are naturally encoded via nested contour integrals , from which one readily identities @xmath586 where @xmath587 this implies that @xmath588 .
step initial data for @xmath0-tasep is a special case of a marginal of a macdonald process ( in which all @xmath589 ) .
half stationary should arise from a `` two - sided '' generalization of macdonald processes ( cf .
@xcite for the schur version ) which has yet to be fully developed .
flat or half - flat initial data for @xmath0-tasep do not seem likely to arise from macdonald processes .
however , as they have been ( non - rigorously ) treated at the kpz equation level , one might hope to explicitly sum the associated series @xmath90 and proceed to asymptotics for corresponding @xmath0-tasep initial data .
[ dualiyresult ] for step initial data , @xmath590 = ( -1)^k q^{\frac{k(k-1)}{2 } } \oint_{\gamma_1 } \frac{dz_1}{2\pi
{ \mathbf i}}\cdots \oint_{\gamma_k } \frac{dz_k}{2\pi { \mathbf i } } \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \prod_{j=1}^{k } ( 1-z_j)^{-n_j } \frac{e^{(q-1)tz_j}}{z_j},\ ] ] where @xmath98 are as in definition [ kqtasepdef ] , with the additional condition that they do not include @xmath99 .
.1 in for half - stationary initial data with parameter @xmath591 such that @xmath592 , @xmath593 = ( -1)^k q^{\frac{k(k-1)}{2 } } \oint_{\gamma_1 } \frac{dz_1}{2\pi { \mathbf i}}\cdots \oint_{\gamma_k}\frac{dz_k}{2\pi { \mathbf i}}\prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - qz_b } \prod_{j=1}^{k } ( 1-z_j)^{-n_j } \frac{e^{(q-1)tz_j}}{z_j-\alpha / q},\ ] ] where @xmath98 are as in definition [ kqtasepdef ] with the additional condition that they do not include @xmath554 ( this is possible due to the restriction on the value of @xmath594 ) .
it suffices to prove the half - stationary result .
assume that @xmath39 are such that there exist @xmath553 so that the hypotheses of lemma [ belowlemma ] are satisfied .
then , applying lemma [ belowlemma ] and corollary [ abovecor ] we arrive at the desired result .
observe that the contours in ( [ fdefqtasephalf ] ) can be freely deformed now to any choice of @xmath39 as in the statement of the proposition .
the step initial data version of this result with all @xmath100 was proved in ( * ? ? ?
* section 3.3 ) using the theory of macdonald processes .
general @xmath6 step and half - stationary result was proved in ( * ? ? ?
* theorem 2.11 ) .
the macdonald process approach was then extended in @xcite to also cover general @xmath6 .
though we have assumed @xmath101 in this paper , many of the results can either be immediately extended ( as well as their proofs ) or easily modified to accommodate general @xmath595 where we define @xmath596 from the perspective of stochastic processes ( like the @xmath0-boson particle system or @xmath0-tasep ) , having @xmath0 outside @xmath597 does not seem to be meaningful .
however , from the perspective of quantum mechanics such an extension is natural .
in fact , a version of the @xmath0-boson hamiltonian ( corresponding with the @xmath598 limit discussed in section [ vandiejen ] ) with @xmath0 on the complex unit circle has been used as an integrable regularization of the real time delta bose gas in the study of quantum quenches ( cf . @xcite and references therein ) . taking @xmath0 on the complex unit circle is akin to taking the ratio @xmath599 in asep to be on the complex unit circle or taking the coupling constant @xmath600 in the heisenberg xxz quantum spin chain ( on @xmath2 ) to be real with @xmath601 .
the plancherel formulas and completeness / biorthogonality for asep and xxz will be the subject of future work @xcite ( see also @xcite ) . in section [ qbosonsyst ] , though the probabilistic interpretation for the backward and forward generators is not valid when @xmath0 is general , the equivalence of the generators to free generators with @xmath31 boundary conditions , and the bethe ansatz eigenfunctions all stay valid . in short ,
all results , definitions and formulas hold for general @xmath595 . in section [ plansec ] ,
the definitions of @xmath62 , @xmath66 and @xmath246 remain valid for general @xmath0 .
recall the conditions on the contours @xmath98 that they be positively oriented , closed contours chosen so that they all contain @xmath40 , so that the @xmath41 contour contains the image of @xmath0 times the @xmath42 contour for all @xmath43 , and so that @xmath44 is a small enough circle around 1 that does not contain @xmath0 . for general @xmath595 , these conditions lead to a rather different picture for the contours than that of figure [ circontours ] . for instance , for @xmath0 on the complex unit circle ( but not a @xmath602 root of unity for @xmath603 ) , figure [ unitcircontours ]
shows contours which satisfy the conditions .
contours when @xmath604 and @xmath0 is on the unit complex circle ( approximately @xmath605 ) .
the contour @xmath606 is a small circle around @xmath40 .
the contour @xmath607 is the union of the small circle around @xmath40 and the slightly larger circle around @xmath0 ( so that it contains the image of @xmath606 when multiplied by @xmath0 ) .
the contour @xmath403 is the union of the small circle around @xmath40 , the slightly larger circle around @xmath0 and the yet larger circle around @xmath608 ( so that it contains the image of @xmath607 and @xmath606 when multiplied by @xmath0 ) . because the image of @xmath403 when multiplied by @xmath0 intersects the original small circle around @xmath40 ,
when defining @xmath367 we take it to be the union of the larger ellipse around @xmath40 and @xmath609 , as well as the circles around @xmath0 and @xmath608 . ] for what follows , fix @xmath44 as a circle around 1 , small enough so as not to contain @xmath0 . for general @xmath595 ,
it is possible that the image of @xmath44 when multiplied by a power of @xmath0 will intersect itself .
this will affect the way that a nested contour integral ( such as in @xmath66 ) expands when contours are deformed to @xmath44 .
indeed , certain residues previously encountered may not contribute . with this in mind
, define the contour ( see figure [ overlap ] ) @xmath610 these contours are similar in nature ( and origin , see @xcite ) to the so - called chebyshev circles considered in @xcite in the context of the xxz model s plancherel formula for @xmath601 . , @xmath605 and a fixed circle @xmath606 around @xmath40 , the contours @xmath611 and @xmath612 are all equal to @xmath606 .
however , @xmath613 is only the portion of @xmath606 which is the complement of the dark black arc .
this arc is the set of @xmath278 such that @xmath614 lies inside of @xmath606 ( grey arc in the figure ) . ]
we may now state the general @xmath0 version of lemma [ expandlem ] .
[ expandlemgeneral ] fix @xmath595 and consider a symmetric function @xmath37 and positively oriented , closed contours @xmath98 such that : * the contour @xmath44 is a circle around 1 , small enough so as not to contain @xmath0 ; * for all @xmath354 , the interior of @xmath41 contains @xmath40 and the image of @xmath42 multiplied by @xmath0 ; * for all @xmath355 , there exist deformations @xmath356 of @xmath357 to @xmath44 so that for all @xmath358 with @xmath361 for @xmath360 , and @xmath370 for @xmath362 , the function @xmath363 is analytic in a neighborhood of the area swept out by the deformation @xmath356 .
then , @xmath615 this follows immediately from a general @xmath0 version of proposition [ 321 ] , using the symmetry of @xmath366 to remove it from the expression for @xmath616 .
the general @xmath0 version of proposition [ 321 ] has essentially the same proof .
let us illustrate the only modification which is required
. as we deformed @xmath405 through @xmath367 to @xmath44 , we picked strings of residues of the form @xmath617 where @xmath618 .
however , if @xmath619 is contained inside of @xmath44 for some @xmath620 , then the deformation will not cross one of the poles which corresponds with this residue string .
thus , the string will be excluded from the residue expansion .
this is the source of the restriction to the @xmath621 contours in the above result .
when @xmath0 is a root of unity more care must be taken due to the fact that @xmath622 and @xmath278 may lie upon the same contour . in the course of proving the above result we encounter denominators which contain terms like @xmath623 .
thus , some additional regularization is needed to deal with this case .
we do not pursue this here .
just as with ( [ kexpform ] ) , we may use the above lemma to record the analogous general @xmath0 formula for @xmath246 : @xmath624 theorem [ kqbosonid ] holds exactly as stated with @xmath62 and @xmath66 as above , and @xmath595 .
this in turn shows that for general @xmath595 , corollaries [ completenesscor ] and [ north ] hold under the one replacement : @xmath625 while we expect that suitable modifications are needed for the spectral orthogonality ( proposition [ specorth ] ) , and consequently the dual plancherel formula ( theorem [ kqbosoniddual ] ) and the plancherel isomorphism ( theorem [ isothm ] ) , we do not pursue this here .
this section deals with two different limits of the results contained in sections [ qbosonsyst ] , [ plansec ] and [ appsec ] . for a parameter @xmath626 define the operator @xmath627 that acts on functions @xmath13 as @xmath628 the first limit ( sections [ epdef ] and [ vandiejen ] ) involves keeping @xmath0 fixed but conjugating @xmath12 by the operator @xmath627 , and then taking @xmath629 .
the limiting system is equivalent to the delta bose gas considered by van diejen @xcite ( for root systems of type @xmath630 ) .
it is also related to the @xmath123 version of the @xmath0-boson hamiltonian discussed earlier in section [ afortiori ] . for this limiting system
we show how the spectral orthogonality of the left and right eigenfunctions follows from the cauchy - littlewood identity for hall - littlewood symmetric polynomials .
this @xmath598 orthogonality is , in fact , the basis for a general @xmath631 orthogonality which is proved in proposition [ trueorththm ] and implies proposition [ specorth ] when @xmath460 .
the second limit ( section [ semidiscsec ] ) involves taking @xmath632 and also performing an @xmath633-dependent conjugation of @xmath12 ( involving @xmath627 as well ) .
the limiting system here is equivalent to the discrete delta bose gas considered in ( * ? ? ?
* section 6 ) and ( * ? ? ?
* section 6 ) .
this system arises naturally from studying moments of a probabilistic system called the semi - discrete stochastic heat equation or equivalently the oconnell - yor semi - discrete directed polymer partition function ( see section [ sdapps ] ) .
there is yet a third limit to the continuum delta bose gas .
this is briefly discussed in section [ deltabosesec ] .
all other results of this section with the exception of propositions [ trueorththm ] , [ epzeroorth ] and the lemmas involved in their proofs can either be proved via suitable limits of our earlier @xmath0-boson particle system results , or can be proved directly for the limiting system ( via the same methods as our earlier results ) . as such ,
we do no include these proofs . in what follows
we recall notation from section [ notations ] and write @xmath210 , @xmath211 , and @xmath212 ( thus suppressing the @xmath6 dependence ) .
we now develop an @xmath633-deformation of the work of sections [ qbosonsyst ] , [ plansec ] and [ appsec ] ( which at @xmath460 corresponds to these earlier results ) .
the main purpose of doing this is to provide a proof of the spectral orthogonality result of proposition [ specorth ] . in section [ orthsecz ]
we prove that if the spectral orthogonality holds true for some @xmath633 , then it holds for all @xmath633 ( by essentially showing the derivative of the relation in @xmath633 is zero ) .
it turns out ( cf . section [ vandiejenorth ] ) that the @xmath629 limit of this @xmath633-deformation is implied by the cauchy - littlewood identity for hall - littlewood symmetric polynomials .
this limit also makes contact with earlier work of van diejen @xcite .
the general @xmath633 deformation corresponds ( via setting @xmath125 ) to the @xmath0-boson hamiltonian discussed in section [ afortiori ] .
fix @xmath101 and recall the operator @xmath627 from ( [ me ] ) . for @xmath626 define the operator @xmath634 by@xmath635 define @xmath636 ( the matrix transpose of @xmath634 ) as @xmath637 and @xmath638 .
it is straightforward to see that @xmath639 define an @xmath633-deformed backward difference operator @xmath640 and forward difference operator @xmath641 which act on functions @xmath135 as @xmath642 define the @xmath633-deformed @xmath0-boson backward free generator @xmath643 in the same way as in ( [ freebwddef ] ) , with @xmath133 replaced by @xmath640 .
likewise define the @xmath31 @xmath633-deformed @xmath0-boson backward two - body boundary conditions in the same way as in ( [ star1 ] ) , with @xmath133 replaced by @xmath640 .
similarly define the @xmath633-deformed @xmath0-boson forward free generator @xmath643 and the @xmath31 @xmath633-deformed @xmath0-boson forward two - body boundary conditions .
the result of proposition [ freetrueequiv ] then holds with all terms replaced by their @xmath633-deformations .
our notation of an eigenfunction remains as in definition [ eigdeffree ] and the result of corollary [ eigcorgen ] applies under this @xmath633-deformation .
using the coordinate bethe ansatz ( cf .
section [ corbetherev ] ) we construct the below eigenfunctions .
for all @xmath644 , set @xmath645 for fixed @xmath16 , these are symmetric laurent polynomials in the variables @xmath646 .
the result of proposition [ prop211 ] then holds with all terms replaced by their @xmath633-deformations , and the same eigenvalues @xmath315 as before .
we define the right and left eigenfunctions for the forward generator @xmath636 as @xmath647 the plancherel formulas proved in section [ plansec ] are readily adapted to this @xmath633-deformation . in order to facilitate this , we must suitably modify definition [ kqtasepdef ] .
[ kqtasepdefe ] define the space @xmath648 of symmetric laurent polynomials in @xmath646 .
fix any set of positively oriented , closed contours @xmath649 chosen so that they all contain @xmath633 , so that the @xmath650 contour contains the image of @xmath0 times the @xmath651 contour for all @xmath43 , and so that @xmath652 is a small enough circle around @xmath633 so as not to contain @xmath653 .
for use later , let us also fix contours @xmath654 and @xmath655 where @xmath654 is a positively oriented closed contour which contains @xmath633 and its own image under multiplication by @xmath0 , and @xmath655 contains @xmath654 and is such that for all @xmath656 and @xmath657 , @xmath658 .
we define @xmath659 and @xmath660 ( as well as their compositions ) in the same manner as @xmath62 and @xmath66 ( see definition [ kqtasepdef ] ) by replacing @xmath661 by their @xmath633-deformations , and replacing @xmath662 by @xmath663 for @xmath138 .
the plancherel formula ( theorem [ kqbosonid ] ) , the dual plancherel formula ( theorem [ kqbosoniddual ] ) and the plancherel isomorphism ( theorem [ isothm ] ) all hold under the above @xmath633-deformation . defining an @xmath633-deformed @xmath0-pochhammer symbol @xmath664
, we likewise find that corollary [ completenesscor ] holds with @xmath665 also replaced by @xmath666 .
modifying in the same manner definition [ bilinearprime ] , it is immediate that corollary [ north ] continues to hold as well .
we turn here to the @xmath633-deformation of the spectral orthogonality .
we will provide a direct proof of this result here .
[ trueorththm ] consider a function @xmath439 such that for @xmath15 large enough , @xmath667 is analytic in the closed exterior of @xmath654 , and consider another function @xmath473 such that @xmath474 is analytic in the closed region between @xmath654 and @xmath655 .
then we have that @xmath668 the above identity may be formally rewritten as @xmath669_{i , j=1}^{k}.\ ] ] where @xmath477 is the dirac delta function for @xmath478 .
we start with the simple example of @xmath670 .
the general @xmath52 proof is more involved and is given after this example .
when @xmath670 , @xmath671 and @xmath672 .
thus , the above orthogonality identity states that @xmath673 consider the left - hand side of ( [ pgfouronestar ] ) .
since for @xmath15 large enough , @xmath674 is analytic outside @xmath654 if @xmath675 then there is no residue at infinity and the integral in @xmath278 vanishes .
therefore the summation can be restricted to @xmath676 . by cauchy s theorem
we can deform the contour of integration for @xmath677 so as to lie on the larger contour @xmath655 , so that @xmath678 for all @xmath656 and @xmath657 .
the summation and integration can now be interchanged since @xmath679 is uniformly absolutely convergent . evaluating this sum shows that the left - hand side of ( [ pgfouronestar ] ) equals @xmath680 we can apply the residue theorem to evaluate the @xmath278 integral . due to the conditions on @xmath681 , outside
@xmath654 there is only a residue at @xmath478 ( and no singularity at infinity ) , hence we immediately arrive at the right - hand side of ( [ pgfouronestar ] ) .
this method of proof does not appear so well adapted to @xmath682 .
let us introduce ( again for @xmath670 ) the method which we will employ for the general @xmath682 case .
in what follows we will work formally , though all manipulations are easily justified by integrating against test functions .
rewrite the orthogonality identity ( formally ) as @xmath683 so that now the left - hand side depends on @xmath633 , whereas the right - hand side does not . here
we have called the above left - hand side @xmath684 .
it suffices then to show that @xmath685 and that @xmath686 for all @xmath631 .
for @xmath670 the proof that @xmath685 is just as above ( see section [ vandiejenorth ] for the @xmath682 version of this result ) .
as far as the derivative , due to telescoping @xmath687 the general @xmath682 proof splits into two steps .
the identity we wish to prove can be rewritten formally as @xmath688_{i , j=1}^{k},\ ] ] where we define @xmath689 and @xmath690 .
here we have called the left - hand side @xmath691 .
the purpose of rewriting the desired identity as above is that @xmath633 only arises on the left - hand side , and not on the right .
the first step is to prove ( [ sixeight ] ) for @xmath598 .
this is proved as proposition [ epzeroorth ] and follows by identifying @xmath692 and @xmath693 with hall - littlewood polynomials ( [ pgstarstar ] ) and ( [ pgstarstarstar ] ) and using the cauchy - littlewood identity ( [ pgfoursevenstarstar ] ) . the second step is to prove that @xmath694 . by the leibnitz rule the derivative either applies to @xmath695 or @xmath696 .
let us calculate how each of these terms transforms under differentiation .
we will show that the derivatives can be written as @xmath697 where @xmath698 and @xmath699 are matrices independent of @xmath633 ( and @xmath174 and @xmath82 ) such that for each fixed @xmath6 , there are only finitely many @xmath700 such that the @xmath701 entry is non - zero . assuming the above decomposition , we can rewrite @xmath702 to show that the above is identically zero , it suffices to show that for all @xmath703 @xmath704 therefore , to complete this proof we must first justify the decomposition ( [ decomp ] ) and then check that the matrices which arise from that decomposition satisfy the relation ( [ crl ] ) .
we will explicitly compute the entries of the matrix @xmath698 ( and likewise those of @xmath699 ) .
observe that @xmath705 where the term @xmath706 is excluded from the product . for @xmath6 such that @xmath707 for all @xmath708 ( i.e. , @xmath6 away from the boundary of @xmath329 )
the above expansion can be easily reexpressed through @xmath709 with coefficients @xmath710 when there are clusters in @xmath6 of equal values , @xmath711 will lie outside of @xmath329 for some @xmath193 .
in order to express the derivative in terms of @xmath709 for @xmath700 inside @xmath329 we will use the following property . for @xmath16 and two functions @xmath712 , we write @xmath713 if @xmath714 [ shiftlem ] if @xmath715 then for any constants @xmath716 and any function @xmath717 symmetric in the @xmath193 and @xmath301 entries , we have that @xmath718 consider the summand ( for a given @xmath297 ) corresponding to the difference in ( [ diffeq ] ) between the left and right sides of the above identity .
all that differs between these two sides of the identity is the terms @xmath719 and @xmath720 .
taking the difference thus yields the factor @xmath721 .
combined with the term @xmath722 coming from the product over @xmath723 , this is symmetric in @xmath724 and @xmath725 . since @xmath715 , and @xmath726 is symmetric in @xmath724 and @xmath725
, it follows that everything except for the denominator of the product over @xmath723 is symmetric in @xmath724 and @xmath725 .
the denominator is antisymmetric , thus , we conclude that the difference between the left and right sides of the desired identity is equal to the symmetrization ( summation in @xmath298 ) of a summand which is antisymmetric in two of its variables .
this implies the symmetrization is zero , and thus so is the difference .
[ cepslem ] consider @xmath727 , then @xmath728 where @xmath729 we will proceed by induction on @xmath52 . for @xmath670 , the decomposition is clearly true with @xmath730 .
fix that for all @xmath52 , @xmath731 .
assume the inductive hypothesis holds for @xmath732 .
then @xmath733 the equality in the first line is from the inductive hypothesis ; the equality between the first and second line is from a repeated application of lemma [ shiftlem ] ; the equality between the second and third line is from rewriting @xmath734 ; and the equality between the third and fourth line is from gathering terms with the same factor of @xmath735 , and using the notational assumption @xmath731 .
the above reasoning shows that the @xmath736 are the solutions to the following recurrence relation : @xmath737 with @xmath731 and @xmath730 .
it is immediate to check that the right - hand side of ( [ checked ] ) uniquely satisfies this relation .
these @xmath736 coefficients combine to form the entries of @xmath738 .
define @xmath739 then it is immediate from lemma [ cepslem ] that for @xmath727 @xmath740 this shows that for @xmath727 , @xmath741 for all @xmath742 with exactly @xmath743 entries equal to @xmath744 ( and the remaining equal to @xmath745 ) . for all other @xmath700 , @xmath746 . to state the general @xmath6 result ,
define @xmath747}^{\pm } = ( n_1,\ldots , n_{a-1 } , n_{a}\pm 1,n_{a+1}\pm 1,\ldots , n_{b}\pm 1,n_{b+1},\ldots , n_{k}).\ ] ] [ rlem ] for @xmath16 , set @xmath748 ( cf .
section [ notations ] ) .
then for all @xmath749 and @xmath750 , if @xmath751}^{-}$ ] then @xmath752 and for all other @xmath700 , @xmath753 .
this same reasoning as was applied in the @xmath754 case may be applied to each cluster of @xmath6 , thus resulting in the statement of the lemma .
we likewise find that [ llem ] for @xmath16 , set @xmath748 ( cf .
section [ notations ] ) .
then for all @xmath749 and @xmath755 , if @xmath756}^{+}$ ] then @xmath757 and for all other @xmath700 , @xmath758 .
this is completely analogous to the @xmath738 case .
note that the negative sign comes from differentiation of terms of the type @xmath759 .
the matrices defined in lemmas [ rlem ] and [ llem ] satisfy ( [ crl ] ) .
let us consider the example of @xmath760 with @xmath761 entries equal to @xmath744 and @xmath762 entries equal to @xmath745 ( thus a total length of @xmath763 ) .
then , the only @xmath700 for which @xmath764 and @xmath765 are non - zero is @xmath766 with @xmath767 entries of @xmath745 and @xmath768 entries of @xmath769 . for these @xmath6 and
@xmath700 we compute @xmath770 from the explicit formula ( [ dexp ] ) for the @xmath771 we immediately confirm the ( [ crl ] ) is satisfied .
the general @xmath16 case follows similarly from the above calculation .
we return to complete the proof of proposition [ trueorththm ] . as we have now justified the decomposition ( [ decomp ] ) and checked that the matrices @xmath698 and @xmath699 satisfy the relation ( [ crl ] )
, it follows that @xmath772 .
since proposition [ epzeroorth ] ( yet to be proved below ) shows that ( [ sixeight ] ) holds in the limit @xmath629 , these two facts together complete the proof of this proposition .
we consider the @xmath629 limit of the @xmath633-deformed @xmath0-boson particle system considered in section [ epdef ] .
our primary purpose is to prove proposition [ trueorththm ] for @xmath598 , as well as to make contact with earlier work of van diejen @xcite .
we also briefly remark on the limiting versions of the plancherel formula .
the @xmath629 limits of the generators @xmath634 , @xmath773 and @xmath636 are straightforward ( set @xmath598 ) , and likewise for @xmath774 and the boundary conditions .
in particular the limits of the backward and forward generators ( @xmath643 and @xmath775 ) applied to a function @xmath776 give @xmath777 and @xmath778 ( respectively ) .
the backward two - body boundary conditions limit to : @xmath779 and the forward two - body boundary conditions limit to : @xmath780 this limiting system arose earlier in work of van diejen @xcite , where he additionally worked in the context of general root systems ( the above system corresponds with type @xmath630 ) . in that work
, van diejen uses bethe ansatz to construct eigenfunctions for these systems and proves a plancherel formula ( for general root systems ) by appealing to earlier work of macdonald @xcite . taking the @xmath629 limit of our results also yields these eigenfunctions and such a plancherel formula ( for the type @xmath630 case ) .
let us take the @xmath629 limit of the left and right eigenfunctions of @xmath636 and call these @xmath693 and @xmath692 . for all @xmath781 ,
we then have @xmath782 we identify these with hall - littlewood symmetric polynomials @xmath783 and @xmath784 ( * ? ? ?
* chapter iii ) : for @xmath6 such that @xmath785 , @xmath786 while for @xmath6 such that @xmath787 , @xmath788 the @xmath629 limits of @xmath659 and @xmath660 ( as well as their compositions ) may thus be expressed in terms of hall - littlewood symmetric polynomials . in taking this @xmath629 limit of @xmath660
, there is no clear way to take the limit of the nested contours @xmath789 .
however , we may just as well take all contours to be @xmath654 and then there is a clear limit of the integrand and the contour .
indeed , at @xmath598 , we may use any contour @xmath790 which encloses @xmath99 and its own image under multiplication by @xmath0 . in this way we can show @xmath629 limits of the plancherel formula ( theorem [ kqbosonid ] ) , the dual plancherel formula ( theorem [ kqbosoniddual ] ) and the plancherel isomorphism ( theorem [ isothm ] ) .
we state and prove the @xmath598 case of proposition [ trueorththm ] by directly appealing to the cauchy - littlewood identity for hall - littlewood symmetric polynomials .
this result then serves as the first step in proving the general @xmath633 result of proposition [ trueorththm ] , and consequently proposition [ specorth ] as well .
[ epzeroorth ] fix a positively oriented circle @xmath791 which contains 0 and the image of @xmath790 under multiplication by @xmath0 , and let @xmath792 be a circle which contains @xmath790 and is such that @xmath793 for all @xmath794 and @xmath795 .
consider a function @xmath439 such that for @xmath15 large enough , @xmath796 is analytic in the closed exterior of @xmath790 , and consider another function @xmath473 such that @xmath474 is analytic in the closed region between @xmath790 and @xmath792 .
then we have that @xmath797 consider a single term in the left - hand side of ( [ epzeroortheqn ] ) with fixed @xmath16 . due to the analyticity condition on @xmath439 , if @xmath798 then the corresponding term integrates to zero ( this follows from the lack of singularity at infinity ) .
thus we can restrict the summation to @xmath799 .
now we can deform the integration contour for @xmath677 from @xmath790 to @xmath792 and we may appeal to the hall - littlewood polynomial representations for @xmath800 and @xmath801 given in ( [ pgstarstar ] ) and ( [ pgstarstarstar ] ) . using homogeneity of the hall - littlewood polynomials , and taking the integration in @xmath278 and @xmath677 outside of the summation , we are left with ( for @xmath15 sufficiently positive ) the left - hand side of ( [ epzeroortheqn ] ) equal to @xmath802 the fact that the summation and the two integrations may be interchanged follows from the fact that for @xmath803 and @xmath804 the summation is uniformly absolutely convergent . by residue considerations
we may change the summation over @xmath805 above to @xmath806 without changing the value of the integrals .
we now utilize the cauchy - littlewood identity for hall - littlewood polynomials ( * ? ? ?
* iii ( 4.4 ) ) to explicitly evaluate the summation @xmath807 therefore , the left - hand side of ( [ epzeroortheqn ] ) is equal to @xmath808 we now apply the residue theorem to evaluate the @xmath278 integrals . for each @xmath809
there are residues outside of @xmath790 at @xmath810 ( there are no singularities at @xmath811 ) .
however , due to the term @xmath812 , no two of these @xmath278 variables can pick the same @xmath677 variable ( otherwise the vandermonde gives zero ) .
thus , the @xmath278 integrals expand as a sum over residues from @xmath813 for @xmath138 and @xmath297 : @xmath814 the integral of this expression times @xmath815 is readily matched to the integral of the right - hand side of ( [ epzeroortheqn ] ) .
let @xmath816 and recall the operator @xmath627 from ( [ me ] ) .
define the operator @xmath817 by @xmath818 then , as @xmath629 , @xmath819 where we call this limit the semi - discrete backward generator ( because of the connection with the semi - discrete stochastic heat equation , see below ) .
below we record ( without proof ) scaling limits ( under the above scaling ) of various results we have proved earlier in sections [ qbosonsyst ] , [ plansec ] and [ appsec ] .
the semi - discrete backward generator @xmath820 acts on functions @xmath13 as @xmath821 the semi - discrete forward generator is the matrix transpose of @xmath820 and acts on functions @xmath822 as @xmath823 where by convention , for @xmath227 we set @xmath228 .
the function @xmath824 depends only on the list @xmath229 of cluster sizes for @xmath6 via @xmath825 we will use the notation of @xmath826 and @xmath827 for multiplication operators by @xmath828 and @xmath829 .
it is clear that these operators commute with @xmath142 ( see section [ notations ] ) .
the semi - discrete conjugated forward generator is defined by @xmath830 and it acts as @xmath831 using the backward and forward difference operators ( section [ notations ] ) we define the semi - discrete backward free generator @xmath832 which acts on functions @xmath251 as @xmath833 where @xmath139 acts as @xmath133 in the variable @xmath141 .
we say that the function @xmath251 satisfies the @xmath31 semi - discrete backward two - body boundary conditions if @xmath834 similarly , the semi - discrete forward free generator @xmath835 acts on functions @xmath251 as @xmath836 where @xmath140 acts as @xmath134 in the variable @xmath141 .
we say that the function @xmath251 satisfies the @xmath31 semi - discrete forward two - body boundary conditions if @xmath837 [ freetrueequivsd ] if @xmath251 satisfies the @xmath31 semi - discrete backward ( respectively , forward ) two - body boundary conditions , then for @xmath16 , @xmath838 [ commremsd ] proposition [ freetrueequivsd ] implies that @xmath839 or equivalently @xmath840 showing that @xmath820 and @xmath841 are related via a similarity transform .
[ eigdeffreesd ] a function @xmath842 is an eigenfunction of the semi - discrete backward free generator with @xmath31 two - body boundary conditions if @xmath843 is an eigenfunction for the semi - discrete backward free generator that satisfies the @xmath31 semi - discrete backward two - body boundary conditions .
we likewise define what it means for a function @xmath844 to be an eigenfunction of the semi - discrete forward free generator with @xmath31 two - body boundary conditions .
the following is a corollary of proposition [ freetrueequivsd ] .
[ eigcorgensd ] any eigenfunction @xmath842 for the semi - discrete backward free generator with @xmath31 two - body boundary conditions is , when restricted to @xmath16 , an eigenfunction for the semi - discrete backward generator @xmath820 with the same eigenvalue .
similarly , any eigenfunction @xmath844 for the semi - discrete forward free evolution equation with @xmath31 two - body boundary conditions is , when restricted to @xmath16 , an eigenfunction for the semi - discrete conjugated forward generator @xmath845 with the same eigenvalue . in turn
, @xmath846 is an eigenfunction for the semi - discrete forward generator @xmath841 with the same eigenvalue .
we may apply the coordinate bethe ansatz of section [ corbetherev ] to the semi - discrete backward and forward free generators with @xmath31 two - body boundary conditions .
the backward generator eigenfunction @xmath847 given below arose in earlier work of @xcite .
[ eigdefnsd ] for all @xmath781 , set @xmath848 these are symmetric laurent polynomials in the variables @xmath849 .
[ prop211sd ] for all @xmath781 , @xmath847 is an eigenfunction for the semi - discrete backward free generator with @xmath31 two - body boundary conditions with eigenvalue @xmath850 .
the restriction of @xmath847 to @xmath16 is consequently an eigenfunction for the semi - discrete backward generator @xmath820 with the same eigenvalue .
similarly , for all @xmath781 , @xmath851 is an eigenfunction for the semi - discrete forward free generator with @xmath31 two - body boundary conditions with eigenvalue @xmath850 .
the restriction of @xmath851 to @xmath16 is consequently an eigenfunction for the semi - discrete conjugated forward generator @xmath845 with the same eigenvalue .
the restriction of @xmath852 to @xmath16 is likewise an eigenfunction for the semi - discrete forward generator @xmath841 with the same eigenvalue .
[ bosonhamiltoniansymsd ] we may extend the eigenfunctions of definition [ eigdefnsd ] so as to be defined for all of @xmath328 ( rather than @xmath329 ) by fixing that the value for a general @xmath131 is the same as the value of @xmath330 where @xmath297 is a permutation of the elements of @xmath6 taking it into @xmath329 .
it is possible to write down an operator on all of @xmath328 for which these extensions ( which we write with the same notation ) are still eigenfunctions .
one finds ( cf .
* proposition 6.3 , ( c ) ) ) that @xmath853 { \ensuremath{\tilde{\psi}^{\textrm{bwd}}}}_{\vec{z}}(\vec{n } ) = \left(\sum_{i=1}^{k } ( z_i-1)\right)\ , { \ensuremath{\tilde{\psi}^{\textrm{bwd}}}}_{\vec{z}}(\vec{n}).\ ] ] this can be thought of as a discrete version of the delta bose gas considered later in section [ deltabosesec ] .
proposition [ prop211sd ] along with the fact that @xmath841 is the transpose of @xmath820 implies that @xmath854 showing the @xmath852 and @xmath847 are ( respectively ) right and left eigenfunctions for @xmath841 with eigenvalue @xmath850 .
this motivates the following .
[ leftrighteigsd ] for any @xmath855 define @xmath856 likewise define @xmath857 these eigenfunctions are limits of @xmath858 and @xmath859 of definition [ leftrighteig ] when @xmath632 and the @xmath860 variables in definition [ leftrighteig ] are replaced by @xmath861 .
we could have just as well defined the right and left eigenfunctions with respect to the operator @xmath12 .
finally , observe the symmetry of @xmath862 and @xmath863 with respect to the space - reflection operator @xmath142 @xmath864 [ kqtasepdefsd ] define the space @xmath865 of symmetric laurent polynomials in the variables @xmath849 .
the semi - discrete transform @xmath866 takes functions @xmath78 into functions @xmath867 via @xmath868 fix any set of positively oriented , closed contours @xmath869 chosen so that they all contain @xmath99 , so that the @xmath870 contour contains the image of @xmath40 plus the @xmath871 contour for all @xmath43 , and so that @xmath872 is a small enough circle around 0 so as not to contain @xmath40 . the ( candidate )
semi - discrete inverse transform @xmath873 takes functions @xmath874 into functions @xmath875 via @xmath876
the composition of the transform and ( candidate ) inverse transform takes functions @xmath78 into functions @xmath877 via @xmath878
the composition of the ( candidate ) inverse transform and the transform takes functions @xmath879 into functions @xmath880 via @xmath881 [ expandlemsd ] consider a symmetric function @xmath882 and positively oriented , closed contours @xmath869 such that : * the contour @xmath872 is a circle around 0 , small enough so as not to contain @xmath40 ; * for all @xmath354 , the interior of @xmath870 contains @xmath99 and the image of @xmath871 plus 1 ; * for all @xmath355 , there exist deformations @xmath356 of @xmath883 to @xmath872 so that for all @xmath358 with @xmath884 for @xmath360 , and @xmath885 for @xmath362 , the function @xmath886 is analytic in a neighborhood of the area swept out by the deformation @xmath356 .
then , @xmath887 where the definition of the notation @xmath888 and @xmath889 is given in section [ notations ] and where , for a partition @xmath169 , @xmath890_{i , j=1}^{\ell(\lambda ) } \prod_{j=1}^{\ell(\lambda ) } \frac{dw_j}{2\pi { \mathbf i}}.\ ] ] we may now record the scaling limits of the plancherel formula ( theorem [ kqbosonid ] ) , the dual plancherel formula ( theorem [ kqbosoniddual ] ) and the plancherel isomorphism ( theorem [ isothm ] ) .
we start with the plancherel formula .
[ kqbosonidsd ] the semi - discrete transform @xmath866 induces an isomorphism between the space @xmath35 and its image with inverse given by @xmath873 .
equivalently , @xmath891 acts as the identity operator on @xmath35 .
a key fact in the proof of this ( which follows in the exact same manner as theorem [ kqbosonid ] ) is the the following property of @xmath891 : for any functions @xmath46 , @xmath892 lemma [ expandlemsd ] yields the following expansion @xmath893 we have shown in theorem [ kqbosonidsd ] that @xmath378 when restricted to functions in @xmath894 .
similarly , @xmath895 let us turn now to the dual plancherel formula .
[ kqbosoniddualsd ] the inverse semi - discrete transform @xmath873 induces an isomorphism between the domain @xmath865 and its image with inverse given by @xmath866 .
equivalently , @xmath896 acts as the identity operator on @xmath865 .
it should also be possible ( as conjectured in remark [ conjrem ] ) to extend this dual plancherel formula to a more degenerate class of functions than presently considered .
we may combine theorems [ kqbosonidsd ] and [ kqbosoniddualsd ] to arrive at the following result .
[ bilinearprimesd ] define a bilinear pairing acting on two functions @xmath439 and @xmath93 by @xmath897 [ isothmsd ] the semi - discrete transform @xmath866 induces an isomorphism between @xmath35 and @xmath865 with inverse given by @xmath873 .
moreover , for any @xmath73 @xmath898 and for any @xmath899 @xmath900 corollaries [ completenesscorsd ] and [ northsd ] are immediate consequences of theorem [ kqbosonidsd ] . proposition [ specorthsd ]
can either be proved directly ( as done for proposition [ specorth ] ) or derived from theorem [ kqbosoniddualsd ] ( assuming that theorem is proved directly via a limit transition from theorem [ kqbosoniddual ] ) .
[ completenesscorsd ] any function @xmath63 can be expanded as @xmath901 and also as @xmath902 [ northsd ] for @xmath466 , regard @xmath903 and @xmath904 as functions @xmath905 and @xmath906 .
then @xmath907 [ specorthsd ] consider functions @xmath899 .
then we have that @xmath908 the following results follow from corollary [ completenesscorsd ] .
we then apply them to the analysis of the semi - discrete stochastic heat equation .
[ bwdappsd ] for @xmath909 the backward equation @xmath910 with @xmath497 is uniquely solved by @xmath911 where @xmath912 .
[ fwdappsd ] for @xmath909 the forward equation @xmath913 with @xmath497 is uniquely solved by @xmath914 where @xmath912 .
we may utilize the above solution to the backward equation to compute integral formulas for the moments of the probabilistic model we now introduce . though we phrase it in terms of a coupled system of stochastic odes , it is equivalent to the partition function of the oconnell - yor semi - discrete directed polymer @xcite ( see also @xcite for more recent developments in the study of this model ) .
[ def : semidiscshe ] the _ semi - discrete stochastic heat equation _ with initial data @xmath915 is the system of stochastic odes @xmath916 where @xmath917 and the @xmath918 are independent standard brownian motions .
we will assume that @xmath919 is compactly supported .
as shown in ( * ? ? ?
* theorem 4.1.26 ) ( see also ( * ? ? ?
* proposition 6.2 ) ) the above system is a @xmath106 scaling limit of @xmath0-tasep with finitely many particles ( cf .
section [ qtasepsec ] ) .
let @xmath920 $ ] . using the feynman - kac representation ( and the so - called replica method ) one shows ( cf .
* section 6.2 ) ) that these moments satisfy @xmath921 \bar{z}(t,\vec{n}).\ ] ] this can be thought of as a discrete version of the delta bose gas considered in section [ deltabosesec ] . as in remark
[ bosonhamiltoniansymsd ] ( see also ( * ? ? ?
* proposition 6.3 , ( c ) ) ) one sees that for @xmath16 , @xmath922 where @xmath923 solves the backward equation @xmath924.\ ] ] it follows from corollary [ bwdappsd ] that we may solve the above equation succinctly in terms of contour integrals . for @xmath925 of compact support ,
@xmath109 is given by the right - hand side of the first and the second lines of ( [ ftvecnsd ] ) .
let us apply this for the example of the semi - discrete stochastic heat equation with delta initial data @xmath926 .
this is the limit of step initial data for @xmath0-tasep .
one should be able to likewise work out the half - stationary limit , though we do not pursue that here .
the following result previously appeared in ( * ? ? ?
* section 5.2.2 ) .
[ dualiyresultsd ] for delta initial data @xmath926 we have @xmath927 = \oint_{\tilde{\gamma}_1 } \frac{dz_1}{2\pi { \mathbf i } } \cdots \oint_{\tilde{\gamma}_k } \frac{dz_k}{2\pi { \mathbf i } } \prod_{1\leq a < b\leq k } \frac{z_a - z_b}{z_a - z_b-1 } \prod_{j=1}^{k } ( z_j)^{-n_j } e^{t(z_j-1)}.\ ] ]
we briefly describe a further scaling limit of the results of section [ semidiscsec ] ( as well as those earlier in the paper ) .
we are informal here and do not give precise statements ( see references for such statements ) .
the analog ( and limit ) here of the @xmath0-boson particle system and the discrete delta bose gas is the continuum delta bose gas , and the analog ( and limit ) of @xmath0-tasep and the semi - discrete stochastic heat equation is the continuum stochastic heat equation ( or equivalently the kardar - parisi - zhang equation ) , see ( * ? ? ?
* section 6 ) , @xcite , or @xcite .
a rigorous treatment of this sort of limiting procedure ( starting from the system considered in section [ vandiejen ] ) is given in @xcite .
the @xmath928 dimensional continuum stochastic heat equation with initial data @xmath929 is the solution to the stochastic partial differential equation : @xmath930 where @xmath931 is space - time gaussian white noise ( see @xcite for more details and background ) . in a similar manner as for the semi - discrete stochastic heat equation , the joint moments ( time @xmath8 fixed , spatial variables @xmath274 varying )
satisfy simple closed evolution equations .
in particular ( cf .
@xcite or ( * ? ? ?
* section 6 ) ) for @xmath932 , defining @xmath933 $ ] @xmath934 \bar{\mathcal{z}}(t,\vec{x}),\ ] ] where @xmath935 is the dirac delta function .
this is often called the ( imaginary time ) delta bose gas with attractive coupling constant , or the ( imaginary time ) lieb - liniger model with attractive delta interaction .
uniqueness of solutions to the above equation ( within a class of solutions with suitable control on their growth in @xmath308 ) should be attainable , but we are unaware of an exact reference in the literature .
it is standard in the physics literature ( cf .
@xcite ) to postulate the following reduction : let @xmath936 then , for @xmath937 , @xmath938 where @xmath939 solves @xmath940 and satisfies the two - body boundary conditions that @xmath941 here @xmath942 means the limit as @xmath943 from below . to our knowledge
, this reduction has not been rigorously justified . in the work of heckman and opdam @xcite , the above system ( laplacian plus two - body boundary conditions )
is called the yang system @xcite ( see also @xcite for other root systems ) corresponding to the type @xmath630 root system and an attractive coupling constant ( of magnitude 1 in our setting ) . following heckman - opdam @xcite
, we now define the analogs of @xmath62 and @xmath66 , though without carefully specifying the domains on which they act .
the fourier - yang transform @xmath950 takes functions @xmath951 into symmetric functions @xmath952 via @xmath953 note that the @xmath954 factor in defining @xmath955 plays no role in this integration as it is identically equal to @xmath956 except on a set of measure 0 . 1 . for all @xmath962 , @xmath963 ; 2 . for all @xmath964 and @xmath965 such that @xmath966 for @xmath360 and @xmath967 for @xmath362 ,
the function @xmath886 is analytic in the complex domain @xmath968 and is bounded in modulus on that domain by @xmath969 for some constants @xmath970 ( depending on @xmath971 but not @xmath860 ) .
heckman and opdam @xcite showed that when restricted to a suitable class of functions , the fourier - yang transform @xmath950 is an isomorphism onto its image with inverse @xmath957 . or ,
equivalently , on this set of functions @xmath975 .
this is the analog of theorem [ kqbosonid ] and the method of proof we employed herein is adapted from their work . just as before , from this plancherel formula and lemma [ expandlemct ] , one proves completeness of the bethe ansatz ( see ( * ? ? ?
* remark 6.2.5 ) for a detailed history of this question )
. there should be an analog of the dual plancherel formula ( theorem [ kqbosoniddual ] ) and the plancherel isomorphism ( theorem [ isothm ] ) in this setting , though we have not seen this in the mathematics literature .
the physics literature contains some work in the direction of spectral orthogonality in this context @xcite . in the past few years
there have been many works in the physics literature using the above delta bose gas to study the moments of the stochastic heat equation with various types of initial data ( flat / half - flat @xcite , stationary @xcite , or more general @xcite ) or at different times @xcite .
these calculations are counterparts to the @xmath0-boson plancherel theorem discussed in section [ qtasepsec ] .
* the contour @xmath44 is a circle around 1 , small enough so as not to contain @xmath0 ; * for all @xmath354 , the interior of @xmath41 contains the image of @xmath42 multiplied by @xmath0 ; * for all @xmath355 , there exist deformations @xmath356 of @xmath357 to @xmath44 so that for all @xmath358 with @xmath361 for @xmath360 , and @xmath370 for @xmath362 , the function @xmath886 is analytic in a neighborhood of the area swept out by the deformation @xmath356 . then we have the following residue expansion identity : @xmath977 where @xmath80 was defined in ( [ wlambda ] ) . here
we define @xmath978 and @xmath979_{i , j=1}^{\ell(\lambda ) } \prod_{j=1}^{\ell(\lambda ) } w_j^{\lambda_j } q^{\frac{\lambda_j(\lambda_j-1)}{2 } } \frac{dw_j}{2\pi { \mathbf i}}\ ] ] where @xmath980 .
we proceed sequentially and deform ( using the deformation @xmath982 afforded from the hypotheses of the theorem ) the @xmath405 contour to @xmath44 , and then deform ( using the deformation @xmath983 afforded from the hypotheses of the theorem ) @xmath984 contour to @xmath44 , and so on until all contours have been deformed to @xmath44 .
however , due to the @xmath985 terms in the denominator of the integrand , during the deformation of @xmath41 we may encounter simple poles at the points @xmath986 , for @xmath43 .
the residue theorem implies that the integral on the left - hand side of ( [ qak ] ) can be expanded into a summation over integrals of possibly few variables ( all along @xmath44 ) whose integrands correspond to the various possible residue subspaces coming from these poles .
our proof splits into three basic steps .
first , we identify the residual subspaces upon which our integral is expanded via residues .
this brings us to equation ( [ rhs2 ] ) .
second , we show that these subspaces can be brought to a canonical form via the action of some @xmath297 .
this enables us to simplify the summation over the residual subspaces to a summation over partitions @xmath81 and certain subsets of permutations @xmath298 in @xmath184 .
inspection of those terms corresponding to @xmath297 not in these subsets shows that they have zero residue contribution and hence the summation can be completed to include all of @xmath184 .
this brings us to equation ( [ above3 ] ) . and third , we rewrite the function whose residue we are computing as the product of an @xmath184 invariant function ( which contains all of the poles related to the residual subspace ) and a remainder function .
we use lemma [ reslemma ] to evaluate the residue of the @xmath184 invariant function and we identify the summation over @xmath297 of the substitution into the remainder function as exactly giving the @xmath987 function in the statement of the proposition we are presently proving . nested contour integral as explained in example [ ex2 ] . on the left - hand side ,
the @xmath403 contour is deformed to the @xmath607 contour and a pole is crossed along @xmath989 at the point @xmath990 ( here the point @xmath991 is drawn as a small black bullet and its location is on the solid line circle ; and the point @xmath992 is drawn as a small white bullet and its location is on the dotted line circle ) . on the upper right - hand side the effect of picking the integral
the @xmath607 contour is represented as a ( doubled ) solid line circle since both @xmath991 and @xmath993 are integrated along it ( these correspond to the two black bullets ) .
the white bullets are along @xmath989 ( the dotted line circle ) and represent @xmath990 and @xmath994 . as the @xmath367 contour is deformed to @xmath607 these residues must be taken .
on the lower right - hand side the effect of picking the initial residue at @xmath995 is shown . as the @xmath367 contour is deformed to @xmath607 a pole
is encountered along @xmath996 at the location @xmath997 ( as before @xmath991 is the black bullet and @xmath997 is the white bullet . ] [ ex2 ] when @xmath988 the integrand on the left - hand side of ( [ qak ] ) contains the fractions @xmath998 times the function @xmath999 which ( by the hypotheses of the proposition ) does not have poles between @xmath357 and @xmath44 ( for @xmath1000 ) .
thinking of @xmath991 as fixed along the contour @xmath607 , we begin by deforming the @xmath403 contour to @xmath607 .
as we proceed , we necessarily encounter a single simple pole at @xmath995 .
the residue theorem implies that our initial integral equals the sum of ( a ) the integral where @xmath993 is along @xmath607 , and ( b ) the integral with only @xmath395 and @xmath991 remaining and integrand given by taking the residue of the initial integrand at @xmath995 .
let us consider separately these two pieces .
for ( a ) we now think of @xmath993 and @xmath991 as fixed along @xmath607 and deform the @xmath367 contour to @xmath607 , encountering two simple poles at @xmath1001 and @xmath1002 . thus ( a ) is expanded into a sum of three terms : the integral with @xmath1003 and @xmath991 along @xmath607 ; and the integral with only @xmath993 and @xmath991 remaining ( along @xmath607 ) and the residue taken at either @xmath1004 or @xmath1002 . for ( b ) we now think of @xmath991 as fixed along @xmath607 and deform the @xmath367 contour to @xmath607 , encountering a simple pole at @xmath1005 .
this is because the residue of ( [ cross ] ) at @xmath1006 equals @xmath1007 thus ( b ) is expanded into a sum of two terms : the integral with @xmath395 and @xmath991 ( along @xmath607 and with the above expression in the integrand ) ; and the integral with only @xmath991 remaining ( along @xmath607 ) and the residue of the above term taken at @xmath1005 .
gathering the various terms in this expansion , we see that the residue subspaces we sum over are indexed by partitions of @xmath52 ( here @xmath988 ) and take the form of geometric strings with the parameter @xmath0 . for example , @xmath1008 corresponds to the term in the residue expansion in which all three variables @xmath1009 and @xmath991 are still integrated , but along @xmath607 . on the other hand
, @xmath1010 corresponds to the term in which the residue is taken at @xmath1011 and the only variable which remains to be integrated along @xmath607 is @xmath991 .
the partition @xmath1012 corresponds to the three remaining terms in the above expansion in which two integration variables remain . in general
, @xmath83 corresponds to the number of variables which remain to be integrated in each term of the expansion .
let us now turn to the general @xmath52 case . by the hypotheses of the theorem
, the function @xmath1013 has no poles which are encountered during contour deformations hence it plays no role in the residue analysis .
as we deform sequentially the contours in the left - hand side of ( [ qak ] ) to @xmath44 we find that the resulting terms in the residue expansion can be indexed by partitions @xmath81 along with a list ( ordered set ) of disjoint ordered subsets of @xmath1014 ( whose union is all of @xmath1014 ) @xmath1015 let us call such a list @xmath1016 . for a given partition @xmath164 call @xmath1017 the collection of all such lists @xmath1016 corresponding to @xmath164 . for @xmath988 and @xmath1012 , in example [ ex2 ]
we saw there are three such lists which correspond with @xmath1018 for such a list @xmath1016 , we write @xmath1019 as the residue of the function @xmath393 at @xmath1020 and regard the output as a function of the terminal variables @xmath1021 .
there are @xmath83 such strings and consequently that many remaining variables ( though we have only written the first two strings above ) . with the above notation in place , we may write the expansion of the integral on the left - hand side of ( [ qak ] ) as @xmath1022 the factor of @xmath1023 arose from multiple counting of terms in the residue expansion due to symmetries of @xmath164 .
for example , for the partition @xmath1024 , each @xmath1025 , corresponds uniquely with a different @xmath1026 in which the @xmath193 and @xmath1027 variables in ( [ ijvar ] ) are switched .
since these correspond with the same term in the residue expansion , this constitutes double counting and hence the sum should be divided by @xmath1028 .
the reason why our residual subspace expansion only corresponds with strings is because if we took a residue which was not of the form of a string , then for some @xmath1029 we would be evaluating the residue at @xmath1030 and @xmath1031 .
however , the vandermonde determinant in the numerator of our integrand would then necessarily evaluate to zero .
therefore , such possible non - string residues ( coming from the denominator ) in fact have zero contribution .
* step 2 : * for each @xmath1025 relabel the @xmath278 variables as @xmath1032 and observe that there exists a unique permutation @xmath297 for which @xmath1033 .
let us also call @xmath1034 , for @xmath1035 .
we introduce a canonical form for taking strings of residues . for a function @xmath1036 define @xmath1037 to be the residue of @xmath1038 at @xmath1039 with the output regarded as a function of the terminal variables @xmath1040 .
we call each sequence of identifications of variables a _ string_. as all poles which we encounter in what follows are simple , the above residue evaluation amounts to @xmath1041 it is also convenient to define @xmath1042 as the function of @xmath1040 , which is the result of substituting the relations of ( [ resqvalues ] ) into @xmath1038 . with the above notation we rewrite the term corresponding to a partition @xmath164 in the right - hand side of ( [ rhs2 ] ) as @xmath1043 where @xmath1044 is defined in equation ( [ resqvalues ] ) . note that the output of the residue operation is a function of the variables @xmath1045
. one should observe that the above expression includes the summation over all @xmath297 , and not just those which arise from an @xmath1025 as above .
this , however , is explained by the fact that if @xmath298 does not arise from some @xmath1025 , then the residue necessarily evaluates to zero ( hence adding these terms is allowed ) . to see this , observe that in the renumbering of variables discussed above , only permutations with @xmath1046 participated .
any other @xmath298 must violate one of these strings of conditions .
consider , for example , some @xmath298 with @xmath1047 .
this implies that the term @xmath1048 shows up in the denominator of ( [ above3 ] ) , as opposed to the term @xmath1049 .
residues can be taken in any order , and if we first take the residue at @xmath1050 , we find that the above denominator does not have a pole ( nor do any other parts of ( [ above3 ] ) ) and hence the residue is zero . similar reasoning works in general .
it is convenient to rewrite the product over @xmath1051 as an @xmath184 invariant function , times a function that is analytic at the points in which the residue is being taken : @xmath1052 since it is only the @xmath184 invariant function that contains the poles with which we are concerned , it allows us to rewrite ( [ above3 ] ) as @xmath1053 we use lemma [ reslemma ] to evaluate the above residue , and we easily identify the substitution on the second line with @xmath1054 as in the statement of the proposition . combining these two expressions and summing the resulting expression over @xmath81 we arrive at the desired residue expansion claimed in the statement of the proposition .
[ reslemma ] for all @xmath1 , @xmath81 and @xmath101 , we have that @xmath1055_{i , j=1}^{\ell(\lambda)},\ ] ] where we have renamed the variables remaining after the residue operator as @xmath1034 , for @xmath1035 . the product on the left - hand side of ( [ reslemmaeqn ] ) involves terms in which @xmath193 and @xmath1027 are in the same string of variables in ( [ resqvalues ] ) as well as terms in which they are in different strings .
we need to compute the residue of the same string terms and multiply it by the substitution of variables into the different string terms . returning to the statement of the lemma
, we see that we can evaluate the desired residue by multiplying the same string terms over all strings in ( [ resqvalues ] ) as well as multiplying all terms corresponding to pairs of different strings . using the above calculations we obtain @xmath1064 it is easy now to rewrite ( using the cauchy determinant ) the above expression so as to produce the equality of the lemma , as desired . from the definition of @xmath337
we can rewrite this identity as @xmath1066 where we recall @xmath211 and @xmath1067 from section [ notations ] .
define @xmath1068 to be the left - hand side of the above desired equality .
we will proceed by induction in @xmath52 .
it is straightforward to check that the identity holds for @xmath670 , and for later use define @xmath1069 .
we may write @xmath1070 the above formula involves some notation which should be explained .
the term @xmath1071 is the summation over all subsets @xmath1072 and @xmath1073 such that @xmath1074 , @xmath1075 and @xmath1076 .
the term @xmath1077 is the summation over all permutations @xmath1078 which permute the elements of @xmath1016 and fix the elements of @xmath1079 , and , conversely , the term @xmath1080 is the summation over all permutations @xmath1081 which permute the elements of @xmath1079 and fix the elements of @xmath1016 .
the final summation is over @xmath1082 with @xmath1083 fixed .
we call @xmath1084 and set @xmath1085 and @xmath1086 .
notice that the summation over @xmath1078 can be readily computed to be @xmath1087 using ( [ mqinverse ] ) .
note that when @xmath1087 is divided by @xmath1088 this yields @xmath1089 .
further rearrangement of terms leads to @xmath1090 in order to prove the inductive step for @xmath52 it suffices to check that the above expression equals the right - hand side of ( [ ik ] ) .
this reduces the proof of the lemma to showing the following equality : @xmath1094 the summation on the left - hand side does not include the term corresponding to @xmath1095 . including this term cancels a portion of the left - hand side and further cross multiplying brings the desired equality to : @xmath1096 where the term @xmath1071 ( without the @xmath1097 ) represents the summation over all @xmath1016 and @xmath1079 such that @xmath1098 and @xmath1076 .
this desired equality is a @xmath0-deformation of the binomial expansion of @xmath1099 . to prove it ,
let @xmath1100 be the vandermonde determinant and let @xmath1101 be the @xmath0-shift operator defined as @xmath1102 . making the change of variables
@xmath1103 we can rewrite the desired equality ( [ desireeqn ] ) as @xmath1104 this eigenfunction relationship for @xmath1105 is easily proved by writing @xmath1106_{i , j=1}^{k}$ ] and observing that @xmath1107_{i , j=1}^{k}.\ ] ] by elementary column transformations the above matrix can be brought to the form @xmath1108_{i , j=1}^{k } = \prod_{j=1}^{k } ( 1-\alpha / q^j ) v(z),\ ] ] thus proving the desired equality ( [ desireeqn ] ) and ultimately the inductive step and the lemma . | we develop spectral theory for the generator of the @xmath0-boson ( stochastic ) particle system .
our central result is a plancherel type isomorphism theorem for this system .
this theorem has various implications .
it proves the completeness of the bethe ansatz for the @xmath0-boson generator and consequently enables us to solve the kolmogorov forward and backward equations for general initial data . owing to a markov duality with @xmath0-tasep ,
this leads to moment formulas which characterize the fixed time distribution of @xmath0-tasep started from general initial conditions .
the theorem also implies the biorthogonality of the left and right eigenfunctions .
we consider limits of our @xmath0-boson results to a discrete delta bose gas considered previously by van diejen , as well as to another discrete delta bose gas that describes the evolution of moments of the semi - discrete stochastic heat equation ( or equivalently , the oconnell - yor semi - discrete directed polymer partition function ) .
a further limit takes us to the delta bose gas which arises in studying moments of the stochastic heat equation / kardar - parisi - zhang equation .
[ section ] [ section ] [ theorem]conjecture [ theorem]lemma [ theorem]proposition [ theorem]corollary [ theorem]claim [ theorem]critical point derivation [ theorem]experimental result # 1 # 1#1 # 1 [ theorem]remark [ theorem]example [ theorem]definition [ theorem]definitions |
bound states are normally associated with basins of attraction , and it is always surprising to find them in the presence of a repulsive interaction .
the most remarkable case is the pairing of electrons in normal superconductors , mediated by lattice distortions .
a new kind was recently observed for ultracold rubidium atoms in an optical lattice [ 1 ] , where the binding arises from pure quantum interference .
the pairs were shown to be stable , thus suggesting that they might form a superfluid phase if the composite is a boson .
the effect was originally predicted for a one dimensional string [ 2 - 5 ] .
the eigenstates decay exponentially in the relative coordinate , while in the center of mass coordinate they are extended over the whole lattice . in higher dimensions one can argue that paired states may exist as well . to see this , consider two particles that repel moving in a two dimensional lattice .
together , the pair combine four spatial degrees of freedom , thus allowing that it be formally described as a single particle moving in a 4d lattice .
a particle - particle interaction appears then as an interface potential in four dimensions , which decays to both sides of the hyperplane x@xmath0= x@xmath1 , y@xmath0= y@xmath1 , where ( x@xmath0,y@xmath0 ) , ( x@xmath1,y@xmath1 ) are cartesian coordinates for the particles position in real 2d space .
surface states bounded to the interface are then expected even if the potential locally rises , as would occur under particle - particle repulsion .
since the interface represents matching coordinates for the two particles in 2d such surface states are paired states in the lattice
. a similar argument applies in 3d .
we find the 2d case particularly interesting since the onset of pairing could be relevant to high temperature superconductors where transport is related to conducting sheets , and to certain 2d experimental probes for which the physics is still controversial [ 6 - 16 ] .
it is believed that in the latter case localization due to disorder on the conducting sheet should produce a divergent resistivity as the temperature is lowered , but this behavior was not observed at all electron densities in metal - oxide - semiconductor field - effect transistors ( mosfet ) . in fact , a metal insulator transition was found in low density 2d samples as the density is slightly increased [ 8 ] .
an early prediction invoked the strongly interacting nature of a low density electron fluid [ 9 - 10 ] , while more recent theories study the effect of formation of a wigner solid [ 13 ] or a superconducting phase [ 14 ] , a strong spin - orbit interaction [ 15 ] and a few classical effects [ 16 ] , none of which has received consensus for describing the physics at the root of the transition [ 17 ] .
as we shall show , paired singlet states are predicted to exist in 2d for repelling particles in a lattice . in section
ii we define the model and show that the density of states reveals the presence of two bands of different character , one corresponding to a bloch particle in 4d and one describing a 2d surface state . in section iii we show that the dynamics of such surface state is that of a composite boson of charge 2e , our main finding , and in section iv we discuss our results .
consider for definiteness two interacting tight - binding electrons in 2d in a uniform external field .
the asociated hamilton operator for such system reads : @xmath2@xmath3where @xmath4 and @xmath5 are the creation and annihilation operators for one electron located at the site @xmath6 with spin @xmath7 , @xmath8 the usual hopping energy parameter , @xmath9 the sum of position vectors , @xmath10 the charge of the electron , @xmath11 and @xmath12 the electric field magnitude and direction , respectively , and @xmath13 the lattice parameter .
the last term represents a two - body hubbard contact interaction potential @xmath14 .
we assume that the interaction is strongly screened and may be ignored for particle - particle separations beyond a lattice constant and discuss only the singlet state [ 3 ] . in the wannier representation , the time dependent wave function for the pair in a singlet state
may be expanded as@xmath15where the sum runs over all lattice sites , @xmath16 is the amplitude for an electron with spin @xmath17 to be at ( @xmath18 ) , and another with spin @xmath19 @xmath17 to be at ( @xmath20 ) , while the ket @xmath21 represents such state .
the amplitude obeys the schrdinger equation of motion@xmath22where @xmath23 is the sum of amplitudes of all nearest neighbors to the site ( @xmath24 in the lattice , and @xmath25 in units of the lattice constant .
we have assumed that the electric field is along the diagonal @xmath26 @xmath27 . by making the replacement @xmath28 the associated eigenvalue equation in the absence of an external field becomes@xmath29 \notag\ ] ] @xmath30 where now @xmath31one can easily check that @xmath32 is an eigenvalue of this equation at the lowest band edge @xmath33 , of eigenvalue @xmath34 .
this solution represents extreme pairing , when one particle is precisely on top of the other . in the absence of interaction solutions of eq .
( 4 ) are plane waves in a bloch band , with eigenvalues @xmath35 where @xmath36 is the center of mass wave number . the band remains when the repulsive interaction is turned on , while additional states appear at positive energy .
this is seen in fig .
1 , where the density of states obtained by numerical diagonalization of the eigenvalue equation for several positive values of the parameter u is shown . data corresponds to a finite lattice with n=49 plaquettes using periodic boundary conditions .
the bell shaped curve represents the density of states of a single particle in 4d , the analog problem in this case , while a 2d - like density of states profile representing surface states in 4d moves away as u grows .
note that the lower edge of this band is at @xmath37 as expected . , it is that of a single tight - binding particle in 4d .
when @xmath38 a new subband emerges , moving to higher energies as @xmath14 is increased . ]
to confirm the presence of two different kinds of states , we have calculated the average distance between particles@xmath39on a finite 2d sample of n plaquetes using periodic boundary conditions . here
@xmath40 labels the eigenstate considered .
the results are shown in fig .
2 , where the quantity @xmath41 is shown as a function of sample size both for states in the b - band as well as states in the u - band .
we see that the average separation of the particles is essentially null for states in the u - band for all cell sizes , as is expected for a bounded pair . for states in the bloch band
the separation is finite , with a value that grows linearly with @xmath42 as expected for extended states in two dimensions . .
the mean separation in a bloch state ( b - band ) grows linearly with the edge size , while for the u - band it stays near zero as expected for paired states . ] a more dramatic confirmation that repelling electrons in a lattice form bounded pairs is found when the electric field is turned on and the time evolution of a given initial state is followed by solving eq .
it is well known that charged particles in a lattice experience bloch oscillations in the presence of the field , with a frequency proportional to the particle charge[18 ] . to find out
if such oscillations are present we studied the time evolution of the average position along the x axis of each particle@xmath43where @xmath44 is the particle index , and a similar equation for the coordinate @xmath45 the numerical work was done using a half implicit numerical method which is second - order accurate and unitary [ 19 ] .
the positions were indeed found to oscillate , not simply performing bloch - type oscillations but a more complex pattern depending upon initial conditions .
figure fig - fourier_3 shows the power spectrum for a 2d lattice of @xmath46 sites and an electric field energy @xmath47 , set rather large in order to avoid reflections from the edges in our reduced numerical sample . in the absence of interactions only the bloch frequency @xmath48 ( @xmath49 ) has any weight , as seen in fig .
we next set the interaction strength at @xmath50 a large value appropriate to have a u - band well separated from the extended states .
for an initial condition with zero amplitude save for points far from the interaction region @xmath51 , @xmath52 , the situation is unchanged .
but for an initial condition with unit amplitude in sites over the interaction hyperplane in the 4d space of coordinates @xmath53 , @xmath54 and either @xmath55 or @xmath56 when @xmath57 representing a portion of the hyperplane and a few nearby points , we see a bloch oscillation with twice the bloch frequency , with a power spectrum as shown in fig .
we interpret this new frequency as arising from motion of a composite particle of charge @xmath58 in an electric field , thus confirming that the pair behaves dynamically as a stable composite in the absence of dissipation .
additional interaction - induced oscillations ( inino ) appear near @xmath59 and @xmath60 shown in more detail in the inset .
these latter oscillations , seen also in 1d systems , are the result of pure correlated electron transport [ 20 ] .
we note that our numerical results show that the features just described persist as the interaction strength enters the region @xmath61 only that hybridization of localized and extended states makes the power spectrum more complex . .
( a ) is in the absence of interactions and ( b ) is for a hubbard interaction strength @xmath62 .
the inset shows details near the hubbard frequency .
frequencies are in units of @xmath63 . ]
in summary , we have shown that two interacting electrons moving in a lattice in 2d have singlet paired states grouped in a band , with the dynamics of a composite particle of charge 2e .
our results were obtained for a highly screened contact repulsive potential , yet studies in 1d show that a coulomb tail does not destroy the pairing [ 2 ] in the singlet state .
although the specific treatment refers to electrons in a lattice , the main results may be extended to other systems such as atoms in an optical lattice [ 1 ] . in order to understand the origin of the paired states we recall that ordinary bloch states may be thought of as atomic states that hybridize with lattice neighbors , forming a tunneling network capable of sustaining extended states that are grouped in bands .
similarly , the paired eigenstates may be described as a high energy two - electron single well state that forms a band of extended states owing to pair tunneling throughout the lattice .
the hubbard model captures the escence of this picture , but details such as possible molecular two - electron extended states may arise if a finite range interaction is included .
the energy of the paired states scales like the interaction strength @xmath14 .
if this latter quantity is of the order of the bandwidth , such states may lie partly or wholly within the band of extended states , allowing pairs to form in the ground state if there are enough electrons in the sample .
the paired states equal in number the extended states , since they correspond to single particle surface states over a 2d planar interface in a 4d lattice .
we thus speculate that if condensation of such pairs indeed occurs then a transition to low resistance should take place as the density is increased , as is actually observed in experiment [ 11 ] .
our model yields a pair whose dynamics is that of a stable composite of charge 2e .
such stability may be reduced by dissipation mechanisms such as phonon excitation . in a static periodic potential such as that generated by an optical lattice phonons
are not present , and a pair of energy @xmath64 @xmath65 above the band center is expected to be stable since its decay into a band state is forbiden by energy conservation . under an external field
@xmath66 , transitions between extended and localized states give rise to oscillations .
if the associated frequencies @xmath67 and @xmath68 were accessible to experiment , then the sum @xmath69 @xmath70 would yield a direct measure of the interaction strength in the 2d sample .
[ 8 ] s. v. kravchenko , g. v. kravchenko , j. e. furneaux , v. m. pudalov , m. diorio , phys
b 50 , 8039 ( 1994 ) ; s.v .
kravchenko , w.e .
mason , g.e .
bowker , j.e .
furneaux , v.m .
pudalov , m. diorio , phys . rev .
b 51 , 7038 ( 1995 ) ; s.v .
kravchenko , d. simonian , m.p .
sarachik , w. mason , j.e .
furneaux , phys .
77 , 4938 ( 1996 ) .
[ 9 ] a. m. finkelshtein , z. phys .
b 56 , 189 ( 1984 ) ; c. castellani , c. di castro , p. a. lee , and m. ma , phys . rev .
b 30 , 527 ( 1984 ) ; c. castellani , c. d. dicastro , and p. a. lee , phys . rev .
b 57 , r9381 ( 1998 ) . [ 16 ] s. he and x. c. xie , phys .
80 , 3324 ( 1998 ) ; b. l. altshuler and d. l. maslov , phys .
82 , 145 ( 1999 ) ; t. m. klapwijk and s. das sarma , preprint cond - mat/9810349 ( 1998 ) ( solid state commun . , in press ) ; s. das sarma and e. h. hwang , phys .
83 , 164 - 7 ( 1999 ) . | we show that two tight binding electrons that repel may form a bounded pair in two dimensions .
the paired states form a band with energies that scale like the strength of the interaction potential . by applying an electric field
we show that the dynamics of such states is that of a composite particle of charge 2e .
the system still sustains bloch - like states , so that if the two bands overlap single and paired states might coexist allowing for a bosonic fluid component that , if condensed , would decrease the resistance at low temperatures .
the presence of two bands allows for new oscillations whose experimental detection would permit a direct measurement of the interaction potential strength . |
a von neumann algebra is a an algebra of bounded linear operators on a hilbert space that is closed under the adjoint @xmath4-operation and that is closed in the weak operator topology .
von neumann algebras arise naturally in the study of groups and their actions on measure spaces .
these constructions go back to murray and von neumann s seminal papers ( * ? ? ?
* chapter xii ) and @xcite .
* if @xmath5 is a countable group , the left translation unitary operators on @xmath6 generate the _ group von neumann algebra _ @xmath7 .
* every action @xmath8 of a countable group @xmath5 by measurable transformations of a measure space @xmath9 and preserving sets of measure zero , gives rise to the _ group measure space von neumann algebra _ @xmath0 .
it is a central problem in the theory of von neumann algebras to classify @xmath7 and @xmath0 in terms of the group @xmath5 or the group action @xmath8 .
more generally , classifying or distinguishing families of von neumann algebras is extremely challenging . in the first part of this exposition ,
i give a panoramic overview of the spectacular progress that has been made in this area over the last years .
the overview is more thematically ordered than chronologically and necessarily incomplete .
several related important topics , including jones theory of subfactors or voiculescu s free probability theory , are not treated . _
all notions that are written in italics are defined in the preliminary section [ sec.prelim ] .
_ the simple von neumann algebras @xmath10 are those that can not be written as a direct sum of two .
equivalently , the center of @xmath10 is trivial and @xmath10 is called a factor .
murray and von neumann have classified factors into three types @xcite and proven that every von neumann algebra can be decomposed as a direct integral of factors @xcite . connes @xcite showed how general factors can be built up from those that admit a finite positive _ trace , _ called _
ii@xmath1 factors .
_ the final form of this decomposition theory is due to connes and takesaki @xcite .
altogether , ii@xmath1 factors form the basic building blocks of arbitrary von neumann algebras . the _ group von neumann algebras _
@xmath7 always admit a finite positive trace and are factorial if and only if @xmath5 has infinite conjugacy classes ( icc ) .
when @xmath8 is essentially free , ergodic , the set of @xmath11 with @xmath12 has measure zero .
ergodicity means that globally @xmath5-invariant measurable subsets have either measure @xmath13 or a complement of measure @xmath13 . ] and probability measure preserving ( p.m.p . ) , the _ group measure space von neumann algebra _
@xmath0 is a ii@xmath1 factor .
moreover , as proven by singer @xcite , its isomorphism class only depends on the _ equivalence relation _ given by the orbits of @xmath8 .
this lead to the study of group actions up to orbit equivalence @xcite and we refer to @xcite for surveys of the recent developments in this area of ergodic theory . [ par.noniso ] two von neumann algebras can be isomorphic in unexpected ways . while all hyperfinite ii@xmath1 factors were already shown to be isomorphic in @xcite , the culmination came with connes uniqueness theorem for _ amenable _ ii@xmath1 factors @xcite implying that all @xmath7 and all @xmath0 are isomorphic when @xmath5 is any amenable icc group or @xmath8 is an arbitrary free ergodic p.m.p .
action of an amenable group . in the early years examples of non - isomorphic ii@xmath1 factors
@xmath10 were obtained by analyzing asymptotically central sequences in a ii@xmath1 factor @xmath10 is called asymptotically central if @xmath14 converges to @xmath13 in the strong operator topology for every @xmath15 . ] of elements in @xmath10 : property gamma factor has property gamma if it admits an asymptotically central sequence of unitaries having trace @xmath13 . ]
@xcite allowed to prove that the free group factors @xmath16 are not hyperfinite , a refinement yielded uncountably many non - isomorphic ii@xmath1 factors @xcite and the @xmath17-invariant @xcite provided the first examples where @xmath10 is non - isomorphic to its opposite algebra @xmath18 and where @xmath19 .
the first rigidity phenomena for von neumann algebras were discovered by connes @xcite who showed that the _ fundamental group of a ii@xmath1 factor @xmath10 consists of the numbers @xmath20 where @xmath21 and @xmath22 run over the projections in @xmath10 satisfying @xmath23 .
here @xmath24 denotes the trace on @xmath10 . ] _ of @xmath7 is countable when @xmath5 is an icc property ( t ) group .
several properties of groups including _ property ( t ) , _ the haagerup property and related approximation properties were shown in @xcite to actually be properties of @xmath7 , leading to remarkable non - isomorphism and non - embeddability theorems for group von neumann algebras .
altogether it became clear that the world of ii@xmath1 factors is extremely rich , but that understanding the natural examples @xmath7 or @xmath0 in terms of the initial group or action is intrinsically very difficult .
[ par.defrig ] a major breakthrough in the classification of ii@xmath1 factors was realized by popa and his discovery of deformation/ rigidity theory @xcite ( see @xcite for a survey ) .
typically , popa studies von neumann algebras @xmath10 that have a rigid subalgebra
e.g. given by the _ relative property ( t ) _ such that the complement has a strong deformation property .
this gives the rigid subalgebra a canonical position within the ambient von neumann algebra and has lead in @xcite to the first example of a ii@xmath1 factor with trivial _
fundamental group : _ @xmath25
. the canonical position of @xmath26 implies that the fundamental group of @xmath10 equals the fundamental group of the _ orbit equivalence relation _ of the action @xmath27 , which is trivial because of @xcite
. in @xcite popa established a striking progress in his deformation / rigidity program by proving the following strong rigidity theorem for group measure space factors .
take an arbitrary free ergodic p.m.p .
action @xmath8 of a property ( t ) group @xmath5 and let @xmath28 be the _ bernoulli action _ of an arbitrary icc group @xmath29 .
if the group measure space factors @xmath0 and @xmath30 are isomorphic , then @xmath5 must be isomorphic with @xmath29 and their actions must be conjugate .
popa s strong rigidity theorem was the first result ever where conjugacy of actions could be deduced from the mere isomorphism of group measure space factors .
[ par.fundamental ] progress in the classification of group measure space factors went hand in hand with major developments in the calculation of invariants of ii@xmath1 factors .
the most well known invariant of @xmath10 is the fundamental group @xmath31 of murray and von neumann @xcite . in one of their long - standing questions they asked what subgroups of @xmath32 might occur as @xmath31 . until 10 years ago ,
progress on this question has been scarce . some ii@xmath1 factors , including the hyperfinite ii@xmath1 factor @xcite and @xmath33 @xcite , were shown to have fundamental group @xmath32 while connes @xcite proved that @xmath34 is countable whenever @xmath5 is an icc property ( t ) group .
a breakthrough in the understanding of fundamental groups came with popa s first examples of ii@xmath1 factors having trivial fundamental group @xcite and having prescribed countable fundamental group @xcite .
it remained a major open problem whether uncountable groups @xmath35 could appear as fundamental group .
in @xcite we solved this problem and proved that @xmath36 ranges over a large family of subgroups of @xmath32 , including all countable subgroups and many uncountable subgroups that can have arbitrary hausdorff dimension between @xmath13 and @xmath37 .
a similar result is true @xcite when @xmath38 is replaced by almost any infinite free product of non - trivial groups , while @xmath39 is necessarily trivial when @xmath8 is an arbitrary free ergodic p.m.p .
action of a free product of two finitely generated groups , one of them having property ( t ) .
so far , @xcite provide the only known constructions of group measure space factors with fundamental group different from @xmath40 or @xmath32 .
the fundamental group @xmath31 of a ii@xmath1 factor @xmath10 can also be viewed as the set of @xmath41 such that the ii@xmath42 factor @xmath43 admits an automorphism scaling the ( infinite ) trace @xmath44 by @xmath45 . in @xcite
we provide examples where @xmath46 , although @xmath43 admits no continuous trace - scaling action of @xmath32 .
another invariant of a ii@xmath1 factor @xmath10 is its outer automorphism group is defined as the quotient @xmath47 , where @xmath48 denotes the normal subgroup of @xmath49 consisting of the inner automorphisms @xmath50 , @xmath51 .
] @xmath52 . in @xcite ioana ,
peterson and popa established bass - serre isomorphism and subgroup ( rather subalgebra ) theorems for amalgamated free products of von neumann algebras . as a consequence they obtained the first calculations of outer automorphism groups and proved that @xmath52 can be any compact abelian group .
in particular , they positively answered the question on the existence of ii@xmath1 factors without outer automorphisms .
later we showed @xcite that in fact @xmath52 can be any compact group .
the results in @xcite are existence theorems involving a baire category argument .
we obtained the first concrete and explicit calculations of @xmath52 in @xcite and proved that @xmath52 can be any countable group .
both the elements of the fundamental group and the automorphisms of a ii@xmath1 factor @xmath10 give rise to _
hilbert @xmath10-@xmath10-bimodules _
@xmath53[0ex]{\scriptsize $ m$}{{\mathcal{h}}}\hspace{-0.2ex}\raisebox{-0.4ex}[0ex][0ex]{\scriptsize $ m$}}}$ ] .
an @xmath10-@xmath10-bimodule that is finitely generated , both as a left and as a right @xmath10-module is said to be of finite jones index @xcite .
the finite index @xmath10-@xmath10-bimodules form a c@xmath2-tensor category @xmath54 and this should be considered as the generalized ( or quantum ) symmetry group of @xmath10 .
whenever @xmath55 is a finite index subfactor , @xmath53[0ex]{\scriptsize $ m$}{{\mathord{\text{\rm l}}}^2(p)}\hspace{-0.2ex}\raisebox{-0.4ex}[0ex][0ex]{\scriptsize $ m$}}}$ ] is a finite index bimodule . in this sense
, @xmath54 also encodes the subfactor structure of @xmath10 .
in @xcite i proved the existence of ii@xmath1 factors @xmath10 such that @xmath54 is trivial , i.e. only consists of multiples of the trivial bimodule @xmath56 .
such ii@xmath1 factors have trivial fundamental group , trivial outer automorphism group and no non - trivial finite index subfactors .
explicit examples were provided in @xcite where also several concrete calculations of @xmath54 were made .
these calculations were exploited in @xcite to give a full classification of all finite index subfactors of certain ii@xmath1 factors . in @xcite
every representation category of a compact group @xmath57 is realized as @xmath54 .
more precisely , for every compact group @xmath57 we prove the existence of a minimal action of @xmath57 on a ii@xmath1 factor @xmath10 such that , denoting by @xmath58 the subfactor of @xmath57-invariant elements , the natural faithful tensor functor @xmath59 is surjective , i.e. an equivalence of categories .
[ par.superrigid ] two free ergodic p.m.p .
actions @xmath8 and @xmath60 are called * conjugate ( or isomorphic ) , if there exists an isomorphism of probability spaces @xmath61 and an isomorphism of groups @xmath62 satisfying @xmath63 for all @xmath64 and a.e .
@xmath11 ; * orbit equivalent ( oe ) , if there exists an isomorphism of probability spaces @xmath61 satisfying @xmath65 for a.e . @xmath11 ; * w@xmath2-equivalent , if @xmath66 .
obviously , conjugacy of actions implies orbit equivalence and singer @xcite proved that an orbit equivalence is the same as a w@xmath2-equivalence sending the _ group measure space cartan subalgebras of a ii@xmath1 factor @xmath10 is a maximal abelian subalgebra whose normalizing unitaries generate @xmath10 .
whenever @xmath8 is a free ergodic p.m.p .
action , @xmath67 is an example of a cartan subalgebra that we call of group measure space type .
not all ii@xmath1 equivalence relations can be implemented by a free action of a countable group @xcite and hence a general cartan subalgebra need not be of group measure space type . ]
_ @xmath68 and @xmath69 onto each other .
rigidity theory for group actions aims at establishing the converse implications under appropriate assumptions .
pioneering oe rigidity results were obtained by zimmer @xcite and the first breakthrough w@xmath2-rigidity theorems were proven by popa @xcite ( see paragraph [ par.defrig ] ) .
our aim here however is to discuss the ideal kind of rigidity , labeled w@xmath2- ( respectively oe- ) superrigidity , where the entire isomorphism class of @xmath8 is recovered from its w@xmath2-class ( resp .
oe class ) .
while a striking number of oe superrigid actions have been discovered over the last 10 years @xcite , the first w@xmath2-superrigid actions were only discovered very recently in my joint paper with popa @xcite .
we found a family of amalgamated free product groups @xmath5 and a large class of w@xmath2-superrigid @xmath5-actions , including ( generalized ) bernoulli actions , gaussian actions and certain co - induced actions . note that w@xmath2-superrigidity for an action @xmath8 is equivalent to the sum between its oe superrigidity and the uniqueness , up to unitary conjugacy , of @xmath68 as a group measure space cartan subalgebra in @xmath0 .
this makes w@xmath2-superrigidity results extremely difficult to obtain , since each one of these problems is notoriously hard .
contrary to the long list of oe superrigid actions referred to in the previous paragraph , unique cartan decomposition proved to be much more challenging to establish , and the only existing results cover very particular group actions .
thus , a first such result , obtained by ozawa and popa @xcite , shows that given any profinite action @xmath3 of a product of free groups @xmath70 , with @xmath71 , @xmath72 , all cartan subalgebras of @xmath73 are unitarily conjugate to @xmath68 . a similar result , covering groups @xmath5 that have the complete metric approximation property and that admit a proper @xmath37-cocycle into a non - amenable representation ,
was then proved in @xcite .
more recently , peterson showed @xcite that factors arising from profinite actions of non - trivial free products @xmath74 , with at least one of the @xmath75 not having the haagerup property , have a unique group measure space cartan subalgebra , up to unitary conjugacy . but
so far , none of these group actions could be shown to be oe superrigid .
nevertheless , an intricate combination of results in @xcite were used to prove the existence of virtually and @xmath76 is up to finite index subgroups of @xmath77 . ]
w@xmath2-superrigid group actions @xmath3 in @xcite , by a baire category argument .
in @xcite we established a very general unique cartan decomposition result , which allowed us to obtain a wide range of w@xmath2-superrigid group actions .
thus , we first proved the uniqueness , up to unitary conjugacy , of the group measure space cartan subalgebra in the ii@xmath1 factor given by an arbitrary free ergodic p.m.p .
action of any group @xmath5 belonging to a large family @xmath78 of amalgamated free product groups . by combining this with kida s oe superrigidity in @xcite
, we deduced that if @xmath79 denotes the group of upper triangular matrices in @xmath80 , then any free mixing p.m.p .
action of @xmath81 is w@xmath2-superrigid . in combination with @xcite
, we proved that for many groups @xmath5 in the family @xmath78 , the bernoulli actions of @xmath5 are w@xmath2-superrigid .
very recently , ioana @xcite obtained the beautiful result that all bernoulli actions of property ( t ) groups are w@xmath2-superrigid .
recall from paragraph [ par.defrig ] popa s strong rigidity theorem for bernoulli actions and note the asymmetry in the formulation : there is a ( rigidity ) condition on the group @xmath5 and a ( deformation ) condition on the action @xmath60 .
one of the novelties of @xcite is a transfer of rigidity principle showing that under w@xmath2-equivalence of @xmath3 and @xmath76 , some of the rigidity properties of @xmath5 persist in the arbitrary unknown group @xmath29 .
note however that property ( t ) itself is not stable under w@xmath2-equivalence : there exist w@xmath2-equivalent group actions such that @xmath5 has property ( t ) while @xmath29 has not ( see section [ sec.cartan ] ) .
related to the uniqueness problem of cartan subalgebras obviously is the existence question .
voiculescu @xcite proved that the free group factors admits no cartan subalgebra , because the presence of a cartan subalgebra forces the free entropy dimension of any generating set to be smaller or equal than @xmath37 .
another application of voiculescu s free entropy theory was given by ge @xcite who showed that the free group factors are prime : they can not be written as the tensor product of two ii@xmath1 factors .
using delicate c@xmath2-algebra techniques , ozawa @xcite proved that for all icc word hyperbolic groups @xmath5 in particular when @xmath5 equals the free group @xmath82 the group factor @xmath7 is solid : the relative commutant @xmath83 of an arbitrary diffuse subalgebra is injective .
obviously non - hyperfinite solid ii@xmath1 factors , as well as all their non - hyperfinite subfactors , are prime .
a combination of techniques from @xcite then allowed ozawa and popa @xcite to introduce a family of ii@xmath1 factors that have an essentially unique tensor product decomposition into prime factors .
@xmath84-rigidity @xcite a ii@xmath1 factor analogue for the vanishing of the first @xmath85-betti number of a group as well as bass - serre rigidity for free products of von neumann algebras @xcite provided further examples of prime factors .
the free group factors have no cartan subalgebra and are solid . in @xcite
both properties are brought together and @xmath16 is shown to be strongly solid : the normalizer of an arbitrary diffuse abelian subalgebra is hyperfinite .
other examples of strongly solid ii@xmath1 factors were given in @xcite .
in so far as the above gave an overview of some recent developments , in the rest of the paper i present the main ideas behind a number of chosen topics : popa s deformation / rigidity theory in section [ sec.defrig ] , computations of fundamental groups in section [ sec.fundamental ] , the ( non-)uniqueness of cartan subalgebras in section [ sec.cartan ] and w@xmath2-superrigidity in section [ sec.superrigid ] .
a _ finite trace _ on a von neumann algebra @xmath10 is a linear map @xmath86 satisfying @xmath87 for all @xmath88 .
we say that @xmath24 is _ positive _ if @xmath89 for all positive operators @xmath90 . a positive trace @xmath24 is called _ faithful _ if the equality @xmath91 implies that @xmath92 .
a positive trace @xmath24 is called _ normal _ if @xmath24 is weakly continuous on the unit ball of @xmath10 . a _ ii@xmath1 factor _ is a von neumann algebra with trivial center that admits a non - zero finite positive trace @xmath24 and that is non - isomorphic to a matrix algebra @xmath93 .
normalizing @xmath24 such that @xmath94 , the trace is unique .
moreover @xmath24 is automatically normal .
we denote by @xmath95 the @xmath84-norm corresponding to @xmath24 . completing @xmath10 w.r.t .
the scalar product @xmath96 yields the hilbert space @xmath56 , which is an @xmath10-@xmath10-bimodule by left and right multiplication on @xmath10 .
we denote by @xmath97 the non - normalized trace on @xmath93 .
occasionally , @xmath97 denotes the infinite trace on positive operators in @xmath98 .
let @xmath5 be a countable group
. then @xmath7 is the unique tracial von neumann algebra generated by unitary elements @xmath99 with the following two properties : @xmath100 for all @xmath101 and @xmath102 for all @xmath103 .
alternatively , we denote by @xmath104 the standard orthonormal basis of @xmath6 , define the translation unitary operators @xmath105 as @xmath106 and define @xmath7 as the von neumann algebra generated by @xmath107 , with @xmath24 being given by @xmath108 for all @xmath109 .
if @xmath110 is a tracial von neumann algebra and @xmath111 is an action of a countable group @xmath5 by trace preserving automorphisms @xmath112 , the _ crossed product _
@xmath113 is the unique tracial von neumann algebra @xmath114 generated by a trace - preserving copy of @xmath115 and unitary elements @xmath99 satisfying the following properties : @xmath116 the map @xmath117 provides an identification @xmath118 and then an explicit realization of @xmath113 as an algebra of bounded operators on the hilbert space @xmath119 . when @xmath8 is a probability measure preserving ( p.m.p . )
action , one considers the corresponding trace preserving action @xmath120 and constructs @xmath121 .
the abelian subalgebra @xmath122 is maximal abelian if and only if @xmath8 is essentially free , meaning that for all @xmath103 the set @xmath123 has measure zero .
when @xmath8 is essentially free , the center of @xmath10 equals @xmath124 , the algebra of @xmath5-invariant functions in @xmath68 .
hence , factoriality of @xmath10 is then equivalent with ergodicity of @xmath8 .
if @xmath5 is an infinite countable group and if @xmath125 is a non - trivial probability space , define the infinite product @xmath126 on which @xmath5 acts by shifting the indices : @xmath127 .
the action @xmath8 is called the _ bernoulli action _ with base space @xmath125 and it is a free ergodic p.m.p . action . more generally , if @xmath5 acts on the countably infinite set @xmath128 , one considers the _ generalized bernoulli action _ @xmath129 given by @xmath130 .
this action is p.m.p . and
it is ergodic if and only if every orbit @xmath131 is infinite . if @xmath125 is non - atomic , essential freeness is equivalent with every @xmath103 acting non - trivially on @xmath128 . if @xmath125 has atoms , essential freeness is equivalent with every @xmath103 moving infinitely many @xmath132 .
an @xmath10-@xmath133-bimodule @xmath53[0ex]{\scriptsize $ m$}{{\mathcal{h}}}\hspace{-0.2ex}\raisebox{-0.4ex}[0ex][0ex]{\scriptsize $ n$}}}$ ] between von neumann algebras @xmath10 and @xmath133 is a hilbert space @xmath134 equipped with a normal unital @xmath4-homomorphism @xmath135 and a normal unital @xmath4-anti - homomorphism @xmath136 such that @xmath137 and @xmath138 commute .
we write @xmath139 instead of @xmath140 .
bimodules should be considered as the ii@xmath1 factor analogue of unitary group representations . based on this philosophy ,
several representation theoretic properties of groups have a ii@xmath1 factor counterpart : amenability , the haagerup property , property ( t ) , etc . to establish the dictionary between group representations and bimodules , take a countable group @xmath5 and put @xmath141 .
whenever @xmath142 is a unitary representation , define the hilbert space @xmath143 and turn @xmath144 into an @xmath10-@xmath10-bimodule by putting @xmath145 .
the trivial representation corresponds to the _ trivial bimodule _
@xmath53[0ex]{\scriptsize $ m$}{{\mathord{\text{\rm l}}}^2(m)}\hspace{-0.2ex}\raisebox{-0.4ex}[0ex][0ex]{\scriptsize $ m$}}}$ ] , the regular representation corresponds to the _ coarse bimodule _
@xmath53[0ex]{\scriptsize $ m { \otimes}1$}{\bigl({\mathord{\text{\rm l}}}^2(m ) { \overline{\otimes}}{\mathord{\text{\rm l}}}^2(m))}\hspace{-0.2ex}\raisebox{-0.4ex}[0ex][0ex]{\scriptsize $ 1 { \otimes}m$}}}$ ] and one defines the notions of _ containment _ and _ weak containment _ of bimodules @xcite in such a way that through the construction @xmath146 these notions exactly correspond to the well known concepts from representation theory .
finally , the _ connes tensor product _ of bimodules ( * ? ? ?
* v.appendix b ) is so that @xmath147 .
[ def.ament ] a tracial von neumann algebra @xmath114 is called _ amenable _
@xcite if the coarse bimodule weakly contains the trivial bimodule .
we say that @xmath114 has _ property ( t ) _
@xcite if any @xmath10-@xmath10-bimodule weakly containing the trivial bimodule , must contain the trivial bimodule .
we finally say that the subalgebra @xmath148 has the _ relative property ( t ) _
@xcite if any @xmath10-@xmath10-bimodule weakly containing the trivial bimodule , must contain the bimodule @xmath53[0ex]{\scriptsize $ n$}{{\mathord{\text{\rm l}}}^2(m)}\hspace{-0.2ex}\raisebox{-0.4ex}[0ex][0ex]{\scriptsize $ m$}}}$ ] .
a linear map @xmath149 is called _ completely positive _ if for every @xmath150 the amplified map @xmath151 maps positive operators to positive operators . in the same way as unitary representations are related to positive definite functions , also bimodules and completely positive maps
form two sides of the same story . whenever @xmath149 is a unital trace preserving completely positive map , the separation and completion of the algebraic tensor product @xmath152 w.r.t .
the scalar product @xmath153 defines a hilbert space @xmath134 that naturally becomes an @xmath10-@xmath133-bimodule . by construction @xmath154
is a cyclic vector , meaning that @xmath155 is dense in @xmath134 , and is a trace vector , meaning that @xmath156 and @xmath157 for all @xmath158 .
every bimodule with a cyclic trace vector arises in this way . whenever @xmath159 is a von neumann subalgebra , we denote by @xmath160 the group of unitaries in @xmath10 that normalize @xmath161 .
a _ cartan subalgebra _
@xmath159 of a ii@xmath1 factor is a maximal abelian subalgebra such that @xmath162 generates @xmath10 .
whenever @xmath8 is a free ergodic p.m.p .
action , @xmath67 is a cartan subalgebra .
we call such cartan subalgebras of _ group measure space _ type .
the relevance of cartan subalgebras in the study of group measure space factors stems from the following theorem .
[ thm.singer ] let @xmath8 and @xmath60 be free ergodic p.m.p .
actions and denote @xmath163 , @xmath164 . assume that @xmath61 is an isomorphism of probability spaces with corresponding trace preserving isomorphism @xmath165 .
then , the following are equivalent .
* @xmath166 is an orbit equivalence : for almost every @xmath11 , we have @xmath65 .
* @xmath167 extends to a @xmath4-isomorphism @xmath168 .
ii@xmath1 equivalence relation _
@xcite on a standard probability space @xmath9 is an equivalence relation @xmath169 with _
equivalence classes such that @xmath170 is a borel subset of @xmath171 and such that @xmath170 is _ ergodic _ and _ probability measure preserving . _ here @xmath170 is called ergodic if every @xmath170-saturated borel set has measure @xmath13 or @xmath37 , while @xmath170 is said to be p.m.p . if every bimeasurable bijection @xmath172 with graph inside @xmath170 , preserves @xmath173 . whenever @xmath8 is an ergodic p.m.p .
action , the _ orbit equivalence relation _
@xmath174 is of type ii@xmath1 .
every ii@xmath1 equivalence relation is of this form , but it is not always possible to choose an essentially _ free _ action implementing @xmath170 ( see @xcite and ( * ? ? ?
* section 7 ) ) .
a variant of the group measure space construction @xcite allows to associate a ii@xmath1 factor @xmath175 to any ii@xmath1 equivalence relation @xmath170 on @xmath9 . by construction @xmath175
contains a copy of @xmath68 as a cartan subalgebra .
every cartan inclusion @xmath159 arises in this way , modulo the possible appearance of a scalar @xmath176-cocycle on @xmath170 .
when @xmath8 is a free ergodic p.m.p .
action , we canonically have @xmath177 .
it is however important to note that both the crossed product and the orbit equivalence relation make sense for non - free actions , but no longer yield isomorphic von neumann algebras .
when @xmath10 is a ii@xmath1 factor and @xmath41 , one defines as follows the amplification @xmath178 . for @xmath179 ,
take a projection @xmath180 with @xmath181 and put @xmath182 . for larger @xmath45 ,
take an integer @xmath150 , a projection @xmath183 with @xmath184 and put @xmath185 .
as such , @xmath178 is well defined up to isomorphism .
one proves that @xmath186 .
fundamental group _
@xmath31 is defined as the set of @xmath41 such that @xmath187 .
the fundamental group of a ii@xmath1 equivalence relation @xmath170 is defined in a similar way . if @xmath188 denotes the associated ii@xmath1 factor , by construction @xmath189 , but this inclusion can be strict ( * ? ? ? * section 6.1 ) .
popa s deformation / rigidity theory , initiated in @xcite , has revolutionized our understanding of ii@xmath1 factors . we explain in this section what kind of deformations popa introduced and how they can be combined with the rigidity given by the relative property ( t ) .
[ def.deform ] a _ deformation of the identity _ on a tracial von neumann algebra @xmath114 is a sequence of normal completely positive maps @xmath190 that are unital , trace preserving and satisfy @xmath191 both group factors @xmath7 and crossed products @xmath113 admit natural deformations of the identity . indeed ,
if @xmath192 is a positive definite function , both @xmath193 extend to normal completely positive maps on @xmath7 , resp .
@xmath113 . if @xmath194 is normalized , i.e. @xmath195 , these maps are unital and trace preserving .
[ ex.haag-freeprod ] if @xmath5 has the haagerup approximation property @xcite , there exists a sequence @xmath196 of positive definite functions converging pointwise to @xmath37 and with @xmath197 for every @xmath150 .
as we discuss below , the corresponding deformation of the identity of @xmath113 plays a crucial role in @xcite .
if @xmath198 is a free product and @xmath199 denotes the natural word length of @xmath64 w.r.t . this free product decomposition and if @xmath200 , then the formula @xmath201 defines a positive definite function on @xmath5 .
if @xmath202 , then @xmath203 pointwise .
the corresponding deformation of the identity is the starting point for @xcite and also this is discussed below .
a second family of deformations of the identity arises as follows .
let @xmath204 be a trace preserving action .
assume that @xmath205 is a deformation of the identity such that @xmath206 for all @xmath64 , @xmath207 .
then , the formula @xmath208 defines a deformation of the identity on @xmath113 . a p.m.p .
action @xmath8 is _ malleable _ if there exists a continuous family @xmath209}$ ] of p.m.p .
transformations of @xmath171 such that for all @xmath210 $ ] , @xmath211 commutes with the diagonal action @xmath212 and such that @xmath213 and @xmath214 is of the form @xmath215 .
the _ bernoulli action _ @xmath216^\gamma$ ] is malleable .
it suffices to construct a continuous family @xmath217}$ ] of p.m.p .
transformations of the square @xmath218 \times [ 0,1]$ ] such that @xmath219 and @xmath220 . this can be done by rotating the square counterclockwise over 90 degrees .
next , one identifies @xmath218^\gamma \times [ 0,1]^\gamma = ( [ 0,1 ] \times [ 0,1])^\gamma$ ] and defines @xmath221 . by construction
, @xmath211 commutes with the diagonal @xmath5-action .
other examples of malleable actions are _ generalized bernoulli actions _ @xmath216^i$ ] given by an action @xmath222 or _ gaussian actions _ given by an orthogonal representation of @xmath5 on a real hilbert space
. [ ex.malleable ] if @xmath211 is a malleable deformation , put @xmath163 and define the corresponding automorphisms @xmath211 of @xmath223 given by @xmath224 . by definition @xmath213 and @xmath225 for all @xmath226 .
view @xmath227 and denote by @xmath228 the trace preserving conditional expectation ( which corresponds to integration w.r.t .
the second variable ) .
the formula @xmath229 defines a continuous family of unital trace preserving completely positive maps with @xmath230 and @xmath231 . when @xmath232 , we get a deformation of the identity on @xmath0 that is at the heart of @xcite .
a variant of the malleable deformation for bernoulli actions is the _ tensor length deformation _ @xcite .
indeed , given a base probability space @xmath125 and a countable set @xmath128 , put @xmath233 and identify @xmath234 with the infinite tensor product @xmath235 , where @xmath236 .
we write @xmath237 . whenever @xmath238 is a subset , we view @xmath239 as a subalgebra of @xmath240 .
we then define for every @xmath200 , @xmath241 then @xmath242 is a well defined unital trace preserving normal completely positive map and @xmath243 when @xmath202 . by construction , @xmath242 commutes with the generalized bernoulli action @xmath244 whenever @xmath222
. let @xmath148 be an inclusion with the relative property ( t ) , see definition [ def.ament ] .
whenever @xmath190 is a deformation of the identity , it follows that @xmath245 converges uniformly in @xmath246 on the unit ball of @xmath133 .
we illustrate this combination of deformation and rigidity by indicating the main ideas behind two of popa s theorems .
[ thm.popatrivfund ] the ii@xmath1 factor @xmath247 has trivial fundamental group .
write @xmath248 , @xmath249 and identify @xmath250 .
the group @xmath5 has the haagerup property and hence admits a sequence of positive definite functions @xmath245 such that @xmath197 for all @xmath150 and @xmath251 pointwise . as in example [ ex.haag-freeprod ] , we get a deformation of the identity on @xmath10 given by @xmath252 .
assume that @xmath148 has the relative property ( t ) and choose @xmath253 . because @xmath197 , we find @xmath207 and a finite subset @xmath254 such that for all @xmath255 in the unit ball of @xmath133 , the @xmath246-distance of @xmath255 to @xmath256 is smaller than @xmath257 . the only obvious subalgebras of @xmath258 with such an approximation property are those that are unitarily conjugate to a subalgebra of @xmath161 , i.e. @xmath259 for some @xmath260 .
popa s _ intertwining - by - bimodules , _ that we recall below , ensures that this feeling is indeed ( almost ) correct .
it is even exactly correct when moreover @xmath148 is a cartan subalgebra .
now observe that @xmath159 has the relative property ( t ) and is a cartan subalgebra .
so , whenever @xmath261 is an automorphism of @xmath10 , the subalgebra @xmath262 still has the relative property ( t ) and the previous paragraph implies that , up to a unitary conjugacy , every automorphism of @xmath10 globally preserves @xmath161 .
this means that every automorphism of @xmath10 induces an automorphism of the orbit equivalence relation of @xmath27 .
a similar statement is true for isomorphisms @xmath263 and therefore , the fundamental group of @xmath10 equals the fundamental group of the equivalence relation , which is trivial by @xcite .
[ thm.strongrigid ] let @xmath5 be a property ( t ) group and @xmath8 any free ergodic p.m.p .
let @xmath29 be any icc group and @xmath264 the bernoulli action .
put @xmath265 . if @xmath66 , then the groups @xmath266 are isomorphic and their actions are conjugate
. put @xmath163 and @xmath267 .
assume that @xmath268 and consider on @xmath269 the tensor length deformation @xmath242 defined after example [ ex.malleable ] . since the subalgebra @xmath7 has property ( t ) , for @xmath270 close enough to @xmath37 , we get that @xmath242 is uniformly close to the identity on the unit ball of @xmath7 . when @xmath271 , the norm @xmath272 is small when @xmath255 can be written as a linear combination of short elementary tensors .
the only obvious short subalgebras of @xmath269 are those that can be unitarily conjugated into either @xmath273 or @xmath274 for some finite subset @xmath275 .
the abelian algebra @xmath274 can never house the property ( t ) algebra @xmath7 and popa indeed manages to prove that @xmath7 must be unitarily conjugate to a subalgebra of @xmath273 .
so we may assume that @xmath276 .
in a second and analytically very delicate part , popa basically proves the following : if a subalgebra @xmath161 of @xmath269 is both abelian and normalized by many unitaries in @xmath273 , then @xmath161 must be unitarily conjugate to a subalgebra of @xmath277 .
this brings us in the situation where @xmath161 can be unitarily conjugated into @xmath277 and @xmath7 can be unitarily conjugated into @xmath273 .
popa proves that automatically both unitary conjugations can be done with the same unitary , yielding isomorphism of the groups and conjugacy of the actions . in @xcite popa
developed a powerful technique to approach the following question : when are two subalgebras @xmath278 unitarily conjugate ?
a detailed explanation and motivation for his method can be found in ( * ? ? ?
* section 5 ) and ( * ? ? ?
* appendix f ) .
so , we are rather brief here .
first consider the case where @xmath279 for some trace preserving action @xmath280 .
every element @xmath90 has a unique _ fourier expansion _
@xmath281 converging in @xmath246 .
we call the @xmath282 the fourier coefficients of @xmath283 .
[ thm.intertwine ] let @xmath284 be a von neumann subalgebra .
then the following two conditions are equivalent .
* there exist projections @xmath285 , a normal unital @xmath4-homomorphism @xmath286 and a non - zero partial isometry @xmath287 satisfying @xmath288 for all @xmath289 . * there is no sequence of unitaries @xmath290 whose fourier coefficients converge to @xmath13 pointwise in @xmath246 , i.e. @xmath291 for all @xmath64 . when @xmath292 is no longer the core of a crossed product , there is no notion of fourier coefficients and their convergence to @xmath13 has to be replaced by the condition @xmath293 for all @xmath88 , where @xmath294 is the unique trace preserving conditional expectation . if @xmath279 , note that @xmath295 .
the first condition in theorem [ thm.intertwine ] is of course not saying that @xmath296 is a unitary satisfying @xmath297 .
the left support projection of @xmath296 lies in the relative commutant of @xmath298 which in concrete applications is usually known but the right support projection of @xmath296 lies in the relative commutant of @xmath299 which is of course a priori unknown , since we do not know @xmath194 .
several techniques based on _ mixing properties _ have been developed to take care of this relative commutant issue and we refer to ( * ? ? ?
* section 5 ) for a more detailed explanation .
in order to construct ii@xmath1 factors with a prescribed fundamental group @xmath300 , you first need a cute construction of a ii@xmath1 factor @xmath10 such that all @xmath301 obviously belong to @xmath31 and then you need a powerful theory to make sure that no other @xmath41 belong to @xmath31 . as an illustration ,
we first briefly explain two constructions that produce ii@xmath1 factors with prescribed countable fundamental group .
let @xmath125 be an atomic probability space and @xmath5 a countable group .
put @xmath302 and define on @xmath9 the following ii@xmath1 equivalence relation : @xmath303 if and only if there exists a @xmath64 and a finite subset @xmath304 such that @xmath305 for all @xmath306 and such that @xmath307 .
whenever @xmath308 , define the subset @xmath309 as @xmath310 .
given @xmath311 , the map @xmath312 changing @xmath313 from @xmath314 to @xmath255 and leaving the other @xmath315 untouched is an isomorphism of the restricted equivalence relations .
hence , @xmath316 belongs to @xmath317 for all @xmath318 .
denote by @xmath188 the ii@xmath1 factor associated with @xmath170 and denote by @xmath300 the subgroup of @xmath32 generated by all the ratios @xmath316 .
we have seen that @xmath319 . in @xcite popa
shows that taking @xmath320 , one has the equality @xmath321 , with one of the ingredients of the proof being the triviality of the fundamental group of @xmath7 .
recall from section [ sec.prelim ] the notation @xmath178 for the amplification of a ii@xmath1 factor @xmath10 . in @xcite dykema and rdulescu
established the following remarkable formula for an infinite free product of ii@xmath1 factors @xmath322 : @xmath323 so , whenever @xmath324 is a countable subgroup different from @xmath40 and whenever @xmath10 is a ii@xmath1 factor , we put @xmath325 and conclude that both @xmath300 and @xmath31 are subgroups of @xmath326 .
as a consequence of their bass - serre theory for free products of ii@xmath1 factors , ioana , peterson and popa @xcite manage to prove that @xmath327 when you take @xmath328 .
a more elementary variant of the previous construction was proposed in @xcite .
both constructions presented so far can be used to produce ii@xmath1 factors with arbitrary fundamental group @xmath300 . however , for uncountable subgroups @xmath324 the resulting ii@xmath1 factors do not have separable predual .
again we start with a construction of a ii@xmath1 factor sowing a number of elements into the fundamental group .
let @xmath329 be a free ergodic action preserving the _ infinite _ non - atomic measure @xmath330 .
put @xmath331 and note that @xmath133 is a ii@xmath42 factor .
whenever @xmath166 is a non - singular preserves sets of measure zero . ]
automorphism of @xmath332 satisfying @xmath333 a.e . , the radon - nikodym derivative between @xmath334 and @xmath330 is @xmath5-invariant and hence constant a.e .
so , @xmath166 scales the infinite measure @xmath330 by a factor that we denote by @xmath335 .
denote the group of all these automorphisms @xmath166 as @xmath336 .
every @xmath337 gives rise to an automorphism @xmath167 of @xmath133 scaling the trace with the same factor @xmath335 .
whenever @xmath338 is a projection of finite trace , put @xmath339 and note that @xmath10 is a ii@xmath1 factor . by construction
we have @xmath340 in the following paragraphs we explain the three main aspects of @xcite : how to get an equality in , how to make sure that @xmath341 is itself a group measure space ii@xmath1 factor and finally , how wild @xmath342 can be . at the end this will give a feeling for the validity of the following theorem .
let @xmath343 be a non - trivial group and @xmath344 an infinite amenable group .
put @xmath345 .
there exists a free ergodic p.m.p .
action @xmath8 such that @xmath0 has a fundamental group of arbitrary prescribed hausdorff dimension between @xmath13 and @xmath37 .
actually , @xmath39 can be any group in the family @xmath346 take @xmath5 of the form @xmath347 and assume that @xmath8 is a free p.m.p .
action with @xmath348 acting ergodically .
let @xmath60 be a free ergodic action preserving the infinite non - atomic measure @xmath349 .
put @xmath350 and consider the action @xmath351 given by @xmath352 if @xmath353 and @xmath354 if @xmath355 .
it is easy to see that @xmath342 equals @xmath356 .
we now make the following assumptions on the action @xmath357 . 1 .
the action @xmath358 is rigid , i.e. @xmath359 has the relative property ( t ) in the sense of definition [ def.ament ] .
the group @xmath29 is amenable .
3 . if @xmath360 is a non - singular isomorphism between the non - negligible subsets @xmath361 satisfying @xmath362 for a.e .
@xmath363 , then @xmath364 for a.e .
@xmath363 . for the following heuristic reasons ,
these assumptions imply that the inclusion in actually is an equality .
consider the ii@xmath42 factor @xmath365 and let @xmath261 be an automorphism of @xmath133 .
the bass - serre theory of @xcite and the relative property ( t ) of @xmath68 imply that after a unitary conjugacy , @xmath261 globally preserves @xmath366 . in a next step , the relative property ( t ) of @xmath68 together with the amenability of @xmath29 roughly implies that @xmath261 globally preserves @xmath367 .
then the third assumption above implies that we may assume that @xmath368 for all @xmath369 and hence that @xmath261 is induced by an element of @xmath370 .
it is highly non - trivial to find group actions @xmath357 satisfying the conditions 1 , 2 and 3 above .
actually , in @xcite we use a baire category argument to prove their existence whenever @xmath348 is an infinite free product @xmath371 with @xmath343 being an arbitrary non - trivial group . replacing @xmath343 by @xmath372
, we may assume that @xmath343 is infinite .
the proof roughly goes as follows : start with arbitrary free ergodic p.m.p .
actions @xmath373 and @xmath374 and view @xmath77 as subgroups of @xmath375 .
by @xcite , there exists an automorphism @xmath376 such that the subgroups @xmath343 and @xmath377 of @xmath375 generate a _ free and rigid _
action of @xmath378 on @xmath9 .
by @xcite there exists @xmath379 such that together with @xmath380 , we obtain a free action of @xmath381 on @xmath9 .
we now start adding copies of @xmath343 acting as @xmath382 for well chosen @xmath383 . at stage @xmath150 , given the free action @xmath384 , the rigidity of the action @xmath385 implies that there are essentially only countably many partial isomorphisms @xmath386 that map @xmath387-orbits into @xmath388-orbits .
by @xcite there exists @xmath389 such that @xmath390 is free w.r.t .
these countably many partial isomorphisms . adding this new @xmath343-action
, we obtain a free action of @xmath391 .
continuing by induction , we get the free action @xmath392 .
we finally prove that there exists an infinite subset @xmath393 containing @xmath394 such that @xmath395 satisfies condition 3 .
since @xmath396 and @xmath385 is rigid , condition 1 is satisfied as well .
take @xmath397 as above , with @xmath29 being infinite amenable .
let @xmath344 be any infinite amenable group . by ( * ?
* lemma 3.6 ) , the restriction of the orbit equivalence relation @xmath398 to a set of finite measure is implemented by a free action of @xmath399 .
so , @xmath341 is a group measure space factor . in @xcite
the notion of an _ ergodic measure _ @xmath400 on the real line is introduced : it is a @xmath401-finite measure on the borel sets of @xmath402 such that for every @xmath403 , the translation @xmath404 of @xmath400 by @xmath283 is either singular w.r.t .
@xmath400 or equal to @xmath400 , and such that denoting @xmath405 the action of @xmath406 on @xmath407 by translation is ergodic . more precisely ,
if @xmath408 is a borel function and if for every @xmath409 we have @xmath410 for @xmath400-a.e .
@xmath411 , then @xmath412 is constant @xmath400-a.e .
the easiest examples of ergodic measures are the lebesgue measure and the counting measure on a countable subgroup of @xmath402 .
in @xcite it is shown that all subgroups of @xmath32 of the form @xmath413 arise as @xmath414 . allowing more general amenable groups
@xmath29 instead of @xmath415 , this can be easily seen as follows .
assume that @xmath416 .
viewing @xmath406 as a closed subgroup of @xmath417 one turns @xmath406 into a polish group .
take a countable dense subgroup @xmath418 and define the additive subgroup @xmath419 as @xmath420 $ ] .
then @xmath421 acts on @xmath422 through multiplication by @xmath423 and we put @xmath424 . define the measure @xmath425 on @xmath402 such that @xmath426 and denote by @xmath427 the lebesgue measure on @xmath402 .
put @xmath428 .
the action @xmath76 given by @xmath429 is measure preserving and ergodic .
one checks easily that @xmath430 .
it is also shown in @xcite ( cf .
@xcite ) that @xmath406 can be uncountable without being @xmath32 and that actually @xmath406 can have any hausdorff dimension between @xmath13 and @xmath37 .
give an intrinsic description of the subgroups of @xmath32 that are of the form @xmath31 where @xmath10 is a ii@xmath1 factor with separable predual .
the only known a priori restriction on @xmath31 is given by ( * ? ? ?
* proposition 2.1 ) : @xmath31 is a borel subset of @xmath32 and carries a unique polish group topology whose borel sets are precisely the ones inherited from @xmath32 .
it is however hard to believe that all polishable borel subgroups of @xmath32 can arise as the fundamental group of a ii@xmath1 factor with separable predual . roughly speaking , the presence of property ( t ) forces fundamental groups to be countable . when @xmath5 is an icc property ( t ) group , connes @xcite proved that @xmath34 is countable and the same method @xcite yields the countability of @xmath39 and @xmath431 , for all free ergodic p.m.p .
actions @xmath8 . for non - icc property ( t )
groups , ioana @xcite could still prove the countability of @xmath431 , but in @xcite we proved that if @xmath5 is a property ( t ) group whose virtual center is not virtually abelian , then @xmath5 admits a free ergodic p.m.p .
action such that @xmath0 is mcduff and in particular , has fundamental group @xmath32 .
we were informed by ershov that such groups exist as quotients of golod - shafarevich groups with property ( t ) .
zimmer @xcite introduced a notion of property ( t ) for ii@xmath1 equivalence relations which is such that for free ergodic p.m.p .
actions , the orbit equivalence relation @xmath174 has property ( t ) if and only if the group @xmath5 has property ( t ) .
using techniques from @xcite we proved @xcite that for @xmath432 , the restriction of the ii@xmath42 relation @xmath433 to a subset of finite lebesgue measure , is a ii@xmath1 equivalence relation with property ( t ) in the sense of zimmer and with fundamental group @xmath32 . in particular
, this ii@xmath1 equivalence relation can not be implemented by a free action of a group .
connes and jones @xcite have given the first examples of ii@xmath1 factors @xmath10 having more than one cartan subalgebra up to conjugacy by an automorphism of @xmath10 .
their construction goes as follows .
take a finite non - abelian group @xmath434 , build the countable group @xmath435 and the compact group @xmath436 that we equip with its haar measure .
consider the action @xmath437 by translation .
finally , let @xmath5 be any non - amenable group .
put @xmath438 and consider the diagonal action @xmath439 which commutes with the bernoulli action @xmath3 .
we obtain a free ergodic p.m.p .
action @xmath440 and hence , the cartan subalgebra @xmath163 of @xmath441 . taking non - commuting elements @xmath442 and defining @xmath443 as being @xmath444 in position @xmath150 ,
one obtains two non - commuting central sequences in @xmath10 .
hence , @xmath10 is a _ mcduff factor _ @xcite , which means that @xmath445 , where @xmath422 denotes the hyperfinite ii@xmath1 factor . choosing any cartan subalgebra @xmath446 , one transports back the cartan subalgebra @xmath447 to a cartan subalgebra of @xmath10 whose associated equivalence relation is not strongly ergodic .
the initial equivalence relation given by @xmath159 is strongly ergodic , so that both cartan subalgebras are non - conjugate by an automorphism of @xmath10 .
ozawa and popa @xcite provided examples of ii@xmath1 factors where one can explicitly see two cartan subalgebras .
we first explain the general procedure and provide a concrete example below .
assume that @xmath5 is a countable group and @xmath448 an infinite abelian normal subgroup .
assume that @xmath448 has the following relative icc property : for every @xmath449 , the set @xmath450 is infinite .
let @xmath451 be a dense embedding of @xmath344 into a compact abelian group @xmath57 .
equip @xmath57 with its haar measure and define the action @xmath437 by translation .
assume that we are given an extension of this action to an essentially free p.m.p .
action @xmath452 .
then , @xmath453 and @xmath454 are non - unitarily conjugate cartan subalgebras of @xmath455 .
an interesting concrete example is given by @xmath456 and its natural affine action on @xmath457 .
we get @xmath458 if @xmath459 , the group @xmath460 has the haagerup property while @xmath461 does not .
if @xmath462 , the group @xmath461 has property ( t ) while @xmath460 does not .
so , neither property ( t ) nor the haagerup property are stable under w@xmath2-equivalence .
since they are stable under orbit equivalence , it follows that the cartan subalgebras @xmath463 and @xmath464 are non - conjugate by an automorphism of the ambient ii@xmath1 factor .
we mentioned in section [ sec.panorama ] that ozawa and popa @xcite established the first and up to now , only uniqueness results for cartan subalgebras up to unitary conjugacy .
recall that a _ profinite _ p.m.p .
action @xmath8 is by definition the inverse limit @xmath465 of a directed system of actions @xmath466 on _ finite _ probability spaces .
[ thm.uniquecartan ] let @xmath467 and let @xmath468 be an ergodic profinite p.m.p .
put @xmath163 and @xmath469 .
it the action is free , @xmath159 is the unique cartan subalgebra up to unitary conjugacy .
if the action is not free , @xmath10 has no cartan subalgebras .
the proof of theorem [ thm.uniquecartan ] consists of two parts .
the group @xmath82 has the _ complete metric approximation property _ ( cmap )
this means that there exists a sequence @xmath470 of finitely supported functions converging to @xmath37 pointwise and such that the maps @xmath471 are _ completely bounded _ with @xmath472 .
then , @xmath473 automatically extends to an ultraweakly continuous map @xmath474 without increasing @xmath475 .
since @xmath468 is profinite , it follows that @xmath10 has cmap as a ii@xmath1 factor : there exists a sequence of finite rank , ultraweakly continuous , completely bounded maps @xmath476 converging pointwise in @xmath477 to the identity and satisfying @xmath472 .
ozawa and popa prove the following general statement : let @xmath10 be a ii@xmath1 factor with cmap and @xmath292 a diffuse amenable ( in particular , abelian ) subalgebra ; denote by @xmath78 the group of unitaries in @xmath10 normalizing @xmath115 .
then , the action of @xmath78 on @xmath115 by automorphisms @xmath50 is necessarily _ weakly compact .
_ we do not explain this notion here but just note that the slightly stronger notion of _ compactness _ means that the closure of @xmath478 inside the polish group @xmath479 is compact .
so far , the reasoning in the previous paragraph works for any profinite action of a cmap group . in the second part of the proof , ozawa and popa
prove the following .
let @xmath468 be any ergodic p.m.p .
action , @xmath480 a subalgebra and @xmath78 a group of unitaries normalizing @xmath115 such that the action @xmath478 on @xmath115 is weakly compact .
then either @xmath78 generates an amenable von neumann algebra or there almost exists a unitary conjugacy of @xmath115 into @xmath68 ( in the sense of theorem [ thm.intertwine ] ) .
the two parts together yield theorem [ thm.uniquecartan ] .
let @xmath467 and let @xmath468 be _ any _ free ergodic p.m.p .
does @xmath481 always have a unique cartan subalgebra up to unitary conjugacy ?
let @xmath467 and let @xmath10 be an arbitrary ii@xmath1 factor .
is it always true that @xmath482 has no cartan subalgebra ?
by @xcite , this is indeed the case if @xmath10 has the complete metric approximation property .
another breakthrough unique cartan decomposition theorem was obtained in @xcite . on the one hand
it is weaker than theorem [ thm.uniquecartan ] since we are only able to deal with group measure space cartan subalgebras , but on the other hand it is stronger since we consider arbitrary group actions .
[ thm.pvcartan ] let @xmath198 be a free product where @xmath348 admits a non - amenable subgroup with the relative property ( t ) and where @xmath483 is any non - trivial group .
let @xmath8 be any ergodic p.m.p .
action and put @xmath121 . if the action is free , @xmath68 is , up to unitary conjugacy , the unique group measure space cartan subalgebra .
if the action is not free , @xmath10 has no group measure space cartan subalgebra .
write @xmath163 .
so , @xmath250 and we denote by @xmath484 the unique fourier expansion of @xmath90 , with @xmath485 for all @xmath64 .
assume that @xmath486 is any other group measure decomposition .
denote by @xmath487 the canonical unitary elements that correspond to this decomposition .
the first step of the proof of theorem [ thm.pvcartan ] consists in transferring some of the rigidity of @xmath5 to @xmath29 .
assume that @xmath245 is a deformation of the identity of @xmath10 .
if instead of @xmath5 , our unknown group @xmath29 would have a non - amenable subgroup @xmath488 with the relative property ( t ) , this would imply that on the unitary elements @xmath489 , the deformation @xmath245 converges uniformly to the identity : for @xmath150 large enough and all @xmath490 , we get that @xmath491 is small .
moreover , the abelian algebra @xmath161 can not contain a copy of the non - amenable algebra @xmath492 so that theorem [ thm.intertwine ] provides a sequence @xmath493 in @xmath488 such that the fourier coefficients of @xmath494 tend to zero pointwise in @xmath477 : for all @xmath64 , we have @xmath495 as @xmath496 .
obviously , we can not prove that @xmath29 automatically has a non - amenable subgroup with the relative property ( t ) .
nevertheless , we have the following transfer of rigidity result .
[ lem.transfer ] let @xmath497 be two crossed product decompositions of the same ii@xmath1 factor with @xmath161 and @xmath498 being amenable .
assume that @xmath5 admits a non - amenable subgroup with the relative property ( t ) and let @xmath245 be a deformation of the identity of @xmath10 .
for every @xmath253 there exists an @xmath207 and a sequence of group elements @xmath499 such that * @xmath500 for all @xmath501 , * the fourier coefficients of @xmath494 tend to zero pointwise : for all @xmath64 , we have @xmath495 as @xmath496 .
lemma [ lem.transfer ] is proven by playing the following positive - definite ping - pong : the formula @xmath502 defines a sequence of positive definite functions on @xmath29 , which in turn define completely positive maps @xmath503 , which in their turn define positive definite functions @xmath504 . by construction , @xmath505 pointwise and hence uniformly on the subgroup of @xmath5 with the relative property ( t ) . from this
, one can deduce the conclusion of lemma [ lem.transfer ] . the starting point to prove theorem [ thm.pvcartan ]
is an application of lemma [ lem.transfer ] to the word length deformation @xmath506 on @xmath507 given in example [ ex.haag-freeprod ] .
the fact that @xmath508 is small means that @xmath283 lies close to a linear combination of @xmath509 , @xmath226 and @xmath64 with @xmath199 not too large .
we refer to such elements @xmath90 as being short .
lemma [ lem.transfer ] provides a sequence of short unitaries @xmath494 .
since @xmath498 is abelian and normalized by the short unitaries @xmath494 , a combinatorial argument implies that the elements of @xmath498 are themselves uniformly short .
but ioana , peterson and popa , in their study of rigid subalgebras of amalgamated free products @xcite , proved that this implies that @xmath498 can be unitarily conjugated into one of the obvious short subalgebras of @xmath510 , namely @xmath511 or @xmath512 .
it follows that @xmath498 can actually be conjugated into @xmath161 since otherwise the normalizer of @xmath498 , i.e. the whole of @xmath10 , would get conjugated in @xmath513 as well .
does the transfer of rigidity lemma hold for arbitrary cartan subalgebras ?
more precisely , let @xmath514 be a cartan subalgebra .
does there exist an @xmath150 and a sequence of unitaries @xmath515 , normalizing @xmath498 and satisfying the same two properties as the unitaries @xmath494 in the formulation of lemma [ lem.transfer ] ? in the affirmative case , one can replace group measure space cartan by cartan throughout the formulation of theorem [ thm.pvcartan ] .
as explained in paragraph [ par.superrigid ] , w@xmath2-superrigidity of an action @xmath8 arises as the sum of oe superrigidity and the uniqueness of the group measure space cartan subalgebra in @xmath0 , up to unitary conjugacy .
this makes w@xmath2-superrigidity theorems extremely hard to obtain . unfortunately
, none of the profinite actions covered by the uniqueness of cartan theorems in @xcite is known to be ( virtually ) oe superrigid .
also actions @xmath516 are not oe superrigid so that we can not directly apply theorem [ thm.pvcartan ] . but theorem [ thm.pvcartan ] can be generalized so that it covers arbitrary actions of certain _ amalgamated free products _ @xmath517 over a common amenable subgroup @xmath344 which is sufficiently non - normal inside @xmath5 ( see @xcite ) . in combination with popa s oe superrigidity for bernoulli actions @xcite or kida s oe superrigidity @xcite , we obtained the following examples of w@xmath2-superrigid actions .
a general statement and more examples , including gaussian actions and certain co - induced actions , can be found in @xcite .
let @xmath518 be an arbitrary group and @xmath519 an infinite subgroup .
put @xmath520 .
then , the bernoulli action @xmath521 is w@xmath2-superrigid . more generally , whenever @xmath222 and @xmath522 is infinite for all @xmath132 , the generalized bernoulli action @xmath129 is w@xmath2-superrigid .
again , because of popa s oe superrigidity for bernoulli actions @xcite , the issue is to prove that @xmath10 has a unique group measure space cartan subalgebra up to unitary conjugacy .
write @xmath524 and @xmath250 .
assume that @xmath486 is another group measure space decomposition .
the first step of the proof of theorem [ thm.io-superrigid ] is similar to the transfer of rigidity ping - pong technique that i explained after lemma [ lem.transfer ] .
denote by @xmath487 the canonical group of unitaries in @xmath525 .
so , ioana considers the @xmath4-homomorphism @xmath526 given by @xmath527 for all @xmath528 , @xmath529 .
since @xmath530 , we rather view @xmath166 as an embedding of @xmath10 into @xmath531 .
but @xmath10 arises from a bernoulli action of a ( relative ) property ( t ) group . as a corollary of theorem [ thm.strongrigid ] ,
every automorphism of @xmath10 is of a special form , i.e. the composition of an inner automorphism and automorphisms induced by a character @xmath532 and by a self - conjugacy of the action @xmath533 . in an amazing technical tour de force ioana manages to generalize popa s methods from automorphisms to embeddings and to give an almost complete picture of all possible embeddings of @xmath10 into @xmath531 , when @xmath10 arises from a bernoulli action of a property ( t ) group .
this picture is sufficiently precise so that applied to the embedding @xmath166 constructed in the previous paragraph , one can conclude that @xmath161 and @xmath498 are unitarily conjugate . , deformation and rigidity for group actions and von neumann algebras .
in _ proceedings of the international congress of mathematicians ( madrid , 2006 ) _ , vol .
i , european mathematical society publishing house , 2007 , p. 445 - 477 . , cocycle and orbit superrigidity for lattices in @xmath538 acting on homogeneous spaces . in _ geometry , rigidity and group actions , _ eds . b. farb and d. fisher .
the university of chicago press , to appear .
arxiv:0810.3630 , on the fundamental group of ii@xmath1 factors and equivalence relations arising from group actions . in _ quanta of maths , proceedings of the conference in honor of a. connes 60th birthday
, clay mathematics institute proceedings , to appear .
arxiv:0810.0706 , circular and semicircular systems and free product factors .
in _ operator algebras , unitary representations , enveloping algebras , and invariant theory ( paris , 1989 ) , _ progr .
math . * 92 * , birkhuser , boston , 1990 , p. 45 - 60 . | we give a survey of recent classification results for von neumann algebras @xmath0 arising from measure preserving group actions on probability spaces .
this includes ii@xmath1 factors with uncountable fundamental groups and the construction of w@xmath2-superrigid actions where @xmath0 entirely remembers the initial group action @xmath3 .
primary 46l36 ; secondary 46l40 , 28d15 , 37a20 .
von neumann algebra , ii@xmath1 factor , measure preserving group action , fundamental group of a ii@xmath1 factor , outer automorphism group , w@xmath2-superrigidity . |
consider a finite composite quantum system described by the complex hilbert space @xmath0 , which is the tensor product of @xmath1 ( @xmath2 ) finite dimensional hilbert spaces @xmath3 of dimension @xmath4 ( @xmath5 ) ; and has dimension @xmath6 .
given any hamiltonian @xmath7 acting on @xmath8 , the thermal equilibrium ( or gibbs ) state @xmath9 for temperature @xmath10 is @xmath11 .
since the quantum system is assumed to be finite , negative temperatures are possible , but they will be disregarded here for reasons of economy although the effects described below are also present for negative temperatures .
+ since the system has components , the question of the entanglement of the gibbs - states arises .
we recall that an arbitrary state @xmath12 ( mixed or pure ) of the composite system is said to be _
separable _ or _ unentangled _ if it can be written as a mixture ( finite convex sum ) of pure product states .
if this is not the case , the state is said to be _
entangled_. + at present there are finite algorithms deciding whether a given state is entangled or not only for two qubits ( @xmath13 with @xmath14 ) ; and for @xmath13 with @xmath15 and @xmath16 .
the entanglement studies presented in the literature have , perforce , been restricted to investigations of the entanglement of the gibbs - state reduced to the possible two - component subsystems , or to the study of particular so called entanglement witnesses or other entanglement monotone functions @xcite . in section
ii , we present some features of the entanglement issue for thermal equilibrium states for the bipartite systems mentioned .
the rest of this introduction collects basic information which is generally valid , i.e. , for any @xmath1 and @xmath17 , and some of which seems to have been overlooked previously .
+ in what follows we often identify the state @xmath12 viewed as a linear positive functional acting on operators @xmath18 and giving their expected - value @xmath19 , with the density operator @xmath12 such that @xmath20 .
if the hamiltonian is a real multiple of the identity , @xmath21 , then @xmath22 for all @xmath10 , or viewed as a functional , the normalized trace which we denote by @xmath23 . since @xmath24 the normalized trace @xmath23 is a separable state .
we assume henceforth that the hamiltonian @xmath25 is not a multiple of the identity , and let @xmath26 be the spectral projection associated with the minimal ( ground - state energy ) eigenvalue @xmath27 whose multiplicity we denote by @xmath28 . for the ground - state @xmath29 , one has @xmath30 where the limit can be taken in various ways . as the limit of matrix elements in any orthonormal basis
you wish ; or @xmath31 where @xmath32 denotes the absolute value of the operator @xmath33 ; etc .
moreover @xmath34 in the same sense , and the map @xmath35 is continuous .
+ for what follows it will be important to observe that the energy @xmath36 as a function of temperature , is a monotone increasing continuous function with @xmath37 and @xmath38 .
a critical or threshold temperature above which entanglement is impossible has been observed in all cited studies .
it is shown in @xcite that for every hamiltonian there exists a finite critical temperature @xmath39 which satisfies : ( i ) @xmath40 ; ( ii ) @xmath41 is separable for every @xmath42 $ ] ; and ( iii ) for every @xmath43 there are entangled thermal states with temperatures in the interval @xmath44 .
we call @xmath39 the _ lower separability temperature_. even if @xmath25 is not a multiple of the identity , it could still be trivial in the sense that it contains no interactions whatsoever between the component subsystems , @xmath45 , @xmath46 acting only on the @xmath47-th component . in this case
@xmath48 is a product - state , hence unentangled and @xmath49 .
a very poor upper bound on @xmath39 was given in @xcite ; and the question of what happens below @xmath39 was posed and left open .
it will be essentially answered in section ii .
+ it is important to stress the fact that the convex set of separable states is closed , and this implies that the set of non - negative temperatures @xmath10 for which @xmath9 is unentangled , is closed . accordingly , the non - negative temperatures @xmath10 for which @xmath9 is entangled is an open set in @xmath50 . + by their very definition , entanglement witnesses or entanglement monotone functions will always provide temperatures for which the thermal states are entangled and these temperatures are lower bounds on @xmath39 .
a particularly simple witness is the energy itself as observed in @xcite .
let @xmath51 , that is the lowest energy expectation value obtainable with an unentangled state .
the infimum is assumed , and since the map @xmath52 is convex - linear , it can be taken over the pure product - states .
+ if for some @xmath53 we have @xmath54 , then @xmath55 is entangled . by the monotone increase of @xmath56 and the intermediate - value theorem
there is a unique @xmath57 such that @xmath58 and all gibbs states with temperatures in @xmath59 are entangled ( in particular @xmath60 is entangled ) . in @xcite , @xmath61 is denoted by @xmath62 and called the entanglement - gap temperature .
it follows that @xmath63 .
it does not follow that gibbs states with temperatures ( immediately ) above @xmath61 are necessarily separable .
although @xmath64 does indeed signal the presence of thermal entanglement at low enough temperatures , its importance should not be overrated .
the correct critical value is : @xmath65 ; for which one can prove , as in @xcite , that : ( i ) @xmath66 , and @xmath67 is separable ; and ( ii ) @xmath68 if and only if @xmath60 is entangled , and in this case all thermal states with temperatures in @xmath69 are entangled .
alternatively , @xmath62 could be defined as the greatest temperature such that all gibbs - states with temperatures below it are entangled ; we call @xmath62 the _ upper entanglement temperature_. obviously , @xmath70 , but one should expect that , in general , @xmath61 can be a rather poor lower bound on @xmath62 .
the following example should serve as illustration .
+ consider two qubits ; and suppose the minimal energy @xmath71 of your hamiltonian is doubly degenerate with ground - state vectors @xmath72 , and @xmath73 where @xmath74 ( resp .
@xmath75 ) is an eigenstate of @xmath76 to the eigenvalue @xmath77 ( resp .
@xmath78 ) for one qubit .
then the ground - state is @xmath79 , and it is entangled ( the partial transpose of @xmath60 has a negative eigenvalue ) ; @xmath80 and thus @xmath81 ; but @xmath82 . the theory of the more mixed thanpartial ordering of states of a quantum system , @xcite , can be put to use in the discussion of entanglement of gibbs states .
it is a result of wehrl and uhlmann ( cf
@xcite ) , that @xmath83 implies @xmath84 , for every unitarily invariant , concave , continuous real - valued functional @xmath85 defined on states .
it is shown in @xcite , that for every such functional @xmath85 , for which @xmath86 implies @xmath87 , there is a critical constant @xmath88 such that : ( i ) if the state @xmath12 satisfies @xmath89 then @xmath12 is separable ; and ( ii ) for every possible value @xmath90 of @xmath85 below @xmath91 there is an entangled state @xmath92 with @xmath93
. there is an analogous version of this for unitarily invariant , convex , continuous real - valued functionals .
thus , every unitarily invariant , continuous real - valued functional which isolates @xmath23 and is either convex or concave , acts as a separability detector and can be used to obtain an upper bound on @xmath39 .
indeed , take such a concave separability detector @xmath85 .
then , @xmath94 is a non - decreasing continuous function for which @xmath95 . by the intermediate - value theorem
there is @xmath96 such that @xmath97 and all gibbs states with temperatures in @xmath98 $ ] are unentangled .
it follows that @xmath99 .
+ many of the available studies , e.g. @xcite , observe that the bipartite subsystem entanglement of thermal states need not be monotone in temperature .
+ the above results allow one to distinguish various entanglement scenarios .
the uninteresting scenario occurs when @xmath100 as happens when the hamiltonians present no interactions whatsoever .
but this scenario is possible even when interactions are present , cf .
section ii .
+ the next scenario is that in which the ground - state is separable , i.e. , @xmath101 , but @xmath102 .
then as temperature increases away from zero , separability is lost at some temperature @xmath103 . for temperatures in the separable segment
@xmath104 $ ] all gibbs states are unentangled , and the segment @xmath105 contains temperatures for which the corresponding thermal states are entangled .
the possibility arises for various closed separable segments ( @xmath104 , [ t_2 , t_3]$ ] , @xmath106 $ ] ) alternating with open entanglement segments ( @xmath107 , @xmath108 ) .
we will present examples for this abnormal scenario in section ii .
+ the other scenarios occur
when the ground - state @xmath60 is entangled , i.e. , @xmath68 .
the normal case is @xmath109 and this is what has been observed in most studies we know of .
although , entanglement usually decreases with increasing temperature , one can have non - monotonous behaviour , cf .
section ii .
the abnormal scenario is @xmath110 , and then there is a temperature @xmath111 with @xmath112 such that for @xmath113 $ ] , @xmath9 is unentangled but there are temperatures @xmath114 for which @xmath115 is entangled . again
the way is open for closed separable segments alternating with open entanglement segments .
examples of this behaviour are given in section ii .
+ consider any state @xmath12 of the composite system and consider the segment joining the normalized trace @xmath23 to @xmath12 , i.e. , @xmath116 with @xmath117 .
since @xmath118 is separable , one will ask in case @xmath119 is entangled , when as @xmath120 increases does one lose separability ?
this question has been analyzed by many authors , notably by yczkowski , horodecki , sanpera , and lewenstein , @xcite , who develop it to obtain a method of estimating the `` size '' of the separable states ; and by vidal , and tarrach , @xcite , who give a virtually complete treatment of the problem . in ref .
@xcite , the modulus of separability of @xmath12 was defined as @xmath121 , whereas the quantity considered by vidal and tarrach is @xmath122 and called by them _ the random robustness of entanglement_. here we only need to observe that @xmath123 , with @xmath124 if and only if @xmath125 is unentangled .
moreover , the upper - semicontinuity property of @xmath126 obtained in @xcite , guarantees that the map @xmath127 is continuous .
this in turn , proves the claims made in [ crit ] about the sets of temperatures where the gibbs state is separable , respectively entangled .
vidal and tarrach , @xcite have computed the modulus of separability for a qubit / qubit ( @xmath128 ) or a qubit / qutrit ( @xmath129 ) ; their beautiful formula is @xmath130 where @xmath131 is the minimal eigenvalue of the partial - transpose of @xmath12 with respect to the qubit . the plots which we will exhibit
show @xmath127 for selected hamiltonians which exemplify the distinct scenarios .
+ the critical values ( or good upper and lower bounds ) are known for some functionals ( e.g. partial ordered eigenvalue sums , von neumann entropy ) from ref .
@xcite and can thus be used to provide upper bounds on @xmath39 as described in [ succ ] .
the map @xmath132 is strictly convex ( cf . ref .
@xcite , p. 47 ) , unitarily invariant and is easily seen to isolate @xmath23 .
furthermore , the critical value for the trace of the square is known to be @xmath133 for two qubits and it lies between @xmath134 and @xmath135 for a qubit / qutrit system ( @xcite ) .
it turns out to be a rather useful separability detector because it is easy to calculate .
although it often provides a rather poor upper bound on @xmath39 , in all the examples to be presented below it gave better bounds than those obtained using the partial eigenvalue sums , or the von neumann entropy .
we denote by @xmath136 , the numerically obtained value of @xmath137 recall [ succ] for @xmath85 equals minus the trace of the square ( using the bound @xmath134 for the qubit / qutrit case ) .
we have not invested a lot of effort in the computation of @xmath138 of [ crit ] .
the temperature value @xmath61 obtained from the numerical , possibly inexact , value of @xmath138 need not necessarily be a lower bound for @xmath62 or even @xmath39 ; but in our experiments this never happened .
the eigenvalues of the hamiltonian counting multiplicities are given as a row - vector @xmath139 . since in our definition of @xmath9
we have incorporated boltzmann s constant in the hamiltonian , the components of @xmath139 have the same dimension as the temperature .
since thermal equilibrium states are invariant with respect to addition of a multiple of the identity to the hamiltonian , we choose @xmath140 .
the eigenvector to the @xmath141-th eigenvalue @xmath142 is listed as a row vector @xmath143 , where the coordinates are with respect to the canonical orthonormal tensor - product basis built from the orthonormal basis @xmath144 of @xmath145 , and @xmath146 of @xmath147 .
+ the hamiltonians to be presented are specifically chosen to exhibit transitions from entanglement to separability below @xmath39 , that is in the interval @xmath148 .
the general idea is , obviously , to choose the eigenvector associated with the non - degenerate ground - state energy to be either separable or entangled and then the eigenvector associated to the first excited state to be , correspondingly , either entangled or separable .
the minimal value of the separability modulus for two qubits is @xmath133 .
+ figure 1 , shows the @xmath149 scenario with a separable segment @xmath104 $ ] followed by an entanglement segment @xmath150 . using the very same eigenvectors as those of fig . 1 , but with @xmath151
, @xmath9 is unentangled for every @xmath152 , that is @xmath49 .
+ figure 2 shows two instances of the standard normal scenario ( @xmath153 ) @xmath109 .
+ in the qubit / qubit system , we have not found the scenario where @xmath154 ( cf . figure 3 for a qubit / qutrit ) .
the minimal value of the separability modulus for a qubit / qutrit systems is @xmath155 .
+ figure 3 shows the @xmath154 scenario ; the entanglement segment @xmath69 is followed by a separability segment @xmath156 $ ] and a second entanglement segment @xmath150 . in this
an further examples not shown here , we have found that once a region of the @xmath139-space is found where the pertinent scenario is present , the values @xmath157 are quite stable with respect to changes in @xmath139 which are large with respect the temperature values .
the inset of figure 3 shows this effect .
moreover , the same eigenvectors used in fig . 3 , but with @xmath158 give @xmath159
. + figure 4 shows the @xmath160 scenario ( cf .
figure 1 for the two qubit system ) with four transitions : @xmath104 $ ] , and @xmath161 $ ] are separable segments alternating with entanglement segments @xmath107 and @xmath162 .
we have introduced two characteristic temperatures @xmath62 and @xmath39 which organize the entanglement behaviour of the thermal state associated to any hamiltonian of an arbitrary composite system . for qubit / qubit and qubit / qutrit systems ,
we have exemplified the possibility of various transitions from entanglement to separability as temperature increases from zero to @xmath39 , above which entanglement is impossible .
one could expect that the features found here will persist and be enhanced as @xmath1 or @xmath17 increase ( although at present there is no manageable criterion to decide when a given state is entangled or not ) .
thus , in general , there will be many separable temperature segments alternating with entangled ones for multipartite systems of higher dimensions .
+ the main motivations for the present study came from certain problems posed in ref . @xcite . from that point of view ,
the main conclusion to be drawn from our findings here are theoretical and concern the results briefly mentioned in [ succ ] .
it was asked in ref .
@xcite : given a separable state @xmath12 which is not pure , does there exist a unitarily invariant , concave , continuous real - valued functional @xmath85 defined for the states of the composite system which isolates the trace and such that @xmath89 ? the answer given here is definitely no ! take any unentangled thermal state @xmath55 such that for some @xmath163 the gibbs state @xmath164 is entangled .
then there can not exist an @xmath85 with @xmath165 because by wehrl s result ( cf .
[ succ ] ) , @xmath166 and by @xcite , the separability of @xmath164 would follow . + 20 m.a .
nielsen : _ quantum information theory . _ ph.d .
thesis , university of new mexico , albuquerque , new mexico , usa 1998 .
arnsen , s. bose , and v. vedral : _ natural thermal and magnetic entanglement in the 1d heisenberg model .
lett . * 87 * , 01790/1 - 4 ( 2001 ) .
d. gunlycke , v.m .
kendon , and v. vedral : _ thermal concurrence mixing in a one - dimensional ising model .
a * 64 * , 042302/1 - 7(2001 ) .
x. wang : _ effects of anisotropy on thermal entanglement .
a * 281 * , 101 - 104(2001 ) .
x. wang : _ entanglement in the quantum heisenberg xy model .
_ phys . rev .
a * 64 * , 012313/1 - 7(2001 ) .
x. wang : _ thermal and ground - state entanglement in heisenberg xx qubit rings .
a * 66 * , 034302/1 - 4(2002 ) .
osborne , and m.a .
nielsen : _ entanglement in a simple quantum phase transition .
a * 66 * , 032110/1 - 14(2002 ) .
x. wang : _ threshold temperature for pairwise and many - particle thermal entanglement in the isotropic heisenberg model .
a .*66 * , 044305/1 - 4(2002 ) .
u. glaser , h. bttner , and h. fehske : _ entanglement and correlation in anisotropic quantum spin systems .
a * 68 * , 032318/1 - 8(2003 ) .
g. toth : _ entanglement witnesses in spin models .
a * 71 * , 010301/1 - 4(2005 ) .
dowling , a.c .
doherty , and s.d .
bartlett : _ energy as an entanglement witness for quantum many - body systems .
_ quant - ph/0408086 ( v3 ) .
wu , s. banyopadhyay , m.s .
sarandy , and w.a .
lidar : _ entanglement observables and witnesses for interacting quantum spin systems .
_ quant - ph/0412099 .
h. fu , a.i .
solomon , and x. wang : _ critical temperature for entanglement transition in heisenberg models .
_ quant - ph/0401015 ( v1 ) .
i. bose , and a. tribedi : _ thermal entanglement properties of small spin clusters .
_ quant - ph/0503170 ( v2 ) .
raggio : _ spectral conditions on the state of a composite quantum system implying its separability .
_ quant - ph/0505044 ( v3 ) .
p.m. alberti , and a. uhlmann : _ stochasticity and partial order .
_ veb deutscher verlag der wissenschaften , berlin 1982 .
m. ohya , and d. petz : _ quantum entropy and its use .
_ springer - verlag , berlin 1993 .
k. yczkowski , p. horodecki , a. sanpera , and m. lewenstein : _ volume of the set of separable states .
a * 58 * , 883 - 892 ( 1998 ) .
g. vidal , and r. tarrach : _ robustness of entanglement .
a * 59 * , 141 - 155 ( 1999 ) . | we revisit the problem of entanglement of thermal equilibrium states of composite systems .
we introduce characteristic , viz .
critical , temperatures and bounds for them marking transitions from entanglement to separability .
we present examples for the various possible thermal entanglement scenarios in bipartite qubit / qubit and qubit / qutrit systems . |
the hard - sphere fluid at a hard wall is a useful reference model for the solid - liquid interface between chemically dissimilar materials .
its simple , but non - trivial , nature has made it a standard reference model to test theories of inhomogeneous fluids , such as integral equation theories and classical density - functional theories . as such , there have been a large number of simulation efforts to study the detailed thermodynamics and structure of this system , both to provide data for the testing of theoretical methods and to provide insight into the generic phenomenology of solid - liquid interfaces .
of particular interest in such studies is the surface free energy , @xmath0 , which measures the work required to create a unit area of interface , and the excess adsorption , @xmath1 , which measures the number of particles in the interfacial region relative to that in a region of equal volume in the bulk . in terms of the single - particle density profile , @xmath2 ,
the excess adsorption is given by @xmath3\ ; dz\ ] ] where @xmath4 is the bulk fluid density and @xmath5 is the cartesian coordinate normal to the surface .
the excess adsorption and the surface free energy are related through the gibbs adsorption equation @xmath6 where @xmath7 is the chemical potential .
a more convenient relationship for use in molecular simulation studies is one derived from the gibbs - cahn procedure@xcite for the excess volume , @xmath8 : @xmath9 the excess adsorption is directly related to the excess volume by the relation@xcite @xmath10 for the hard - sphere / hard - wall system , we can express both @xmath0 and @xmath1 in dimensionless form : @xmath11 and @xmath12 , where @xmath13 and @xmath14 is the hard - sphere diameter . in what follows , we will drop the @xmath15 and assume that all quantities are in dimensionless form . for the hard - sphere fluid / hard - wall system , the values of @xmath0 and @xmath1 are dependent upon the choice of the reference point for measuring distance between the wall and the fluid spheres . in this work
, we will adopt the `` edge - centered convention '' , where the coordinate of the center of a fluid sphere in contact with the wall is @xmath16 . in contract ,
a number of other studies - especially many of the early works - use the `` sphere - centered '' convention with @xmath17 begin the coordinate of the center of such a sphere .
these two conventions give different values of the system volume and other characteristics , but the relationship between them is easy to establish . if we denote the interfacial free energy for a system using the edge - centered and sphere - centered conventions as @xmath0 and @xmath18 , respectively , then we have @xmath19 where @xmath20 is the bulk fluid pressure .
correspondingly , if we denote the edge - centered and sphere - centered values of the excess adsorption as @xmath1 and @xmath21 , respectively , we have @xmath22 over the past four decades , there have been a number of simulation studies focused on the calculation of @xmath0 and @xmath1 for the hard - sphere / hard - wall system .
one of the earliest is that of henderson and van swol@xcite , who calculate @xmath0 using a mechanical definition of the surface tension via the kirkwood - buff equation@xcite @xmath23\ ; dz \label{eq : kirkwood - buff}\ ] ] where @xmath24 and @xmath25 are the normal and transverse components of the pressure tensor and @xmath5 is the direction normal to the wall . in that early work , @xmath0 ( and also @xmath1 )
are calculated at relatively few values of the packing fraction @xmath26 .
the statistical uncertainties in these calculation are also quite high , due both to convergence issues inherent in eq .
[ eq : kirkwood - buff ] and computational power at the time .
more recently , de miguel and jackson using an improved kirkwood - buff algorithm calculated @xmath0 with higher statistical precision than the results of ref . , but the values at the highest packing fractions are in disagreement with most other later studies .
heni and lwen@xcite used a more accurate thermodynamic integration technique with square - barrier and triangular cleaving potentials to determine @xmath0 . while considerably improved over the henderson and van swol results , the relative statistical error in @xmath0 at the highest packing fraction studied ( near fluid - solid coexistence ) was still high ( nearly 10% ) .
this thermodynamic integration method was improved upon by fortini and dijkstra@xcite with significant improvement in the overall statistical error .
the most precise published measurements to date are those of laird and davidchack@xcite and recent calculations by yang , et al.@xcite .
the former were calculated using the gibbs - duhem integration technique@xcite from data for the excess volume ( from which @xmath1 can be easily calculated ) , whereas the latter utilizes a grand - canonical transition matrix monte carlo simulation method .
recently , we have refined the calculations in ref . to be more precise .
these new results are given in the supplemental information.@xcite these simulation data are identical to those in ref .
within the original statistical estimates , but are considerably more precise . for many theoretical and practical applications using the hard - sphere / hard - wall as a reference model , it is useful to have an accurate parametrized form of the wall surface tension @xmath27 and of excess adsorption @xmath28 that has a simple form and accurately accounts for the simulation data . in this work ,
we critically review previous theoretical and empirical expressions for @xmath0 and @xmath1 and propose our own highly accurate parameterization that fits the most accurate simulation data within the statistical error .
the earliest functional form for @xmath0 is that derived from scaled particle theory ( spt):@xcite @xmath29 where @xmath26 is the packing fraction - the fraction of the total volume occupied by the spheres .
. [ eq : spt ] can be obtained from the expression for @xmath18 given in refs . using the spt expression for the pressure and eq .
[ eq : gammabar ] @xmath30 given an expression for @xmath27 , the corresponding expression for the excess adsorption can be obtained using a modified version of eq .
[ eq : excess_v ] @xmath31 for the spt , the expression for @xmath1 is quite simple @xmath32 these spt expressions work well for very low packing fractions , but exhibit significant deviation from the simulation values at intermediate and high packing fractions - see fig .
[ fig : comparison ] .
comparison of the spt and wbii expressions for the hard - sphere / hard - wall @xmath0 with the simulation results of laird and davidchack.@xcite the error bars on the simulation data in this an all subsequent figures represent 95% confidence estimates.,scaledwidth=70.0% ] a more - accurate theoretical expression for @xmath0 can be derived from density - functional theory ( dft ) .
it has been shown@xcite that an approximation for the surface tension of a hard - sphere fluid at a planar hard wall can be obtained within a class of dfts known as fundamental measure theories ( fmt).@xcite in these theories , the surface free energy can be found from the excess free energy density of a binary bulk ( homogeneous ) mixture . for the white bear mark ii ( wbii ) version of fmt one obtains @xmath33
the wbii performs significantly better than the spt expression and shows significant deviations only at the highest packing fractions - see fig [ fig : comparison ] .
this wbii expression will be the starting point of our later parameterization .
the excess adsorption within wbii can be obtained from eq .
[ eq : gamma_partial ] using the carnahan - starling equation of state for hard spheres @xmath34 using @xmath35 and eq .
[ eq : wbii ] in eq .
[ eq : gamma_partial ] gives @xmath36 the spt and wbii theoretical expressions for @xmath0 and @xmath1 make convenient starting points for designing empirical expressions to represent the simulation data . using the spt form as a reference , henderson and plischke@xcite proposed the following empirical functional form for @xmath18 to fit the molecular - dynamics simulation results of ref .
@xmath37 using eq .
[ eq : gammabar ] together with the carnahan - starling equation of state ( eq . [ eq : cs ] ) gives @xmath38 with the corresponding equation for @xmath1 : @xmath39 more recently , urrutia@xcite , also starting with the spt expression , but using the more precise simulation results,@xcite suggested a more accurate expression @xmath40 the corresponding expression for @xmath0 is @xmath41 this expression has the property that the first three virial coefficients for @xmath27 are exact , as discussed in the next section .
urrutia does not provide a corresponding expression for @xmath1 , but one can easily derive one using eq .
[ eq : excess_v ] and the carnahan - starling equation of state @xmath42
another popular parameterization for any thermodynamic quantity is the so - called virial series , where the quantity of interest is expanded in a taylor series with respect to the density , pressure or packing fraction . with respect to the packing fraction
, the virial expansion for @xmath0 can be written as @xmath43 the virial expansion coefficients , @xmath44 , for @xmath18 can be calculated from those for @xmath0 using the usual virial expansion coefficients , @xmath45 , for the pressure ( usually given as an expansion in @xmath4 ) : @xmath46 the coefficients , @xmath45 , are known analytically up to @xmath47 and have been calculated numerically up to @xmath48.@xcite using eq .
[ eq : gammabar ] gives @xmath49 like the virial coefficients for the pressure,@xcite the virial coefficients @xmath50 can be written as a sum of cluster integrals@xcite and the first three coefficients of eq .
[ eq : virial ] are known analytically.@xcite recently , yang , et al.@xcite used monte carlo sampling techniques to evaluate the cluster integrals for @xmath51 to 7 , obtaining approximate estimates for the exact cluster expansion expression for @xmath50 . in this work , the virial coefficients , @xmath44 for @xmath18 were determined .
the virial equation for @xmath18 and the corresponding one for @xmath1 , truncated at @xmath52 , were shown to give very good agreement to their simulation results at low to intermediate packing fractions ( up to about @xmath53 = 0.4 ) .
each of the theoretical and empirical parameterizations in the previous section can be expanded in a virial series .
table i shows the virial expansion coefficients for @xmath0 for @xmath54 to 6 .
( @xmath55 is not shown because of the large statistical error in the monte carlo estimates of the exact coefficient . )
all parameterizations get the exact first two virial coefficients ( @xmath56 and @xmath57 ) correctly , but of the parameterizations presented so far , only that of henderson and plischke and that of urrutia also get the correct third virial coefficient ( @xmath58 ) . [ cols="^,^,^,^,^,^,^",options="header " , ]
the henderson - plischke and urrutia parameterizations use as their starting point the spt expression for @xmath1 . in this work , we begin with the more accurate wbii expression to build an empirical parameterization . in the wbii , the first two virial coefficients of @xmath0 agree with the exact values . to preserve this , one can write @xmath59 where @xmath60 and @xmath61 are fitting parameters . for @xmath62
, the fitting function also gives the exact third virial coefficient .
for this value of @xmath60 , a weighted least - squares fit to the simulation data gives @xmath63 . while this form is a good fit at low to moderate @xmath53 , there are still deviations beyond the statistical simulation uncertainties at the very highest packing fractions
. the high packing fraction data can be well fit by introducing a high power term in the numerator of the first term on the right hand side of eq .
[ eq : fitting ] - a similar approach was used by kolafa , labk and malijevsk in their development of a highly accurate equation of state for hard spheres.@xcite the actual high - power exponent ( @xmath64 ) used is not too important as long as the term does not significantly affect the value of @xmath0 at low to intermediate packing fractions - here we use @xmath65 : @xmath66 to develop our final parameterization , we note that fixing @xmath60 to give the correct third virial coefficient does not lead to an expression that fits the data at intermediate @xmath53 within the error bars . relaxing this condition
gives us our final parameterization and the main result of this work .
[ fig : gamma_diff ] shows the percentage relative deviation of eq .
[ eq : dlr ] from the simulation values , along with the deviations for the other parameterizations considered here .
only the new parameterization fits the data within the statistical error over the full range of @xmath53 : @xmath67 percentage deviation of the parameterizations for @xmath0 considered here from the simulation results.@xcite , scaledwidth=45.0% ] comparison of the spt , wbii and dlr parameterizations for the hard - sphere / hard - wall excess adsorption , ( @xmath1 ) with the simulation results.@xcite the inset shows the high packing fraction data at higher resolution .
, scaledwidth=45.0% ] using the carnahan - starling equation of state and eq .
[ eq : excess_v ] , the corresponding dlr expression for @xmath1 is @xmath68 however , this expression does not quite fit the data for @xmath1 within the error bars - presumably due to slight inaccuracies in the carnahan - starling equation of state ; however , slight modification of the high - order terms gives an expression that is accurate over the whole range of the simulation data within statistical error .
@xmath69 fig .
[ fig : adsorption_comparison ] shows the excess adsorption as a function of @xmath53 for the dlr , spt and wbii expressions ( eq . [ eq : dlr_gamma ] ) together with the simulation results@xcite . to compare the adsorption for all of the methods presented here
, we plot in fig .
[ fig : adsorption_diff ] the percentage deviation from the simulation values as functions of @xmath53 of all of the adsorption expressions presented here .
note that the dlr expression is the only one to capture the simulation data within the statistical error over the entire @xmath53 range studied .
percentage deviation from the simulation data@xcite for each of the parameterizations of the adsorption ( @xmath1 ) presented here.,scaledwidth=45.0% ]
the thermodynamics of an inhomogeneous fluid at a wall can be characterized by the surface free energy @xmath0 and the excess adsorption @xmath1 , two thermodynamic quantities that are related by eq .
[ eq : gibbs ] , the gibbs adsorption theorem . for a hard - sphere fluid at a planar hard wall both @xmath0 and @xmath1
have been measured very precisely in recent computer simulations .
these simulation data can be employed as benchmark data for theories , such as scaled particle theory ( spt ) , or expressions derived from the white bear mark ii ( wbii ) density functional theory in certain limits , or from empirical parameterizations .
each of the expressions for @xmath0 and @xmath1 that we have tested here perform well at low to intermediate packing fractions ( @xmath53 ) of the hard - sphere fluid . however , at higher values of @xmath53 close to freezing there are , however , significant deviations visible in all existing expressions see figs .
[ fig : comparison][fig : adsorption_diff ] .
these deviations are most prominent for the excess adsorption @xmath1 . because explicit expressions for @xmath0 and @xmath1 for the hard - sphere hard - wall model system can be useful in theoretical studies we suggest a new empirical parameterization for these quantities . to this end
we start with the surface free energy @xmath0 .
it is interesting to note that the known virial coefficients for @xmath0 are of limited use only in constructing an accurate expression .
we employ the functional form of the surface free energy from wbii with some added fitting parameters . the key observation , however , is that an additional high power ( in @xmath53 ) term is required to reproduce the correct behavior of @xmath0 at high fluid densities .
our parameterization for @xmath0 is given in eq .
[ eq : dlr ] .
we obtain the parameterization for the excess adsorption in two steps .
first we employ a modified gibbs adsorption theorem , eq .
[ eq : gibbs - cahn_v ] , and the carnahan - starling equation of state to find the functional form of @xmath1 from that of @xmath0 .
because the carnahan - starling equation of state shows small deviation from simulation results at very high @xmath53 , we can improve the agreement of our parameterization with the simulation data by slightly adjusting the higher - order terms , which leads to our final parameterization for @xmath1 ( eq . [ eq : dlr_gamma ] ) . both new parameterizations for @xmath0 and @xmath1 agree within the very precise statistical estimates with simulation data over the entire fluid density range .
an interesting question that arises is whether or not is it possible to obtain also improved expressions for the free energy of a homogeneous hard - sphere fluid and to derive a corresponding density functional for the inhomogeneous hard - sphere fluid based on the requirement of a high - power term ( in @xmath53 ) in @xmath0 and @xmath1 .
bbl acknowledges support from the national science foundation ( nsf ) under grant che-0957102 .
the improved simulation data@xcite were obtained using the alice high performance computing facility at the university of leicester . | the inhomogeneous structure of a fluid at a wall can be characterized in several ways . within a thermodynamic description
the surface free energy @xmath0 and the excess adsorption @xmath1 are of central importance . for theoretical studies
closed expression of @xmath0 and @xmath1 can be very valuable ; however , even for a well - studied model system such as a hard - sphere fluid at a planar hard wall , the accuracy of existing expressions for @xmath0 and @xmath1 , compared to precise computer simulation data , can still be improved . here , we compare several known expressions for @xmath0 and @xmath1 to the most precise computer simulation data .
while good agreement is generally found at low to intermediate fluid densities , the existing parameterizations show significant deviation at high density . in this work ,
we propose new parameterizations for @xmath0 and @xmath1 that agree with the simulation data within statistical error over the entire fluid density range . |
if we ignored gravitational entropy , we would think that the early universe had high entropy because it was uniform and hot @xcite , and that the cold and clumpy universe of today has low entropy .
the evolution of the universe would then violate the second law of thermodynamics , which states that entropy never decreases .
however , gravitational entropy can not be ignored because it increases with clumpiness and ends up dominating the other forms of entropy .
gravitational entropy is usually identified with a quarter the area of a horizon , which suggests that the entropy arises in some way from the information that the horizon hides from us . in the case of the cosmological horizon , the expansion of the universe carries the distant object beyond contact with observers from within the horizon . in the black hole case
, the composition of an object is forgotten once it falls into the hole , and only a few macroscopic variables - mass , spin , and charge - remain .
it would then be easy conclude that gravitational entropy exists only in the presence of the horizon and is proportional to the information hidden by the horizon . however , there are many counter - examples that suggest that gravitational entropy is non - zero even when a horizon is absent .
the gravitational - wave equivalent of the cosmic microwave background should have a temperature of about a kelvin , and should have an entropy roughly in line with that of the cmb .
does a collapsing star , which is @xmath1 away from forming a black hole and does not have a horizon yet , have gravitational entropy ?
black holes are the most lasting objects in the universe .
they can be fully described by three observables - mass , spin and charge and thus have a huge number of indistinguishable configurations . in general , entropy is proportional to the number of indistinguishable configurations in a system . as a gas cloud collapses to form a star , and eventually forms a black hole , entropy increases .
it is not surprising then that the three largest repositories of entropy in the universe are gravitational : the cosmological horizon , the supermassive black holes and the cosmic microwave background .
the local supercluster is gravitationally bound and is expected to collapse to a supermassive black hole @xcite with a schwarzchild radius of @xmath2 light years and a mass @xmath3 @xcite .
the hawking temperature of this black hole would be @xmath4 k @xcite . in @xmath5 years when the cmb cools below @xmath6 k , this black hole will start to evaporate .
heat will flow from the hotter black hole to the colder cosmological horizon producing much more entropy that was lost by the black hole . as the black hole evaporates , the total entropy increases by a factor proportional to @xmath7 , which is much larger than the entropy of the black hole itself : @xmath8 where @xmath9 @xcite .
in due time , nothing will be left other than slowly evaporating supermassive black holes making black hole thermodynamics the physics of the far future ! the evaporation process satisfies the second law of thermodynamics since the heat flows from hot ( black hole ) to cold ( cosmological horizon ) generating entropy .
the emptying of the universe is its approach to equilibrium @xcite .
the state of maximum entropy of a gas in a box is its equilibrium state .
it is also the simplest state the gas can be in , i.e. , uniformly distributed and featureless . in the case of the universe , the simplest , most uniform state is flat space .
empty desitter is maximally symmetric .
it posses a full set of ten symmetries expressed by killing vectors ( one time and three space translations , three rotations , and three boosts ) . however , it has a preferred scale given by the cosmological constant @xmath10 @xcite , which suppresses very long wavelength modes of the weyl tensor also known as gravitational waves .
flat space minknowski space therefore has even higher entropy than desitter because it has the same symmetries , but no suppression factor .
flat minkowski can be thought of as desitter with an extremely small cosmological constant @xmath0 . in this limit , the entropy is divergent @xcite , which introduces an unregulated infinity in the calculation when we use flat space as a background in black hole thermodynamics . for this reason ,
we propose that desitter space - time is a better background than minkowski .
gravitational entropy has to taken into account to explain the evolution of the universe .
we do not know how to compute gravitational entropy in general situations when no horizon is present . a first attempt to compute gravitational entropy inside a star from brown - york @xcite quasi - local quantities
is performed by lundgren _
et al._. @xcite .
flat space has higher entropy than desitter space .
the second law of thermodynamics therefore favors the decay of desitter to flat space , which may happen when quantum effects are included @xcite .
alternatively , if the cosmological constant is mimicked by a scalar field , our universe could cascade through a series of configurations with increasing entropy and decreasing apparent dark energy density , and eventually reach flat space .
10 komatsu , e. _ et al . _
( 2011 ) , apjs * 192 * , 18 .
a summary of all the wmap results is provided on nasa webpage : http://map.gsfc.nasa.gov/. krauss , l. m. and starkman , g. d. ( 2000 ) , astrophys .
j. * 531 * , 22 .
_ et al . _
( 2007),a&a * 476 * , 697 , arxiv:0706.1122 .
bekenstein , j. d ( 1973 ) , phys .
d * 7 * , 2333 . hawking , s. w. ( 1974 ) , nature ( london ) * 248 * , 30 .
gibbons , g. w. & hawking , s. w. ( 1977a ) , phys .
d * 15 * , 2738 .
lundgren , a. p. , bondarescu , m and bondarescu , r. , arxiv:1201.1298 .
s. m. carroll , _ from eternity to here _ ,
dutton adult ( 2010 ) .
davies , p. c. w. ( 1988 ) , classical and quantum gravity * 5 * , 1349 .
davis , t. m. , davies , p. c. w. , and lineweaver , c. h. ( 2003 ) classical and quantum gravity * 20 * , 2753 , arxiv : astro - ph/0305121 .
curir a. & demiaski , m. ( 1979 ) , acta physica polonica * b*10 , 763 .
brown , j.d . and
york , j.w .
jr . , phys .
d * 47 * , 1407 ( 1993 ) .
lundgren , a. , bondarescu , r. , and bondarescu , m. ( 2013 ) arxiv:1304.0594 .
we thank the session chair for emphasizing this point . | we observe the past and present of the universe , but can we predict the far future ?
observations suggest that in thousands of billions of years from now most matter and radiation will be absorbed by the cosmological horizon .
as it absorbs the contents of the universe , the cosmological horizon is pushed further and further away . in time , the universe asymptotes towards an equilibrium state of the gravitational field .
flat minkowski space is the limit of this process .
it is indistinguishable from a space with an extremely small cosmological constant ( @xmath0 ) and thus has divergent entropy . |
microlens parallaxes measure a vector quantity , the `` projected velocity '' @xmath0 , which is the projection of the lens - source relative velocity on the plane of the observer .
another way of writing this quantity is @xmath1 where @xmath2 and @xmath3 are the lens - source relative parallax and proper motion .
it is a useful quantity to measure because it depends only on the kinematic properties of the lens and source , and is independent of the mass .
once it is measured , one can also determine the projected einstein radius " , @xmath4 where @xmath5 is the einstein timescale " ( which is almost always well - measured ) and @xmath6 .
hence , an ensemble microlens parallax measurements can distinguish statistically between different kinematic populations .
microlensing parallax measurements are feasible with _ spitzer _ toward two classes of targets , the large and small magellanic clouds ( lmc and smc ) and the galactic bulge .
the science cases and technical challenges are substantially different for the two classes .
the nature of the microlensing events detected toward the magellanic clouds by the macho @xciteand eros @xcite collaborations is unknown .
they are definitely too infrequent to make up most of the dark matter , but could make up of order 20% .
the basic problem is that , while a few of the lenses have special properties that allow their distances to be determined , for most lenses it is not known whether they are in the milky way ( mw ) halo , the mw disk , or in the clouds themselves . while it would be tempting to use the lenses with known locations to address this question , actually the very properties that allow their distances to be measured predispose them to be in the clouds or the mw , so they do not constitute a fair sample
. it would be better to find a way to choose typical " events , not selected by any characteristic , and determine their position .
_ spitzer _ has already been used to measure the microlens parallax of one lens .
@xcite found that its projected velocity was @xmath7 , which is typical of halo lenses , but much bigger than expected for mw disk lenses and much smaller than expected for smc lenses .
however , they also found that ( with no priors on the location or properties of the lens ) that it could be in the smc with 5% probability .
thus , to draw rigorous scientific conclusions , it will be necessary to obtain parallaxes for at least 3 , and more comfortably 5 , such events .
@xcite conducted a systematic analysis of what could be learned about lenses observed toward the bulge by obtaining parallaxes for an ensemble of them .
they found that such parallaxes greatly enhance microlensing as a probe of the stellar mass function .
for example , comparing the precision of mass estimates ( for lenses assumed to be in the disk in both cases ) , they found that without parallaxes the masses could be individually measured " with a @xmath8 precision of 0.6 dex ( so , essentially no measurement ) whereas with parallaxes , the uncertainty was reduced to 0.2 dex .
of particular interest would be to measure the frequency of disk brown dwarfs ( bds ) .
of course , bds can be directly detected , but only while they are young , so that inferences about their global frequency depend sensitively on both their history of formation and their assumed cooling properties .
it would be nice to constrain their frequency based solely on their mass .
the main characteristic of disk bds from a parallax standpoint is that they have small @xmath9 ( see eq . [ [ eqn : retilde ] ] ) .
for example , a bd with @xmath10 and distance @xmath11kpc would have @xmath12 . in order for an m dwarf with mass @xmath13 to produce the same @xmath9 would require a distance @xmath14kpc .
this is much less likely , but of course not impossible .
so an ensemble of measurements would be needed to estimate to estimate the bd frequency .
in fact , as i will discuss in the [ sec : bulgetech ] , parallax measurements are actually easier for lenses with smaller @xmath9 ( up to a point ) .
spitzer _ is well - placed to probe this key regime .
the technical challenges in the two directions differ significantly , although they are related .
as mentioned , there is already one _ spitzer _ microlens parallax toward the clouds , which is a proof of concept .
however , the source was an @xmath15 clump giant ( i.e. , quite cool , so large @xmath16 ) , which is what permitted excellent photometry with just 2 hours of integration time in each of the four measurements required .
such microlensing sources toward the clouds are extremely rare ( this was the only one detected so far during the entire _ spitzer _
mission ) . to obtain the 35 events that are needed for the science goal ,
requires acceptance of more typical turnoff stars , which are both fainter and bluer , and hence demand integration times that are 1020 times longer .
such integrations were prohibitive for cold _ spitzer _ but might be possible for warm _ spitzer _ when the observatory is under less severe demand .
if we imagine an average of 60 hours per event for 4 events , this would take about 250 hours over 4 years .
note that the individual exposures can not be made any longer than the 27 second exposures for the clump giant because of artifacts from very bright stars .
here there are many challenges .
first , because the bulge is near the ecliptic , it can only be observed for two 38-day periods per year ( in late spring and late autumn ) .
moreover , the autumn window is almost useless for microlensing because no events can be discovered then from the ground .
second , in contrast to the clouds , a large number of events must be monitored to obtain scientifically interesting results , close to 50 and preferably 100 .
third , there are technical challenges related to reconciling ground - based and _ spitzer _ photometry .
this is a classic problem first discussed by @xcite .
basically , unlike trigonometric parallax , microlens parallax is actually a vector , @xmath17 , which is related to observations by @xmath18 where @xmath19 is the measured difference in times of peak of the event as seen from _ spitzer _ and the ground , @xmath20 is the measured difference in the impact parameter , and @xmath21 is the magnitude of the earth-_spitzer _ separation vector projected onto the plane of the sky .
the problem is that while @xmath19 can be robustly measured ( because the time of the peak can be read directly from the lightcurve ) , @xmath22 from each observatory is much harder to measure because it is a fit parameter that is partially degenerate with the amount of blended light .
@xcite showed this degeneracy could be strongly constrained by arranging for the space and ground observatories to have identical cameras ( and so identical blending ) but that is obviously impossible for _ spitzer _ observations .
@xcite evaded this problem by making _
responsible only for measuring the first component of @xmath17 : the events toward the clouds are long enough that the acceleration of the earth toward the sun ( roughly perpendicular to _ spitzer _ ) allowed ground - based measurement of the other component .
however , most events toward the galactic bulge are too short to make this trick work . in order to understand whether the first two challenges can be overcome , i plotted ( not shown ) , all 90 ogle events that peaked during a random 38-day period in spring 2007 , with events plotted in red before they were alerted and in black afterward .
events must be alerted before peak or it is impossible to measure @xmath23 from _ spitzer_. the diagram proved to be quite a mess , but much of this mess is caused by 24 events that were alerted after peak because they are so faint .
the remaining 66 events are shown in figure [ fig : ogle_poss ] . based on experience with _ spitzer _
observations of the smc event , it should be possible to obtain better than 1% photometry in one hour exposures of events that reach @xmath24 .
( the bulge sources are intrinsically bluer than the smc clump giant , but the fields are heavily reddened , so the _ observed _
@xmath16 colors are similar .
) there are 24 such events .
hence , over 5 years , a sample of 120 events could yield measured parallaxes . before addressing the challenge of aligning the photometry of the two observatories
, one should take note of the bottom panel in figure [ fig : ogle_poss ] .
it shows the earth-_spitzer _ separation projected onto the bulge , during each of the 5 38-day intervals of the warm _
mission , red for the first year and blue for the last .
the average value is about 0.6 au , about 2.5 times the leverage available for the smc event @xcite .
this is important because , until the projected separation gets to be the same order as @xmath9 , the signal - to - noise ratio of the measurement scales directly as the projected separation . a possible approach to aligning the ground and _ spitzer
_ photometry would be to obtain @xmath25 band images during the event from the ground with a high - resolution , large aperture telescope .
such measurements would yield an accurate @xmath26 color , independent of blending . while @xmath27 is certainly not identical to _ spitzer _
@xmath28 m , it should be possible to use this color , together with the colors of other stars in the image , to align the two photometry systems based on stars of similar color .
only empirical testing will determine whether this is a practical method .
one potential problem with this approach is that there may not be enough bright bluish stars to perform the alignment for the ( typically bluish turnoff ) stars that dominate the event rate . in this case
, one might be forced to stick with the brighter stars , which comprise 7 stars in the interval shown in figure [ fig : ogle_poss ] .
even if it proves difficult to measure @xmath20 very accurately , it is important to point out that for one of the most interesting applications , the bd mass function , @xmath20 does not have to be measured with great precision .
with typical @xmath29 and @xmath30 , measurements of @xmath20 accurate to 0.1 would be quite adequate .
finally , there is the question of the practical organization of bulge observations .
target - of - opportunity ( too ) observations are notoriously disruptive , and the warm _ spitzer _
mission is likely to have a lower level of staff support to deal with these disruptions than the cold mission does .
however , given the very large number of observations required during each 38-day observing season " , it should be possible apply the concept [ originally developed for the _ space interferometry mission ( sim ) _ ] of a regular too " or rtoo ( as opposed to a disruptive too " or dtoo ) . for an rtoo ,
one plans in advance to make observations of the bulge for a designated set of times , but without specifying exactly what the targets are .
then a day or so before the observations , instructions are uploaded to the spacecraft specifying the targets .
if it really proves possible to measure 24 targets , and each one requires 7 1-hr measurements ( plus 1 baseline measurement that can be done at leisure during the autumn 38-day interval ) , then this amounts to 168 hours of observing over 38 days , or about 18% of observing time during this interval .
all of these observations could be scheduled in advance , without knowing what the targets were . | because _ spitzer _ is an earth - trailing orbit , losing about 0.1 au / yr , it is excellently located to perform microlens parallax observations toward the magellanic clouds ( lmc / smc ) and the galactic bulge .
these yield the so - called `` projected velocity '' of the lens , which can distinguish statistically among different populations .
a few such measurements toward the lmc / smc would reveal the nature of the lenses being detected in this direction ( dark halo objects , or ordinary lmc / smc stars ) .
cool _ spitzer _ has already made one such measurement of a ( rare ) bright red - clump source , but warm ( presumably less oversubscribed ) _ spitzer _ could devote the extra time required to obtain microlens parallaxes for the more common , but fainter , turnoff sources . warm _ spitzer _ could observe bulge microlenses for 38 days per year , which would permit up to 24 microlens parallaxes per year .
this would yield interesting information on the disk mass function , particularly old brown dwarfs , which at present are inaccessible by other techniques .
target - of - opportunity ( too ) observations should be divided into rtoo / dtoo , i.e. , `` regular '' and `` disruptive '' toos , as pioneered by the _ space interferometry mission ( sim)_. lmc / smc parallax measurements would be dtoo , but bulge measurements would be rtoo , i.e. , they could be scheduled in advance , without knowing exactly which star was to be observed . |
in the last few decades , the fields of aerodynamics and aero - elasticity have bloomed significantly due to advances in computation power and experimentation . as a result ,
the implementation of analytical methods has receded slightly . despite their limitations ,
analytical techniques have contributed significantly in the development of the fields of aerodynamics and aero - elasticity .
an important example is the pioneering work of theodorsen @xcite who derived closed form expressions of the aerodynamic loads on thin airfoils in incompressible flow using tools from complex analysis . despite the relative simplicity of the expressions derived by theodorsen
, they have been intensively used by a significant number of researchers to study the aero - elastic stability and control of wing structures ( see for example @xcite ) .
moreover , researchers have used theodorsen s work as a basis to develop more accurate aerodynamic models ( see for example @xcite ) .
it must to be noted that there were also other early works , prior to theodorsen work , that have considered analytical expressions of aerodynamic loads in subsonic flow such as @xcite .
another important example is the relatively recent work of a.v .
balakrishnan that has revived the interest in analytical techniques in aerodynamics and aero - elasticity ( see for example : @xcite ) and in particular the possio equation of aeroelasticity ( a generalization of the classical airfoil equation ) .
balakrishnan has implemented functional analytic techniques intensively to derive and solve singular integral equations from which the aerodynamic loads on thin airfoils in compressible flows can be obtained . additionally , he studied the aero - elastic stability of wing structures using continuum models . besides the work of balakrishnan
, there has been a series of recent mathematical works ( see for example : @xcite ) which have studied the mathematical aspects ( existence , uniqueness , obtaining solutions , and stability ) of different aerodynamic and aeroelastic problems . in conclusion , implementation of analytical techniques , despite their limitation and complexity , is significant in studying different aerodynamic and aeroelastic phenomena .
ground effect is one of the aerodynamic phenomena that received attention from engineers and scientists for many years due to its significance in many applications @xcite .
the ground effect is an aerodynamic phenomenon that can be observed when a flying object is near the ground as the induced lift on the flying object becomes relatively high compared to the lift induced in an open flow . in this paper , approximations of analytical formulas for the aerodynamic lift and moment on a thin airfoil near the ground
are obtained .
the formulas are obtained by first deriving a singular integral equation , also known as the possio equation @xcite .
the integral equation is derived based on linearized compressible potential flow theory , typical section theory , fourier and laplace transforms , and theory of mikhlin multipliers @xcite .
the solvability of the integral equation is then discussed for the steady state case and approximate solutions are then obtained for that case . finally , the divergence speed of a continuum wing structure in steady compressible flow near the ground is calculated based on the derived approximate solutions .
in this work , subsonic flow over a thin airfoil that is located near the ground is considered .
the aim of the study is to derive formulas from which the aerodynamic loads on the airfoil can be calculated .
the airfoil is described by the set @xmath0 \times \ { z = z_{0 } \ } \subset \mathbb{r}_{xz}^{+}$ ] at distance @xmath1 from the ground @xmath2 ( see figure ) , and is subject to an airflow with a free stream velocity of @xmath3 in the positive direction , so that @xmath4 signifies the trailing edge and @xmath5 is the leading edge of the chord . the flow over the airfoil is assumed to be isentropic and following the behavior of ideal gases . additionally , the movement of the airfoil is assumed to be small compared to the free stream velocity .
the flow is assumed to be two dimensional based on typical section theory which is valid for wings with high aspect ratios .
accordingly , the flow over the airfoil is defined by a disturbance potential @xmath6 that satisfies the differential equation @xmath7 where @xmath8 , @xmath9 , @xmath10 , @xmath11 is the free stream speed of sound and @xmath12 is the free stream mach number .
note that the assumption of potential flow neglects the effects of the boundary layer which become significant when the airfoil is very close to the ground .
therefore , the implementation of the theory discussed in this work is assumed to be valid when the airfoil is at an elevation range in which the boundary layer effects are small but the ground effect is still present .
additionally , equation is valid for ranges of mach number between 0 and 0.7 @xcite . to calculate the pressure jump along the airfoil
, a reasonable approximation in terms of the acceleration potential @xmath13 is used where @xmath14the pressure jump term @xmath15 is then defined by @xmath16where @xmath17 moreover , equation is supplemented with the following boundary conditions which describe zero normal velocity at the ground , matching normal flow - structure velocity on the airfoil ( flow tangency ) , zero pressure jump off the wing ( kutta -joukowski condition ) , zero pressure jump at the trailing edge of the airfoil ( kutta condition ) , and vanishing disturbance potential far from the airfoil .
[ eq : boundary conditions ] & = 0 , z=0 , + & = w_a , z = z_0 |x|b , + & a(x , t)=0 , |x|>b + & _ xa(x , t)=0 , + & _ x , z(x , z , t)=0 , where @xmath18 is the downwash or the normal velocity on the airfoil surface .
note that the kutta condition insures the uniqueness of solution to the problem under consideration as shall be discussed in the later sections .
in this section , an equation that relates the pressure jump along the airfoil chord to the airfoil downwash , which is referred to as a possio equation , is derived .
the derivation process starts with applying the laplace transform in the t variable and the fourier transform in the @xmath19 variable to both sides of equation ( [ eq : flow equation ] ) to obtain @xmath20 where @xmath21 , @xmath22 and @xmath23 . rearranging equation ( [ eq : transformed flow equation ] )
results in @xmath24 where @xmath25 and @xmath26 is the reduced frequency . since , @xmath27
, it can be shown that the function @xmath28 is never @xmath29 .
next , the method of images is used to account for the ground effect ( see figure ) .
the method assumes an open flow ( no ground ) and an image of the airfoil to be located at a distance @xmath30 from the ground axis . due to the linearity of equation ,
the solution of the flow problem is obtained by studying the open flow over the airfoil and its image separately .
the separate solutions are given by @xmath31 the normal velocities about the axis @xmath32 are given by @xmath33 the solution to the flow problem should insure the continuity of the velocity field . therefore , @xmath34 .
using the equations in and , the difference in @xmath35 and @xmath36 is expressed in terms of @xmath37 as @xmath38 next , the fourier and the laplace transforms are applied to the acceleration potential in ( [ eq : acceleration potential ] ) and the pressure jump term in ( [ eq : pressure jump ] ) to obtain @xmath39 [ ] using equations and , the pressure jump term @xmath40 is represented in terms of the normal velocity @xmath37 as @xmath41 rearranging the above equation yields : @xmath42 equation is the desired possio equation that relates the pressure jump to the normal velocity at @xmath43 in the fourier domain .
the next step is to obtain an integral equation , based on equation , that relates the pressure jump along the airfoil to the airfoil downwash . to accomplish this , we appeal to the theory of mikhlin multipliers in the next section .
in this section , the necessary definition and theory of mikhlin multipliers used in this work are presented .
the readers are refereed to @xcite for detailed discussions and proofs .
[ def : mikhlin multiplier ] let @xmath44 and @xmath45 be two functions in @xmath46 where @xmath47 and let their fourier transforms @xmath48 and @xmath49 be related by @xmath50 where @xmath51 is @xmath52 and satisfies @xmath53 for all @xmath54 except maybe at @xmath55 . then @xmath51 is called a mikhlin multiplier if @xmath44 and @xmath45 are two functions defined and related as in definition [ def : mikhlin multiplier ] , then there exists a bounded linear operator @xmath56 where @xmath47 such that @xmath57 the detailed proof is discussed in the work of mikhlin @xcite .
the next step is to show that the multiplier @xmath58 in equation ( [ eq : possio equation in fourier domain ] ) , which we denote by @xmath59 , is a mikhlin multiplier .
@xmath60 is a mikhlin multiplier @xmath61 can be written as @xmath62 where @xmath63 and @xmath64 .
it was shown in @xcite that @xmath65 is a mikhlin multiplier .
therefore , it remains to show that @xmath66 is a mikhlin multiplier .
in particular , the function @xmath67 satisfies the estimate @xmath68 the term @xmath69 can be written as @xmath70 for a fixed value of @xmath71 , it can be verified , by calculations that are omitted in this paper , that * @xmath72 is asymptotically equivalent to @xmath73 .
* @xmath74 is asymptotically equivalent to @xmath75 .
* @xmath76 is asymptotically equivalent to @xmath77 . * therefore , @xmath69 is continuous and has convergent limits at infinity , and hence @xmath69 is bounded .
therefore , @xmath66 is a mikhlin multiplier .
based on the previous discussion , there exists a bounded linear operator @xmath78 corresponding to equation such that @xmath79 applying the projection operator @xmath80 $ ] to both sides of equation results in @xmath81 which is the possio equation that relates the pressure jump to the downwash .
the existence and uniqueness of solution to equation depends on the properties of the operator @xmath82 . in the next section , the existence and uniqueness of solution to equation
is discussed for the steady state case .
in this section , the solvability of equation is considered in the case of @xmath83 , which corresponds to the steady state case . setting @xmath83 in equation results in @xmath84 the multiplier @xmath85 corresponds to the hilbert operator @xmath86 which is defined as @xmath87 additionally , the multiplier @xmath88 corresponds to an integral operator @xmath89 that is defined as @xmath90 where @xmath91 note that the parameter @xmath92 plays an essential role in the existence of solution argument .
based on equation , the operators defined in and , and applying the projection operator , we obtain following integral equation : @xmath93 note that @xmath94 vanishes off the chord .
therefore , @xmath95 . additionally , the operators @xmath86 and @xmath96 commute as their product in the fourier domain correspond to a mikhlin multiplier . implementing the above points and the distributive property of operators in equation results in @xmath97 the operator @xmath98 is given by @xmath99 changing the order of integration in results in @xmath100 the term @xmath101 corresponds to the hilbert operator applied to @xmath102 .
we then note that @xmath103 and the positive shifting and scaling are preserved under the hilbert transform .
finally , @xmath104 can be written as @xmath105 where @xmath106 .
the operator @xmath107 corresponds to the finite hilbert operator @xmath108 which is given by @xmath109 the inverse of the finite hilbert operator exists if the kutta condition described by the third equation in ( [ eq : boundary conditions ] ) is imposed .
therefore , the possio equation has the form @xmath110 in this work , the tricomi operator @xmath111 defined by @xmath112 is used as an inversion formula of the finite hilbert transform .
[ inversion ] given @xmath113 $ ] with @xmath114 there exists a solution @xmath115 $ ] for all @xmath116 to the equation @xmath117 moreover , for @xmath47 , any solution @xmath45 has the form @xmath118 with @xmath119 being an arbitrary constant .
this is a classical result due to tricomi and sohngen , @xcite .
next , a bound for the norm of the tricomi operator is obtained in the following lemma .
[ tricomi bounded ] the tricomi operator @xmath111 defined in is bounded from @xmath120 $ ] to @xmath121 $ ] for every @xmath122 and @xmath123 .
[ lemma : t bounded ] @xmath124 can be written as the following : @xmath125 where @xmath126 @xmath127 and @xmath128 next , we estimate the norm of @xmath129 in @xmath121 $ ] given @xmath113 $ ] .
using minkowski s inequality , we have @xmath130 } + |\pi_{0}|\left|\left|\frac{1}{\sqrt{b^2-x^2}}\right|\right|_{l^{r}[-b , b]}.\ ] ] in the following , the definitions of @xmath131 and @xmath132 are unchanged but the terms @xmath133 and @xmath134 are redefined for each subsection .
next , we consider each term in separately the first term @xmath135}$ ] can be estimated by hlder s inequality so that @xmath136}^r\leq\left|\left|\frac{1}{\sqrt{b^2-x^2}}\right|\right|_{l^{q'}[-b , b]}^{\frac{q'}{r } } \left|\left|\pi_{1}\right|\right|_{l^{p'}[-b , b]}^{\frac{p'}{r}}\ ] ] where @xmath137 .
note that @xmath138}$ ] is finite if @xmath139 .
moreover , the term @xmath140}$ ] can be estimated using hlder s inequality @xmath141}^{p ' } , \end{split}\ ] ] where @xmath142 .
the double integral can be shown to be finite for @xmath143 , @xcite . to illustrate that , the double integral is estimated as the following : @xmath144 using polar coordinates , the term @xmath145 can be written as @xmath146 using the substitution @xmath147 , the integration term @xmath148 can be estimated as @xmath149 from the inequality , we have finiteness when @xmath150 and hence @xmath151 . since @xmath152 , we have @xmath153 and @xmath154 .
the term @xmath156 corresponds to the finite hilbert operator applied on @xmath44 . using riesz theorem @xcite , the finite hilbert operator is bounded on @xmath157 $ ] for any @xmath158 .
consequently , we have @xmath155 } \leq c_{p , r } \vert f \vert_{l^{p}[-b , b ] } $ ] for @xmath159 . finally , for the last term @xmath160}$ ]
, we have that @xmath161 is a finite constant since @xmath162}^{p},\\ % & \leq \left(\int_{-b}^{b}\left(\frac{b+y}{b - y}\right)^\frac{p'}{2}dy\right)^{\frac{1}{p ' } } \end{split}\ ] ] where @xmath142 .
note that the first term on the right side of the inequality is finite when @xmath163 which corresponds to the case of @xmath122 .
moreover , @xmath164}$ ] is finite when @xmath165 without imposing any conditions on @xmath132 .
based on studying the boundedness of the three terms in , @xmath111 is bounded when @xmath122 and @xmath153 and that completes the proof .
we next characterize the solution of the equation @xmath166 using the tricomi operator @xmath111 defined in .
[ t ] given @xmath113 $ ] with @xmath167 , there exists a solution to the equation @xmath168 given by @xmath169 $ ] .
moreover , the solution is unique in the class of functions satisfying the kutta condition @xmath170 . starting with the inversion formula from lemma [ inversion ]
, we can express any solution @xmath171 as @xmath172 where @xmath173 is finite since @xmath113 $ ] for @xmath122 , which implies all possible solutions @xmath171 can be expressed as @xmath174 with @xmath175 arbitrary .
now , since we seek a solution @xmath171 satisfying the kutta condition , this implies @xmath176 , and the proof is complete .
* remark * : a solution to the equation and its uniqueness in a class of functions satisfying the kutta condition can also be obtained using the fact that @xmath177 is a bijective isometry from the weighted hilbert space @xmath178 on itself where @xmath179 , @xcite . in this section
, the existence and uniqueness of solution to equation is established through the framework of contraction mappings theory .
applying the tricomi operator to each side of equation yields @xmath180 the equivalence of equation and follows from lemma [ t ] , under the assumption of @xmath181 $ ] and @xmath182 $ ] with @xmath94 satisfying the kutta condition .
note that the tricomi operator defined in maps functions from @xmath183 $ ] to @xmath184 $ ] while the operator @xmath104 maps functions from @xmath185 $ ] to @xmath186 $ ] boundedly .
hence , we establish existence of a solution @xmath94 in @xmath185 $ ] to .
if there exists values of @xmath92 such that @xmath187)}<1 $ ] , then the solution to equation is given uniquely by @xmath188 the expression in can then be rearranged to give @xmath189 equation shows that the pressure jump is a linear combination of two terms .
the first term is equivalent to the pressure jump along the airfoil in open flow ( represented by the first term on the right hand side of equation ) .
moreover , the second term couples the effect of the downwash with the effect of the elevation from the ground .
next , we require a preliminary lemma in order to establish the existence and uniqueness of the solution to the derived possio equation .
[ lemma : plhp bounded ] the operator @xmath104 given by equation is bounded from @xmath120 $ ] to @xmath186 $ ] for any @xmath190 and hence into any @xmath120 $ ] with a bound inversely proportional to @xmath92 .
estimating @xmath191 using hlder s inequality results in the inequality @xmath192 } \vert f \vert_{l^{p}[-2b,2b]}\ ] ] where @xmath193 and @xmath194 with a norm estimated by @xmath195}=\left(\int_{-2b}^{2b}\left|\frac{t}{t^2+c^2}\right|^{q}dt \right)^{\frac{1}{q}}\leq\left|\frac{c}{c^2+c^2}\right|(4b)^{\frac{1}{q}}=\frac{(4b)^{\frac{1}{q}}}{2c},\ ] ] and that completes the proof .
finally , the existence - uniqueness theorem of equation is introduced and proved .
given @xmath196 $ ] with @xmath122 , there exists a value @xmath197 such that for all @xmath198 the possio equation has a unique solution @xmath199 $ ] satisfying the kutta condition .
consider the equivalent equation .
given @xmath200 $ ] , we have @xmath201 $ ] for every @xmath202 .
it is sufficient to show that the operator @xmath203 is a contraction on the space @xmath121 $ ] .
now , @xmath204 is bounded on @xmath121 $ ] since it is a composition of @xmath205\rightarrow l^{2+}[-b , b]$ ] and @xmath206\rightarrow l^{\frac{4}{3}-}[-b , b]$ ] as established in lemmas [ tricomi bounded ] and [ lemma : plhp bounded ] respectively .
the norm @xmath207)}$ ] is estimated using the bounds on @xmath208 and @xmath209 , from lemmas [ tricomi bounded ] and [ lemma : plhp bounded ] respectively , to obtain @xmath210 where @xmath211 , and @xmath212 .
choosing @xmath213 , we have that @xmath207)}$ ] is a contraction , and hence has a unique solution @xmath214 $ ] for any @xmath202 or @xmath215 $ ] .
moreover , @xmath94 clearly satisfies the kutta condition . from lemma [ t ] ,
@xmath94 must satisfy , and on the other hand , a solution to which satisfies the kutta condition must necessarily satisfy .
this establishes that @xmath94 is the unique solution to in the class of @xmath184 $ ] functions satisfying the kutta condition .
next , an approximate solution to is obtained . from equation and and changing integration order ,
the operator @xmath216 can be written as : @xmath217 the expression in brackets in corresponds to applying the tricomi operator on @xmath218 .
the function @xmath218 can be linearly approximated by @xmath219 and that reduces equation to @xmath220 but @xmath221 and @xmath222 .
therefore , equation is simplified to be : @xmath223 where @xmath224 and @xmath225 . applying @xmath226 and @xmath227 on equation respectively
results in the set of equations @xmath228 where @xmath229 , @xmath230 , @xmath231 , and @xmath232 . solving equations and for @xmath233 and @xmath234 results in @xmath235 finally , substituting in and solving for @xmath94 yields @xmath236.\ ] ] for the steady state case ,
the airfoil downwash @xmath237 is a function of the angle of attack @xmath238 . in particular , @xmath239 note that the downwash given in is not a function of @xmath19 and therefore , only @xmath240 and @xmath241 are required to compute the pressure jump for the steady state case with a downwash given by .
the pressure jump is calculated to be @xmath242.\ ] ] now , we calculate the aerodynamic lift @xmath243 on the airfoil using the formula @xmath244 using equation , we obtain the following expression for @xmath243 @xmath245 note that in equation ( [ eq : lift ] ) , the first term on the right hand side represents the aerodynamic lift assuming an open flow and the second term introduces the aerodynamic lift due to the ground effect .
next , the aerodynamic moment @xmath246 on the airfoil is calculated using the following equation @xmath247 where @xmath248 is the location of the center of rotation of the airfoil .
the aerodynamic moment is given by @xmath249 similar to equation , equation separated the open flow moment ( first term on the right hand side ) from the moment due to the ground effect .
for the steady state incompressible case @xmath250 , the present problem can be studied using complex analytic methods which include joukowski transformations to construct flow potential and contour integration to obtain the aerodynamic loads @xcite . despite the relative ease of obtaining flow potential using joukowski transformations , up to the knowledge of the authors ,
there are no closed expressions for the aerodynamic loads in the present case . in this section , the divergence speed @xmath251 of a wing structure with thin airfoil ( see figure in subsonic flow
is calculated @xcite .
the steady state configuration of the wing is governed by the equations @xmath252 where @xmath253 , @xmath254 is the span of the wing , @xmath255 is the torsion angle and corresponds to the angle of attack in , @xmath256 is the deflection of the wing , @xmath257 is the torsional stiffness of the wing , and @xmath258 is the bending stiffness .
the clamped - free boundary conditions satisfied by @xmath259 and @xmath238 are given by @xmath260 the divergence speed is obtained by solving the eigenvalue problem of finding the minimum free stream velocity that satisfies system and the boundary conditions .
the divergence speed is now calculated as the following .
the moment expression is rewritten as @xmath261 where @xmath262 then , the governing equation of the torsion is given by @xmath263 the general solution to equation is @xmath264 where @xmath265 and @xmath266 are positive constants that depend on the boundary conditions .
for the boundary conditions , the following relation is obtained @xmath267 the lowest speed that satisfies equation ( divergence speed ) is then given by @xmath268 the present work can be extended to obtain lift and moment expressions for the incompressible transient case @xmath269 and consequently a flutter analysis can be conducted in a fashion similar to @xcite .
in this paper , the problem of subsonic flow of thin airfoil near the ground is studied .
a singular integral equation , namely possio equation , is derived to relate the pressure jump over the airfoil to its downwash .
the existence and uniqueness of the new possio equation is verified for the case of steady subsonic flow .
moreover , an approximate solution of the derived possio equation in this particular case is established and closed form expressions of aerodynamic loads and the divergence speed are obtained .
g. avanzini and a. de santis , _ beyond possio equation : the legacy of camillo possio to flight dynamics and hydrodynamics_. in : ceragioli f. , dontchev a. , futura h. , marti k. , pandolfi l. ( eds ) system modeling and optimization .
csmo 2005 .
ifip international federation for information processing , vol 199 .
springer , boston , ma , 2006 .
i. chueshov , e. h. dowell , i. lasiecka , and j. t. webster , _ nonlinear elastic plate in a flow of gas : recent results and conjectures _ , applied mathematics and optimization , vol .
73 , no . 3 , pp .
475500 , 2016 .
g. n. lance .
_ the kernel function of the integral equation relating the lift and downwash distributions of oscillating finite wings in subsonic flow _ , journal of the aeronautical sciences , vol .
9 , pp . 635 - 636 , 1954 .
watkins , h.l .
runyan , and d.s .
woolston , _ on the kernel function of the integral equation relating lift and downwash distributions of oscillating finite wings in subsonic flow _ , naca report 1257 , 1953 . | in this paper , the problem of compressible flow over a thin airfoil located near the ground is studied .
a singular integral equation , also known as possio equation @xcite , that relates the pressure jump along the airfoil to its downwash is derived .
the derivation of the equation utilizes laplace transform , fourier transform , method of images , and theory of mikhlin multipliers .
the existence and uniqueness of solution to the possio equation is verified for the steady state case and an approximate solution is obtained .
the aerodynamic loads are then calculated based on the approximate solution .
moreover , the divergence speed of a continuum wing structure located near the ground is obtained based on the derived expressions for the aerodynamic loads .
department of mathematics and statistics + american university of sharjah + sharjah , uae + e - mail : mohamedserry91100gmail.com + e - mail : atufaha100aus.edu |
the study of superconducting and superfluid systems has recently attracted wide attention , among other things because of the realization of ultracold fermi gasses in optical lattices @xcite .
these systems can be considered as quantum simulators that can be used for probing fundamental problems in condensed - matter physics @xcite , for instance the search for exotic new phases in strongly magnetized superconductors .
ultracold fermi gasses offer important advantages over conventional superconductors , mainly because of their extensive tunability . in a superconductor ,
the number of spin - up and spin - down electrons is equal and the interaction strength is fixed . in ultracold fermi gasses , one can not only tune the interaction strength by use of feshbach resonances @xcite , but also the population imbalance can be freely adapted .
this experimental freedom has led to the study of a variety of new phenomena in imbalanced ultracold fermi gasses @xcite .
one fundamental question , that is still not settled , concerns the nature of the ground state of an imbalanced fermi gas .
when population imbalance between the spin - up and spin - down components is introduced into these systems , complete pairing is no longer possible .
clogston and chandrasekhar suggested that above a critical imbalance , the superfluid system would undergo a transition into the normal state @xcite .
this effect has been observed experimentally by the mit @xcite and rice @xcite groups .
however , their observations were not in exact agreement , and there still exists some controversy @xcite about the exact nature of the phases of the superfluid system at high levels of imbalance . in 1964 fulde and ferell @xcite and independently larkin and ovchinnikov @xcite proposed that a superfluid system can accommodate population imbalance , by making a transition into a state with a finite center - of - mass momentum ( and thus a spatially modulated order parameter ) .
this state is the so - called fulde - ferell - larkin - ovchinnikov - state ( fflo - state ) .
recently , there has been an ongoing theoretical search for this exotic state @xcite . in 1d
, the fflo - state was predicted to exist in superconducting systems @xcite .
this work was confirmed numerically @xcite and elaborated further through theoretical studies of the ground state and the phase diagram of a 1d fermi gas @xcite .
furthermore it has been shown that the 1d analogue of the fflo - state is stable in a large section of the bcs - bec crossover phase diagram @xcite , compared to the case of a 3d fermi gas .
although this 1d fflo - state has not yet been observed directly , a recent paper reports the experimental observation of density profiles that agree quantitatively with theoretical predictions at low temperature @xcite . in three dimensions however , the experimental observation of the fflo - state has so far remained elusive .
one of the main reasons for this is that the fflo - state in three dimensions only occurs in a tiny section of the bcs - bec crossover phase diagram @xcite .
this then begs the question , is there a way to stabilize the fflo - state in a 3d fermi gas ?
the purpose of this paper is twofold : first we develop a path - integral description for a superfluid fermi gas which can accommodate the fflo - state , and second we propose a method to stabilize the fflo - state through an optical potential . in two recent papers
it was suggested to stabilize the fflo - state by the use of a 3d optical lattice @xcite . in this paper
we investigate the stabilizing effect of a 1d optical potential in order to investigate the interplay between the wavevector of the fflo - state and the wavevector of the laser which creates the optical potential . in the present work
, the 1d optical potential provides a periodic modulation in one direction .
we emphasize that we do not look at the fflo - state in a one - dimensional gas @xcite , but in a 3d gas with a superimposed one - dimensional periodic potential . in section [ the free fermi gas ]
we describe the fflo state in an imbalanced fermi gas in 3d within the path - integral framework . as a test for the correctness of this description ,
we calculate the zero temperature phase diagram for this system and compare our findings with recent theoretical results @xcite . in section [ modelling a one - dimensional external potential ]
we investigate two models to account for the effect of a one - dimensional optical potential .
we show that the presence of such a potential leads to a substantial increase of the stability region of the fflo - state .
finally in section [ conclusions ] we draw conclusions .
the partition sum of an imbalanced fermi gas in 3d can be written as a path integral over the fermionic fields @xmath0 and @xmath1 : @xmath2{l}\mathcal{z}={\textstyle \int } \mathcal{d}\bar{\psi}\mathcal{d}\psi \text { } \exp \left ( -{\displaystyle \sum \limits_{\mathbf{k},n } } { \displaystyle \sum \limits_{\sigma } } \bar{\psi}_{\mathbf{k},\omega_{n},\sigma}\left ( -i\omega_{n}+\mathbf{k}^{2}-\mu_{\sigma}\right ) \psi_{\mathbf{k},\omega_{n},\sigma}\right . \\
\qquad \left .
-\dfrac{g}{\beta l^{3}}{\displaystyle \sum \limits_{\mathbf{k},n } } { \displaystyle \sum \limits_{\mathbf{k}^{\prime},n^{\prime } } } { \displaystyle \sum \limits_{\mathbf{q},m } } \bar{\psi}_{(\mathbf{q}/2)+\mathbf{k},\omega_{m}+\omega_{n},\uparrow}\bar { \psi}_{(\mathbf{q}/2)-\mathbf{k},\omega_{m}-\omega_{n},\downarrow}\psi_{(\mathbf{q}/2)-\mathbf{k}^{\prime},\omega_{m}-\omega_{n^{\prime}},\downarrow}\psi_{(\mathbf{q}/2)+\mathbf{k}^{\prime},\omega_{m}+\omega_{n^{\prime}},\uparrow}\right ) .
\end{array } \label{toestandssom begin}\ ] ] here @xmath3 is the wavevector , @xmath4 are the fermionic matsubara frequencies and @xmath5 denote the two different hyperfine states .
furthermore , @xmath6 is the inverse temperature given by @xmath7 , @xmath8 is the lateral size of the system , the chemical potential of a particle with spin @xmath9 is denoted by @xmath10 and @xmath11 is the renormalized interaction strength .
we use units such that @xmath12 .
this partition sum ( [ toestandssom begin ] ) can be made more tractable by introducing the hubbard - stratonovic transformation , which decouples the fourth - order interaction term into second order terms by introducing two auxiliary bosonic fields @xmath13 and @xmath14 , interpreted as the pair fields .
as a first approximation , only the saddle - point is taken into account in the path integral over the bosonic fields . to describe the fflo - state , we propose to use a saddle point at which the atomic pairs have a finite wavevector @xmath15:@xmath16 by using this particular form of the saddle - point , we choose to describe the ff - state , which has an order parameter given by a plane wave @xmath17 .
it is also possible to describe the lo - state , which is a superposition of two plane waves with wavevector @xmath15 and @xmath18 . in this paper
, we will not consider the lo - state .
for the remainder of the article , the ff - state is referred to as the fflo - state .
when @xmath15 is set equal to zero in ( [ zadelpunt ] ) , the description of the normal superfluid is recovered @xcite . in expression ( [ zadelpunt ] ) , the prefactor @xmath19 ensures that @xmath20 has units of energy . using ( [ zadelpunt ] ) the fermionic fields can be integrated out in expression ( [ toestandssom begin ] ) for the partition function , leading to an effective action @xmath21 , through @xmath22 + \dfrac{\beta l^{3}}{g}\left \vert \delta \right \vert ^{2}\right ) = \exp \left ( -\mathcal{s}_{sp}\right ) , ~ \label{saddle - point partition sum}\ ] ] with @xmath23 the inverse nambu propagator which is given by@xmath24{cc}-i\omega_{n}+(\mathbf{q}/2+\mathbf{k})^{2}-\mu_{\uparrow } & -\delta \\ -\delta^{\ast } & -i\omega_{n}-(\mathbf{q}/2-\mathbf{k})^{2}+\mu_{\downarrow}\end{array } \right ) .\ ] ] expression ( [ saddle - point partition sum ] )
can be simplified further by performing the sum over the matsubara frequencies .
also , it is useful to express the results as a function of the total chemical potential @xmath25 and the imbalance chemical potential @xmath26 . furthermore , the interaction between particles is modeled with a two - body contact potential @xmath27 .
the renormalized interaction strength @xmath11 can then be written as follows @xcite : @xmath28 where @xmath29 is the 3d s - wave scattering length . as a final step the continuum limit is taken , and a thermodynamic potential is associated with the effective action @xmath30 .
this then results in@xmath31 -\xi_{\mathbf{q},\mathbf{k}}-\frac{\left \vert \delta \right \vert ^{2}}{2k^{2}}\right ) -\frac{\left \vert
\delta \right \vert ^{2}}{8\pi \left ( k_{f}a_{s}\right ) } , \label{thermodynamic potential free fermi gas}\ ] ] where the following notations were introduced @xmath32{l}\xi_{\mathbf{q},\mathbf{k}}=k^{2}-\left ( \mu-\dfrac{q^{2}}{4}\right ) \\ e_{\mathbf{k}}=\sqrt{\xi_{\mathbf{q},\mathbf{k}}^{2}+\left \vert \delta \right \vert ^{2}}\\ \zeta_{\mathbf{q , k}}=\zeta+\mathbf{q\cdot k}\end{array } \right . .
\label{korte notaties vrij fermi gas}\ ] ] the resulting form of the thermodynamic potential ( [ thermodynamic potential free fermi gas ] ) has a similar form as the original result for the homogeneous 3d fermi gas derived by iskin and s de melo @xcite and coincides with it for @xmath33 . in order to test the correctness of ( [ thermodynamic potential free fermi gas ] ) , the thermodynamic potential is used to calculate the zero temperature phase diagram of an imbalanced fermi gas in 3d .
this can be done for a fixed number of particles or for fixed chemical potentials . to transform between these two descriptions , the number equations , given by@xmath34 have to be solved .
as an example , the phase diagram is calculated for a fixed density @xmath35 and for a fixed imbalance chemical potential @xmath36 . to do this , the first number equation ( [ first number equation ] )
is solved , given values of @xmath37 and @xmath38 , to determine the chemical potential @xmath39 .
the set of values @xmath40 is then substituted in the free energy @xmath41 .
the minima in the free energy landscape determine which state is the ground state of the system , for a given imbalance chemical potential @xmath36 and a given interaction strength @xmath42 .
there are three local minima that can be identified in the free energy landscape : the bcs - state ( spin - balanced superfluid ) with @xmath43 , @xmath44 , the fflo - state with @xmath43 ,
@xmath45and the normal state with @xmath46 figure [ contours_geenzeta.eps ] shows that the fflo - state can indeed be the ground state of an imbalanced fermi gas in three dimensions ( at zero temperature ) . in this figure ,
the free energy of the system is shown , as a function of the bandgap @xmath20 and the wavevector @xmath15 , relative to the fermi energy @xmath47 and the fermi wavevector @xmath48 respectively .
[ h ] contours_geenzeta2.eps for relatively small values of the imbalance chemical potential @xmath36 , the system is in the bcs ground state ( a ) .
when @xmath36 increases , the system undergoes a first order transition into the fflo - state ( b ) .
when the imbalance chemical potential increases further , the system continuously goes over into the normal state ( c ) .
figure [ contours_geenzeta.eps ] only shows these phase - transitions for one specific value of the interaction strength , near the bcs - limit @xmath49 .
however , since the first number equation is used to calculate the value of the chemical potential @xmath39 , the description is also valid for the complete bcs - bec crossover regime .
it must be noted that the mean field approximation breaks down in the unitarity limit .
however , the fflo - state is expected to form only in the bcs region of the bcs - bec crossover , where we expect the mean field description to give qualitatively correct results .
the phase diagram of the system is shown in figure [ fd_vrijfermigas ] .
this diagram shows that , theoretically , the fflo - state can be formed in an imbalanced fermi gas in 3d , but since it only occurs on a tiny section of the phase diagram , it may be hard to observe this state experimentally .
our results coincide with recent theoretical results @xcite .
[ h ] fd_vrijfermigas.eps
as shown in section [ the free fermi gas ] , the problem with detecting the fflo - state in an imbalanced fermi gas in 3d is that it exists only in a relatively small section of the bcs - bec phase diagram . in this section ,
we describe the 3d imbalanced fermi gas in a 1d optical potential .
there are two main reasons why such a potential can stabilize the fflo - state .
the first reason is that in an imbalanced fermi gas in 3d , the fflo - state can have a wavevector @xmath15 in an arbitrary direction .
this freedom of choice leads to low - energy bosonic excitations , or goldstone modes , which render the fflo - state unstable . in the presence of a 1d optical potential however
, it will be energetically favorable for the fflo - state to form in the direction of the optical potential .
this will limit the choice for the wavevector @xmath15 to just one value , thus suppressing the goldstone modes , which is expected to stabilize the fflo - state .
the second reason is that the optical potential will enhance the 1d modulation of the fflo - order parameter .
we therefore expect the enhancement to be largest when the wavevector of the fflo - state is equal to the wavevector of the optical potential .
the present mean - field treatment does not include the effect of excitations such as the goldstone modes , but it does include the energy lowering of the modulated order parameter due to the optical potential .
in this section we propose two approaches of modeling a 3d imbalanced fermi gas in a 1d optical potential . in both approaches ,
the optical potential is described by using a modified dispersion relation . in section [ anisotropic masses ]
we model the optical potential by introducing an anisotropic effective mass in the direction of the potential @xcite ( from here on this is supposed to be the @xmath50-direction ) .
this approximation is valid when the fermi energy of the system lies near the bottom of the lowest bloch band , i.e. in the case of low density or a short - wavelength optical potential . in section [ bloch - dispersion ] ,
we model the optical potential by treating the full lowest bloch band in the tight binding approximation @xcite .
this model is valid when the fermi energy lies in the lowest bloch band ( otherwise more bands have to be considered ) , but contrary to the first case , it does not need to lie at the bottom of the band . in both section
[ anisotropic masses ] and section [ bloch - dispersion ] the optical potential is supposed not to forbid tunneling , as this would inhibit the formation of the fflo - state .
this implies that we will treat the imbalanced fermi gas in 3d in a 1d optical potential as a three dimensional system with a one - dimensional periodic modulation . in this section
, the 1d optical potential is modeled through the use of a modified effective mass of the fermionic particles in the direction of the optical potential @xmath51 where @xmath52 is the effective mass of the particles in the @xmath50-direction . here and for the remainder of the paper
, @xmath53 denotes the magnitude of the in - plane wavevector .
this is the wavevector which lies perpendicular to the laserbeam which creates the 1d optical potential .
the derivation of the thermodynamic potential for this case is analogous to the derivation in section [ the free fermi gas ] . in the present derivation
however , it is assumed that the fflo - state will form in the @xmath50-direction @xmath54 , because this is energetically favorable to other directions , due to the anisotropy introduced through ( [ dispersie anisotroop ] ) .
the resulting thermodynamic potential , in the limit for temperature going to zero , is given by@xmath55 -\xi_{\mathbf{k}}-\frac{\left
\vert \delta \right \vert ^{2}}{2\left ( k^{2}+\dfrac{k_{z}^{2}}{2m_{z}}\right ) } \right ) -\frac{\left \vert \delta \right \vert ^{2}}{8\pi \left ( k_{f}~a_{s}\right ) } , \label{thermodynamic potential anisotropic masses}\ ] ] with the modified notations @xmath32{l}\xi_{\mathbf{q},\mathbf{k}}=k^{2}+\dfrac{1}{2m_{z}}\left ( k_{z}^{2}+\dfrac{q^{2}}{4}\right ) -\mu \\ e_{\mathbf{k}}=\sqrt{\xi_{\mathbf{q},\mathbf{k}}^{2}+\left \vert \delta \right \vert ^{2}}\\ \zeta_{\mathbf{q , k}}=\dfrac{1}{2m_{z}}k_{z}~q-\zeta \end{array } \right . .\ ] ] the number equations are still given by ( [ first number equation ] ) and ( [ second number equation ] ) , but the density @xmath35 has changed to @xmath56 because of the modified dispersion relation ( [ dispersie anisotroop ] ) . in the limit @xmath57 the thermodynamic potential ( [ thermodynamic potential anisotropic masses ] ) and the density ( [ dichtheid ] )
converge to the corresponding expressions in the case of an imbalanced fermi gas in 3d , described in section [ the free fermi gas ] .
expression ( [ dichtheid ] ) implies that when the effective mass @xmath52 changes , the density @xmath35 changes with it
. it would be interesting however , to compare the phase diagrams for fermi gasses with different effective masses at equal density .
this can in fact be achieved , because we have found that the thermodynamic potential of the system with an effective mass @xmath52 can be rescaled to the thermodynamic potential of the system with effective mass @xmath58 ( [ thermodynamic potential free fermi gas ] ) , using the following scaling relation@xmath59 from a theoretical point of view , this rescaling property is time - saving for calculations and gives a deeper insight into the role of the effective mass @xmath52 .
the main advantage is however , that all physical properties can be studied at the same density .
this property relates to experiment , because when an external potential is turned on , the effective mass is altered , but the average density will remain the same .
the effect of changing @xmath52 on the bcs - bec crossover phase diagram is shown in figure [ mz_ok.eps ] .
this figure shows the fflo phase boundaries of the imbalanced fermi gas in 3d for different values of the effective mass @xmath52 , before rescaling according to ( [ herschaling ] ) ( and hence at different densities ) .
[ h ] mz_ok.eps there is a tilting of the fflo - region about a fixed point at unitarity .
furthermore , there is an increase in the width of the fflo - region ( relative to the abscissa ) as the effective mass increases . by using the scaling - relation ( [ herschaling ] ) , the phase diagrams in figure [ mz_ok.eps ] can be rescaled to equal density .
after rescaling , the phase diagrams for the different effective masses maps onto the phase diagram of the imbalanced fermi gas with isotropic effective mass ( @xmath58 ) , described in section [ the free fermi gas ] . from this
we conclude that the fflo - state is not fundamentally influenced by an optical potential in which the fermi energy lies near the bottom of the first bloch - band .
this can be explained by the fact that no fundamental anisotropy is introduced into the system by altering the effective mass , because independent of the effective mass , the system can be scaled back to the case of the imbalanced fermi gas where the effective mass equals 1/2 . in section [ anisotropic masses ]
it was shown that a more fundamental anisotropy is needed , in order for the optical potential to have an effect on the fflo - state . in this section ,
we model a 1d optical potential in the tight - binding approximation , using the first bloch - band . for this purpose
, the quadratic dispersion in the @xmath50-direction is replaced by a periodic dispersion @xcite @xmath60 ~. \label{bloch - dispersie}\ ] ] here @xmath61 is the wavevector of the optical potential and @xmath62 is a prefactor with units of energy , given by @xcite@xmath63 with @xmath64 the depth of the potential and @xmath65 the recoil energy given by @xmath66 , with @xmath67 the wavelength of the optical potential and @xmath68 the mass of the fermionic particles . in the limit for small @xmath69 expression ( [ bloch - dispersie ] ) simplifies to ( [ dispersie anisotroop ] ) , with@xmath70 given the new dispersion ( [ bloch - dispersie ] ) , the thermodynamic potential for this system can be calculated . the result is@xmath71 -\xi_{\mathbf{k}}-\frac{\delta^{2}}{2\left \ { k^{2}+\delta \left [ 1-\cos \left ( \frac{\pi k_{z}}{q_{l}}\right ) \right ] \right \ } } \right ) -\frac{\delta^{2}}{8\pi \left ( k_{f}~a_{s}\right ) } \label{thermodynamic potential bloch dispersion}\ ] ] with the following notations@xmath32{l}\xi_{\mathbf{q},\mathbf{k}}=k^{2}+\delta \left [ 1-\cos \left ( \frac{\pi}{2}\frac{q}{q_{l}}\right ) \cos \left ( \frac{\pi k_{z}}{q_{l}}\right ) \right ] -\mu
\\ e_{\mathbf{k}}=\sqrt{\left \
{ k^{2}+\delta \left [ 1-\cos \left ( \frac{\pi}{2}\frac{q}{q_{l}}\right ) \cos \left ( \frac{\pi k_{z}}{q_{l}}\right ) \right ] -\mu
\right \ } ^{2}+\delta^{2}}\\ \zeta_{\mathbf{q , k}}=\zeta-\delta \sin \left ( \frac{\pi}{2}\frac{q}{q_{l}}\right ) \sin \left ( \frac{\pi k_{z}}{q_{l}}\right ) \end{array } \right.\ ] ] it can easily be shown that expression ( [ thermodynamic potential bloch dispersion ] ) is equal to the corresponding thermodynamic potential ( [ thermodynamic potential anisotropic masses ] ) of the anisotropic effective mass case , in the limit @xmath72 , @xmath73 with @xmath52 held constant , according to ( [ relatie mx ql delta ] ) .
the two number equations again are given by ( [ first number equation ] ) and ( [ second number equation ] ) and the density @xmath35 can be calculated using the general expression@xmath74 \right \ } \right)\ ] ] which yields@xmath75{c}\dfrac{q_{l}}{2\pi^{2}}\left ( 1-\delta \right ) ~\left ( 1\geq2\delta \right ) \\ \dfrac{q_{l}}{2\pi^{3}}\left [ \left ( 1-\delta \right ) \arccos \left ( \dfrac{\delta-1}{\delta}\right ) + \delta \sqrt{1-\left ( \dfrac{\delta -1}{\delta}\right ) ^{2}}\right ] ~\left ( 1<2\delta \right ) \end{array } \right .
~. \label{n}\ ] ] here it must be noted that our derivation is only exact if @xmath76 , because otherwise more than one bloch band has to be considered . as in the previous sections ,
the phase diagram for this system can be constructed by studying the local minima of the free energy .
figure [ fd_mettekst.eps ] shows a comparison between the phase diagram of an imbalanced fermi gas in 3d and the phase diagram of an imbalanced fermi gas in 3d subject to a 1d optical potential , modeled in the tight - binding approximation using the first bloch band .
[ h ] fd_mettekst.eps this figure shows that the fflo - region is enlarged by a factor 3 to 6 , due to the stabilizing effect of the 1d optical potential .
figure [ fd_mettekst.eps ] further shows that at @xmath77 a transition point occurs , where the fflo - region reaches a maximum width ( relative to the abscissa ) , and narrows quickly for larger values of @xmath36 .
this effect finds its origin in the magnitude of the wavevector of the fflo - pairs @xmath78 .
when the imbalance chemical potential @xmath36 increases , @xmath78 increases likewise to accommodate for the widening gap between the fermi surfaces of the two spin - species . at a certain level of imbalance ( in the case of figure [ fd_mettekst.eps ] at @xmath77 ) @xmath78 equals the wavevector of the optical potential @xmath61 .
at this point , the fflo - state is optimally enhanced , because the spatial modulation of the fflo - state is equal to the spatial modulation of the optical potential .
this results in a maximal width of the fflo - region .
when @xmath36 increases further , @xmath78 retains the constant value @xmath61 , and is not able to grow any further .
this effect is shown in figure [ deltas_ql12.eps ] .
[ h ] deltas_ql12.eps hence , we can conclude that , although the imbalance has increased further , the optical potential forces the fflo - state into a state where the form of the fflo - order parameter matches the form of the optical potential .
this results in a narrowing of the fflo - region , because the wavevector of the fflo - state is not sufficiently large anymore to bridge the gap between the fermi - surfaces of the spin - up and spin - down particles .
the value of the transition point , where @xmath78 becomes equal to @xmath61 , rougly increases linearly with the value of @xmath62 , as shown in figure [ delta050607.eps ] .
[ h ] delta050607.eps qualitatively , this is because the rate of change of the fflo - wavevector @xmath78 with increasing imbalance chemical potential @xmath36 , decreases when @xmath62 becomes larger .
this means that @xmath36 has to be larger ( compared to the case of lower @xmath62 ) for @xmath78 to reach the limiting wavevector of the optical potential @xmath61 .
the advantage of this resonant enhancement of the fflo - state is that , for a given level of imbalance , an optimal stability region for the fflo - state can be created , simply by tuning the wavelength of the 1d optical potential .
it should be noted that , when considering an imbalance chemical potential @xmath36 smaller than @xmath79 , @xmath62 has to become smaller than @xmath80 and more bands have to be taken into account in order for our description to be exact . we do not treat this case in the present paper . to obtain a more direct link with experimental parameter values , we convert our units back to si units for the three situations depicted in figure [ delta050607.eps ] . a possible choice of atoms which we considered is @xmath81 atoms in a one - dimensional harmonic trap , with a density of @xmath82 . for instance
, the case with @xmath83 and @xmath84 corresponds to the case of an optical potential with wavelength equal to @xmath85 , a recoil energy of @xmath86 and an optical potential depth of @xmath87 .
the numerical values for these experimental parameters in the case of @xmath88 and of @xmath89 are depicted in figure [ delta050607.eps ] panels ( a ) and ( c ) respectively .
for the illustrative cases of figure [ delta050607.eps ] we use @xmath84 , but theoretically , any choice of @xmath61 was possible because we found that the wavevector of the optical potential @xmath61 acts as a scaling parameter , according to the following scaling relation:@xmath90 in principle this means that we can vary @xmath61 from zero to infinity . however , there exist some limitations on this parameter .
first there is a lower limit for @xmath61 because below a certain value of @xmath61 no value of the optical potential depth @xmath64 can satisfy equation ( [ tight binding delta ] ) , given values for @xmath62 and for the recoil energy @xmath65 .
second , when the depth of the optical potential becomes too large , particles will be confined in the direction of the optical potential , thus inhibiting the formation of fflo - states .
this sets an upper limit for the ratio of @xmath91 and subsequently for the value of @xmath61 . during the course of our work , loh and trivedi published their results on the lo - state in a 3d cubic lattice @xcite .
they found that in a 3d cubic lattice , the lo - state was more stable than the ff - state .
since we already find a substantial increase in the ff - state using a 1d optical potential , we expect that the effect on the lo - state will be similar or larger . it would be interesting to apply our 1d - potential scheme also to the lo - case .
we have described the fflo - state in an imbalanced fermi gas in 3d within the path - integral framework , by choosing a suitable saddle - point at which the atomic pairs have a finite centre - of - mass momentum . as a platform to address the case of a 3d imbalanced fermi gas in a 1d optical potential and to validate our path - integral description , we rederived the zero - temperature phase diagram for an imbalanced fermi gas in 3d . for this case ,
our results coincide with recent theoretical results . as a proposal to stabilize the fflo - state we have studied an imbalanced 3d fermi gas in a 1d optical potential .
this potential was modeled in two different ways .
for the first model , where we considered anisotropic effective masses , we have found that this model is a rescaling of the case of a 3d imbalanced fermi gas with isotropic effective mass . in the second model
, we described the effect of the 1d optical potential using the first bloch band in the tight - binding approximation . in this case
we have found a substantial increase in the stability region of the fflo - state , as compared to the case of the 3d fermi gas without the 1d optical potential .
related results were recently found in the case of a 3d cubic optical lattice @xcite . the advantage of our 1d optical potential scheme , compared to a 3d cubic optical lattice , is that it allows to find an optimal stability configuration for the fflo - state with a given level of imbalance , by tuning the wavelength of the optical potential .
this resonant enhancement of the fflo - region occurs when the wavevector of the fflo - pairs is equal to the wavevector of the optical potential .
this tunability makes a 1d optical potential a suitable experimental configuration for the stabilization of the fflo - state .
we therefore propose that this concept can facilitate the experimental observation of the fflo - state in an imbalanced fermi gas in 3d .
t. mizushima , k. machida , and m. ichioka , phys .
lett . * 94 * , 060404 ( 2005 ) ; d. e. sheehy and l. radzihovsky , ibid . * 96 * , 060401 ( 2006 ) ; j. kinnunen , l. m. jensen , and p. trm , ibid .
* 96 * , 110403 ( 2006 ) ; k. machida , t. mizushima , and m. ichioka , ibid .
* 97 * , 120407 ( 2006 ) ; p. castorina , m. grasso , m. oertel , m. urban , and d. zappal , phys . rev .
a * 72 * , 025601 ( 2005 ) ; n. yoshida and s .- k . yip , ibid . *
75 * , 063601 ( 2007 ) ; w. zhang and l .- m .
duan , ibid . *
76 * , 042710 ( 2007 ) ; t. k. koponen , t. paananen , j .-
martikainen , m. r. bakhtiari , and p. trm , new j. phys .
* 10 * , 045014 ( 2008 ) .
k. b. gubbels , m. w. j. romans , and h. t. c. stoof , phys .
lett . * 97 * , 210402 ( 2006 ) ; t .-
dao , m. ferrero , a. georges , m. capone , and o. parcollet , ibid .
* 101 * , 236405 ( 2008 ) ; c .- c .
chien , q. chen , y. he , and k. levin , ibid .
* 97 * , 090402 ( 2006 ) ; * 98 * , 110404 ( 2007 ) . | we describe an imbalanced superfluid fermi gas in three dimensions within the path - integral framework . to allow for the formation of the fulde - ferell - larkin - ovchinnikov - state ( fflo - state ) , a suitable form of the saddle - point
is chosen , in which the pairs have a finite centre - of - mass momentum . to test the correctness of this path - integral description
, the zero - temperature phase diagram for an imbalanced fermi gas in three dimensions is calculated , and compared to recent theoretical results .
subsequently , we investigate two models that describe the effect of imposing a one - dimensional optical potential on the 3d imbalanced fermi gas .
we show that this 1d optical potential can greatly enlarge the stability region of the fflo - state , relative to the case of the 3d fermi gas without 1d periodic modulation .
furthermore it is show that there exists a direct connection between the centre - of - mass momentum of the fflo - pairs and the wavevector of the optical potential .
we propose that this concept can be used experimentally to resonantly enhance the stability region of the fflo - state . |
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.